Risk and Morality 9781442679382

Table of contents :
Contents
Acknowledgments
1. Risk and Morality
PART ONE. Theories of Risk and Morality
Introduction
2. Risk and Dirt
3. The Rise of Risk
4. Risk and Morality: Three Framing Devices
PART TWO. New Moralities of Risk and Responsibility
Introduction
5. New Moralities of Risk and Political Responsibility
6. Risk Management and the Responsible Organization
7. Risk and Moralization in Everyday Life
8. From Risk to Precaution: The Rationalities of Personal Crime Prevention
PART THREE. The Moral Risks of Insurance
Introduction
9. Moral Uncertainties: Contract Law and Distinctions between Speculation, Gambling, and Insurance
10. Containing the Promise of Insurance: Adverse Selection and Risk Classification
11. Insurers as Moral Actors
12. The Moral Risks of Private Justice: The Case of Insurance Fraud
PART FOUR. Governing Risky Desires
Introduction
13. Risking Rescue: High Altitude Rescue as Moral Risk and Moral Opportunity
14. The Neurochemical Self and Its Anomalies
15. Targeted Governance and the Problem of Desire
Contributors

Citation preview

RISK AND MORALITY Edited by Richard V. Ericson and Aaron Doyle

Risk and Morality examines how decisions about risk and uncertainty relate to moral principles and ethical conduct. In this volume editors Richard Ericson and Aaron Doyle have brought together a selection of original essays on the topic by renowned scholars in the disciplines of philosophy, sociology, law, political science, geography, criminology, and accounting from Canada, the United States, England, France, and Australia. Presenting cutting-edge theory and research, the essays analyse the broader social, political, economic, and cultural dimensions of risk and morality. The concept of risk has become pervasive in recent years in political discourse, popular culture, organizational communications, and everyday life. The contributors' respective research projects on risk and morality in politics, business, legal regulation, crime prevention, insurance, extreme sports, and biotechnology provide original empirical evidence to substantiate their theories and address the ideological and policy relevance of their work. Collectively, the contributors explain why risk is such a key aspect of Western culture, and demonstrate that new regimes for risk management are transforming social integration, value-based reasoning, and morality. Further, they illustrate that these new regimes do not necessarily foster more responsible conduct or greater accountability in institutions. RICHARD ERICSON is Principal of Green College and Professor of Law and Sociology at the University of British Columbia. AARON DOYLE is Assistant Professor of Sociology at Carleton University.

This page intentionally left blank

Green College Thematic Lecture Series

Risk and Morality EDITED BY RICHARD V. ERICSON AND AARON DOYLE

UNIVERSITY OF TORONTO PRESS Toronto Buffalo London

www.utppublishing.com University of Toronto Press Incorporated 2003 Toronto Buffalo London Printed in Canada ISBN 0-8020-8760-4 (cloth) ISBN 0-8020-8563-6 (paper)

Printed on acid-free paper

National Library of Canada Cataloguing in Publication Risk and morality / edited by Richard V. Ericson and Aaron Doyle. (Green College thematic lecture series) Includes bibliographical references. ISBN 0-8020-8760-4 (bound). ISBN 0-8020-8563-6 (pbk.) 1. Social ethics. 2. Risk - Moral and ethical aspects. 3. Risk Sociological aspects. I. Ericson, Richard V., 1948- II. Doyle, Aaron III. Series. HM1101.R55 2003

303.3'72

C2003-900680-8

This book has been published with the help of a grant from Green College, University of British Columbia. The University of Toronto Press acknowledges the financial assistance to its publishing program of the Canada Council for the Arts and the Ontario Arts Council. University of Toronto Press acknowledges the financial support for its publishing activities of the Government of Canada through the Book Publishing Industry Development Program (BPIDP).

Contents

Acknowledgments

vii

1 Risk and Morality RICHARD V. ERICSON AND AARON DOYLE

Part One: Theorizing Risk and Morality Introduction

1

11

13

2 Risk and Dirt IAN HACKING

22

3 The Rise of Risk DAVID GARLAND

48

4 Risk and Morality: Three Framing Devices JOHN ADAMS

87

Part Two: New Moralities of Risk and Responsibility Introduction

105

107

5 New Moralities of Risk and Political Responsibility VIOLAINE ROUSSEL

117

6 Risk Management and the Responsible Organization MICHAEL POWER

145

vi Contents 7 Risk and Moralization in Everyday Life ALAN HUNT

165

8 From Risk to Precaution: The Rationalities of Personal Crime Prevention KEVIN D. HAGGERTY

193

Part Three: The Moral Risks of Insurance Introduction

215

217

9 Moral Uncertainties: Contract Law and Distinctions between Speculation, Gambling, and Insurance PAT O'MALLEY 231 10 Containing the Promise of Insurance: Adverse Selection and Risk Classification TOM BAKER

25h8

11 Insurers as Moral Actors CAROL A. HEIMER

284

12 The Moral Risks of Private Justice: The Case of Insurance Fraud RICHARD V. ERICSON AND AARON DOYLE

Part Four: Governing Risky Desires Introduction

317

365

367

13 Risking Rescue: High Altitude Rescue as Moral Risk and Moral Opportunity JONATHAN SIMON

375

14 The Neurochemical Self and Its Anomalies NIKOLAS ROSE

407

15 Targeted Governance and the Problem of Desire MARIANA VALVERDE

Contributors 459

438

Acknowledgments

This book grew out of a conference entitled 'Risk and Morality' that was held at Green College, University of British Columbia, in May 2001. We are most grateful to both Green College and the Social Sciences and Humanities Research Council of Canada conference grant program for their generous financial support of the conference and publication. Richard Ericson also thanks The Canada Council for the Arts, which awarded him a Killam Research Fellowship to pursue a research program of which this book is an important part. We would also like to thank Carolyn Andersson, Dene Matilda, and Diana Ericson of Green College for being so helpful in organizing the conference and assisting with the preparation of this book, and Allyson May for her skilful copyediting. As always, Virgil Duff of the University of Toronto Press has been a most supportive and helpful sponsoring editor. RICHARD V. ERICSON AND AARON DOYLE

This page intentionally left blank

1 Risk and Morality RICHARD V. ERICSON AND AARON DOYLE

This book examines the ways in which risk discourse and practice relate to moral principles and ethical conduct. It brings together scholars in the disciplines of philosophy, sociology, law, political science, geography, criminology, and accounting from Canada, the United States, the United Kingdom, France, and Australia. In the papers they have written especially for this volume, these scholars advance the most recent theories of how modern societies are governed through risk and morality. Their respective research projects on risk and morality in politics, business enterprise, legal regulation, crime prevention, insurance, extreme sports, and biotechnology provide original empirical evidence to substantiate their theories and address the ideological and policy relevance of their work. In fostering transdisciplinary and transnational conversations about risk and morality, this collection is eclectic. Rather than working within a single paradigm, each contributor offers a chapter on her or his current research and thinking. We have not imposed any given framework — for example, a conventional sociological analysis of gender, race, and/or class - on the book as a whole. Instead, each author has chosen her or his own framework for addressing the relation between risk and morality, including considerations of stratification when relevant to their specific theoretical and empirical concerns. Collectively, the contributors to this book present up-to-date, cuttingedge theory and research on risk society. They analyse the broader social, political, economic, and cultural dimensions of risk and morality. They explain why risk is a key aspect of Western cultures. They demonstrate how discourses and practices of risk are always imbued with moral language and ethical clauses. In grounding their analyses in specific

2 Richard V. Ericson and Aaron Doyle institutions - politics, business enterprise, law, insurance, and medicine - they demonstrate that new regimes for risk management are transforming social integration, value-based reasoning, and morality. They also reveal that these new regimes do not always foster more responsible conduct or greater accountability in institutions. This book is organized according to four broad themes: theorizing risk and morality (Part I), new moralities of risk and responsibility (Part II), the moral risks of insurance (Part III), and governing risky desires (Part FV). A summary of the contributions each chapter makes to these organizing themes, and of the relation among them, is provided in the introduction to each part. The concept of risk has become pervasive in political discourse, popular culture, organizational communications, and everyday practices, and risk has consequently become central to debates across a broad range of academic disciplines, whereas it hardly appeared at all even twenty years ago. Although in this volume leading scholars treat risk as a problematic concept, the basic idea is simple: risk refers to threats or dangers attributed to persons, technologies, or nature. It also refers to the chance or probability that these threats or dangers will result in adverse consequences for a specified party. Assessment of the chance of adverse consequences depends on risk communication systems. These systems include rules, formats, and technologies for risk classification, selection, and response. It is through these systems that risks are given meaning and acted upon. As the analyses in subsequent chapters reveal, communication systems are not simply conduits through which knowledge of risk is transferred. Risk is called into being, made visible, and responded to through the rules, formats, and technologies available in the communication system. The communication system makes risks real. Its logics and autonomous processes govern social relations and circumscribe what people and their organizations can do in both taking chances and taming them. Assessment of the chance of adverse consequences also depends on morality. Indentification of a threat or danger, and of adverse consequences, is based on judgments about 'goodness' and 'badness' and distinctions between right and wrong. Moreover, the risk communication system itself- its rules, formats, and technologies - is constructed in terms of moral principles and designed to produce moral conduct. The risk communication system is also a system of morality. It produces systematic knowledge for the evaluation of risky conduct. It produces

Risk and Morality 3 knowledge of risk as a capacity to act in principled ways so as not to cause adverse consequences to specified parties. In addressing the myriad ways in which risk and morality are entwined, contributors to this volume reject the view that the concept of risk and practices of risk management are increasingly divorced from morality. Some risk society theorists, such as Beck (1992), argue that the sheer magnitude of modern risks with catastrophic potential precludes the isolation of responsible parties, and therefore effective attributions of moral accountability. Others (e.g., Giddens 1990, 1991) argue that the ascendancy of abstract risk communication systems spells the end of broader moral orders. These systems force people to focus instrumentally on specific knowledge of risk in their immediate environment to routinize their activities and feel secure. Giddens (1991: 145) even contends that such abstract systems of risk entail the 'evaporation of morality' and asserts that 'moral principles run counter to the concept of risk and to the mobilizing dynamics of control. Morality is extrinsic so far as the colonizing of the future is concerned.' Consistent with Giddens, many analysts view risk management systems as amoral, or grounded only in a utilitarian morality. As moral responsibility for material and social well-being is transferred from the state and other broad collectivities to local organizations, communities, and individuals, morality vaporizes out of responsibility. Responsibility becomes merely an instrumental articulation of risk management practices: whatever will simultaneously allow profitable forms of risk taking, minimize harms, and distribute negative consequences. Risk management is reduced to the scientific identification and verification of objects called 'risk,' and to the utilitarian obligation of those deemed responsible to do something about such risk. For example, some criminologists decry a decline in institutions of moral order, such as welfare and the criminal law, and the ascendancy of risk communication systems based on crime management and private justice. Garland (1997: 182; see also Garland 2001) observes that these systems are 'forward-looking, predictive, oriented to aggregate entities and concerned with the minimization of harms and costs, rather than with the attribution of blame or dispensation of individual justice.' Shearing and Stenning (1983, 1984) view private security systems strictly from the instrumental eyes of the client, rather than seeing them as a response to externally defined moral standards. They admit that moral discourse does arise in relation to private security, but this they read as

4 Richard V. Ericson and Aaron Doyle mere 'moral rhetoric ... a control strategy' (Shearing and Stenning 1984: 340). Here private security is simply a product of instrumental consumption that protects other practices of consumption. It is said to be consistent with all other consumption, that is, founded on the 'economic amoralism' (Slater 1997: 28) of cost-benefit analysis, risk calculus, and rational choice. The risk management systems of insurance have been analysed as another case in point. Lowi (1990: 32) believes that in the social insurance systems of the welfare state, 'the very concept of risk had already taken the morality out of responsibility and had made the consequences of risk - injury - a merely instrumental matter.' The same has been said about private insurance systems. Simon (1987: 62; 1992: 44) and Reichman (1986: 151-2) believe that insurance systems disempower the individual as a moral and political subject and create instrumental communities of interest rather than communities based on collective conscience and moral solidarity. Loss prevention systems governed in part through private insurance are viewed in the same way. For example, Simon (1987: 75) asserts that the motor vehicle transportation system represents 'a field of conduct to which little moral judgment attaches. It has become demoralized ... because our practices for handling the negative consequences of driving don't rely on moral judgments about the practices involved.' A related illustration is the tort liability system in civil law, which has shifted responsibility to the party in the best position to prevent injuries or, failing that, to spread the costs of them. Less regard is paid to the moral intentions of litigants, and more emphasis is placed on insurance as the primary institution of prevention, risk distribution, and indemnification. This use of the tort liability system for risk management hardly results in 'the evaporation of morality' (Giddens 1991: 145), but 'the moral foundation of the new regime is relentlessly utilitarian' (Priest 1990:215). While the rise of risk discourse and risk management practices has undoubtedly altered morality, especially in utilitarian directions, it does not spell the end of morality. To the contrary, the rise of risk entails an attendant rise in new moralities of responsibility and accountability at multiple levels of society: institutions, organizations, communities, and individuals. The work of Mary Douglas (1990, 1992) is especially important in establishing that risk is part of a political vocabulary of moral responsibility and accountability. The moral dimension of risk arises in particular from the fact that 'risk is not only the probability of an event but also the

Risk and Morality 5 probable magnitude of its outcome, and everything depends on the value that is set on the outcome. The evaluation is a political, aesthetic and moral matter' (Douglas 1990: 10). Thus risk is connected to moral assessments of threat or danger. It is used in political culture to mobilize moral communities for dealing with danger in particular ways, and to force accountability. The word 'danger' sufficed in the past, but now the scientism of risk analysis has greater rhetorical effect. The word 'sin' also sufficed in the past to translate dangers into moral and political issues, but sin like danger is becoming obsolete as a mobilizer of moral community compared to 'the modern, sanitized discourse of risk ... A neutral vocabulary of risk is all we have for making a bridge between the known facts of existence and the construction of a moral community ... risk, danger and sin are used around the world to legitimize policy or to discredit it, to protect individuals from predatory institutions or to protect institutions from predatory individuals' (ibid.: 4-5). Douglas observes that risk as a rhetoric of moral responsibility is especially strong in the individualistic culture of contemporary liberal regimes. Risk is invoked as a defensive mechanism to protect individuals from encroachment by others. To be at risk is to be sinned against, vulnerable to the threats of imposing institutions, communities, or individuals. The sinning individual is singled out as a suitable enemy to uphold the moral solidarity of the community (Durkheim 1964), but the sinning community is now used to uphold the moral autonomy of the individual as well. Risk codes danger as a threat to liberty. In this view risk is simultaneously a discourse of responsibility, accountability, justice, and retribution. Uncertainty is not entertained in public discourse because to argue uncertainty is to eschew responsibility and accountability (Douglas 1990: 9). Instead, risk is deployed as a discursive tactic to identify the danger of future damage caused by political opponents. Each side mobilizes experts in scientific risk analysis to articulate the cause of danger. Causal responsibility is then linked to political responsibility. Political responsibility is assessed in relation to the response ability of those charged with the moral duty to do something about the danger. In this process risk also becomes a discourse of justice, of who has both the capacity and obligation to protect individuals from danger. Of course, this tactical use of risk can be abused, enhancing liberty for some at the expense of others. In particular, it can offer 'automatic, self-validating legitimacy to established law and order ... Those who fear the taboos see dangers and their connection with morality as part of how the world works' (Douglas 1990: 8).

6 Richard V. Ericson and Aaron Doyle Morality is also embedded in the minutiae of risk communication systems. These systems constitute risk objects as moral objects in the very process of making classifications and probability assessments. Statistics of risk are produced on the basis of moral classifications (Douglas 1986, 1992). Risk and morality are also entwined at the level of probability calculations and risk management practices. Statistical knowledge of risk provides a pragmatic discourse of moral normalcy. With the advent of probability statistics, the word 'normal' came to connote objectivity about human beings. Probability statistics describe what is normal about a population, and who within it is normal and who poses a risk. However, the word 'normal' functions simultaneously to evaluate what ought to be. 'It uses a power as old as Aristotle to bridge the fact/value distinction, whispering in your ear that what is normal is also all right ... The norm may be what is usual or typical, yet our most powerful ethical constraints are also called norms' (Hacking 1990: 160, 163). Quantitative risk assessments do not offer a unique rationality that pinpoints a single right course of action, but rather probabilities that require moral assessment for action (Adams 1995: ix). Moreover, efforts to commodify risks, for example, by state regulatory agencies or insurers, also entail moral preferences and reveal relative values (ibid.: 96). Insurance is a moral technology of responsibility that quantifies and commodifies commitments in every aspect of underwriting, preventive security, and indemnification (Ewald 1991: 206-7; Baker and Simon 2002; Ericson, Doyle, and Barry 2003). Risk communication systems effect moral regulation by making people think of risk objects in terms of their own ethical conduct with respect to these objects. They must become knowledgeable about risks and do their part to prevent, minimize, and distribute them. No one can escape being subject to risk discourse and risk management practices because all actions are under a description of risk (Hacking 1986). Everyone takes risks (Adams 1995: 16), including the risk of inaction (Giddens 1990: 32). Knowledge of risk - of probability and normalcy-is used in action, and the action taken feeds back into the risk management system. Furthermore, because risk cannot be separated from morality, all actions and risk assessments of them are simultaneously under a moral description. This is what Hacking refers to as the 'looping effects' of knowledge of risk. 'Such social and personal laws were to be a matter of probabilities, of chances. Statistical in nature, these laws were nonetheless inexorable; they could even be self-regulating. People are normal if they conform to the central tendency of such laws, while those

Risk and Morality 7 at the extremes are pathological. Few of us fancy being pathological, so "most of us" try to make ourselves normal, which in turn affects what is normal' (Hacking 1990: 2). Risk communication systems affect the interplay of normalization, restraint, security, and risk taking. This interplay occurs simultaneously on the level of risk as a discursive strategy of blame (Douglas 1990, 1992), and on the level of risk as routine probability assessments (Hacking 1990): The ethics of lifestyle maximization, coupled with a logic in which someone must be held to blame for any event that threatens an individual's 'quality of life,' generates a relentless imperative of risk management not simply in relation to contracting for insurance, but also through daily lifestyle management, choices of where to live and shop, what to eat and drink, stress management, exercise and so forth. Of course, this inaugurates a virtually endless spiral of amplification of risk ... these arrangements within which the individual is re-responsibilized for the management of his or her own risk produces a field characterized by uncertainty, plurality and anxiety, thus continually open to the construction of new problems and the marketing of new solutions. (Rose 1996: 343, 346)

In order to sustain the 'logic in which someone must be held to blame for any event that threatens an individual's quality of life,' risk must be construed as a product of human agency and therefore controllable through attributions of responsibility and processes of accountability. Institutions such as science, law, medicine, and insurance participate in the search for human causes of risky activities. Singling out a cause renders risky activity subject to attributions of moral responsibility for indemnification after the fact. It also makes risky activity predictable and therefore subject to attributions of moral responsibility for prevention before the fact. This process is exemplified in the shifting moral language of the accidental (Green 1997). The word 'accident' is a moral term, denoting that the event in question was not motivated by bad intentions. The event is instead deemed to have arbitrary causation, to be unique and unpredictable. Thus no one can be held responsible for it. Indeed, those involved are victims who deserve sympathy and compensation. However, with the rise of risk discourse, 'accident' is being replaced by other terms that allow the attribution of responsibility. For example, vehicle insurers substitute 'collision' or 'crash.' This new moral language is accompanied

8 Richard V. Ericson and Aaron Doyle by claims that over 90 per cent of all 'crashes' result from individual driver actions. Road and vehicle conditions are dismissed, as is the view that accidents 'just happen' (Insurance Corporation of British Columbia 1996; Ericson, Doyle, and Barry 2003). Gusfield (1981) points to this trend in his brilliant analysis of how impaired drivers are represented in American political culture. He shows that the cultural organization of vehicle accidents is based on the belief that they result from individual driver performance. This view is underpinned not only by the strongly individualistic character of American culture but also by scientific research and legal discourse that individualize social problems in order to attribute responsibility and justify intervention. 'Multicausality weakens the capacity and purposefulness which make control seem possible' (ibid.: 74). Thus, when alcohol is involved in an accident it becomes the cause of the accident, regardless of other factors. The attribution of a primary cause is crucial to pinpointing responsibility and taking punitive or other remedial action. One major consequence of this framework is blaming the victim. Someone who might otherwise have been seen as the victim of an accident is now seen as a contributor to the crash because her driver performance was faulty. The driver failed to take adequate preventive measures by learning and acting upon knowledge of risk. 'As responsibility is divorced from motivation, all victims of accidents are potentially culpable. Culpability arises not from the motivations of the victim or other agents, but from their miscalculation or ignorance of risks' (Green 1997: 199). The logic of seeing accidents as a product of individual failure is extended to other areas of risk management. For example, the victims' movement has in some instances ironically blamed victims of crime and succeeded in responsibilizing potential victims to reduce their exposure to risk. Potential victims are treated as suspects: they are suspected of not doing enough to prevent criminal victimization in the first place. Actual victims are treated as offenders: they are responsible for the security breach that allowed criminal victimization to occur. The moral meaning of misfortune is also a part of risk taking in everyday life. Risk taking is crucial to the process by which people give moral significance to their lives. It confirms moral autonomy, liberty, and individuality. Thus efforts to make people more responsible through risk management strategies are always subject to the looping effect of risk compensation: an individual may take more risks because of the new level of security provided, in part because the assertion and confirma-

Risk and Morality 9 tion of individual autonomy is a major reward for risk taking in individualistic cultures. The contributors to this volume analyse the relationship between questions of risk discourse and practice and moral language and ethical conduct. In defining risk as danger, in mobilizing risk as a discursive tactic, and in practices for risk management, moral issues are omnipresent. References Adams, J. 1995. Risk. London: UCL Press Baker, T., and J. Simon, eds. 2002. Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Douglas, M. 1986. How Institutions Think. Syracuse: Syracuse University Press - 1990. 'Risk as a Forensic Resource.' Daedalus 119 (14): 1-16 - 1992. Risk and Blame. London: Routledge Durkheim, E. 1964. The Rules of Sociological Method. New York: Free Press Ericson, R.V., A. Doyle, and D. Barry. 2003. Insurance as Governance. Toronto: University of Toronto Press Ewald, F. 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality, 197-210. Chicago: University of Chicago Press Garland, D. 1997. 'Of Crimes and Criminals: The Development of Criminology in Britain.' In M. Maguire et al., eds., The Oxford Handbook of Criminology, 11-56. Oxford: Oxford University Press - 2001. The Culture of Control: Crime and Social Order in Contemporary Society. Oxford: Oxford University Press Giddens, A. 1990. The Consequences of Modernity. Cambridge: Polity - 1991. Modernity and Self-Identity: Self and Society in the Late Modern Age. Stanford: Stanford University Press Green, J. 1997. Risk and Misfortune: The Social Construction of Accidents. London: UCL Press Gusfield, J. 1981. The Culture of Public Problems: Drinking-Driving and the Symbolic Order. Chicago: University of Chicago Press Hacking, I. 1986. 'Making up People.' In T. Heller et al., eds., Reconstructing Individualism, 222-36. Stanford: Stanford University Press - 1990. The Taming of Chance. Cambridge: Cambridge University Press Insurance Corporation of British Columbia. 1996. Motor Vehicle Insurance in

10 Richard V. Ericson and Aaron Doyle British Columbia at the Crossroads: Volume 1: The Case for Change. Report prepared by KPMG, Exactor Insurance Services Inc., Eckler Partners Ltd. (19 December) Lowi, T. 1990. 'Risks and Rights in the History of American Governments.' Daedalus 119 (4): 17-40 Priest, G. 1990. The New Legal Structure of Risk Control.' Daedalus 119(4): 207-27 Reichman, N. 1986. 'Managing Crime Risks: Towards an Insurance-Based Model of Social Control.' In S. Spitzer and A. Scull, eds., Research in Law, Deviance and Social Control, 151-72. Greenwich, CT: JAI Press Rose, N. 1996. 'The Death of the Social: Re-figuring the Territory of Government.' Economy and Society 25: 327-56 Shearing, C., and P. Stenning. 1983. 'Private Security: Implications for Social Control.' Social Problems 30: 493-506 - 1984. 'From the Panopticon to Disney World: The Development of Discipline.' In A. Doob and E. Greenspan, eds., Perspectives in Criminal Law, 33548. Aurora, ON: Canada Law Book Simon, J. 1987. The Emergence of Risk Society: Insurance, Law and the State.' Socialist Review 95: 61-89 - 1992. 'In Place of the Parent: Risk Management and the New Social Control on Campus.' Paper to the Annual Meeting of the Law and Society Association, Philadelphia, 31 May Slater, D. 1997. Consumer Culture and Modernity. Cambridge: Polity

PART ONE Theories of Risk and Morality

This page intentionally left blank

Introduction

In Part I, Ian Hacking, David Garland, and John Adams theorize different conceptions of risk and their relation to morality. Ian Hacking analyses how collectivities choose the risks they worry about, and how their choices shape experience. David Garland considers the rise of risk as a dominant concept in modern societies, and how it is used in quite different ways by scholars from different theoretical traditions. John Adams presents three cultural frameworks through which people understand and respond to risks in moral terms. Ian Hacking observes that the word 'risk' originates in European languages of the sixteenth century. From the outset the concept of risk was used to instil prudence, the moral duty to act today with regard to what might happen in the future. In the seventeenth century, this new way of thinking and being was overlain with the idea of calculating the chance that something might occur and assessing the utility of action in relation to that chance. Risk subsequently emerged as a calculating concept that brokers between fear and harm and spans fact and value. It became central to the moral habits through which people consume the future: investing, purchasing insurance, and installing loss prevention technologies in the perpetual effort to minimize and distribute harm. Probability, statistics, insurance, and decision theory have fostered prudence for centuries. Therefore the contention of some, such as Ulrich Beck, that we are only now experiencing a 'new modernity' based on risk is erroneous. Modern societies have always focused on future events and interactions with a concern for danger, and they have always devised risk technologies to yield safety as the normal, desired state. At the same time, Hacking recognizes that there has been substantial social change in the past three decades. This change is perhaps substan-

14 Part One: Theories of Risk and Morality tial enough to justify use of the term 'risk society,' albeit in a way markedly different from Beck's use of it. As David Garland shows in chapter 3, there has been an exponential rise in risk discourse, and practices in the name of risk, over the past three decades. Hacking attributes much of this development to the institutionalization and professionalization of risk analysis, and later of risk management, as expert fields. Risk analysis focuses on fear, in particular why people fear new risks and how 'rational' those fears are in terms of utility and probability. It began to proliferate in the 1970s because of new experiences of collective non-distributable risks, especially those related to environmental pollution. It was fuelled not only by the iatrogenic effects of expertise, but also by social movement activism around collective non-distributable risks. Risk management, in contrast, is located in the corporate world, where it continues the 'calculating good sense of the seventeenth century' (31) of trying to engineer the future in ways that maximize utility. Given the corporate focus of risk management, it typically expresses a moral utilitarianism that is coincident with maximization of profit. Hacking turns to the scholarship of Mary Douglas to understand the contemporary cultural significance of risk discourse. The key question is how the specific risks analysed by a collectivity are chosen. Focusing on risks that are difficult to distribute, he notes that the decision of whether to include such a risk in the portfolio of a collectivity changes markedly over time and place, is not in general determined by expert knowledge about objective or real probabilities and disutilities of the risk, and cannot be determined by a general theory, because risk selection is dependent on too many local contingent facts and stories. A given collectivity is highly selective of the risks it includes in its portfolio of hopes and fears. The risks of literal pollution - unclean air, contaminated water, impure food - are a clear case in point. But metaphorical or figurative pollution, which Hacking terms 'dirt,' is involved in collective efforts to define cultural borders and a sense of place and security within them. For example, the strongest Canadian cultural defence of its border with the United States is universal health care. 'The United States is impure, evil, immoral, dirty, untouchable, because it does not have universal health care. We are pure, righteous' (35) because we do (or, in light of current transformations in our health care system, still like to think that we do). Crime, especially violent crime, is another risk central to the Canadian portfolio for articulating the Canada/United States, purity/pollution border divide.

Introduction

15

Hacking provides many other rich illustrations of how purity and pollution are two sides of the coin of culture, spent on fears in the hope that moral community will be built and strengthened. He emphasizes that the risks featured in a collectivity's portfolio define the centre in political as well as cultural terms. The risk of crime, for example, is not only about actual loss - some of which can be indemnified through insurance - but also entails central values such as the sanctity of property, privacy, and liberty. Reaction to crime thus extends well beyond the 'real' harms done and includes punishment of symbolic pollution to sustain a sense of purity. The punishment of symbolic pollution reaches extreme forms in the 'war on drugs': risks are dramatized as evil dangers and punished accordingly. Underscoring the selective nature of risk portfolios, Hacking notes that handguns are a symbol of purity rather than pollution in the United States. That thousands of people die each year from handgun shootings has no bearing, because 'symbols rule the roost of risk' (39). The right to bear arms is enshrined in the constitution; handguns defend liberty as the political essence of the nation, and they are defenders of privacy (both the home and body). As symbols of purity, origin, and identity, handguns preserve symbolic borders of transgression although the cost is excessive death and injury. Lest Canadians feel self-righteous for viewing handguns as symbols of pollution, not purity, Hacking turns to nuclear energy production. Germany has recently banned the construction of future nuclear power plants and is phasing out its existing plants. But bordering France embraces nuclear energy production as the way for the future, and Canada not only embraces this risk but exports its dangerous CANDU reactors outside its borders. Clients include some unstable regimes that no doubt appreciate the availability of CANDU's weapons-grade tritium byproduct to help manage risks of a different kind. The nuclear energy example illustrates the convergence of real, figurative, metaphorical, and symbolic pollution. The real pollutants are particled with symbols, and the symbolic pollutants have real effects. Terrorism is another case in point. The September 11, 2001, terrorist attacks were assaults on the symbols of capitalism (the World Trade Center) and militarism (the Pentagon). However, the subsequent discourse about terrorism risk focused primarily on the terrorists as pollutants who assaulted purity, in the form of American liberty. In defence of liberty, security operations of both government and private corporate institutions have involved enormous constraints on liberty. The anthrax

16 Part One: Theories of Risk and Morality scare compounded this dynamic. An exemplar of pollution, the anthrax scare transformed the real threat into an assault on purity and thereby embedded it in the common sense of liberty. Hacking reveals that the culture of risk has changed dramatically. Risk portfolios are in flux because of rapidly changing responses to nondistributable risks that bear uncertainties: a water supply is contaminated and kills people; a child is abducted and sexually assaulted; a nuclear energy plant is leaking and will cause intergenerational human defects; terrorists are in our midst and strike. The intense concern is not driven by new knowledge. Rather, it is stirred by fear and other emotions: the feeling of defilement, the pain of unspeakable filth. It is in these terms that each highlighted risk is conceptualized as pollution, given a particular place in a risk portfolio, and experienced as dirt. The topography of risk is now experienced as a kaleidoscope, assaulting the senses at every turn. Risk is surrounded by uncertainty, prudence by extreme precaution, security by insecurity. In chapter 3, David Garland asks why intellectual, political, and popular currents have converged around risk. While risk has long been salient to organizational and management tasks - albeit identified under different terminology - it has now become an essential component of modern culture. Garland explores seven reasons for the rise of risk. First, consistent with the ideas of Mary Douglas, risk as danger is a cultural code for expressing and defending interests. Nothing is inherently risky, but anything can be treated as a risk if a party claims that it has the potential to adversely affect their interests and values. Second, risk expresses the urge to address uncertainties and the fears they generate. As a probability statement, risk calculations can only make uncertain knowledge claims. Yet people require knowledge of risk as a capacity for action, to feel that they are taming chance even as they are taking it. Furthermore, uncertainty often increases as action is taken under descriptions of risk, because risk is interactive: it is conditional and continuous in response to the attitudes and behaviour of the participants in the risk communication system. Uncertainty resides in the values, interests, and intentions of others in the system. As Garland remarks pithily, 'Drive in rush-hour traffic and you'll get the idea' (55). Third, there is a long-term trend in which modern governments have become more responsible for risk management through legal regulation, welfare programs, and compulsory insurance schemes. While welfare provision has been rolled back in recent years, it is still considerable. The state is also involved in private insurance markets through compul-

Introduction 17 sory insurance schemes, as well as regulation of insurance company solvency and market conduct. More broadly, the welfare state has become the risk management state, addressing security in all spheres of life. The risk management state is concerned with population management through risk categorization. It addresses who is most at risk and the security provisions they require regarding health, ageing, crime, relative deprivation, and so on. This focus has given rise to a new politics of identity, in which different risk groupings compete for limited resources and other forms of preferential treatment in risk management. Fourth, risk discourse has become central to liberalism. In particular, risk taking and its management are seen as essential to expressions of freedom, choice, and autonomy. The actual expansion of choice also entails an expansion of risk: of actual danger, as well as risk communication systems to address it. Risk taking - in the financial markets, in the casino, in flying to Las Vegas, in choosing where to stay or what to do increases the probability of adverse consequences. Risk communication systems also proliferate, in an attempt to make people aware of risks so that possible unanticipated consequences can be turned into anticipated ones. Fifth, again consistent with the work of Mary Douglas, Garland argues that risk is integral to culture. Risk discourse and practice are revealing of the people who engage in them, and of their communities, organizations, and institutions. Risks singled out tell people not only what to do but who to be and are therefore central to the politics of group identity and rights. We live in a new culture of risk and security. Risk society is experienced not so much as a juggernaut, as Giddens has called it, as a hurdling society. Risk communication systems keep people focused on possible obstacles and therefore hypersensitive to the myriad risks they run. Risk is ubiquitous in planned action, day-to-day practices, intellectual ruminations, and cultural imaginations, all at once. Sixth, the rise of risk has been accelerated by advances in risk communication systems. Risk can be a ubiquitous aspect of planned action because of the omnipresence of risk communication systems. Advanced technologies are especially important in this regard. For example, probability statistics, computers and other electronic communication media, and auditing formats have together made formal risk management a key institution of modern society. Seventh, risk ebbs and flows with our emotions. In the more cataclysmic vision of Beck, risk society is driven by a negative logic and fearful emotions. A limitless dread of catastrophes can result from the normal

18 Part One: Theories of Risk and Morality accidents of complex scientific and technological systems, and from the immoral uses of science and technology. In the new precautionary environment depicted by Ewald and others, risk society is driven by the deployment of more and more resources to preventive security, which simultaneously play on fear and insecurity even as they try to reduce these emotions. These negative emotions are exacerbated as government and other institutions seek to govern on the basis of local knowledge of risk, and to make discrete organizations, communities, and individuals responsible for specific precautionary measures. In chapter 4, John Adams takes up the social psychology of risk and precaution and its relation to morality. He conceptualizes decisions about risk as moral decisions made in the context of uncertainty. The process of making moral decisions about risk involves three framing devices. First, risk definition and response vary by the media through which risk is perceived. Second, analysis is based on what Adams calls 'the risk thermostat': a calculus of costs and benefits encompassing not only financial considerations but a full range of possible rewards and losses among all parties involved in, or who could be affected by, the risky activity. Third, this calculus necessarily includes 'ethical filters,' assessing the contending moralities of risk and responsibility. The framing of risk depends on the media used to perceive it. Many risks are directly perceptible. Activities such as climbing a tree or driving a car are managed instinctively and intuitively, although of course a learning process has been involved. It is in this learning process that moral questions arise. For example, during early childhood development parents constantly make moral judgments about the imposition of safety and danger in allowing their children to take or avoid risks. Moral quandaries also arise when various institutional authorities, especially governments, act in loco parentis. Contemporary political culture is concerned to a great extent with moral debates about who has the right to intervene to manage risks on behalf of whom. Such debates fuel political conflict because conflicting moral codes are involved. The authorities must make hard choices and establish priorities, deciding to intervene in some areas of safety or self-harm prevention more than others. Adams offers the example of people in the United States being three thousand times more likely to be killed by a gun than by fireworks, yet in most states fireworks are regulated more strictly than guns. Some risks are mediated through science. Science discovers dangers that are invisible to the naked eye of non-specialists. They can only be seen with special instruments and a trained capacity to look and inter-

Introduction

19

pret via these instruments, but they are also invisible in that they are typically based on suspicions and require inferential leaps to convert them into probability statements. Moreover, probabilities are rendered in the form of moral statistics by institutional risk managers charged with specific mandates. The example of driving behaviour and road accident risk used above is a case in point. In this field probabilities are based on limited data from past years; a large number of driver characteristics; and multiple factors involving road, weather, and vehicle conditions. Any full, multiple factor analysis turns the probability calculus into an expression of ignorance. All the authorities can do in this context is select one or a few factors, such as those related to driver performance, and render them in moral terms for practical preventive action. As Adams observes, 'The veneer of scientific authority imparted by quantified probability often can withstand little scratching' (91). Finally, there are 'virtual risks' in relation to which scientists, experts, and institutional risk managers either confess ignorance or make contradictory claims. Examples may be found in environmental risks (e.g., global warming), health risks (e.g., passive smoking) or financial risks (e.g., equity market performance). Virtual risks conform to the old adage of W.I. Thomas: if something is treated as real it will be real in its consequences. Virtual risks have a liberating quality because if scientists, other experts, and institutional authorities cannot successfully foreclose their meaning, everyone is able to impose their own meaning, whether based on convictions, prejudices, or superstitions. Indeed, it is often a good thing that people do not put too much faith in science, especially its embodiment in technological systems for safety. A healthy scepticism about such systems helps us to avoid the Titanic effect: the complacency that results from believing that a system is infallible. The framing of risk also depends on the risk thermostat - Adam's system for understanding how risk-taking and -taming subjects balance their propensity to take risks, their perception of risks, likely rewards, and possible accidents. All risk taking transpires with a degree of uncertainty regarding rewards and losses. When risks are directly perceived, the balancing of propensity, perception, rewards, and accidents transpires through the mental processes of the individual risk taker. When risks are mediated by institutional risk managers, there is typically a bias towards the taming of chance and accident minimization. The rewards of risk taking are downplayed, the rhetoric of safety at any cost appears, and resources are poured into the taming of chance and the ever-elusive prospect of greater control. The risk manager rarely asks whether there

20 Part One: Theories of Risk and Morality are enough accidents in the context of the rewards gained through risk taking. The major exception is found in the financial industry, where the bias goes the other way: the focus is on the rewards of risk taking. Bad news of an 'accident' is minimized by claiming that it is an aberration or plain bad luck, and sometimes by dismissing the risk manager who gambled and lost. Dismissal is not a penalty for gambling with other people's money, but only for losing. The third component in the framing of risk is what Adams refers to as ethical filters. Sets of moral principles and ethical codes for risk taking are embedded in every framework for perceiving and responding to risk. Ethical filters intensify in moving from directly perceptible risks and rewards to the uncertainties of virtual risk. Adams offers four ideal types of risk takers and their respective approaches to the ethical filtering of risk. The individualist sees the world as robust, benign, and cornucopian and believes that risk taking will result in safe and equal outcomes. The egalitarian sees the world as fragile, ephemeral, and threatening, and prescribes collectivized approaches to risk that yield greater equality. The fatalist sees the world as dangerous and unpredictable, and risk taking as a matter of individual (mis)fortune and attendant inequality. The hierarchist sees the world as predictable within limits, but believes caution must be exercised in not pushing the limits in order to collectivize the impact of risk. Adams notes that the perception of any type of risk is shaped by whether the risk is voluntarily assumed or imposed. When risk is voluntarily assumed it is intimately tied to a sense of moral autonomy, and this sense is one of the greatest rewards of risk taking. This sensibility surely explains why people pay substantial sums to take extreme risks in adventure sports or gambling. When risk and safety are imposed, people will scrutinize and often challenge the authority of the risk managers involved. There is a bias in favour of being one's own risk manager, because being in control is essential not only to moral autonomy but also to identity, well-being, and personal growth. However, in the modern world of complex systems, scientific expertise, and virtual risk, interdependency is more operative and there seems to be less room for moral autonomy. In this context trust becomes the crucial 'ethical rudder.' Adams cites survey evidence to show that when it comes to knowledge about risk, people are unlikely to trust major institutions such as corporations and governments, deferring instead to family and friends. The problem is that family and friends are usually not very knowledgeable about scien-

Introduction

21

tific and virtual risks: they are trusted to tell the truth about what they do know. As for institutions, distrust about the safety systems they design and impose tends to make people fatalistic and selfish about the risks involved. Most people 'shrug off responsibility for big risks such as global warming. Meanwhile, they still engage institutions to help them seek individual compensation for losses they have suffered through risky activities. Litigation-based compensation on behalf of fatalists is on the rise, while welfare-based compensation on behalf of egalitarians is in demise.

2 Risk and Dirt IAN HACKING

This paper should really be titled 'Risk and Filth,' but I thought that sounded over the top. While dirt can be clean, filth cannot, and in part I shall be talking about defilement - hence my preferred but rejected title. I could also have used a polite title: 'Risk Portfolios.' How do we, individually and collectively, choose the classes of dangerous events and harms that we worry about? Most of us have a standard set of precautions, many of them dinned into us at an early age - if you are going to jaywalk, look both ways before you step into the street, and if the street is busy, be nippy. Many risks are ignored, after routine safety measures are taken. Driving a car, for instance, is dangerous, but few of us worry about it. Yet in recent history a small number of risks have given great cause for concern and have provoked intense political activism, less within political parties than outside them. The 'portfolio' of risks changes markedly over time. The choice of risks to worry about is rarely determined by what experts assure us are the 'objective,' 'real' probabilities and disutilities of the dangers. No general theory about what determines a choice of risks can be offered, for too many contingent facts and local stories affect the choice. No theory of risk-in-general could have predicted that the Germans, through the strategic importance of Greens, would decide to dismantle their entire nuclear power industry in the next few years, while the French are determined that within the next decade 90 per cent of their electricity will be nuclear sourced. There are important historical and socio-political explanations for these two phenomena, but no theory of risk can account for them on its own. What we might hope for is an understanding of necessary conditions under which a collective risk should become high priority, part of the communal risk portfolio of causes to work for (or against), or at least to worry about.

Risk and Dirt 23 I take the idea of a portfolio, much like a portfolio of stocks or of drawings, from a remarkable anthropologist, Mary Douglas, and her collaborator, a distinguished political scientist, the late Aaron Wildavsky. Douglas is the hero of this paper, for it is from her that I learned the connection between risks and pollution, both 'real' pollution and what the firm-minded among us would call symbolic pollution: literal and figurative filth. Douglas, the anthropologist, makes us a little queasy about this distinction. Of course there is a difference between incest, on the one hand, where the 'pollution' is figurative, and the literally polluting residues of industrial mercury in the drinking water available to some Japanese or Canadian Native peoples. But after reading Douglas we can begin to notice that the literal pollution of mercury can be invested with symbolic content. Conversely, reconsider incest, which is loathed in most communities, according to their own set of kinship relations (in my Anglican tradition the relations are set out in the Book of Common Prayer). Incest, especially when it is parent/child-oradolescent, the bogey of 'intrafamilial child sexual abuse,' is viewed with intense horror in our own world. When that kind of incest does not move into cruelty, the chief effects derive from the meaning, the symbolism of the behaviour, and they can be devastating for the child. Yet incest, when it is heterosexual, also carries the 'real' risk of inbreeding and defective children. In this discussion I shall be concerned not with risks to individuals but with some of the collective risks that in a certain sense cannot be distributed. Most forms of insurance are devices for distributing risks. Homeowners buy house insurance to share the risk of houses burning down, burglars entering, or a passer-by slipping on a bit of sidewalk that the owner has failed to clear of ice. Institutions of insurance - not just insurance companies, but also legislation, assessors, types of contract, and so forth - are relatively recent. In fact, I would say that the invention and practise of various kinds of insurance in order to distribute risks is one hallmark of 'modernity.' Since I am interested in risks that cannot readily be distributed, my topic is rather timeless, or at any rate not 'modern' in the sense just suggested. Although I shall make comments on insurance, I am chiefly bypassing that modern thing, distribution of risk through insurance. Mary Douglas's most famous book, Purity and Danger: An Analysis of the Concepts of Pollution and Taboo, was based on her early study of a group of people in the formerly Belgian Congo (Douglas, 1966). But she included in this book a chapter about Central Africa, which discussed the Abominations of Leviticus. That suits my purposes: the fears of the Lele

24 Ian Hacking of the Congo, of the Israelites of the Pentateuch, and today's panic about global warming are all centred on risks that cannot well be distributed or insured against. These three examples are drawn from extraordinarily different places and times, but this need not surprise us: members of almost any group of people experience collective dangers. The fears and dangers may well be organized, with variations, in much the same way in many social groups, cultures, and civilizations. Here I am not taking a polemical 'universalist' stance. Once we conceptualize something as a collective risk that cannot readily be distributed, we are likely to find many commonalties that are indifferent to the place or date at which a group of people establishes itself. Words

But first things first. For an analytical philosopher like me, the first things are the words that we use. 'Risk,' says the OED, is 'exposure to mischance or peril.' 'The possibility of suffering harm or loss; hazard, danger' - that's the American Heritage Dictionary's definition. For French, Le Petit Robert has 'danger eventuel plus ou moins previsible'; on the CDROM version, we are electronically directed to the cognate words hasard, peril, danger. Each dictionary's definition flags a slightly different aspect of the idea of risk. The French reminds us of the core idea of prevision. When we do not have foresight we become foolhardy. The OED points to mischance and the close connection between many risks and accidents. 'Accident' itself achieved its most common current meaning of a grave mishap only in the steam age, with the advent of industrial coal mines and the railways, whose boilers were always blowing up, whose locomotives went off the tracks, and whose cuttings caved in. Possible disasters, catastrophes, and threats are more nebulous than actual harm. Risk, exposure to possible harm, is more shadowy still. Are we concerned with real exposure, of which we may be ignorant? Or are we concerned, as Le Petit Robert implies, with what we can more or less foresee? These words, 'danger,' 'peril,' 'risk' (and concomitant 'accident' and 'liability') run around each other. The Oxford linguistic philosophers used to play a game called 'Vish' - for 'vicious circle.' Each player has a dictionary, and a word is chosen. Each contestant looks up the word, then the words in the definition, and so on, until someone finds the original word and cries 'vish!' On our subject, the Shorter Oxford has: Risk: Danger, (exposure to) the possibility of loss, injury, or other adverse circumstance.

Risk and Dirt h25 Danger: Liability or exposure to harm or injury; risk, peril. (Vish!) Peril: Liability or exposure to the possibility of imminent injury or destruction; jeopardy, danger. (Vish!)

And so we run around in circles, but not quite, for each word has its own family of connotations. Notice that these words incorporate two ideas, one of which refers to something relatively distant from us, something that we think of as out there: harm, injury, loss. Richard Ericson and his collaborator Kevin Haggerty (1997: 4) write that 'Risk refers to external danger, such as natural disaster, technological catastrophe, or threatening behaviour by human beings.' This list of what risk 'refers to' is a list of types of harm that may befall us, all coming from outside, external. There are internal dangers too, illness, mental and physical, not to mention emotional catastrophe: 'If he goes on working overtime like this, there's a real risk that his wife will leave him.' My own wife crossed a continent and an ocean after her mother's stroke. Mother seemed unlikely to make much of a recovery, but was likely to lie there for months, more or less out of contact. My wife made the journey because of the risk that her mother would die, without the daughter being there. Death of another may be external, but the emotional damage of not being there is internal. Notice that such risks - losing your wife, not being present at the time of your mother's death - cannot be distributed, but those are not the non-distributable risks that will concern me here, for they are personal, not collective. Peril, danger, risk, foresight: the historical topography of these different words is fascinating, but this is not the occasion to indulge in philosophical philology. I should just say that despite the apparent nearuniversality of collective risk and danger to which I have alluded, when we analyse risk we are employing a quite recent family of ideas. The word 'liability,' which plays so large a part in equitable litigation about risk, is of eighteenth-century coinage. 'Risk,' from French 'risque,' derived in turn from Italian 'rischio' dates from the sixteenth century, and seems first to have been used among the new merchant classes, especially in connection with commercial navigation. I do not mean that it was first used by merchants when discussing the risks of their business ventures, or ship's masters engaged in nautical adventures. It was so used, but it was used at the same time to connote a new sense of life and morality. Quentin Skinner has pointed out that prudence, both the word and the idea, changed meaning about the sixteenth century. In courtly language, prudence denoted cowardice, the lowliness of the frugal, lack of honour, selfishness. From the sixteenth century the prudent became the

26

Ian Hacking

wise who accepted the moral duty of attending to the future, of saving for a rainy day, the virtue of foresight. Although we know the etymologies of those older words 'danger' and 'peril,' the Italian 'rischio' is 'of uncertain origin.' It is as good a word as any to signal a new way of being, connected both with commerce and religion, a new way of seeing and experiencing the world. It is connected with our place in the world and what we ought to do about it, one of many words that span fact and value, 'is' and 'ought.' You need not be one of those people who endlessly reflect on the nature of modernity in order to see that something new entered our being. The new prudence and the idea of risk lead to insurance, which 'forms an institution peculiar to the modern world, the origin and growth of which attest a remarkable change in men's ideas and habits of thought.' Those are the words of my trusty if tattered 1911 Encyclopaedia Britannica (llth ed., 14: 656). Hope and Fear To return to prudence: European travellers, merchants, missionaries, settlers, conquerors, and administrators constantly noted and usually bemoaned the lack of foresight among the peoples whom they encountered and whom they labelled as savage or primitive. In any particular case we think it likely that a people dismissed in this way was meticulous in foreseeing and predicting things that were important to them. We do not expect a lack of prevision, only a difference in interests from those of explorers and those who came in their wake. There is also a general and rather a priori point to make. It is strange that any group should lack prudence and a sense of risk. The cognitive science of the emotions, which has only just became fashionable (in the 1990s) focuses on fear. To take one example among many, the book published in 1998 by the distinguished neurologist and cognitive scientist Joseph LeDoux (1998) is titled The Emotional Brain: The Mysterious Underpinnings of Emotional Life, but it is not about emotions, or the emotional life. It is about fear. Fear is said to have its physiological roots in the amygdala, deep in the nervous system. Fear, we are told, is innate if anything is. Another new and modish science, evolutionary psychology, makes this an a priori truth, for without instinctive fear to warn sentient beings like ourselves, we should never have survived. We procreate slowly and take a long time to grow up, and fear helps the young run and hide. If all peoples are born with fear and are aware of dangers, why did the

Risk and Dirt 27 European travellers find no sense of risk among the 'savages'? Because risk is the calculating concept that modulates the relations between fear and harm. The 'primitive' did not calculate. Calculation began to dominate instinct, tradition, and collective wisdom, in a measured way, only in seventeenth-century Europe. Of course there were community councils elsewhere in which the merits of various strategies of policies were debated - but not by the calculations that Europeans invented in the seventeenth century. One of the earliest authors who told us how to calculate, using games of chance as model, rather than to weight specific ventures, was the physicist Christiaan Huygens. In 1658 he published the first textbook of probability theory, first in his native Dutch and then in learned Latin. He chose 'Hope' and 'Fear' for what we now call positive and negative expectation. We commonly define expectation as a weighted sum of probabilities. Expectation, probability: those are sanitized Latin words with which rational choice theories, today's rendition of man as a rational animal, conceal our emotions, our hopes, and our fears. Huygens went for the gut feelings: he defined probability in terms of hope and fear. Mathematically that is unproblematic: from a formal point of view either expectation or probability can be used to define the other. But emotionally, Huygens's frank words are more in touch with our souls, and I shall use his words quite often. What harms or injuries should we fear? About the same time as Huygens wrote his textbook, the authors of the Port Royal Logic, or the Art of Thinking, undoubtedly advised by Blaise Pascal, made a classic statement of what people ought not to be afraid of. 'Many people are exceedingly frightened when they hear thunder ... If it is only the danger of dying by lightning that causes them this unusual apprehension, it is easy to show that this is unreasonable. For out of two million people, at most there is one that dies this way. We could even say that there is hardly a violent death that is less common. Our fear of death ought to be proportional not only to the magnitude of the harm, but also to the probability of the event. Just as there is hardly any kind of death more rare than being struck by lightning, there is also hardly any that ought to cause less fear' (Arnauld and Nicole 1994 [1662]: 273-4). Next, they present the same considerations for hope, using the lottery as an example. They had not quite arrived at conditional probability. Their observation provides cold comfort if you are hiking exposed over bare rock in the Sierras while a thunderstorm is raging, never more than three seconds between a bolt and the sound; there were twelve deaths by

28 Ian Hacking lightning in these parts last year! Nevertheless, the conceptual leap had been made in print. Fear and hope should be a function of two things, utility and probability. That had already been known in an intuitive way to anyone who dabbled in the insurance of, for example, merchant ventures. And there, in a certain sense, things rested, from 1662 to 1969. Well, not quite! A few fundamental developments occurred, as described below. The Distribution of Risk After 1662, there was the growing realization that risk could be distributed. This can be achieved in two ways. There is the Lloyd's model. A risky voyage is to be undertaken. The owner of the vessel pays a sum to a syndicate, a group of wealthy men. If the vessel founders, the syndicate reimburses the owner a set amount to cover his losses. If it returns safely, the members of the syndicate divide the premium among themselves. Individual Lloyd's-type insurance is always a gamble, as Lloyd's itself discovered a few years ago, when inexperienced 'Names' of insufficiently deep pockets were, for reasons of greed, allowed to form syndicates whose risks they could not afford. And there is always the danger of a totally unexpected catastrophe. I do not mean something like an earthquake, which is unpredictable but a known risk; I mean catastrophes we have never thought of, all too familiar now with the destruction of the World Trade Center in New York, which has thrown the insurance world into disarray. Lloyd's then, provides one model, chiefly applied to situations in which by the nature of things there cannot be much actuarial experience. In the beginning of the space age, commercial rocket launches were routinely insured at a premium of 25 per cent of the payload. An individual or a corporation was covering one particular risk among a totality of rare events, namely rocket launches. But the premium was a straightforward bet; with a few such bets the insurer hoped to profit, and the rocket launcher simply could not afford total ruin if the rocket crashed. There were not enough launches to constitute an actuarial collectivity - we had reverted to the old days when a ship owner insured against the pirates in the Straits of Malacca or off the Barbary Coast. Then there is another model, the collective model, exemplified by life assurance, fire insurance, and the like, in which a large number of people with similar risks pay premiums and are reimbursed if they suffer actual harm. Life annuities are the converse of life insurance. They can even turn into a zero-sum game, as in the tontine, in which a hundred

Risk and Dirt 29 people of equal age contribute one hundred dollars to a bank account. The last surviving member collects $10,000 plus interest. But we should not underestimate the difficulty of sound actuarial practice. You need a lot of data, and that was available only by the mid-nineteenth century, as a consequence of the event that I call the avalanche of printed numbers (Hacking 1983; 1990: chap. 4). Insurance changed the perception of risk. Collective insurance and the state have usually gone hand in hand. Sometimes the state itself is the insurer, as in the social net insurance inaugurated in Prussia and then in Bismark's Germany in the latter half of the nineteenth century: unemployment, workmen's compensation, mortgage insurance for labourers. This type of insurance developed into what in Canada is our biggest enterprise, health insurance. Health care now consumes 6.5 per cent of our GNP, still cheap in comparison to the United States. Sometimes the state is merely the regulator of commercial insurance companies. But there is far more to the link between the state and insurance than that. The state has become the primary manager of risk. The German Sozialnet was invented partly out of a conservative sense of responsibility and of the duties owed by the rich to the poor, and partly from a need to keep the working classes stable. Pierre Rosanvallon, however, makes the more striking suggestion that distribution of risk was one of the immediate consequences of the French Revolution (Rosanvallon 1998). In 1986 Ulrich Beck (1992) dubbed our modern form of society the 'Risk Society,' and the name has taken hold even in the title of Ericson and Haggerty's book, Policing the Risk Society (1997). But Rosanvallon is in effect claiming that the risk society was well underway by 1800. Thus however we understand the term, we should not think of the risk society as simply a 'New Modernity,' a phrase from Beck's English subtitle. Michalis Lianos, in collaboration with Mary Douglas, has coined the term 'dangerization' to describe the ways in which future events and interactions are conceived in terms of danger, subject to a system of automatic controls that will remove the danger and lead to safety, which is defined as the normal state (Lianos with Douglas 2000; as I refer in this paper chiefly to Douglas and Wildavsky 1982, this is a useful reminder that she does not stand still). Risk Analysis

From the date of the Port Royal Logic, probability, statistics, insurance, and decision theorizing developed apace. The greatest of European revolutions brought the risk society in its wake. Insurance was taking in

30 Ian Hacking new territory, especially as new technologies, such as steam, brought new kinds of risk into being. There were brilliant innovations in probability theory, and substantial developments in the theory of utility. Economics emerged as a far from dismal science. And probability took over our lives, in ways that I tried to describe and account for in The Taming of Chance (Hacking 1990). But I would like to propound the paradox that as far as risk was concerned, nothing much really new happened until 1969. Of course I exaggerate, but I do so in order to highlight an event that will return us to Mary Douglas and pollution, and in particular her joint book with Aaron Wildavsky, Risk and Culture: An Essay on the Selection of Technological and Environmental Dangers (1982). What happened in 1969? Here I am following the recommendations of one methodology much encouraged among today's historians of science - look for institutions and the creations of professions. Nineteensixty-nine is the preferred year for dating the emergence of risk analysis as a newly institutionalized activity, with a new professional: the risk analyst. Risk analysis entered the scene in 1969 in a reactive way. It turned its attention to a relatively small class of possible dangers. This is because it was reacting against fears played upon by militant activists. Why choose 1969? It is not my choice. In the folklore of the profession of risk analysts, 1969 was singled out because it is the year that Chauncey Starr (1969) published a ground-breaking popular paper in Science. At the time that Douglas and Wildavsky were writing their book, they could have read 'Modern risk-benefit thinking had its real birth with the classic paper of Chauncey Starr' (Okrent 1980: 375, n. 3). For me, 1969 is a good enough date to mark the establishment of the profession as profession, with numerous specialized journals, organizations, meetings, and all the usual paraphernalia of establishment coming in train. Note that I say risk analysis, not risk management, and I should pause to explain the difference. The OED cites a title from 1963, Risk Management in the Business Enterprise. That turn of phrase is easy, but a selfdefined profession of risk management came later. 'Risk management' is the topic of Michael Power's paper in this volume, and he cites 1995, not 1969, as the definitive year marked by a wave of institutions, journals, and in-house corporate restructuring. Exactly so: for risk management. (This is by coincidence confirmed by another author in this volume, Nikolas Rose (1998), who has written a paper about the introduction of risk management into British hospitals - in 1995.) So what's the difference? I have been told by a major independent consultant that corporations divide risk analysis and risk management

Risk and Dirt 31 rather sharply in the corporate hierarchy. Risk analysts and risk managers report to entirely different superiors. The crass distinction between the two occupations is simple. Risk analysts are concerned with fear. Risk managers are concerned to balance hope and fear, where hope is bottom-line profit. To illustrate, a classic of risk analysis is Acceptable Risk (Fischoff et al. 1981), a set of essays by four authors; it is contemporary with Risk and Culture. A typical problem posed is why are people afraid of new risks (nuclear meltdown, or, in our day, genetically modified foods), while the known risks of driving an automobile have great disutility and far higher probability? How are we to calculate what risks are acceptable to citizens, consumers, or corporations? In Risk Acceptability According to the Social Sciences, Mary Douglas (1986) was to argue in detail, case by case, that most of these questions were wrong-headed, but the earlier Risk and Culture contains her more fundamental critique. Although 'risk-benefit' (analogous to cost-benefit) analysis was used as another name for risk analysis, benefits played almost no role except as boundary conditions. Risk analysis was the analysis of fear, and of the fears that we can accept, and accommodate. But it was a short-lived enterprise, as Professor Power implied. The calculating good sense of the seventeenth century was restored to corporate planning. There is nothing new in risk management: at bottom it is just the assiduous creation of models within which one can maximize utility (read profit). Sometimes the risk managers make horrible mistakes. Take the recent events connected with Firestone tires. It is widely believed that the risk managers put into their calculations the cost of a significant number of deadly rollovers, including the costs of subsequent litigation (thought of as the chief measure of the risk to the corporation of a significant number of deaths and injuries). The costs of fixing the design problems were (the risk managers are thought to have concluded) greater than the expected loss in lawsuits. Big mistake. The risk managers (according to this unkind conjecture) did not factor in investigative reporting, television, and the possibility of a massive public outcry. As we know, the tire scandal hurt Firestone sales across the board, necessitated massive recalls, and attracted vastly greater penalties than would have arisen from individual civil cases or buyouts of victims' families. Risk analysts might have done a better job and demonstrated that this risk was not acceptable. Maybe they did do a better job, but reported to another and less important part of the corporate structure. Note that I am here reporting widespread popular belief, and not affirming such allegations.

32 Ian Hacking But we have excellent reason to believe such a thing happened with the Ford Pinto. In the case of Grimshaw v. Ford Motor Company (119 Cal. App. 3d 757, 174 Cal. Rptr 358 (1981)), a Los Angeles jury found that the company had deliberately weighed the cost of possible deaths against the expenses and decided to economize - it awarded $325 million in punitive damages. These damages were ruled excessive on appeal, but the facts were not in question (Hampton 1992: 1687-91). A distinction has, then, evolved between risk analysis and risk management. Risk management has taken over a duchy in corporate structure, and, to a lesser extent, in the more disorganized and hands-on approaches of politicians and governments. Yet even a date-cruncher like myself can see continuity behind the mutations. In the recommendations at the end of Fischoff et al. (1981) we read that the authors would like to help 'create a profession of risk management.' Risk and Culture Why did risk analysis emerge in 1969? Had we entered a new and frightening epoch in the history of industrial civilization? No, we had not, but that is what it felt like. Collective non-distributable risks were experienced as never before. That in turn was part of the larger picture of politics and activism that characterized the late 1960s. This was a 'cultural' phenomenon. Douglas and Wildavsky's slim Risk and Culture appeared a little more than ten years after the advent of risk analysis. The book was completed in 1981, the year in which the Society of Risk Analysts was founded, and a new journal was launched, Risk Analysis: An International Journal. Our two authors did something quite out of the ordinary. They looked not at how risk should be analysed, but at how the risks to analyse are chosen. Both authors cast their nets widely. Illustration: Douglas's eightieth birthday was feted at the British Academy on 25 March 2001. The daylong series of talks in her honour was organized by two younger anthropologists and admirers, one of whom has written her biography. But only one of the seven speakers addressed her fundamental contributions to anthropological theory and practice. We spoke about, well, everything, including the classification of Chinese ceramics. The nearest there was to an overlap in topics occurred when a postmodern professor of religion and a near-Eastern textual scholar both made remarks bearing on Douglas's most recent book, Leviticus as Literature (Douglas 1999), a successor to her In the Wilderness: The Doctrine of Defilement in the Book of

Risk and Dirt 33 Numbers (Douglas 1993). I have already indicated what Biblical exegesis and fieldwork in the Congo, the Abominations of Leviticus and the pollution rules of the Lele, have in common. Pollution and danger are the theme of Risk and Culture too. In what follows I am going to speak as if Douglas were the sole author of Risk and Culture, because it was she who brought to it her thoughts about pollution, upon which I shall build. So I must emphasize now that it was a joint work that came into being thanks to a fairly chance intersection of the authors' professional lives, when both, under slightly eccentric circumstances, found themselves in the basement of the Sage Foundation building in New York. Wildavsky, who died in 1993, published a prodigious amount of work on public budgeting, policy analysis, presidential elections, and public administration. Several of his books are rather fiercely anti-Soviet. It happens that he too wrote a thoroughly iconoclastic book about the Pentateuch, The Nursing Father: Moses as a Political Leader (1984). The fact that I shall not mention him again owes solely to a desire for brevity and a focus on pollution. Risk Portfolios Whichever author first thought up that useful phrase, 'risk portfolio,' I expect that Wildavsky liked it a good deal. It is already a capitalist metaphor, or rather, an adaptation of a capitalist metaphor. Around 1930 stockbrokers began to refer to a collection of stocks owned by a person or an estate as a portfolio. The metaphor is a perfect extension from stockbroking, for of course a portfolio of equities is a portfolio of risks, a portfolio of hope and fear. A risk portfolio is just that set of hopes but especially fears that moves you, that concerns you, that you feel strongly about. If you are an activist, it is the set of risks that engage you, that you combat. And it was the activists' risk portfolios that fascinated Douglas in 1980. She was living through the great epoch of environmental groups: Greenpeace, the Sierra Club, and many, many more ephemeral groups, such as the Clamshell Alliance. What were their risk portfolios? How and why did they choose them? The match with the anthropologist's idea of pollution was inevitable, for literal pollution was one of the key ingredients in the risk portfolio of many environmental groups. When I first read Risk and Culture in 1982 I drew up a triangle of contemporary fears (Hacking 1982). I am not sure now that my table was a good representation of the tableau of risks at that time, but it stands as an eyewitness report. It looked something like this:

34 Ian Hacking Nuclear war (Armageddon) Cancer Nuclear meltdown Polluted air Radiation Toxic waste Threatened species Smoking Lead Asbestos Saccharine Poisoned waterways Auto emissions This is a table drawn from a particular perspective, drawn from where I myself was. I was near, but not on, the border, in a sense to be explained. I was trying to be impartial, to describe risk as it is, or rather was then, but I was not actually objective. If I had positioned myself nearer the centre, crime, drugs, terrorists, teenage pregnancy, and drunken drivers would have been high up on my triangle of fear. Border and Centre

The reference to borders and centre refers not to landmasses but to groups. Anthropologists study groups of people. They do not just study Anthropos, that is, 'Man,' human beings, they study human beings in groups, as grouped. So a first question is, what keeps people together in a group? The question took a long time to ask, because the anthropologists of the imperial era simply went overseas and found people living in a place, together. Douglas asked, what keeps people together in a group? When does a group fall apart? She developed a whole theory that I shall not go into. The code names are 'group' and 'grid,' not words that I find happily chosen, although what they refer to is deeply important. By group she meant 'the outside boundary that people have erected between themselves and the outside world. Grid means all the other social distinctions and delegations of authority that they use to limit how people behave to one another' (1982: 138). Three familiar kinds of society can be characterized in these terms. Hierarchical societies have many group-encircling regulations, and many grid constraints on how to act. Japan is just such a society: it is virtually impossible to become Japanese; marriage and residence and even living in Japan for seven generations are not sufficient for entry into the group. And Japan remains extremely hierarchical. Individualistic societies, at the extreme, 'leave to individuals maximum freedom to negotiate with each other and have no effective group boundaries' (1982: 139),

Risk and Dirt 35 while an extreme 'sectarian society would be recognisable by strong barriers identifying and separating the community from non-members, but it would be so egalitarian that it would have no rules of precedence or protocol telling people how to behave' (ibid.). No nation is strictly individualistic, because nations as we now identify them have border controls; idealists have dreamt of truly sectarian nations. Douglas has represented societies on a 'group-grid axis.' The United States, for example, is one of the lowest on both group and grid. Compared to most nations it has been quite generous with respect to admission, the wall on the border with Mexico notwithstanding. And it is low on grid, more so in egalitarian principle than in practice. Douglas also made a surprising, powerful, and insightful application of her group/ grid model to environmental organizations, explaining or predicting which ones would grow into vast bureaucracies (the Sierra Club) and which would fall apart, splitting into sects. Border and pollution are powerfully connected. Douglas began her research in the Congo, working with a people who had what seemed to us very strange pollution rules governing what you could eat, whom you could meet, what could be touched, what must be avoided at all costs. She proposed that pollution is an almost universal device for defining and maintaining a society, and above all for creating a society's sense of who it is, what it is. A Canadian has to find this insight rather embarrassing. For, with a blush, we all know our primary purity/pollution concept. It is universal health care. That is what makes us not American. The United States is impure, evil, immoral, dirty, untouchable, because it does not have universal health care. We are pure, righteous, even if the expenses of advanced medicine combined with incompetent management, or, as in Ontario, gratuitous governmental viciousness and cruelty in the name of efficiency and corporate profit, means that the system does not work as well as it once did. It would be a total disaster for Canada if the United States were able to devise something like nationwide, universally available health care. Happily it will not happen, for, aside from the entrenched power bases, health care is simply becoming too expensive for all risks and all treatments to be distributed uniformly to all members of the population. Subsidiary pollutions also fix our borders. One is crime. I never cease to be astonished at how deeply the pollution rules are embedded in even our most alert and critical young minds. I strongly support an undergraduate's attempts to enter the University of California at San Diego (UCSD). He succeeds; he wins a Regent's scholarship. Marvellous. But

36 Ian Hacking he has second thoughts. How can he go live where violent crime is so rampant? I do not convince him that UCSD is vastly more safe than Parkdale, the scruffy, drug-infested Toronto neighbourhood where he lives. The fear and the pollution are what we analytic hard-liners call symbolic, not realistic. My student will get excellent health coverage from his university, and he need never enter a dangerous neighbourhood. At UCSD he is far more likely to become a paraplegic from a surfing accident, or to break his ankle playing beach volleyball, than he is to be mugged. But I have come to be very wary of the distinction between realistic and the symbolic pollution. That Triangle of Fear

Let us return to the problem of groups. Douglas was fascinated by the protest groups that were proliferating around her. Here we have to notice two things. First, the connection with pollution, which will preoccupy me; but second, more interesting to her, the way in which different groups defined themselves, and how those definitions affected how they stayed together or fell apart. First, think about my triangle in terms of pollution. At the top is apocalypse and Armageddon, nuclear war thought of as the end of the world. Strictly speaking, 'apocalypse' means revelations, and named the book of Revelation and similar contemporary warnings; Armageddon, the final battle between good and evil, is foretold in Rev. 16:16. But Coppola's film Apocalypse Now somehow confirmed the way in which 'apocalypse' has recently garnered most of the connotations of Armageddon and the end of the world. A great many people whom I have asked think that 'apocalypse' means the end of the world, so I suppose that it is what it means, by a curious leap of metonymy. But the original Apocalypse, related by Stjohn of Patmos, provides the veritable archetype of pollution. When the seven angels blow their seven trumpets pollutions invade the earth. The fifth angel opened the bottomless pit; and there arose a smoke out of the pit, as the smoke of the smoke of a great furnace; and the sun and the air were darkened by reason of the smoke of the pit. And there came out of the smoke locusts upon the earth; and unto them was given power, as the scorpions of the earth were given power. And it was commanded them that they should not hurt the grass of the earth, nor any green thing, neither any tree, but only those men which have not the seal of God on their foreheads.

Risk and Dirt 37 And to them it was given that they should not kill them, but that they should be tormented five months; and their torment was as the torment of the scorpion, when he striketh a man. And in those days shall men seek death, and shall not find it; and shall desire to die, and death shall flee from them. (Rev. 9:2-6) That is pretty much how people saw the after-effects of nuclear war. In those days we had the neutron bomb, which leaves buildings and plants intact, exactly as foretold, but mercilessly destroys humans and other animals over an agonizing few months. Nuclear war may sound like a lot more than pollution, but it is Armageddon. This is true even for those in our midst who do not know where the word comes from, for they have embedded within them - deeper than mere memory - the words of St John of Patmos telling of ultimate pollution. Note, incidentally, another feature of Revelation: the good group, the 144,000 with the seal upon them. That was the model for environmental groups, especially the more extreme. They were the chosen, closer to purity than thee. Purity and pollution are intimate, the obverse and the reverse of the coin of culture. Look at the items below Armageddon. In reality, cancer is a horrible disease of ageing, with a few forms such as leukaemia that attack the young. But it is experienced as the ultimate personal pollution, an alien invader that grows within the confines of one's body. Nuclear meltdown is feared because of that most insidious of all pollutions, invisible, odourless radiation that seeps into our bodies, and rains down on us as strontium 90, imperceptibly poisoning our children's milk. The remainder of my triangle is likewise a list of pollutants, with the exception of threatened species. Here the focus is not on pollution but its partner, purity. There is a certain amount of rhetoric deployed to argue that destroying plants in the Amazon reduces the potential for miracle cures. Logically - but not emotionally - that has nothing to do with worrying about a rare and rather drab owl in the rainforests of the Pacific Northwest. The owl impedes the livelihoods of literally hundreds of men and families who live off the woods. Even I, a complete diversity-nut, who positively cultivates the amazing variety of insects that populate my tiny plot of inner city garden, can scarcely credit the utility attached, in the popular and political mind, to preserving a stupid if uncommon owl. I must put at least part of it down to a desire to preserve purity, a conception of nature-as-it-is as pure. Notice that you cannot insure against species extinction by distributing the risk, no more than you can insure against nuclear war. In fact of

38 Ian Hacking the first three layers in my 1969-82 triangle of fears, cancer, as an individual illness, is the only one of the risks against which you can insure (at least for medical costs, loss of income, and death). The financial risk of cancer can be distributed among prospective victims, even if the suffering cannot. With enormous ingenuity Canada and the United States have worked out a way to do something like distributing the risks of fossil fuel-use/ozone-depletion. We exchange carbon credits with less profligate parts of the world. The Centre My risk triangle was composed of fears that moved the border. I left out the fears that move the centre, the concerns that preoccupy both the middle classes and the broad mass of ordinary workers, farmers, and homemakers. Crime looms largest. It is essential to realize that for most people the fear of crime is not a fear of loss by theft. Of course we fear murder, grievous bodily harm, and violent threats. But what we really fear after brute physical injury is the invasion of our privacy. Ordinary people seldom suffer much from burglary, but they are much pained. (Nothing is simple: breaking and entering is far more common in Canada than in the United States, but almost never results in injury, whereas in the United States with its armed citizenry, burglary victims are far more likely to be injured or killed than in Canada.) The private world of the burgled household is invaded, and made dirty by foul, probing hands. The burgled feel defiled. The punishment is indeed made to fit the crime. The recidivist burglar is imprisoned, where he in turn will be defiled, and perhaps raped. All the vapid protest about poor prison conditions, a world of violence, smuggled drugs, and rape, quite ignores the point: heavy duty prisons serve their purpose: they defile the defilers. The embezzler who has helped himself to ten thousand times the loot is put away in relative safety in a white-collar prison, quite likely a farm. The official excuse is that the embezzler is not a threat to society, but I believe the real reason is that he has not defiled us, so we will not return him to filth. The marijuana user, who is usually no danger, is usually put away with the filth, for he has engaged in ritual pollution. You will have noticed that I put drugs second on my list of items not found in my triangle. We have great debates about decriminalizing both soft and hard drugs. In the present state of affairs the heroin addict is indeed a danger, which he would seldom be if kept sedated on his drug

Risk and Dirt 39 of choice and kept from driving. Drugs cause real harm, especially to underprivileged people such as those in the black ghettos of America. They cause real harm not only to users but also to all those who live around them. Not to mention the civil wars in Columbia. But the war on drugs is not presented primarily as about real harm; its rhetoric is the rhetoric of symbolic pollution. Teenage pregnancy, as it was still called in 1982, falls under breach of purity rather than pollution. Drunken driving? That concern is epitomized by one of the powerful activist groups coming entirely from the centre. MADD: that is, Mothers Against Drunken Driving. Douglas and Wildavsky did not mention MADD, but lest we forget look at the dates. They published in 1982. MADD was founded in California in 1980. President Reagan announced a Presidential Task Force on drunken driving in 1981, and invited MADD to serve on it. MADD had over a hundred chapters in the United States by the end of 1981. Mothers and Reagan in alliance: that's the centre all right. Who could not support the aims of MADD? To be cynical would be to be cynical about Mom and Apple Pie. Drunken driving, death and injury on the roads associated with alcohol, is a real peril. But at exactly the same time that MADD was founded, Joseph Gusfield's (1981) pioneering, iconoclastic, and I think rather daring study of agenda setting in politics showed by means of detailed statistics (to which no one pays the least attention) that the reality of the peril is not the real issue. It is, in the sense of what I am saying, the symbolic power of the evil danger, driven by the pollution that is alcohol, that mobilizes action. Handguns and Purity

This is perhaps the moment to state the obvious, that the reality and the symbolism of pollutants make a complex story. In most of the industrial world, guns are pollutants, both real and symbolic dangers. Handguns have almost no use except in the hands of police to control armed violence, and rifles are of no use except for those who legitimately hunt game. But guns are stupendous symbols of power and virility. We might say that for the 'left' handguns evoke sinister killers feared by women and the effeminate, for the 'right,' male justice, proud members of the patriarchy, proud tools. No amount of reality talk about their dangers is going to make the slightest difference in the United States. As is so often the case, symbols rule the roost of risk. But notice that in America, guns are defined by purity that overrides danger. First, they are part of the

40 Ian Hacking original purity, embedded in the definition of the identity of the United States, namely its constitution. In Great Britain, Magna Carta is a souvenir written in funny letters. In Canada, our Charter has a built-in 'notwithstanding clause' - notwithstanding what the Charter says, we can do the opposite. In America, the constitution serves as holy writ, interpretable but inviolable. Second, handguns defend the proclaimed essence of America, liberty. Third, they defend the individual against intrusion by the other, invasion of the private place, the home or the body. If guns were merely phallic, Americans would have dumped them long ago, just like the rest of us. But guns are far more than that. They are symbols of purity, of origin, of identity, of what preserves the border against transgression, at all costs. Pollution as a Necessary Condition for Entry into a Risk Portfolio At the beginning of this paper I said that no theory of risk could explain or predict the highly contingent factors that from case to case and time to time determine what goes into a collective portfolio of nondistributed risks. I gave the contrast between France and Germany on the issue of nuclear power as an example. Everywhere opponents of nuclear power point to pollution: disposal of nuclear waste, transport of nuclear waste, the excess of leukaemia in the proximity of nuclear power stations due to insidious radiation pollution, thousands of acres poisoned by Chernobyl, and the ultimate fear of meltdown. These are pollutions all right, and they have moved the German state and its citizens. Yet they are completely ineffective as cause for concern in France. No theory of risk can explain that fortuitous, contingent, historical difference. Closer to home, we present (as is our talent) the unique Canadian heavy water CANDU system as a benign reactor that cannot melt down like a thermo-nuclear one. But even diligent activists like Energy Probe have not tried to work up its pollutions. It is a serious heat pollutant, for it cools itself by slowly raising the temperature of Lake Ontario, making the lake even less inhabitable by fish. Worse, no one mentions that it is one of the most dangerous polluters in the whole world, for it is the prime source of tritium, used for weapons-grade materials (the best!). Instead, our government and its agencies endeavour to export CANDU technology, in ways subsidized by the Canadian taxpayer, to some of the most unstable parts of the world. No one is upset. CANDU satisfies an apparent necessary condition for entering a risk portfolio - it can be made out to be a very dangerous polluter. But that was not sufficient for

Risk and Dirt 41 it to become a source of concern, perhaps because of contingent facts about the exceptionally conservative demeanour of the Canadian public. (I mean 'conservative' in the proper, traditional, don't rock a floating boat, sense of the word.) All that is to say that we cannot establish generic conditions for entry into a risk portfolio. The thesis I extract from Douglas and Wildavsky is that pollution is a necessary condition for the entry of a non-distributable risk into a collective risk portfolio. Of course, we have to be careful about how generous we are at construing something as 'pollution.' The relevant question is not whether it 'really' is pollution, but whether in the rhetoric of debates something is presented either as pollution, or as an assault from across the border on purity. As a cold analytic philosopher I may wish to make firm distinctions between real, figurative, metaphorical, and symbolic pollution, but in social life the boundaries are blurred. Real (for example chemical) pollutants become symbolic. Conversely, symbolic pollutants are totally real in their effects. There is an incredibly rich play of symbols and realities. In Hacking (1982) I noted how two young mountaineers climbed the 550-foot smokestack of the Magma Copper Company of Arizona, and just sat there, drawing attention to the poison spewing out across the desert. They recall St Simeon Stylites and all the other fifth-century saints who fought for purity by sitting alone atop columns in the desert. Test Case: Terrorists

The obvious test case for a bond between pollution and risk portfolios is the current terrorism scare. I refer specifically to the fear of terrorists, not to fears about the evolving story of reactions and reactions to reactions. The successful attacks of September 11 clearly aimed at (a) American capitalism and commercial hegemony, in the form of its prime symbol, the World Trade Centre, and (b) against American military hegemony in its symbolic representation by the Pentagon. It is not known where the unsuccessful plane-missile was headed, but it is presumed that it was aimed at another symbol. But in the immediate reactions we did not hear a word about capitalism or military might as the target of attack. No, the target was liberty itself. This is despite the fact that the only internal constraints on American liberty have been imposed by the government itself. Fear of terrorists has had unbelievable consequences for civil liberties in North America. Everyone from the U.S. President down announced, within minutes,

42 Ian Hacking that the attack was an attack on liberty, which in American language is, among other good things, code for purity. In contrast, American terrorists terrorizing Americans are not part of anyone's risk portfolio, for they are not pollutants. They are a little embarrassing, because in their demented way they present themselves as defenders of purity. They are on the border, trying to liberate the United States, by violent means, from its own self-pollution and impurity. This is true of both the extreme left border — the Unabomber - and the extreme right border - the Oklahoma bombing. But in addition to the terrible deaths of innocent people, which you would have thought would be sufficient to stir anger, the external terrorist attacks were plainly painted as some kind of defilement, an assault on the home of purity, defined as 'liberty.' The immediate follow-up was no figurative pollution, but the anthrax scare, caused, it seems by another demented American, but a perfect exemplar of pollution. And we are warned of the most likely subsequent pollutions, either a biological attack by terrorists, or a freighter with a 'dirty bomb' of ordinary explosives packed inside masses of low-grade nuclear contaminants sailing into New York harbour and contaminating the city forever. If you read the rhetoric after reading Risk and Culture, you will see that although potential loss of life is taken very seriously, what makes these possibilities so monstrous is that such attacks would pollute, contaminate, defile, poison. I am not saying that there is no real threat. I am saying that in being made central to a risk portfolio, the threat is transformed into an assault on purity. And it is experienced that way, as a matter of common sense. (This is connected with the doctrine of Antonio Gramsci and Stuart Hall, according to which hegemony is now achieved by having the demands of hegemonic domination experienced as the desire of all as a matter of common sense.) Sceptics and stiff-upper-lippers have pointed out that England (not the whole of Great Britain) has been subjected to terrorist attacks for thirty years, and although the terrorists are loathed by most of the population, there is no panic; terrorists do not enter the risk portfolio. True, terrorists from Ireland are not as dangerous as holy warriors willing to choose to die in order to kill others, so the 'real' risk is less. But the necessary condition for entry into a risk portfolio is not met: the terrorists are not seen as pollutants. Interestingly, jailed (found guilty or simply detained) Irish terrorists went out of their way, perhaps unwittingly, to portray themselves as pollutants. They spent months rubbing their faeces on the walls, as if to cry out, yes, we are polluted, we are pollutants! Get worried about us; take us into your risk portfolio! It did not work.

Risk and Dirt

43

At first sight one might think that risk of terrorism is not grist for my mill, because the risks are distributable in the traditional way. Take out terrorism insurance. It won't save your building or your life, but it will cover the financial losses of destruction or death. But here the insurance companies moved in to prevent distribution of the risks of terrorism. They temporarily stopped selling terrorism insurance, and revised their coverage of terrorist acts. This decision was partly forced on them. Insurance companies commonly distribute their risk by taking out massive reinsurance policies to cover the unexpected catastrophic loss of one insurance company with premiums paid by more fortunate firms. Towards the end of 2001 all reinsurance in terrorism was blocked, and previously existing policies cancelled at the time of renewal. Thus thanks to purely contingent historical factors, such as the prudent self-interest of insurance companies and reinsurance syndicates, risk of terrorism was temporarily made a non-distributable risk. New Pollutions The Lele of the Congo had a fixed risk portfolio, only slowly adapting to interactions with other peoples, especially representatives of colonial power. It is a striking feature of our times that our risk portfolio is constantly in flux. We ought not to be blindsided by the terrorism mania, which we shall properly understand only a few years from now. Since 1969, there have been substantial changes in risk portfolios and what we might call the public structure of risks. Some of the reasons are straightforward. As I write we may or may not be terrified of the prospect of a local (?) war between the two nuclear powers of South Asia, but we are no longer worried, for the foreseeable future, about a U.S.-Soviet Armageddon. The real risks have changed from what they were in 1969 or 1981. New risk portfolios emerge in ways less definitively linked to changes in the course of events. Just before the conference in May 2001 on which this book is based, the National Post headlined the claim that half of all Canadians fear that their drinking water is unsafe - polluted by E-coli. This concern is a case of immediate overreaction to a much-publicized incident in which the corrupt water officers of a small town in Ontario systematically lied about their town water and their own actions, in the process killing seven people and making a great many others very ill. But that is just a passing overreaction, which is not to say there are no problems with water supply. Sales of bottled water continue to increase, but there are other odd, new social practices that mostly account for this, including the new ritual of going everywhere with a plastic bottle of

44 Ian Hacking water in hand, formerly bottled spring water, now, increasingly, tap water bottled by the Coca Cola company. By the end of 2001, not many people put unsafe water in their risk portfolios, not that much had changed with water inspection. Similarly, Americans returned to their unfriendly skies not so long after September. The most striking new long-term fear is fear for our children. I used to park my children in their baby carriage outside the supermarket, and never thought twice about it. No matter what I believe I could not do that with my baby grandchildren. Aside from resultant ostracism from my own children, now parents, I would be put in jail for negligence and abuse. When my wife and I were children, we were out on the streets alone all the time. Today children of the middle classes and the respectable poor are not allowed on the street alone. They must be walked or driven from a visit to a playmate, a late afternoon soccer match, a dance class, or a hockey game. This is a truly radical change in the life-world of children. It can also be totally destructive of parental life; yesterday a divorced mother of two children, aged ten and fifteen, said, near to tears, 'I have to have a break, I am on duty 24/7' (twenty-four hours, seven days a week). My mother never had that problem. This fear-for-our-children is an overwhelming addition to risk portfolios, and is above all a fear that our children will be defiled, subjected to unspeakable filth. The intense concern is not caused by new knowledge. We do not know that children are molested more commonly today than they were in the past. We have endured a period of great and by-andlarge valuable consciousness raising about child abuse, most of which occurs at home. This has undoubtedly alerted us to the problem of strangers in parks. But there is no evidence that molesters are more prevalent than before. Yes, more children are abducted than in days of yore, but most of the abductions are by parents, one divorcee abducting children assigned by law or fact to the other parent. So here I would find confirmation of the thesis I derived from Douglas, that pollution or a threat to purity is of cardinal importance in placing a danger in a risk portfolio. As I have insisted, it is not sufficient. In conversation with me, Nikolas Rose has protested that a far better explanation of the phenomenon of never-letting-your-child out-of-your sight has to do with the way in which our world has become anonymous our children are the final fundamental of our identity. I do not quarrel with Rose's insight. My thought is that the story of tying children to the figurative apron strings of mother and father illustrates once again the fact that when a new risk comes over the horizon, it is conceptualized as

Risk and Dirt 45 a pollution fear. This does not explain the emergence of the new risk, but rather predicts how it is experienced. It predicts its place and role in the portfolio. Mary Douglas did provide explanations, but not of why, contingently, a risk became a risk - which is exactly what Rose wants to do with his reference to anonymization of life. Douglas explains something quite different. She has a theory that uses the group/grid model to account for the ways in which different environmental groups evolve differently. Some hang together, some expand, some contract, while some split up into sub-sects. She identifies a necessary condition for a risk entering a portfolio of risks, but does not explain further which polluting risks are chosen. She does, however, offer an explanation of what happens when they are chosen. The New Risks Are at the Centre, While Many Pollutions on the Border Are Not Presented as Risks The new fears for our children are located in a place in society that is quite different from the position of the fears examined by Douglas and Wildavsky some twenty years ago. Those concerns about pollution were cultivated on the border. They served to hold activist cells, sects, and organizations together. But the fears of today are located squarely in the centre. That centre consists of responsible parents of all socio-economic strata. Aside from the all-too-common dysfunctional families, only someone wilfully occupying a position on the border will say to little Mary and Johnny, or even quite big Hilary and Madge, 'Stop pestering me, go over to the vacant lot five blocks away and play.' The arrangement of border and centre, with respect to pollution and risk, is substantially different in other ways as well. Take vegetarianism, a movement that in North America mostly thrives on borders. That is contingent. Whole civilizations are vegetarian, and some of the most loathsome leaders who made themselves the official centre were also vegetarian: Napoleon, Hitler, and their ilk. Nevertheless, in the here and now, vegetarians are on the border. They regard the eating of meat as pollution. But they do not thereby make it a risk. They are happy to cite today's medical folklore that eating red meat is bad for your heart, while lots of vegetables are good for your colon. But that is not the reason they are vegetarians. Those who call themselves deep vegetarians despise such arguments. They are vegetarians because eating meat is wrong, in part but only in part because of the mind-numbing cruelty with which edible meat is produced. They are stern moralists, subscribing to or

46 Ian Hacking going along with the beliefs of all but the most extreme animal rights activists. The intake of meat defiles a person. But even though eating meat is a manifest pollution taboo, it is not, except incidentally, a risk. These examples show how the topography of risk has changed in the two decades since Douglas and Wildavsky first wrote about it. Now we have sectarian groups built around pollution, but perceived risk and danger are not essential to them. That's modern vegetarianism. Conversely, we have pollution fears where risk and danger are constantly experienced, but sectarianism prevails. That's the apron-string model of North American child protection. Another food taboo, or at any rate restriction, is working itself out very differently in Europe from what has happened in North America. Members of the European Community have taken a fairly strong stand against the general and unfettered use of genetically modified (GM) food. One of the arguments is that seed from such crops will cross with nonmodified crops and overtake them: the seed is a wind-borne pollutant. Or the foods themselves will somehow damage our health; while not exactly a poison, they are like radiation, a possibly insidious unnoticed carrier of disease and death. In Europe an alliance has arisen between activists on the border and consumers in the centre. In North America, on the other hand, the battle against GM food never began in earnest. Opponents to genetic modification remain on the border, or in the wilderness. This returns us to the contingency of the actual choice of a risk portfolio, while leaving intact the thesis of the necessity of pollution for a risk to slot into a portfolio, and, when there, to seem almost an inevitable. References Arnauld, A., and P. Nicole. 1994. Logic, or the Art of Thinking, Containing, besides Common Rules, Several New Observations Appropriate for FormingJudgment. Trans. and ed. from the French of 1662 byj. Buroker. Cambridge: Cambridge University Press Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Douglas, M. 1966. Purity and Danger: An Analysis of the Concepts of Pollution and Taboo. London: Routledge and Kegan Paul - 1986. Risk Acceptability According to the Social Sciences. London: Routledge and Kegan Paul

Risk and Dirt 47 - 1993. In the Wilderness: The Doctrine of Defilement in the Book of Numbers. Sheffield: Journal for the Study of the New Testament, Supplement Series 158 - 1999. Leviticus as Literature. New York: Oxford University Press Douglas, M., and A. Wildavsky. 1982. Risk and Culture: An Essay on the Selection of Technological and Environmental Dangers. Berkeley: University of California Press Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press Fischoff, S. et al. 1981. Acceptable Risk. Cambridge: Cambridge University Press Gusfield,J. 1981. The Culture of Public Problems: Drinking-Driving and the Symbolic Order. Chicago: University of Chicago Press Hacking, I. 1982. 'Why Are You Scared?' (A review of Douglas and Wildavsky 1982, and Fischoff et al. 1981). New York Review of Books. 23 September: 30-2, 41-2 - 1983. 'Biopower and the Avalanche of Printed Numbers.' Culture and History: 279-95 - 1990. The Taming of Chance. Cambridge: Cambridge University Press Hampton, J. 1992. 'Correcting Harms versus Righting Wrongs: The Goal of Retribution.' UCLA Law Review 39: 1659-1702 LeDoux, V. 1998. The Emotional Brain: The Mysterious Underpinnings of Emotional Life. London: Weidenfeld and Nicolson Lianos, M., with M. Douglas. 2000. 'Dangerization and the End of Deviance.' British Journal of Criminology 40: 261-78 Okrent, D. 1980. 'Comment on Social Risk.' Science208: 372-5 Rosanvallon, P. 1998. The New Social Question: Rethinking the Welfare State. Trans. B. Harshaw. Princeton: Princeton University Press Rose, N. 1998. 'Governing Risky Individuals: the Role of Psychiatry in New Regimes of Control.' Psychiatry, Psychology and Law 3: 177-96 Starr, C. 1969. 'Social Benefit Versus Technological Risk.' Science 165: 1232-8 Wildavsky, A. 1984. The Nursing Father: Moses as a Political Leader. University, Ala.: University of Alabama Press.

3

The Rise of Risk DAVID GARLAND

Anthony Giddens likes to begin public lectures by posing some version of the following question: 'What do the following have in common? Mad cow disease; the troubles at Lloyd's Insurance; the Nick Leeson affair; genetically modified crops; global warming; the notion that red wine is good for you; anxieties about declining sperm counts' (Giddens 1999a, 1999b). The answer, of course, is that they are all about 'risk' and how it has come to permeate our lives and our politics. Each item on Giddens's list illustrates the impact of science and technology on our daily lives and material environment. Each one reflects the new tensions and ambivalence of our relationship to scientific expertise. Each one is an exemplar of what Giddens and Beck mean when they talk of 'the risk society' and claim that risk is a defining characteristic of the world in which we live. I would like to start this chapter by posing a somewhat different question. 'What do the following have in common? (a) the social theories of Ulrich Beck and Anthony Giddens; (b) Peter Bernstein's worldwide best-seller, Against the Gods; (c) the cultural anthropology of Mary Douglas and Aaron Wildavsky; (d) the Foucauldian analyses of power produced by Francois Ewald and Robert Castel; (e) the historical studies of tort and legal regulation of George Priest and Theodore Lowi; (f) the law and economics analyses of Guido Calabresi and Mark Geistfeld; (g) the normative jurisprudence of Steven Perry and Jules Coleman; (h) the welfare state studies of Peter Baldwin and Pierre Rosanvallon; (i) the prediction studies of John Monaghan and Peter Greenwood; (j) the "prospect theory" developed by psychologists Daniel Kahneman and Amos Tversky; (k) recent work in the sociology of insurance, of accounting, of governance, and of social control; (1) all of the chapters contained in this volume.'

The Rise of Risk 49 The short answer to my question is that they have nothing much in common. Each group of authors on this list is engaged with a separate set of questions and forms of inquiry. Each has its own research agenda, its own disciplinary affiliation, and its own distinctive object of analysis. The list could be a page from Borges's Chinese encyclopedia, with no obvious theme linking its oddly disparate items. Except, of course, that all the authors listed here write about something called 'risk.' I am struck by the fact that what is sometimes referred to as 'the risk literature' is in fact several distinct literatures, involving different projects, different forms of inquiry, and different conceptions of their subject matter, all linked tenuously together by a tantalizing, four-letter word that has, out of nowhere, come to stand centre stage in contemporary politics and social theory. The convergence of these various works on the concept of risk, in spite of their theoretical and disciplinary diversity, reinforces my impression that, suddenly, everyone seems to be talking about risk. If Raymond Williams were still alive and minded to update his classic account of the culture's Keywords (1983), the word 'risk' would be at the top of his list for new inclusions. The idea of risk has come to appear indispensable for understanding our times. As Giddens (2000: 39) puts it, 'this apparently simple notion unlocks some of the most basic characteristics of the world in which we now live.' Yet only ten or fifteen years ago, 'risk' had barely a marginal place in the vocabularies of social thought or cultural commentary and was rarely discussed outside of scientific journals and managerial reports. Today's accounts of risk are remarkable for their multiplicity and for the variety of senses they give to the term. Risk is a calculation. Risk is a commodity. Risk is a capital. Risk is a technique of government. Risk is objective and scientifically knowable. Risk is subjective and socially constructed. Risk is a problem, a threat, a source of insecurity. Risk is a pleasure, a thrill, a source of profit and freedom. Risk is the means whereby we colonize and control the future. 'Risk society' is our late modern world spinning out of control. Whatever one makes of these claims, it seems clear that risk and its management have outgrown the domain of the technical specialists and are becoming increasingly pervasive features of the contemporary world. Risk continues to be a major focus of scientific, economic, and managerial concern, but it has also become the subject of a whole variety of cultural, historical, and political inquiries, as well as a prominent theme of the social theories we generate to interpret our world. In this chapter I aim to do three things: (i) clarify the terms of discussion; (ii) speculate

50 David Garland about why so many intellectual and political currents nowadays converge around the idea of risk; and (iii) discuss the most influential account of the place of 'risk' in modern society - Ulrich Beck's 'risk society' thesis. What Is Risk? 'Risk' has a range of different meanings. In common, everyday usage it refers to 'the possibility of loss, injury, disadvantage or destruction' (Webster's 1971). Sometimes we characterize persons or things as 'risks' when they are prone to create hazards, or as being 'at risk' when they are more than usually vulnerable to being adversely affected by some problem or danger. We also use the word in its active verb form to mean 'to expose to hazard or danger,' a usage that reminds us that we sometimes engage rather than evade dangers in the world around us. 'Risk' has also acquired various more technical meanings in probability theory, in insurance contracts, in risk management manuals and, at least since Beck's best-selling book, in sociological theory. Risk and Danger In colloquial use, and sometimes even in theoretical texts (see Beck 1992 and Douglas 1992: 28) 'risk' and 'danger' are often taken to be the same thing, but it is helpful to distinguish the two terms and clarify their distinct meanings. A danger is a 'contingent evil' (Webster's), something that is liable to cause harm, injury, or adversity of some kind. A risk is the possibility of some such loss or injury. Risk is a measure of exposure to danger, of the likelihood and the extent of loss. So the ideas of risk and danger are closely connected, but clearly distinguishable. 'Danger' (or its synonyms, 'peril' or 'hazard') is the potential for harm that inheres in a thing, a person, or a situation. Risk is a measure of that potential's likelihood and extent. Put at its simplest, risks are estimates of the likely impact of dangers. Dangers and Our Relation to Them Neither risks nor dangers are simply brute facts in nature to which we have given a proper name, despite our tendency to talk about 'natural hazards' such as floods and lightning. In their central sense, both risk and danger have relational meanings. Each term defines a perceived

The Rise of Risk 51 relationship between ourselves and our world. Each involves a vision of what might happen and how it might affect us. There are countless things in our environment that have the potential to do harm, cause damage, or bring about economic loss. Meteorites and volcanoes, toxic emissions and speeding cars, market crashes and urban riots, sex offenders and abusive parents, collapsing demand and overheating economies are all familiar examples. Some of these things are generated by human activity and are amenable to social control. Some are not, or at least not yet. Whether 'man-made' or 'natural' (a distinction that is increasingly hard to sustain) most of these things exist in the world, and continue to do so whether or not we worry about them.1 They become 'hazards' or 'dangers' only when they relate to us in ways that might adversely affect us or our interests. Dangers are dangers for someone - for specific individuals or groups or species, under certain conditions - nothing is dangerous as such, not even floods and lightning. On the other hand, anything and everything has the potential to become a danger to something or someone. All that is required is that there are interests or values that the thing may adversely affect. The balance between humanly produced hazards and naturally occurring ones has shifted towards the former, as our capacity to act upon the world (and upon each other) has increased from the industrial revolution to the present day. A potential for unwanted side effects and dangerous repercussions is a by-product of the extended scope and increased interdependency of human action. The increase in our powers and technological capability brings with it new possibilities for the misuse or faulty deployment of that increased power. Increases in the range of human actions, investments, and interests will tend to increase our exposure to loss, harm, and hazard. Efforts to combat dangers, reduce our exposure, or remedy harmful externalities have tended to lag behind technological and economic innovation. Today we seek to 'design in' safety features that anticipate risk, but for much of human history, damage limitation has been 'retrofitted' after the fact. Risks and Uncertainty

Potentially dangerous things become 'risks' only when we assess the likelihood that these adverse effects will indeed occur and weigh the harms that they are liable to bring about. Risks are not 'material' things

52 David Garland in the nineteenth-century sense of that term. Risks are estimations of possible events. Unlike dangers or hazards, risks never exist outside of our knowledge of them. They are the product of future-oriented human calculations - assessments made by people in the face of an uncertain world and the possibilities that it holds for them. Risk always exists in the context of uncertainty. Where there is certainty about an event, where we know for certain that it will or will not occur, we do not talk of risks. Risk begins where certain knowledge ends. Claims about risk are, literally, uncertain knowledge claims - impressionistic guesses, informed estimates, and probabilistic predictions about a future that cannot fully be known. As John Adams (1995: 25) puts it, paraphrasing Frank Knight's classic work, Risk, Uncertainty and Profit ([1921] 1965), 'if you don't know for sure what will happen, but you know the odds, that's risk, and if you don't even know the odds, that's uncertainty.' Statements about risk are claims to know something about a future that can only be 'known' in probabilistic terms. In an important sense then, a risk is not a first order 'thing' existing in the world. It is rather a specific way of assessing and categorizing the (hazardous) relationship that these things have to us, to our plans, our interests, and our well-being. The number of risks that exist at any time is a function of the number of risk identifications and assessments that human beings make. (Dangers are a different matter.) If we are becoming a 'risk society,' as Beck (1992) and others have claimed, then this is, in an important sense, a development for which we (or some of us) are directly and actively responsible. If our orientation towards life's dangers were suddenly to shift from active concern to fatalistic acceptance, our world would be no less hazardous but the risk society would disappear. The idea of an unforeseen hazard makes sense, at least in retrospect there may be dangers around the corner of which we are unaware, and familiar phenomena may contain an unknown potential to do us harm. But there is no such thing as an unforeseen risk, since, properly speaking, something becomes a 'risk' only to the extent that its potential for adverse consequences has been brought to notice and subjected to some kind of estimation, however rough and ready. One might go further, as Adams does, and say that there is no such thing as an unmanaged risk, because as soon as a risk is identified, as soon as it is noticed, we tend to take steps to manage or reduce its adverse consequences. '[RJisk perceived is risk acted upon. It changes in the twinkling of an eye as the eye lights upon it' (Adams 1995: 30). Risk, risk perception, and risk management are thus interdependent, interactive, and mutually constituting.

The Rise of Risk

53

This mutuality should be reassuring. The modern world may be crammed full of risks and dangers, but, as we will see, it is also populated by practices and institutions designed to identify and manage risks, reduce uncertainty, and cope with danger. Risks Are Conditional Risks are relationships of possible adversity, calculated and assessed by someone for some specific purpose using some specific means. The idea of a risk is therefore thoroughly conditional. A risk is always a risk of something, for someone, estimated for a certain exposure (specified in terms of time, intensity, or amount), and calculated using specific units and instruments of measurement. 'Risks' are stated measures of possible hazard, usually expressed in some basic metric such as monetary values or fatality rates. Such measures are calculated by multiplying the probability of the adverse event by the amount of the loss that such an event would entail: as Webster's dictionary defines it, risk is 'the product of the amount that may be lost and the probability of losing it.' Measurements of risk can be more or less accurate, more or less valid, more or less scientific. That risks are conditional does not detract from the possibility of accurate measurement within defined parameters. That they are 'socially constructed' should be obvious, though quite how this social construction works is a central topic of sociological, economic, and psychological research. The important practical question is whether they are well or badly constructed as estimates of the event in question. Risks Are Reactive Our capacity to identify, evaluate, and manage uncertain future events is intrinsically limited, even where there is good data, good science, and a settled perception of the hazard in question. Past experience is our best predictor of future events, especially when that experience can be quantified, understood, and statistically analysed. But even at its best the past is an unreliable guide. Extrapolations from past experience are always inferences from a limited data set using premises (about cause and effect, about the factors involved, about ceteris paribus) that may be disproved by subsequent events. And if this is true of 'natural hazards' such as meteorites, volcanoes, and earthquakes - that are largely unresponsive to human action, it is truer still when the hazard in question is 'conduct-related,' or somehow affected by human behaviour.

54 David Garland Many of the conduct-related risks that we seek to manage - fire and financial loss, crime and cancer, early death and late delivery - are altered as soon as we identify them as such. Research on skin cancer prevalence and etiology changes our sunbathing patterns and so alters rates and risks. Reports that crime is rife in a particular neighbourhood prompt law-abiding people to stay away and thus increase the risks of victimization for those who remain. Householders who worry about burglary become more vigilant arid thus reduce the risk that they worry about. In all these cases, risk is 'reactive' (Heimer 1985) - it responds to the attitudes and actions that people adopt towards it. Risks Are Continually Calculated and Compensated If the worried householders decide to take out insurance, preferring to pay a small regular premium rather than run the risk of a large, uncompensated loss, they may relax their former vigilance and increase the risk to its previous level. Insurers refer to this phenomenon as 'moral hazard' and seek to combat it by imposing minimum care requirements in order to reduce its effects. (See Heimer, this volume, on insurance as discipline.) 'Moral hazard' describes the temptations to bad behaviour (false claims, carelessness, wilful damage, etc.) that the promise of compensation can produce for an insured party. But the underlying phenomenon to which this points is the tendency of actors continually to weigh the costs and benefits of their situation and behave accordingly. Drivers who speed along at 80 mph may drive even faster if their car is subsequently fitted with seat belts or anti-lock brakes. Otherwise cautious skiers will chance steep slopes in the company of an instructor that they would never attempt on their own. Investors who have profited from a bull market may buy high-risk stocks that they would previously have avoided. This tendency to adjust behaviour in response to changing perceptions of risk is usually called 'risk compensation.' Safety officials may strive to reduce risks towards zero, but most individuals find a level of risk with which they are comfortable and stick to it, even as their environment changes. As a result, measures intended to create safer roads may be offset by the increased recklessness of drivers who now feel less at risk. Railway tracks that become markedly dangerous may prompt rail companies to impose speed restrictions and so produce fewer accidents. Airlines are one of the 'safest' forms of transport, but a moment's reflection reveals that this is because of the meticulous standards of

The Rise of Risk

55

care that have been put in place, not the absence of danger involved in flying at 30,000 feet. Risk compensation means that as perceived risks alter, so too do our levels of vigilance and exposure. The dynamically interactive process that results often confounds attempts to measure risk or to increase safety. Risks Are Interactive If it is true that actors perform a kind of balancing act as they constantly adjust their conduct in response to the changing hazards in their environment, it is equally true that individuals differ in respect of where they set that balance. Each of us has our own 'risk thermostat' (Adams 1995) - our own comfort level in the face of particular risks and particular kinds of risks. Individuals vary in their propensity to take risks - whether as result of personal dispositions, cultural bias, past experience, or available sources of security - and so do institutions. Some of us are cautiously risk averse, others are reckless risk takers, and we necessarily act in the company of others whose behaviour we do not control. Drive in rush-hour traffic and you'll get the idea. Unless they live on a desert island, individuals and institutions are usually responding to hazards in company with other individuals and institutions whose collective decisions and conduct will affect the risk that any one of them will face. The risks we run depend on the actions of others and the risks they take. Steering clear of dangerous drivers is the simplest example, but there are more indirect ways in which the risk perceptions of others affect the risk you face. If the people on your street purchase security alarms for their houses, you had better do so too, lest you become the easiest target for burglars. If the property market is going through the roof, anyone who fails to buy a house risks being left behind by the prosperity of others. It is this complex interaction between multiple, reflexive actors and particular reactive risks that John Adams has in mind when he offers the image of 'the dance of the risk thermostats.' Game theory makes the same point more prosaically when it says that 'the true source of uncertainty lies in the intentions of others' (Bernstein 1996: 232) .2 Risk is, in short, profoundly interactive, which is to say, profoundly social. The old natural science idea of 'objectively' measuring 'actual' risks is increasingly being redefined by a more complex, more social, understanding of the processes involved (Royal Society 1992).

56 David Garland 'Objective Risk' and Perceived Risk When people talk of 'objective risk' they mean a risk that has been scientifically established using the best available data and knowledge, as opposed to 'perceived risk,' which is based upon merely 'subjective' impressions. Authorities often respond to public anxiety by seeking to ascertain the 'objective' or 'actual' risk of a particular hazard (such as road accidents, HIV infection, or criminal victimization) and using this measure in attempts to persuade citizens to alter their attitudes. The space between expert-defined measures of 'objective risk' and the public's 'subjective' impressions is a problematic one, particularly if the conduct of governments or members of the public bears upon the risk in question. Debates about 'actual risk' and its implications for action are a recurring theme - perhaps the central theme - of contemporary politics. Whether we accept the idea of 'objective' risk and risk measurements depends on whether we believe any conventions of categorizing and counting are sufficiently well-established, widespread, and uncontested to deserve that name. But even when strong conventions do exist, the notion of 'objectivity' tends to cloud rather than clarify the issues at stake. Like all measurements, risk assessments depend for their validity upon a prior system of categorizations and metrics, which are in turn, grounded in specific conventions, institutions, or ways of life. Natural science, and organizations like the BMA, the Royal Society, or the National Academy of Science, are powerful institutions but their assumptions are not beyond challenge, nor are they devoid of contestable judgments. Objective versus subjective risk is a false opposition. The contrast is more often between different conventions for observation, measurement, and evaluation. As Fischoff says, 'what is commonly called the conflict between actual and perceived risk is better thought of as the conflict between two sets of risk perceptions: those of ranking scientists performing within their field of expertise and those of everybody else' (quoted in Royal Society 1992: 97). Because risk statements carry consequences, the representation of risk is subject to political manipulation and tendentious presentation. The background condition of uncertainty, limited knowledge, and differential evaluation often makes it impossible to separate out 'factual' claims from politically loaded perceptions. When risks are politically contentious, it is often interpretation all the way down.

The Rise of Risk

57

Risk and Public Perception

Governments and scientists have learned that the risk perceptions of ordinary people are not easily transformed by reference to 'objective' measures of the risk in question. Subjective impressions can be strong, scientific claims may be regarded with scepticism, and individuals often place greater trust in their own experience, or that of friends and neighbours, than in the claims of government experts (Adams, this volume). More importantly, any assessment of risk entails an assessment of consequences as well as of likelihood. People who worry about becoming victims of crime, cancer, or nuclear fall-out are expressing personal views about the event's (dis) utility as well as its probability. They are making emotionally laden value judgments as well as cognitive claims. And while scientific evidence may counter misleading estimates of the likelihood of an event, it can do little to change public views about the undesirability of its occurrence. Democratic governments have learned that they must listen to public concerns of this kind, rather than brush them aside as 'uninformed,' 'irrational,' and 'unscientific.' As the Royal Society Report (1992: 91) states, 'the public's viewpoint must be considered not as error but as an essential datum.' Or, as Beck (1992: 77) puts it, Tf people experience risks as real they are real in their consequences.' The challenge for democratic politics is to create the conditions for informed debates and decision making in which citizens as well as specialists can participate effectively (Beck 1992; Beck, Giddens, and Lash 1994; Pildes and Sunstein 1995; Franklin 1998). Risk Communication

Governments and corporations are learning that risk communication is itself a risky business, prone to producing counter-productive effects and unintended consequences (Nelkin 1984). Efforts to inform the public about the estimated risks associated with identified hazards must be carefully judged to avoid alarmism on the one hand and complacency on the other. Hazardous situations are easily made worse by official advice that is poorly judged or misunderstood, so there is an understandable reluctance to 'go public' when possible new risks are first identified. But public officials are under great pressure to provide full and accurate information as soon as possible, and will be held to account if they fail to do so. Hence the modern predicament - authorities must disclose the

58 David Garland facts, clearly and comprehensively, even when they are themselves unsure of what these facts are. The difficulty of making critical decisions in conditions of uncertainty is thus exacerbated by the difficulties of representing these uncertainties to the public, particularly a public untutored in the interpretation of risk measures and prone to unrealistic expectations about the character of scientific knowledge.3 Research suggests that there are systematic biases in the public's perceptions of risks, with most people being prone to overestimate the risk of low frequency, high magnitude events (catastrophes such as nuclear explosions or airplane disasters) and underestimate the risk of high frequency, low magnitude events (routine accidents in the home, road accidents, etc.) (Royal Society 1992). Such biases ensure that 'rational' choices and politically feasible choices often fail to coincide. But the most basic problem today is the problem of trust, and the public's relation to 'the authorities.' Sections of the public have come to distrust government officials and large-scale organizations, believing that their interests and motives do not match their own. Similarly, the claims of 'scientific experts' are often regarded with some scepticism, particularly if other experts offer competing advice. Whether this new distrust reflects unrealistic expectations on the part of the public, the poor safety records of particular organizations, or, as I will discuss later, because of underlying changes in social organization, the result is that government officials and scientific experts can no longer assume that their advice will be regarded as authoritative. Risk Experts and Institutions

Risk assessment and risk management are activities that all mature human beings constantly undertake in their daily lives, often without much conscious reflection. We all use know-how and information (of varying quality) to address the hazards that confront us - sometimes trying to steer safely around them, sometimes engaging with them for pleasure and profit, most often trying to balance risks against rewards and trade off the effectiveness of precautions against the effort of taking them. Common-sense knowledge about how to handle risk draws upon all sorts of superstitions, practical recipes, and folk wisdom but in modern societies it also relies heavily upon experts, scientific knowledge, and risk-managing institutions specially designed for this purpose. This relationship between lay actors and scientific expertise is a central characteristic of modern society, where expert opinion on topics

The Rise of Risk 59 such as diet, health, relationships, finance, and investment quickly filter down into public consciousness and the daily routines of individuals (Giddens 1991). The immediacy of this relationship between lay and expert knowledge, and the extent to which daily conduct is now governed by reflexive knowledge rather than habit and tradition, has ensured that science is now in the public view much more than was previously the case - with consequences for the credibility of scientific expertise. Dealing with risk may be a routine task for everyone, but in modern societies this reflexive attitude to the world's dangers has given rise to a whole series of occupational specialisms whose task is the formal measurement and management of specific risks. All manner of specialists civil engineers, chemists, toxicologists, epidemiologists, traffic planners, security experts, military intelligence, actuaries and financial analysts, fire officers, psychiatrists, social workers, criminologists, health and safety officers, and countless others - are employed to identify, analyse, and control specific risks. For these experts, working in diverse industries and areas of government, the management of risk is a socio-technical problem, governed by scientific protocols as well as the needs of the organization in which they are employed. The Government of Risk The question of who is responsible for regulating risk is a fundamental and recurring one. Should governments be responsible for the safety of their citizens and for managing the risks that regularly affect them? Should individuals manage their own risks wherever possible, relying upon private insurance, tort law, and simple prudence? Should corporations be legally obliged to ensure the safety of their employees, customers, and stakeholders? Or is the market an adequate mechanism for delivering the levels of safety and quality that consumers prefer? If government regulation is required, what kind of regulatory regime is most appropriate? These questions frame many of today's political and legal issues, ranging from environmental policy and pollution control to employment law and welfare state benefits. In practice, responsibility for risk management is parcelled out among individuals, corporations, and government agencies, with the particular distribution of responsibilities being a rough indicator of the political and economic structure of the society in question. In more welfarist societies, the state tends to act as the general risk manager, governing risk by means of extensive regulation, standard-setting, and inspection,

60 David Garland and operating as an insurer of last resort for otherwise uncompensated losses. In more free-market societies, individuals bear more of the burden of risk management, relying upon private insurance or private litigation to compensate losses caused by the actions of others. The long-term historical trend in modern societies is for governments to become ever more responsible for risk management, governing risks by means of statutory regulation, legal norms and standards, compulsory insurance, and the provision of benefits to those in need (Friedman 1981). Risk management - usually known by some other name - has always been a basic function of government, if by this one means the securing of internal peace, law and order, and defence against foreign invasion (Lowi 1990). And at least since the early modern period, the 'police powers' of the state have extended to rudimentary sanitation, containing epidemics, controlling idle workers, and regulating weights and measures (Foucault 1991). But in modern states, of whatever stripe, the expert management of risk has become an essential task of government that reaches into practically every domain: from the regulation of food and drugs to manufacturing standards; from health and safety at work to environmental regulation; from social insurance and economic policy to the legal norms of tort and civil liability.4 By means of their actions and inactions, governments everywhere allocate risks, distribute dangers, and place costs, in particular political configurations. But the long-term trend has been for governments to assume more and more of the job of risk management themselves. Freedom and Security: Insurance as a Technique of Government Ever since the end of the nineteenth century, insurance in its various forms has become an important element of social and economic policy. Governments have used its techniques to secure citizens against social and economic risks, to reduce social conflict, and to enhance economic performance. These insurance arrangements - which include workers' compensation, old age pensions, unemployment insurance, family allowances, and so on - illustrate what Michel Foucault (1991) means when he talks of the importance of 'apparatuses of security' for the 'liberal' mode of government. Insurance is an important tool of government in modern, liberal societies because it preserves the free play of autonomous action within the economic and social spheres (allowing individual decisions about work, marriage, childbearing, purchase and sale, investment, etc.) while adding a safety net that removes some of the risks associated with these freedoms. Thanks to the operation of statisti-

The Rise of Risk 61 cal laws and probabilities, populations can be governed through insurance in a way that leaves scope for individual freedom but ensures a level of security and stability for the whole (Hacking 1990). This need for mechanisms of security to underpin the exercise of freedom is also apparent in the build-up of regulatory law that has occurred over the same period. The historical forces that produced this risk-governing state are many and various, but a major reason is the need to generate trust and confidence in everyday processes and products that are increasingly beyond the control and understanding of their users (Giddens 1990, 1991). Our willingness to drink the water in our taps, to swallow prescribed medications, to board transatlantic airplanes, or to allow the siting of a nuclear reactor in our region, depends upon our being confident that the systems that deliver these products are properly regulated and meet acceptable safety standards. Governmental authorities seek to provide a basis for such trust by generating a framework of regulation and inspection intended to ensure that health and safety standards are met. Where these regulatory schemes work well, they allow us to take routine risks without pausing to think about them (Manning 1989), thus greatly facilitating the flow of social and economic life. The Welfare State Is a Risk Management State

The modern state is a welfare state, a regulatory state, an insurance state.5 It insures its citizens, indemnifies them against losses, regulates economic risks and environmental dangers, protects individuals from social harm and economic disaster. Its basic tools are those classic techniques of risk management: statistical enumeration, insurance, discipline, regulation, standardization, norm-setting, and inspection (Hacking 1990; Ewald 1991). Most of its specialist agencies - in social work, criminal justice, mental health, environmental health - have as their primary function the management and reduction of particular kinds of risk. Alongside these statutory agencies and regulatory frameworks, a system of consumer law, tort law, and product liability law has emerged that functions not so much to find fault and allocate blame but rather to distribute risk in patterns that are efficient as well as fair (Priest 1990; Lowi 1990) .6 Behind this new distribution of social liability for harms are new conceptions of responsibility, new relations between social groups, and new techniques of insurance, all of which converged at the end of the nineteenth century to produce a distinctively 'social' mode of managing risk and promoting solidarity (Donzelot 1979; Ewald 1991). Social and economic historians have debated the social basis of the

62 David Garland welfare state, with some pointing to the working classes as the key actors and others identifying the needs of capital and a more top-down process of change. But the comparative evidence does not fully support either of these standard interpretations. As the work of Baldwin (1990) and Rosanvallon (2000) suggests, thinking of the welfare state as a riskmanagement state shifts our attention away from conflicts over the means of production and towards conflicts over the means of security. In this analysis, the key historical actors are not so much social classes as risk categories - social groups defined by their relationship to a particular policy such as social security, old age pensions, or health care - and the social policies that came to be established in different nations were the ones that worked to serve the interests of the most powerful risk groupings. The contrast between support for old age pension schemes (for which all adults are 'at risk') and support for income support schemes (which benefit only the most needy) demonstrates the power of riskdefined interests to shape political behaviour. Neo-liberalism's Re-allocation of Risk

Over the last two decades, neo-liberal governments have sought to move away from the classic post-war model of the risk-managing state (Yergin and Stanislaw 1998). Arguing that the insurance state gives rise to its own version of moral hazard — the 'culture of dependency,' the erosion of individual responsibility, the decline of the entrepreneurial spirit, a 'norisk society' (Aharoni 1981) - they have deregulated markets and financial institutions, emphasized free enterprise, and taken steps to shift risk and responsibility onto individuals (Rose 1996; Barry, Osborne, and Rose 1996; Ericson, Barry, and Doyle 2000) .7 This return to the market is driven, in part at least, by highly paid sections of the population for whom state-provided benefits (such as pensions, education, health care, income support) are no longer essential and no longer adequate. In effect, certain demographic groups have calculated that they stand to gain little from the welfare state insurance scheme and now prefer to pay reduced taxes and take their chances in the private market. They perceive themselves as risk categories for whom state insurance demands premiums that are too high and offers benefits that are too low, so they look to private insurance and investment as more effective ways of managing their risk. (Rosanvallon 2000) .8 Under the banner of market freedom and individual responsibility, the better risks depart the state insurance scheme, leaving an impoverished collective fund to deal with

The Rise of Risk 63 their more vulnerable, higher risk fellow citizens. Ulrich Beck (1992) stresses the importance of environmental risk in contemporary politics and claims that the economic conflicts associated with 'industrial society' have tended to fade. In fact, it seems clear that debates about economic risk and its management still stand centre stage of domestic politics. Risk and Morality

Choices concerning the allocation of risk involve moral-political choices (Perry 1996). Social security provisions, worker protections, consumer laws, tort law principles, environmental regulations, criminal law, insurance contracts - all of these carry moral components insofar as they distribute risks and compensations according to some scheme of fairness and efficiency.9 Risk can be managed in a whole variety of ways, ranging from deterrent approaches that hold individuals strictly liable for harms and the payment of damages, to 'no-fault' systems that collectivize risks, privilege prevention, and draw compensation from general funds. The Royal Society (1992: 159) describes seven distinctive approaches, each of which allocates responsibilities, risks, and incentives to different actors according to different principles. When we talk of 'risk management' as a mode of governance, we are generally referring to a distinctive form of practical morality that is quite at odds with traditional forms of moral and judicial reasoning. Instead of going in for post hoc blame allocation and holding the individual actor fully responsible, a risk management approach tends towards a more structural account of responsibility and is less concerned with fixing blame or imposing penalties. Loss prevention, harm reduction, and efficient compensation are, in many circumstances, more easily achieved by cooperative action and shared responsibility than by individualistic moral codes. One sees this in modern civil law (and especially products liability), where there has been a historical shift from fault to risk, from blame-worthy individuals to risk-spreading corporations (Priest 1990; Lowi 1990) .10 And while some critics see this as the abolition of responsibility, it is, in reality, an attempt to create a more complex, more social, imposition of duties and responsibilities - one that is more appropriate for a complex social system. The language of risk might be thought of as the self-evident moral discourse of an interdependent differentiated society. It is the liberal harm principle raised to a new social level. The imperative to 'do no harm' now takes on a more social, more complex aspect, since our

64 David Garland actions typically affect so many more people. As chains of interdependence grow longer, and shared moral codes grow thinner, risk management has become a necessary moral technology, operationalizing liberalism's twin concern to maximize freedom of action and to reduce that freedom's harmful consequences. The Risk Industries

If governments today are in the risk-management business, so too is the private sector. Some of today's largest commercial enterprises are, in effect, risk industries, trading in risk and its management. The insurance industry is a prime example (see Heimer, Ericson, O'Malley, and Baker, this volume). Insurance offers products that protect against individual calamity by collectivizing risks and rendering them predictable. Insurance pools the risks faced by large populations of clients, so that each individual bears only the cost of the insurance premium and need no longer worry about additional loss. Freud once said that psychoanalysis is not a cure for life's suffering, but rather a means of replacing neurotic suffering with common unhappiness. In the same way, insurance is no miracle cure for life's dangers. What it offers the individual is the possibility of dealing with these hazards in advance, and at an averaged cost, by paying a standardized premium. Insurance stabilizes risk and renders its costs predictable - it does not do away with risk altogether. With the help of statistical tables that establish the regularity of certain events, actuaries can ascertain the probability that such events will affect particular classes of persons. As Ewald (1991: 202) says, 'When put in the context of a population, the accident which taken on its own seems random and avoidable ... can be treated as predictable and calculable.' For the insurance industry, risks are opportunities for business, raw materials to be purchased, traded, and transformed into profits and capital. Since its emergence in the seventeenth-century seaports, commercial insurance has expanded to become a pervasive institution that has transformed the risk environment in which modern life proceeds, with deep social consequences.11 Gambling is another 'risk industry' that has become a major part of contemporary economies (see Reith 1999), one that is a mirror image of insurance. Like insurance it trades in risks, pools losses, and makes profits out of closely calculated chance events. But whereas the insurance industry depends upon risk aversion and anxiety about the pros-

The Rise of Risk 65 pect of loss, the gaming industry thrives on the embrace of risk and the possibility of gain. In the world of casinos, lotteries, off-track betting, and football pools, the risks of losing and winning are transformed into exciting, pleasurable experiences (Skolnick 1978; Reith 1999). Much of gambling's appeal lies not in the promise of big winnings (the odds of which are usually remote) but in the thrill of playing the odds, taking risks, courting danger. Like extreme sports (Simon 2002) and venture holidays, the gaming industry trades on the willingness of individuals to embrace risks for the intrinsic pleasure of doing so. Finally, one should certainly include here the biggest risk industry of all, the financial sector and the various institutions of which it is composed - stock markets, investment banking, the insurance industry, finance companies, pension funds, and so forth. All capital investment involves risks: any producer of goods and services must rely on forecasts about market demand, product competitiveness, production and distribution efficiency, and so on. But for finance capital, risk is the specific commodity in which it deals. Risks, futures, anticipated returns, managed uncertainty - this is the very stuff of the industry, the raw material out of which it carves its considerable profits. This secondary form of commerce - it does not produce goods and services, but underwrites the risks of those who do - has developed sophisticated techniques of risk forecasting and management as well as sophisticated 'risk-hedging' instruments such as derivatives, options, and swaps. Little surprise then, that the financial sector has been a driving force in the computerized applications of probability theory and the formation of massive data sets (Bernstein 1996; Dunbar 2001). Risk and the Responsible Corporation Companies and organizations are today becoming increasingly selfconscious and reflective about the risks that they face (Power, this volume) . As well as the risks of monetary loss that commercial companies run whenever they enter the market, all organizations generate risks as a by-product of their activities. Any firm in any business is capable of mismanaging its resources and exposing them to unnecessary risks, or alternatively, of wasting resources by taking too many precautions (e.g., purchasing too much insurance, or devising overly cautious investment strategies). Any firm is capable of doing unintended harm to its customers, employees, or stakeholders, and liable to be sued or to lose profits as a consequence. This is true of newspaper proprietors who can be sued

66 David Garland for libel, physicians who face malpractice suits, or lawyers and accountants who misrepresent their clients, but it is a particular problem for companies that manufacture food, medicines, or pharmaceuticals, where a whole host of health regulations must be met before products can get to market. Similarly, for manufacturers of products where user safety is a major issue (which is to say any manufacturer, but especially the makers of motor vehicles, airplanes, cigarettes, and guns) their products will be tested against industry standards and subjected to government inspection designed to monitor, measure, and minimize risk. The willingness of courts to extend the traditional principles of tort law in products liability cases in ways that increase the duty of care owed by manufacturers to customers has done much to hasten these developments (Priest 1990; Lowi 1990). Health and safety, the avoidance of legal liability, and the management of financial risk are now essential components of any substantial business undertaking. Every responsible company or institution now has a risk management office charged with the task of putting systems in place that will ensure safety procedures are observed, minimize legal liability, and safeguard the company against harms to its reputation or goodwill (Power, this volume). The recent spread of 'managerialism' - with its emphasis upon systems management, performance indicators, and audits - into the public sector has ensured that risk management is as much a feature of institutions like local government, universities, and hospitals as it is of commercial corporations. Institutional Orientations to Risk

Psychological research has developed tools to measure the risk propensity of individuals, and the factors that affect these propensities (Kahneman and Tversky 1979; Wildavsky and Drake 1990; Royal Society 1992: chap. 5). But our institutions for dealing with risk can also be regarded as risk averse or risk embracing. In the risk management literature of health and safety, risks are typically viewed as a problem to be minimized, an evil to be eradicated. Similarly, in social services, criminal justice, or health care, where individuals and communities are characterized as 'at risk' (of social problems, crime, or ill health), the overarching aim is to eliminate risk wherever possible. Very much the same approach emerges in the environmental literature and in Beck's book, The Risk Society, where risks are presented as overwhelmingly negative in their import - as events to be avoided, catastrophes to be contained.

The Rise of Risk 67 In this way of thinking, risk is an evil with no redeeming social value and so there is no such thing as being too careful. One sees this attitude vividly embodied in versions of the 'precautionary principle' that opposes any undertaking that entails the risk of environmental damage, or in current attitudes towards the release from custody of certain kinds of criminals (above all, sex offenders) where any level of risk, however minimal, is deemed unacceptable. Today's common usages often make us think of risk as danger, as a bad thing to be avoided. We don't talk of people being 'at risk' of winning the lottery, or about high intelligence being a 'risk factor' that predicts university entrance. Negatives such as ill health, social problems, or large-scale disasters have dominated our discourse about risk to the exclusion of a more balanced, more realistic, approach. This zero-tolerance approach may be embedded in certain institutions, but it is a poor guide to our general attitudes towards risk and risk taking. As Adams points out, the figure of homo prudens - zero risk man is a figment of the risk manager's imagination rather than a description of characteristic human attitudes. Adventurous, enterprising individuals are drawn to risk. They embrace it, deriving excitement from the prospect of uncertain outcomes, affirming their autonomy by tackling danger or experiencing the unknown (see Simon, this volume; Baker and Simon 2002). Even the least venturesome among us are prepared to take certain risks, particularly if we assume the risks voluntarily, if we can, to some degree, control the outcomes and the risk involved, and if our gambling carries the chance of some reward. Taking risks, acting boldly, refusing to follow the dictates of prudent rules and statistical averages are, for many people, a measure of their agency, their will, their essential subjectivity (Baker and Simon 2002). Enterprise and the Embrace of Risk

One way of countering overly negative discussions of risk is to bring economics into the discussion. According to the Oxford English Dictionary, a secondary definition of risk is 'the chance that is accepted in economic enterprise and considered the source of (an entrepreneur's) profit.' The foundation of the economic system that dominates the Western world production for exchange in the market in pursuit of profit - is built upon the willingness of investors and entrepreneurs to take risks with their assets in the uncertain expectation of financial reward in the future. As Bernstein (1996: 3) puts it, 'the capacity to manage risk, and

68 David Garland with it, the appetite to take risk and make forward-looking choices, are key elements of the energy that drives the economic system forward.' No risk, no profit, no capitalism. In that sense, risk and its management have been central to our social arrangements ever since modern capitalism came to dominance. When economists or business people talk about 'managing risk' they do not mean eliminating it entirely. In a competitive market economy, zero risk is a formula for zero returns. Managing risk means steering it, controlling it, minimizing its detrimental effects while making the most of its positive potential. Control of this kind is possible because the uncertainties faced by firms are not truly random events but instead patterns of human action (demand levels, consumer attitudes, exchange rates, production costs ...) that can often be predicted with some degree of accuracy. The essence of risk management lies in maximizing areas where knowledge - and hence control - are possible, while avoiding areas that are less known and less predictable (Bernstein 1996: 197). Consequently, the capacity to make accurate risk estimates and forecasts is a major commercial asset that companies nurture and develop. The spectacular earnings made by day traders and dot.com start-ups in the 1990s, and the equally spectacular collapse of many of these seat-ofthe-pants entrepreneurs, illustrate the rewards of taking big risks and the penalties for failing to manage them effectively. It also illustrates the role of time and regression towards the mean, both of which shape the odds of any ongoing enterprise. Risks and Benefits Reference to economic issues brings out a point that is often overlooked in areas such as criminal justice, where public safety considerations dominate discussion. The point is simply that risks are usually a corollary of activities or decisions that are otherwise beneficial in their consequences. Many of the risks inherent in modern life - from car crashes to global warming, from air pollution to obesity - grow out of activities that bring important benefits to individuals, communities, and national economies. In an important sense, risk is the necessary accompaniment of freedom and choice. Where there are no choices to be made there are no risks to run. Fatalism and determinism exclude risk calculation. Choice, on the other hand, involves weighing options. The expansion of choice is necessarily the expansion of risk - if only the risk of making the suboptimal choice.

The Rise of Risk

69

Thinking about risks must always include thinking about the benefits attached to the risky activity, as well as the costs (including opportunity costs) of reducing these risks by restricting the activity in question or the choices of the actors involved. The best risk management strategies are those that reduce risks in ways that have fewest consequences for the desired activity to which the risks attach. It is for these reasons that costbenefit analysis of risks, rather than risk elimination, is the style of thinking found in most areas of government and industry today. Thinking about Risk As I noted earlier, risk used to be something of a technical topic. It was studied chiefly by scientific and managerial experts concerned to identify, measure, and manage the specific risks that fall within their purview, as well as by psychologists, economists, and decision theorists oriented towards the practical science of risk management. In recent years, however, the study of risk has burst out from these technical confines and become a major topic in the social sciences and in cultural commentary. It is not surprising that social scientists should find risk an attractive object of inquiry. Precisely because of its conditionality - risk is always a risk of something for someone - and its reliance upon conventions of perception, judgment, and measurement, 'risk' is a thoroughly social, thoroughly cultural, thoroughly psychological phenomenon. The selection of risks to worry about, the judgments that are made about them, the distribution of responsibility for their management, and the methods employed to deal with them vary greatly, among individuals, among groups, and among cultures. As Mary Douglas and her colleagues have shown, the risks that we identify 'out there' reveal as much about us about our psychological traits, our cultural biases, our structures of perception, and our institutional affiliations - as they do about the hazards and contingencies in our environment (Douglas 1992; Douglas and Wildavsky 1982; Thompson, Ellis, and Wildavsky 1990). All of this variation is of great interest to social scientists, who seek to understand why it is that different individuals, groups, or societies differ in the ways they perceive, select, and respond to the risks that they construct. Thus alongside the cast of characters that we might call risk professionals (scientists, engineers, health and safety executives, clinicians, actuaries, and risk managers) whose job it is to identify, measure, and manage risk, there is another group of investigators that we might call risk scholars (sociologists, anthropologists, historians, political scien-

70 David Garland tists). The project of this second, more recently assembled group - some of whom are listed in the introduction to this article - is to understand risk not in order to improve its management (or at least not primarily for that purpose) but instead to improve our understanding of the individuals and institutions involved. There is no hard and fast boundary between these different kinds of intellectual activity, and certain writers move back and forth between the two undertakings. A number of psychologists, economists, and decision theorists (notably Kenneth Arrow (1971) and Daniel Kahneman and Amos Tversky (1979)) have made contributions of great significance for the practice of risk management while also advancing the theoretical agendas of their respective disciplines. Risk scholars such as Ulrich Beck, Anthony Giddens, and Mary Douglas engage with policy makers and policy thinking, even if their projects are primarily intended as contributions to social theory and cultural anthropology. And a researcher and writer of so eccentric a cast as John Adams altogether confounds this distinction, since he is both a technical expert who has written about traffic control and a theorist who has made important contributions to the psychology and sociology of risk. The Rise of Risk

The recent emergence of a body of risk scholarship that is not avowedly practical in its orientation seems to me to be a development that signals something important about the significance of risk in contemporary society. That a diverse group of historians, sociologists, philosophers of science, cultural anthropologists, analysts of power, students of government, welfare state scholars, and social theorists should suddenly find themselves converging, quite unexpectedly, upon questions of risk, strongly suggests that these questions have become important features of contemporary culture and society. Of course academic discourse is often prone to fads and fashions, and a concept made famous by a prominent theorist will sometimes be taken up and repeated whether or not it does any theoretical work in the hands of the imitators. No doubt the word 'risk' - like 'discourse' or 'postmodernism' or 'globalization' - has become a buzzword that is sometimes overused. But behind this penumbra of merely modish use, there is now an impressive core of serious work that takes 'risk' as its object of inquiry and which discovers this phenomenon playing an important role in numerous institutions and social settings.

The Rise of Risk

71

Risk and Modernity

What is it about contemporary society that makes discourse about risk so resonant? What social forces have prompted the rise of risk and reflections upon it? Are we now a 'risk society' and if so, precisely what does that phrase imply? In answer to these questions, and by way of a conclusion, I will sketch out an argument built around the following assertions: (a) Risk and our attempts to control it are corollaries of purposeful action and are thus ubiquitous elements of human experience; (b) Modern societies have become more successful in assessing and managing risk thanks to the development of probability theory, statistics, and systematic techniques of measurement and control; (c) As a result, techniques of systematic risk management have become a pervasive element of modern organizations and institutions; (d) Questions that bear upon risk management have become a source of considerable anxiety in contemporary culture because of increased expectations, decreased levels of trust, and new sources of insecurity; (e) We are not a 'risk society' in the sense of being exposed to more, or more serious dangers. If we are a risk society it is because we have come to be more conscious of the risks that we run and more intensely engaged in attempts to measure and manage them. Risk Is Ubiquitous

Human beings have always engaged with chance, with uncertainty, and with the risks involved in an unknown future. Risk is a corollary of planned action. Whenever human beings engage in purposeful, futureoriented action, they encounter the possibility of mischance and try to control for that possibility. Hunting and gathering, planting crops, making tools, making promises, forming families, bearing children - all the most basic forms of human activity - involve attempts to 'colonize the future,' to realize present-day plans in some anticipated future time. Knowledge about what that future might bring, and methods of controlling it, have consequently been among the basic tools of human life, and among the most sought-after of human assets. The search for ways to know the future and techniques to control it has accompanied human life from the beginning, giving rise to auguries, omens, prophecies, and propitiating sacrifices, as well as to more rational methods such as induction from past experience and the careful observation of seasonal patterns and temporal cycles. This search has been energized, at least since

72 David Garland the nineteenth century, by the discovery that social conduct can be better governed once its patterns are discerned and its norms established (Hacking 1990; Foucault 1991). Peter Bernstein (1996: 1) may exaggerate when he claims that 'the revolutionary idea that defines the boundary between modern times and the past is the mastery of risk: the notion that the future is more than a whim of the gods and that men and women are not passive before nature.' But his basic point - that our capacity to colonize the future and to control risk has been revolutionized by the development of rational techniques derived from mathematics and statistics - is unquestionably true. Our capacity to develop reliable predictions about the future by analysing data from past experience has been utterly transformed by the development of statistical inference and methods of systematic enumeration. Census counts, book-keeping, registers of births and deaths, mortality tables, crime figures, accident rates, the whole avalanche of printed numbers - these have been raw materials for forecasting and planning the future. (Hacking 1990; Bernstein 1996). Large data sets, carefully sampled and sorted, systematically analysed by reference to normal distributions, standard deviations, and other statistical concepts, together with a growing scientific understanding of how things work, have altogether transformed our capacity to tame chance and deal with uncertainty. What distinguishes modern society from its predecessors is not the attempt to master risk and to colonize the future, but the invention and widespread adoption of rational, systematic methods for formally and effectively doing so. Rationality, Reflexivity, and Risk

Max Weber (1978) and Michel Foucault (1977) have characterized modern society as being driven by processes of rationalization and discipline that bring norms of purpose-rational action to bear in all areas of social activity, from production and commerce to government and the conduct of everyday life. Anthony Giddens (1990; 1991) has elaborated on this theme by describing the importance of reflexivity in modern organizations and modern life more generally, with institutions and individuals increasingly monitoring their conduct in systematic ways, and bringing scientific knowledge and expertise to bear on their decision making. More recently, Michael Power (1994; 1997) has described the 'audit explosion' - the rapid spread of techniques of inspection, verification,

The Rise of Risk 73 and control throughout businesses and public sector organizations in an effort to increase the transparency and accountability of these organizations and thus enhance their effectiveness. Each of these interlinked processes - rationalization, reflexivity, and regulation through auditing - is characteristic of modernity, and each one is designed to identify, measure, and manage the various risks faced by the particular individuals or organizations in question. Risk, its monitoring and its management, is a built-in element of reflexive action. And as society becomes more organized, more rational, and more managed; as audits and inspections spread out from the business world into other organizations and areas of life, systematic risk assessment and management become ever more pervasive. Formal risk management is, in that sense, one of the characteristic institutions of modern society. Its spread - like the spread of discipline, rational management techniques, reflexivity, and auditing - is part of what we mean when we talk of modernization. The will to measure and evaluate risk in a systematic way, the occasions for its monitoring and management, and the tools and techniques with which to perform these functions, are increasingly pervasive features of the modern landscape (Reiss 1989). The spread of risk management into new settings has significant consequences for the organizations and institutions concerned. As actuarial or algorithmic styles of risk management are taken up in institutions where risk was previously managed in less formal, or more clinical, ways, the institutions change. Scholars like Robert Castel (1991), Nikolas Rose (2002), Pat O'Malley (1998); Malcolm Feeley and Jonathan Simon (1992) have traced the ways in which the introduction of actuarial risk management has transformed decision making in mental health and criminal justice settings, with consequences for the ways in which these institutions conceive of the problems they address and exercise the powers at their disposal.12 Scholars such as Sparks (2001) have noted that risk discourse or even risk management protocols, when adopted, do not necessarily govern decision making in these settings, since other considerations are also in play. But it is clear that as scholars begin to investigate their particular area of expertise, they will find more and more evidence of the spread of risk technologies and their increasing influence in modern decision making. As Francois Ewald (1990: 147) puts it, 'society has come to understand itself and its problems in terms of the principles of the technologies of risk.' Modern societies are riskmanaging societies.

74 David Garland Beck's 'Risk Society* Thesis The proliferation of formalized risk management in all sectors of modern society, together with the importance of risk industries, may explain the empirical resonance of the new wave of sociological risk scholarship. But it does not account for the urgency and anxiety that now seems to drive much of the public discussion of risk — a cultural mood that is clearly expressed in Beck's 'risk society' thesis (Beck 1992; see also Simon 1987). If risk management is a routine element of the lifeworld, why does it generate so much anxiety? Beck's claim is that we are 'living on the volcano of civilization' as we enter upon a new stage of modernity. This latest phase of the modernization process (which runs from pre-modernity to the 'simple modernity' of industrial society to the 'reflexive modernity' of contemporary 'risk society') is fraught with new risks and hazards that we ourselves have manufactured and that are massive in their potentially global impact. 'Risk society is a world society ... Risk society is a catastrophic society' (Beck 1992: 22 and 24): By risks I mean above all radioactivity, which completely evades human perceptive abilities, but also toxins and pollutants in the air, the water and foodstuffs, together with the accompanying short- and long-term effects on plants, animals and people. They induce systematic and often irreversible harm, generally remain invisible, are based on causal interpretations, and thus initially only exist in terms of the (scientific or anti-scientific) knowledge about them. They can thus be changed, magnified, dramatized or minimized within knowledge, and to that extent they are particularly open to social definition and construction. Hence the mass media and the scientific and legal professions in charge of defining risks become key social and political positions. (Beck 1992: 23)

Beck's risk society thesis is a disutopia of unintended consequences. History has caught up with modern societies, causing them to focus less and less upon technical and economic development, and more and more upon the problem of managing the hazards that this development entails. In this context, individuals and institutions are heavily dependent upon expert advisers as guides to the risks that they run, the hidden dangers they face, and the safest or healthiest course of action available. But experts provide conflicting information. And the more the public learns about science, the more it realizes that science is fallible, provi-

The Rise of Risk 75 sional, always subject to doubt and revision. 'Science becomes indispensable, and at the same time devoid of its original validity claims' (Beck 1992: 165). This, together with the cumulative experience of disasters and poorly managed risks, has produced an anxious public that no longer trusts scientists to get it right and no longer trusts governments to keep us informed. Risk professionals no longer monopolize risk discourse - a situation that both democratizes debate and makes it more explosive. The result is a new kind of society in which 'risk production' overtakes 'wealth production' and struggles over the distribution of hazards ('bads') displace struggles over the distribution of wealth ('goods') as the central theme of political conflict. Beck emphasizes the unprecedented size and catastrophic potential of today's manufactured hazards, together with the intrinsic unknowability of their likelihood. 'Now manufactured uncertainty means that risk has become an inescapable part of our lives and everybody is facing unknown and barely calculable risks. Risk becomes another word for "nobody knows." We no longer choose to take risks, we have them thrust upon us. We are living on a ledge - in a random risk society, from which nobody can escape' (Beck in Franklin 1998:12). In this respect, his analysis seems diametrically opposed to that of Peter Bernstein (1996), who celebrates the 'mastery of risk' that separates modern society from its predecessors, and to that of Ian Hacking (1990), who talks of 'the taming of chance' and emphasizes how social control - over both people and events - has been greatly extended by modern techniques of classifying, counting, and calculating. Whereas Bernstein and Hacking document the ways in which modern science has succeeded in extending our capacity to manage risks and control outcomes, Beck is convinced that modern science has run up against its limits, having manufactured dangers - 'uncontrollable risks' that will forever remain beyond human control. More Risk or More Management?

Beck's cataclysmic analysis seems overstated, both in its historical judgments and in its account of dangers we now face.13 It is true, of course, that modern science and engineering have created substances, technologies, and forms of energy with the power to bring death and destruction to large parts of the planet. It is true that global commerce and communications intensify our interdependencies, increase our mutual exposure, and make regulation harder to impose and implement. It is

76 David Garland true that industrial production, the burning of fossil fuels, and the spread of motorcars have brought about changes in the earth's atmosphere, though it is not clear that these changes are as detrimental as some environmentalists fear (for a sceptical account, see Lomborg 2001). And it is undeniable that disasters such as Bhopal, Chernobyl, Three Mile Island, and September llth, 2001 have shown the appalling injury and destruction that chemical, nuclear, or terrorist incidents can inflict. But human societies have always faced massive threats to life and wellbeing - whether from 'nature,' in the shape of plagues, famines, floods, and earthquakes, or from other people, in the form of war, pogroms, and genocides. Most of these threats are now better understood and better controlled than at any time in human history. That banks no longer fail with the calamitous frequency that they did a century ago is a mark of this change (Braithwaite and Drahos 2000). That mortality and morbidity rates in the developed world have been continually improving for at least the last century (British Medical Association 1987) is crude but compelling evidence that many risks are now being better managed than ever before. That international organizations have been established to manage threats of nuclear aggression, ecological disaster, and financial collapse hardly guarantees that we will not be convulsed by some future disaster, but it does reduce that risk considerably. Of course, many of the risks to which Beck refers are the result of new processes and technologies, and may yet prove devastating in their future impact.14 But it is not hopelessly optimistic to believe that the same scientific and engineering skills that manufactured these new risks will be capable of designing and putting in place technologies and control systems that will manage them effectively, preventing their misuse, avoiding accidents, and reducing their harmful side effects to tolerable levels. And where such safety measures and controls are lacking, or are perceived as inadequate, contemporary risk awareness is liable to push for more conservative, precautionary approaches. The growing importance in contemporary law and policy of 'the precautionary principle' (Geistfeld 2001) suggests that the onus of proof is shifting from the victim of a hazard to the actor who might bring it about.15 Our relationship to science and technology, and to the modernist project of transforming the world and colonizing the future, is now much more ambivalent than it once was. What has changed most, it would seem, is not the risks we face but the perceptions and sensibilities we bring to bear upon them. Countering Beck's world-historic pessimism with a Panglossianism of equal proportions does not get us very

The Rise of Risk

77

far, however. Debates that trade generalized claims about aggregate risk levels quickly lead away from the facts towards the cultural and political variables that Mary Douglas describes. These are, of course, global risks that require global solutions, but for the most part we would do better to avoid impressionistic analysis that talks of 'society as a whole' or 'world risk society' and aim for greater specificity. Individualization and Insecurity Beck seems more convincing when he develops a subsidiary thesis about the impact of a series of social and cultural transformations that changed the social position of individuals in the second half of the twentieth century and altered the ways in which they relate to their world and to the threats that it contains. Beck uses the ideas of 'individualization' and 'the end of tradition' to refer to the various ways in which the structures of class, gender, work, and locality have increasingly relaxed their grip upon individuals. This process - which has been extensively documented by sociologists such as Putnam (2000), Fukuyama (2000), and Garland (2001) - has given individuals more freedom, more mobility, more choice, and more power to shape their personal identities in non-traditional ways, relatively unconstrained by other people. But the liberating effect of these social trends has come at a cost - not just to the stability of families and communities or to those left behind, but also to the affluent individuals who exercise these newfound freedoms. The individualized world is a world of choice, of multiple options and endless possibilities. A person born into a particular class or faith, growing up in a particular neighbourhood, developing a particular sexuality, working in a specific job, or married to a specific spouse need not be forever defined by these circumstances. He or she is now freer than before to leave these ties and identities behind and create new and different ones elsewhere. But this post-traditional world of choice is also, and necessarily, a more uncertain world, replete with risks and dangers. People and places change. Friends move. Families fall apart. Jobs come and go. Identities are no longer set. The self is no longer stable. In this new context, there are fewer settled traditions and established group norms to which the individual can turn for guidance when hard choices need to be made. He or she must face these risks alone. Of course, people can turn to experts, therapists, and professional advisers, and they do so in increasing numbers (Giddens 1991). But conflicts of expert opinion and the variability of goals and values often render diis

78 David Garland recourse a source of frustration rather than reassurance. The result is an increasingly endemic sense of insecurity - experienced even by well-todo individuals who are, by historical standards, healthier and more affluent than ever before (Sennett 1998; Bauman 2000). Individualization has been driven by the demand for greater choice and individual freedom, and it has made these goods available to more of the population than ever before. But it has also been a recipe for insecurity - for low levels of commitment and diminishing social trust. Today's freed-up individuals enjoy their freedoms against a background of a newfound dependency upon expert systems and newfound uncertainty about the lives they choose. Risk and the Cultural Imagination

With this cultural context in mind, it is easier to understand the sense of dread produced by the existence of large-scale, unforeseeable, insidious risks - particularly those risks that cannot be controlled by individuals. If risk perception and fear levels are shaped by subjective orientations as well as by objective evidence, it should not surprise us that people focus upon risks that mirror the structure of their personal fears and anxieties. Beck reminds us that risks can be 'changed, magnified, dramatized or minimized within knowledge, and to that extent they are particularly open to social definition and construction.' That risks are socially constructed makes them no less real, but it does mean that they can be magnified and dramatized in the public imagination, as projections of a structure of personal anxieties created by the new uncertainties of social life. In this highly reflexive, low-trust culture, one of the key functions of safety regulation - its tendency to blind us to the risks we routinely take (Manning 1989) - is increasingly undermined. Security and insecurity, risk management and risk awareness, urban fortification and fear of crime - our responses to risk and the risks we perceive frequently intensify one another instead of cancelling one another out. Citizens of contemporary Western nations are, on the average, healthier, live longer, and are better protected from economic risks than those of any previous societies (British Medical Association 1987). And if they face more in the way of self-imposed environmental hazards, or international terrorist threats, then they are also more actively engaged with the management of these risks than ever before. But success in the regulation and reduction of risk has not prevented the emergence of new sources of insecurity that have their roots in destabilized social relations

The Rise of Risk

79

rather than environmental dangers. And insecurities, like risks, are interactive, feeding off one another and corroding the trust upon which modern social life increasingly depends. If we live in a 'risk society' it is not in the sense of one that is more dangerous by any objective measure. If we live in a risk society it is because we are now more democratically engaged, more reflexively rational, and more prone to distrust in our engagement with the phenomenon than were previous generations. Coda Anyone wishing to understand the place of risk in contemporary culture can do no better than read Don DeLillo's novel White Noise (1985), which explores this psychic current and its social roots. The publisher's blurb, which could have been written by Ulrich Beck, portrays the risk society as the dark underside of modernity, and suggests its intimate links to the new structures of family life and individualized freedom: Jack Gladney teaches Hitler studies at a liberal arts college in Middle America ...Jack and his fourth wife, Babette, bound by their love, fear of death, and four ultramodern offspring, navigate the usual rocky passages of family life to the background babble of brand-name consumerism. Then a lethal black chemical cloud floats over their lives, an 'airborne toxic event' unleashed by an industrial accident. The menacing cloud is a more urgent and visible version of the 'white noise' engulfing the Gladney family- radio transmissions, sirens, microwaves, and TV murmurings - pulsing with life, yet heralding the danger of death. In his fictional account of this 'airborne toxic event' and his characters' reaction to it, DeLillo anticipates every single one of the 'negative' factors that, according to a later report by the Royal Society (1992) are known to prompt individuals to recoil from risks: (1) Involuntary exposure to a risk; (2) Lack of personal control over outcomes; (3) Uncertainty about probabilities or consequences of exposure; (4) Lack of personal experience with the risk (fear of unknown); (5) Difficulty in imaging risk exposure; (6) Effects of exposure delay in time; (7) Genetic effects of exposure (threatens future generations); (8) Infrequent but catastrophic accidents ('kill size'); (9) Benefits not visible; (10) Benefits go to others (inequality); (11) Accidents caused by human failure rather than natural causes. (Royal Society 1992: 101)

80 David Garland In matters of such social and psychic salience, perhaps science can only discover what the culture already knows. Acknowledgments I wish to thank the following friends and colleagues for their advice: John Adams, John Braithwaite, Richard Ericson, Mark Geistfeld, Clay Gillette, Dorothy Nelkin, Rick Pildes, Ricky Revesz, Richard Stewart, and the participants at the Green College conference on Risk and Morality. I am also grateful to the Filomen D'Agostino and Max E. Greenberg Research Fund for its support.

Notes 1 Though some are a product of our perceptions - for instance, where 'we have nothing to fear but fear itself 2 Carol Heimer (this volume: 284ff) makes a similar point when she says that 'insurance is an ongoing strategic game ... [SJtrategic considerations pervade every stage of the unfolding insurance contract - the early application, offer, and negotiation of the contract; the subsequent actions of the policyholder, now protected from some of the financial consequences of a possible loss; the investigation following a reported loss; the bargaining about exactly how much a policyholder will be compensated when a loss occurs. We should therefore think of insurance as an ongoing strategic interaction, a game in which insurers and policyholders both continue to adjust to each others' moves ...' 3 America's recent experience with anthrax in the mail revealed all of these problems. Government officials struggled to find the right balance between safety warnings and panic-inducing alarms while casting around for accurate information about exposure levels and health consequences. Meanwhile, citizens in the affected cities struggled to figure out what precautions to take - wearing gloves, taking prophylactic doses of antibiotics, inspecting the mail, leaving town. 4 Baker and Simon (2002: 11) develop the idea of'governing through risk,' by which they mean 'the use of formal considerations about risk to direct organizational strategy and resources.' This idea, which refers to the actions of both public and private authorities, is closely related to the themes of governmentality developed by Foucault (see Burchell et al. 1991). 5 The extent of welfare provision, the character of social regulation, and the indusiveness of social insurance, are of course key variables that distinguish

The Rise of Risk

81

different states and shape social politics. For a discussion of variation, see Esping-Andersen (1990). 6 'The predominant function of modern law is to allocate risk ... In the field of contract law, for example, contract litigation only a few decades ago turned chiefly on differing interpretations of the terms of underlying written contracts. In modern contract litigation, in contrast, the issues have been completely reoriented around the issues of risk. The fact that some change in conditions led one of the parties to breach the contract is only the beginning of the inquiry. The issue before the court is which party should bear the risk of the change in conditions that impelled the breach' (Priest 1990: 209). 7 The neo-liberal assumption that one encourages entrepreneurial risk taking by removing protections runs up against the evidence. ' [A] review of... riskcontrol policies tends to lead to a conclusion virtually opposite the prevailing wisdom, which holds that there is an inverse relationship between risk control and risk taking, as though people become more and more risk averse as risk-control policies spread ... [T]he long history and extent of public risk-control policies ... suggest that the relationship between risk control and risk avoidance, rather than being inverse, is direct, the one depending on the other' (Lowi 1990: 26). Successful entrepreneurs do take risks, but they do so against a background of extensive protections and hedges, many of them - such as the $500 billion savings and loan bailout in the United States - supplied at the taxpayers' expense. 8 '[T]he welfare state functioned under a 'Veil of ignorance," an insurance principle that presupposed the equality of individuals with regard to various risks. The opaque nature of society was thus an implicit condition of the sense of equity. All members of society could be considered interdependent insofar as they perceived the nation as a relatively homogeneous class of risks. But this perception is ceasing to correspond to reality ... as society gains more and more knowledge of its differences, a considerable change in the perception of fairness tends to be produced' (Rosanvallon 2000: 29). This shift from universalism to risk segmentation in respect of welfare is paralleled in private insurance: 'The recent trend towards more risk segmentation or unpooling within private insurance stems in part from the increasingly detailed risk information (e.g. financial, medical) which is available to actuaries and underwriters, both concerning individual insureds and concerning trends in population' (Ericson, Barry, and Doyle 2000: 534). 9 On insurance and morality, see Zelizer (1979); Stone (2002); Heimer (this volume). 10 'Our courts have defined two basic principles of decision-making to inter-

82 David Garland nalize costs to create incentives to reduce the risk level as much as is practicable. First, if the injury could have been practically prevented, liability will be placed on that party in the relatively better position to prevent it. Second, if the injury could not have been practicably prevented, liability will be placed on that party in the relatively better position to spread the risks of the injury' (Priest 1990: 216). 11 The private security industry - which is one of today's fastest growing commercial sectors - is another important risk industry. See Jones and Newburn (1998). 12 Describing a shift 'from dangerousness to risk' in the practice of French mental health institutions, Castel (1991: 288) says this: 'What the new preventive policies primarily address is no longer individuals but factors, statistical correlations of heterogeneous elements. They deconstruct the concrete subject of intervention [which used to be the dangerous individual] and reconstruct a combination of factors liable to produce risk.' Feeley and Simon (1992) identify similar shifts in criminal justice and talk of the emergence of'actuarial justice.' 13 David Goldblatt (1996) provides a valuable critique of Beck's thesis. See Beck's response in Franklin (1998). 14 The major risks in this respect would seem to lie not so much in technology and its development, but rather in the collapse of the control systems designed to safeguard hazardous substances, either because of economic collapse (as in the former Soviet Union) or because of the actions of rogue states and terrorist organizations. The management of political risks of this kind calls for solutions that rely less on scientific knowledge than on international cooperation and strategic intelligence. 15 Franklin (1998: 4) describes this precautionary principle as 'a political mechanism which enables politicians to take precautionary action where there is any reasonable doubt of risk to public health or to the environment, even in the absence of scientific evidence.' In international environmental law, the principle dictates that 'any uncertainty regarding the hazardous properties of a substance or activity ought to be resolved in a manner that favors regulation (and the associated possibilities of risk reduction), with cost considerations of secondary importance' (Geistfeld 2001: 119). For a detailed discussion of the principle's various versions and practical applications, see Stewart (2002). This principle operates differently in different contexts. We typically think of it as a progressive environmentalist principle, preventing big business from despoiling the earth or subjecting the common people to unnecessary dangers. But this same principle is at work in

The Rise of Risk

83

criminal justice and mental health, where an overweening concern for public safety prevents the release of individuals who may or may not present a risk of violence. The avoidance of risk to some is usually purchased at the cost of freedom of others (Garland 2001).

References Adams, J. 1995. Risk. London: UCL Press Aharoni, Y. 1981. The No Risk Society. London: Chatham Arrow, K. 1971. Essays in the Theory of Risk-Bearing. Chicago: Markham Publishing Baker, T, and J. Simon. 2002. 'Embracing Risk.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Baldwin, P. 1990. The Politics of Social Solidarity: Class Basis of the European Welfare State 1875-1975. Cambridge: Cambridge University Press Barry, A., T. Osborne, and N. Rose. 1996. Foucault and Political Reason. London: UCL Press Bauman, Z. 2000. 'The Social Uses of Law and Order.' In D. Garland and R. Sparks, eds., Criminology and Social Theory. Oxford: Oxford University Press Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Beck, U., A. Giddens, and S. Lash. 1994. Reflexive Modernization: Politics, Tradition and Aesthetics in the Modern Social Order. Cambridge: Polity Press Bernstein, P.L. 1996. Against the Gods: The Remarkable Story of Risk. New York: Wiley Braithwaite,J., and P. Drahos. 2000. Global Business Regulation. New York: Cambridge University Press British Medical Association. 1987. Living With Risk. London: BMA Publications Burchell, G., C. Gordon, and P. Miller, eds. 1991. The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press Calabresi, G. 1970. The Cost of Accidents. New Haven: Yale University Press Castel, R. 1991. 'From Dangerousness to Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press Coleman, J. 1992. Risks and Wrongs. New York: Oxford University Press Defert, D. 1991. 'Popular Life and Insurance Technologies.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press

84 David Garland DeLillo, D. 1985. White Noise. New York: Viking Penguin Donzelot.J. 1979. The Policing of Families. London: Hutchinson Douglas, M. 1992. Bisk and Blame: Essays in Cultural Theory. London: Routledge Douglas, M., and A. Wildavsky. 1982. Risk and Culture: An Essay on the Selection of Technological and Environmental Dangers. Berkeley: University of California Press Dunbar, N. 2001. Inventing Money: The Story of Long-Term Capital Management and the Legends Behind It. New York: Wiley Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press Ericson, R.V., D. Barry, and A. Doyle. 2000. 'The Moral Hazards of Neoliberalism: Lessons from the Private Insurance Industry.' Economy and Society 29(4): 532-58 Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press Ewald, F. 1990. 'Norms, Discipline and the Law.' In R. Post, ed., Law and the Order of Culture. Berkeley: University of California Press - 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press - 2002. 'The Return of Descartes Malicious Demon: An Outline of a Philosophy of Precaution.' In T. Baker andj. Simon, eds., Embracing Risk The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Feeley, M., and J. Simon. 1992. 'The New Penology: Notes on the Emerging Strategy of Corrections and its Implications.' Criminology 30: 449-74 Foucault, M. 1977. Discipline and Punish: The Birth of the Prison. London: Allen Lane - 1991. 'Governmentality.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press Franklin, J., ed. 1998. The Politics of Risk Society. Cambridge: Polity Press Friedman, L. 1981. 'Legal Culture and the Welfare State.' In G. Teubner, ed., Dilemmas of Law in the Welfare State. Berlin: de Gruyter Fukuyama, F. 2000. The Great Disruption: Human Nature and the Reconstitution of Social Order. New York: Free Press Garland, D. 2001. The Culture of Control: Crime and Social Order in Contemporary Society. Chicago: University of Chicago Press Geistfeld, M. 2001. 'Reconciling Cost-Benefit Analysis with the Principle That Safety Matters More Than Money.' New York University Law Review 76(1): 114-87 Giddens, A. 1990. The Consequences of Modernity. Cambridge: Polity Press

The Rise of Risk 85 - 1991. Modernity and Self-Identity: Self and Society in the Late Modern Age. Cambridge: Polity Press - 1999a. 'Risk and Responsibility.' The Modern Law Review 62(1): 1-10 - 1999b. 'Runaway World: The Reith Lectures Revisited.' The Director's Lectures. London: LSE - 2000. Runaway World. New York: Routledge Goldblatt, D. 1996. Social Theory and the Environment. Cambridge: Polity Press Greenwood, P. 1982. Selective Incapatitation. Santa Monica, CA: Rand Hacking, I. 1990. The Taming of Chance. Cambridge: Cambridge University Press - 1991. 'How Should We Do the History of Statistics?' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality. Chicago: University of Chicago Press Heimer, C. 1985. Reactive Risk and Rational Action: Managing Moral Hazard in Insurance Contracts. Berkeley: University of California Press Jones, T, and T. Newburn. 1998. Private Security and Public Policing. Oxford: Oxford University Press Kahneman, D., and A. Tversky. 1979. 'Prospect Theory: An Analysis of Decision under Risk.' Econometrica47(2.): 263-91 Knight, F. 1965 [1921]. Risk, Uncertainty and Profit. New York: Harper Row Lomborg, B. 2001. The Skeptical Environmentalist. Cambridge: Cambridge University Press Lowi, T. 1990. 'Risks and Rights in the History of American Governments.' Daedalus 119 (4): 17-40 Manning, P. 1989. 'Managing Risk: Managing Uncertainty in the British Nuclear Installations Inspectorate.' Law and Policy 11 (3): 350-69 Monaghan,J. 2001. Rethinking Risk Assessment: The Macarthur Study of Mental Illness and Violence. New York: Oxford University Press Nelkin, D. 1984. Science in the Streets: The Communication of Technical Risk. New York: Priority Press O'Malley, P., ed. 1998. Crime and the Risk Society. Aldershot: Dartmouth Perry, S. 1996. 'Risk, Harm and Responsibility.' In D. Owen, ed., Philosophical Foundations of Tort Law. New York: Oxford University Press Pildes, R., and C. Sunstein. 1995. 'Reinventing the Regulatory State.' University of Chicago Law Review 62: 1-131 Power, M. 1994. The Audit Explosion. London: Demos - 1997. The Audit Society: Rituals of Verification. Oxford: Oxford University Press Putnam, R. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon and Schuster Priest, G. 1990. The New Legal Structure of Risk Control.' Daedalus 119(4): 207-27

86 David Garland Reiss, A. 1989. The Institutionalization of Risk.' Law and Policy 11(3): 392-401 Reith, G. 1999. The Age of Chance: Gambling in Western Culture. London: Routledge Rosanvallon, P. 2000. The New Social Question: Rethinking the Welfare State. Princeton: Princeton University Press Rose, N. 1996. 'Governing "Advanced" Liberal Democracies.' In A. Barry et al., eds., Foucault and Political Reason. London: UCL Press - 2002. 'At Risk of Madness.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Royal Society. 1992. Risk: Analysis, Perception, Management. London: Royal Society Sennett, R. 1998. The Corrosion of Character: The Personal Consequences of Work in the New Capitalism. New York: Norton Simon, J. 1987. 'The Emergence of a Risk Society: Insurance, Law and the State.' Socialist Review95: 61-89 - 2002. 'Taking Risks: Extreme Sports and the Embrace of Risks.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Skolnick, J. 1978. House of Cards: Legalization and Control of Gambling. Boston: Little Brown Sparks, R. 2001. 'Degrees of Estrangement: The Cultural Theory of Risk and Comparative Penology.' Theoretical Criminology 5(2): 159-76 Stewart, R. 2002. 'Environmental Regulatory Decisionmaking under Uncertainty.' Research in Law and Economics 10: 71-126 Stone, D. 2002. 'Beyond Moral Hazard: Insurance as a Moral Opportunity.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Thompson, M., R. Ellis, and A. Wildavsky. 1990. Cultural Theory. Boulder, CO: Westview Weber, M. 1978. Economy and Society. Berkeley: University of California Press Webster's Third New International Dictionary. 1971. Springfield, MA: G. and C. Merriam and Co. Wildavsky, A., and K. Drake. 1990. 'Theories of Risk Perception: Who Fears What and Why?' Daedalus 119: 41-60 Williams, R. 1983. Keywords: A Vocabulary of Culture and Society. London: Fontana Yergin, D., andj. Stanislaw. 1998. The Commanding Heights: The Battle between Government and the Marketplace that is Remaking the Modern World. New York: Simon and Schuster Zelizer, V. 1979. Morals and Markets: The Development of Life Insurance in the United States. New York: Columbia University Press

4

Risk and Morality: Three Framing Devices JOHN ADAMS

Risk-management decisions are moral decisions made in the face of uncertainty. This paper proffers three framing devices that it is hoped will assist an understanding of the process. The first — three kinds of risk — is a typology of uncertainty; any attempt to relate risk to moral principles and ethical conduct should be clear about the type of risk under discussion. The second - the 'risk thermostat'- characterizes risk management as cost-benefit analysis without the dollar signs. It calls attention to the diversity of incommensurable risks and rewards, the importance of being clear about who gets the rewards and who bears the costs, and the consequence of ignoring the reasons that people have for taking risks. The third - a typology of ethicalfilters- makes the point that in the realm of risk management there is no such thing as 'society'; there is no moral consensus about the right way to manage risk, but rather a set of contending moralities. Directly Perceptible Risk Directly perceptible risks are managed instinctively and intuitively - we all duck if we see something about to hit us. Our ability to deal with such risks successfully has been acquired by evolution. We manage them intuitively; we do not undertake a formal quantitative risk assessment before we cross the street. We seek to manage these risks ourselves. We monitor our environment for signs of safety and danger and respond to what we see. Attempts by institutional risk managers to impose upon us safety standards that differ from our own are resisted. This risk category can be used to introduce three interesting sets of ethical problems: risks to children, risks imposed on others, and the role of the state.

88 John Adams Figure 4.1: Three Kinds of Risk e.g., cholera: need a microscope to see it and a scientific training to understand

Scientists don't know or cannot agree: e.g., BSE/vCJD, global warming, low-level radiation, pesticide residues, HRT, mobile phones, passive smoking, stock market...

e.g., climbing a tree, riding a bike, driving a car

The Problem of Children

Newborn infants have all their risks managed for them by their parents or guardians. The process of development is one in which responsibility for managing risk is progressively handed over until the child reaches the age of responsibility. A study of this process can be rewarding for ethicists because it is here that one often finds the rules of moral behaviour most explicitly enunciated. The question of how the process of development ought to be managed raises a host of ethical issues. Parenting strategies range from overprotective to reckless and irresponsible - with each extreme being defined by the other. If the handover of responsibility is too slow, physical fitness and the acquisition of social skills and a sense of responsibility will all be impaired (Hillman, ed. 1993). If it is too fast, more accidents will result. Inculcating a sense of responsibility for risk management involves the development of a sense of self-preservation, but also a sense of duty to others. The Problem of Imposed Safety and Danger

Moral behaviour, responsible behaviour, respects the rights of others however they might be defined. In a study of children's independence

Risk and Morality: Three Framing Devices 89 (Hillman, Adams, and Whitelegg 1990) we discovered that these rights were defined differently in England and Germany, with interesting consequences. We tried to chart the ages at which certain responsibilities for managing risk were transferred from adults to children by recording the ages at which children acquired parental 'licences' to go to school on their own, use buses, ride their bikes in the street, and so forth. We found that the age of licence in Germany was much lower than in England; German parents allowed their children more independence because they felt that they could rely on adults in the community to discipline other people's children in public if they saw them misbehaving. This might be construed as evidence of a greater sense of community engagement in Germany - or perhaps as evidence of a culture requiring a strict adherence to social rules. Further Anglo-German comparisons throw up other differences in the ethical codes governing behaviour on the road. Germany's road death rate is about twice that of England's, and the explanation is not to be found on the autobahns which, despite fewer speed limits, have fatality rates similar to those on English motorways. The Germans appear to have much stricter rules, more strictly enforced. A pedestrian in Germany crossing against a red light, for example, is committing a criminal offence, whereas in England such signals are merely advisory. Germans of my acquaintance, byway of explaining their high road death rate, quote the inscription on the mythical German tombstone: T had the right of way.' Rule-bound cultures, it seems, are not necessarily safer. The Role of the State The rules that ought to govern state intervention in the management of risk are the subject of endless debate. With respect to directly perceptible risks the state often acts in loco parentis, for adults as well as children. Despite reaching the age of responsibility, adults are frequently not trusted to behave responsibly. There is little disagreement with the view that the criminal law should be invoked to punish behaviour that puts others at risk - such as speeding, or ignoring red lights and stop signs. But should it also declare criminal mental or physiological states that predispose people to behaviour that imperils others, such as driving while under the influence of drink or drugs? A person 2.5 times over the permitted alcohol limit is twenty times more likely to be involved in a fatal road accident, but people diagnosed as having personality disorders are ten times more likely than 'normal' people to die in a road

90 John Adams accident. Should they also be forbidden to drive? Or should young men, who, on average, are a hundred times more likely to be involved in serious road accidents than middle-aged women (Adams 1985, 1999; Evans 1991)? Should blood-alcohol limits be supplemented by testosterone limits? The search for consistency in defining the role of the state in the management of risk encounters contending moral codes. In the United States, where the ratio of people killed by guns to people killed by fireworks is about 3000:1, fireworks in most states are more strictly controlled than guns.1 Increasingly, the state now criminalizes self-risk in the form of behaviour such as driving without a seatbelt, or riding a motorcycle or bicycle without a helmet. Should it, to be consistent, ban all behaviour that carries with it an elevated risk of self-harm - including smoking, drinking, and eating too many cream buns? Risks Perceived through Science Such questions become increasingly pertinent when we enter our second category of risk. While directly perceptible risks are still managed for the most part by individual perceivers, risks that can only be seen with the help of science usually have institutional risk managers. Advancing science routinely discovers risks invisible to the naked eye. The cause of cholera was discovered, in a preliminary fashion, by the father of modern epidemiology, Drjohn Snow, some decades before its microbial cause was revealed by researchers with microscopes. This is an early example of the application of science to the management of risk. Dr Snow removed the handle of the Broad Street pump, a well in Soho in the west end of London. He was acting upon his suspicion that the source of the illness was a contaminated well. Not everyone who drank from the well succumbed to cholera, only a percentage. This form of risk management has become much more statistically sophisticated. Suspicions, now called hypotheses, are converted into probabilities. Risk is a close relation to uncertainty. Where we cannot be certain of the connection between cause and effect we clutch at the straw of probability. At present, for most illnesses we must content ourselves with probabilities. Ultimately, genetic science may be able to identify the causes of certain diseases precisely; we may be able to say with certainty who will, or will not, develop a particular illness. But difficulties in measuring exposures, ignorance about dose response relationships, and variability in human susceptibility will continue to make work for actuar-

Risk and Morality: Three Framing Devices 91 ies and epidemiologists, who deal in probabilities, as will accidents that are the consequence of human fallibility or unpredictable in nature. Estimates of the probability of particular harms are quantified expressions of ignorance. One can estimate that the average Briton has a 1:17,000 probability of dying in a road accident this year. This risk estimate is calculated by dividing the number of people killed in road accidents last yearly the total population, and assuming nothing much will change. However, as we noted above, the road accident literature reveals that the average young man is about a hundred times more likely than the average middle-aged woman to be involved in a serious road accident; that you are ten times more likely to die, on average, if you have a personality disorder, and twenty times more likely, on average, if you are 2.5 times over the alcohol limit. And all these averages have variances that must be explained by invoking further variables such as the condition of the brakes, state of the road, and so on. The veneer of scientific authority imparted by quantified probability often can withstand little scratching. Virtual Risks

When we reach our third category the veneer has been stripped away completely. Here we can no longer pretend to sufficient knowledge to ascribe probabilities. When scientists admit to ignorance, or reputable scientists contend with each other in ways that mystify the rest of us, we are in the realm of virtual risk. Virtual risks may or may not be real, but they have real consequences if sufficient numbers of people believe in them. Virtual risks are liberating. If scientists cannot pronounce convincingly, we are free to act upon our convictions, prejudices, and superstitions. Figure 4.2, borrowed (and amended) from the risk management manual of a major airline, describes a common form of virtual risk. Down Victorian coal mines, in conditions of high-dose exposure to radiation or other known toxins, or in transport systems with high accident rates, danger is obvious, and the measures required to reduce it usually equally so. But when all the obvious measures are in place accidents will still, occasionally, happen. One hundred per cent safety is a Utopian goal. Indeed, it is possible to have too many safety measures. So long as there is a residual dependence on the vigilance of fallible humans, their level of vigilance will depend on their belief that something can go wrong. The impressive safety record of civil aviation, and all

92 John Adams Figure 4.2: The Human Reliability Curve andVVirtual Risk

the safety redundancy built into modern aircraft, have created a problem of keeping pilots awake on long flights across time zones. Why should anyone stay alert in anticipation of something they believe will never happen? When you are on the flat part of the curve you do not have a clue whether further safety precautions will have any beneficial effect, but there are circumstances where they can have a perverse effect • where the belief in safety measures can induce complacency - the Titanic Effect. We do not respond blankly to uncertainty. We impose meaning(s) upon it. And in so doing we impose ethical principles upon it. Before considering how we do this with the help of our third framing device let us look first, with the help of our second framing device, at the process of risk management. The Risk Thermostat Risk management involves balancing the rewards of actions whose outcomes are uncertain against the losses. Figure 4.3 is a model of this balancing act. The model postulates that • everyone has a propensity to take risks • this propensity varies from one individual to another

Risk and Morality: Three Framing Devices 93 Figure 4.3: A Risk Thermostat with Filters

• this propensity is influenced by the potential rewards of risk taking • perceptions of risk are influenced by experience of accident losses one's own and others' • individual risk-taking decisions represent a balancing act in which perceptions of risk are weighed against propensity to take risk • accident losses are, by definition, a consequence of taking risks; the more risks an individual takes, the greater, on average, will be both the rewards and losses he or she incurs. Figure 4.3 is a conceptual model, not a quantifiable one. Both the rewards and accidents boxes contain very large numbers of incommensurable variables that defy reduction to a common denominator. With individual risks directly perceived this balancing act takes place inside the head of the individual risk taker. The risks that I will take dodging traffic in order catch a bus approaching on the opposite side of the road will depend on how urgently I want the reward - catching the bus. In the case of risks perceptible only with the help of scientists, engineers, actuaries, or epidemiologists, the risk manager is usually a government regulator or a safety officer in a public or private institution. The institutionalizing of risk management commonly leads to an important bias. Institutional risk management is usually considered synonymous with accident reduction.

94 John Adams The institutional risk manager's job frequently forbids any contemplation of the rewards of risk taking; judgments about safety are to be protected from corruption by concerns about profits. Institutional risk management is commonly confined to the bottom loop of the risk thermostat model. This can lead to the fruitless pursuit of safety at any cost. Rarely does the risk manager ask, 'Do we have enough accidents?'2 A Typology of Ethical Filters

In figure 4.3 the risk thermostat is fitted with filters. The influence of these filters increases as we move from clearly perceptible risks and rewards to the uncertainty we are calling virtual risk. Most debates about whether risk is 'real,' and capable of objective measurement, or something that is socially constructed, usually arise out of a failure to be clear about the type of risk under discussion. The filters operate for all three types of risk. Even with clear and directly perceptible risks - such as those taken by a rock-climber clinging to a sheer face - filters operate. His (it is usually a young man) perception of the risks and rewards that motivate his behaviour will be very different from that of his nervous mother. With invisible and possibly non-existent risks, such as those associated with pesticide residues or low-level radiation, the filters will be all-determining. The typology presented in figure 4.4 has been fully described elsewhere. It has been variously called a 'typology of rationalities,' a 'typology of social solidarities,' and a 'typology of perceptual filters.' For purposes of this chapter it is relabelled a typology of ethical filters because, inextricably bound up with every rationality, solidarity, or perceptual framework, one finds sets of moral principles and ethical codes that inform risk-taking behaviour. The characters in this typology - the hierarchist, fatalist, individualist, and egalitarian - all adhere to different myths of nature represented by the icons. The individualist myth, the ball in the cup, stands for nature robust, benign, and cornucopian; you can shake it about and the ball always comes to rest safely and securely in the bottom of the cup. The egalitarian sees nature as everywhere fragile, ephemeral, and threatened - as represented by the ball perched precariously on the overturned cup. The fatalist sees nature as untrustworthy and unpredictable. The hierarchist sees it as reliable and well behaved within limits, but cautions against pushing the ball over the rim. Like all models, figure 4.4 is a simplification of a rather more complex

Risk and Morality: Three Framing Devices 95 Figure 4.4: A Typology of Ethical Filters Prescribed Inequality

The Hierarchist

The Fatalist Individualized

Collectivized

The Individualist

The Egalitarian

Prescribed Equality

reality, but the typology nevertheless captures not only significant differences in the way nature is perceived, but also significant differences of opinion about what constitutes moral behaviour both with respect to the physical environment and to those with whom we share it. An Example: Genetically Modified Food

Some of the risks associated with food can be assigned to the category of risk directly perceptible. Our senses of sight, smell, and taste form our first line of defence against food that might make us ill. Putrid food offends all three senses and is rejected. Commonly the rewards are also directly perceptible; eating is one of life's pleasures and we are attracted to foods that look, smell, and taste delicious. Hunger and our sense of repleteness also govern, more or less satisfactorily, the quantities we consume. Science also plays an important role in what we eat. Folk science, in the form of accumulated knowledge about which plants are poisonous, or curative, has assisted direct perception for many millennia. Increasingly, the range of direct perception is being extended by the printing on

96 John Adams packaging of use-by dates and other advice relating to preparation and nutrition. Modern science in the form of knowledge about poisons, vitamins, allergies, metabolism, and genetic susceptibilities also guides the regulators of the food chain. But at the same time that science is illuminating and reducing old risks, it is creating new ones. It produces impressive rewards - in the form of nuclear power, new materials, effective pesticides, new crops, and so forth - but often accompanied by uncertain, and potentially catastrophic, side effects. The appendix applies the typology of figure 4.4 to responses to the perceived risks (or lack of them) of genetically modified food. The perception of all three types of risk is strongly influenced by whether they are seen as imposed or voluntarily assumed. In the lists of contents of the Rewards box and Accidents box in figure 4.3, control and loss of control have been put in bold type because a sense of being in control over the choice of what risks one takes is essential to a sense of moral autonomy. Some risk-taking behaviour appears to be explicable only as the pursuit of confirmation of moral autonomy. Dostoevsky suggests that such confirmation in itself might be considered the ultimate reward for risk taking. Only by invoking such a reward can one account for behaviour that would otherwise be seen as perverse and selfdestructive (Adams 1995). Dostoevsky (1960) puts it this way: What man wants is simply independent choice, whatever that independence might cost and wherever it may lead ... Reason is an excellent thing, there's no disputing that, but reason is nothing but reason and satisfies only the rational side of man's nature, while will is a manifestation of the whole life ... I repeat for the hundredth time, there is one case, one only, when man may consciously, purposely, desire what is injurious to himself, what is stupid, very stupid - simply in order to have the right to desire for himself even what is very stupid and not be bound by an obligation to desire only what is sensible. Of course, this very stupid thing, this caprice of ours, may be in reality, gentlemen, more advantageous for us than anything else on earth, especially in certain cases. And in particular it may be more advantageous than any advantage even when it does us obvious harm, and contradicts the soundest conclusions of our reason concerning our advantage for in any circumstances it preserves for us what is most precious and most important - that is, our personality, our individuality.

While this may seem a plausible insight into the self-destructive behaviour of some rebellious young people asserting their autonomy

Risk and Morality: Three Framing Devices 97 Figure 4.5: Whom Do You Trust?

by defying authority, most of us of more mature years would probably reject it as a description of our own mental processes when confronting risk. But, while our behavioural responses may be less extreme, we are all to some degree sensitive to, and resentful of, both imposed risk, and imposed safety. We want wherever possible to be our own risk managers, and we scrutinize very closely the motives of those who would do it for us. Trust: An Ethical Rudder Confronted by virtual risks, what one believes depends on whom one believes, and whom one believes depends on whom one trusts. Figure 4.5 presents the results from an English survey of trust; it records the percentage of respondents who said they would 'often' or 'always' trust institution X to 'tell the truth about risks.' X referred to the government, companies, and the media, among others (Marris, Langford, and O'Riordan 1996). Least trusted were companies (9 per cent) and the government (6 per cent). These are the main producers and regulators of threats to the environment, and the people likely to have the most useful knowledge about them. Most trusted are friends and family, 78

98 John Adams per cent and 86 per cent respectively. Unfortunately, these are the people least likely to have useful knowledge about threats to the environment. Expressions of trust and distrust are statements about perceived motives. Assuming that the views summarized in figure 4.5 come from a representative cross-section of the English public,3 the message is interesting. The motives of government (hierarchists) and big business (individualists) are viewed with suspicion by the great majority; indeed, in controversies such as those raised by Brent Spar and BSE government and big business were widely held to 'be in bed with each other.' Environmental organizations (egalitarians) appear to have captured the popular moral high ground. Doctors are seen (mostly) as working in the interest of their patients. When the scientist category is unpacked, trust is revealed to be highest for those scientists working for environmental organizations, and much lower for those in the pay of government and big business. The relatively low position of trade unions probably reflects the view that they are defenders of narrow sectional interests rather than the wider public interest. The relatively lower levels of trust enjoyed by religious organizations possibly reflects a suspicion, in a secular age, of religious dogma and those who 'spin' its message. And the lowly position of 'the media' (there are numerous honourable exceptions) may reflect 'risk fatigue' - a state of cynicism engendered by the popular media's habit of sensationalizing every newly discovered virtual risk. The high levels of trust enjoyed by family and friends are noteworthy. It is unlikely that most respondents trust their families and friends to tell them the scientific truth about risks; few people have scientists among their families and friends competent to judge the science. It is more likely that these high scores indicate that they trust them not to lie. Perhaps figure 4.5 has captured the ethical perspective of what might be called the fatalistic majority. According to this perspective • where the truth threatens profit, profit will prevail, • where the truth threatens the electoral prospects of government, government spin will prevail, • you should place your provisional trust only in those who have no obvious motive for lying to you, • you can only trust with confidence those whom direct personal experience tells you you can trust.4 Paranoia flourishes in conditions of hypermobility (Adams 2001). Hypermobile societies are characterized by high levels of anonymity

Risk and Morality: Three Framing Devices 99 and low levels of trust. Both the generators of risk and the regulators of risk are seen by ordinary citizens as remote, self-interested, and unresponsive to their concerns. As mobility - both physical and electronic continues to increase, the size of the fatalistic majority is likely to grow larger still. We are all risk managers and the moral ground on which we stand while performing this task is shifting. The decline in civic engagement and social solidarity documented by Garland and Putnam (Garland 2001; Putnam 2000) is producing a response to risk that is more fatalistic, more cynical, and more selfish. We are increasingly mistrustful of, and resentful of, large institutions that seek to impose either danger or safety upon us. Deference is a thing of the past. In Britain MORI reports 'a dramatic decline in public faith in the way companies use their profits over the past 30 years' (MORI Polls & Surveys 1999). But in 1939 Douglas Jay (who subsequently became president of the Board of Trade) could write, without fear of universal derision, 'in the case of nutrition and health, just as in the case of education, the gentleman in Whitehall really does know better what is good for people than the people know themselves' (Oxford Dictionary of Quotations 1999: 404, citing The Socialist Ca*?1939:ch30). As risk managers most of us, most of the time, are realistic fatalists. We duck if we see something about to hit us; we vote in declining numbers for the hierarchists who oversee our safety, and with diminishing expectations that our vote will make a difference; and we tend to mistrust the scientific advisers of both governments and big business. Large virtual risks such as global warming are met by most with a fatalistic shrug; the Prisoner's Dilemma is played out by strangers who do not trust each other — there is dwindling confidence that individual acts of self-sacrifice will be reciprocated. At the level of individual risks, what used to be seen with the benefit of foresight as risks worth taking are increasingly seen with the benefit of hindsight and the help of clever lawyers as culpable negligence. Our first instinct is not to sue a close friend or relation who negligently harms us, but the rise of the blame-litigation-compensation culture encourages us to sue strangers, especially ones with deep pockets. The result is a society increasingly fearful of litigation and encumbered with pettifogging rules and regulations - a world beset with the practice of defensive medicine, the withdrawal of school trips and sports and recreational facilities, and demands for assessments of every conceivable risk.5 Unless and until we can rebuild mutual trust and widen its scope, these unattractive trends look set to continue.

100 John Adams APPENDIX Ecological Risks and Prospects of Transgenic Plants: A Typology of Bias

Ql) Fatalist

(~) Hierarchist

The whole world is powerless to countermand the actions of powerful, profitdriven corporations: '[GMOs are] being inflicted on unwilling people like myself by Monsanto's unwelcome inclusion of GMOs in the world's food supply ... There are no benefits for the consumer by the inclusion of GMOs, only greater profits for Monsanto.'1

Genetically modified organisms constitute a management problem, soluble by science and regulation. 'We conduct a full scientific risk evaluation. Once we are satisfied, we recommend to Ministers, who have always accepted our advice and who then issue Government approval.' Derek Burke, Chairman of the Advisory Committee on Novel Foods and Processes, explaining how genetically modified foods gain approval in Britain.2 'We had no safety concerns [about genetically modified soya] and the Food Advisory Committee did not require labelling.' Ibid. Government and the scientists it employs know best - but there is a risk communication problem. 'We used to think that all we had to do was to decide whether a novel food or process was safe or not, and a grateful public would accept what we said. We should have known better! Food irradiation, a process I and many others believe to be safe is unusable because of fears connected with the word "irradiation," which go back to the atomic bomb and are fed by concerns about nuclear power stations.' Ibid.

Austin, The Guardian, 16 December 1997 Gallows humour is a common fatalist response to perceived powerlessness.

Risk and Morality: Three Framing Devices 101 \& Individualist • The new technologies are environmentally friendly and will lead to health benefits, an end to world hunger and reduced use of pesticides. "There's no crop or person that cannot benefit. There's a tide of history turning. You can look back, or ask how you're going to feed the world," Monsanto said.'3 • 'Biotechnology is, and has always been, used to make bread, bacon, beer, wine, cheese, yoghurt, pickles and sauces. Humans have been manipulating plant and animal genes for about 8000 years, by breeding and cross-breeding. The difference is that, since Crick and Watson worked out the structure of the genetic code in 1953, it is now possible to work out exactly what is going on when an animal or plant grows faster, taller, or straighter, or withstands rust or blight or brucellosis.'4 • If you can't prove its dangerous assume it's safe: 'Do you cease to approve all new technologies until everything you could conceivably imagine as a risk has been evaluated to the nth degree? ... I am confident it is safe. It is not possible to prove that it is entirely safe.' Monsanto5

\y Egalitarian Abhors 'unnatural'practices; is averse to unpredictability; fears technology dependence and the polarizing socio-economic consequences of the concentration of the ownership of the new technology in a small number of hands. 'Robert Shapiro [CEO of Monsanto]... has to find a market for the products his company has spent billions developing ... The wants and needs of ordinary humans are incidental. This "growth at any costs" attitude on the part of the world's corporate giants is destroying not just our physical environment but the social environment that nurtures human community ... The biotech industry [seeks] to prohibit labelling of genetically modified foods ... The premium now is clearly on ignorance ... Whatever the multi-million dollar spin merchants care to tell us, the scientists cannot guarantee their results ... man's tampering with nature in this way is a recipe for disaster straight out of a horror movie. And you know what comes next. Nature fights back.'6 If you can't prove it's safe assume it's dangerous: 'We cannot just release these things into the environment and hope for the best.' Greenpeace7

SOURCES 1 Lynette Anderson, Food Magazine, November 1997. A true fatalist would not trouble to write to a magazine because there is no point, but this quotation exemplifies what might be termed an informed-fatalist perspective. A recent study of public attitudes in Britain to genetically modified foods discovered that fewer than half the people recruited for focus group discussions of GMOs had even heard of biotechnology in the context of food (R. Grove-White, P. Macnaghten, S. Meyer, and B. Wynne, An Uncertain World: Genetically Modified Organisms, Food and Public Attitudes in Britain (Centre for the Study of Environmental Change, Lancaster University, 1997). Thus fatalists can be assumed to outnumber by a wide margin all the active participants in debates about GMOs. 2 Derek Burke, The Regulatory Process and Risk: A Practitioner's View,' in Science, Policy and Risk (London: Royal Society, 1997) 3 The Guardian, 15 December 1997 4 Bernard Dixon, editor of Medical Science Research, in The Guardian, 18 December 1997 5 The Guardian, 17 December 1997 6 Anita Roddick, Body Shop International, in letter to The Guardian, 19 December 1997 7 The Guardian, 17 December 1997

102 John Adams Notes 1 In Kennesaw in Georgia, one of the states in which the sale of fireworks is prohibited, a local ordinance requires heads of households to keep at least one firearm in the house. 2 Less commonly, usually in the realm of financial risk, we encounter risk management incentives that are biased by a preoccupation with rewards. The annual bonuses of successful financial managers, speculating with other people's money, can be large enough to retire on comfortably for life. If they have an 'accident' the worst that is likely to happen is that they will need to find another job. 3 Figure 4.5 combines the results from two samples in Norwich: A, stratified by housing type (N= 127), and B, three groups: scouts, Chamber of Commerce, and environmentalists (N- 70). 4 In The Origins of Virtue, Matt Ridley describes the conditions necessary for altruism to flourish and concludes that a social scale sufficiently small for individuals to recognize each other is of central importance (Ridley 1997). 5 At the time of writing this conclusion British newspapers were chronicling some of the consequences: • The High Court ruled that a school was 50 per cent responsible for the injuries suffered by a seventeen-year-old boy on a school skiing trip. The boy was injured skiing off-piste, despite having been reprimanded for doing it previously. The court held that a reprimand was insufficient and that the supervising teacher should have confiscated his ski pass. As a result of this and similar judgments the National Association of Schoolmasters and the Union of Women Teachers are now advising their members not to organize school trips: 'our advice is stark. These trips are so fraught with difficulty that we advise our members not to go on them. If something goes wrong, they place their jobs at risk and may face prosecution' (The Times, 26 July 2001). • The implementation of new safety rules was preventing the touching of wildlife in zoos and aquariums. The cost of meeting the requirement that washbasins be provided for hand-cleaning afterwards has led a number of aquariums to stop touching and holding sessions. The reward foregone for obviating a minuscule risk? 'Touching the creatures brings the whole thing to life for children. It makes it more memorable, and that helps the learning process' (Mark Oakley, spokesman for Sea Life Centres, quoted in the Sunday Telegraph, 29 July 2001). • A seventeen-year-old boy was paid £100,000 compensation for injuries

Risk and Morality: Three Framing Devices 103 caused by a 'negligent tackle' in a school rugby match, because 'the time he spent out of school recovering from his injuries meant the grades he received were well below the level predicted by his teachers.' As a result of his low grades he could not pursue a course in dentistry and had to do business studies instead (The Times, 8 August 2001).

References Adams, J. 1985. Risk and Freedom: The Record of Road Safety Regulation. London: Transport Publishing Projects - 1995. Risk. London: UCL Press - 1999. Risky Business: The Management of Risk and Uncertainty. London: ASI (Research) Ltd. - 2001. 'Hypermobility.' Prospect (March): 27-31 http://www.prospect-magazine.co.uk/highlights/hypermobility/index.html Dostoevsky, F. 1960. Notes from the Underground. New York: Doubleday Anchor Evans, L. 1991. Traffic Safety and the Driver. New York: Van Nostrand Rheinhold Garland, D. 2001. The Culture of Control: Crime and Social Order in Contemporary Society. Oxford: University of Oxford Press Hillman, M., ed. 1993. Children, Transport and the Quality of Life. London: Policy Studies Institute Hillman, M.,J. Adams, andj. Whitelegg. 1990. One False Move: A Study of Children's Independent Mobility. London: Policy Studies Institute Marris, C., I. Langford, and T O'Riordan. 1996. Integrating Sociological and Psychological Approaches to Public Perceptions of Environmental Risks: Detailed Results from a Questionnaire Survey. Norwich: Centre for Social and Economic Research on the Global Environment, University of East Anglia, CSERGE Working Paper GEC 96-07 MORI Polls & Surveys. 1999. British Public Increasingly Cynical about Business. http://www.mori.com/polls/1999/mori30th.shtml Putnam, R.D. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster Ridley, M. 1997. The Origins of Virtue. London: Penguin

This page intentionally left blank

PART TWO New Moralities of Risk and Responsibility

This page intentionally left blank

Introduction

In Part II, Violaine Roussel, Michael Power, Alan Hunt, and Kevin Haggerty analyse shifting moralities of risk and responsibility at different levels of society. Violaine Roussel documents how political authorities in France are being held legally responsible for failures in risk management. This process of legalization has created a new politics of responsibility and precaution involving powerful institutions, in particular the judiciary, science, news media, and government. Michael Power examines the shift to a new corporate social responsibility regarding risks: a concern with the external social and environmental effects of organizations, and efforts to address this concern through partnerships with 'stakeholders,' reputation management programs, and recognition of the importance of intangibles such as reputation and intellectual capital. Alan Hunt describes how the increasing pervasiveness of risk discourse has intensified the moralization of everyday life. Such moralization leads to a proliferation of bureaucratic regulation and new self-governing responsibilities for individuals. Kevin Haggerty takes up Hunt's analysis and applies it to the specific case of how individuals manage the risk of criminal victimization. He outlines a new precautionary logic that has emotional, symbolic, and cultural elements well beyond the prevention of crime per se. Violaine Roussel begins with the observation that many risk society theorists, most prominently Beck, focus on major catastrophes to claim that it is becoming increasingly difficult, and in some cases impossible, to hold individuals responsible for risks. However, a case study of France in the 1990s reveals the opposite to be true. Political scandals over major financial, health, and environmental risks were responded to through criminalization of the politicians and government officials targeted as

108 Part Two: New Moralities of Risk and Responsibility responsible. As each new risk was made visible - HFV-tainted blood transfusions, technological failures, Creutzfeld-Jakob disease, and so on - new responsible actors were designated and scandalized through an accelerating process of criminalization. Over the decade, even local politicians and civil servants were increasingly prosecuted for the unanticipated consequences of disasters and accidents and condemned for failing to manage the risks involved. Circumstances previously seen as beyond their control became redefined as imprudence and negligence for which they were responsible. Roussel describes the political processes involved and explanations of them. Magistrates had recently become more activist in pursuing politicians involved in financial scandals; such politicians were prosecuted as part of an effort to change institutionalized patterns of privilege and corruption. Success in these cases made prosecution of politicians regarding other risks and responsibilities seem more plausible. 'Political risk taking' became a central denomination, and failure to take precaution a central denunciation. Heterogeneous cases were merged under the labels of risk and precaution, with attendant attributions of criminal and political responsibility, leading to the progressive redefinition of their meaning and deployment. While many analysts have focused on the increasing criminalization and responsibilization of poor and marginal populations under the banner of risk and responsibility, Roussel demonstrates that these categories have also been effectively deployed in discursive struggles among the powerful. Roussel shows that by the mid-1990s, this field of criminal responsibility was effectively reorganized around conceptions of imprudence and precaution. The precedents that had been set by this time were not in the law per se, but in the successful political mobilization of the law by a range of actors, only one of which was the magistrates themselves. The success of these actors was not so much based on their enunciation of broader moral principles; rather, the articulation of professional ethics focused on the procedural requirements of their respective institutions. For example, judges invoked the ethics of legal due process, journalists the ethics of transparency and neutrality, and medical researchers the objectivity of their scientific procedures in identifying risks and their possible effects. All of these actors thereby claimed that they were procedurally and technically constrained, and therefore only doing their jobs rather than engaging in morally motivated political struggles. As successful prosecutions evolved, it became increasingly difficult for political actors to deny criminal responsibility. Instead, they were forced to

Introduction

109

develop discourses of precaution and make preventive measures visible to address the new risks for which they were now deemed fully responsible. They articulated new rules of precaution, demonstrated that they were playing by these rules, and thereby argued that they were beyond reproach. Roussel's analysis gives life to Mary Douglas's observation that risk is used discursively to both discredit and legitimize particular policies, and through policies to protect people from predatory institutions and institutions from predatory people. Risk, responsibility, and precaution are discursive moral labels with adhesive qualities in the mobilization of political scandals. Their vagueness, polysemy, and elasticity make them especially effective. For example, Roussel argues that 'irresponsibility' was attributed to politicians in order to make them play the political game in accordance with new, tacit normative rules favoured by magistrates and other actors involved in the mobilization of scandal. For politicians, the new route to responsibility and respectability was to avoid being caught in acts that appear imprudent or negligent - especially acts that seemed to prioritize their own interests - rather than actually to assume direct control over public risks in uncertain situations. Tacit normative rules that frame what can be done and said in the political field were now formatted in the language of risk, responsibility, and precaution. Roussel shows that it was not only politicians who became locked into particular courses of action in these discursive struggles. As other discourses - for example, those of law, science, and news - were invoked politically, 'validation networks' were formed among the institutions involved. Each actor was in part constrained to think and act through the language of risk, responsibility, and precaution that had become publicly validated through his or her previous participation in risk controversies. Nevertheless, the effects of participation in validation networks were not equally distributed across the actors and institutions involved. For example, when political actors maintained that they were relying on scientific judgments about risk, the scientific sphere increased its hold on political actors. Similarly, when political actors spoke in the name of new legal responsibilities for precautionary measures, they gave up some autonomy to magistrates and the legal institution more generally. Indeed, a major effect of these developments over the 1990s was an extension of political accountability to new forms of criminal responsibility that authorized more judicial control over political action. Questionable political practices previously considered normal or untouchable

110 Part Two: New Moralities of Risk and Responsibility were now judicially interrogated, and political control over judicial activities unravelled somewhat. Roussel observes that the new public definition of risks, and of political decision makers' responsibility for them, may serve as an alternative means to achieve welfare in an era of neo-liberalism. Criminalization of political actors extends the moral boundaries of public intervention in dealing with risks and legitimates a broader role for the state in risk management and preventive security. This reactivation of welfarism does not result from ideological intentions of the actors involved, but is rather the effect of discursive struggles which have myriad contexts and goals. New quasi-public agencies, linked to private corporatist logic and interests, have arisen to provide expertise and assist in the precautionary project. Addressing risk management and precautionary measures, these agencies are also expected to absorb responsibility and avoid prosecutions of political actors. They are the locus of a new politics of visible precaution, in which dramatic displays of extreme precautionary measures are intended to symbolize responsibility. One example is the decision to destroy 50,000 animals that had been in contact with English sheep at the beginning of the foot and mouth disease crisis, even though no ill sheep had been identified in France. As Roussel concludes, we live in a risk society to the extent that actors from various spheres assimilate different risks in their speeches and behaviours, make them equivalent, and thereby give them form and reality. Michael Power shifts analysis to the level of organizations, especially corporations. He notes that corporate social and ethical responsibility has traditionally been based on internal concerns, such as the health and safety of workers. However, in the past three decades there has been increasing concern with external social and environmental effects of organizations, creating the new corporate social responsibility movement. This movement, far from unified, is an assemblage of ideas and practices based on critiques of corporate activity. It has been accentuated through major technological disasters, such as the Exxon Valdez and Bhopal cases. It has been connected to the social audit movement, which is aimed at making corporate activity and impact more transparent, controllable, and amenable to change. In this connection there has arisen a new breed of 'organizing organizations' which create standards of corporate conduct. In the United Kingdom there is now even a Cabinet post of Minister for Corporate Responsibility. The corporate social responsibility movement is directed not only at changing corporate behaviour, but also at changing the concept of the

Introduction

111

corporate entity itself, its very licence to operate in society. Similar to what Roussel documents in the case of new moralities of risk and responsibility in the political sphere, in the corporate sphere there are strong efforts to import societal concerns about major risks into organizations. This effort includes behavioural change: altering the motives of workers and constructing a morally culpable and blameworthy management. External participants do not take the moral high ground as much as cohere with the utilitarian morality of enlightened corporate selfinterest. Social concerns are aligned with business values, but with the intention of creating good social effects. For example, tax incentives may be used to induce desirable policy outcomes, such as reduced carbonmonoxide emissions, thereby effecting internalization of external social goals. Power's project is to understand how internalization occurs through risk management systems: the habits they form and the routines they support. Risk management systems can embed 'interested' behaviour in an organization, sensitive to ideals of social responsibility, without directly appealing to either individual morality or economic self-interest. The goal is to 'hard-wire' a kind of'enlightened self-interest' into organizational life in a way that blurs the boundary between self-interested and other-oriented behaviour. Power asks, 'How does risk management process morality?' The organizational imperative to manage risk has the potential to internalize external interests and align them with the organization in new ways. A new 'window of responsiveness' has been opened, braiding ethical aspirations with operational pragmatism and incorporating them into an internally programmed system for corporate responsibility. Power identifies core features of the new risk management. For the first time risk management has become a senior management task. One reason for this development is an emphasis on standardization and integration of risk management practices across the organization. This emphasis in turn creates a body of knowledge of risk across sub-units of the organization (treasury, health and safety, etc.). Many organizations have created the new post of corporate risk officer to coordinate and address this body of knowledge. Moreover, various managerial posts are re-articulated around the language of risk. The new managerial discourse of risk is directed at making new sets of organizational actors responsible and accountable for their actions. This involves an evaluation of managerial competence itself in terms of the ability to manage risk and holding managers blameworthy when risk management fails.

112 Part Two: New Moralities of Risk and Responsibility Again, some parallels to Roussel's analysis of the political sphere are evident. Power sees this new management agenda respecting risk as a way of forcing particular responsible people to act in the interests of broader constituencies and human welfare. Power considers three dimensions of the new risk management. First, the concept of 'stakeholder' is becoming integral to the enterprise. This concept is attractive because it has both strategic and ethical meaning and constructs a pragmatic bridge between managerial and corporate social responsibility agendas. The everyday routines of risk management can de facto enfranchise external groups by building in responsiveness to their concerns. On the other hand, when incorporated into risk management, the stakeholder is often treated as a risk. The question becomes who might blame and thereby damage the organization. The answer is to shape risk management so that it incorporates the stakes of others without allowing them to hold rights, and without creating a company duty to them. Second, there is a shift from the old, defensive 'public relations department' approach to a proactive 'reputational assurance' approach in organizational communications. This new approach emphasizes greater information accessibility and disclosure, social and environmental impact studies, proactive mechanisms for listening to stakeholders, and staff training and codes of conduct. Reputation forms a basis for the structural coupling of business and society, because reputational assets are relational with stakeholders and add substantial value to the company. Reputation as mediated by risk management allows the corporation to negotiate and define its boundaries with external parties. Nevertheless, reputation assurance is a highly problematic field because reputation is a product of disparate corporate practices that are always percolating with trouble. As Power observes, reputation assurance involves 'managing an outcome independently of its multiple causes' (157). Third, intangibles have become the most substantial component of company assets, and tangible assets are the residual. Intangibles pose new challenges for risk management, especially the problem of how to measure and value assets. For example, in knowledge-based economies 'human assets' and 'intellectual capital' are rich resources yet difficult to quantify. Reputation is itself a source of value creation, but lies outside the organization's boundary in interdependent relationships. These relationships are full of webs of economic and social significance, yet difficult to capture as an asset. Nevertheless, in a world of mergers and acquisitions, intangibles such as intellectual capital and reputation can be translated into very tangible differences in asset value.

Introduction

113

Power sees stakeholder, reputation, and intangible asset risk management as interconnected and crucial vehicles for the de facto responsibilization of organizations. They yield potential for a new corporate social responsibility that internalizes wider values and concerns into everyday organizational life. They make it possible to move beyond binary oppositions such as 'profit' and 'morality' in assessing corporate activity by forcing organizational actors to be self-governing regarding risk, response ability, responsibility, and responsiveness. In so doing they create potential for new forms of public accountability regarding the corporate sphere. This potential arises in new 'organizing organizations' such as AccountAbility, an institute of social and ethical accountability. As the name signifies, accountability is entwined with risk management systems that provide an ability to account for one's actions in organizationally sanctioned ways. Alan Hunt locates his analysis at the level of individuals and their experiences with moralization through risk discourses. This moralization is experienced through the proliferation of bureaucratic regulation in the everyday world and the attendant expansion of responsibility for self-governance. Far from the deregulation urged in contemporary neoliberal discourse, the rise of risk and its moralization entail greater regulation, reaching into the minutiae of everyday life. In this respect contemporary liberal risk regimes exemplify the paradox of liberalism more generally: concern with governing too much, but prone to govern more. Consistent with Mary Douglas, Hunt observes that moral discourse used to be more explicit but now functions more regularly through other discursive forms, in particular the discourses of harm and risk. In the hybridization of risk and morality, the moral dimension is made utilitarian and aimed at benign choices regarding harm to others and self. For example, in the name of risk and harm, alcohol consumption is itself consumed by risk and morality discourses about harm to others (e.g., Mothers Against Drunk Driving), harm to self (e.g., drinking during pregnancy), and the need for more principled governance (e.g., by parents of children, employers of employees, bar managers of patrons, partygoers of drivers, and so on). Risks do not go away. They perpetually circulate among the alternating registers of security, danger, anxiety, and fear. This circulation can only be frozen momentarily through the immediacy of choices about how to act under descriptions of risk. These choices are not based in freedom (complete autonomy of decisions), but in whatever agency (discretion or relative autonomy of decisions) is enabled through the

114 Part Two: New Moralities of Risk and Responsibility regulatory apparatus of risk. Choice is difficult for the individual because of the pluralization of expertise about risk, which provides not only different calibrations of uncertainty but also paradoxical and sometimes contradictory recommendations about how to act responsibly. As the need to know about risks spirals, experience of the limits of knowledge capacity increases, which in turn shapes perceptions of danger and the insecurity of choice itself. Made responsible for an expanding set of choices about risk, individuals are compelled to become moral entrepreneurs of the self. They must become their own policy makers, charting principled courses of action that exhibit their capacity for self-improvement in self-controlled ways. They use risk to problematize everyday life morally, to articulate who they are and what they do in relation to the threatening world in which they live. Since risk is based in probabilistic thinking about the likelihood of future events, it engages the individual in consumption of the future. That is, it brings the future into the present to choose what to conserve, what to insure, what gambles to take, and so on. Consuming the future involves both action and inaction, and it is particularly motivated by fear defined as heightened sensitivity to risk. In this respect, risk is a kind of objectification of anxiety. In the everyday world of risk and moralization, actions that fail to overcome the risks associated with identified harms are declared irresponsible. Thus the proliferation of risk discourses is accompanied by an expansion of the contexts in which individuals can be held irresponsible. Through new discourses of risk and responsibility, even big risks such as environmental threats can be located in individual choices: save a tree, use a bicycle rather than a car, and so on. In the moral regulation of risks the individual is held to the responsible standards established by the institutions concerned, regardless of whether she accepts responsibility on the personal level. Of course, the individual can deny responsibility, blame others, and try to transfer responsibility and the risk that goes with it. Individual smokers suing tobacco companies provide a case in point, as do crime victims who demand greater compensation and protection from private insurers and governments. The discourses of risk feed into blaming the victim, and thereby making her responsible, and blaming the risk-pooling institution, and thereby making them responsible. These two faces of risk and responsibility are both projects of moral regulation. Another paradox arises. The politics of displacing responsibility for risk entails 'deresponsibilization' and 'demoralization' as well as their

Introduction

115

opposites. While there are sporadic outbursts of moral rage - for example, regarding a paedophile being released from prison, or a government that fails to supplement private insurance compensation following a disaster - the targets are selective and partial. Otherwise, broader discourses of risk and morality do not echo in the corridors of private interest and convenience, and moral indifference erases concern for others. Hunt emphasizes the ways in which risk is not so much a quantified and calculating basis of rational action as a trope for experiencing and expressing anxieties about threats. This is precisely the point taken up by Kevin Haggerty in his analysis of the rationalities of personal crime prevention. Haggerty makes a distinction between risk and precaution. Risk is based in statistical reasoning about the probability of unwanted outcomes. Risk management addresses rational means for reducing that probability, as well as insurance-based schemes of compensation when things go wrong. Precaution involves emotional, symbolic, and cultural dimensions of how threats are perceived and responded to, beyond the rational probability calculus of how to minimize harms. Precaution is based in different rationalities and logics and speaks to matters in addition to the prevention of harm per se. Haggerty uses the substantive example of personal crime prevention to illustrate four features of precautionary logic. First, precautionary logic is non-calculative. Individuals cannot be all-calculative about the specific crime risks they face because of the general problem of moving from aggregate data to situational decisions about an individual circumstance. Moreover, they depend on information sources that have systemic biases. Experts on crime risks speak from institutional positions that reflect particular ideologies, interests, and agendas. Mass media portrayals of crime risks are replete with overreactions to events that are statistically improbable. They induce 'moral panics' based on fear, and subsequent precautionary measures among people who are least likely to be victimized. Second, precaution emphasizes catastrophic potential. Crime is pictured in worst case scenarios: homicide or other serious violence, abduction of loved ones, or loss of prized possessions whose sentimental value cannot be compensated in monetary terms. Preventive security technologies are purchased, but they never seem quite enough, in part because their very presence is a reminder of (in) security. Compensation for loss through normal insurance mechanisms is also not enough, because the loss is too serious and personal to be remedied in cash.

116 Part Two: New Moralities of Risk and Responsibility Indeed, such compensation makes little sense compared to the logic of preventive security that promises avoidance of events with irreparable and irreversible consequences. Precautionary activity is aimed not at reducing probabilities but rather at eliminating possibilities. Third, precaution is driven by a permanent state of anxiety and fear. This point is exemplified in how security products are marketed. The responsible party wishing to avoid catastrophe is sold surveillance technologies that promise to cover every angle. In an especially telling example, Haggerty cites an advertisement urging that nannies be watched through a 'surveillance camera inside a teddy bear, a move that hardwires the technology of insecurity into the icon of childhood security' (205). Many other technologies similarly hit home, for example, home surveillance cameras viewable from an Internet connection, or the Timex Ironman sports watch, which allows a jogger to monitor a home security system from the watch's display. In such marketing of (in)security, fear is promoted as a valuable emotion that should be perpetually kept alive to motivate the elimination of catastrophic potential. Fear is celebrated 'as an effective pre-rational warning system' (206). Marketing (in)security can only promise peace of mind by failing to provide it. The mind is kept at war. Fourth, precautionary decisions are culturally loaded. Strident efforts to avoid catastrophic consequences of crime signify personhood, especially being a responsible moral agent. For example, women's subjectivities are fashioned around how they address the threat of sexual assault, parent's subjectivities are fashioned around how they address the threat of child abduction, and so on. Again, marketing (in) security features ads that play on lifestyles and identities, with market segment distinctions and aesthetics. Security products become technological gadgets marketed just like computers or stereo systems, seductive toys in search of desires, needs, and uses. Security products also fit into zero-sum games of consumer society. Each individual gain in personal security is made at the expense of neighbours, who are thereby demoralized and made unequal. Neighbours respond by similarly overspending on security to avoid being the odd, insecure one out.

5 New Moralities of Risk and Political Responsibility VIOLAINE ROUSSEL

Everywhere we search for responsible people, it is generally the instinct of judging and punishing that is at work. Nietzsche, Twilight of the Idols, 1888

A number of sociologists (e.g., Giddens 1990; Beck 1992) suggest that specifically modern risks exclude the identification of responsible parties. For these authors, the current age faces life-threatening risks of high intensity, but these are not risks anyone chooses to run and there are no 'others' who could be held responsible. These features of contemporary risks explain the sense of foreboding characteristic of the current age (Giddens 1990: 131). Political scandals regarding risks to health and security in France during the 1990s suggest the opposite. These scandals indicate the simultaneity between the rise of new visible risks and the designation of new responsible actors. New problems of political responsibility were addressed by various actors, including magistrates, political authorities, journalists, and activists. The major scandal of the decade was blood transfusion with the HIV virus. This scandal was the first in a series that accelerated a process of political responsibilization and criminalization. With this case, political decision makers were increasingly prosecuted in relation to their management of new risks to public health and security. Nevertheless, the new discourses and activities connecting public risks and political responsibility did not appear spontaneously out of nowhere when the blood scandal broke. They originated in a larger series of scandals related to financing of political parties and to personal enrichment of political

118 Violaine Roussel actors that emerged in the early 1990s. Over the decade, judicial practices converged with other social activities to create a new definition of political responsibility that incorporated criminal responsibility and paved the way for prosecutions. The new risks were also addressed in the news media, where journalists pointed to political crisis. Political actors shared the crisis framework, including the discourse of 'facing up to their responsibilities.' This convergence in the assignment of responsibility to political decision makers was not the result of some natural political response ability in this field. Rather, it was the consequence of various social mobilizations focused on public health and security matters. The collective construction of a new political responsibility for public risks has to do with the changing relations between several institutions during the scandals, especially with transformations in the relationships between justice and politics. This chapter addresses how such changes occurred and were progressively labelled in terms of risk and responsibility. It focuses on the strategies through which political actors tried to sustain their legitimacy. In particular, it shows how political actors related the new discourses on health and security risks to judicial efforts to criminalize their responsibility for such risks. Primary data are drawn from interviews with magistrates and political actors involved in political scandals in France during the 1990s (Roussel 1999). Criminalization and Responsibilization of Political Action Public discourses about the new political responsibility for risk arose during the blood contamination scandal. This scandal had a strong impact on the French political world and made the front page of leading newspapers for several months. Financial scandals which had developed in the beginning of the 1990s, by contrast, did not carry the same emotional load. In the blood contamination case, three ministers, including the former Prime Minister, Laurent Fabius, as well as state employees, were prosecuted for criminal offences. The ministers appeared before a special court, the Haute cour de justice, which was reformed in 1993 in the context of the blood scandal. The new Cour de justice de la Republique discharged two ministers and condemned lightly the third. One of the discharged, the Health Minister, Georgina Dufoix, was quoted widely by the media as claiming to be 'responsible, but not guilty,' meaning politically but not criminally responsible. The question of a new political responsibility was, from this moment, at stake in

New Moralities of Risk and Political Responsibility 119 political debates, in both formal parliamentary and informal arenas. Several strategies of justification or disqualification from politicians, journalists, lawyers, and associations developed around the problematic of risk and responsibility. The link forged between political and criminal responsibility did not arise spontaneously in 1992, when blood contaminated with HIV was revealed in Le Monde. This correspondence was prefigured in previous successful prosecution of political leaders for financial crimes. The blood contamination scandal was placed in the context of other scandals, through which prosecuting political actors had become a possibility for judges or people appealing to courts. Nevertheless, after the blood scandal, the new forms of prosecution against political actors proliferated. Local political actors have been increasingly prosecuted for the unanticipated consequences of natural disasters or accidents and often condemned for failing 'to manage the risk' related to the technological facilities installed in their towns or areas. Previously, catastrophes were understood as natural disasters or accidents caused by the irrepressible forces of circumstance, and therefore impossible to control. After 1995, they were more likely to be depicted in court or in public debates as questions of imprudence or negligence, for which public decision makers must be held responsible. For example, a case in which children drowned during a school trip to Grenoble, after a dam broke, was constituted in court as a negligence case. The town of Grenoble was fined 500,000 francs, upheld on appeal. Several other accidents with public facilities have since given rise to criminal prosecutions. All of these various cases have in turn been related to other financial, health, and security matters in a global movement of responsibilization and criminalization of political actors. Political Criminalization and the New Relation between Justice and Politics Until the middle of the twentieth century, French judges and political actors had a common history and belonged to the same small world of local notables. They came mainly from the same background and shared the same social life and perceptions. Judges were highly politicized and often directly involved in political action (Royer, Martinage, and Lecocq 1982). After the Second World War, and particularly after 1960, this social and political proximity gradually decreased. Systems of local domination changed, and clusters of notables gradually disintegrated.

120 Violaine Roussel The social status of magistrates, and their position in relation to other groups, placed them at a new distance from political actors. This distancing was reinforced by the creation of a national school for judges in 1958. With the establishment of the Ecole Nationale de la Magistrature, middle-class individuals acceded to judicial positions. The school provides a homogeneous training in law and prepares a unique profession (Bodiguel 1991; Boigeol 1989). All magistrates are henceforth civil servants and follow a common career with standard professional stages. Individuals are expected to move, during their career, from one position to another and to be, for instance, prosecutor, and later judge (or vice versa). The judges recruited through the school share similar experiences and hold similar beliefs about their role. New professional models and a new judicial ethics are the result. There is a strong emphasis on judicial independence from politics and social influences, which no magistrate would have conceived in the nineteenth century. This emphasis includes a view that political actors should be subject to prosecution just as any other individual. Such perceptions of the judicial profession and the political world functioned as preconditions for the criminalization of certain political behaviours that occurred in the 1990s. However, these perceptions are not sufficient explanations, since they do not elucidate why only some judges committed themselves in these prosecutions and why these prosecutions only began in the 1990s. What was new in France in the 1990s was not the occurrence of scandals per se, but the massive accumulation of successive scandals involving political actors, including members of Parliament and successive governments. This trend began in 1991, with the Urba scandal over the illegal financing of the socialist party. Many other scandals followed, involving both right- and left-wing organizations. The conditions of judicial action in the first political scandals of the decade are quite specific with regard to the context, the type of actors involved, and the modes of action chosen. The Urba case is typical. In 1990, several collective mobilizations, demonstrations, and strikes of magistrates developed, which had never happened before. The examining magistrate in the Urba case was himself one of the activists of these mobilizations. The action chosen was collective and embedded in a trade union activity. The Urba judge was a member of the Syndicat de la magistrature, the left-wing trade union of magistrates, as were most of the other judges with whom he prepared his protest strategies before and during the Urba case. The judges of this early period saw their actions as political, anti-establishment activities, similar to some previous mobilizations the Syndicat had initiated in the 1970s and the 1980s. In

New Moralities of Risk and Political Responsibility 121 interview, the Urba judge explained that he had wanted to contest an unfair system of privilege, beneficial to political decision makers. 'Our problem was to move the judicial action to the Political field, with a capital P, saying, "Is it possible to have in our country, in a state supposed to follow the rule of law, two systems of justice?" We only wanted to say no.' Thus, the first prosecutions of the 1990s were rooted in a more traditional type of behaviour, invented two decades earlier in the judicial world. Moreover, the magistrates involved at this stage were not typical members of the judiciary. The difference is especially striking regarding the Urba judge's career path. He is one of the very small number of magistrates selected not through the national examination open to students after a master in law, but directly from within the civil service system. Furthermore, he did not possess the standard training provided by the Ecole Nationale de Magistrature. He also imported the culture of the professions he practised before becoming a judge, in which he used to be more autonomous and was less subject to control through a hierarchical system. This judge defined his role in ways that were closely linked to the conditions of his recruitment in the judicial profession and to his preliminary experiences. However, being appointed later than most of the judges, he also knew he had little hope of becoming a high-ranking magistrate. Thus he found it possible and appropriate to behave in a way considered by other judges as 'too risky' or as 'professional suicide.' At this time, the common perception was that judges who encouraged the prosecution of political scandals would be stopped and penalized so that the scandals would not develop further. A lot of people are saying that the case [I investigate] will have some repercussions on my career. Anyway, two possibilities: if I fail... I will have to stagnate professionally during a period, and if I don't fail, it will have created some jealousies, so that my career will not necessarily benefit from it. (Examining magistrate, 1993) I think that there will be a renewed attack and that politicians will react... There is a terrible anti-judiciary discourse today in political circles. (Examining magistrate, 1993)

Such anticipations proved to be wrong. Judges were not stopped, although they expected to be, because they benefited from the cooperation of other actors. The judges' activity converged with the action of some journalists, most of the police officers involved in the investigations, and

122 Violaine Roussel with that of middle-rank magistrates in Appeal Courts. Here we witness an example of multisectorial mobilization, that is, mobilization developing simultaneously in various social sectors (Dobry 1996). Such mobilization involves different groups of actors with heterogeneous perceptions, interests, aims, modes of action, and legitimization principles. The unexpected judicial successes in the first scandals, in which some national decision makers were convicted and sentenced, lead to the diffusion of similar modes of action within the judiciary. Prosecuting political actors appeared more and more possible, with a lowering level of risk for judges. 'When we see the other going further, we know we can also go further. It is obvious that we move the limits step by step. It is very important not to be alone, even if it is in totally different fields, it is important to know where we can go and up to which limit' (Examining magistrate, 1993). After the middle of the decade, judges were even able to anticipate some symbolic profits related to media coverage of the scandals. Prosecutions were no longer the protest behaviour of a few heterodox and atypical magistrates: the success of the first prosecutions made it thinkable for ordinary judges to engage in the same kind of behaviour. Prosecuting political actors increasingly became part of the usual way of doing the judicial job, whereas some years before judges had restrained themselves from initiating such prosecutions. 'Nobody can imagine that you simply do your job! The more we go up, the more we come close to the political spheres, the more it is difficult. In these conditions, I don't see what could be the interest of a magistrate when he investigates political cases. Except if you consider that we do our job. Some of my colleagues perhaps only want to do their job, they do not want to be hindered all the time' (Examining magistrate, investigating a political scandal, 1996). In this process, the more scandals developed and politicians were convicted, the more new magistrates committed themselves in new prosecutions. Judicial embedded behaviours constituted, as well as the series of scandals themselves, a self-reinforcing process. The first results of previous scandals - resignation of prosecuted members of the government, judicial decisions sentencing political actors to ineligibility or sometimes to prison - carried an effect of authorization to act for other judges, police officers, lawyers, associations suing for damages against politicians prosecuted, journalists disclosing new cases, and so on. The multiple uses of justice in the scandals and their successive outcomes hung together in a developing process (Roussel 2000, 2002). During the scandals, the actors' practical knowledge about strategies

New Moralities of Risk and Political Responsibility 123 opened or excluded was continuously reshaped by current activities and confrontations. Judicial and non-judicial activities together constituted a wide system of embedded actions through which cognitive and discursive frames were produced and changed. By the end of the decade, judicial actors and other protagonists of the scandals shared a common perception of the foreseeable future and currently possible actions, according to which judges had 'won' against political actors. Controlling prosecutions and judicial activities, at least in a visible way, appeared more and more difficult and potentially dangerous for political actors. 'Today, it would be a madness to imagine compelling an examining magistrate to investigate or, on the contrary, not to investigate ... The politicians who would play such a game would pay an exorbitant price for it' (Examining magistrate, investigating a political scandal, 1996). The changes that occurred in the relationships between justice and politics in the context of financial scandals were directly connected to the extension of these transformations through health and security scandals. The later scandals contributed to the self-reinforcing process described above. For the magistrates involved, the possibility of prosecuting government members in the context of the blood contamination scandal derived from the shift in perceptions already under way. Moreover, judicial decisions in early cases encouraged not only other magistrates to commit themselves in such modes of action, but also actors outside the legal institutions to refer new matters to court. Lay people increasingly pressured judges to satisfy their claims against political decision makers. These processes explain why individuals who have cancer they attribute to radioactivity felt authorized and found it appropriate only in 2001 to lodge a complaint against members of the government who had previously denied the crossing of the radioactive cloud from Chernobyl over France and the existence of a threat to public health. In these mobilizations, the field of possible and legitimate behaviours for magistrates, as well as for political actors, journalists, or complainant associations, is enlarged. In particular, the recurring uses of the categories of political and criminal responsibility, and of risk and precaution, lead to the progressive redefinition of their meanings. Judicial and Social Categorization THE JUDICIAL INVENTION OF 'EXPEDIENTS'

The chief denomination used in the blood contamination scandal - and later in other cases regarding security and health matters - centred on

124 Violaine Roussel political risk taking. Local and national representatives of government, state employees, and sometimes scientists from pharmaceutical laboratories, were denounced for not having exercised sufficient precaution. These actors were often prosecuted according to a common range of legal offences related to the idea of imprudence or negligence in the management of risky situations. Various heterogeneous cases were merged. Talking about risk and precaution in the blood scandal, or later in the mad cow disease scandal, might appear natural. But such a unified reading of diverse situations did not arise immediately; it was constituted through interactions and in the course of the interdependence process already mentioned. The legal offence of 'endangerment of other's life,' which belongs to offences involving imprudence or negligence, has gradually come to be seen by judicial actors as the relevant legal charge to prosecute successfully political decision makers in public health cases as well as in public security cases. However, this legal category was only created through a reform in 1996 and therefore was not available at the beginning of the blood contamination scandal. When that scandal first emerged, the decision was taken to prosecute certain political, economic, and medical decision makers under legal provisions regarding 'poisoning' and 'complicity of poisoning.' This decision originated mainly in the desire to secure a severe sentence on conviction, corresponding to an intentional rather than involuntary offence. The accused rejected the idea of any criminal responsibility, admitting only to the civil offence of selling defective products. The examining magistrate failed to put the criminal offence of poisoning before the higher courts, and finally in June 1998 before the Cour de cassation, the highest judicial court in France. A number of actors learned from this case that the charge of 'poisoning' does not achieve the expected results. Simultaneously, the successes in using 'endangerment of other's life,' an offence raised in 1996 with the penal code reform in the prosecution of three ministers involved before the Cour de justice de la Republique, stabilized a specific interpretation of this new text and made it the relevant device for similar cases. Of course, the 1996 reform was not intended to give rise to the prosecution of politicians in health and security scandals. The 1996 law is an exception to a fundamental rule of criminal law - the rule of mens rea or intentionality of criminal offences - extending the scope of criminal responsibility to imprudence or negligence offences. The court appreciates if the behaviour prosecuted has respected conditions of 'normal diligence.' The invention of a criminal offence of this type was intended in particular for situations of road and occupational safety. But

New Moralities of Risk and Political Responsibility 125 the successful mobilization of this offence in the blood contamination case modifies the meaning of the law itself, rendering it the most appropriate means by which to bring decision makers to court in public health and security cases. When a large number of local political actors and state employees were prosecuted for negligence or imprudence in cases of natural or technological accidents - for example, regarding accidents in mountain sports (Caille 2000) - after 1996, it was not the result of any immediate and natural implementation of legal texts. Rather, the initial successes with negligence and imprudence prosecutions led to new uses of the same legal tool. Through successive appropriations in various cases involving accidents and disasters, this legal move to criminalization became the most expedient route to obtain a conviction. If a judge wants to prosecute a mayor criminally in relation to an accident with public installations in his town, he now knows that it is possible and could be efficient. Moreover, he knows how to do it successfully. This practical knowledge also entails sidestepping the problems posed by other legal moves, more adapted, theoretically, to the case. The same type of mechanism led most of the judges involved in financial scandals during the period to use the legal category of 'misuse of a company's assets' in order to obtain the conviction of political actors. For example, in political financing cases, since the offence of 'corruption' had been impossible to prove, the examining magistrates commonly turned to the 'misuse of a company's assets' as a useful expedient. Thus, they suggest that the law itself should be changed: 'We would not have to rack our brains to find expedients. The expedient for corruption, it is the misuse of company's assets, and if it shall be suppressed, it would be impossible to prosecute, that's all' (Examining magistrate, investigating a political scandal, 1996). As these examples show, judicial activities during the scandals contributed to the stabilization of new meanings of legal texts and shared perceptions of the 'good uses' of the rules. They illustrate the building of jurisprudence in action. The creation of a completely new offence in court is extremely rare in the French system, but the reinvention of a previously existing offence through new uses is common. The actions of magistrates thus objectify, in legal form, new definitions of criminal responsibility of political actors. In other words, magistrates have altered the frontiers between what is 'only' part of moral or political responsibility and what belongs to the scope of criminal responsibility and carries judicial effects. After the mid-1990s, this field of criminal responsibility was reorga-

126 Violaine Roussel nized around the notions of imprudence and precaution. Accusations of lack of 'precaution,' which is not a legal offence, were increasingly invoked to stigmatize imprudent, risky, irresponsible behaviours. By the end of the nineties, 'precaution' (or 'lack of precaution') became the appropriate terms to employ in celebrating or denouncing political actors, and to describe the situations they were involved in. These discursive strategies were strongly linked to increasing referrals to court. They entered in a cumulative process in which the blood contamination scandal served as a precedent for later scandals (Chateauraynaud and Torny 1999; Roussel 2000, 2002). The new mobilizations of the precaution rhetoric, already proven efficient, fostered its success and the chances of future uses. DIFFUSION OF CATEGORIES AND STABILIZATION OF REPRESENTATIONS

With the development of the prosecutions, the political refusal of criminal responsibility became increasingly difficult and illegitimate: the public discourses of various actors were reshaped around the question of precaution. Confronting various denunciations, political actors felt compelled to develop systematic discourses on precaution and visible preventive measures in order to show that they always play by the rules and that no reproach should be addressed to them. Heterogeneous activities, deriving from various social mechanisms, were cross-assimilated under recurring accusations of 'imprudence,' 'risk,' and 'lack of precaution.' 'Faulty' activities of civil servants, consultants, ministers, scientific actors, doctors, and members of pharmaceutical laboratories were pointed out. The precautionary principle also functioned as a general formula — an equivalence format—used to establish an equivalence among various situations and what they imply in several social spheres (Thevenot 1986). The obligations of a mayor with respect to sport installations in his town, what a minister must decide regarding the support of pharmaceutical laboratories or in terms of agriculture policies and food safety, what private laboratories or heterogeneous industries are expected to do, and so on, are merged under the unique denomination of 'precaution.' Despite the diversity of these scandals and proceedings, the homogenizing labels of 'lack of precaution' and 'new political responsibility' apply in relation to defaulting management of risks. Behind this apparent unity, precautionary behaviours, as well as risky actions, reflect heterogeneous aims, interests, beliefs, and practices, depending on the type of actors involved. The discourses denouncing

New Moralities of Risk and Political Responsibility 127 the lack of precaution in dealing with risky situations also reflect different argumentative registers. Inappropriate political behaviour is stigmatized by political challengers as political irresponsibility, byjudicial actors as a legal offence, by scientific experts as technical and scientific error, by the media as a moral scandal. The various activities that together define a political action as irresponsible risk taking are not equally moral crusades. On the contrary, they refer to different cognitive frameworks, different ranges of signification, to establish the (dis) qualification of the situation. Moral responsibility was specifically addressed in the news media, through the dramatization of the risks taken and responsibilities defaulted. Journalists sometimes became moral crusaders (Becker 1963), emphasizing the corruption of a political system 'in crisis.' However, associations or individuals involved as complainants in health and security cases also acted as moral crusaders. Moreover, judicial ethics converged with journalistic and other professional ethics to demand a new 'right of inspection' for political behaviours. None of these moralities displayed an abstract moral form. Rather, they were articulated at the level of practical ethics. Judicial ethics appeared as a legal and technical requirement; media ethics meant transparency and neutrality in releasing facts; medical ethics implied scientific procedure. Since their legitimacy was at stake, all the actors denied that their action was a matter of moral and personal choice. On the contrary, they shared the same kind of legitimization discourses, asserting that they had no choice but to behave according to legal rationality, science, technical constraints, or neutrality. Since they were only 'doing their job,' their actions should not be criticized outside of their professional world. A close connection developed between the representations and actions of the various protagonists for reasons beyond these ethical differences. The blood contamination case was converted, through its construction in the news media, into a 'focal point,' the standard or reference mark to evaluate the existence of new risks associated with an original need of precaution and a new problem of responsibility. Judges made jurisprudence through the news media, and judicial and political actors adjusted their behaviours to the events described in the media. The news media thereby structured other practices (Ericson 1995; Manning and Hawkins 1990). Through the media, risks and necessary precautions were also constituted for a wider audience. With the disclosure of experts' activities made in the health scandals context, some aspects of the ordinary

128 Violaine Roussel functioning of the scientific, technical, and political worlds, usually unnoticed, were suddenly made public. The news media participated, together with the other social actors in the public sphere, in the transformation of abstract probabilities defined in precise conditions into individual risks, subjectively experienced. The discourses about risk and responsibility exposed the equivocality of science, as well as the fact that political representatives, who are supposed to be materially disinterested, have financial dealings with public and private laboratories. When media, judicial, or political actors publicly denounced contradictions in the management of a question - for example, in the blood contamination case, the conditions for contamination with the HIV virus were scientifically still uncertain, but political actors claimed that the situation was under their control - they contributed to its denaturalization and its formation as a problem of risk. As a consequence, behaviours or situations previously taken for granted became 'risky' activities. Through such publicity, 'realisation of the areas of ignorance which confront the experts themselves, as individual practitioners and in terms of overall fields of knowledge, may weaken or undermine that faith on the part of lay individuals' (Giddens 1990: 130). The expansion of discourses on risk, precaution, and responsibility originate in such a collective and public construction. Various social groups converge, for heterogeneous reasons, in the mobilization of these categories, often without any intention of participating in the selfreinforcing process. On the contrary, political actors would probably have chosen, if it had been possible, to avoid these designations. In shouldering the stigma of risks for which they have to take precautions, political actors ensure the locking in of the process, the stabilization of a common reading of the situation. Since political actors cannot, in the current balance of powers - in particular between justice and politics simply deny and reject the discourses about new risks, they develop strategies to escape from a difficult situation through the promotion of the precautionary principle. Agreeing tacitly that the problem is one of risk, political actors become, paradoxically, more vulnerable to denunciations while developing these activities and discourses around precaution. For political actors, the precautionary principle means practically that scientific uncertainties have henceforth to be translated into visible public health policies or other systematic measures to show that 'all the precautions have been taken to be close to zero risk' (an expression of the French Minister of Agriculture about the foot and mouth disease [Savey 1997; Kiss 1996]). Independent of the efficiency of action or the

New Moralities of Risk and Political Responsibility

129

control really exerted on a situation, such precautionary discourses certify publicly there is no prejudicial event for which the political actor should be held responsible. Labelling Processes in the Political Field EFFICIENT LABELS AND THE RULES OF THE POLITICAL GAME

At first sight, risk, precaution, or responsibility are only efficient labels used as discursive weapons in scandal mobilizations. These discourses proliferate particularly through political struggles, in which denouncing the irresponsibility of an adversary is usually an efficient move. The vagueness of the labels, their polysemy, the elasticity of their meaning, guarantee the efficiency of their uses: this loose meaning allows all the reversals in the discourses and the 'returns to sender' using the theme of irresponsibility. In the course of such discursive struggles, in which the definition of (il) legitimate behaviours and the meaning of situations are at stake, the actors involved often proclaim 'the end of polities' and 'the decomposition of political responsibility.' 'What is irritating finally, is that the Cour de Cassation [highest judicial court] pronounces judgments for lack of political decisions. I think that the legislator, the political class have to take their responsibilities. Either they decide to fight against corruption, against misuses of a company's assets, or they continue with half-measures, precedents that are in a sense for one case, and are different for another case' (Examining magistrate, investigating political scandals, 1997). Political actors themselves commonly emphasize their dependence upon expert authorities in decision-making processes in order to avoid proceedings in scandals. This strategy - whether used as a defence in court, or as a discursive strategy in political debates - was denounced by the public prosecutor in the Court of Justice of the Republic that had to judge three ministers in a blood contamination case in 1999: 'Who governs, asks the victim of social risks? A science that is or claims to be irresponsible? An administration assigned to implement texts without controlling them? Political consultants with an uncertain status or without any status? Politicians that did not know? Who still governs if the judge becomes the arbiter of these elusive responsibilities?' Of course, political actors use the label of (ir) responsibility in selfinterested ways. For example, in February 2001, the French Agency for Sanitary Security of Food (AFSSA) suggested a ban on selling some parts of sheep, to avoid the risk of diffusion of the variant Creutzfeldt-Jakob

130 Violaine Roussel disease, even if infection through ill sheep was not yet scientifically established. The French President, Jacques Chirac, immediately criticized such an unpopular measure (for the sheep farmers) and denounced the 'irresponsible behaviour' of the agency. In the same way, high-ranking magistrates and political actors often stigmatize, through the news media, the 'irresponsibility' of magistrates who participate in the prosecution of public actors. [In the Gigastore case, regarding illegal relationships between a private firm and local political representatives, and in which the president of a Regional Council has been in pre-trial custody] I think we really saw the arbitrariness at work, one can go so far as to behave arbitrarily, one can dispose, without any consideration, without responsibility, without being responsible of anything, of one's freedom, of one's honour, even of one's dignity, because we have seen a very honourable person dragged through the town, with handcuffs! So, it is true that justice sometimes goes too far. To irresponsibility. (High-ranking magistrate, 1996)

These contradictory uses of the notion might discourage analysis of responsibility. Responsibility might appear to be merely a strategic weapon, without stable content, an empty shell. In practice, discourses about risk and responsibility are numerous but composite, vague, historically changing. They reveal that an essence of responsibility cannot be found in the real practices of social actors regarding the categories. But concluding that (ir) responsibility simply does not exist would be abandoning explanation. Irresponsibility refers, in the examples quoted just above, to behaviours that appear to transgress some implicit rules of the game, rules intended to guarantee that things happen in an orderly manner in a given social sphere. A responsible magistrate is one who does not cause a scandal. He or she is expected not to subvert the usual functioning of the game, even if there is some discrepancy between its internal rules and legal rules, and even if those internal rules do not completely coincide with general values publicly professed. The experience of what happened in the initial scandals inclined most political actors to behave according to a new pragmatic rule of the political game: not being caught in the act of imprudence or negligence - which often means not being caught in the act of prioritizing their own economic or political interests - rather than anticipating public risks in uncertain situations (Bailey 1969).

New Moralities of Risk and Political Responsibility 131 New representations of what can be said and done in the political field, formatted in the language of precaution, became integrated not only into pragmatic rules but also into normative rules of the political game. Such rules echo certain aspects of the ethics of responsibility, which, combined with an ethics of conviction, defines the 'authentic political actor' (Weber 1959). This idea that a political actor has to take into account the consequences of his or her behaviour is central to the precaution rhetoric. Political actors must act with prudence, they have to take all possible precautions, especially in the field of public health and security. This normative rule refers to legitimate representations or values - historically and socially produced, and not derived from some 'nature' of responsibility - grounding the manifest political game. These norms require, for example, that political actors shall not prioritize the pursuit of economic benefit to the detriment of public health preoccupations. The normative rule of the game here stipulates the disinterestedness officially characteristic of the political game: the political actor serves public interests, particularly public security, disregarding his or her own self-interest. JUDICIAL RISK

By the end of the nineties, a number of political representatives sometimes prosecuted themselves - conceived and repeated a new statement: magistrates have become the producers of a new risk for political decision makers, and more broadly for the normal functioning of the democratic political system. Denouncing this 'judicial risk,' political actors tried to transfer the 'risk stigma' back to magistrates. 'The representative takes his responsibilities, so he takes risks. He now risks being given a roasting by the voters first, and then, going to prison!' (Political actor prosecuted in a financial scandal, President of a Regional Council, 1997). Judicial risk comes, in such discursive strategies, to counterbalance the risks for public health and security advanced in the prosecutions. According to this new risk, political actors can define themselves as victims of risks from which they must guard themselves. Concretely, political actors develop activities and devices to protect themselves from potential prosecutions seen as permanent 'swords of Damocles.' These strategies are first specific and local. Official and legal actions - for example, the development of local policies of risk prevention - are taken to show publicly that local representatives did their best to avoid risks and are totally blameless. These strategies are often coupled with unoffi-

132 Violaine Roussel cial ones, such as turning to a criminal judge for advice with respect to possible judicial troubles. The strategy of 'judicial risk' became part of national political mobilizations in 1999 and 2000. Left- and right-wing political actors agreed, despite their usual conflicts, on the necessity to legislate in order to force judges to abandon previous efforts at criminalization. Confronting the judicial practices in the health and security scandals, political actors promoted, with a wide consensus, an alternative way of reading the situations. They denounced a dangerous, unfair, and undemocratic judicial process. They urged changes in law to eliminate 'judicial risk.' The relative naturalization of this kind of legislative strategy derives from its simultaneous adoption in the political class as a whole. In March 1999, the big towns' mayors' association addressed a public letter to Prime Minister Lionel Jospin in this respect. In June, Minister of Justice Elisabeth Guigou appointed a government 'Study Group on the Criminal Responsibility of Public Decision-makers,' that was to make a convergent report in December 1999. In November, during the French Mayors, Association Congress, Lionel Jospin stated that 'an elaborate analysis about the notion of involuntary offence' was again necessary and suggested that criminal responsibility for involuntary offence should not constitute a serious offence. Against the possibility of responsibility without fault - which does exist in French law - political protagonists often underline the moral and logical need for the equation of the two, which results in a restriction in prosecutable behaviours. This kind of argument was also developed in the discourses of local representatives. 'The Mayors of big towns also protest against the institution of what could be called an "office responsibility" mayors are systematically charged with, even without any fault of the local representative, but only because he is the executive power of the local community' (Grandes villes hedbo, March 1999: 9). These discourses suggest that political actors would constantly be threatened with legal proceedings. 'Judicial risk' or legal insecurity would thus lead to renunciation of the political profession. What is interesting here is not so much the reality of this danger of political disaffection - 70 per cent of previous mayors ran again in the elections of March 2001 but the way political actors built and publicly asserted this argument at the same time that they carried a new bill designed, at least in part, to reduce the judges' power and the development of scandals about public health and security. Indeed, fifty-four local representatives in Parliament

New Moralities of Risk and Political Responsibility 133 were committed for trial, and fourteen convicted and sentenced, between April 1999 and March 2000. This legal insecurity is taking on such dimensions for local representatives that many mayors are discouraged and demobilized ... The whole society, and public and private decision-makers particularly, worry about the evolution of the criminal risk, that sometimes leads to prosecuting like criminals, in dreadful cases, highly covered by the media, people who have never committed any guilty action, and whose responsibility seems to be indirect and uncertain ... In particular, mayors feel as if they had been made responsible for everything; they are used as 'scapegoats,' before being delivered up, without consideration, to the news media 'pillory law': the risk of discouraging local representatives is very real and can finally lead to a deficit of candidates, a professionalization of the function, and a paralysis of local management. (March 2000, Rene Dosiere, Report on the bill about the new definition of unintentional offences, National Assembly (Parliament), Law Commission)

This discourse, defining a risk of political crisis, was coupled with the denunciation of wider risks for democracy inherent in the growth of judicial power and the judicialization of social relations. Behind the speeches about the political risk of inequitable conviction and political crisis, and behind the controversies, apparently on legal questions, about the fairness of assigning criminal responsibility to political actors, two conflicting visions of politics can be detected. On one hand, in mobilizing the precaution paradigm in their uses of old and new involuntary offence laws, magistrates, associations appealing to courts, and news media covering these events promoted and justified a reading of the political role in which a regulating state must act in a preventive way in order to avoid dangerous situations. A tacit obligation of vigilance was introduced. On the other hand, even if they promoted the idea of necessary precaution and developed preventive systems or measures (in order to show they were not suspects), political actors simultaneously furthered a role conception in which they will only be held responsible for what they know and can anticipate. According to this view, prosecution for unforeseeable risk would be unfair. In this respect, the then French Minister of Justice, Elisabeth Guigou, defended a bill explicitly proposed 'to fight against inequitable convictions.' 'Current principles of our criminal law regarding the involuntary offences sometimes lead

134 Violaine Roussel to shocking situations, particularly for public or private decision-makers. It is not fair ... that a mayor could be sentenced for homicide because a child had been electrocuted with a lamp post, installed 20 years before by his predecessors, even when he had never been alerted to the problem of this lamp post's maintenance' (Elisabeth Guigou, Minister of Justice, Parliament, June 2000). The bill on The Definition of Unintentional Offences,' adopted in July 2000, was conceived as a risk-reducing technology, in the sense that it reduced judicial risk for political actors. The creators of this text expected a decrease in proceedings and some discharges in current cases. With this codification political actors also tried to formalize a new representation of the political role. The political decision-maker should be legally responsible, not because of a preventive vigilance one could expect from him, but in relation to what he actually knew. 'The faulty and blameworthy characteristic of a behaviour is connected to the more or less foreseeable nature of its prejudicial consequences' (Elisabeth Guigou, Minister of Justice, Parliament, June 2000). In opposition to the current law, the new bill stipulates that a person can be convicted only if it has been established that: the person's behaviour constituted a serious offence, the person put others in a danger he or she could not ignore, and the danger in question was particularly serious. The new correspondence the law establishes between anticipation and responsibility affects, as expected, judicial decisions and compels judges to change their behaviours. However, the informal constraints produced by the discourses and practices of various actors -journalists, union and association representatives, lawyers, scientists - still weigh on the political game. What was especially at stake with the law of July 2000 is a definition of responsibility that limits the influence of magistrates, and of non-political actors more generally, on political activities. While restricting the field of criminal responsibility, political actors tried to confine inside the political sphere itself the mechanisms available to hold decision makers responsible. Political actors finally worked to restore their control, to guarantee the closure of the political system and to accentuate its autonomous functioning. The same result is expected from activities that do not imply legal codification, for example, the creation of risk surveillance networks, or the repeated appointment of parliamentary or expert commissions on involuntary offences, political morality, and public health. This result also derives from bills that do not directly confront the question but reform, for example, the judicial system, or create quasipolitical courts, such as the Cour de justice de la Republique.

New Moralities of Risk and Political Responsibility

135

Changing Action: The Effects of New Representations

The Constitution of Validation Networks The discourses of scandal respond to different validity registers depending on the institution - judicial, political, scientific, media - involved. Therefore actors think and act according to social markers that are not standardized or unified. The representations of precaution and risk in medical research refer, for example, to the unavoidable existence of uncontrolled factors, in the context of scientific controversies and uncertainties, whereas other visions of uncertain situations and their effects compel other actors to different behaviours. Nevertheless, a statement established in one specific sphere can be mobilized and transferred to another. It can be presented as taken-forgranted knowledge and become a pre-condition for interpretation and action. On one hand, such uses of local knowledge may form selfinterested strategies. When political actors mobilize, for example, in the Creutzfeldt-Jakob disease case, the fact that scientists cannot prove the possibility of human contamination through ill sheep, and assure us that 'we can eat sheep, there is no risk,' they convert a still-controversial scientific result into an efficient resource for political action. On the other hand, political beliefs in a scientific fact - for example, that BSE (Bovine Spongiform Encephalopathy) can contaminate humans - structure and constrain political action and modify the possible frame of action. Actors conceive a situation in a certain way because they attribute similar representations to experts, thereby entering into the circle of interdependent activities in defining situations. These political uses also consolidate scientific assessments, not only because they are stated publicly, but because they are brought to, and reaffirmed by, official authorities, such as epidemiological or sanitary agencies and judicial bodies. Through such appropriation mechanisms in several spheres, validation networks are constituted, crossing various interdependent institutions that participate in a controversy.1 The chain of embedded political, administrative, judicial, scientific, and media validations, and the importation of specific statements from one sphere into another, progressively lock in a configuration of action. The mechanisms at work in such situations exemplify path dependency. Path dependency has been used to explain changes in political regimes and especially transitions to democracy: it shows that the destiny of revolutions and political transformations depends on the embeddedness of small historical events specific to each phenomenon observed. This

136 Violaine Roussel system of interdependent activities delimits a particular 'social path' through which all the anticipations and actions are oriented, and following which an historical result is finally produced (Arthur 1989; Pierson 1997). In the present case, the connections between institutions built through interdependent systems also tend to unify the readings of situations. For example, in the mad cow case, all the participants finally share the idea that only some parts of beef are contaminating. In the course of scandals, political actors build cooperation and interaction arenas explicitly in order to produce common markers. Governmental agencies and expert committees are instituted to say what has to be thought and done with respect to blood products, which meats are risky, and so on. Official and public representations of risk are constructed in such arenas. The production of more homogeneous definitions of risk and responsibility, and the management of contradictory and complex views, are possible in particular because scientific conflicts are erased. The science appears unified. Scientists are elevated to the status of neutral experts, able to reconcile all the conflicting protagonists. In his work on medical expertise, Nicolas Dodier opposes two expert figures in industrial medicine (Dodier 1993): the 'lawyer-expert' who defends a public health cause, and the 'neutral expert' who incarnates the 'sole' and 'objective' scientific knowledge. In the cases studied here, some lawyer-experts may be identified, especially in the blood contamination case. Furthermore, some scientists denounced the political failings and spoke in the name of the immeasurable suffering of individuals contaminated with the HIV virus after a blood transfusion. More often, however, neutral experts appeared. They appealed to the progress of scientific mechanisms, and to the small statistical probability of becoming contaminated with HIV. The expert figure's speech derives strength from an apparent technical nature and neutrality that makes it less subject to debate than a political or moral choice. The neutral expert figure is widely instrumentalized in politics, and political actors often seek the authority of scientific experts to sanction their actions. At the same time, when they maintain they rely on scientific judgment, political actors contribute involuntarily to an increase in the real influence of the scientific sphere and knowledge on political activities. Various tools to manage the scandals were produced in quasi-political bodies such as health agencies and commissions, and resulted in concrete devices. The precautionary principle and the related measures and policies are examples of such tools. The notion of tracabilite (trackability)

New Moralities of Risk and Political Responsibility

137

- for example, the possibility of knowing the origin of animals and all the places they have been before their commercialization - fulfils the same functions. Its reactivation followed the precedents of contaminated blood and mad cow disease. Before these cases, trackability referred to the necessity of isolating a product in order to avoid contamination with parasites, and the associated method was quarantine (Torny 1998). Prior to the 1990s, these mechanisms were primarily taken into account as problems of quality management, especially in food and pharmaceutical industries. Trackability now invokes the new requirement of avoiding potential threats, and not only established dangers, and belongs to the modes of government of things and individuals analysed by Foucault (1975) with respect to disciplinary mechanisms. New Definitions of Possible and Legitimate Political Action The scandals are often described in the news media, as well as in the works of many French social scientists, as symptoms of new kinds of problems in democratic functioning. They are said to reveal a crisis of political legitimacy, sometimes considered as specific to French politics, sometimes seen as a larger problem (Sine 2000). The impossibility to talk about 'responsibility' in civil or administrative law explains the necessity to have recourse to criminal law. Criminal decision identifies definitely someone responsible who cannot escape. It is no longer the case elsewhere: under the Fifth Republic, political responsibility does not mean anything more, when nobody still resigns ... In the contaminated blood case, the judge finally exceeds pure legal reasoning, saying: 'I cannot behave differently because people will not accept it.' It is a political reasoning. But this excess comes only secondarily: when the ones who could have given a political judgement, according to their responsibility, have not done it. (Engel and Garapon 1997)

Two points arise. The first has to do with the relations established between the problematics of responsibility and the problematics of political legitimization. But reducing the debates on political responsibility to a symptom, the sign of a deeper and less visible legitimacy problem, hinders the explanation of responsibility discourses and means that they are no longer carefully analysed. The conflicting discourses on responsibility that took place during the scandals do not simply reflect a loss of legitimacy that occurred previously, through other practices, but consti-

138 Violaine Roussel tute in themselves (de)legitimization struggles. Legitimacy lies, in this perspective, in the tacit recognition that various social groups, in dominant positions, grant to political actors (Stinchcombe 1968). The crossdenunciations of political irresponsibility, and the political importation of this thematic in the way public problems are formatted, converge to substantiate the idea of a lack of political legitimacy, what is often described as a political crisis. The joint denunciations from journalists and judges produce efficient results, despite the fact that they respond to very different social determinations, and compel political actors to believe that they could not avoid a legitimacy problem. Political actors are drawn into justification discourses in order to demonstrate their political responsibility. Nevertheless, these conflicts are not only word-struggles. Political actors feel compelled to act in order to ward off, especially through the building of precaution devices, the danger of a loss of legitimacy. As a result of the mobilizations through which risk and responsibility discourses develop, the scope of possible behaviours for various actors involved fluctuates. For many political scientists, current political 'failings' inevitably result in the increasing intervention of magistrates, who finally perform a political function. This view is reflected in the concept of the 'judicialization' of political - or even social - systems, often conceived as a widespread process, and not specific to the French situation (e.g., Tate and Valinder 1995; Commaille, Dumoulin, and Robert 2000). Understanding the criminalization of some political behaviours as a judicialization of politics can lead to two opposite views. On one hand, magistrates are described as new social arbiters and guardians of public values, and their new democratic role is highly celebrated (Ludet 1995; Garapon 1996). On the other hand, the recent increase of judicial power is seen as a dangerous dysfunction threatening the democratic equilibrium based on political representation. In both interpretations, neither the perceptions of magistrates themselves nor their effects are taken seriously. In interviews, judges always deny scrupulously that they are fulfilling such a political role. In the context of political scandals, they feel particularly the necessity of demonstrating that they do not exceed their professional authority and act under the cover of this legitimacy. T think that the judge decides according to his file ... We do not have to substitute for the legislator, we have to know where our responsibilities are, we are not legislators, so, we have to apply a law, and if we do not agree, we leave the job, that is clear' (Examining magistrate, investigating political scandals, 1993).

New Moralities of Risk and Political Responsibility

139

If, in the specific circumstances of the scandals, this judge has to assert and repeat what is usually taken for granted in the definition of the judicial function, it is first because he often confronts, in this context, political discourses mentioning the spectre of the judicial will to govern. But it is also because the definition of magistrates' professional field of action, of what doing the judicial job means, is actually at stake, just as the scope of political action carries the marks of the scandals. Determining what a political actor is responsible for, or the responsibilities of a magistrate, means defining and challenging the limits of explicit, efficient, and legitimate action by these social groups. The capacity of one group to constrain another's social activities - the existence of a judicial right of inspection of some political behaviours, as well as the political control of judicial functioning - is at stake. Thus, for the magistrate quoted below, political responsibility, the importance of the political role, impels political actors to account for their activities and justify themselves. 'Democracy means that everyone has to be accountable to justice for the facts they are blamed for. But that means that all people have to be considered in the same way, and a politician does not have to be considered in a better or worse manner ... We have to consider these people with all due respect, but also the responsibilities vested in them that imply their accountability, especially to justice' (Examining magistrate, investigating political scandals, 1993). From this judicial point of view, politicians should be accountable to the judicial system and the criminal courts, and not only to voters (to whom political actors claim they owe their chief political responsibility). Criminal responsibility (accountability to the judges) is dismissed by many politicians in the name of political responsibility (accountability to the electorate). Far from being only legal and technical problems, the conflicts manifested in each scandal contribute to the reconfiguration of possible actions open to each actor involved in this relational system. Extending political accountability to new forms of criminal responsibility authorizes a new judicial control over political action. The transformation of the relationship between powers constitutes one of the major effects of the recent French scandals. Judges are now less likely to turn a blind eye to questionable political practices previously considered as normal. Moreover, political control over judicial activities becomes looser. The image of magistrates in public culture, especially of examining magistrates, has also improved considerably. These phenomena likewise entail a shift in the locus of responsibility of individuals for risk taking regarding some dangers defined as public

140 Violaine Roussel health questions or public security problems. Events are defined as new public risks and a new responsibility is attributed to political decision makers if they do not succeed in protecting citizens from these risks. Labelling as responsible political representatives and civil servants is here highly significant. It reflects a conception of the welfare state, currently in decline, which legitimates a broad definition of the field for political action, since health and security questions are state responsibilities. In opposition to the neo-liberal vision that makes individuals responsible for risk taking, fostering reliance on a private insurance system (Ericson and Haggerty 1997; Ericson, Barry, and Doyle 2000), in the welfare state model, political actors must plan the management of statistically foreseeable risks (Habermas 1998). Of course the neo-liberal paradigm, currently dominant in France, can be coherent with some mechanisms of criminalization through which the scope of the state's action is reduced to its repressive functions. But in the criminalization of political behaviours analysed in this chapter, something more interesting is at stake: the extension of the field of public intervention in dealing with risks. The definition of a new political responsibility makes sense in relation to the evolution of the moral bounds of public action, and legitimates a broader role for the state. Without any ideological motivation or goals on the part of the social actors involved, the social activities and denunciations at work in this context participate in the partial reactivation of welfare state mechanisms, over and against the global process of withdrawal of the welfare state. Nevertheless, this expansion of the field of public action under cover of precaution requirements does not correspond to the classical responses following from the welfare state model. The new forms it takes are illustrated by the creation of quasi-public agencies - such as the French Agency for Sanitary Security of Food - simultaneously defined as expert bodies relatively autonomous from the state and its political powers. Such agencies are distinct from organizations made up of state employees. They also have greater links to private and corporatist logic and interests. The creation of an agency of this kind belongs to the various technologies imagined by political actors to evaluate, control, and if possible exclude risks, and concurrently to avoid prosecutions. Through such tools - including local risk prevention plans, epidemiological surveillance networks, and committees on public health or security problems political decision makers contribute to the production of visibility mechanisms, which constitute the reality of risks for everyone, including themselves. The extreme multiplication of the politics of visible precaution reveals,

New Moralities of Risk and Political Responsibility 141 and makes concrete these effects of objectification of new risks. The diffusion of such representations of risks and the related changes in the perception of possible action make understandable some French precautionary measures, at first sight strange because disproportionate: for instance, the one cited above, which entailed destroying, at the beginning of the foot and mouth disease crisis, almost 50,000 animals that had been in contact with English sheep suspected of the disease, although no ill sheep had yet been identified in France. After several health scandals, precaution appeared as a new mode of government corresponding to the management of uncertain situations specific to the current times, a new model of action which challenges the classical theory of political decision (Gallon, Lascoumes, and Barthe 2001). Nevertheless, such a conception of the risk and precaution society attributes too much homogeneity to the social mechanisms merged into these designations. Borrowing the notion of precaution from the European recommendations regarding environmental pollution and sustainable development, the French users of the word embued it with new meanings and implications, unifying heterogeneous issues and activities through this designation. A fortiori, recent discourses referring more generally to a new risk society - with new risks generating original responsibilities - link together elements from several heterogeneous registers. The private and individual risk taking formed by the insurance industry; professional risks - not only compensated through private insurance, but also recognized publicly by collective agreements or by the state - that entitle professionals, such as teachers, to a higher salary for working in highrisk zones; the 'ontological' risks pointed out by Giddens (1990); the possibility of nuclear war, pollution, climatic changes, genetic engineering, and so on, are unified. These risks are moreover supposed to be connected to new problems of democratic functioning: in particular a crisis of political decision-making mechanisms, a decrease of public trust in the political system, and an uncontrolled rise in judicial power. These mechanisms are said to shape new situations and to correspond to a new social demand for responsibility. These various risks have in common only a family likeness, as Wittgenstein says about apparently univocal concepts. But, when public actors assimilate implicitly the different risks in their speeches and behaviours, and equate them through the unity of the word, they give form and reality to the risk society. Heterogeneous risks are amalgamated. Shared perceptions of the situation carry effects for social practices, particularly readable in the generalization of visible

142 Violaine Roussel political precaution, but also in the multiplication of research on risks publicly financed, in judicial interest in the repression of risk taking, and so on. In this sense at least we do live in a risk society. Nothing is more likely to guarantee the success and the consolidation of this representation than its simultaneous appropriation - even with partly different meanings - in various social spheres. Moreover, actors can refer, when they mobilize the notion, to the authority of science. In France, the appropriation of risk and responsibility labels as interpretative concepts in social science has reinforced their success (Garapon and Salas 1996; Beaud 1999). Some social scientists have left the scientific register to apply the notions in a political register and participate in political and economic processes (e.g., Ewald and Kessler 2000). What is more interesting, however, is the sudden increase of hybrid meanings among political actors, members of the economic sphere (especially from the private insurance industry), and scientists, under the auspices of public research institutions (such as the National Centre of Scientific Research, or CNRS).2 Even conceptualizations of risk that are more distant from lay uses of the word - in which risk is often associated with the analysis of growing anticipation mechanisms, and with the development of communication and surveillance technologies and social control - involuntarily foster political appropriations which will then respond to other meanings and interests. Notes 1 These networks are depicted by Pierre Lascoumes to show how important financial groups, such as Elf-Acquitaine, could have become involved in projects that should have been recognized from the beginning as vain and expensive (for example, in the 'sniffer planes' scandal) (Lascoumes 1997b). 2 International conference of the review: Politiques et management public, 'Les nouvelles exigences de la responsabilite publique,' March 2000: 9-10; Conference: 'Risques collectifs et situations de crise. Bilan et perspectives,' Programme 'Collective Risks and Crisis Situations.' CNRS, February 2001: 7-9, Paris.

References Arthur, W.B. 1989. 'Competing Technologies, Increasing Returns, and Lock-in by Historical Events.' EconomicJournal99: 116-31

New Moralities of Risk and Political Responsibility

143

Bailey, F.G. 1969. Stratagems and Spoils: A Social Anthropology of Politics. Oxford: Basil Blackwell Beaud, O. 1999. Le sang contamine: Essai critique sur la criminalisation de la responsabilite des gouvernants. Paris: PUF Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Becker, H. 1963. Outsiders. Glencoe: Free Press Bodiguel,J.L. 1991. Les magistrals, un corps sans dme? Paris: PUF Boigeol, A. 1989. 'La formation des magistrals, de 1'apprentissage sur le tas a 1'ecole professionnelle.' Actes de la recherche en sciences sociales, 76-7 Caille, F. 2000. 'L'action des magistrals dans la regulation des risques colleciifs: Texemple des sporls de monlagne.' Droit et societe44—5: 179-97 Gallon, M., P. Lascoumes, and Y. Barlhe. 2001. Agir dans un monde incertain: essai sur la democratie technique. Paris: Seuil Champagne, P., and D. Marchelli. 1994. 'L'information medicale sous conirainle. A propos du "scandale du sang contamine."' Actes de la recherche en sciences sociales 101-2: 40-62. Chateauraynaud, F., and D. Torny. 1999. Les sombres precurseurs: Une sociologie pragmatique de Valerie et du risque. Paris: Editions de 1'EHESS Commaille, J., L Dumoulin, and C. Robert, dir. 2000. La juridicisation du politique. Lecons scientifiques. Paris: LGDJ. Dobry, M. 1996. Sociologie des crises politiques. Paris: PFNSP Dodier, N. 1993. L'expertise medicale: Essai de sociologie sur I'exercice dujugement. Paris: Metailie. Engel, L., and A. Garapon. 1997. 'La fonction publique saisie par le droit.' Entretien, Esprit, oclobre: 108-9 Ericson, R.V. 1995. 'The News Media and Accouni Ability in Criminal Justice.' In P. Slenning, ed., Accountability for CriminalJustice. Toronto: University of Toronto Press Ericson, R.V., D. Barry, and A. Doyle. 2000. The Moral Hazards of Neo-Liberalism: Lessons from the Private Insurance Indusiry.' Economy and Society 29(4): 532-58 Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press Ewald, F., and D. Kessler. 2000. 'Les noces du risque el de la politique.' Le debat, juin 109: 55-72 Foucault, M. 1975. Surveiller et punir. Paris: Gallimard Garapon, A. 1996. Le gardien des promesses: Justice et democratie. Paris: Odile Jacob. Garapon, A., and D. Salas. 1996. La Republiquepenalisee. Paris: Hachelle Giddens, A. 1990. The Consequences of Modernity. Cambridge: Polity Press

144 Violaine Roussel Habermas, J. 1998. 'Les paradigmes du droit.' Mouvements 1: 93-101 Kiss, A. 1996. 'The Rights and Interests of Future Generations and the Precautionary Principle.' In D. Freestone and E. Hey, eds., The Precautionary Principle and International Law. The Hague: Kluwer Law International Lascoumes, P. 1996. 'La precaution comme anticipation des risques residuels et hybridation de la responsabilite.' I'Annee sociologique 46: 359-82 -" 1997a. 'La precaution, un nouveau standard de jugement.' Esprit novembre: 129-40 - 1997b. Elites irregulieres. Essai sur la delinquance d'affaires. Paris: Gallimard Ludet, D. 1995. 'Quelle responsabilite pour les magistrals?' Pouvoirs 74: 119-37 Manning, P.K., and K. Hawkins. 1990. 'Legal Decisions: A Frame Analytic Perspective.' In S. Riggins, Beyond Goffman: Studies on Communication, Institution, and Social Interaction, 203—33. Berlin: Mouton de Gruyter. Meny, Y 1996. 'Metamorphoses de la responsabilite.' Droit et cultures 31: 7-131 - 1997. 'La corruption : question morale ou probleme d'organisation de 1'Etat.' RevueFrancaise d'Administration Politique 84: 585-91 Pierson, P. 1997. 'Path Dependence, Increasing Returns, and the Study of Politics.' Unpublished manuscript, October Poncela, P., and P. Lascoumes. 1998. Reformer le code penal Paris: PUF Roussel, V. 1999. Les magistrats dans les scandales politiques en France - 1991-1997. These de doctoral, Universite Paris X - 2000. 'Scandales politiques et transformation des rapports entre magistrature et politique.' Droit et Societe 44-5: 13-39 - 2002. Affaires dejuges: Les magistrats dans les scandales politiques en France. Paris: La Decouverte Royer, J.P., R. Martinage, and P. Lecocq. 1982. Juges et notables au XIXe siecle. Paris: PUF Savey, M. 1997. 'Les lecons de la 'vache folle.' Esprit novembre: 101-17 Sine, A. 2000. 'Les nouvelles figures de la responsabilite politique.' Communication au colloque. Les nouvelles exigences de la responsabilite publique. 10e Colloque de la Revue. Politiques et Management Public, mars Stinchcombe, A.L. 1968. Constructing Social Theories. New York: Harcourt, Brace and World. Tate, C.N., and T. Valinder. 1995. The Global Expansion of Judicial Power. New York: NYU Press Thevenot, L. 1986. 'Les investissements de forme.' Conventions economiques. Cahiers du GEE, 29, Paris, CEE-PUF: 21-71 Torny, D. 1998. 'La tracabilite comme technique de gouvernement des hommes et des choses.' Politix44, 4e trim: 51-75 Weber, M. 1959. Le savant et le politique. Paris: Plon

6

Risk Management and the Responsible Organization MICHAEL POWER

Issues to do with 'corporate social and ethical responsibility' have a long history and can be traced back to concerns with the health and safety of workers and related legislation. In the United Kingdom these concerns took legal shape within various Factory Acts (Hutter 1997). However, while this inner 'regulatory space' of companies was originally (and remains) preoccupied with the prevention of accidents at work, the last three decades have witnessed a broadening of the corporate social responsibility (GSR) agenda to encompass a diverse array of external social and environmental factors. Indeed, it should be assumed that GSR is an evolving assembly of ideas and practices broadly united as a critique of existing forms of corporate activity, but by no means a coherent unity. This critical agenda which has had, and continues to have, many different kinds of enthusiasts and supporters, has left a trail of initiatives and often disappointed hopes. For example, it is over twenty years since the first corporate 'social audit' movement organized to provide criticisms of the consequences of business activity (See Gray, Bebbington, and Walters 1993: chap. 13), and over time there have been varying degrees of faith in the ability of different forms of accounting to change corporate behaviour and to promote CSR. Thus, value added statements (Burchell, Clubb, and Hopwood 1985), social indicators, eco-balance sheets and, most recently, triple bottom line and sustainability reporting (GRI 2000), have been at the heart of an evolving project to make the activities and impacts of companies somehow more transparent, auditable (Accountability 2002), and thereby amenable to change. These experiments in CSR reporting have been stimulated by numerous and often high-profile scandals and disasters. Exxon Valdez and Bhopal have become iconographic cases, creating pressures for reform

146 Michael Power in the way large corporations think about and process social and environmental responsibility issues. Such cases generate heterogeneous explanations and excuses (Jasanoff 1994), and provide increasingly wellorganized critics with, at least briefly, the ear of government and the regulatory policy process. Somewhere between corporate fears of loss in the face of legal liability and hope in the form of enhanced reputation, this critical discourse of GSR has been transformed (extreme critics would say co-opted and diluted) and has acquired greater currency within large organizations. The rise of 'business ethics' as an academic discipline over this same period is symptomatic of the ambivalence of a critical spirit which tries increasingly to work with the grain of a business logic that is part of the problem. The normalization of the CSR agenda is in part symbolized in the United Kingdom by the Labour government's creation of the post of 'Minister for Corporate Responsibility,' while on the international stage, avenues for developing CSR outside legislative frameworks have become increasingly institutionalized in the form of multiple networks and alliances that link large businesses, government, pressure groups, and non-governmental organizations (NGOs). The rise of these 'organizing organizations,' which seek to define and write standards for corporate conduct, has been conspicuous. At the heart of these many and varied CSR initiatives is the fundamental problem of internalization and this in turn raises the question as to what a company is, over and above the people who work in it, and where the boundaries of its logic of organization begin and end (Power 1994). From this point of view, CSR is about changing not merely corporate behaviour but the very concept of the corporate entity and, in particular, the concept of its role in society, or what is loosely called 'the licence to operate.' At its simplest, at least analytically, CSR is concerned with how management decision-making processes and practices in companies can come to accommodate and be sensitive to a disparate range of social and environmental values and concerns. Not least among these interests are those that organize themselves in the name of sustainability (e.g., Business in the Environment 2000; ACRE 2000a; ACBE 2000b). This problem of internalization in relation to CSR, of bringing 'society' into organizations, can be constructed idealistically as a project to alter the motives of those who work in organizations and to construct morally capable and blameable management. Many discussions and proposals in this vein culminate in calls to enhance organizational loyalty and ethics, and to create a new civic space for the corporation based on visions of corporate citizenship (e.g., Deetz 1992; Ackerman and

Risk Management and the Responsible Organization 147 Alstott 1999) and the rights of 'stakeholders,' as they are now known. In contrast, other constructions appeal to forms of enlightened corporate self-interest in which social, environmental, and business values might be aligned with business objectives as they currently exist. For example, economic tools, such as tax instruments, are indifferent to the moral qualities of corporate decision makers and are intended only to achieve desirable policy outcomes such as reductions in carbon monoxide emissions. In this way the internalization of external social and environmental goals is attempted via economic, rather than moral, incentives. This essay explores, via a discussion of risk management, a third angle on the problem of internalization, beyond the normative imperatives of the idealistic and the self-interested projects described above. If we assume that (risk) management systems and the routines they support are active ingredients in the construction of 'interested' behaviour in organizations, then the project of internalizing ideals of social and environmental responsibility is as much about embedding micro-systems and categories of information gathering as it is about appealing to individual morality or economic self-interest. From this point of view a kind of'enlightened self-interest' may become 'hard-wired' into organizational life in such a way that the distinction between self-interested and other-oriented behaviour is empirically unclear. And if what ultimately matters are ethically justifiable outcomes rather than 'good' people in organizations (here economic incentives may have the edge), then the routines and habits of organizational life by which the enactment of CSR is de facto possible, rather than the intentions of organizational participants, become the more fruitful empirical and analytical focus. This chapter is thereby concerned with a new phase in thinking about CSR which involves attention to the re-articulation of enlightened corporate self-interest around the ideas and practices of risk management. In short, how does risk management process morality? In the next section, it is suggested that the corporation and its responsibilities are co-defined and this idea is important for the subsequent overview of recent trends in risk management, the so-called new risk management. Specific attention is then paid to the rise of 'stakeholder thinking' and its significance in internalizing external interests and in challenging the boundaries of the corporation. The subsequent section deals with an important implication of stakeholder thinking for corporations- the need to manage reputation and the emergence of 'reputation assurance' as a discrete strand of corporate risk management advice. Finally, the chapter focuses more generally on the problem of corporate

148 Michael Power intangibles, such as brands and intellectual property, and recent pressures for greater management attention to them in the name of shareholder value. Overall, it is argued that these three closely related and convergent themes in risk management thinking - stakeholders, reputation, and intangibles - emerging in the mid-1990s, provide a jointly distinctive potential to reconstruct and redefine earlier ideals of 'enlightened corporate self-interest.' Indeed, 'risk' in the management setting reveals a certain 'organizing potential' for GSR at the same time as traditional distinctions between moral and instrumental thinking are blurred. Defining Responsibility, Defining Corporations The concept of corporate responsibility is a hybrid and encompasses both a morally approbative sense of 'acting responsibly' and a more forensic and legal sense of answerability and blamability. The fact that GSR can be defined as the 'adopting by organizations of responsibilities for actions which do not have purely financial implications but which are demanded of the organization under some kind of - explicit or implicit- contract' (Woodward, Edwards, and Birkin 1996) does not really enlighten this hybrid and leaves many difficult questions intact, not least the fact that corporations and their responsibilities cannot be defined independently of one another. Understood narrowly in a forensic sense, there is a need to answer the question: 'who should be responsible for corporate misdeeds?' According to Bovens (1998), large complex organizations pose this problem in a particularly acute form. While thinking of the corporation as actor or person is analytically convenient, there are difficulties in extending notions of 'motive' and 'intention' from the level of human agents to that of organizations. Consequently, Bovens argues for the organization as a 'semi-autonomous field,' an idea which squares broadly with 'structuration' and similar concepts in social theory. This direction leads in turn to a conception of senior management as responsible agents (Wells 1993). Indeed, senior management are increasingly held to create and support the 'climate' in which organizational deviance is, or is not, possible. However, the debate takes a further twist, since hierarchical conceptions of responsibility as answerability, which corporate regulators and other parties may prefer for clear blaming purposes, do not fit easily with prevailing ideas of agency or organizational design that value the au-

Risk Management and the Responsible Organization

149

tonomy of sub-units and which, for many regulatory projects, play an increasingly important constitutive role. Bovens (1998) argues that this difficulty can be partly resolved by 'collective responsibility' schemes, but the problem remains that, other than for exceptional organizational forms (e.g., small partnerships), making individuals responsible and blameable for events remote from their actions is unfair and creates free riders. And yet exempting them entirely and validating typical excuses, such as 'it would have happened anyway' or 'I knew nothing about it' is also unsatisfactory. This delicate and unresolved set of issues exists at the point where notions of responsibility imputation and blaming meet questions of organizational design and responsiveness relevant to the broader agenda of CSR. At this point it also becomes clear that notions of responsibility, social or otherwise, cannot be articulated independently of a debate about what the corporate organization is. The analytical concern of the arguments which follow is not with morality or even corporate 'responsibility' directly or substanlively, in the sense of coming to a determination about what is right and wrong in corporate life. As suggested above, corporate responsibility is at best a hybrid and contested concept, a mix of ideas based on moral notions of duty on the one hand and answerability or accountability on the other. Rather, this chapter turns the jurisprudential focus on meaning on its head and argues for the de facto and defining significance of the array of mechanisms, tools, belief systems, and organizational habits that constitute risk management as a technology and by which a set of preoccupations with CSR might be represented and internalized by corporations. In other words, regardless of desirability and other normative concerns, I assume that it is also useful to understand how notions of 'responsible' organization and CSR might come to be embedded in, and thereby defined by, the everyday routines of corporate life and, specifically, by risk management systems. The general argument is that because risk management creates and elaborates an inner 'regulatory space' (Hancher and Moran 1989; Power 2000a), it creates a potential 'window of responsiveness' through which ethical aspirations and operational pragmatism can become intertwined and corporate responsibility in a broader sense internally programmed. The extent to which ideas of managing risk lead in practice to an alignment between CSR and profit motives is an empirical question which this essay does not address systematically. The purpose in what follows is simply to provide an overview of those recent developments in the discourse of risk management which create the conditions (neces-

150 Michael Power sary but not sufficient) of possibility for such effects. Unlike earlier attempts to graft 'social responsibility' onto corporate agendas, the concept of corporate risk has a boundary shifting potential. As we shall see, risk and the organizational imperative to manage it appear, at least on the surface, to be able to internalize external interests and align them with corporate imperatives in a way that was previously impossible. Risk is the basis for corporations to process morality (see also Sison 2000). A New Risk Management

Something happened to risk management ideas in the 1990s to make the GSR agenda thinkable and operational in a new and distinctive way. This set of changes in thinking, evidenced by an explosion of conferences, new professional journals, and endless commentaries, is significant and can be suggested by the motif of the 'new risk management' (NRM) (Power 2000b). Although highly programmatic in form, representing an incomplete bricolage of ideas and aspirations, the NRM derives its momentum from a relentless emphasis on the systematization and integration of risk management practices. In addition, it is characterized by an ideal of consolidation in which risk and its management can be understood for the first time as a senior management task. Thus, in addition to becoming integrative and encompassing, risk management is also to shift its position in the corporate hierarchy. No longer simply the preserve of a variety of (humble) experts in different subfields with little in common (e.g., health and safety, treasury management, etc.), a comprehensive body of knowledge is under construction which inscribes this more systematic risk management agenda. Organizations have created, and are creating, the post of corporate risk officer or its functional equivalent (Butterworth 2000), a role that continues to evolve to overcome the 'risk archipelago' (Hood and Jones 1996). This is consistent with other projects of internalization based on new functional roles in the organization (e.g., Rehbinder 1991). Indeed, risk officers are likely to be reconstructions of existing internal agents, such as compliance officers (Weait 1993) and internal auditors (Power 2000a). Given this broader agenda risk management is becoming an explicitly managerial project for the first time, with many different subfields being re-articulated and co-related around the language of risk (e.g., Tschoegl 1999). That many practitioners assert that their work has always been 'about risk' does not alter the fact that making this newly visible and explicit is a change with substantive consequences. In particular, new

Risk Management and the Responsible Organization 151 organization roles in the name of risk have a forensic purpose in making sets of agents internally responsible and answerable for corporate actions. This means that the new risk management is also a project of heightened responsibilization within the organization (cf. Douglas 1992, 1999). Two themes or rhetorics were prominent in the U.K. professional conference circuit briefings of the late 1990s.1 First, fear is generated and attention sought by the claim that many risks - typically a range of messy issues assembled under the (largely residual) label of'operational risk' — are under-managed by virtue of not even reaching the risk management agenda. Second, it is also claimed that other potential hazards may be over-managed by unnecessary insurance and failure to recognize 'natural hedges' within organizations. These two claims constitute an extended sales pitch for a more systematic and generic approach to managing risks. The promotion of tools such as risk maps provides not only a 'technical' basis for improving organizational risk awareness, responsiveness, attention, and communication at senior levels, but is also a powerful symbol of rational integration itself through which insurers and auditors are reinventing their roles. What risk 'mapping' and 'landscaping' may lack in terms of a quantitative legitimacy of clearly calculated exposures is compensated for by the new legitimacy of their integrative and enterprise-based scope. Indeed, while such maps have been in use for several decades, they were not promoted as a template for management in general until the NRM wave of the 1990s. The NRM remains an incomplete assembly of ideas and practices which has been taking shape since the mid-nineties and acquiring legitimacy via the publication of generic risk management documents. In providing very general frameworks for risk management, these and other documents establish a new management agenda around risk, one in which awareness of responsibility as blameability and answerability is heightened while the evaluation of managerial competence is becoming closely linked to the ability to manage risk. For the purposes of this chapter, the NRM must be understood as a general project of internalization and responsibilization. At the very centre of this project is the idea and practice of an internal control system which analyses, evaluates, controls, and communicates across a wide spectrum of risks facing the organization. Such a system is no mere technical artifact with low organizational status but projects an ideal blueprint about how responsibility is rationally and legitimately distributed internally. Since the internal control and communication system has always been the great hurdle for projects

152 Michael Power of corporate social responsibility, it is directly relevant to the discussion that follows. I argue that a 'responsibility space' for organizations is being created at the intersection of the three risk management preoccupations identified above: stakeholders, reputation, and intangibles. The Rise of the Corporate Stakeholder

CSR is increasingly elaborated as responsibility to 'stakeholders.' This was not always the case. A U.K. discussion document published in 1975 recommending a broader concept of corporate reporting does not employ the word 'stakeholder' at all, but refers instead to the rights of 'user groups' (CCAB 1975). A narrow and strategic concept of stakeholder can be traced back to Freeman (1984). He included all parties who can influence, or are affected by, the advancement of a firm's objectives. However, this idea does not seem to have entered mainstream management discourse until the late 1990s. One possible reason for this, at least in the United Kingdom, is that the concept acquired currency and legitimacy in political policy circles (Hutton 1995; cf. 'stakeholder' pensions in the United Kingdom) and was then re-imported into discourses of CSR and subsumed within a broader 'third way' mood of partnership. The concept of 'stakeholder' is attractive precisely because of its morally ambivalent mixture of strategic and ethical meaning, and the pragmatic bridge it promises between managerial and CSR agendas. This bridging potential is particularly visible in risk management. During the second half of the nineties, initiatives to codify and develop generic risk management guidance put stakeholder analysis at their centre (e.g., CSA 1997; AS/NZS 1999; BSI 2000), thereby normalizing, at least at the level of policy discourse, the need for such analysis as a crucial part of risk identification. Echoing earlier strategic definitions, a stakeholder for this purpose is defined by BSI (2000) as any 'individual, group or organization having a vested interest or influence on the business or its projects.' Stakeholder analysis in this sense recognizes that 'whatever the nature of their interest, the existence of that interest means that they are a potential source of risk to the project and perhaps to the business' and can be 'difficult to understand, control or influence.' Without doubt there is an element of fear (see Hacking this volume) in these formulations; whereas 'users' or 'interest groups' occupy a space beyond the boundary of the organization, the concept of the 'stakeholder' threatens to blur the distinction between boundary and centre; corporations are being 'polluted' by society as never before.

Risk Management and the Responsible Organization 153 Accordingly, these risk management formulations can be regarded as a way in which the 'centre' tames the stakeholder concept. For example, it should be noted that this operational concept of the stakeholder is based neither on the rights of individuals and groups external to the company nor on the duties of the company to such groups. It becomes defined as part of the expanded risk management mandate described above, which prescribes that management adopt a broad view of threats to the business (or project). In short, the risk management stakeholder problem can be formulated as: 'who might blame and thereby damage the organization?' The normative operational task is to know ex ante who these stakeholders might be and this requires a broader view of the potential boundaries of a business or project. Indeed, the question of boundaries lies at the very heart of the issue; stakeholders of different kinds effectively alter the boundary between business and society. As Shell learned in disposing of Brent Spar, and as is now more generally evident from the Seattle and Genoa riots, risk management must be broader in scope than a 'cool' technological and objective assessment allows. According to Lofstedt and Renn (1997), Shell executives and the U.K. government mistakenly believed that the public, and pressure groups such as Greenpeace, would automatically understand the environmental superiority of water-based disposal. Their technocratic understanding of the problem overrode their sensitivity to the diversity of risk perceptions. While the academic literature on risk perception and the relation between risk and values is extensive, companies have only relatively recently come to recognize that discussions of risk cannot simply assume that the public is stupid or irrational, and that therefore their perceptions have no relevance for corporate policy processes. Rather, such publics and their organized representatives can have de facto and increasingly legitimate interests and claims. Moreover, they can adversely affect corporate reputation. Thus even false beliefs about environmental and social impacts of corporate activity must be taken seriously and managed in the stakeholder economy by 'upstream social analysis' (GroveWhite, Macnaughton, and Wynne 2001). In this way, the mundane routines of risk management may enfranchise groups external to organizations as management systems build in greater responsiveness to these diverse constituencies. Via the corporate recognition of stakeholders, a distinctive and explicit relation between risk and responsibility is under construction, although the extent to which the boundaries of the corporation are genuinely changing is unclear.

154 Michael Power In the United Kingdom, a review of company law recommended that legislation to enshrine the rights of other stakeholders in company law would be inappropriate. Instead, it was proposed that directors of companies should have a general duty to consider other stakeholders, and an evolving disclosure document, the Operating and Financial Review, is envisaged as a medium through which companies can report on these matters to such stakeholders. The relatively recent institutionalization of the 'stakeholder' as an element of risk management thinking therefore reflects a broader project to make large corporations more responsive to interests other than those of creditors and shareholders. Even if formal corporate legal systems, such as that found in the United Kingdom, tend to reject the institutionalization of stakeholder rights (ignoring for the moment the complication of workers' representation in continental European countries, and the fact that some stakeholders have rights in other parts of the legal system), companies themselves have begun to take very seriously these potential claims, filtered via the discourse of reputation rather than morality or regulation. According to Browne (2000a; 2000b), stakeholders are 'here to stay' and where regulatory systems and other experiments have failed in the past, companies will, for reasons of self-interest, enfranchise external groups with rights to information and transparency, and even rights of inquiry and questioning. However, despite extensive interest (Clarkson 1998), evidence to date also suggests that stakeholder engagement by companies in Europe is patchy (Hibbett and Kamp-Roeland 2001). Nevertheless, if stakeholders may not enter the corporation through the front door of legislative reform, they are entering via the back door of risk management and 'reputation assurance.' Reputation Assurance and Morality

The shock of various adverse events has led large multinational companies to seek to be less aloof. Rather than skilled defensive communications strategies housed in public relations departments, these companies are beginning to recognize the economic sense (prompted by regulation in some cases) of understanding and mapping 'drivers' of corporate reputation, and of developing various tools and practices (e.g., reputation scorecards). Information disclosure and accessibility; explicit codes of conduct linked to staff training; the measurement and disclosure of social and environmental impacts; and the development of proactive mechanisms for listening to stakeholders of varying kinds all add up to

Risk Management and the Responsible Organization

155

what has been called a new mood of 'expressionism' in organizations (Schultz, Hatch, and Mogens 2000). Fombrun and Rindova (2000) analyse the post-Brent Spar and Nigeria reputation recovery and management process at Shell, a process in which understanding stakeholder expectations in the environment and embodying this in a 'stakeholder management system' were central to a paradigm shift in reputation management and in the role of the Investor Relations Department at the company. The integration of stakeholder risk analysis and management into long-term plans, and an emphasis on dialogue and two-way information flows, is simultaneously the reinvention of corporate public relations. These emerging practices differ from older concepts of 'corporate citizenship' that prompted large organizations to establish charitable trusts and, although 'reputation' as an analytical concept within economics has a long history, it has only recently become a concept 'in use' and an object of managerial concern. As we shall see, many of the concerns identified above have been present and represented imperfectly within traditional accounting systems in residual categories, such as 'intangibles.' But corporate reputation per se, and its economic meaning as an 'asset,' is emerging as a distinctive object for management attention, and inevitably as a new object for advisory markets (see ICAEW 2000). According to Peters (2000), reputational capital is a 'new intangible' expressing the societal value that an organization puts back into the economy as well as a return to an old concept: the corporate 'licence to operate.' The public mandate of business has been a prominent theme since the mid-1990s and the flashpoint of various anti-global capital protests. Triple bottom line accounting is one tool for expressing and addressing this mandate on the understanding that 'what gets measured gets managed' (Wilson 2000). One can also begin to see the germs of change with the development of new vocabularies and the increasing internalization of concepts of 'sustainability' and responsibility reflected in high-level corporate support for experiments like the Global Reporting Initiative (GRI2000). This initiative recommends the disclosure of the basis by which stakeholders are defined and the involvement of such stakeholders in the development of performance indicators relevant to them. Large professional services firms have developed 'reputation assurance' functions and conferences on the subject are being held. For example, one such event was held in May 2001 at the Institute of Chartered Accountants in England and Wales entitled, 'Measuring and managing corporate reputation: making corporate social responsibility

156 Michael Power count and assessing its impact on the balance sheet.' Co-sponsors of this event were KPMG (a large professional services firm), AccountAbility (the institute of social and ethical accountability), Business in the Community, and Control Risks Group (a consulting firm). Sessions included such themes as 'winning with integrity'; 'defining and measuring reputational capital'; and 'reputational risk management- an investors' perspective.' Reputational 'assets' or 'brand equity' have become part of the reconstruction of marketing as a management practice in the service of shareholder value (Ambler, Barwise, and Higson 2001), but the rise of the stakeholder is broader than this. Reputation is not regarded solely as a function of, say, marketing and related expenditures, but also of perceptions of the corporation with a CSR content. Accordingly, 'reputation,' for all its operational ambiguity, is becoming a corporate window on society or, to choose a systems-theoretic metaphor, a basis by which society and business are 'structurally coupled.' Reputational assets understood explicitly as relational in form (companies-stakeholders) require active intervention and management, including an understanding of the risks to reputation and policies for managing these risks. Reputation assurance, as a subset of NRM, prescribes that companies must understand their constituencies, consider their significance and possible impact, and develop and implement a strategy for communicating with them. Stakeholder analysis thereby becomes an advisory service in the service of reputation. These nascent developments raise a number of questions and problems. For example, how do such constituencies of stakeholders exist in a sense independent of the corporate representative mechanisms for them? Reference was made above to Shell's stakeholder 'management system' and this raises the possibility that stakeholders are at least in part an internal product of managerial discourses, a new window for mediating complex environmental disturbances and shocks to large companies, created in the face of previous failures to understand the degree of organization of external critical groups. In addition, not all stakeholders recognized by such systems are equally important from a risk management point of view, in the sense of being equally capable of directly damaging reputation, even if they may have rights in some broader sense. Accordingly, companies face a complex balancing act and also run the risk of over-managing risks to reputation (Dukerich and Carter 2000). A second difficulty is that 'reputation assurance' as an independent practice is far from being well established and is characterized by a high degree of supply-side aspiration. Accordingly, the interest in corporate

Risk Management and the Responsible Organization 157 reputation may be more symptomatic of the general rise of the stakeholder and, in particular, of the institutionalization of pressure groups like Greenpeace, which have shifted from a purely antagonistic towards a more strategic relation with corporations (Friedman and Miles forthcoming). Reputation is in essence an outcome of myriad other corporate practices and managing an outcome independently of its multiple causes strongly suggests that the mutation of public relations into reputation assurance has a long way to go before it becomes a coherent and semilegitimate management practice. Nevertheless, directors of reputation assurance have been created, and consultancy advisory products reflect new and ingenious ways of catching the attention, and playing on the fears, of senior corporate executives. To summarize: organizations newly conscious of the imperative to 'express' their asset base and their social responsibilities are also fearful of the effects that certain stakeholders can have on their economic value. A linkage between risk and morality is emerging from the language and aspirations of risk management and not from some superimposed form of command and control regulation. While the liability and reputational risks of non-cornpliance with environmental and social regulation have always been important for large corporations, society is exerting its influence over the 'licence to operate' by other mechanisms. Indeed, the question which has historically preoccupied CSR in its various manifestations has been to do with what companies must be like to take their social responsibilities seriously. The new risk management turns this preoccupation on its head to pose a similar question: taking corporate objectives as given, what must society be like for CSR to become normalized? Framed in this way, risk management is not simply a window through which a set of external preoccupations might become internal, but a practice that negotiates and defines the boundaries of the corporation itself. Intangibles and the Porous Corporate Boundary The third strand of practices that support the internal responsibitization of large organizations relates to the idea of 'intangibles' and their increasing importance and co-production within the so-called new economy. Traditional 'old' economy accounting systems focused on the tangibility of assets. Tangibility is an important quality, enabling identification and verification, despite the fact that these features had and have little to do with any economic value that such assets may possess. Intangible assets

158 Michael Power emerged as a residual and conceptually problematic category of sources of value that could not be attributed to tangible assets. The label for these diverse sources of value was 'goodwill' and goodwill accounting has a long and controversial history. It was always understood that goodwill stood for such 'things' as a skilled work force, customer relations, and corporate reputation, but these were not matters that could be measured or managed with any degree of accounting skill. The value of such hazy assets only crystallized, and then only as a problematic aggregate, when a business was sold. While the value attributed to this 'goodwill' was small relative to corporate net worth, it posed a theoretical problem of understanding rather than a practical difficulty. However, in the last twenty years, it became increasingly apparent that, for many large companies, intangibles are by far the largest component of corporate value (as measured by market capitalization) and the tangible assets are in reality the residual. Despite abandoning tangibility as a relevant category for thinking about assets, the problem of intangibles haunts accounting systems and their claims to represent economic reality. Various efforts to deal with these problems, such as the independent valuation of brands, have been made within traditional accounting thinking. More radical calls have also been made for 'value reporting' and 'triple bottom line' accounting, which break with tradition and report on a variety of non-financial measures of corporate performance and on the management of a reputational asset called 'brand equity' (Ambler, Barwise, and Higson 2001). Whatever the realities underlying notions of the new knowledgebased economy, a serious challenge has been posed to traditional systems of entity-based reporting and control. Sources of value creation (such as reputation, discussed above) may lie outside the traditional boundary of the organization in a complex set of social and economic interdependencies from which corporations derive their mandate. So the accounting problem of intangibles reflects a more fundamental issue to do with sources of corporate value arising from relationships, relationships with the general public as much as with suppliers, investors, and customers. The driver for change in the corporate recognition, measurement, and management of intangibles is not necessarily ethical in intent or process, but it is potentially aligned with ideas of CSR in terms of overall direction. For example, the rise of the so-called individualized corporation reflects the increasing attention paid to the 'human' asset dimension of intangibles (Ghoshal and Bartlett 1997). And while there is a

Risk Management and the Responsible Organization 159 reasonably long tradition of thinking about human asset management and accounting (particularly in Scandinavia), such thinking has only recently received new momentum from the demise of the traditional accounting model of the corporation. Experiments, supported by the state, to measure and manage intellectual capital in organizations have a new urgency in practice despite a long history of concern (Mouritsen 2001). Accordingly, where companies are unable to process social concerns directly, the 'human capital' dimension of intangibles is set to become another category through which the boundary between society and organizations will be negotiated as a site for risk management. Companies are becoming conscious as never before of the risks of not managing human capital and are responding to the need to report creatively (internally and externally) on their efforts in this direction, supported by new schemes for disclosure and reporting, such as the U.K. Operating and Financial Review. Does the new visibility of intangibles and their potential role in redrawing the border between society and corporation have anything to do with morality as it is commonly understood? In terms of Habermas's famous, and probably overused, distinction between 'instrumental' and 'practical' reason, one could argue that these programmatic changes in managerial sensitivities and systems are simply a case of the former, just a more advanced and subtle mode of Taylorism. In contrast, the thrust of this chapter is to suggest that, at the level of practice, the management of risk in organizations provides an operational potential, or condition of possibility, to transcend crude oppositions between profit and morality. Micro-routines of risk management provide an as yet unrealized potential for new vocabularies and sensitivities to be internalized by organizational agents, and for the development of practices that negotiate and extend the boundary of the organization around the problem of intangibles. Conclusion At a time when it is being suggested that organizational loyalty is in decline and professional ethics a mere sham, it may seem implausible to make the case for a potential for corporate ethics in the name of risk management. After all, Enron claimed to have a fully integrated enterprise risk management system. However, I have argued here that the categories of 'stakeholder,' 'reputation,' and 'intangible asset' provide

160 Michael Power crucial and complementary avenues for a new de facto responsibilization of organizations, one which demands as a survival imperative greater sensitivity to the value systems of parties in the internal and external organizational environment (workers, interest groups). In short, the new risk management projects a rewriting of enlightened self-interest that blurs the strategic, technical, and normative dimensions of corporate decision making. As a project of internalization linking the management of intangible reputation to stakeholders, the NRM potentially re-inscribes society within corporate strategy in such a way that 'competent' management and 'socially responsible' management will no longer be the subjects of automatically incommensurable discourses. 'Responsible' and 'competent' organizations, and their senior management, are becoming defined by corporate governance frameworks in terms of managing, and not unknowingly taking, risks with their (stakeholders') assets. These 'responsibilized' organizations do not necessarily contain responsible or ethical individuals, however we define such terms, and the way they process risk will never be identical with directly processing socially desirable outcomes. However, risk management systems increasingly seek to build certain forms of responsible behaviour into the organization as habits or routines, supported by information and control systems. These systems, and the data collection which supports them, define the boundary of an organization which is increasingly porous to social values and interests. It has been suggested (e.g., Hood and Jones 1996: chap. 3) that a risk management system which embodies a blame-centred concept of responsibility will inhibit vital information flows within the organization, flows which characterize the good intelligence and responsiveness for preventative action. So in normative terms nothing less is at stake here empirically than the role of risk and its management in the construction of a new corporate responsibility that overcomes the negative side effects of blaming and succeeds in internalizing wider values and concerns into the decision-making routines of organizational life. From this point of view, the problem of making an organization responsible is not so different from making it 'strategic' or 'efficient.' The responsibilization of organizations requires the installation of a certain kind of self-observing, self-regulating capacity, a point which suggests important links between 'responsibility' and 'responsiveness.' The responsible organization no longer requires the necessary fiction of 'good' people within it, it needs rather systems that are responsive to ethical pressures in the organizational environment. The rise of ethics

Risk Management and the Responsible Organization 161 committees and ethics officers in some companies suggests that an important reconfiguration of management practice may be taking place. However, this essay is intended only to signal the potential created by the new risk management thinking. As noted above, earlier experiments with, for example, value-added reporting similarly offered a new vision of stakeholders and corporate responsibility, one which quickly vanished (Burchell, Clubb, and Hopwood 1985). And it must be remembered that the extent to which the new risk management does in fact bind social values into the heart of business logic has as much to do with developments in society as a whole, and its representative institutions, as with changing corporations. Understanding how corporations process and define their responsibilities to society through their risk management systems is part of a larger puzzle. Acknowledgment I am grateful to Bridget Hutter for comments on an earlier draft of this essay. Notes 1 I attended a number of these conferences between 1998 and 2000 and collected their promotional literature over this period. 2 For more on the rise of the stakeholder concept see Clarkson (1998); Ackerman and Alstott (1999). Although management articles on stakeholders have appeared earlier, indications of the late-nineties 'managerial arrival' of the 'stakeholder' are the special issues of Critical Perspectives on Accounting 9(2) 1998 and the Academy of Management Journal &(b) 1999.

References Accountability. 2002. AA1000Assurance Standard: Guiding Principles. London: Institute for Social and Ethical Accountability ACEB. 2000a. Internalizing Sustainable Development. London: Advisory Committee for Business and Environment - 2000b. 'Value, Growth, Success - How Sustainable is Your Business? Briefing Note for Directors. London: Advisory Committee for Business and Environment Ackerman, B., and A. Alstott. 1999. The Stakeholder Society. New Haven: Yale University Press

162

Michael Power

Ambler, T., P. Barwise, and C. Higson. 2001. Marketmetrics: What Should We Tell the Shareholders? London: London Business School/Institute of Chartered Accountants in England and Wales AS/NZS. 1999. Standard 4360: Risk Management. Sydney: Australian and New Zealand Standards Bovens, M. 1998. The Quest for Responsibility: Accountability and Citizenship in Complex Organizations. Cambridge: Cambridge University Press Browne, J.M. 2000a. 'Building an e-reputation.' European Business Forum 3: 13-15 - 2000b. 'Walking the Reputation Tightrope.' Accounting and Business 3 (1): 12-13 BSI. 2000. BS 6079-3: 2000 Project Management - Part 3: Guide to the Management of Business Related Project Risk. London: British Standards Institute Burchell, S., C. Clubb, and A. Hopwood. 1985. 'Accounting in Its Social Context: Towards a History of Value Added in the UK.' Accounting, Organizations and Society 10(4): 381-413 Business in the Environment. 2000. Internalising Sustainable Development. London: Business in the Environment Butterworth, M. 2000. The Emerging Role of the Risk Manager.' InJ. Pickford ed., Mastering Risk, 21-5. London: FT Books CCAB. 1975. The Corporate Report. London: Consultative Committee of Accountancy Bodies Clarkson, M. ed. 1998. The Corporation and Its Stakeholders. Toronto: University of Toronto Press CSA. 1997. CAN/CSA-Q850-97 Risk Management: Guidelines for Decision Makers. Canadian Standards Association Deetz, S. 1992. Democracy in an Age of Corporate Colonization. Albany: State University of New York Douglas, M. 1992. 'Risk and Blame.' In Mary Douglas, ed., Risk and Blame: Essays in Cultural Theory, 1-21. London: Routledge - 1999. 'The Risks of the Risk Officer: Diversity of Institutions and the Distribution of Risks.' Paper presented to the Conference on Risks, Strasburg Dukerich, J., and S. Carter. 2000. 'Distorted Images and Reputation Repair.' In Schultz, M., MJ. Hatch, and H.L. Mogens, eds., The Expressive Organization: Linking Identity, Reputation and the Corporate Brand, 97-112. Oxford: Oxford University Press Fombrun, C., and V. Rindova. 2000. 'The Road to Transparency: Reputation Management at Royal Dutch Shell 1977-96.' In Schultz, M., MJ. Hatch, and H.L. Mogens, eds., The Expressive Organization: Linking Identity, Reputation and the Corporate Brand, 77-96. Oxford: Oxford University Press Freeman, R.E. 1984. Strategic Management: A Stakeholder Approach. Boston: Pitman

Risk Management and the Responsible Organization

163

Friedman, A., and S. Miles, forthcoming. 'Developing Stakeholder Theory: A Social Realist Perspective. 'Journal of Management Studies Ghoshal, S., and C. Bartlett. 1997. The Individualized Corporation. London: Random House Gray, R.,J. Bebbington, and D. Walters. 1993. Accounting for the Environment. London: Paul Chapman Publishing Ltd. GRI. 2000. Sustainability Reporting Guidelines. Boston: Global Reporting Initiative Grove-White, R., P. Macnaughton, and B. Wynne. 2001. Wising Up: The Public and New Technologies. Lancaster: Centre for the Study of Environmental Change Rancher, L., and M. Moran, eds. 1989. Capitalism, Culture and Regulation. Oxford: Clarendon Press Hibbitt, C., and N. Kamp-Roeland. 2001. Prudently Protecting Profits. Amsterdam: NFVRA. Hood, C., and G.Jones, eds. 1996. Accident and Design: Contemporary Debates in Risk Management. London: UCL Press Hutter, B. 1997. Compliance: Regulation and Enforcement. Oxford: Clarendon Press Hutton, W. 1995. The State We're In. London: Jonathan Cape ICAEW. 2000. Human Capital and Corporate Reputation: Setting the Boardroom Agenda. London: Institute of Chartered Accountants in England and Wales. Jasanoff, S., ed. 1994. Learning from Disaster: Risk Management after Bhopal. University of Pennsylvania Press Lofstedt, R., and O. Renn. 1997. 'The Brent Spar Controversy: An Example of Risk Communication Gone Wrong.' Risk Analysis 17(2): 131-6 Mouritsen.J. 2001. 'Intellectual Capital and the 'Capable Firm': Narrating, Visualising and Numbering for Managing Knowledge.' Accounting, Organizations and Society 26: 735-62 Peters, G. 2000. 'Reputational Capital: Where the New Economy is Heading.' ICAEW. Human Capital and Corporate Reputation: Setting the Boardroom Agenda, 12. London: Institute of Chartered Accountants in England and Wales Power, M. 2000a. The Audit Implosion: Regulating Risk from the Inside. London: Institute of Chartered Accountants in England and Wales - 2000b. 'The New Risk Management.' European Business Forum 1(1): 26-7 - 1994. 'Constructing the Responsible Organization: Accounting and Environmental Representation.' In G. Teubner, L. Farmer, and D. Murphy, eds., Environmental Law and Ecological Responsibility, 369-92. London: John Wiley and Sons Ltd. Rehbinder, E. 1991. 'Reflexive Law and Practice: The Corporate Officer for Environmental Protection as an Example.' In A. Febbrajo and G. Teubner, eds. State, Law, Economy as Autopoetic Systems: Regulation and Autonomy in a New Perspective, 579-608. Milan: Guiffre Schultz, M., M.J. Hatch, and H.L. Mogens, eds., 2000. The Expressive Organiza-

164

Michael Power

tion: Linking Identity, Reputation and the Corporate Brand. Oxford: Oxford University Press Sison, A.J. 2000. 'Integrated Risk Management and Global Business Ethics.' Business Ethics: A European Review 9: 288-95. Tschoegl, A. 1999. 'The Key to Risk Management: Management.' Paper 99-42, Financial Institutions Center, Wharton School, University of Pennsylvania Weait, M. 1993. 'Icing on the Cake? The Contribution of the Compliance Function to Effective Financial Regulation. 'Journal of Asset Protection and Financial Crime 1(1): 83-90 Wells, C. 1993. Corporations and Criminal Responsibility. Oxford: Oxford University Press Wilson, M. 2000. 'Corporate Reputation and the Triple Bottom Line: The Social Responsibilities of Business.' Risky Business, PricewaterhouseCoopers 2a: 6-10 Woodward, G., P. Edwards, and F. Birkin. 1996. 'Organizational Legitimacy and Stakeholder Information Provision.' British Journal of Management1!: 329-47

7

Risk and Moralization in Everyday Life ALAN HUNT

This chapter explores the role played by risk discourses in the moralization of everyday life. I will argue that the increasing prominence of risk analysis has generated an expansion and intensification of such moralization. While the recent escalation of risk discourses has produced an expansion of calculative approaches to uncertainty and stimulated insurance and actuarial rationalities, there has been a tendency for studies of risk to overlook the moralization that accompanies the way in which the invocation of risk intervenes in social life. This moralization in turn leads to a proliferation of both bureaucratic regulation in the everyday world and an expansion of the responsibilities that burden citizens in a way that reinforces and even multiples the regulatory impact. It is at this point that we rediscover a link between the everyday risks and the macrorisks of the present. The inability of existing forms of economic and political organization to address effectively the 'big risks' is, if not masked, then deflected by the overwhelming regulatory urges that characterize modern governmentality in both the public and private spheres.1 This tendency yields further evidence of the gap between the dominant political discourses of neo-liberalism, with their mantra of deregulation, and the actual governmental practices pursued. The topic of risk has become a prominent focus of contemporary interest. I begin by distinguishing my position from the two major positions within risk theory. First, I depart from Ulrich Beck's influential account of 'risk society' (Beck 1992). While he avoids the sterile debates of the 'post-isms' that predominated in the 1980s, his claim about the novelty of the risks of late modernity is unconvincing. That hazards are socially created is not new; the death toll from the spread of syphilis through trade and colonialism was a pandemic in much the same way

166 Alan Hunt that AIDS is today. There is evidence that original human settlement and movement had massive impacts on ecology and caused mass extinctions of both humans and other species. Undoubtedly the potential impact of today's risks are greater, but there seems to be no reason to view this as a qualitative distinction that justifies the identification of an epochal change to a 'risk society' that Beck proposes. The second approach I distinguish myself from, although it will become clear that I am much closer to it than to Beck, is that associated with Foucauldian governmentality theory, perhaps most explicitly articulated by Mitchell Dean (1999). Dean and others focus on the part played by responses to risk and the role of risk discourse in governmental technologies. This line of inquiry yields important results when applied to detailed studies of various technical forms of institutional government, such as the relationship between risk and insurance (Castel 1991; Ewald 1991) and actuarial practices (Power 1997; Simon 1988). My reservation is that this focus has tended to promote a view of government as a technical activity pursued by formal or expert institutions and thus to narrow attention to official documents and discourses. Such an orientation runs the risk of obscuring the part played by risk discourses in everyday life. The more positive aspect of the governmentality approach is the attention that it pays to the interaction between the practices of governing others and governing the self. This dimension provides the central focus of my paper: the role of risk discourse in everyday life. Risk analysis is deployed to moralize a wide range of social behaviour and is thus a potent source of moral regulation. I have previously argued that over the last century there has been a significant shift in the way in which moral regulation functions (Hunt 1999a). In the nineteenth century 'the moral' was a distinctive genre; things were condemned as 'wrong' or 'immoral.'2 Increasingly, morality has come to function through proxies, not in its own voice, but in and through other discursive forms, the two most important and closely related being the discourses of 'harm' and 'risk.' I suggest that there is an historical transition. 'Harm' emerged during nineteenth century, and is exemplified in J.S. Mill's classic formulation, 'the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others' (Mill 1910: 72). The concept of harm occupied centre stage for much of the twentieth century, but by the end of the century the discourses of risk had come to predominate. The point that needs emphasis in explicating the thesis that moral

Risk and Moralization in Everyday Life

167

discourses function through proxies is that the moral dimension is not excluded, rather it becomes subsumed within discourses whose characteristics have a utilitarian guise. The outcome of this interconnection between moral discourse and risk discourse constitutes an instance of hybridity, the combination of two types of discourse in such a way as to merge their characteristics into a distinctively new form. The most striking feature of the hybridization of morals and risks is the creation of an apparently benign form of moralization in which the boundary between objective hazards and normative judgments becomes blurred. Projects articulated and justified in terms of 'risk' that invoke 'harm' are frequently also projects of moral regulation. For example, the consumption of alcohol has long been a target of regulatory intervention. This intervention used to be because drinking was deemed 'wrong' for religious or some other moral reason, but during the twentieth century the discourses underwent a fundamental transformation. There has been an increasingly utilitarian focus on 'harm to others,' as in the material produced by Mothers Against Drunk Driving (MADD). Alternatively, harm is medicalized as a risk factor in discourses invoking the care of the self. Thus the standard U.S. Surgeon General's warning on wine bottles reads, 'According to the Surgeon General women should not drink alcoholic beverages during pregnancy because of the risks of birth defects.' Individuals come to be assessed, or are invited to assess themselves, as being 'at risk.' A new expansionary cycle of responsibilization sets in.3 No longer is it sufficient to take responsibility for one's own drinking; people are now presumed responsible for the guests they entertain. Responsibilization is further compounded by what may be described as waves of expanding legal liability accompanied by prudential insurance and further litigation liability that seeks to protect against the consequences of the risks that arise from the consumption of alcohol. In sum, we witness an expanded problematization of alcohol within which a remoralization occurs: the responsible employer should no longer serve alcohol at office parties. Alcohol has increasingly become a moral issue and a 'new temperance' movement has emerged (Wagner 1997). Risk and Everyday Life

The 'big risks' which galvanize Beck's concerns, from Chernobyl to global warming, coexist with a host of 'everyday risks': whether to drink bottled water, to allow a teenage offspring to go downtown alone, to

168 Alan Hunt tackle tendinitis with yoga or cortisone, and dozens more, all of which share a common discursive formation as risks. This expansion of invocations of risk has occurred not only because there are more risks, but also because a diverse range of problems have come to be conceptualized in terms of risk. Beck's treatment starts out with 'big risks' and takes these as the model for all risks. Is it necessary to consider the relationship between 'big risks' and 'everyday risks'? Giddens suggests that social awareness of the big risks produces a state of ontological insecurity and it seems to follow that this insecurity breeds a generalized state of insecurity and anxiety in the everyday realm (Giddens 1991: 37). Lest we too summarily conclude that contemporary life is suffused with insecurity, it should be noted that many features of our lives have been rendered more secure and predictable. We should recognize the existence of a complex dialectic of insecurity and security. The globalization of food franchises, McDonaldization, ensures that I can get a standardized and thus predictable food or drink item in almost any city, and I no longer have to worry whether my car will start on a cold morning (Ritzer 1996; Barber 1995). My computer reliably puts me in easy contact with people and organizations around the world; that contact may or may not be imperilled by fear of millennial computer crashes or global viruses. Our world is not necessarily safer or more secure, it is transected by dimensions of security and danger. Safety and danger stand in some non-linear relation to security, anxiety, and fear. Discourses of risk play a significant part in how risks are experienced and lived.4 It is important to bear in mind that risk involves a continuum that encompasses not only risks outside the possibility of individual control, but also those where responsibility is individualized, and not only real and pressing dangers to people, individually and collectively, but a mass of entirely trivial risks as well. If we inhabit a social world characterized by this dialectic of security and insecurity it is important to consider how the boundaries between these dimensions become established. A key feature of both is that we are confronted by choices. The choice as to which soft drink to consume has little intrinsic significance: we choose between Pepsi and Coke as lifestyle consumers. But we may also choose between carbonated drinks and water as lifestyle consumers, and here issues of risk may be implicated in such a way that they appear on both sides. Water may not carry the risks associated with caffeine and other additives, but tap water may involve risks associated with impurities or fluoridization. We then consider a move to bottled water, only to learn that its purity is far from

Risk and Moralization in Everyday Life

169

certain. Everyday risks present us with the necessity of making a seemingly never-ending set of choices. The significance of these choices is compounded by the disparate pressures of the mechanisms of responsibilization that demand that we make them in a context that requires us to treat our lives as a project over which we should exercise a deliberate and long-term calculative effort. Parents used to send their children to the local school, and to take them to the local doctor if they were ill. That was the most that was expected of responsible parents. Today they are required to exercise choice about the appropriate type of school and style of medicine. The stakes involved in being responsible have been rising steadily. This situation is well captured by those who have noted the emergence of a 'new prudentialism' that requires us to fashion a prudential relationship of the self towards the self (O'Malley 1992; Rose 1999). The dialectic of security and risk manifests itself in an unstable interaction of familiarity and alienation, trust and indifference, privatism and participation. The exercise of choice in the face of risk has been linked in an important debate about whether there has been a decline in trust in contemporary society (Giddens 1991; Putnam 2000). Trust makes it easier to live with risk; in developmental psychology the root of trust occurs in the process whereby the infant learns that the caregiver who departs will return. For present purposes the more significant issue is the distinction between interpersonal trust through reliance on intimates and impersonal trust in expert systems (Luhmann 1988). Trust, it is argued, is 'like a lubricant that makes the running of any group or organization more efficient' (Fukuyama 1999: 16). We 'trust' the electronic computation on our shopping such that few of us bother to check our receipts. We trust the knowledge of contractors that the electrical systems they install in our homes not only comply with regulations but, more generally, are safe. The question of whether there has been a general decline in trust in both interpersonal and expert systems has excited much controversy. Significant for present purposes is the consequence of the pluralization of expertise. Experts are no longer satisfactorily legitimated by their location in state institutions or traditional professional organizations. They are increasingly dispersed across rival institutional sites and disseminate a plethora of incompatible knowledge and advice. The medical field is riven by competition between opposing bodies of knowledge. Medical controversy is not new, but its disagreements are increasingly fought out in public. Experts disagree about what the risks are, how risky

170 Alan Hunt they are, how these risks can or should be quantified, and how they are to be minimized. The fracturing of expert knowledge has coincided with a massive growth in non-expert expertise. A proliferation of single-issue movements have developed their own expertise through which they dispute traditional forms of state and professional expertise and act as what may be usefully referred to as 'risk-promotion' movements. Such movements not infrequently succeed in effecting the 'regulatory capture' of expert systems, in that it is often easier for them to 'go along with' the risk entrepreneurs of such risk-promotion movements than to fight them.5 The consumer of expert knowledge is confronted by the problem of how to weigh and evaluate conflicting knowledge claims. At this stage the distinction between trust in experts and interpersonal trust becomes blurred. Only a decade or two ago doctors used to conduct their diagnosis, write out a prescription, and wave the patient away with a cheery, 'Just take the pills,' probably without even naming the medication in question. Today doctors tend to engage patients in their own diagnosis and treatment by itemizing a range of possible treatments and asking them to choose. This can be a disconcerting experience for a patient in the absence of relevant knowledge. 'Trust' in doctors is unlikely to be based on attempts to evaluate their credentials or the medical science on which they found their advice, but is much more likely to be formed on the basis of their style and other interpersonal characteristics. How can I decide whether to obtain a second opinion or which course of treatment to opt for? It is necessary to exercise caution before concluding that trust serves as an antidote to risk. The problem is that trust is inherently experiential and provides limited means of validation of the reliance that results from that trust. The growing distrust in expert knowledge leads in two possible directions. On the one hand, there is every reason to be concerned about increased dependence on the quasi-expertise generated by singleissue movements, such as those that tell us that all our ills are allergic or are caused by chemical food additives. Such movements tend to rise and fall rapidly, leaving many in an increased state of anxiety. On the other hand, distrust of expertise can result in the closing of the radius of trust. As Giddens argues, the investment of greater attachment to intimate relations, founded in openness and intimacy (1990: 121), involves increased 'risks' from the strains that such relations produce. This dimension is well illustrated by Carla Willig (1997), whose interviews with people in long-term sexual relations revealed that the majority were

Risk and Moralization in Everyday Life

171

reluctant to ask their partners to use condoms because to do so would violate trust. She concludes that trust itself can be 'risky.' The responsible and prudential choices that individuals are called upon to make are rendered difficult and problematic by the pluralization of expertise. In the context of the weakening of the authority of expertise, the contemporary responsibilized individual must choose between competing versions. Decisions about which brand of expert knowledge to follow have to be made, like so many other lifestyle choices, on the basis of subjective preferences whose roots and motivations we are rarely able to perceive. Individuals are thus expected to engage in the risk management of everyday life in the context of escalating competition between experts and quasi-experts. Into this destabilization of trust in expert knowledge individuals, to some extent, but only in a limited number of fields, seek to acquire a degree of expertise for themselves. While I may acquire some knowledge of food labelling, I will be unlikely to learn much about the safety of automobiles, or the risks associated with use of garden pesticides. The gap between the knowledge we need and the knowledge we can acquire widens probably at much the same rate as the knowledge sources available to us increase. The most immediate outcome is that for many the spiral of the need to know and the limits of knowledge capacity can only lead to an amplification of the experience of risk and the insecurity of choice. Individual risk management is never simply a matter of utilitarian risk assessment and risk avoidance. Risk has been projected by 'risk theory' as a set of technical issues, of 'risk analysis,' 'risk assessment,' and 'risk management,' but the technical dimensions of risk do not eliminate moral and normative judgments. It is a key component of my argument that risk practices are deeply moralized. As Mary Douglas argues, risk, danger, and sin have been reconnected in contemporary discourses (Douglas 1992: 26). By moralization I mean that social practices are subjected to scrutiny in moral terms requiring judgments about whether practices are 'right' or 'wrong.' This moralization involves appeal to some set of criteria that work through a continuum of evaluation that moves from conduct that is morally neutral to varying degrees of wrongness or immorality of the practice. This moralization process is a social process in that it is always located within social contexts and relations. It is significant that despite the scientization of risk assessment, explanations of risks are still frequently grounded in moral discourses that call into play issues of ethnicity, sexuality, and other social stereotypes. The classic exemplar in the case of HIV/AIDS has been the remoralization of homosexuality.

172 Alan Hunt The processes that make us responsible for an expanding set of choices within the matrix of risk require individuals to engage in a moral enterprise of the self. It may be more productive to substitute 'ethical' for 'moral' to avoid any implications that this process involves considering the moral judgments of others. An ethical self-governance refers to the significant consequence of responsibilization, namely, the requirement that individuals seek out and fashion an ethical life for themselves. This requirement does not impose prescriptions about the correct conduct of that life. It only demands that life be conducted reflexively, or that the individual be capable of advancing some justification of the choices made. 'What to do? How to act? Who to be? These are focal questions for everyone living in circumstances of late modernity' (Giddens 1991: 70). An ethical enterprise of the self requires the employment of knowledge and self-knowledge to advance projects the crucial element of which is to display a capacity for self-control and self-improvement. It immediately becomes apparent that the discursive formation of ethical self-governance bears remarkable similarities to an earlier paradigm of self-governance that reached its most developed form in the late nineteenth century. This paradigm was embedded in the discursive formation of 'character,' whose main attribute was the securing of mastery over the self through self-control and self-improvement. The difference between the old and newer form of self-control is captured in an historical transition from 'character' to 'personality.' 'Character,' which required conformity to externally defined and legible attributes (proper bodily deportment, a firm handshake, etc.), was displaced by the twentieth-century discourse of 'personality,' which required individuals to fashion a distinctive and attractive identity (Susman 1984; White and Hunt 2000). The quest for personality required an engagement in a moral enterprise of the self. The related dimensions of moral and ethical self-governance exist in tension. While the self-fashioning of an ethical identity has become more important, the external pressure of moral regulation persistently asserts itself. Consider the current situation surrounding tobacco use. While its problematization exhibits a characteristically modern form in being heavily medicalized, it also reveals other dimensions, the most important of which is the need for individuals to exercise self-control in the name of personal health. But the demonstration of self-control is at the same time a display of respectability. To be a respectable citizen requires the same kind of self-control that Elias (1978) noted for an earlier period, when it became necessary to abstain from eating with the

Risk and Moralization in Everyday Life

173

fingers or belching in company. Respectability requires compliance with external standards, proving that one can 'win the battle' against an everyday addiction. Problematizations and Anxieties

Foucault argues that the central question of the history of thought is 'to define the conditions under which human beings "problematize" what they are, what they do, and the world in which they live' (1985: 10). Problematization thus concerns the ways in which issues become the focus of concern, or in an earlier sociological tradition, how and in what way 'social problems' are generated. Risk provides a major form in which many aspects of life are problematized, which Foucault refers to as attending to 'the history of the way in which things become a problem' (Foucault 1999: 414). It is in this sense that we may speak of the social construction of risk in a way that avoids the unhelpful and misleading issue of whether or not risks are 'real.' The identification of risks provides one form in which problems are experienced, grasped, and articulated. This construction also avoids the necessity of debating potential distinctions between 'risks' and 'dangers.' Risk problematization brings into play a discursive formation that provides a way through which the future is framed in the present and which connects some anxiety-inducing feature of the present to the fear of future harm. It is this forward projection of potential harm that gives risk discourse the power not just to induce anxiety, but to stimulate personal or social action. That is, such forward projection gives the discourses of risk a different temporal extension than that generated by the immediacy of danger. While the discursive forms of risk discourses have no fixed or even stable content, once in existence they are readily applied to newly 'discovered' risks. The adoption of problematization as a starting point generates two main lines of inquiry. The first concerns the subjective experience of problematization: how do problems come to be experienced and articulated? One of the most significant forms in which risks are experienced is as anxieties. The second avenue of inquiry concerns the forms of action or inaction that anxiety gives rise to. Anxiety is a psychic condition of heightened sensitivity to some perceived threat, risk, peril, or danger. A distinction between anxiety and fear seems both possible and attractive, but it is not ultimately sustainable. One possibility is to define fear as an immediate response to

174 Alan Hunt danger, while anxiety is a generalized non-immediate apprehension. This line of inquiry is unhelpful because it requires a normative judgment that rests upon a presumed natural distinction between fear as reasonable and anxiety as neurotic. If people are concerned about the dangers of nuclear war or alien capture we merely reveal our own normative position if we label such responses as 'realistic fears' or 'hysterical anxieties.' Anxiety has the effect of personalizing risk. Anxiety also concretizes risk: it is no longer an abstract threat, but is converted into an apprehension confronting individuals. It is next helpful to distinguish between individual anxiety and social anxiety. Individual anxiety has no social importance unless it is a shared anxiety (social anxiety), one that becomes additionally significant if it results in some discernible action by significant numbers of people. This also allows us to put aside any debate about whether the risk anxiety in question is 'real,' since it has already become so if it gives rise to observable social action.6 This distinction is, however, becoming increasingly unstable. Individuals with unusual anxieties can now find others that share them as a result of the extended communicative capacity of electronic media. It is characteristic of the present period that the term 'risk' is used in a way that tends to conflate it with 'anxiety' and to define the goal as 'safety' or 'security.' Parents 'anxious' about the safety of their children tend to conceptualize their anxiety in terms of the 'risks' their children are exposed to, such as bullying in school, sexual predators, and 'drug pushers.' That these discursive constructions have significant practical consequences may be observed by the paradoxical outcome that keeping children 'safe' often means keeping them dependent. Anxiety about children is amplified the more that children are treated as a special category of 'innocents' and, significantly, results in a dual process in which the infantilization of children exists alongside the promotion of a precocious adulthood. However, it is important to notice that this involves more than a mere conflation of 'risk' and 'anxiety.' It gives rise not only to specific conduct (driving children to all their activities), but also leads to the launching of regulatory projects (imposing post-detention restrictions on sex offenders). Risk discourse transposes anxieties into an objectivist problematic. Is X a risk? How much of a risk is X? Risk discourses thus imply that the risks they articulate are real, potentially measurable, and calculable. The subjectivism inherent in discourses of anxiety is displaced; each anxiety becomes conceived as a 'risk' that is 'real' and is thereby provided with a

Risk and Moralization in Everyday Life

175

rational form of legitimation that opens up the possibility of adopting practices of risk avoidance and risk management. In contrast, anxieties alone tend to give voice to the concern that 'something ought to be done,' without necessarily specifying any concrete course of action. Yet there is no guarantee that risk analysis yields reliable quantification. While holding out just such a promise, it all too often functions as a form of normative judgment. Most significantly, risk serves to objectify anxiety. The Supreme Court of Canada in R. u Butler ([1992] 1 S.C.R. 452) has concluded that pornography constitutes a 'significant risk of harm' to women. This formulation provides a more pragmatic justification for censorship than the traditional assertion that pornography is obscene. Yet the court's formulation ensures that it is not necessary to be able to demonstrate that any identifiable harm is caused; all that is required is a risk of such harm. Thus 'risk' serves merely to widen the catchment of censorship and to weaken the prosecution's burden of proof. This formulation exemplifies the conflation of 'risk' and 'harm,' which is a widespread feature of contemporary risk discourses. That anxieties are today commonly articulated as risks does not provide any guarantee of rational grounding or objectivity. Thus there is, ultimately, no escape from the question about the reality of any specific risks. But as Ian Hacking has demonstrated, there is equally little mileage in counterposing claims of reality and social construction (Hacking 1999). My strategy is to refuse such a debate by avoiding arguments that rely on constructionist claims. This can be achieved by drawing attention to the ever-present link between risk discourses and epidemiological techniques. Epidemiology is a technique of calculation that avoids simple questions of causation. It establishes instead what may be termed 'thin' connections that seek to unravel the links between multiple risk factors. Epidemiological practices rely on some version of the metaphor of a 'web of causation,' which views events as being produced through a complex interaction of multiple conditions (risks) and countervailing protective or mitigating factors (e.g., sexual conduct exposes individuals to the risk of contracting STDs, while condoms protect against, but do not eliminate the risk). What is important is that such a model challenges all simple models of direct causation in which X 'causes' Y.7 However, as epidemiological reasoning is imported into everyday forms of risk discourse it seems to confer the mantle of causality upon the phenomena being discussed. Thus, for example, if tofu eaters have a lower rate of heart attacks than others this will probably be because tofueating is linked to many other factors that distinguish between tofu and

176 Alan Hunt non-tofu eaters, such as vegetarianism or exercise. Discursively, the result is claims that 'tofu-eating reduces the risk of heart attack,' and not uncommonly the even more extreme claim that tofu 'prevents' heart attacks. Such invocations of epidemiological reasoning in risk discourses imports a form of expert knowledge that further facilitates the proliferation of linkages between anxieties and potential remedial strategies. Epidemiological reasoning becomes party to a potential expansion of regulatory interventions in many spheres of daily life about what is 'good' and 'bad' for individuals or social aggregates ('middle-aged men should ...' or 'pregnant women should ...'). An important feature of risk discourses is that they are both individualizing and totalizing. On the one hand, risk assessment serves as a factor in the calculative discourses of individual life chances. We change our patterns of activity and consumption to avoid risks and to promote some conception of our 'well-being.' On the other hand, 'risk' discourses, especially in their technical forms (statistics, actuarial tables, epidemiology), totalize aggregated populations (pregnant women, middle-aged men, drug users). There is, however, a constant movement back and forth between individual and totalizing logics in that what holds at the aggregate does not necessarily provide secure guidance with respect to individual practices. Thus it is important to bear in mind the differences between the everyday and technical discourses of risk. Everyday risk talk is much more directly concerned with identifying direct causal connections (X causes Y) and with assigning 'blame.' Such rhetoric, although invoking the quantitative and calculating logics of technical risk discourses, is generally systematically non-calculative. There is a persistent movement back and forth between individualizing and totalizing discursive forms that occur under the umbrella of risk. One of the results of the generalization of risk discourses is that there has been a tendency to advance theses about epochal shifts. In addition to Beck's metahistorical thesis about the rise of a 'risk society,' others have suggested that an historical transition is occurring in which disciplinary society is being displaced by insurance technologies (Donzelot 1979; Ewald 1991). Castel goes so far as to speak of a 'post-disciplinary order' (1991: 281). This approach suggests that risk technologies are less coercive, more efficient, and significant aids to governmental projects in being less likely to provoke resistance.8 In contrast to this view I suggest that attention should be focused upon the way in which different techniques of governing form shifting configurations of elements that do not replace each other, but which find specific linkages between the tech-

Risk and Moralization in Everyday Life

177

niques themselves and the socio-political contexts in which they are mobilized (O'Malley 1996). I have previously sought to draw attention to the way in which discourses interact in such a way that the resultant combination is distinct from the component elements.9 I now want to consider the role of such combinations with respect to institutional practices. The governmentality school has pointed to the way in which contemporary liberal governmental practices are dispersed through a diverse range of institutions functioning through expert knowledges in a calculative manner, driven by a concern with drawing those to be governed into the process by promoting various techniques of self-governance, and attentive to the danger of governing too much. I do not want to argue that this general line of argument is wrong, but rather that it does not tell enough of the story of 'real liberalism.' Such rational liberal governance coexists with another face that I will term liberal governance in the name of risk,10 which while sustaining a concern with governing too much is prone to an ever-present urge to govern more. These dual tendencies to govern both less and more constitute a paradox at the heart of liberalism. One face of the governance of risk has been extensively explored in the governmentality literature epitomized by rational forms of risk management through insurance technologies. The other face of governance in the name of risk is characterized by risk avoidance rather than risk management. Risk avoidance is dominated by a preoccupation with 'safety.' The goal of safety seems to have an apparently self-evident normative value in that it is better to promote safety than to chance danger. This mix of anxiety and a desire for safety can result in a stultifying and restrictive mentality that, rather than serving to reduce anxiety, can fuel anxieties and promote further waves of regulatory interventions. An expanding panoply of regulation in the name of safety mandates, for example, helmets for cyclists and floatation devices for boaters. It also generates calls for restrictions on 'risky' pursuits, such as potholing and rock-climbing, and produces a social world scarred by visual warnings of risk and danger. The preoccupation with risk and safety further underlines the negativity that suffuses so much risk discourse. This negativity precludes the possibility of positive association between risk and pleasure. Risk theory employs a negative construction of risk as bad, something to be avoided, and thereby ignores or downplays the positive face of risk as excitement and pleasure.11 Risk discourses can be viewed as weapons in the struggle for normali-

178 Alan Hunt zation and conformity by identifying the non-risky as normal. Deviance or transgression is constructed as 'risky' and becomes moralized as a possible prelude to instituting some regulatory intervention. For example, the recent increased preoccupation with the risks associated with sexual activity has normalized monogamous sexual relations and stigmatized those who have multiple sexual partners as deviant. This in turn has revived calls for registration and notification of morally sensitive diseases which had last been raised in the controversies over syphilis in the late nineteenth century (Brandt 1985). As is so often the case with risk discourses, they give rise to paradoxes. Anxieties about the risks of STDs has induced a wide-ranging 'safe sex' program as a major element of contemporary public health policy. But at the same time this concern has come to dominate much of sex education in such a way that the dangers rather more than the pleasures of sex are promoted. Intentionally or otherwise, the result is a climate of sexual conservatism (Corrigan 1997). In his essay on governmentality, Foucault refines some early formulations by insisting that he does not claim that the nineteenth century witnessed the replacement of a society of sovereignty by a disciplinary society, to be replaced in its turn by a governmental one. Rather, he suggests, 'we have a triangle; sovereignty-discipline-government, which has as its primary target the population' (1991: 102). The deployment of risk partakes of precisely the characteristics of Foucault's 'triangle.' It results in an expansion of interventions which are increasingly characterized by their 'mixed' character, in that incitement to individual risk reduction is coupled with welfarist regulation of aggregate populations. The most prominent instance is smoking, where individuals are called upon to 'butt out' and regulations restrict the opportunities to light up. In some instances, such as the proliferation of safety offences (e.g., health and safety, motor vehicles, etc.), they take the form of the legalization of risk, which induces detailed managerial regulation. In other instances the result is an expansion of disciplinary practices associated with the management of public spaces, as in the proliferation of regulatory signs such as 'No Skateboards.' The disciplinary character of such regulation is ironically underlined in the TV advertisement for skateboard gear that shows police hassling tennis players while pointing to a 'No Tennis' sign. In all of these combinations in which risk discourses are mobilized is the constant possibility of a moralization of 'risky' conduct: skateboards are 'risky' in a way in which tennis is not deemed to be.

Risk and Moralization in Everyday Life

179

However, one amendment to Foucault's triangle thesis seems necessary. He was correct in identifying the mechanisms of welfare or security whose target was an aggregated population as the major manifestation of the governmental network by the end of the nineteenth century. But the expansion of government through risk effects a significant 'doubling' of the target of governance. Not only are risk practices directed at populations (e.g., anti-tobacco health campaigns), they also facilitate the targeting of individuals (those identified as being 'at risk'). One of the major characteristics of risk governance is that it facilitates the movement back and forth between populations and individuals, and is thus both individualizing and totalizing. It is significant that the new entrepreneurial discourses of neo-liberalism significantly downplay the positive associations of risk taking. This second face of risk in the guise of safety is fundamentally a bureaucratic set of responses concerned to generate and enforce sets of rules that are not simply intended to promote safety, but more importantly to demonstrate that an attempt has been made to bring the future under control. If accident, injury, or loss occurs the rule makers are able to demonstrate that they had shown foresight in identifying potential risks and had taken reasonable steps to prevent or minimize harm. One of the prominent themes developed in recent governmentality studies has been the identification of the concern not to govern too much as a distinctive feature of liberal modes of governance (Rose 1993; Burchell 1993; Gordon 1991). This line of inquiry builds on Foucault's observation, 'It is here that the question of liberalism comes up. It seems to me that at the very moment it became apparent that if one governs too much, one did not govern at all - that one provoked results contrary to those one desired' (Foucault 1984: 242). This conception of liberalism involves a continual questioning of the activity of governing. It requires that a 'constant critical scrutiny be exercised over the activities of those who rule - by others and by authorities themselves' (Rose 1993: 292). Valuable though these insights may be with regard to the general characteristics of contemporary liberalism, it is important that political liberalism's own self-conception in terms of the promotion of regimes of deregulation should not be taken at face value. It is necessary to engage with the evidence that liberal encounters with problems that are posed as issues of risk, and that elicit concern with safety, lead to an expansion of regulatory interventions. These responses are frequently not only inefficient, but socially alienating. The outcome is an intensification of

180 Alan Hunt regulation of the type identified by Habermas (1986) as thejuridification of the lifeworld. The rules and prescriptions for individual conduct devised by alliances of bureaucrats and experts, frequently pressured by one or more social movement organizations, are highly formalized and take little or no account of the practices of the everyday world. Some of these rules are simply silly and unnecessary, such as the ubiquitous 'No Bill Posting' signs that denote the best sites to post bills (Hermer and Hunt 1996). Others cause unnecessary inconvenience. 'No jaywalking' inconveniences pedestrians on the assumption that the normal individual drives a car. The classic instance of juridified, that is law-like, regulation is the ever-present 'No Trespassing' signs. It should be noted that such regulation frequently explicitly invokes the claim of danger, as in 'Danger - No Entry' as a modified form of 'No Trespassing' signs. While safety rules find their fullest development in institutional settings, they have been penetrating the private realm. Prohibitory signs ('No Smoking,"Perfume Free'), often employing the standard red prohibitory circle, have found their way into private cars and homes. The presumed imperative of litigation avoidance has come to provide the major legitimation of the regulatory explosion. Institutions characteristically claim to regret the need to expand regulation, attributing it to fiscal responsibility, which requires the introduction of regulations that serve to minimize the possibility of litigation or to limit its impact. This development is itself a direct product of the expansion of risk discourses. Risk warnings are found everywhere. Most familiar are the warnings in restaurants that the management does not accept responsibility for customers' coats and hats, and the many public venues that warn those about to enter that they do so 'at their own risk.' Such risk warnings have expanded to include the more modern and ridiculous warnings that coffee is hot, that ice is cold, that knives are sharp, ad infinitum. It should be noted that these liability avoidance techniques are closely linked to other forms of risk-handling practices. Thus criminal law and insurance practices combine with risk management and self-governance to mutually interact, as is illustrated in such everyday practices as driving a car, where wearing a seat-belt is not only a legal requirement but can have major implications for insurance benefits in the event of personal injury. The permutations are legion, but for present purposes it is sufficient to draw attention to the massive increase in the generation and dissemination of rules and regulations in an expanding array of social sites. My own university posted notices prohibiting faculty members (despite an official 'cyclist friendly' policy) from taking their bikes into

Risk and Moralization in Everyday Life

181

their offices. Inquiry elicited the 'explanation' that bikes might constitute an 'obstruction,' which in the event of fire might result in the university incurring legal liability to anyone injured. What is significant about the explosion of such petty regulation is that it fundamentally contradicts the apparently attractive feature of risk analysis, namely, that it makes possible the quantification of risk and thus enables a scientific risk management. Instead, the expansion of risk discourses provides compelling evidence of a liberalism that governs too much. Risk and Moralization

It is not only the proliferation of regulation in the name of risk management that provides grounds for scepticism about the rationalism of risk discourses. The more potential 'causes' of risks are identified, the more potential 'solutions' are available. Perhaps the most important implication of the proliferation of risk discourses is that they have facilitated an expansionary logic of responsibilization. As more potentially 'risky' substances are identified, the list of items to be reduced or eliminated from everyday patterns of consumption becomes longer. This expansion of responsibilization through which individuals are encouraged, provoked, and incited to engage in taking care of themselves in turn results in an expansion of the moralization of individual conduct. Behaviour that fails to incorporate risk-avoidance practices comes to be viewed as irresponsible; not only is such conduct unwise, it is increasingly viewed as 'wrong.' The recent history of the tobacco wars illustrates this escalation from risk avoidance to moralization and regulation. When the epidemiological links between smoking and multiple forms of disease became widely disseminated it became prudent for individuals to quit smoking. However, this individual responsibilization was soon transformed into a moral issue. Not only was it wise to quit, but quitting smoking increasingly became a major signifier of respectability, a testament to the capacity for self-care through self-control. This in turn produced a form of externalized normative judgment that affected relations with others. The posting of 'No Smoking' signs in both public and private sites became a potent moral discourse that is further reinforced by the degradation ceremony in which those lacking the appropriate self-control reveal that lack while huddled outside buildings indulging their vice. The particular significance of posing risks as 'moral questions' is that it acts as a mechanism of closure. Discussion of complex and controversial issues is closed down by excluding or refusing other forms of discur-

182 Alan Hunt sive interrogation. In the first stage moralizing discourses homogenize a wide variety of different issues. For example, in recent child pornography debates the moralizing response has insisted that any photograph of a naked child is pornographic, such that parents now make sure that their children are clothed before taking holiday snapshots at the beach. The second step is the act of closure. The assertion that the issue is a 'moral question' excludes considerations other than issues of moral judgments. Thus if child pornography is defined as inherently evil there can be no space to consider matters such as artistic merit or speech rights. Perhaps the clearest instance is provided by the slogan 'abortion is a moral issue.' This formulation acts as a closure that disallows consideration of any other dimension, such as choice or entitlement to exercise self-control over fertility. While such moral closures are commonplace, it is also necessary to interrogate the Equally common linking of moral issues to considerations of 'harm.' It is necessary to attend to the way in which 'harm' is invoked. The invoking of harm not only draws attention to material consequences but also involves an implied assertion of moral harm by identifying some conduct as inherently 'evil' and thus again acts as a closure. More generally moral-harm discourses take a more distinctively modern form by employing a supplemental argument: some conduct is judged immoral because it causes some identifiable harm.12 However, in an epoch in which overt moralization runs the risk of being greeted with some suspicion, risks are frequently posed in such a way as to downplay a moral dimension by highlighting medical discourses. This has produced a powerful new force, that of medico-moralization. One of the more significant manifestations of medico-moralization was provided by the 'discovery' of secondary or passive smoking. This expansion of moralization was effected through the amplification of the risk discourses levelled at tobacco. From here it was but a short step to the introduction of public regulatory prohibitions on the permissible sites of the now thoroughly moralized practice of smoking. Similar medico-moralization is increasingly being directed against alcohol. While a simple return to a politics of prohibition is unlikely to succeed, the reemergence of the temperance mentality is a significant component of the expansion of regulatory endeavours pursued in the name of health promotion.13 In late modernity not to engage in risk avoidance constitutes a failure to take care of the self. Risk avoidance becomes a 'moral enterprise relating to issues of self-control, self-knowledge and self-improvement'

Risk and Moralization in Everyday Life

183

(Lupton 1999a: 91). Whereas in the early modern era the 'dangerous classes' were disciplined through coercive strategies, risk-based crime control strategies focus less on the individual than on the environment, as exemplified by the introduction of video surveillance of subway platforms. What impact does this shift from 'dangerousness' to 'risk' have upon the forms and practices of moral regulation? The provisional answer is that the techniques of moral regulation remain much the same as they were in the classical period of moral regulation at the end of the nineteenth century. Then and today moralization operates both to individualize the wrongdoer and to constitute deviant types so as to provide legitimations for intervention. In so far as there is a difference between these historical forms of moralization, contemporary moralization appears, as argued above, superficially more benign, less overtly moralizing. Thus the conjunction of risk discourse with long-standing strategies of moral regulation serves only to change its surface. A society of risk discourses, as distinct from a 'risk society,' is far less 'new' than the more extravagant claims of 'risk society' theory would have us believe. Risk and Deresponsibilization

The proliferation of risk discourses has expanded the range of social situations in which individuals can become responsibilized.14 While the 'big risks' change slowly, as witnessed by the way in which concern with 'global warming' has slowly inserted itself in popular consciousness, the minor risks spring up more suddenly and tend to generate logics which urge the individual to take responsibility by responding to them. It should also be noted that there is considerable discursive activity that seeks to translate responses to 'big risks' into the arena of individual responsibility. Slogans such as 'save a tree' to encourage recycling or 'one less car' for cyclists undoubtedly play a significant role in stimulating a sense that an individual can 'do something' about the big risks. We may leave aside the question of whether such action serves to allay anxieties or contributes meaningfully to a solution. Responsibilization is always double-sided. It lays down a norm against which individuals, groups, or institutions may evaluate their own conduct. But it also opens up the possibility of moralization in so far as others may seek to hold individuals to that standard, regardless of whether they have accepted the responsibility. In this sense responsibilization always involves moralization. If the expansion of responsibilization simply led to further pressure

184 Alan Hunt for individuals to take on board a greater responsibility for their own conduct and their care of the self in responding to risks, such developments might well be regarded as unproblematically positive.15 However, the proliferation of risks and their concomitant responsibilities produces an opposite response, which can involve either a refusal of responsibility or, perhaps more significantly, the transfer of responsibility to others. This alternative reaction I will term deresponsibilization, which flags those responses to risk in which individuals refuse to accept responsibility for tackling risks resulting from their own choices and instead transfer responsibility and blame onto others. Perhaps the classic instance of deresponsibilization is witnessed among individual smokers who are suing tobacco companies, seeking compensation for smokingrelated ill-health.16 We need have little sympathy for the tobacco companies in order to note the significance of the denial of personal responsibility for lifestyle choices such as smoking, which has long been recognized as a significant health risk and which, perhaps more significantly, has been the target of long and sustained campaigns of selfresponsibilization. More emblematic examples of deresponsibilization are those of the woman who successfully recovered damages when she spilled hot coffee on herself and the employee who recovered damages for injury suffered when she caused an automobile accident after drinking alcohol at an office party.17 Such cases highlight the more absurd manifestations of deresponsibilization, but they are also evidence of a pervasive trend in both judicial and legislative policy making in a number of jurisdictions that are seeking to transfer the burden of responsibility from those who engage in 'risky' conduct to those who provide the means to undertake the activity. A similar process can be detected in another mechanism of deresponsibilization which allows individuals to shift responsibility to others by claiming the status of victim. Victims are enmeshed in the discourses of risk and anxiety. The emergence of crime-victim movements attests to the increasing sense that the world is an irrational, threatening place, enhancing fear, generating anxieties, and feelings of vulnerability. The common feature is the construction of the 'innocent victim' and the parallel constitution of the offender as moral outsider in a process that 'glorifies the values of one group and demeans those of another' (Gusfield 1967: 177). The typical victim is identified as weak, incapable of self-defence, but an essentially good and unsuspecting individual (Weed 1995).18 The most important feature of the victim

Risk and Moralization in Everyday Life

185

movement is that it often becomes captured by state institutions or 'law-and-order' organizations. One of the most significant manifestations has been the pervasive dilemma of feminist politics in having to choose between conceiving women as 'agents' or as 'victims.' While some currents have enthusiastically embraced woman as victim, others have sought to problematize 'victim' analysis by focussing on how women become active participants in the very processes that objectify and dominate them. The special significance of the proliferation of the victim status is that the risks articulated serve as a legitimatory discourse for calls to regulate others and, significantly, harmonize with projects of moral regulation. It is important to consider the logic that underlies policies that permit or encourage the various forms of deresponsibilization. It may be that it is only in the context of the expansion of liability insurance that such developments can be understood. Such a policy is arguably part of a pervasive legal strategy to impose the economic burden of accidents and injuries on those who are in a position to bear the cost and to defray their liability through insurance (Ericson and Haggerty 1997: 51). Suing an institutional party for injury arising from risky conduct might be regarded not so much as attempting to shift moral responsibility as merely taking advantage of an economic opportunity. While this may provide insight into motivations in specific instances, in aggregate such practices are diametrically opposed to what has been identified as one of the most significant features of neo-liberal forms of governance, namely, that individuals are subject to a diversity of strategies and tactics to induce them to engage in practices of self-governance. Thus there is a contradiction at the heart of modern forms of rule. While individuals are activated to take responsibility for mitigating an expanding inventory of everyday risks, there is, at one and the same time, an increasing tendency to deny or refuse responsibilities either by shifting them onto others or simply by blaming others. There are two faces of responsibilization for modern risks. That these two faces cohabit is the outcome of the process whereby accepting risk responsibilities frequently involves becoming encumbered with the moralization which, as has been argued, is the penalty for the assumption of responsibility for engaging with the risks of modernity. In brief, taking on board responsibility for risks is itself a 'risky' undertaking. Deresponsibilization provides an explanation of the processes that conservatives such as Gertrude Himmelfarb have described as 'demoralization' (1995).19 We do not have to mourn the decline of character

186 Alan Hunt and 'the virtues' to understand how it is that people both seek responsibilities and just as strenuously refuse to accept responsibilities, processes that are the result of viewing life through the lens of risk discourses. These conditions of existence have similarities to the process that Giddens (1991: 8) has sought to capture by means of the concept of 'sequestration,' in which people lose direct contact with events and situations that involve broad issues of morality and finitude. But sequestration does not yet capture the strange and paradoxical sense in which an intensification of moralization goes hand in hand with demoralization. There is a coexisting tendency for intensive but shifting outbursts of moral rage, with the ominous figure of the paedophile currently the most demonized. Such intensity coincides with wide fields of moral indifference in which private interest and convenience override concern for others. The awareness of risk generates a calculating, actuarial attitude on the part of the late modern citizen that is manifest in a wariness towards others. On the one hand, this is expressed in calculating responses through insurance and litigation, but on the other hand, there is evidence of a more pervasive suspicion, hostility, and intolerance, the extreme forms of which have become bracketed under the labels of ethnic cleansing and genocide. Risk in Everyday Life

Risk appears in a number of different guises. It is important to avoid narrowing our conceptualization of risk in such a way as to consider only its technical calculative dimension. Such a tendency is most marked in the risk assessment and risk management fields that find their most developed expression in the fabrication of management as science. The sociological 'risk theory' associated with Beck and Giddens has tended to operate with this somewhat restricted sense of risk by focusing predominantly on the 'big risks.' However, risk discourses play a pervasive role in everyday life. The trope of 'risk' has become one of the widespread discursive forms within which a range of social, environmental, and ecological concerns are expressed. In these domains risk discourses are significant contemporary forms through which anxieties are experienced and articulated. Anxieties are not mired in the controversy which has counterposed 'realist' and 'constructionist' views of risk, commonly presented as a clash between 'risk theory' and 'sociocultural' theory of risk (Crook 1999; Lupton 1999b). Anxieties are real and it is a matter for subsequent analysis to decide whether a particular anxiety is rational or reasonable.

Risk and Moralization in Everyday Life

187

Risk in everyday life does not necessarily involve quantitative or calculative forms of reasoning. While these specific forms of analysis play distinctive roles as forms of governmental rationalities, it should not be presumed that everyday risk discourses share any or all of these characteristics. Indeed, their most distinctive role is to provide contemporary forms of moralizing judgments. This usage is ideologically significant in that, unlike standard moral judgments ('X is wrong'), the normative dimension is not immediately apparent. Thus, while 'having unsafe sex is risky' does not directly state that 'unsafe sex is wrong,' the moral judgment lurks very close to the surface. Risk discourses provide an apparently utilitarian grounds for regulatory intervention without appearing directly to institute a form of moral regulation.

Notes 1 By employing the conceptualization of 'everyday life' I seek to avoid the macro-micro distinction between 'big' and 'little' risks that haunts so much of the literature on risk. 2 This form of moral discourse was itself not very far removed from a more specific theological form that generated moral codes ('Thou shall not steal,' 'Six days shalt thy labour,' etc.), whose authority was grounded in some privileged text or institution. 3 Responsibilization refers to a form of governing that discursively imposes specific responsibilities on individuals relating to their own conduct or that of another for whom they are presented as being responsible. Much responsibilization is entirely conventional: parents for their children, employers for the work of their employees. But the technique lends itself to expansion. Increasingly, patients are deemed responsible for the management of their recovery from illness, expected to adopt the right activities. Pregnant women are responsibilized for the well-being of their foetuses, expected to abstain from an expanding range of 'risky' behaviour. 4 This line of thought has links to Luhmann's (1993) couplets of risk/security and risk/danger. Both approaches share the view that risk is not something that can be eliminated, but always involves the making of decisions and choices. My thanks to Melanie White for drawing my attention to this parallel. 5 This should not be taken to imply that expert systems are always captured. Indeed, it would be interesting to study the question of under what circumstances expert systems resist the risk promotion projected by risk entrepreneurs. For example, efforts to ban fluoridization have been resisted and it

188 Alan Hunt currently appears that the 'experts' are not inclined to capitulate to opposition to the genetic manipulation of staple crops. 6 I have argued elsewhere that a number of very different social theories have directly or indirectly made use of some conception of anxiety to advance causal explanations of why it is that particular circumstances have given rise to social problems and how these problems may be understood. Appeal to anxiety is not misleading, but the concept of anxiety is made to carry more causal weight than it can sustain (Hunt 1999b). 7 One way of avoiding a preoccupation with linear causality is found in the strategy of actor-network theory that focuses on the interaction and coexistence of diverse elements. This approach is put to good use by Crook (1999). 8 The thesis that there has been a transition from discipline to risk technology has surprising resonances with the older 'social control' theory, in which generalized strategies of power were viewed as operating through bureaucratic institutions, such as schools, medicine, and in particular the welfare state, to establish all-encompassing regimes of surveillance and control over pacified populations (Van Krieken 1991). 9 The concept of 'combinatory articulation' or 'hybridity' allows attention to focus on discursive linkages and slippages. Moral discourses are at their most potent when they succeed in linking two or more different discursive strands, as when a discourse appeals to both economic interest and moral values. Valverde observes that 'discourses of moral reform ... can be analysed as a continuous series of slippages, a potentially endless movement between conflicting definitions of apparently unitary "social problems'" (1990: 64). The discursive elements not only 'slip' to embrace new constellations of meaning, they also recombine to constitute new problematizations which achieve a degree of stability until disrupted by some subsequent slippage. It is only in their combination and interaction that they stimulate a level of social reaction that single concerns would rarely generate. 10 An approach that has close parallels with that employed in this chapter is advocated by Pat O'Malley (1992), the major difference being that I add the part played by moralizations in linking risk and regulatory techniques. 11 One of the paradoxes of risk discourses in neo-liberal society is that the generally negative connotations of risk conflict with the valorization of entrepreneurial risk taking. 12 It should be noted that in some specific instances the making of a moral judgment is intentionally avoided. Thus Mothers Against Drunk Driving (MADD) strive to avoid any suggestion of favouring alcohol prohibition by focusing attention exclusively on 'harm' in the form of death and injury

Risk and Moralization in Everyday Life

189

caused by drivers under the influence of alcohol. They moralize the conduct but not the substance. 13 The expansionary logic of moralized risk is evident in a flyer on 'Sex and Alcohol' circulated to university students, which insists that 'People under the influence of alcohol (or drugs) are absolutely incapable of giving sexual consent.' 14 While responsibilization has become closely associated with neo-liberalism, with its inducement that individuals become entrepreneurs, it should be noted that responsibilization played a significant role in other historical periods. But the responsibilities created were concentrated within social roles rather than individualized. For example, a long-standing responsibilization is that associated with motherhood; a more historically specific instance was the responsibilization of the male 'breadwinner' during the nineteenth century (Dean 1991: 13). 15 It is in this sense that Etzionian communitarianism presses for the expansion of responsibilities and duties to the community as the price to be paid for individual rights. His 'new golden rule' is that 'The basic social virtues are a voluntary moral order and a strong measure of bounded individual and subgroup autonomy held in careful equilibrium' (Etzioni 1997: 244). 161 distinguish between individual litigation and those cases in which health departments and authorities are suing to recover the costs incurred in the provision of medical treatment to those suffering from illness attributable to tobacco. 17 In February 2000, an Ontario court ordered a real estate company to pay US$200,000 for their 12.5 per cent responsibility for injuries suffered by the defendant, an employee, who was involved in an accident after she had consumed alcohol at the firm's Christmas party. The employer had offered to call the defendant's husband. The defendant had subsequently visited a bar, where she consumed more drinks, and three hours after leaving the firm's party she was severely injured in a motor vehicle accident. The defendant was deemed 75 per cent responsible and the bar owner 12.5 per cent responsible for the accident. 18 It is only a short step from this construction of the victim as 'innocent' to one in which the victim is viewed as being to blame for the risks suffered. For example, Margaret Thatcher argued that 'we only have ourselves to blame' for the crimes that befall us if, for example, we fail to help the police or neglect to lock our doors and windows. That the discourses on victimization can be used in different ways should not persuade us to return to a simplistic view of the victim as always innocent and thus lacking in responsibility.

190 Alan Hunt 19 A more modulated conservatism is expressed by Christopher Lasch (1984) whose 'minimal self within consumerist capitalism experiences a sense of powerlessness and victimization that encourages a defensive contraction of the self.

References Barber, B.R. 1995. Jihad versus McWorld. New York: Time Book Beck, V. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage. Bourdieu, P. 2001. Masculine Domination. Trans. Richard Nice. Stanford: Stanford University Press Brandt, A.M. 1985. No Magic Bullet: A Social History of Venereal Disease in the United States Since 1880. New York: Oxford University Press Burchell, G. 1993. 'Liberal Government and Techniques of the Self.' Economy and Society 22(3): 267-82 Castel, R. 1991. 'From Dangerousness to Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality, 281-98. Chicago: University of Chicago Press Corrigan, P.R. 1997. 'What Do They" Want From Us? Moral Regulation Gets Real in England and Wales in the 1990s.' In K. Rousmaniere, K. Dehli, and N. de Coninck-Smith, eds., Discipline, Moral Regulation and Schooling: A Social History, 247-71. New York: Garland Publishing Crook, S. 1999. 'Ordering Risks.' In D. Lupton, ed., Risk and Sociocultural Theory: New Directions and Perspectives, 160-85. Cambridge: Cambridge University Press Dean, M. 1991. The Constitution of Poverty: Toward a Genealogy of Liberal Governance. London: Routledge - 1999. 'Reflexive Government.' In Governmentality: Power and Rule in Modern Society, 176-97. London: Sage Donzelot, J. 1979. 'The Poverty of Political Culture.' Ideology & Consciousness, 5: 71-86 Douglas, M.T. 1992. Risk and Blame: Essays in Cultural Theory. London: Routledge Elias, N. 1978. The Civilizing Process. Oxford: Basil Blackwell. Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press Etzioni, A. 1997. The New Golden Rule: Community and Morality in a Democratic Society. New York: Basic Books Ewald, F. 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds. The Foucault Effect: Studies in Governmentality, 197-210. Chicago: University of Chicago Press

Risk and Moralization in Everyday Life

191

Foucault, M. 1984. 'Space, Knowledge and Power.' In P. Rabinow, ed., The Foucault Reader, 239-56. New York: Pantheon Books - 1985. History of Sexuality: Vol. II The Use of Pleasure [1984]. New York: Random House - 1991. 'Governmentality.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 87-104. Chicago: University of Chicago Press - 1999. 'What Our Present Is.' In S. Lotringer, ed., Foucault Live: Collected Interviews, 1961-84, 2nd ed. New York: Semiotext(e) Fukuyama, F. 1999. The Great Disruption: Human Nature and the Reconstitution of Social Order. New York: Free Press Giddens, A. 1990. The Consequences of Modernity. Stanford: Stanford University Press - 1991. Modernity and Self-Identity: Self and Society in the Late Modern Age. Cambridge: Polity Press Gordon, C. 1991. 'Governmental Rationality: An Introduction.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 1-51. Chicago: University of Chicago Press Gusfield, J.R. 1967. 'Moral Passage: The Symbolic Process in Public Designations of Deviance.' Social Problems 15: 175-93 Habermas, J. 1986. 'Law as Medium and Law as Institution.' In G. Teubner, ed., Dilemmas of Law in the Welfare State, 203-20. Berlin: de Gruyter Hacking, I. 1999. The Social Construction of What? Cambridge: Harvard University Press Hermer, J., and A. Hunt. 1996. 'Official Graffiti of the Everyday.' Law and Society Review 30: 501-26 Himmelfarb, G. 1995. The De-Moralizat ion of Society: From Victorian Virtues to Modern Values. New York: Knopf Hunt, A. 1999a. Governing Morals: A Social History of Moral Regulation. Cambridge: Cambridge University Press - 1999b. 'Anxiety and Social Explanation: Some Anxieties About Anxiety.' Journal of Social History 32: 509-28 Jackson, P. 1994. 'Passive Smoking and Ill-Health: Practice and Process in the Production of Medical Knowledge.' Sociology of Health and Illness 16: 423-47 Lasch, C. 1984. The Minimal Self: Psychic Survival in Troubled Times. New York: W.W. Norton Luhmann, N. 1988. 'Familiarity, Confidence, Trust: Problems and Alternatives.' In D. Gambetta, ed., Trust: Making and Breaking Cooperative Relations, 94-107. Oxford: Blackwell. - 1993. Risk: A Sociological Theory. New York: de Gruyter Lupton, D. 1999a. Risk. London: Routledge

192 Alan Hunt - ed. 1999b. Risk and Sociocultural Theory: New Directions and Perspectives. Cambridge: Cambridge University Press Mill.J.S. 1910. 'On Liberty.' In Utilitarianism, Liberty and Representative Government. London: Everyman O'Malley, P. 1992. 'Risk, Power and Crime Prevention.' Economy and Society 21: 252-75 - 1996. 'Risk and Responsibility.' In A. Barry, T. Osborne, and N. Rose, eds., Foucault and Political Reason: Liberalism, Neo-Liberalism and Rationalities of Government, 189-208. Chicago: University of Chicago Press Power, M. 1997. The Audit Society: Rituals of Verification. Oxford: Oxford University Press Putnam, R.D. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster Ritzer, G. 1996. The McDonaldization of Society: An Investigation into the Changing Character of Contemporary Social Life. Thousand Oaks: Pine Forge Press Rose, N. 1993. 'Government, Authority and Expertise in Advanced Liberalism.' Economy and Society 22: 283-99 - 1999. The Powers of Freedom: Reframing Political Thought. Cambridge: Cambridge University Press Simon, J. 1988. 'The Ideological Effects of Actuarial Practices.' Law and Society Review 22: 772-800 Susman, W.I. 1984. Culture as History: The Transformation of American Society in the Twentieth Century. New York: Pantheon Valverde, M. 1990. 'The Rhetoric of Reform: Tropes and the Moral Subject.' International Journal of the Sociology of Law 18: 61—73 Van Krieken, R. 1991. 'The Poverty of Social Control.' Sociological Review 39: 125 Wagner, D. 1997. The New Temperance: The American Obsession with Sin and Vice. Boulder: Westview Press Weed, FJ. 1995. Certainty of Justice: Reform in the Crime Victim Movement. New York: Aldine de Gruyter White, M., and A. Hunt. 2000. 'Citizenship: Care of the Self, Character and Personality.' Citizenship Studies 4: 93-116 Willig, C. 1997. 'Trust as Risky Practice.' In L. Segal, ed., New Sexual Agendas, 125-35. London: Macmillan

8

From Risk to Precaution: The Rationalities of Personal Crime Prevention KEVIN D. HAGGERTY Although crime prevention has become increasingly prominent over the past twenty years, it can be a slippery and overly inclusive concept. There are, in fact, few traditional criminal justice practices that could not be justified as a means of preventing crime. That said, new developments are occurring under the rubric of crime prevention. These largely relate to the locus of anti-crime decision making - who is responsible for preventing crime, and what means are available to accomplish this goal. Such changes can themselves be connected with wider trends in the practice of governance. The 1980s witnessed a sweeping decline of welfare-based strategies of government in various social spheres. Under attack from the political right for being inefficient, fostering dependency, and reducing individual initiative, Keynesian welfare state policies were put to rest as the failed legacy of 'social engineering.' During the same period neoliberalism arose to promote self-governance. It constituted individuals as integral to the management of the risks they face. Citizens now undergo a process of responsibilization in which they are enticed, persuaded, and coerced into viewing their life as a perpetual process of risk taking and risk management. This life project involves adopting a calculative attitude towards security, health, finances, and career. Neo-liberalism valorizes the individual as the rational manager of his or her own risk portfolio; he or she is expected to adopt a calculative attitude by becoming knowledgable about risk. Such knowledge is presumed to enable the individual to make prudent decisions about what practices, tools, and services to embrace in efforts to take and manage risk. The process of managing risks is significantly different from efforts to avoid danger. Danger refers to a host of untoward social or natural

194 Kevin D. Haggerty eventualities that are often 'out there' in our experiential environment. Risk, in contrast, exists entirely 'in here,' in assorted aggregate knowledges that delineate the probability of future unfavourable events based on an analysis of previous occurrences (Ewald 1991). As Douglas (1986: 19) observes, risk is about 'trying to turn uncertainties into probabilities.' While some dangers can be avoided risk is constant, in that all actions (even inaction) carry some probability of an unfavourable outcome. The goal, then, becomes one of prospectively managing risks through measures that: 1) reduce the statistical probability of an untoward outcome and 2) employ insurance-based schemes that pool risk to financially compensate the unfortunate few who have experienced a loss. A prominent working assumption of neo-liberalism in relation to criminal justice is that 'the rational subject of risk takes on the capacity to become skilled and knowledgeable about crime prevention and crime risks' (O'Malley 1996: 201). This neo-liberal scenario underlies the rise of a more active role for individuals in securing themselves against crime risks. Another key aspect of neo-liberalism has been the equation of individual responsibility with the individualized logic of the market place. With the diminished role of the state as agent of social welfare, individuals are encouraged to govern themselves through protective services and products offered for sale. A slew of product providers have embraced this more prominent role for the market and rushed in to fill the void left by the withdrawal of the state. Personal security has become commodified, with assorted products now claiming to be able to reduce crime or crime-related anxiety (Ericson and Haggerty 1997; Marx 1995). The following list is by no means complete but it provides an appreciation for the range of anti-crime products currently being marketed to private citizens: personal alarms, access controls, steering wheel locks, pepper spray, gated communities, guard dogs, bullet proof vests, cellular telephones, instruction in martial arts, car alarms, surveillance cameras, handguns, motion-sensitive lighting, vehicle geographic positioning systems, home alarms (infrared, ultrasonic, photoelectric, and audio sensor), personal electronic monitors, motion detectors, missing child kits, private security services, light timers, window bars, fencing, safes, and tasers. Most recently we have also seen the development of 'antipaedophile' internet software, personal geographic positioning systems, encryption programs, and home-based drug tests (Moore and Haggerty 2001). All of these products promise to provide greater security for some

From Risk to Precaution

195

combination of self, home, family, or vehicle. 'Security' has become the promised end-state of purchasing tools and services, much in the same way that the market has commodified beauty and health. This chapter argues that the neo-liberal model of the responsibilized citizen inaccurately depicts how people manage their risk profile. Using the example of crime risks, I argue that neo-liberalism idealizes the rationality of individual decision makers. While responsibility for taking actions to forestall criminal victimization increasingly falls to potential victims, it is not clear that such efforts are exclusively 'skilled and knowledgeable.' Decisions taken to avoid personal victimization do not necessarily involve a calculative form of rationality, but can be informed by a host of other emotional, symbolic, and cultural considerations. Any understanding of the rationality of responsibilized decision makers must take into consideration a broader array of social purposes served, and meanings conveyed, by anti-crime efforts beyond a utilitarian effort to reduce crime. After outlining the rationally calculative standard of risk that the public is expected to employ, the majority of this paper details an alternative logic of precaution that emphasizes the limited scope for the operation of a calculative rationality. It underscores how crime prevention decisions can be influenced by more informal knowledges and assorted considerations pertaining to a desire to reduce anxiety and engage with the social meanings conveyed by using anti-crime technologies and services. The conclusion addresses what types of limits might be placed on precautionary decision making. One of the places where this precautionary logic is articulated and amplified is in advertisements for security products and services. Consequently, the following analysis refers to advertisements for assorted personal security products to demonstrate some of contours of the precautionary logic. The nexus formed between the responsibilized individual and the market makes such advertisements a valuable site in which to examine the discourse of precaution. Such a data source is also particularly revealing as it demonstrates the myriad cultural factors that can impinge on an individual's decisions about how to manage crime risks. Advertisements from security industry trade magazines such as Security Management, Canadian Security, Cahners' Security, and Intersec for the years 1999-2000 were analysed. In addition, I visited more than a hundred personal security websites and collected assorted promotional material from the security industry in order to analyse how these products are marketed.

196 Kevin D. Haggerty Rational Standards for Crime Prevention

Where advocates of crime prevention through environmental design or routine activities theory posit the existence of a 'rational criminal actor,' critics contend that such individuals are fabrications (Garland 2001). The ideal of a rational criminal actor conveniently caricaturizes the complex routes that can lead individuals to commit crimes, a caricature that legitimates classical models of crime control and punishment. Yet there has been no comparable analysis of nee-liberalism's assumption of a 'rational prudent citizen' who strives to reduce his or her risk of criminal victimization. While individual citizens are increasingly constituted as the vital cog in situational crime prevention efforts, there has been no explicit discussion of how such individuals are expected to make such decisions. In particular, we are left to wonder about what assorted factors are apt to influence a person's decisions about the extent and content of their anti-crime efforts. It is simply assumed that individuals will embrace a calculative attitude to determining the risks that they face and adopt appropriate measures in light of their probability of being victimized. As Crawford observes, in the model of situational crime prevention 'individuals appear logically capable of voluntary choice, free to act in a rational self-interested fashion, be they potential offenders or victims' (1998: 72). We can gain a glimpse of this logic in the calls by some crime prevention experts for the use of quantitative crime data to determine the 'hot spots' of crime. Citizens are also encouraged to conduct security audits or risk analysis of their homes, cars, and routines. However, the calculative logic that purportedly motivates such efforts is generally more a background assumption of situational crime prevention than an explicitly articulated attribute. One of the best ways to gain an appreciation for the extent to which individual citizens are presumed to act in a rationally calculative manner is therefore to examine what happens when individual crime-related concerns and actions diverge from that ideal. One of the most important insights derived from victimization surveys is that the groups who most fear crime are often the least likely to be victimized. Where a person's fears and actions do not reflect their statistical likelihood of victimization, there is often an explicit or implicit criticism that they have failed to live up to a rationally calculative standard. The extensive sociological literature on moral panics frequently displays such a logic, suggesting that there is evidence of a moral panic when the degree of public and media attention to a problem does not

From Risk to Precaution

197

correspond to the likelihood of being harmed by that problem. Some authors have explicitly claimed that a 'moral panic' cannot ensue unless the public and media pay 'disproportionate' attention to a problem (Goode and Ben-Yehuda 1994: 36). Disproportionality is demonstrated by the fact that harms that are more statistically likely, or produce greater measurable social ills, do not receive comparable attention. One of the most prominent authors working in this tradition is Joel Best, who has consistently demonstrated that the public's fears and actions relating to such phenomena as Hallowe'en sadists (Best and Horichi 1985), child abductions (Best 1990), or gang violence (Best 1999) are routinely media-stoked overreactions to statistically improbable events. This form of analysis has become common outside of the academy, as evidenced by the popular success of Glassner's (1999) The Culture of Fear. In the first few pages of this book the author notes disapprovingly that Americans are terrified of crime despite a declining crime rate. Concerns about unemployment are high even though the unemployment rate dropped below 5 per cent for the first time in twenty-five years. Health-related anxieties are also prominent, despite the fact that life expectancy doubled during the twentieth century. The irrationality of these and other fears is demonstrated time and again by showing that they do not correspond with their statistical probabilities and historical trends. Glassner subtitled his book, Why Americans Are Afraid of the Wrong Things. The implicit assumption of his and comparable studies is that to be afraid of the 'right' things would involve some kind of actuarial calculus whereby what we fear and seek to prevent closely corresponds with what is likely to harm us. Individuals are consequently held to a standard of calculative rationality where they are expected to attend to and mitigate the real risks they face as demonstrated by the statistical likelihood that something will cause them harm. When they take actions to mitigate a risk that is probabilistically low, they face the prospect of being accused of acting irrationally or having been swept away by a moral panic. This risk-based approach is neither an accurate nor fair standard for judging individual crime-related fears and actions. We need to broaden our understanding of how such decisions are made by attending to the various non-calculative motivations, anxieties, and desires that inform anti-crime efforts. Actions to avoid criminal victimization are embedded in a situational rationality of precaution rather than a calculative rationality of risk. This terminological shift from risk to precaution signifies an attempt to understand the epistemological and social position of

198 Kevin D. Haggerty responsibilized decision makers. Precautionary decision making has a different relationship to scientific knowledge, is concerned with a broader range of potentialities, and works in a separate logical register than decisions based on a template of risk. The remainder of this paper provides an overview of the precautionary model of responsibilized anti-crime decision making. It is inspired by Francois Ewald's (2002) arguments about recent transformations in the management of uncertainty in specific decision-making contexts. However, this is a creative appropriation of Ewald's approach, not an application, as at least two significant factors distinguish his preoccupations from my own concerns. First, for Ewald, the precautionary principle applies only to decisions about serious and irreversible potentialities, for example, catastrophic risks posed by the chemical, nuclear, and genetic sciences that can fundamentally transform our relationship to nature (Beck 1992; Giddens 1990). The second distinction concerns the decision-making agents. While decisions about nuclear or chemical risks are made within complex organizational processes, I am interested in the decision-making practices of responsibilized individuals. Precautionary Logic Is Non-calculative

The idealized rational subject of neo-liberalism manages her risk profile in light of a knowledge of the types of risks she faces, often with the assistance of expert advice. This model of responsibilized governance is stated most succinctly by O'Malley: 'Guided by actual data on risks and on the delivery of relevant services and expertise, the rational and responsible individual will take prudent risk-managing measures' (1996: 200). Such a risk-based model of governance 'presupposes science, technical control, the idea of possible understanding, and objective measurement of risks' (Ewald 2002: 282). However, such a depiction raises serious questions about the extent and quality of the public's knowledge of crime risks, the reliability of the advice offered by crime-prevention experts, and what 'calculation' might amount to in practice. It can, in fact, be very difficult for individuals to discern the 'the facts of the matter' about their risk of criminal victimization. Risk is a statistical construct, and many individuals lack the skills and training to work with even rudimentary statistics (Paulos 1988). Moreover, basic statistical skills might not be sufficient to determine your particular crime risk. Knowing the crime rate for your city, for example, might not be a good indicator of the crime rate for your neighbourhood, block, or workplace.

From Risk to Precaution 199 Your personal risk of victimization differs for different crimes, and can be influenced by complicated interactions pertaining to your age, gender, race, income, and occupation. The myriad limitations on official crime data make interpreting such numbers all the more difficult, even for statistically literate citizens (Haggerty 2001). These complexities are compounded if a person wants to calculate the degree to which employing a particular crime prevention practice might reduce her risk. Even when probability statistics are available for a population of which the individual is a member, there is no direct translation to how the particular individual should act. For instance, Ian Hacking (1990: 86) observes that when practical medicine deals with a given patient, 'obtaining more data about different individuals is irrelevant to the particular case of the patient we wish to treat.' Likewise Porter (1995: 102) discusses how insurers facing a particular local situation cannot directly address what to do on the basis of population statistics alone. Individuals who must decide about potential criminal victimization are in a similar situation, in that there is always a void when moving from population data to deciding about how they should act. To add to these epistemological problems, crime is only one among the plethora of risks that an individual risk manager faces, meaning that she is expected to be knowledgeable about the assorted facts that have a bearing on her entire risk portfolio. Given that this is not a likely scenario, this epistemological void is filled by other types of knowledge. Advertisers for security products and services capitalize on the public's inability to make critical sense of most numbers. Shocking statistics about crime present an image of a society in which criminal danger is assumed to exist in any imaginable situation. These numbers depict high crime rates, increasing violent crime, and the spectre of random victimization. They are numbers from nowhere, devoid of qualifications or operational definitions. The advertisement for an automobile steering wheel lock informs us that 'More vehicles are stolen in January than in any other month' and then notes that a vehicle is stolen every twenty-five seconds in the United States. It then moves away from the quantitative into the fantastic: 'A carjacking can happen anytime a thief sees the opportunity.' The promotional material for a woman's self-defence book lists a series of alarming crime statistics, concluding: 'These shocking facts show that violent crimes can affect anyone at anytime, regardless of where they live or work ... Based on these statistics it's possible at some point in your life you might be a victim of violent crime.' None of these advertisements mention the steady decreases in the crime rates that

200 Kevin D. Haggerty most North American jurisdictions have experienced over the past decade. Moreover, as these numbers are aggregates, their demonstration of fluctuations in the crime rate is also of no specific use to individual citizens. Not only are many of these statistical claims partisan, they are often highly dubious. One advertisement suggests that 75 per cent of all rapes are perpetrated by a man known to the victim, while another for a home door lock claims that 60 per cent of all rapes occur during a home invasion. Were both of these statistics true it would mean that a sizeable minority (35 per cent) of rapes would occur during a home invasion by someone known to the victim - a highly unlikely figure. A host of statistical inconsistencies are also apparent across different advertisements, with one advertisement warning that a woman is raped in America every five minutes, while another proclaims, 'A woman is raped every 46 seconds in America ... that's 78 rapes each hour!' One of the more common scenarios invoked in these advertisements is the prospect of child abduction. Such advertisements often follow the strategy used by political claims-makers of employing shocking statistics to draw attention to this problem (Best 1989). In the 1980s the problem of 'missing children' surged to the front of the political agenda. Provoked by the recurrent claim that 50,000 children were missing in America, various state and private agencies swung into action in an effort to solve this problem. However, it soon became apparent that there was considerable latitude in what advocates were counting as an instance of a 'missing child.' Different data sources employed different operational definitions of both 'missing' and 'child.' This meant that at times any or all of the following were included in the activists' claims about the number of 'missing children': youths who were momentarily out of their parents' sight, missing adults, teenagers who did not come home, youths living with a non-custodial parent or grandparent, and a significant number of runaways. Depending on what data source you used, the count of 'missing children' in America could be as high as 50,000 or as low as 70 (Best 1989). Not surprisingly, activists routinely employed the largest and most alarming numbers. Such inflated claims are prominent in advertisements for security products, such as the website for a child protection kit that cautions that in America a child goes missing every twenty-four seconds. Were this true it would mean that the impossibly high number of 3,600 children went missing every day, 1,314,000 children every year.

From Risk to Precaution 201 The relation of these numbers to the truth of crime is not important. Such statistics are a form of rhetoric that reinforces and augments public anxieties: they repeatedly communicate the simple message that everyone should be afraid of crime. Neo-liberal models of governance assume that individuals will seek out the assistance of experts to help them understand the nature of the risks they face and how these might be mitigated. Experts help to outline strategies of risk reduction and risk maximization for individuals who cannot personally calculate these probabilities (Johnson 1993; Rose 1993). Crime prevention advice is offered by the police, the insurance industry, criminologists, security consultants, locksmiths, and social advocates, among others. A host of mundane decisions about where to live, what to wear, who (and how) to date, how to construct homes, transport children, secure vehicles, and configure computers are now informed by expert-mediated considerations about crime and victimization. However, the notion that there is a coherent realm of 'expertise' about crime risks poses its own problems. Expert opinion about crime is incredibly fractured. Experts often cannot agree on the dimension of some of the most high-profile criminal risks, such as sexual assault, child abduction, or impaired driving (Best 1989; Gilbert 1997; Gusfield 1981). Many 'experts' who offer advice about crime prevention also have a financial or ideological interest in inflating the incidence of crime in order to sell products and services or advance a law and order ideology. Moreover, some of the most high-profile claims makers about crime risks have structural incentives to inflate the perceived level of risk of certain types of crimes in order to capture public attention, push for legislative reform, or secure government funding. Accepting expert advice involves a leap of faith, a trust in that individual and the expert system she is aligned with (Giddens 1991; Shapin 1994). In attempting to compute their crime risks, individuals rely on assorted cultural cues to decide which experts to consult, what risks to focus on, and how to alleviate these risks. Such determinations are influenced by 'common sense' and long-standing practice, as individuals draw on pre-existing emotional, ideological, and cultural discriminations to guide their behaviour. As Hunt (see chapter 7) observes, 'decisions about which brand of expert knowledge to follow have to be made, like so many other lifestyle choices, on the basis of subjective preferences whose roots and motivations we are rarely able to perceive' (171). While individuals can align themselves with expert systems that are based on an

202 Kevin D. Haggerty actuarial logic - property insurance, for example - the individual decision about whether to purchase such insurance or not is hardly based on an actuarial calculation. Precaution Deals with Catastrophic Potentialities

Precaution differs from risk in terms of the nature of the potential loss. Ewald's model of precaution is concerned with the dangers inherent in large-scale science that may fundamentally and irrevocably change our future relationship with nature. That such potentialities transcend the risk paradigm is evidenced by the fact that individuals faced with such possibilities are largely unsatisfied with financial compensation but seek to hold individuals criminally liable. Compensation, in fact, makes little sense in precautionary situations, as the only logical objective can be to avoid the occurrence of events with irreparable and irreversible consequences. The question of what an individual might consider to be a serious and catastrophic potentiality, however, immediately leads us into the realm of subjective perceptions. From the standpoint of the individual risk manager, a whole series of possibilities related to criminal victimization would count as serious and catastrophic. The types of harms that 'crime' conjures up in the public imagination are often the most gruesome and personally devastating events, such as murder, physical/sexual assault, and child abduction. Moreover, even forms of victimization that an outsider might view as relatively minor can be catastrophic for the victim. Domestic burglaries are a prime example. While an actuary might see the resultant losses as particularly well-suited to financial compensation, money cannot easily make up for the loss of sentimental items, nor can it alleviate the fear and sense of personal violation invoked by a burglary (Nicholson 1994). The logic of precaution is concerned with avoiding such subjectively defined catastrophic potentialities. Precautionary decision makers exist in a fundamentally different epistemological context than risk-based decision makers. Precautionary actions are taken to avoid the worst eventuality that can be imagined, not to reduce the likelihood that a probable event will occur. Advertisements help to extend the range of imagined catastrophes by presenting an extraordinary array of potential personal disasters that can ostensibly be thwarted by purchasing new security products. An extreme example can be found in the millennial issue of Security Sense. This self-consciously hip glossy magazine is sold as 'The complete magazine of vehicle security

From Risk to Precaution

203

and convenience.' In one of its extended 'articles' readers are invited to imagine themselves as a modern-day James Bond, complete with a vehicle fully loaded with assorted high-end security products: 'Say that your secret agent buddy is waiting for you in your car, while you are inside the government consulate for Istanbul doing some late night research. All of a sudden your buddy in the car sees someone is coming ... Luckily your friend can press the summons button inside the car and you will automatically receive a page.' The same article suggests that one 'secret agent' advantage of a remote ignition system is that it 'allows you to start your car remotely to first see if anyone accidentally planted a bomb in your car before you get in. Pretty clever.' As the technological abilities of security tools outstrip the likely circumstances of personal victimization, marketers construct highly implausible scenarios to sell their products. Precautionary logic operates in a context of subjective uncertainty about the probability of victimization. As such, precaution is not really concerned with 'risk' at all. Precaution 'invites one to anticipate what one does not yet know, to take into account doubtful hypotheses and simple suspicions' (Ewald 2002: 288). A personal anecdote helps illustrate this subjective dimension of precautionary concerns. Years ago a friend of mine lived with her parents in a wealthy section of Toronto's relatively quiet suburb of Scarborough. After a highly atypical shooting in the area, her parents insisted that we keep the blinds drawn for fear of 'snipers.' No amount of reassurance or appeals to logic could convince them that they did not need to take precautions against being shot through their window. The contrast between a logic of risk and a logic of precaution in crime prevention is further demonstrated in the following three hypothetical, but not uncommon, scenarios. A woman purchases a handgun to protect her family, ignoring the objections of friends who point out that the weapon is much more likely to be used to harm her, or her family, than a criminal intruder. Second, a father purchases a 'Child Identification Kit,' which allows him to maintain his daughter's records in one location, including photographs, fingerprints, dental records, and hair samples, in the event that she is abducted. Again, friends note that child abductions are extremely rare. Third, a single woman installs a sophisticated home alarm and monitoring service after her home was burgled of sentimental objects. Friends inform her that the likelihood of being burgled twice in her neighbourhood is remarkably low and, moreover, that the cost of the alarm and security service would quickly add up to more than she lost in the original burglary. In all of these scenarios the

204 Kevin D. Haggerty responsibilized citizen is working on a logic of precaution that is concerned with the 'just-in-case' subjectively defined worst-case scenario, while the 'friends' are evaluating those behaviours in light of a calculative logic that focuses on statistical probabilities of victimization. Pointing out that a person's fears have a low statistical probability of occurring is small comfort to someone operating on the precautionary principle. The usual response is 'Well, if it stops only one (murder, assault, child abduction ...) then it is worth it,' or, 'It makes me feel safer.' Precaution Is Motivated by Anxiety and Fear

O'Malley refers to the model of individual responsibilized governance as 'prudentialism.' This implies the use of rational calculation to avert or mediate risks, as indicated by Webster's definition of 'prudence' as 'the ability to govern and discipline oneself by the use of reason.' However, as the hypothetical examples used above demonstrate, the type of reason employed to manage crime risks can be imbued with highly emotional considerations. Anxiety figures prominently among the factors that shape what crime risks people attend to and the types of responses they embrace. Fear and anxiety can be a powerful catalyst for action. As fear is not necessarily rational, however, the issues that individuals attend to are not likely to accord with a probabilistic 'hit list' of likely victimization. We should not be surprised by the lack of correspondence between those crimes that people fear and those that they are apt to experience. Public anxieties about crime are too symbolically loaded, politically aligned, and psychically embedded to ever conform to an actuary's risk profile. How these fears become manifest will vary among individuals and will depend on people's 'unique biographies, especially their histories of anxiety and how they have come to handle the circumstances of their lives in light of these' (Hollway and Jefferson 1997: 265). Any discussion about crime anxiety must acknowledge the influence of the popular media. Newspapers and television produce a steady stream of crime-related narratives which are the public's primary source for information about crime. While these accounts do not necessarily cause viewers to fear crime in a direct causal fashion (Gunter 1987; Sparks 1992), they nonetheless have powerful and diffuse consequences. The significant proportion of news and entertainment programming that is dedicated to crime, particularly violent crime, can foster an environment where the embrace of a host of strategies and technologies to negate the prospect of highly improbable crimes seems reasonable and justified.

From Risk to Precaution

205

After the release of the movie The Hand That Rocks the Cradle, which portrays a demonically evil nanny, a rash of parents approached the police to have their nannies security checked (Ericson and Haggerty 1997). Subsequent television shows have aired disturbing videos of nannies surreptitiously captured in the act of hitting children. This newly crystallized fear was turned into financial profit with the development of the 'nanny cam': miniature cameras placed in household objects such as lamps, digital clocks, stereos, and smoke detectors. One of the more popular models places a surveillance camera inside a teddy bear, a move that hardwires the technology of insecurity into the icon of childhood security. In the process nannies, who were often de facto extensions of the nuclear family, are transformed into objects of suspicion and surveillance. An advertisement for a nanny cam boasts: 'Make sure your home and family are safe, even if you're not there! Install the Pan/Tilt Remote Camera in any room in your home, and you can check on your loved ones from any Internet connection in the world. Make sure the nanny is taking care of your children properly.' A different nanny-cam advertisement uses parental guilt as a marketing strategy. Two angelic children stare at the viewer from a surveillance camera photograph. Above them a caption reads 'My Mommy and Daddy say that I'm the most important and precious thing in their lives..." The implication is that if the children are indeed so precious, no measure should be spared to enhance their security. Fear is frequently used to sell security products. Advertisements help to channel and focus diffuse and amorphous anxieties about crime and other forms of social breakdown into distinct and personalized fears about someone entering your bedroom window, stealing your car, or attacking you on your evening jog. To do so they use vivid testimonials, unsettling images, and shocking statistics that simultaneously reflect and reproduce popular crime anxieties. Consider the following advertisement for the women's self-defence book Kick the Rapist Where It Hurts, Although the author claims that she has no intention of fear mongering, the entire advertisement is designed to have precisely that effect: I've seen it all in the movies and read it in the papers. But those things happen to others and not to me - or so I thought... Just think about the world we live in today. I don't want to scare you, I just want you to realize the fact - No one is immune to violence. Doesn't it make sense to know how to be AWARE and also be PREPARED before it's too late? (original emphasis).

The valorization of fear is becoming a pervasive trope in the crime prevention literature. Rather than being denigrated for being irrational,

206 Kevin D. Haggerty fear is now held out as an important emotion that can help keep you alive, or at least reduce your chances of criminal victimization. Nowhere is this more apparent than in De Becker's (1997) popular book, The Gift of Fear, which portrays fear as an effective pre-rational warning system. Individuals are encouraged to attune to, and nurture, their intuitive fears of crime. Advertisements for crime prevention tools increasingly capitalize on this anxiety, suggesting that their products provide 'peace of mind.' Devices such as pinhole cameras and pepper spray are not marketed simply as instrumental anti-crime technologies, but as devices that fulfil a host of psychic needs. This is nicely articulated in a website banner that reads: 'Even if you never have to use that pepper spray or stun gun you are carrying, you will feel more secure and confident knowing that it's there' (original emphasis). Criminologists should not be surprised by this public embrace of the fear of crime. It is a continuation, in a different register, of trends in criminal justice policy over the past two decades. As fear of crime has become recognized as a social problem, it has, in turn, been singled out as an object of governance in its own right. The result is that the range of potentially pertinent 'crime-related' policies extend well beyond those that promise to reduce crime to include those that strive to shape public perceptions about crime and disorder. Such policies clash with corporate advertisements and popular media accounts that augment the concern about crime in efforts to sell products or capture market share. Cultural Considerations Inform Precautionary Decisions

As precautionary decision making is generally not based on scientific types of evidence and calculation, such decisions are opened up to considerable influence from local cultural factors. As John Adams (see chapter 4) points out, it can be liberating when decision making moves out of the realm of knowledge, as we are then 'free to act upon our convictions, prejudices, and superstitions' (91). This is increasingly apparent in relation to 'large-scale' scientific risks, where the public and scientists alike must now take into consideration opinions that are recognized as scientifically marginal or dissident. Ewald suggests that this dimension of the precautionary principle 'invites one to take the most far-fetched forecasts seriously, predictions by prophets, whether true or false' (Ewald 2002: 288). The range of crime prevention prophets includes assorted institutional actors that have a vested financial or ideological interest in promoting a fear of crime.

From Risk to Precaution

207

Freed from a purely rational calculus, precautionary decisions are opened up to the seductions of security commodities and the meanings these tools convey about ourselves and our lifestyles. Individuals are integrated into a consumer society by defining their identities, desires, and interests in terms of possessing commodities. The commodification of security suggests that the objects we purchase and the services we employ in the name of combating crime also communicate meaning about who we are, both to ourselves and to others (Loader 1999). As a result, personal anti-crime decisions can involve an assessment of what such initiatives will signify about you as a person. A note of caution is in order at this point. Although it is important to attend to the meanings communicated by owning or employing a particular security product, such connotations are themselves polysemic. For the purposes of this paper, however, it is not as important to provide a detailed analysis of the multiple significations that crime prevention efforts can convey as it is to stress that consideration of the social meanings of anti-crime devices and actions can influence choices about crime prevention initiatives. The individual consumer is one of the most important audiences for the significations of crime prevention. Such efforts reflect reassuring messages back to us about our capacity as responsible moral agents. Personal security products can be some of the most intimate and familiar objects in our day-to-day lives. The pepper spray canister is only useful if it is proximate - in your purse, on your belt, your key chain, or attached to your car's visor where, as one advertisement proclaims, 'it is needed the most.' Routine encounters with such products are a blunt material testament that the individual is doing her part to avoid victimization. The marketing of personal security products capitalizes on this intimacy, and extends it into visions of individualized empowerment, as in the stun-gun advertisement that urges you to 'Experience the difference of knowing you can protect yourself.' Potential purchasers of a 'Taser' are invited to 'Watch them shake on the ground after using the Talon 250C 250,000 volt defense.' This empowerment can be wrapped up in jingoism, as in the advertisement that proclaims: 'Since the dawn's early light, self-defense is an American right! God Bless America!' Elements of our social identities are formed through the risks we attend to and how we attenuate those risks. Security technologies help to form these identities through their sign value, a form of prestige that is appropriated and displayed in the act of being consumed (Loader 1999). The techniques used to counter risk can communicate unique charac-

208 Kevin D. Haggerty teristics about your personality. For example, a familiar strategy to transform a mass-produced consumer good into a testament of its owner's individuality is to offer the product in a range of different colours, which purportedly speaks to an owner's unique style. This tactic was used to great effect in the marketing of Macintosh iBook and iMac computers and is now used to sell security products. Surveillance camera casings now come in assorted brilliant colours, as do key chain fobs for car alarms and pepper spray canisters. Purchasers of 'The Club' steering wheel lock can accessorize their car's security in the patriotic American colours of red, white, or blue. Gendered identities are also fostered through the consumption of security products. As Stanko (1997) has observed, concern about crime risk, particularly risk of sexual assault, has long helped to fashion women's subjectivities. Women's anti-victimization strategies have traditionally involved practical advice about attire, comportment, and response to hazardous situations. Today's responsibilized woman is encouraged to be more active in her self-defence, and security products play a prominent role in such strategies. Some tools, such as pepper spray and personal alarms, are highly gendered and targeted predominantly at female consumers. Women are also encouraged to embrace more lethal technologies, such as handguns designed specifically for a female market, or the 'Taser' with a 'slim grip' to make it easier for women to use. The construction of masculinity in security product advertisements usually involves an image of the patriarch who spares no expense in securing his home and family. Alternatively, men are encouraged to engage in ultra-masculine self-defence or revenge fantasies. The meanings of commodities are relational; they situate the consumer in the context of a 'lifestyle' that speaks to that individual's broader consumption community (Featherstone 1991; Sarup 1996; Slater 1997). Each lifestyle has a distinctive constellation, or habitus, of preferences in consumer goods, art, sports, and food (Bourdieu 1984). Part of the appeal of some security commodities relates to their ability to mesh with such lifestyle communities (Spitzer 1987). This is apparent in the way in which jogging, that distinctive marker of middle-class preoccupations with fitness and an athletic aesthetic, plays into the marketing of security products. Personal alarms, and mace and pepper spray advertisements make frequent reference to protecting joggers. A major security company has developed a 'Timex Ironman' sports watch, which allows consumers to monitor their home security system from the watch's display. Such strategies integrate the security product into a

From Risk to Precaution

209

collection of interests and preoccupations characteristic of middleclass market segments. The meanings of objects help to mark the social order by communicating social relationships and classifications (Bourdieu 1984). One important set of meanings communicated by security commodities relates to the class hierarchy. Such tools offer 'distinctions' that delineate different class and consumption communities from one another. Dinner party conversations that gravitate towards the strategies and technologies of home protection, for example, are not purely information-gathering small talk, but contribute to the formation of a contemporary middleclass sensibility. Burglar detectors and car alarms are now as much markers of middle-class status as they are anti-crime devices. Individuals incorporate such devices into a class-based habitus. For the middle and upper classes, this habitus can include an intimate familiarity with the routines and practicalities of operating a home alarm or securing a car stereo. Moreover, the class distinctions that these tools delineate are continually in transition. Where previously only the very rich had car alarms, now the vehicles of the middle classes come factory-equipped with alarms, while the wealthy install aftermarket security systems that can disable the fuel system, lock the ignition, or track the car from space on a geographic positioning system should it be stolen. As these more fanciful tools become more affordable to the middle classes, the class distinctions that they currently perform will again be transformed. Commodity seduction also plays a role in the security market. Among certain segments of the population, security products are desirable objects to possess in their own right. The role of seduction in the security market is perhaps most apparent in relation to the handgun. Legitimated as a means of self-defence, handguns are marketed as physically seductive objects that are imbued with meaning related to freedom, sex, and power in American culture. Other security products are characterized by their technologically sophisticated electronic infrastructures, which are prominently displayed in advertisements, thereby helping to equate security with owning yet another consumer gadget. The more fanciful options become instilled with a degree of panache, as in the case of the car alarm advertisement that enthuses, 'the new remotes are cool!' Even cooler is the ability to access and control your car's security system from anywhere on the planet with a cellular phone, although it would be difficult to argue that this technology meets a pressing security need. Such gadgets are seductive toys in search of a problem. The security industry now markets the aesthetic qualities of some of its

210 Kevin D. Haggerty products. This is alternatively an attempt to alert people to the seductions of these products or an effort to overcome the highly functional aesthetics that security tools have traditionally displayed. So, the 'Big Jammer' door lock is now described as having a 'clean, baked enamel finish' that 'makes it attractive for in-house use,' while brass padlocks provide 'stylish protection.' A home alarm and motion detector are both described as 'compact and attractive' while a device that prevents individuals from kicking down your door is 'an attractive accessory to your door - available in brass and antique finishes to add to the beauty of your home.' Surveillance cameras provide 'the finishing touch to any wellappointed room' and a motion detector is marketed on the basis of its 'decor-enhancing shape.' Pepper spray comes with an 'attractive leatherette holster' and is advertised on television during the Christmas season as 'a great stocking stuffer' (Doyle 1998: 108). Perhaps the most prominent example of the seductions of commodified security are the attractive wrought iron window bars that are both stylish objects of art and efficient crime-preventing devices. Conclusion

Precautionary decisions about crime operate according to a situational rationality that is shaped by informal knowledges, personal history, anxiety, and the cultural meanings of crime prevention. The role of such factors in decision making suggests that there might be few calculatively rational constraints on crime prevention measures. This brief conclusion considers where the limits of precautionary decision making might be drawn in practice. Part of the larger context for the rise in personal crime prevention efforts has been the transformation in the social position of the middle classes in Western societies. As Garland (2000) has observed, starting in the 1960s there was a change in the middle-class experience of crime. Where previously the 'crime problem' was predominantly a concern for the poor and dispossessed, increasing levels of crime and a sustained public discourse about crime brought the issue home to the middle and upper classes: 'rising crime rates ceased to be a statistical abstraction and took on vivid personal meaning in popular consciousness and individual psychology' (Garland 2000: 360). Public attitudes towards crime and punishment hardened and the market in personal security expanded. Where the poor had traditionally been forced to rely predominantly on state efforts to secure them against crime, the now anxious and

From Risk to Precaution 211 newly responsibilized middle classes turned to the market in hopes of enhancing their own personal security profile. However, the private embrace of market solutions to crime raises questions about how much security is enough. To answer that question we must first recognize that the world of precautionary decision making is populated by atomized individuals. Not that there is such a space in reality, but the factors individuals are encouraged to attend to in making such decisions seldom extend beyond their immediate family. Situational crime prevention envisions a society populated by people who will do bad things should the opportunity arise. The best approach to preventing criminal behaviour is therefore a strategy of crime diversion. You should make your car, home, self, or family such a difficult or unappealing target that the potential criminal moves on to another, less well-protected target. Advertisements for security devices and police promotional material repeatedly stress this message. Therefore, the parameters on precautionary crime prevention efforts are local. Individuals need to employ enough security measures to ensure that the potential criminal goes to the next house or neighbourhood. One paradox of this local standard of reasonableness is that it introduces a deeply anti-social dimension into precautionary anti-crime decisions that works against the more communitarian focus of other approaches to crime prevention. These latter approaches encourage communities and organizations to work together to solve neighbourhood social problems. Precautionary decisions work at cross-purposes to such communitarian efforts because the precautionary approach assumes that security is a zero-sum game in which gains in personal security are made at the expense of one's immediate neighbours and neighbouring community. The standard of precautionary security is shifting, contextual, and competitive, such that if your neighbour installs a higher fence, or super-sensitive alarm, your home ostensibly becomes a more attractive criminal target in the situational ecology of insecurity. Neighbour is pitted against neighbour, with private investments in security products and services providing that vital extra boost that elevates your security the additional notch beyond the security measures established by the state, your community, or your neighbours. In light of this escalation, it becomes increasingly difficult or irrelevant to suggest that a certain anti-crime measure is excessive or misguided. No external rational cost/benefit calculus on anti-crime security is apt to prospectively constrain decisions about personal security. Precaution employs a fluid, subjective, and contextual standard of reasona-

212 Kevin D. Haggerty bleness. What a responsibilized citizen will deem to be reasonable will be influenced by his or her personal experience of victimization, the number of security measures established by neighbours, sensitivity to mediated fears, and the amount of money available to spend on services and technologies. For some, this will mean a relatively small investment in personal security, for others the sky will be the limit. As Davis (1990: 248) has noted, some of the homes of Los Angeles' super-rich are now built with terrorist-proof security rooms, accessible by sliding panels and secret doors. Ultimately, this means that we must rethink the point of studies that accentuate the gap between what the public fears and the statistical likelihood of being harmed by such eventualities. Such studies employ a calculatively rational benchmark to evaluate public fears and actions that seems to have only a minimal bearing on why individuals choose to embrace a particular constellation of security measures. Acknowledgments I would like to thank Richard V. Ericson and Serra Tinic for their helpful comments on this chapter, and Michelle Robinson for her research assistance.

References Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Best, J. 1989. 'Dark Figures and Child Victims: Statistical Claims About Missing Children.' In Images of Issues: Typifying Contemporary Social Problems, 21-37. New York: Aldine de Gruyter - 1990. Threatened Children: Rhetoric and Concern about Child-Victims. Chicago: University of Chicago Press - 1999. Random Violence: How We Talk About New Crimes and New Victims. Berkeley: University of California Press Best, J., and G.T. Horichi. 1985. The Razor Blade in the Apple: The Social Construction of Urban Legends.' Social Problems 32: 488-99 Bourdieu, P. 1984. Distinction: A Social Critique of the Judgement of Taste. Cambridge: Harvard University Press Crawford, Adam. 1998. Crime Prevention and Community Safety. London: Longman Davis, M. 1990. City of Quartz. London: Verso.

From Risk to Precaution 213 De Becker, G. 1997. The Gift of Fear: Survival Signals that Protect Us from Violence. Boston: Little Brown Douglas, M. 1986. Risk Acceptability According to the Social Sciences. London: Routledge and Kegan Paul Doyle, A. 1998. '"Cops": Television Policing as Policing Reality.' In M. Fishman and G. Cavender, eds., Entertaining Crime: Television Reality Programs, 95-116. New York: Aldine de Gruyter Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press Ewald, F. 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality, 197-210. Chicago: University of Chicago Press - 2002. The Return of Descartes' Malicious Demon: An Outline of a Philosophy of Precaution.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Featherstone, M. 1991. Consumer Culture and Postmodernism. London: Sage Garland, D. 2000. The Culture of High Crime Societies: Some Preconditions of Recent "Law and Order" Policies.' British Journal of Criminology 40: 347-75 - 2001. The Culture of Control. Chicago: University of Chicago Press Giddens, A. 1990. The Consequences of Modernity. Cambridge: Polity - 1991. Modernity and Self-Identity: Self and Society in the Late Modern Age. Cambridge: Polity Gilbert, N. 1997. 'Advocacy Research and Social Policy.' Crime and Justice 22: 101-48 Glassner, B. 1999. The Culture of Fear: Why Americans are Afraid of the Wrong Thing. New York: Basic Books Goode, E., and N. Ben-Yehuda. 1994. Moral Panics: The Social Construction of Deviance. Oxford: Blackwell Gunter, B. 1987. Television and the Fear of Crime. London: John Libby Gusfield, J. 1981. The Culture of Public Problems: Drinking-Driving and the Symbolic Order. Chicago: University of Chicago Press Hacking, I. 1990. The Taming of Chance. Cambridge: Cambridge University Press Haggerty, K.D. 2001. Making Crime Count. Toronto: University of Toronto Press Hollway, W., and T.Jefferson. 1997. The Risk Society in an Age of Anxiety: Situating Fear of Crime.' British Journal of Sociology 48: 255-66 Johnson, T. 1993. 'Expertise and the State.' In M. Gane and T.Johnson, eds., Foucault's New Domains, 139-52. London: Routledge Loader, I. 1999. 'Consumer Culture and the Commodification of Policing and Security.' Sociology 33: 373-92 Marx, G.T. 1995. The Engineering of Social Control: The Search for the Silver

214

Kevin D. Haggerty

Bullet.' InJ. Hagan and R. Peterson, eds., Crime and Inequality, 225-46. Stanford: Stanford University Press Moore, D., and K.D. Haggerty. 2001. 'Bring It On Home: Home Drug Testing and the Relocation of the War on Drugs.' Social and I^egal Studies 10: 377-95 Nicholson, P. 1994. The Experience of Being Burgled. Sheffield: University of Sheffield O'Malley, P. 1996. 'Risk and Responsibility.' In A. Barry, T. Osborne, and N. Rose, eds., Foucault and Political Reason: Liberalism, Neo-Liberalism and Rationalities of Government, 189-207. Chicago: University of Chicago Press Paulos,J.A. 1988. Innumeracy: Mathematical Illiteracy and Its Consequences. New York: Hill and Wang Porter, T. 1995. Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton: Princeton University Press Rose, N. 1993. 'Government, Authority and Expertise in Advanced Liberalism.' Economy and Society 22: 283-99 Sarup, M. 1996. Identity, Culture and the Postmodern World. Athens: University of Georgia Press Shapin, S. 1994. A Social History of Truth: Civility and Science in Seventeenth-Century England. Chicago: University of Chicago Press Slater, D. 1997. Consumer Culture and Modernity. Cambridge: Polity Sparks, R. 1992. Television and the Drama of Crime: Moral Tales and the Place of Crime in Public Life. Buckingham: Open University Press Spitzer, S. 1987. 'Security and Control in Capitalist Societies: The Fetishism of Security and the Secret Thereof.' InJ. Lowman, R. Menzies, and T. Palys, eds., Transcarceration: Essays in the Sociology of Social Control, 43-58. Aldershot: Gower Stanko, E.A. 1997. 'Safety Talk: Conceptualizing Women's Risk Assessment as a "Technology of the Soul."' Theoretical Criminology 1: 479-99

PART THREE The Moral Risks of Insurance

This page intentionally left blank

Introduction

In Part Three, Pat O'Malley, Tom Baker, Carol A. Heimer, Richard V. Ericson, and Aaron Doyle use the insurance industry as a substantive vehicle for understanding risk and morality. Pat O'Malley examines the history of efforts to tame chance through contract law. He documents how these efforts were surrounded by moral uncertainties about the boundaries of speculation, gambling, and insurance. Tom Baker explores insurers' claims to be classifying risks in order to maintain the integrity and viability of insurance pools. These claims reveal ideological and moral commitments that privilege individual interests over the common good, as well as insurers' interests. Carol A. Heimer takes up the question of insurers as moral actors, focusing on how insurance contracts and relationships shape moral conduct. Richard V. Ericson and Aaron Doyle use an empirical study of insurance fraud to examine the moral conduct of insurers. In the private justice system of insurance, fraud is tolerated to the extent that it contributes to premium revenue and efficient claims control. Moral judgments about insurance fraud are made in strictly utilitarian terms. Pat O'Malley observes that in contemporary globalizing economies there is renewed emphasis on speculation, which does not lend itself to statistical calculation. Predictability is said to be a thing of the past, and people are therefore encouraged to embrace uncertainty, become entrepreneurs, and take risks for capital gains. This embracing of uncertainty revives key aspects of nineteenth-century economic reasoning, which was embedded in practical aesthetics and common sense more than statistics. It was in part the use of social sciences in the twentieth century that changed the emphasis on speculation. Social sciences, and the institutions they underpinned, promised rational calculation based not

218 Part Three: The Moral Risks of Insurance only on statistics but also through bureaucracy, surveillance, and law. In the present-day world of globalizing economies we are perhaps retreating in the direction of uncertainty and speculation, and again encountering the troubling moral issues raised when 'gambling' enters into commercial relationships. In order to illuminate the present conjuncture, O'Malley offers a genealogy of uncertainty as seen through the lens of contract law. In nineteenth-century liberal theory, calculative rationality was the key to rendering subjects free and able to govern themselves. Because the process of calculative rationality was invisible, its operation was judged by effects, that is, the predictability of the behaviour of rational subjects. Predictability of behaviour became the subject of liberal theory and later of social science, shifting uncertainty into the realm of risk as a probability calculus. But in the pragmatic world of legal decision making about contractual relations among subjects, there was an uncoupling of uncertainty and risk. Through doctrines such as 'reasonable foreseeability' the foresight expected of a reasonable person in relation to a practical decision - contract law formed subjects who were expected to govern in the face of uncertainty. Beyond the abstractions of liberal philosophers and economists, the law recognized and enforced speculative contracts and articulated moral distinctions among legitimate speculation and gambling. In the early modern period, the moral concern with gaming was centred on its unproductive relationship with commerce. It was not speculation or reliance on fortuna that was of concern, but whether the transaction was simply a game, a wager, and whether the law should be involved in providing remedies or guarantees in such cases. A primary locus of working out the distinction between games and commerce was the insurance industry. Until the mid-1700s insurance could take the form of wagering, with no requirement of insurable interest: life insurance contracts could be taken out on the life of anyone, marine insurance contracts could be taken out on the safe completion of a voyage in which one had no investment, and so on. Legislation was subsequently introduced to require insurable interest, in other words that the policyholder must suffer a compensable loss in the case of a specified event occurring. The legal concern was less moral rectitude than a view that a commercial exchange should include material value, or title to value, which itself is an essential ingredient of advantage to commerce. By the mid-nineteenth century a substantial shift had occurred in the legal conception of the relation between gambling and exchange. With

Introduction

219

'difference' contracts, only formal legal title distinguished commerce from gambling. That is, the exchange was established within the discourse of law rather than in the actual movement of materials. The legal emphasis was now on the intent of the parties to the commercial relationship: did they intend to gamble? If not, they were left to speculate with regard to 'normal' levels of business competence in the particular case. Speculation thus became distinct from intentional wagering and was seen as an inevitable dimension of good contracts. In this development of contract law, there was little consideration of how gambling and speculation influenced the moral character of the parties to commercial relationships. The primary focus was the imagined effect on the economy of 'empty' exchanges that seemed to contribute to commerce. In the late nineteenth century concern arose about the morality of gambling in particular, while speculation on investment - which almost exclusively involved a separate class of risk-takin entrepreneurs - was seen as good for the economy. However, the rise of futures trading led to popular depictions of traders as morally corrupt gamblers, especially when they engaged in practices such as short selling. But in the legal world the concern remained the intent of the actors, in this case the intent to deliver on the futures contract. Moreover, the onus of proof of no intent to exchange resided with the challenger. Arguments evolved about the positive benefits to the economy enabled by futures trading, for example, its risk-spreading ability, income and cost spreading efficiency, and capacity to stabilize markets. In offering what O'Malley terms 'uncertainty spreading,' these insurance-like features allowed futures traders to reinvent speculation both morally and technically. Speculation was not only legitimated as a source of benefit, but through a series of legal decisions a loosening of the benefit to commerce/material exchange nexus emerged. Gambling could now be reconciled with insurance and investment. The practical view in law and commercial life was that prosperity would be threatened if trading were confined to actual material products. This view was only later taken up in liberal economic theory, exemplifying O'Malley's view that practical law and commerce helped produce the discourse of liberal theory. At the turn of the twentieth century, there was growing moral concern about fraud, and efforts were made to protect investors through expertise and regulation of 'speculative competence.' However, regulation was limited because of a laissez-faire belief that it is wrong to interfere with the right to invest, and because of entrepreneurial common sense about market uncertainties and the limits of expertise. Uncertainty was not

220 Part Three: The Moral Risks of Insurance reconciled with rational calculation, and speculation advanced as an important aspect of governance of the economy. O'Malley concludes with examples of a dramatic contemporary shift regarding what is deemed to be for the 'good of commerce.' Governments are dropping many legal and tax-based controls on financial risk taking, blurring the line between speculation and gambling. For example," in Australia the legal requirement of insurable interest is no longer essential to insurance contracts. The gambling industry itself has proliferated in many countries, backed by governments thirsty for a share of the take. Individual subjects are urged to gamble not only on lottery tickets and in casinos but in myriad investment markets as well. O'Malley shows that uncertainty and risk have co-existed as technologies of governance. Governance through uncertainty as 'reasonably' addressing a future that is largely unformed and incalculable is as vital as governance through risk technologies that render probable futures in statistical terms. Indeed, in the present period, uncertainty is ascendant, and it is certainly becoming more hybridized with risk. Tom Baker pursues the question of how liberalism is embedded in the practices of institutions, through an analysis of the ways in which the insurance industry addresses problems of adverse selection. Adverse selection refers to the possibility that low-risk individuals will choose not to participate in insurance pools, leaving such pools with a disproportionate percentage of high-risk individuals. For example, where health insurance is not compulsory, the young and healthy will tend to take their chances and not buy insurance, while the ageing and unhealthy will likely buy insurance and make frequent claims. Baker shows that insurers' response to adverse selection problems through risk classification reveals their ideological and moral commitments. In the name of actuarial fairness in risk classification, insurers are actually advancing a brand of liberalism that privileges both individual interests over the common good and the interests of the insurance industry itself. As a result, the proper design of a socially responsible insurance institution is inhibited. Through risk classification, higher risks are charged higher premiums and given less favourable contract conditions, such as higher deductibles and more excluded clauses. These practices exclude many from the insurance pool altogether: those who insurers will not underwrite at any price, those who cannot afford the premium set, and those who feel the premium set is not worth the price, given the unfavourable contract conditions. Through risk classification practices, insurers assign and

Introduction

221

maintain social status, not only for the insured but for others with an insurable interest. For example, the children of parents refused life or health insurance will lose socio-economic standing if the parent dies or is permanently disabled. Risk classification is entwined with market competition among insurance companies. Insurers are incessantly fine-tuning their classifications, on the one hand, to sell new products in newly defined market niches referred to in the industry as 'features creep' - and, on the other hand, to unpool and eliminate the most risky. Eliminating the worst risks reduces the cost of insuring those who remain and therefore allows the company to sell insurance at a more competitive price. Moreover, if the worst risks are picked up by other companies thirsty for premium revenues, the average risk of competitors is heightened and they are relatively disadvantaged in the market. In this respect risk classification to address adverse selection problems is a game of high risk transfer among insurance companies. Ironically, risk classification fuels adverse selection. By adding more and more gradations to the classification system, ostensibly to promote the integrity of the risk pool, insurers are actually making selections that are adverse to the broader pooling or spreading of risk that insurance is supposed to enable. They build an extension ladder of risk classification and attendant market segments, based more on the unpooling of risk than on broadly based inclusion. They also create a competitive game of snakes and ladders among insurance companies, in which risk classification is intended to produce inefficiencies for competitors by forcing the worst risks on to them. The insurance industry is constantly struggling with the quagmire of its pooling practices in order to avoid descent into a cesspool. Baker reviews four alternatives to the problem of adverse selection. First, universal insurance can be provided through a single insurer. Second, the state can legislate compulsory insurance that disallows underwriting based on risk classification. Third, in non-compulsory insurance state prohibition or limitation can still be imposed on specified risk classifications. Fourth, in voluntary insurance, products can be designed to change the incentives of low risks to keep them in the pool. In justifying the least restraint possible on their risk classification underwriting, insurers invoke the rhetoric of science as well as making moral appeals. They rely upon three fundamental justifications. First, in the absence of risk classification, low risks are unfairly forced to subsidize high risks in the pool. Second, risk classification promotes social efforts

222 Part Three: The Moral Risks of Insurance to prevent losses from occurring in the first place. Third, risk classification advances the core value of individual responsibility for risk: it makes individuals engage in preventive efforts and requires them to pay differentially for the consequences of loss. Baker illustrates each of these justifications through case studies. Subsidization of high risks by low risks is illustrated through the historical example of debates concerning age rating in fraternal life and sickness insurance. In this case the question was simple: should the young subsidize older members of the pool by paying the same premium for the same schedule of benefits? Responsibility for loss prevention is illustrated through the example of experience rating in American unemployment insurance. Unemployment insurance is provided through government programs for two reasons. First, unemployment insurance organized through government provides broader benefits to the insured, her family, and the economy than would be available through market-based insurance. Second, an obvious problem of adverse selection makes market-based unemployment insurance unattractive to the private insurance industry. If insurance is available voluntarily on a market basis alone, it will be especially attractive to employees expecting to lose their jobs, and to employers expecting to terminate employees in the near future. In the face of adverse selection problems, government unemployment insurance programs also invoke risk classifications on the basis of experience: firms with high levels of past lay-offs and dismissals are charged higher rates. In what is in effect a form of co-insurance, the employer is rewarded for preventing unemployment by organizing his business to sustain stabilized employment rolls. Individual responsibilization for risks is illustrated through the example of efforts by American life and health insurers in the 1980s to exclude battered women from insurance benefits. Insurers argued that their rating system that excluded battered women was actuarially fair. One possible line of argument here is the 'liberty interest' of low-risk individuals to join an insurance pool that is not unduly threatened by obvious high risks. In particular, if the low-risk status is somehow deserved, then the moral claim is that individuals are entitled to benefit from that status. However, if that status is not deserved - for example, because being the victim of domestic violence is not deserved and indeed is morally arbitrary - then the liberty interest claim is weak. Should people be burdened with the consequences of their undeserved high-risk status to the point where they are blamed as victims? Should

Introduction

223

people who are more fortunate be entitled to all the benefits of being low risk, whether they deserved that status or not? Baker notes that in any case, fairness arguments used by insurers in public policy debates have not been invoked to protect low-risk individuals as much as to promote the freedom of insurance companies to classify as they wish. The operable moral principle is utility, not liberty. In the utilitarian morality of the insurance industry, risk classification reflects and creates a moral vision of individual responsibility and individual protection. In consequence the insurance pool itself is seen only as a collection of individuals without responsibility to one another. It is therefore not surprising that insurers are incessantly (un)pooling in search of competitive advantage and profit. Carol A. Heimer considers insurers as moral actors. She observes that most analysts have focused on how insurers shape the behaviour of insureds through contract incentives, disciplinary mechanisms, and political policies on the administration of risk sharing. She advances the need to study parallel questions about how the behaviour of insurers is also shaped through contracts, disciplinary mechanisms, and political policies that make up insurance relationships. Heimer views insurance as a pooling of group resources to cushion the effects of individual losses. Insurers organize insurance pools as groups with a common fate and act as agents on behalf of these groups. They seek policyholders whom they can value: as a lucrative source of premiums to invest and as morally responsible individuals who will not be careless about loss or make unjustified claims that affect the integrity of the pool. Similarly, prospective policyholders look for insurers they can trust, concerned about the moral risks or bad faith of an insurer who fraudulently sells the wrong product or denies justified claims. Heimer outlines four elements of insurance company operations and their relation to insurance contract and incentive systems. First, insurers are not only risk managers but also financial institutions seeking profits through investments. The central concern in risk management is the loss ratio: premium revenue offset by claims paid. The goal is not simply to ensure some reasonable surplus of premium revenue after claims have been paid; in most types of insurance the profits lie in using the premium revenue to invest for more substantial returns. Hence a company may be able to profit more by, for example, accepting high losses if it is also obtaining high premiums to invest, compared to a low premium/low loss approach. Second, an insurer should not be viewed as a unitary actor. An insur-

224 Part Three: The Moral Risks of Insurance ance company is a complex organization, with many actors involved in myriad relationships both internally and externally. The organization has multiple utility functions and therefore will be inconsistent in its reactions. It has many units with conflicting interests, staffed by people whose needs are different from those of the organization. These units comprised of sales agents, claims adjusters, claims service providers, investment managers, reinsurance specialists, and so on - must be governed through incentives, training, and audits to build moral environments in their occupational cultures. But insurers' self-interest can also lead them to foster systematic immoral conduct in their agents. For example, commission incentive structures for sales agents can lead them to sell policyholders products that they do not need and/or which are excessively priced. Auditing systems that force claims adjusters to minimize payments can lead them to deny what the claimant deserves. The fact that the organization is not unitary can be made into an advantage for insurers. Third, insurance contracts are not two-party but rather multi-party. For example, an insurance contract encompasses third parties such as an employer who negotiates a group health plan for employees and medical specialists who supply the health care. Insurers strike bargains with these third parties, who also pose moral risks. For example, medical service providers have incentives to provide extra services and over-bill, to the detriment of the insurer's loss ratio. Insurers therefore audit medical service providers, and try to mobilize them as agents of claims minimization. In contemporary managed care systems, insurers penetrate and mediate the once-sacrosanct relationship between physician and patient. Fourth, insurance is a perpetual strategic game. All parties to insurance contracts have incentives to avoid full compliance and to cheat. These incentives, and strategic considerations about them, are pervasive at every stage: application for insurance and negotiation of contract terms; how the insured behaves with knowledge of the insurance contract and especially its terms of compensation in the event of a loss; investigation of a loss; and negotiations over compensation once a claim is accepted. As Heimer observes, while policyholders may have a propensity to pad claims, their deception seems less morally reprehensible if seen as a strategic move to deal with the fact that insurers are believed to be paying claims below what is deserved and therefore not abiding by the spirit of the contract. In effect, the insured is making de facto rate adjustments in the circumstances, just as insurers adjust their rates to compensate for the moral risks posed by the insured.

Introduction

225

In the strategic game, insurers are not only incessant classifiers - as Baker's chapter shows - but also incessant rule makers with respect to how the game should be played. They seek to discipline not only policyholders but service providers and themselves through fine-grained rules about what is covered, at what rate, requiring what evidence, and according to what procedures. In turn, insurers are disciplined by the other parties, for example, by service providers who hold clout through expertise and service infrastructure (e.g., physicians and hospitals), and by other institutions that require compulsory insurance of their clients (e.g., banks that demand mortgage insurance). The strategic interaction and rules of the game transpire principally among powerful players in interinstitutional fields with insurance at the hub, rather than simply between the insurer and insured. The body politic is another institutional arena through which insurers are judged and governed as moral actors. Following the ideas of Deborah Stone, Heimer points out that insurance provides a 'moral opportunity' to publicly debate what is owed to other members of communities. However, the fact remains that insurance involves multifaceted organizations with myriad internal and external relationships. This fact poses conundrums about who actually holds what visions of what moral opportunities, what influences them, and what consequences they face if their visions are blurred or tarnished. Moral visions espoused by insurers, governments, or other parties may help legitimate insurers and smooth their insurance relationships, but they have little value otherwise. They compete and collide. Moral visions are unevenly incorporated into organizational routines, again because the insurance company has multiple units without unified goals, reactions, or coherence. Moral visions are also technologically mediated, for example, through the risk classification formats of actuarial science, the risk communication formats of computerized data systems, and the risk procedure formats of rules concerning coverage and claims. Moral sensibilities are embedded in these technologies, and to the extent that they express moral utilitarianism they are unlikely to articulate moral vision beyond the insurer's level of satisfaction with loss ratios. Reform of insurance and other institutions for public good is only likely to occur in a moral environment of broad social solidarity. Otherwise, the structured inequality of (un)pooling will accentuate, with insurance compounding rather than offsetting the fragmentation of risk society. Richard V. Ericson and Aaron Doyle explore insurers as moral actors, through an empirical investigation of insurance fraud. They point out

226 Part Three: The Moral Risks of Insurance that the epidemiology of insurance fraud is difficult to establish because of problems of definition and evidence. Some industry sources advance a broad definition of fraud as any act or omission that results in illicit benefit payments. Examples include false statements on applications, padded or inflated claims, and fabricated claims. Since intentionality is difficult to establish, others argue that there must be factual evidence, for example, false representation of fact in documents or material evidence that belies a claim. Still others make a distinction between exaggeration of claims and fraud. In investigative practice, fraud determinations are made on the basis of some combination of intentionality, the magnitude of the loss involved, and the likely pay-off of allocating resources to an investigation. These considerations indicate that fraud is an artifact of how the insurance industry organizes to deal with it. That is, its incidence and prevalence are unknown except to the extent that it is made into a problem for the governance of insurance relationships. One can speak of 'fraud through governance' because fraud is made, managed, suppressed, and visualized on the terms and conditions established by insurers. Within the private justice system of the insurance industry, fraud is posed not as illegal behaviour defined by the moral codes of law, but as a moral risk defined by utilitarian criteria of the industry. The governing mechanism is therefore not legal control over unwanted conduct, but surveillance networks that produce knowledge useful in managing fraud as a moral risk. The concern is not order as imagined through the regularities of law, but the security of loss ratios: the predictable flow of premium revenue and claims costs that provide enough surplus to invest and make profits. The logic of loss ratio security, introduced in the previous chapters by Baker and Heimer, is crucial to understanding how insurance fraud is addressed. Loss ratio security can be provided by minimizing claims costs. Through contract conditions and other incentives, the insured can be mobilized to act as agents of loss prevention. The insurer can also establish claims management practices that control the iatrogenic tendencies of claims service providers, the 'social inflation' tendencies of claimants, and the fraudulent tendencies of both. Loss ratio security can also be provided by generating more premium revenue. Higher premiums can be generated for higher 'substandard' risks through processes of risk classification. Thus Ericson and Doyle quote the president and CEO of an insurance company as saying, 'My attitude is a good risk is one

Introduction

227

that pays high enough premium' (321). Premium revenue can also be generated through lucrative commissioned-based incentives to sales agents. Market conditions are crucial: if there is a highly competitive 'soft' market in which it is difficult to raise premiums for fear of adverse selection, the alternative is to focus on a claims crackdown including fraud control to improve loss ratio security. Within this logic of loss ratio security, fraud is but a factor in the moral risk calculus, something taken into account in deciding upon acceptable levels of loss. It is policed more vigorously if it is seen as having a substantial impact on loss ratio security. Otherwise it is tolerated, or even deemed acceptable because, ironically, it can actually contribute to loss ratio security. The governance of fraud cannot be too intrusive because it can upset insurance relationships and the status quo. Governance-at-adistance is preferred, attempts to mobilize each party involved to be selfgoverning with respect to the moral risk of fraud they pose, and to be vigilant about others who may take advantage of them. At the same time there is 'governing through fraud.' The industry regularly dramatizes the massive extent and cost of fraud as a threat to all members of its risk pools in order to justify the expansion of surveillance that serves purposes beyond the control of fraud per se. Both fraud through governing, and governing through fraud, are illustrated through analyses of sales fraud, claimant fraud, and claims services fraud. Sales fraud is a product of how agents are governed. In the traditional life insurance sales arrangement, the sales staff are commission-dependent. They therefore have a compelling incentive to sell products that yield higher premiums for the company and higher commissions for themselves, even though those products may be inappropriate and too costly to consumers. For example, expensive investment-related products may be sold to low-income families who would derive greatest advantage from simple term life policies that carry only death benefits. There are similar incentives to 'twist' or 'churn' existing policyholders into more expensive replacement policies. These practices are accentuated in competitive soft markets, where the 'feature creep' of increasingly specialized risk classifications and market segments drives agents to sell new policies to existing policyholders that are too expensive and do not meet the policyholders' needs. The industry has difficulty transcending this systemic fraud because the life insurance product is not consumer driven: it is expensive, it addresses remote future possibilities people prefer not to dwell upon, and it does not have the commodity

228 Part Three: The Moral Risks of Insurance appeal of material goods. Life insurance must therefore be sold proactively and aggressively through commission incentives and a sales culture motivated to missell. Faced with recurrent sales fraud scandals, life insurers must respond through new governance mechanisms. Governing through sales fraud often includes new training programs. However, the sales system would not work if rigorous training were required because the agent is a commissioned seller in a buyer/seller relationship, not a well-trained professional in a client/professional relationship. She is paid for sales, not for advice, and is herself churned out of a job if she fails to make sales. Insurers also use fraud scandals as a justification for 'rightsizing' their sales force and shifting to alternative distribution channels. They also introduce new means of governing through technology, for example, computer-based sales packages that control and monitor each stage of the sales process, or voice recordings of call centre sales, which have the same effect. Claimant fraud poses four problems in governance. First, there are evidence problems, especially in fields such as personal injuries, where diagnoses of medical problems are easy to imagine but impossible to prove. Second, even when it may be possible to generate evidence of fraud, it is often more expedient not to bother and simply to make a 'nuisance payment.' Third, there is a concern with policyholder relationship management to protect the premium end of loss ratio security. The fraudulent claimant will still be paying premiums in the future - indeed at higher rates, because of the claim - and therefore may pay off the loss and then some over time. The policyholder may also be a client in the cross-selling of other insurance and financial products. As Heimer notes in her chapter, some insurers operate on a high premiums/high claims basis to upmarket customers. They do not want to bite the hand that feeds, even if their own hand is nibbled at. There is also the lingering fear of making a false allegation of fraud, which, given the ambiguous nature of evidence, is an omnipresent concern. Fourth, also in keeping with Heimer's analysis, broader social and economic relations with third parties make allegations of fraud difficult. For example, workers' compensation insurers face sensitive labour and management issues that limit their capacity to police fraud too heavily. Insurers govern through claims fraud. Since they know only too well that fraud is a product of their own governance practices, insurance operatives suspect every claimant. This presumption of fraud entails a strong blaming-the-victim approach, to the point where the 'squeeze' is

Introduction 229 put on many claimants. The strategic game to which Heimer refers is set in play. On the insurer's side, there may be intensification through additional surveillance systems and new regimes for claims management. Ericson and Doyle examine new 'work-hardening' regimes for disability claimants that have arisen in markets where disability premium revenues have been restricted by competition and/or government controls. Ericson and Doyle also use disability insurance practices to illustrate the dynamics of claims services fraud. Fraud by disability claimants is entwined with the professional practices of lawyers and health service providers, who build claims for mutual benefits. The definition of disability is itself an insurance construct negotiated in the pragmatic contexts of claims settlement. Like illness more generally, it involves a moral battlefield fought through risk classification schemes, figurative language, and sanctions. For insurers, the goal is to limit an iatrogenic system where service providers build claims through questionable diagnoses, unnecessary treatments, and over-billing. Beyond the situational interest of particular selfish people, systemic processes of professional jurisdiction, privilege, and material benefits are involved. The result can be what Ivan Illich described as 'disabling professions.' Insurers respond by governing through claims services fraud. Fraud is used as a justification to 'prevent medicine' for fear that too much surplus value will pass into the hands of medical service providers. Preventive measures are multiple and interconnected. First, authorized service providers are subject to limitations, often through designated assessment centres. Second, education programs for medical practitioners are established to train them to be self-governing regarding the iatrogenic effects of their diagnoses and treatments. Third, medical practitioners are constituted as claims investigators on behalf of insurers, with the emphasis on fraud by both insureds and their colleagues. Fourth, medical services compensation levels and incentive structures are altered to reduce costs. Fifth, the classifications and reporting forms are altered to effect greater standardization and auditing of diagnoses and treatments. Sixth, related dataveillance systems are developed, for example, regarding billing practices. Seventh, professional liability insurance conditions are altered to further discipline service providers into managed care system efficiencies. Ericson and Doyle document the differential response to insurance fraud, which is tied to its relation to loss ratio security. The onus is placed on each and every participant in insurance relationships to contribute to

230 Part Three: The Moral Risks of Insurance the surveillance network as it fosters loss ratio security. This onus means that there are less benefits for those who fail to contribute fully and who in that sense pose a moral risk. Indeed, someone who is not a full contributor may be treated as a 'fraud' even if he or she has committed no fraudulent act. Meanwhile, those who do commit fraud - but who are otherwise deemed worthy contributors within the moral utilitarianism of the insurance institution - continue as valued participants in the system.

9 Moral Uncertainties: Contract Law and Distinctions between Speculation, Gambling, and Insurance PAT OMALLEY

While the risk society thesis represents governance as focused on the task of rendering the future calculable with ever more precision, a current politics of uncertainty has challenged technologies of risk in the very heartland of the neo-liberal economy. Perhaps best expressed by gurus such as Tom Peters, it suggests that the statistically based modelling of economics is profoundly in error, not least because the planned economy of the mid-twentieth century has given way to a global economy in which a form of speculative calculation that does not lend itself to statistical calculation has become the norm. 'Predictability is a thing of the past,' is the claim: economists and their precise models of rational calculability have become the nemesis of this vision of the entrepreneurial world. 'The entrepreneurs sustain the world. In their careers there is little of the optimizing calculation, nothing of the delicate balance of markets ... The prevailing theory of capitalism suffers from one central and disabling flaw: a profound distrust and incomprehension of capitalists. With its circular flows of purchasing power, its invisible handed markets, its intricate plays of goods and moneys, all modern economics, in fact, resembles a vast mathematical drama, on an elaborate stage of theory, without a protagonist to animate the play' (Peters 1987:245). This current centring of uncertainty as the technique of successful capital accumulation in many ways reproduces the theories of nineteenth-century economists, such as Emery and von Thunen, who argued that profit emerges out of the creativity of entrepreneurs. In words that foreshadow Tom Peters, Emery (1896) argues that '[s] peculation has

232 Pat O'Malley become an increasing factor in the economic world without receiving a corresponding place in economic science ... yet it is one of the chief directive forces in trade and industry. But treatises in the English language on economic theory and conditions have given very little space to this influence which is fundamental in practice.' For von Thunen and others 'the rewards of the entrepreneur ... [are] the returns for incurring those risks which no insurance company will cover because they are unpredictable, and in this sense, the entrepreneur is necessarily an "inventor and explorer in his field."' (Blaug 1988: 461). For these thinkers, risk technologies assume that the world will repeat itself in statistically calculable and thus predictable ways, while entrepreneurs, on the other hand, are innovators creating a new and uncertain future. Their innovating activities are described as 'inspiration,' 'foresight,' and 'acumen': the discourse less of a statistical science than of a practical aesthetics or common sense (see also Knight 1921; Reddy 1996).1 It is a discourse that sat well with the nineteenth-century vision of the laissez-faire economy, in which individualistic creativity could be subjected to expert state regulation only at the peril of its extinction. It sits less well with the current sociological vision of a risk society, and with earlier models of interventionist state governance of the economy that sought to tame the 'irrationality' of the market place and the speculator. After the end of the nineteenth century, the discourse of uncertainty— although never extinguished — was increasingly submerged under the discourses of such positive human sciences as economics and econometrics, where it was linked with the 'planned economy' (Reddy 1996). The submerging of speculation and economic uncertainty likewise appears in sociology, notably in Weber's vision of modern capitalism as the 'rationally calculative' approach to profit and accumulation. It is discernible in Weber's discussions of the Protestant ethic and its formative role in 'Western' capitalism; in his views on the nexus between capitalism, formal rational law, and legal domination; and in his accounts of the mutually supportive characteristics of capitalism and bureaucracy (Weber 1958, 1965). In each case, stress is placed upon the structuring of this-worldly activity in such a way that the accumulation and increase of capital is pursued through the precisely or maximally calculated structuring of profitable opportunities, ideally linked with stable and axiomatic - and thus maximally predictable - legal and administrative techniques. This rational activity is imagined especially as the polar opposite of medieval economic adventurism reliant on

Contract Law and Distinctions between Speculation, Gambling 233 fortuna - a view repeated by more recent risk theorists such as Giddens (1990). Weber thus creates a binary that excludes a vision of 'uncertain' capitalist activity, a form of refined common sense or reasonableness based upon inspiration and tempered by pragmatic rules of thumb. In all of this, Weber's legacy mirrors the assumptions of contemporaneous economic theorists - for built into his theses is a realist assumption that, specifically vis-a-vis the 'adventurism' of previous eras, rationally calculating and logically predictive activity accounts for the material success of modern capital accumulation.2 This modernist view overtook the 'uncertain' vision of those proponents of enterprise in the late nineteenth century who regarded the economic benefits of capital as being based in speculation. It has in turn been challenged by the postmodern economics of Peters and the new champions of entrepreneurship. It may be no coincidence that Weber wrote at a time characterized by acute concerns over stock market speculation and its nexus with gambling. As will be seen, the German state at the turn of the twentieth century even sought to regulate speculation on the stock market, turning it into a form of licensed activity practised only by those with demonstrated expertise in market prediction. In this period, across Britain, Europe, and America, much moral, legal, and theoretical effort was made to distinguish between gambling and speculation - a distinction that had been effected a century or more earlier with respect to insurance and gambling. This 'crisis' was founded in the invention and development of contract law and of debates around it during the nineteenth century, which legally legitimated speculation and established a governmental discourse of uncertainty (O'Malley 2000). Rather than simply following liberal political and economic theory, the law of contract in Britain, for example, became a site in its own right for the development of 'liberal' ideas of uncertainty and economics. Much of this appears to have been formulated out of legal interpretations of 'normal' business relations. In the case of the distinction between gambling and speculation - the principal subject of this chapter - legal formulations governing economic activity were grounded partly in lawyers' understandings of the 'good of the economy' and partly in economic theories and justifications developed by stock exchange officials and their allies. Liberalism, risk, and uncertainty, from this perspective, may be imagined as individually and collectively possessing a number of distinct but interrelated genealogies. This essay continues a genealogy of uncertainty through the exploration of debates in and around the law of contract begun in a previous

234 Pat O'Malley paper (O'Malley 2000). In the first part, I will briefly outline the ways in which uncertainty appears in contract law as a distinct technology from that of risk, endorsing the speculative 'creativity' of entrepreneurs, but also imposing a discipline of 'reasonable foreseeability' upon them. The second part will begin to examine how this endorsement of uncertainty and speculation created difficulties for the eighteenth-century legal distinction between gambling and business speculation, and how a variety of 'economic' knowledges were generated and mobilized in order to attack and sustain this distinction. The final section will explore one of the unforeseen 'interventionist' assemblages built out of (or upon) this platform - the attempt to create regulatory expertise in uncertainty. The Governmental Foundations of Contract3 In the constitution of such liberal subjects as Adam Smith's homo economicus, or Bentham's performer of the felicity calculus, what renders the subject 'rational,' and thus eligible to participate fully in liberal governance, is the cognitive process of calculating the future consequences of courses of action. This capacity for calculative 'rationality' renders such liberal subjects 'free' - that is, able to govern themselves while 'irrationality' vitiates freedom. Nevertheless, while assumed to be 'logical' and 'coherent,' the process of rationality is invisible. Its operation is known by its effects - the fact that the behaviours of rational subjects are predictable, while those of the irrational are not. In short, calculative rationality and behavioural predictability are locked together in such liberal subjects. These formulations approximate far more closely to the 'risk' modality of calculation-as-prediction than to the less precisely or more qualitatively calculable modality of uncertainty. Bentham's and Smith's subjects are calculators of the cost-benefit kind: the model of the calculus is mathematical in form, even if the values with which it deals are subjective. For Bentham of course, a critical part of the exercise of governing liberally was to increase the certainty of outcomes (such as the certainty of capture and punishment, the certainty that sentences will marginally exceed the benefits of crime, and so on) in order to render the predictability of the felicity calculus more transparent and precise. These liberal subjects, whether Smith's entrepreneurs or Bentham's criminals, appear as subjects of 'risk' considered quite narrowly as an 'arithmetic' calculative rationality. It is difficult to see here any foundation or even justification for a genealogy of uncertainty distinct from that of risk. However, there is a

Contract Law and Distinctions between Speculation, Gambling 235 second genealogy that runs alongside and braids into it. It is to be found not so much in the rationalist writings of early liberal philosophers, political theorists, and economists, but rather in the development of that branch of common law that is inseparably linked to the growth of liberalism and the formation of its subjects: the law of contract. In this genealogy we can trace the development of certain ways of envisioning problems for government, certain governmental categories linked to these problems, and some techniques of the self for governing them that are often much less precise and calculable in their nature than is suggested by terms such as 'rational calculation' (see generally O'Malley 2000). One of the key elements of the genealogy of the law of contract is the formation of subjects who are expected to govern an uncertain, imprecisely foreseeable future. Prior to the nineteenth century, the law of exchange (for it cannot accurately be referred to as 'contract') was focused on the exchange of titles to material things or services, imagined as a more or less instantaneous process. This focus shaped its governing categories in specific ways. It could recognize 'debt,' for example, where some good or service had been provided, and it could recognize situations where payment had been made but the service or goods not provided. Its remedy in both cases often was specific performance - the proper completion of the exchange agreed to and already partly completed. As well, it incorporated many elements now understood in terms of the idea of the 'moral economy.' For example, it assumed that there is a fair or 'just' price for commodities. The courts could thus examine exchanges to see how far they conformed with this price and they would overturn contracts that departed appreciably from the norm. As well, the law implied that a just price was warranty upon goods and services, so that the agreement could be annulled if the goods were not sound. In these senses the early eighteenth-century law sought to govern certainty. The development of the law of contract, slowly during the eighteenth century and much more rapidly in the nineteenth century, changed much of this. The law shifted in the direction of recognizing exchanges that were understood in terms of the free-willed assessments of the contracting parties as to the benefits bestowed by the exchange. This transition was reflected in the erosion of the courts' investigation of the conformity of contracts to standards such as a 'just price.' Increasingly, it came to be assumed that while the parties might be wrong in their assessments of benefit - for example, that one party had purchased at well above the market value of a commodity - their error was a reflection

236 Pat O'Malley of the different, subjective, evaluations of the parties. Among such 'subjective' matters now more or less placed out of the courts' reach was the relative ability of different parties to assess the direction or performance of the market in pricing goods. Between about 1780 and 1850, then, the law of contract came to reflect a market in which prices are not fixed by tradition or fiat, but rather by what came to be understood as 'supply and demand.' And since supply and demand in turn reflected the subjective desires of those entering the market place, the value of commodities was 'naturally' prone to fluctuations that could not be predicted precisely. We thus witness the development of a legal subject who is expected to possess foresight that might well be based on a thorough knowledge of the market, but to assess the future only with a form of 'reasonable' or common-sense judgment. During the nineteenth century, this expectation was given legal expression in the idea of 'reasonable foreseeability.' Reasonable foreseeability came to be understood as 'foresight in relation to a practical decision,' foresight 'such that in all the circumstances a reasonable man would adopt or refrain from a particular course of action,' and assessment of what is 'foreseeable in the practical sense' (Hart and Honore 1985:263). Here we witness die governmental formation of a technique of 'uncertainty,' a legally required skill of the subjects of an emerging law of markets. Such subjects are also understood to inhabit an economic environment characterized by uncertainty, in which markets - and the relevant future - are mobile and only loosely predictable rather than statistically 'calculable.' Early in the nineteenth century, this understanding was registered in the development of new legal categories - notably that of 'expectation interest.' In effect, an expectation interest is the right of a subject to the speculative profits that flow from a legally binding exchange. In the event of a breach of contract, a plaintiff could now sue not primarily for die completion of the exchange, as under old law. Indeed, this 'specific performance' of the contract might be no remedy at all, as the opportunity for profit may have disappeared by the time the contract was enforced by the court. Rather, the plaintiffs could sue for the hypothetical profit that would have been realized had the exchange occurred - even though the legally accepted uncertainty of the market often means that this estimation of lost profit was recognized as only an 'informed guess' by the court (Kercher and Noone 1990; Kercher 1996). As this might suggest, the emergent law reflected not so much the elegant theorizations of liberal economists and philosophers such as

Contract Law and Distinctions between Speculation, Gambling 237 Smith, Ricardo, and Bentham, founded in hard-edged visions of rational calculability, but interpretations by lawyers and judges of the nature of emergent contractual relations in the 'reality' of the economy (Hamburger 1984; Simpson 1979). By the middle of the nineteenth century, uncertainty - a pragmatic form of reasonably estimating the future - was thus ensconced in the legal heartland of liberalism. While many of the profound implications of this development cannot be dealt with here (see O'Malley 2000), it will not have escaped notice that the law had come to embrace, or at least to recognize and enforce, speculative contracts, whereas in the eighteenth-century law of exchange speculation was problematic. Prior to the nineteenth century, speculative contracts were often overturned by the courts - especially where a party bought low and soon after sold high - since this implied that one or another of the contracting parties must have been cheated of the just price. Whatever other concerns this change created at the time, it opened up a space in law for an array of new dilemmas, ambiguities, and contests. To begin with, while speculation was being licensed in the market, gaming was already subject to moral and legal restriction, and attempts to legally distinguish between gambling and insurance had already been set in train. More than ever, distinguishing between gambling and legitimate speculation, and setting the 'proper limits' to speculation, became a problem for law, and for regulatory concerns in such market institutions as the stock markets. Gaining and the 'Good of Commerce'

Viewed from the platform of the twenty-first century, the 'problem' of gambling is coloured by the nineteenth-century moral panics over the lack of diligence and thrift among the working class and the need to abandon the corrosive temptations of a get-rich quick mentality. It was in this environment that the Gaming Act of 1845 (8 & 9 Viet c. 109) rendered all wagering and gaming contracts unenforceable at law.4 But in the seventeenth and eighteenth centuries, gambling was viewed as problematic even more because of its perceived threats to good order, because of a need to protect the fortunes of the wealthy against the impact of gambling, and because of its 'unproductive' relationship to commerce.5 The present analysis focuses on the last of these concerns. Statutes specified that the law would not provide guarantee or remedy with respect to securities related to games: that is, 'all notes, bills, bonds, judgements, mortgages or other securities or conveyances whatsoever'

238 Pat O'Malley that were transferred in relation to gaming debts (1710, 5 & 6 Wm IV., c. 41). Thus in the foundational Act against Deceitful, Disorderly and Excessive Gaming (1664, 16 Car. II c. 7, s. 3), the distinction was made between contracts and debts on events other than games, which were enforceable at law, and those associated with 'cards, dice, tables [i.e., backgammon], tennis, bowls, skittles, shovel-board, cock-fightings, horse races, dog matches, foot races, or other games or pastimes, game or games whatsoever.' The operation of the statute was judged not to be limited to games of chance, for cricket and any game, whether of chance or skill, came under its concern (Jeffreys v. Walter (1748), 1 Wils KB 220). They were thus not associated with any condemnation of reliance upon fortuna.6 Rather, the nature of games was at issue. The importance of the 'game,' rather than the calculability or predictability of the investment, is to be seen precisely in the fact that legitimate insurance at this time could take the form of betting without incurring the same legal disability. This included insurance contracts taken out on the life of another, or in the nature of a tontine in which the survivor takes all, or even of insurance on an event - such as the safe completion of a voyage in which one had no investment. All such exchanges remained enforceable at law until late in the eighteenth century. However, by the mid-eighteenth century certain species of marine insurance were already coming under the scrutiny of the state. Under the Statute of 1745 (19 Geo II., c. 37) marine insurance contracts in which the insurer had no 'interest' were rendered null and void. The reasons for this were several, and certainly included concerns to prevent the fraudulent destruction of ships and their cargoes. However, as Blackstone - reflecting earlier common law decisions (Sadler's Coy v. Baldock (1743), 2 Atk. 556; Godin v. London Assurance Coy. (1758), 1 Burr. 492) - indicates, the legislation was concerned with the distinction between games and commerce: 'Being much for the benefit and extension of trade by distributing the loss or gain among a number of adventurers, they [i.e., insurance policies] are gready encouraged and protected both by common law and acts of parliament. But as a practice had obtained of insuring large sums without having any property on board which were called insurance interest or no interest... which were a species of gaming and without any advantage to commerce and were denominated wagering policies... (such policies are rendered null and void)' (Blackstone 1769; emphasis added). In addressing species of economic investment that were not in themselves associated with games, but which were taken to be a form of

Contract Law and Distinctions between Speculation, Gambling 239 gaming, the Act simultaneously extended the definition of gaming and thus of the kinds of 'contracts' that were not enforceable in law. But the legislation and the common law cases from which it developed also made clear that gaming or wagering - in which this species of insurance was now to be included - was itself regarded as in some way an unproductive or commercially valueless activity ('without any advantage to commerce') (Windeyer 1929). The Act required that the beneficiary of a marine insurance had to have an insurable - that is fiscal or equivalent - interest that could be indemnified. As is well known, this was to be the effect of the more general Gaming Act of 1774 (14 Geo. Ill c. 48), which finally provided the general distinction between gambling and insurance on the basis of the absence or presence of an 'insurable interest': that is, that the policyholder will suffer a compensable loss in the case of a specified event occurring (Blackburn J. in Wilson v. Jones (1867), L.R. 2 Ex ISO).7 What becomes clear in this process is that for the law of commerce the critical issue is not so much the question of the moral rectitude, or otherwise, of wagering on games. Rather, the central issue is an assumption that the act of exchange must include some element of material value or title to value, which in itself constitutes an essential element of advantage to commerce, and that value becomes the critical distinction between gaming and enforceable economic exchanges, including insurance. Based on this view it was reasoned that even if the security given by a to b in a gaming contract was then passed to a third party, c, it still had no value and could not be enforced at law. However, if in that exchange consideration (e.g., money) was given by cto b for the security, the third party did have a right of recovery from b: the 'worthless' property of b has achieved economic or commercial value by the fact of c's investment in it, regardless of whether cknew that the security was based on a gaming debt (Fitch v. Jones (1855), 5 E&B 238) .8 Once consideration (value) is given that is not itself per medium of a gaming contract, any disability is removed: it is a matter imagined as one of material commerce, not a question of the intentions of actors. Like the vision of the 'entrepreneur' and his or her capacity of 'reasonable foresight' that was emerging in contract law at this time, such thinking can only loosely be related to the theoretical ideas of liberal economists. As in other areas of contract law after 1800, we search in vain for any judicial reference to the writings and theories of Smith or Bentham. As far as we can tell from reading these cases, we confront yet another strand of the braided genealogy of modern liberalism and the

240 Pat O'Malley technology of uncertainty, one that is critical to the distinction of gambling and insurance, and which is to become still more critical - and perhaps still more unstable - in relation to the distinction between legitimate business speculation and gambling. It is not that these strands exist in isolation of each other, for it would be implausible to imagine a judiciary unaware of theoretical developments in liberalism. Rather, the judiciary interpret economic activity through their own lens, and in terms of legal theories of economic benefit that have their roots in decisions laid down well before the systematic formulations of liberal theorists (see O'Malley 2000). 'Difference' Contracts: Gambling and Exchange As suggested, in such thinking there is a focus on the material transfer of value that distinguishes between that which is a legally unenforceable gamble and enforceable insurance contracts. This fitted well with the eighteenth-century law of exchange, under which the court's business was simply to enforce the transfer of title where consideration had been paid, or to enforce the debt where actual transfer of title had occurred. But the development of contract law in the nineteenth century rendered enforceable those exchanges which were purely promissory in nature, where no actual exchange of value had occurred, and in which the key issue was recognized as the profit (expectation interest and expectation damages) rather than the exchange itself (enforceable as specific performance, but not 'as of right'). If the contractually enforceable aim of the exchange is profit, and this goal could legitimately be the only interest in acquiring and selling the commodities involved, why should the exchange of material commodities be required to occur at all? Clearly, however, this view would collide with Blackstone's principle and its underlying materialist theory of exchange and the good of commerce. The courts, conceivably, could have abandoned the materialist theory, just as they had abandoned the anti-speculative principle of a 'just price' elsewhere in contract law. Instead, they seemingly retained the old theory as developed in relation to insurance, but developed it in turn by focusing on the intentions of the contracting parties. In an action involving the transfer of shares (Grizewold v. Blane (1851), 11 C.B. 526), the court determined that if the jury believed that neither party intended to purchase or sell the shares, the contract constituted a gambling contract and was thus void. Hence in Ashton v. Dakin (1859), 7 W.R. 385) it was argued without any apparent irony that 'If no actual purchase had been

Contract Law and Distinctions between Speculation, Gambling 241 made, but the object merely was to speculate, without buying or selling, upon the price on a future day that would be by way of gaming and wagering. But I see no objection to a man, confiding in his own judgement, saying to a broker "Will you buy corn (or anything else) for me and let the bargain be so as to the day of payment that you may have an opportunity of re-selling it for me by such a day, when I expect the market will have risen, and then you will pay the seller for me with the money you receive from the purchaser, and I shall receive the gain from you, if any, or pay you the loss" ... That is not gaming' (per Pollock CB) .9 Nevertheless, to all appearances, the focus on the material exchange of commodities was now somewhat removed from any economic theorization of profit or economic benefit being produced in the act of exchange per se, and even from Blackstone's materialist theory. Seemingly, the required exchange that distinguished commerce from gambling had become one only of formal tide, an exchange occurring entirely within the discourse of law rather than in the movement of materials. Within a few years, developments with respect to the contractual 'law of uncertainty' brought further changes. The court in Hadley v. Baxendale ((1854), 156 E.R. 145) determined that in the event that one party to the contract stood to gain or lose in some unusual fashion from the contract, unknown to the other, or as the result of some possibility that the other party 'reasonably' did not consider as bearing on the contract, the award of damages would be restricted to what the court decided was reasonably foreseeable. In short, working on the principle of reasonable foreseeability, the court determined that no-one should be required to take into account that which was not knowable through its own constitution of 'normal' levels of business competence. Building upon this model of uncertain subjects, in Barry v. Croskey ((1861) 2 J.&H. 1) the court determined that in order for a transaction to be considered a gambling contract, a mutual intention to wager must be demonstrated. Consequently, if - in the contemplation of one of the parties - the exchange represents a bona fide sale, then the fact that the other party never intended to complete the exchange will not vitiate the contract.10 Subtly, then, the law had shifted as two streams of legal theorizing about uncertainty came together. Whereas in the eighteenth century gaming was defined in terms of criteria concerning the exchange of material commodities, the intent of the subjects, and not any objective impact on 'the economy,' had become the deciding factor. A gamble in the case of 'difference' contracts, at least, was now defined by the intent to gamble.11 Thus it became possible for the court to claim that' [i] t must be

242 Pat O'Malley made clear that speculation does not necessarily involve a contract by way of a wager, and to constitute such a contract a common intention to wager is essential. A contract may be speculative and yet a perfecdy good contract. It would be disastrous to hold otherwise' (See v. Cohen (1923) 33 C.L.R. 182 per IsaacsJ., emphasis added). 'Uncertainty Spreading' and Economic Benefit The prominence of the 'intent to gamble' is indicative of the ongoing concern with speculation and its contribution to economic activity, notwithstanding that in law the focus was on the question of intent to deliver the contracted goods. At least in principle, the law focused upon the idea diat one party could, hypothetically, enforce a 'genuine' market exchange. Yet it is important to recognize that in contract law questions as to the moral impact of gambling and speculation on the 'character' of the parties to the exchange were little considered. The focus was still primarily upon the imagined effect on the economy, and the notion that 'empty' exchanges contributed nothing to commerce. While it might be considered unlikely that this focus would remain unaffected by the ascendancy of concerns over the moral character of the working class during the latter half of the century, its influence on contract law can neither be denied nor taken for granted.12 In particular, it must be recognized that there was no assumption that the 'working poor' should engage in entrepreneurial activities themselves, least of all that they should or could enter into share transactions on the stock exchanges. Entrepreneurs constituted a separate class whose laudable risk taking benefited all, absent 'difference' contracts and fraud. While positive examples of the 'self-made man' occupied a sacred place in laissez-faire mythology, the avenues of investment for working people were always economically ultra-conservative ventures such as the savings bank and the life insurance policy (O'Malley 1999). In the United States, indeed, it was not until the 1870s that life insurance for the masses began to appear unambiguously as an acceptable 'investment' rather than a form of gambling (Zelizer 1979).13 However, it is certainly the case that in the period after about 1870, in a fashion especially acute in the United States (in Britain the matter was pragmatically resolved by the Royal Commission of 1878), new concerns began to emerge about previously accepted practices of 'short selling' and related forms of transactions in the stock and commodity exchanges. In part, these concerns were related to the increased volume of such activity (Cowing 1965), but for

Contract Law and Distinctions between Speculation, Gambling

243

whatever underlying reason, this period was marked by an anxiety about the morality of speculation. In this emergent moral politics, condemnation of gambling was prominent. However, speculation was also defended as 'good for the economy.' Futures trading has considerable lineage in contract law. Its most frequent form had consisted of buying cargoes 'on forward delivery' or on a 'to arrive' basis - that is, the party purchases title well in advance of taking possession and accepts all risks associated with destruction, loss, or damage to the property en route or associated with fluctuations in the market price of the goods concerned. Futures markets emerged as an extension of such practices, in which - for example - title to a quantity of some commodity (such as grain) is purchased in advance of the material object of the exchange being produced. The expectation, of course, is that prices will have risen above the purchase price by the date of delivery. Such futures may be resold to a third party, also prior to any party taking possession or even producing the material commodity. Related activities, such as 'short selling,' involve parties selling title to a commodity which they do not own, with the intention of purchasing it before delivery is due, and after an anticipated price fall has occurred in the market. Significantly, in law the key difference between a 'normal' transaction (in which parties buy low and sell high) and short selling (in which parties sell high and then buy low) or futures transactions (in which an unknown future value is purchased) lies not in the intent to speculate, nor in the relative uncertainty or predictability of the speculation. Rather, it lies in the fact that the transactions rely upon the distancing of acquisition of title from possession of the commodity. As we have seen, by the 1860s, only the intent to deliver in futures contracts distinguished them from gambling. The possibility of actual exchange (even though driven by speculation on prices) was regarded as economically beneficial, whereas mere speculation on prices alone appeared deleterious (U.S. Congress 1892a: 235-47). Yet 'intent,' evidently, is steps removed from the actual delivery of material goods, and many regarded it as an empty legal fiction that allowed speculation akin to gambling. Increasingly, the opponents of futures trading railed against the courts' seeming acceptance of 'fictitious dealing,' which was regarded as a travesty of the legal theory of the material economy. As one representative argued, 'Those who deal in "options" and "futures" contracts, which is mere gambling, no matter what less offensive name such transactions be designated, neither add to the supply nor increase the demand for consumption, nor do they accom-

244 Pat O'Malley plish any useful purpose by their calling, but on the contrary they speculate in fictitious products' (quoted by Brenner 1990: 108).14 Why, in this light, and given the ghostly persistence of the material theory of exchange in law, should these forms of activity not be regarded as gambling? From the mid-1880s, pressure to curb or prohibit futures speculation increased, and it continued for the next half century with varying degrees of intensity.15 Particularly in the debates and analyses surrounding the 'Hatch bill' (U.S. Congress 1892a, 1892b) and several reports to government over the subsequent twenty years (such as the Hughes Report (Cowing 1965) and the Senate Report on Stock Exchange Regulation (U.S. Senate 1914)), those seeking to defend futures trading largely brokers and other representatives of the stock exchanges - articulated an array of largely successful claims against attacks founded on accusations of immoral gambling, economic sabotage, and parasitism on primary producers and small stockholders. Futures speculators were frequently depicted as morally corrupt and physically degenerate gamblers, especially in the populist press. They were accused of selling commodities they did not own, thereby creating surplus supply and driving prices down so that they could then buy low. Such short selling was said to make prices unstable, to interfere with the normal processes of supply and demand, and most significantly, along with all forms of futures trading, to exploit the commodity producers, and particularly farmers (Cowing 1965). While some opponents sought to ban futures trading outright, the most influential critics were those who advocated the imposition of a tax that would ensure that such trading would be carried out only by those who owned and possessed the commodity at the time of the transaction, thus resolving the practical ambiguities opened up by the issue of intent to deliver.16 In their own light at least, they sought on both moral and economic grounds to address the problem of 'gambling on the stock exchange' by effectively restoring the eighteenthcentury principle of material exchange rather than the 'fiction' of intent it had been transformed into.17 In practice, the dice were heavily loaded against these critics. The emphasis in court on the intent to deliver, coupled with the fact that the onus of proof of no intent to exchange was on the challenger, meant that few attempts at litigation went against the speculators (Lurie 1979: 61). As Lurie points out, 'it was very easy for a commission merchant and broker to "intend" delivery, make a contract, change their minds, and settle between themselves.' However, advocates of speculation did not wish simply to rely on such defences. Rather, they sought to establish that any attempt to regulate would undo major benefits bestowed upon the

Contract Law and Distinctions between Speculation, Gambling 245 economy, and upon various participants in the market, by futures speculation - benefits that were created even where there was no intent to deliver or accept delivery. In a strong sense, they sought to play the abstract principle underlying the distinction between gambling contracts and business contracts - the 'good of commerce' test- against Blackstone's materialist interpretation of it. In the initial debates over primary production and futures markets, three major arguments were raised in defence of futures trading. Income and Cost Spreading For primary producers, a continuing problem has been that crops do not spread their maturation evenly through the year. Harvests, shearing, and culling come usually once or a few times per year, whereas the costs associated with production spread themselves more evenly. Primary producers thus bear considerable carrying costs - the transaction costs created by the length of the production process. Futures trading was said to spread these costs more efficiently. As the commodity was sold in advance, the farmer was able to realize and use capital otherwise tied up in the production process. This meant that it was possible for producers to spread their expenditures over the year without incurring interest charges, and thus to create a more efficient distribution of market exchanges (for equipment, plant, etc.) over time. Suppliers, too, experienced these 'smoothing' benefits, as did others down the line. The costs of effecting such increased efficiencies were regarded as being borne by the speculator. In this way, claims were made that speculators performed an economic and moral function, making the same speculative choice as any other investor on the stock market, with no greater or lesser degree of uncertainty attaching to their decisions. While in some accounts they were made out to be experts in price assessment and market performance, overall they appeared as sometimes winning sometimes losing but in the process translating the carrying costs of producers into their own burden of risk. Such accounts depicted the speculator as a moral agent acting on behalf of the economy and of producers (e.g., U.S. Congress 1892a: 40-1, 156-81; 1892b: 75-85). Risk Spreading Primary producers experience poor harvests and gluts, unanticipated competition from other markets, unanticipated costs caused by the weather, and so on. Futures markets were said to allow producers to

246 Pat O'Malley spread these risks, passing them on in some degree to the speculators as they saw fit. Commodity producers could take the same speculative chances as any other participant in the market, choosing to take the certain prices on offer against the possibility of lower prices later, but running the risk that prices would rise. They were thus regarded as using futures markets as a form risk management, for they could spread risk by prior selling of part of their product according to their estimates of market opportunities in the uncertain future. Once again, the speculators appeared as providing a beneficial opportunity to producers, effectively offering the benefits of a highly flexible crop insurance scheme (e.g., U.S. Congress 1892a: 71-9; 1892b: 78-85). As Cowing (1965: 1213) has argued, such practices were akin to 'hedging' bets or investments, but '[b]y using the respectable word "insurance" instead of "hedging" brokers felt they strengthened their case; in fact this whole line of argument made it possible to view the speculator not only as the incidental benefactor of the farmer ... but also as an indispensable cog in a system of agricultural insurance.' There is, perhaps, no little irony in all of this. Whereas, at this time, the stability of insurance was generated out of the risk-based technologies of actuarialism, and this was deployed in contemporaneous distinctions between insurance and gambling (O'Malley 1999), it was clear to all that the processes at work here were the uncertain techniques of the market place. The risk-spreading technology of insurance was being applied to a domain understood at the time to be governed by forces incapable themselves of being tamed by actuarial techniques (O'Malley 2000): speculation was being morally and technically reinvented as a form of 'uncertainty spreading,' with the speculators providing this service at their own expense. Stabilization

In both of these 'spreading' techniques, futures trading was claimed to stabilize markets. To begin with, the spreading of purchases over time evened out the price fluctuations driven by short-term considerations such as poor harvests or increased costs. As well, because futures markets are national, local variations are likewise reduced. Price fluctuations both over time and space would thus tend to be ironed out. Indeed, it was argued that the more exchanges that occurred, the greater this stabilizing-spreading effect would tend to be, for the overall fluctuation of prices would be divided up into more and more transactions thus reducing the risk to any party (e.g., U.S. Congress 1892a: 156-68).

Contract Law and Distinctions between Speculation, Gambling 247 The aftermath of these debates was marked by the continued refusal of the courts to demand delivery, and to insist that wording indicating intent-to-deliver rendered futures contracts enforceable at law.18 While the distinction between difference contracts and futures contracts was not eroded, the invention of 'uncertainty spreading' as a critical economic function performed by futures speculation represents a major step in the genealogy of uncertainty. It is now not simply that speculation is regarded as a source of benefit, for this had been ensconced in law from the invention of the executory contract and expectation damages. Rather, increasingly the debates over common law generated new understandings that loosened the nexus of 'benefit to commerce' with the existence of material exchange. The view, now officially endorsed by the senior institutions of the stock exchange and widely accepted in Congress, was that - as Cowing (1965: 13) paraphrases it - 'if trading were confined to actual products, prosperity would disappear' (see also U.S. Congress 1892a: 156—68). These arguments generated largely by brokers and their representatives were only later translated into economic theory, notably in Emery's Speculation on the Stock and Produce Exchanges (1896) to become the orthodox position of liberal economists. In the forty years that followed, these debates and the issues they raised were extended to speculative share trading in general. The 'Hughes Report' of 1910 to the Governor of New York state, the flurries over the 'bucket shop era' at the turn of the twentieth century, and the debates over the 'Blue Sky' laws of the 1920s did little more than rehearse these arguments on a new stage - and generally to little effect. For the moment, at least, a kind of stalemate had been reached, in which clear recognition of the de facto status of many futures contracts as gambling was balanced by the vital economic function that they were held to serve. Gambling, insurance, and speculation had reached a crossroads. While the distinction among these remained critical to all parties in the debate, the pre-contractual, materialist theory of exchange had been destabilized. It was now more nearly possible for gambling itself to be reconciled with insurance and business investment.19 Experts in Uncertainty? One of the frequendy mentioned benefits of futures trading and associated speculation was the maximization of entrepreneurial opportunity and activity, and through this, the 'quickening' of the spirit of enterprise (U.S. Congress 1892a: 158-61). But several observations that surfaced during the fin de siecle period conditioned this asserted benefit by

248 Pat O'Malley raising moral and technical concerns. While the defenders of speculation had quite happily assigned the costs of uncertainty spreading to the hoards of small speculators, whose numbers and small holdings were thought to spread the risks with minimal impact, moral qualms began to emerge over their fate. In particular, it was argued that these people (particularly 'professionals' and 'women') were the victims of sharp practitioners who made their career on the market, and who were experts at stock speculation. In this respect, parallels were made with the legitimate gambling industry, which was beginning to make its presence felt in the eastern states (Cowing 1965). One response to these new concerns was a series of efforts over the next thirty years to minimize fraud. A more technical concern also arose: if speculation performed an essential function for commerce, was it right to entrust it to clearly inept or ignorant amateurs, rather than people 'of means and experience' whose activities were 'based on intelligent forecast' (Hughes Report 1910, quoted by Van Antwerp 1913: 228). As Brenner (1990) and Cowing (1965) have suggested, four criteria for 'speculative competence' emerged in this period, largely influenced by the considerations outlined above and by the impact of bureaucratic regulations developed recently in Germany.20 First, as the function of speculators is to effect risk spreading and market stability, then they were to be of sufficient financial means to trade without putting family or business at risk — since lack of means would affect their judgment and the tenure. Second, the speculator had to be distanced from key positions within the stock market, largely as a means to remove opportunities for fraud. The third criterion was that speculators had to have adequate knowledge of the commercial markets, so that their expertise could hold sway over the performance of the economy. The fourth was temperament - that is, speculators had to be possessed of 'character' - a bundle of attributes that insulated them from the emotional impact of successes and reverses and thus operated to effect stability. Such criteria became the object of various legislative experiments and stock market rules during the 1920s.21 Such thinking, inherited from the German regulatory scene (Cowing 1965: 115), reflected the influence of emerging visions of the 'planned economy,' even in the heartland of capitalist speculation.22 These initiatives were not seeking to regulate the market by transforming uncertainty into risk, either through direct control of market prices or by the imposition of decisions made by the human scientists of economics. While they sought to eliminate the haphazard and irrational operation

Contract Law and Distinctions between Speculation, Gambling 249 of commodity markets, their aim was more radical still: to identify and deploy experts in uncertainty. It was believed that uncertainty could be subordinated to calculative rationality, and that the various bodies such as stock exchanges and grain boards could identify those who should properly be excluded from market participation. In this the legislative initiatives met with considerable opposition, which was largely to cripple them. First, these various initiatives interfered with the right of individuals to invest their funds in profitable enterprises, thus tampering with what was regarded as a basic right in a laissez-faire economy, and - by implication - creating a closed and privileged community who would enjoy the fruits of speculation. Second, they collided with entrepreneurial common sense about the uncertainty of the market - something enshrined in contract law, and recently given new strength by enterprise radicals such as Tom Peters and Reuben Brenner. On this view, (t)he administrators had no way of telling better than the rest of the public whether a price change would be followed by a further movement in the same direction, let alone by how much, or for how long, and whether a fluctuation was of shorter or longer duration. Not only did the administrators have no advantage over the speculators in making predictions, they were worse at it. They had less incentive to make an effort. Hidden in the anonymous labyrinth of a government bureaucracy, they had neither reputation nor money at stake. In contrast, the opportunity cost to the speculator of making a bad prediction was higher: either to lose wealth or miss the opportunity of accumulating it. (Brenner 1990: 111)

The point at issue here is not that Brenner, and the historical subjects he is re-presenting, are right or wrong. Rather, it is to recognize the clash of two competing technologies for governing the market. Indeed, the effort to subject speculation to the discipline of rational calculation was perhaps the most far-reaching of its kind. It was not attempted even by the 1940s planners of the British welfare state, such as Beveridge, whose concern was never to rationalize speculation as such, but to capture the power of speculation and enterprise, to regulate production according to a rationale of social need, and in the process sometimes to take sectors of production out of the market. Perhaps not surprisingly, the impact of this initiative was slight and of short duration: the separation of speculation from gambling by the most transparent of legal fictions had not only survived criticism, new 'indigenous' knowledges of economics had rendered speculation a critical resource for 'the good of commerce.'

250 Pat O'Malley Conclusion Currently, the British government is proposing to reduce the tax on financial spread betting, which allows people to bet on the movement of stock exchange prices. In Australia, the requirement of an 'insurable interest' is no longer essential to insurance contracts. In most common law countries gambling has become a legitimate industry that governments seek to foster rather than suppress. At the same time, investment and even speculation have become almost de rigeur for the 'enterprising subject' who has become the new Everyman. In all of these developments, it would appear that the definition of what promotes the 'good of commerce' has shifted dramatically. The episodes with which this paper has dealt did not create this shift, but rather reflect a disturbance that generated new arguments concerning what constitutes the 'good of commerce.' Of equal importance, they centred speculation as a species of uncertainty that could not necessarily be - and did not need to be - reconciled with rational calculation in order to appear as an important feature of the (self) government of the economy. While we could read this shift as having created some of the conditions of existence of the contemporary ideology of enterprise and its revised legality, it is equally the case that these struggles brought into existence arrangements - concerning the formation of experts in uncertainty - that are readily associated with the 'regulated economy' that was coming into being in the 1930s. Certainly their borrowing from the German state initiatives is suggestive of this interpretation. Perhaps we would only find this surprising were we to believe that - in the varied visions of risk society theorists, rationalist sociologists, and those that 'thrive on chaos' - risk and uncertainty are in some way contradictory technologies. It would appear instead that they can be brought together in a variety of ways: by ignoring the difference between the calculable and the incalculable altogether (as in much of the law of contract), by rendering risk the servant of uncertainty (as in current conceptions of enterprise), by regarding risk as a superior technology that harnesses uncertainty (as in the economics of interventionism), by regarding uncertainty as the technology of creation that needs to be governed by risk (as in Keynesian visions), and so on. Risk and certainty appear to be distinct species of governmental technology, but ones that are readily capable of being braided together and perhaps even hybridized. Nevertheless, in all of this, uncertainty - governance through forms of

Contract Law and Distinctions between Speculation, Gambling 251 'reasonably' governing a future that is significantly incalculable and unformed — persists. It has as robust a lineage as do statistically or otherwise rationally calculable techniques of governance. In this, I suggest, we can also trace a critical element of the political rationality that we have come to think of as liberalism. I have argued here and elsewhere (O'Malley 2000) that liberalism was formulated in the courts as well as in the textbooks of liberal theorists. Yet here it becomes clear that it was also formulated in the patterns of struggle that developed around the law, and in the arguments that were generated out of such 'indigenous' fields of government as the stock exchanges and business associations. Each of these borrows from the other, but it is difficult to regard them all as partaking in a single, unilinear evolution, for in each field we may see new inventions emerging to resolve problems specific to their discursive domains. In thinking through the genealogy of this thing called liberalism, we may thus be better advised to follow such diversifying paths of analysis, rather than, say, reducing liberalism to an abstract formula such as the 'critique of government' (cf. Gordon 1991), or even attempting to find that constant thread that links together 'neo-liberalism,' 'classical liberalism,' and so on. The metaphor of liberalism as a woven or braided construct without essence — in which particular stands may fray, be discontinued, or be substituted by others - may be of more use than imagining all liberalisms as related under a covering principle. Perhaps they bear no more than an uncertain family resemblance to each other — sometimes with respect to one feature, sometimes with another - with no simple identity discernible. Notes 1 Such views date back to the early years of liberal economics. McCulloch (1825: 65), for example, argued that 'speculation is really only another name for foresight, and though fortunes have sometimes been made by a lucky hit, the character of a successful speculator is, in the vast majority of instances, due to him also who has skilfully devised the means of effecting the end he had in view and has outstripped his competitors in the judgement with which he has looked into the futurity.' 2 It may be argued that Weber's main concern was with the rational reinvestment of profits. But this, in turn, would only lead to systematic accumulation if investment and reinvestment were themselves more calculable activities, more assured of reproducing profits.

252 Pat O'Malley 3 This section summarizes the account appearing in O'Malley (2000). 4 This did not mean, of course, that all wagering was illegal: only that such contracts were enforceable as matters of honour. Exceptions existed, most notably with respect to state licensed lotteries (which, incidentally, had been a source of considerable state revenue since the seventeenth century). Ironically, the definition of a lottery in law is that prizes must be determined entirely by chance, any element of skill making the process not a lottery (Hallv. Cox, [1899] 1 QB 198). The last act authorizing the raising of money by lotteries was passed in 1823 (4 Geo. IV c. 60). 5 For example, Blackstone (1767: 4: 13) was generally concerned about such matters, noting its 'tending by necessary consequence to promote idleness, theft and debauchery among those of a lower class, and among persons of a superior rank it hath frequently been attended by the sudden ruin and desolation of ancient and opulent families.' 6 Nor was it to do with any idea that gambling is or is not a matter of rational calculation - as indeed may be the case with a skilled game of poker - for this was already the subject of considerable analysis in the realm of statistics (Dastonl985). 7 The Life Insurance Act, 1774 (14 Geo. Ill c. 48), renders illegal all insurances on any event in which a person or persons assured have no interest. The key provisions of the Act are: 1 No insurance shall be made ... wherein the person or persons for whose use, benefit or on whose account the policy or policies shall be made, shall have no interest, or by way of gaming or wagering. And every assurance made contrary to the true intent and meaning thereof, shall be null and void... 2 It shall not be lawful to make any policy or policies on the life or lives of any person or persons or other event or events without inserting in such policy or policies the person or persons' name or names interested therein, for whose use or benefit or on whose account such policy is made or under-wrote. 3 In all cases where the insured hath interest in such life or lives, event or events, no greater sum shall be recovered or received from the insurer or insurers than the amount or value of the interest of the insured in such life or lives, or other events or events. Note that this assumed that the insured event could be compensated in an 'amount or value': which in turn becomes an amount of money or equivalent. 8 This is not the case where the consideration is based on a gamble which is itself illegal.

Contract Law and Distinctions between Speculation, Gambling 253 9 In a masterpiece of understatement, Windeyer (1929: 20) concludes that it 'is often not easy to say whether a particular agreement is a wager or not.' 10 This was confirmed in the leading case of Carlill v. The Carbolic Smoke Ball Co., [1892] 2 QB 491): 'if the evidence is such as to make the intentions of the parties material in the consideration of the question whether a contract is a wagering one or not, and those intentions are at variance, those of one party being such as, if agreed in by the other would make the contract a wagering one whilst those of the other would prevent it from becoming so, this want of mutuality would destroy the wagering element of the contract and leave it enforceable by law as in an ordinary one' (per Hawkins, J.). 11 This concern with the intent to gamble, while still good law, created its own new breed of difficulties. To begin with, how was this intent to be known? Perhaps uncharacteristically in contract law, it was determined as early as 1860 that the court should go behind the contract itself and consider all of the evidence relating to the exchange in deciding whether or not the parties intended delivery (ExpartePhillips and Marnham, re Morgan, ((I860), 2 De G.F.&J. 634). 121 have already indicated the dangers of making such assumptions with respect to the impact of liberal economic theory on contract law (O'Malley 2000). 13 This role division is one of the critical and under-explored differences between the current ideologies of the 'enterprising subject' and the liberal subject of the nineteenth century. 14 This may be clarified if we consider the distinction between a simple option and a futures contract. The former could be settled by the exchange of a settlement of differences, and thus was regarded as unenforceable by the courts. A purchaser of a futures contract could in principle demand delivery of the material product. As Cowing argues (1965: 14), 'In practice, however, the distinction was meaningless, because futures contracts were settled just as simple options were - by the payment of differences. In only 3 per cent of the futures trades was there actual delivery; in fact to demand delivery was to brand oneself a miscreant and led to ostracism by the brokers' (emphasis added). 15 Much of the following discussion draws upon Cowing's (1965) pioneering work. 16 This was the aim of the Hatch bill, for example. While the bill passed narrowly in the lower house, delays in the Senate meant that it was not enacted and subsequent political developments prevented its reappearance in legislation. 17 Sir John Barnard's Act ([1734], 10 Geo. II c. 8) had prohibited the buying and selling of stocks not in the seller's possession at the time of the contract. The 'evil practice of compounding or making up differences' was

254 Pat O'Malley stated in the Act's preamble as having diverted many subjects 'from exercising their lawful trades and vocations to the utter ruin of themselves and their families, to the great discouragement of industry, and to the manifest detriment of trade and commerce.' The Act thus clearly engaged with the traditional vision of trade involving material exchange, and ran against the claims eventually accepted, discussed below. By 1860 it was repealed as it 'imposed unnecessary restrictions on the making of contracts for sale and transfer of public stocks and securities' (Van Antwerp 1913: 227). 18 Between 1883 and 1908 some fourteen state legislatures did prohibit futures trading, or taxed it to such a degree that the effect was to limit the market to dealings involving contracts in which one party actually held the commodity in question. Most of these states were rural states in the South, where the impact on national share market opinion and activity was marginal. Much of the legislative activity that followed was related to reducing fraudulent dealing associated with such markets. 19 It should be noted that in this period racetracks were also being legalized. The impact of this development on the questions of the enforceability of gambling contracts and the benefit of speculation and gambling to the economy has yet to be determined. 20 Germany restricted futures dealing in grain and some other products and established a register of those who were permitted to engage in futures transactions. For a contemporary view of this initiative, see generally Loeb (1897). 21 In the following years, a number of such proposals were put into effect, usually in the form not of legislative requirements but as regulatory schemes administered by stock exchanges. In one sense, this is business as usual, for as Ferguson (1983) has pointed out, stock markets have a considerable history as regulators of contracts below or outside the threshold of contract law, relying heavily on informal business ethics linked with the sanction of exclusion. Such activity has always involved regulating admission and continued membership, and thus the extension of these powers to regulate speculators, while problematic, was not a radical departure. What is of notice, however, is that stock exchanges for many years had resisted attempts to regulate speculation and had not been impressed by the concern that delivery should be a possibility in futures trading; indeed, as Cowing argues (1965: 14), a demand for delivery would lead to ostracism by the brokers. As Lurie indicates (1979: 59), actual delivery on the Chicago exchange occurred in fewer than 10 per cent of transactions, and even the courts noted, 'delivery is the rare exception and the intention to deliver likewise is rarely present' (Board of Trade v. Kinsey & Co. (1904), 130 Fed. 507).

Contract Law and Distinctions between Speculation, Gambling 255 In Britain, much of this fascinating but tortured process was pre-empted by the recommendations of the Royal Commission into the London Stock Exchange (LSEC 1878). It was argued that in the vast majority of cases, at the time of making the contract it was impossible for brokers to distinguish between difference contracts and contracts 'that arise from the desire to invest money or to sell securities.' Thus even though the commissioners were concerned that 'gambling to an enormous extent does exist in the present day, in securities of all kinds,' 'we do not think it is practicable to make bargains entered into for the purposes of speculation or gambling any more illegal than they are at present, and we do not propose any change in the law' (1878: 21). Instead, the commissioners passed regulation over to the Stock Exchange itself. Since brokers would become aware over time of who were engaging in such activities, because of the amounts of stock bought and sold and the way such bargains are dealt with on settling day, the 'Committee of the Stock Exchange might therefore very well hold a restraining hand over their own members who should turn out to have lent themselves to extravagant speculation' (1878: 21). This, basically, was the position eventually reached in the United States. 22 The British government kept a watching brief on these developments in Germany, and sought advice from its ambassadors on how a number of states, including the United States, were responding. In the end, no action appears to have been taken.

References Atiyah, P. 1979. The Rise and Fall of Freedom of Contract. Oxford: Clarendon Press Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage Belasco.J. 1990. Teaching the Elephant to Dance: Empowering Change in Your Organization. London: Hutchinson Blackstone, W. 1769 [1922]. Commentaries on the Laws of England. Philadelphia: Bisel Blaug, M. 1988. Economic Theory in Retrospect. Cambridge: Cambridge University Press Brenner, R. 1990. Gambling and Speculation: A Theory, a History and a Future of Some Human Decisions. Cambridge: Cambridge University Press Cowing, C. 1965. Populists, Plungers and Progressives: A Social History of Commodity Speculation. Princeton: Princeton University Press Daston, L. 1985. Classical Probability in the Enlightenment. Princeton: Princeton University Press Dean, M. 1999. Governmentality: Power and Rule in Modern Society. London: Sage

256 Pat O'Malley Dewey, T. 1905. Legislation Against Speculation and Gambling in Futures. New York: Scribners Emery, H. 1896. Speculation on the Stock and Produce Exchanges of the United States. New York. Ewald, F. 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 197-210. Chicago: University of Chicago Press Ferguson, R. 1983. 'The Horwitz Thesis and Common Law Discourse in England.' Oxford Journal of Legal Studies §: 34-42 Giddens, A. 1990. The Consequences of Modernity. Stanford: Stanford University Press Gordon, C. 1991. 'Government Rationality: An Introduction.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 1-51. Chicago: University of Chicago Press GordleyJ. 1991. The Philosophical Origins of Modern Contract Doctrine. Oxford: Clarendon Press Hamburger, P.A. 1984. 'The Development of the Nineteenth-Century Consensus Theory of Contract.' Law and History Review 7(2): 274-301 Hart, H.L.A., and T. Honore. 1985. Causation in the Law. 2nd ed. Oxford: Oxford University Press Horwitz, M. 1977. 'The Triumph of Contract.' In The Transformation of American Law. Harvard: Harvard University Press Kanter, B. 1989. When Giants Learn to Dance: Mastering the Challenge of Strategy: Management and Careers in the 1990s. New York: Simon and Schuster Kercher, B. 1996. Debt, Seduction and Other Disasters. The Birth of Civil Law in Convict New South Wales. Sydney: Federation Press Kercher, B., and M. Noone. 1990. Remedies. Sydney: Law Book Company. Knight, F. 1921 [1964]. Risk, Uncertainty and Profit. New York: A.M. Kelley Loeb, E. 1897. 'The German Exchange Acts of 1896.' Quarterly Journal ofEconomics. July: 187-214 LSEC (London Stock Exchange Commission). 1878. Report from the Commissioners on the London Stock Exchange. (Command 2157) London: HMSO LurieJ. 1979. The Chicago Board of Trade, 1859-1905. Chicago: University of Illinois Press McCulloch, J. 1825. Principles of Economics. London: Bradshaw Miller, P., and N. Rose. 1990. 'Governing Economic Life.' Economy and Society 19: 1-31. O'Malley, P. 1999. 'Imagining Insurance: Risk, Thrift and Industrial Life Insurance.' Connecticut Insurance Law Journal 5: 675-705. - 2000. 'Uncertain Subjects: Risks, Liberalism and Contract.' Economy and Society 29: 460-84

Contract Law and Distinctions between Speculation, Gambling 257 Osborne, D., and T. Gaebler. 1993. Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector. New York: Plume Books Peters, T. 1987. Thriving on Chaos: Handbook for a Management Revolution. New York: Knopf - 1994. The Tom Peters Seminar: Crazy Times Call for Crazy Organizations. London: Macmillan Reddy, S. 1996. 'Claims to Expert Knowledge and the Subversion of Democracy: The Triumph of Risk over Uncertainty.' Economy and Society 25: 222-54 Rose, N. 1990. 'Governing the Enterprising Self.' In P. Heelas, and P. Morris, eds., The Values of the Enterprise Culture. London: Unwin Hyman Simpson, A.W.B. 1975. 'Innovation in Nineteenth Century Contracts.' Law Quarterly Review 91: 247-78 - 1979. 'The Horwitz Thesis and the History of Contracts.' University of Chicago Law Review 46: 533-601 Tilbury, M., M. Noone, and B. Kercher. 1993. Remedies: Commentaries and Materials. Sydney: Law Book Company U.S. Congress. 1892a. Fictitious Dealing in Agricultural Products. Washington: Committee on Agriculture. House of Representatives. [52 Congress 3 Sess.] - 1892b. Options and Futures. Washington: Committee on Judiciary. United States Senate [52 Congress 2 Sess.] U.S. Senate. 1914. Stock Exchange Regulation: Hearings Before the Committee on Banking and Currency. U.S. Senate 63 Cong. 2. Sess. Washington: Government Printing Office Van Antwerp, W. 1913. The Stock Exchange from Within. New York: Doubleday Page Want, J. 1995. Managing Radical Change: Beyond Survival in the New Business Age. New York: John Wiley Weber, M. 1958. The Protestant Ethic and the Spirit of Capitalism. New York: Scribners - 1965. 'Bureaucracy.' In H. Gerth and C.W. Mills, eds, From Max Weber: Essays in Sociology. London: Routledge and Kegan Paul Windeyer, W. 1929. The Law of Wagers, Gaming and Lotteries. Sydney: The Law Book Company of Australasia Zelizer, V. 1979. Markets and Morals. Princeton: Princeton University Press

10

Containing the Promise of Insurance: Adverse Selection and Risk Classification TOM BAKER Insurance, therefore, takes from all a contribution; from those who will not need its aid, as well as from those who will; for it is as certain that some will not, as that some will. But as it is uncertain who will, and who will not, it demands this tribute from all to the uncertainty of fate. And it is precisely the moneys thus given away by some, and these only, which supply the fund out of which the misfortune of those whose bad luck it is that their moneys have not been thrown away, are repaired. The afflicted finds his money spent to some purpose; and only the fortunate part with it for nothing. From this point of view the whole beauty of the system of insurance is seen. It is from this point of view that it presents society a union for mutual aid, of the fortunate and unfortunate, where those only who need it receive aid, and those only who can afford it are put to expense. Thus, while the aggregate of human suffering and calamity remains undiminished ... thus, while the uncertainty of their visitation remains unremoved, human ingenuity and cooperation equalize the distribution of this fearful aggregate, and alleviate the terrors of uncertainty. By a system of mutual insurance thus generally established, embracing all callings, a great fund, as it were, for the benefit of society, would be created; a fund to which none could be said to contribute gratuitously, from which none but the needy should be aided; a great reserve fund, held in readiness for the uncertain case of want. We thus have the mechanic, the laborer, and the merchant, joined hand in hand in mutual protection against the risks of their callings; we have the masses, above all, shielded from the most blighting evil of the inequality of human condition, the danger of destitution; we have society united on the basis of mutual insurance. (Jacques 1849)

Adverse Selection and Risk Classification 259 Written by an American lawyer and insurance entrepreneur, this midnineteenth-century text offered insurance as a solution to at least some of the contradictions of capitalism. Accepting the inequality that resulted from capitalist modes of production, the author presented insurance as a technology for protecting people from what the sociologist Joseph Schumpeter would later call the 'creative destruction' of capitalism. Against Marx (or, more likely, against the mid-nineteenth-century French socialists who influenced Marx), Jacques offered a less radical vision. Instead of collective ownership of the means of production, there would be collective protection against misfortune. Not quite 'to each according to his need and from each according to his ability,' but rather 'society united on the basis of mutual insurance.' Looking back from our position on the far side of the twentieth century, it is easy to say who held the better crystal ball. The technology and form are not precisely as Jacques predicted, but insurance has become nearly as central a social institution as he imagined. Yet the expansion of insurance has been accompanied by a better sense of the limits of insurance as an engine of social solidarity. Notwithstanding the ubiquity of insurance, the 'unity' it produces is much less complete than Jacques's Utopian image might suggest. Two reasons commonly given for the limits on the promise of insurance are the problems of moral hazard and adverse selection. 'Moral hazard' refers to the change in incentives that can result from insurance protection. 'Adverse selection' refers to the theoretical tendency for lowrisk individuals to avoid or drop out of insurance pools, with the result that, absent countervailing efforts by administrators, insurance pools can be expected to contain a disproportionate percentage of high-risk individuals.1 The problem of moral hazard has received significant attention in the sociology of risk and insurance (Heimer 1985; Baker 1996, 2000; Ericson, Barry, and Doyle 2000; Stone 2002). Although the rhetoric of moral hazard can be misused to disguise the politics of selfinterest, there is little doubt that protection against harm can increase what Jacques called the 'aggregate of human suffering and calamity' (Baker 1996). For this reason, all successful insurance institutions take measures to align individual incentives with the common goal of minimizing the frequency and extent of insured losses (Heimer 1985). The problem of moral hazard contains (in the sense of limits) the promise of insurance by requiring insurance institutions to balance protection and control. The more that a particular risk lies within the

260 Tom Baker control of the insured, the less confidently insurance institutions can insure that risk. To address the problem of moral hazard, insurers require either that the insured relinquish control or retain some risk 'coinsurance.' Taking control away from insureds - as liability insurers commonly do with respect to the defence and settlement of insured claims - has costs. Where control is particularly expensive, even the most risk-averse insured prefers some coinsurance to paying the price of full insurance, creating what Carol Heimer has termed a 'community of fate' between insurer and insured (Heimer 1985). Risks that pose a very high degree of moral hazard typically are not insurable at all. For example, intentional harm is almost always excluded from insurance coverage. This chapter begins to extend to the problem of adverse selection the critical attention provided in prior work to the problem of moral hazard. Like moral hazard, adverse selection is an old insurance concept that was adopted, formalized, and generalized by economists developing the economics of information. As with moral hazard, insurance economics has addressed the phenomenon of adverse selection largely from the insurers' point of view. Thus, part of my project will be to examine the problem from a different point of view: how insurers, too, create and shape adverse selection. Perhaps the most significant theoretical advance presented here is examining adverse selection in terms that can be applied to both sides of the insurance relationship. At least in the context of insurance risk classification, it is useful to think of adverse selection as a 'dual' problem (similar to moral hazard), meaning that actions to address the adverse selection problem can lead to the depooling effect that motivated the actions in the first place. The chapter proceeds in two parts. This first sets forth the case for understanding adverse selection as a dual problem and highlights alternatives to insurance risk classification. The second uses three historical examples to explore moral justifications for insurance risk classification: the nineteenth-century controversy over age-based pricing in fraternal insurance, the mid-twentieth-century controversy over experience rating in unemployment insurance, and the late-twentieth-century controversy over efforts to exclude battered women from life, health, and disability insurance pools. These examples demonstrate that, rather than being a neutral, technical solution to a structural dynamic inherent in the insurance relationship, risk classification reflects moral commitments. My goal in this chapter is not to discredit the practice of risk classification, but rather to focus attention on the morality that is implicit in arguments for and against it.

Adverse Selection and Risk Classification 261 Parti Adverse Selection

Sometimes called 'anti-selection' in the insurance trade literature, adverse selection refers to the theoretical tendency for low-risk individuals to avoid or drop out of voluntary insurance pools, with the result that, absent countervailing efforts by administrators, insurance pools can be expected to contain a disproportionate percentage of high-risk individuals (Rothschild and Stiglitz 1976). For example, adverse selection is said to explain the disparity in prices between group and individual health insurance in the United States. When an employer signs up employees as a group, the insurer acquires both the low and the high risks. Where people decide on their own whether to purchase insurance, those who need it the most are the most likely to purchase it (assuming they have the financial means), and insurers consequently end up with more high risks in the pool and fewer low risks. The phenomenon of adverse selection appears to have been given this name in the life insurance industry in the nineteenth century. In order to decrease the odds of paying a premature death claim, early life insurers, like those today, commonly selected among lives according to the health of the applicant. With the development of mortality tables that represented the average rate of death within the population, insurers began to price their policies on a 'scientific' basis (Van Amringe 1873). By pricing on the basis of average mortality, building in a margin for expenses and profit, and then selecting to beat the average, insurers were sure to turn a comfortable profit, or at least so insurance theorists reasoned. In practice, insurers were confounded to find that the mortality experience of insured lives matched, and in some cases suffered by comparison to, the mortality experience of the general population (King 1876). Given that they were selecting on the basis of health status, the mortality experience of insured lives should be much better than that of the general population! The answer, actuaries reasoned, was another force at work, adverse to their selection efforts. For life insurers, there were two components to this adverse selection. The first operated at the inception of the insurance relationship: 'if the medical examiner did not stand at the entrance gate, the weakest and least desirable lives would be surest and soonest to come in' (Lippincott 1905: 200). The second was the discovery of ill health by those who had already purchased insurance, leading to a

262 Tom Baker decline in the average health status of the insured population. The people who decided to discontinue their policies were disproportionately the healthy. Even if 'the lives at starting are a very select class, they not only lose this advantage, but degenerate till they are on average worse than the general population' (King 1876: 397). The answer was renewed attention to selection (classification) at the entrance door and the design of insurance contracts that inhibited healthy lives from leaving through the exit (ibid.; Lippincott 1905). These two strategies of nineteenth-century life insurers represent the paradigmatic responses to the problem of adverse selection: risk classification and binding risks to the insurance pool. The choice between these responses mirrors that between the responses to moral hazard: coinsurance and control. Like coinsurance, classification leaves individuals exposed to more risk. Like control, binding risks to the insurance pool constrains at least some aspects of the autonomy of the participants in the pool. Insurance Risk Classification

Insurance risk classification is the process of sorting insurance applicants into categories believed to correspond to differences in expected risk. Common examples include sorting life insurance applicants by age, health insurance applicants by health status, workers' compensation insurance applicants by type of industry, and property insurance applicants by the nature of the construction of the property to be insured (e.g., wood versus brick). Risk classification highlights the distributive nature of insurance. Insurance is predicated on the existence of a large percentage of fortunate members of the insurance pool, whose premium dollars go to pay the losses of the unfortunate members. The excerpt from Jacques quoted at the beginning of the chapter presents a 'share and share alike' vision of this distribution. All of society is included in the pool: 'the mechanic, the laborer, and the merchant, joined hand in hand in mutual protection against the risks of their calling.' In this Utopian vision, insurance serves only to preserve the status quo, maintaining the economic and social status of members as it stands before misfortune strikes (cf. Weisbrod 2000). The introduction of risk classification complicates this picture. Some people must pay more than others to enter the pool, and others cannot enter at any price. Thus, insurance institutions not only maintain status,

Adverse Selection and Risk Classification 263 they also assign it. For example, the children of a parent refused life or disability insurance maintain a more tenuous grasp on their position as a result of the insurers having classified their parent as a high risk. Should the parent die or become disabled, the children's resulting loss of social position derives not from the death or disability alone, but also from the insurance risk classification. Because of that classification, there will be no insurance payment to offset the loss of the parent's income. In a world of competitive, voluntary insurance organizations, it is easy to see why insurers classify applicants according to risk. Risk classification is one of the most potent competitive tools. Eliminating the most risky from an insurance pool reduces the average cost of insuring the members of the pool, allowing the insurer to offer a lower price and, possibly, obtain a greater profit. An insurer that discovers a new way to identify and exclude high risks improves its competitive position in two ways: it lowers its average risk and, assuming the people it rejects go elsewhere, it increases the average risk of its competitors. This competitive power of risk classification produces a classification 'arms race,' in which insurers either maintain their classification edge or face the loss of low risks to the competition and the migration of the high risks to their insurance rolls. The use of risk classification to exclude applicants from the pool illustrates most starkly the obvious fact that classification can reduce the degree to which insurance spreads risk.2 Charging different prices or offering different contract terms on the basis of expected risk has a similar, if softened effect, as do other, more indirect, risk classification measures, such as targeted marketing and designing insurance contracts to segregate applicants on the basis of risk.3 Risk Classification Can Create Adverse Selection

As the idea of a risk classification 'arms race' suggests, risk classification itself can create a kind of adverse selection. Risk classification innovations allow insurers to select risks in a manner that is adverse to the insurance pool, by reducing the ability of the insurance pool to spread risk. Referring to this practice as 'adverse selection' may seem like a play on words. But the dynamics of insurer-side adverse selection are similar to those of the insured-side adverse selection that risk classification addresses. Both involve self-interested behaviour based on the risk status of insureds. Both are collective action problems, in which individually rational actions produce a result that is contrary to the inter-

264 Tom Baker ests of the whole. And both inhibit the ability of insurance institutions to spread risk. The fact that risk classification can lead to a collective action problem means that even those who ordinarily trust or believe in the market cannot easily conclude that any particular risk classification is optimal from an efficiency or utilitarian perspective.4 Outside formal economic models, there are no purely technocratic, value-free answers to the question of who should pay how much to jump into the insurance pool. Efforts to address adverse selection share the reactivity of efforts to address moral hazard (Baker 1996). Acting to prevent adverse selection on one side of an insurance relationship can promote adverse selection on the other. Thus, as with moral hazard, there is much to be gained from thinking of adverse selection as a 'dual' problem that can affect both sides of an insurance relationship.5 The Alternative: Binding Risks to the Insurance Pool Just as there is a 'control' alternative to the usual 'coinsurance' prescription for moral hazard, there is also a 'control' alternative to the usual 'risk classification' prescription for adverse selection. In the case of adverse selection, the 'control' is directed at the ability of insureds to opt out of insurance and the ability of insurers to slice up the insurance pool on the basis of risk. A simple alternative solution to the problem of adverse selection is mandating universal insurance, to be provided through a single insurer. A universal insurer can charge everyone the same price without any fear of low risks dropping out or defecting to a competitor. Many forms of government-provided insurance work in this way. A leading example is the U.S. Social Security system, which provides nearly universal retirement, disability, and life insurance to the employed population. A second approach is to mandate (directly or indirectly) that everyone purchase insurance, prohibit insurers from charging prices or underwriting on the basis of risk, but allow multiple insurers into the market. This is the ordinary practice in the United States for employment benefits within large companies,6 and it was the approach of the Clinton administration's universal health insurance proposal. Limits on risk classification in auto insurance - which is mandatory in the United States and many other parts of the world - represent partial adoptions of this approach. This approach cannot completely eliminate adverse selection because individual insurers may have some ability to design or market

Adverse Selection and Risk Classification 265 their services in a way that appeals disproportionately to low risks. Nevertheless, this approach can significantly improve the risk-spreading function of the affected insurance programs as compared to what is otherwise possible through the market, as demonstrated by comparing the large group and individual health insurance markets in the United States.7 A third approach is to prohibit (or limit) risk-based pricing and underwriting without requiring the purchase of insurance. Some might object that this approach does not address adverse selection at all, but rather deprives insurers of the means to combat it. Yet, this objection focuses only on insured-side adverse selection and ignores the role of insurer-side adverse selection in disaggregating insurance pools. Eliminating insurer-side adverse selection may be enough to keep most risks in at least some insurance pools. Faced with the choice between no insurance and expensive insurance, many or even most lowrisk applicants would not change their purchasing decision because of the higher price.8 Purchasing insurance may be the cultural norm, people may be sufficiently risk averse, or there may be institutional arrangements that encourage the purchase of insurance, as in the case of homeowners' insurance in the United States (Heimer 2002). Thev degree to which this approach improves on the risk-spreading function of the insurance market is more difficult to determine than in either of the preceding two approaches. The high risks that would not otherwise have been able to purchase insurance are better off than they would be in the absence of regulation of this kind. On the other hand, the low risks that choose not to purchase insurance at the higher price face their risks alone, so they are worse off than they would be otherwise. In contrast to the 'free market' situation, however, all the people without insurance would have had the opportunity to buy insurance at a reasonable price. Thus, there are autonomy gains from this apparently antiautonomy approach. I am not aware of any complete adoption of this third approach, but there have been many partial adoptions. For example, some forms of health insurance have been required by state law in some U.S. states to charge all individual applicants the same price.9 A more limited approach is to prohibit certain kinds of classifications. Race, religion, and national origin are the most commonly prohibited insurance classifications in the United States, but gender, age, and other, more narrowly defined classifications are also prohibited in some contexts. A recent variation on this approach is to prohibit the use of information from genetic testing by insurance companies.

266 Tom Baker A fourth approach to addressing adverse selection is designing insurance products to change the incentives of low risks that are already in the insurance pool. This addresses the problem identified by the nineteenth-century insurers: 'even if the lives at starting are a very select class, they not only lose this advantage, but degenerate till they are on average worse than the general population' (King 1876: 397). One example is level premium life or disability insurance products. With a level premium product the insured, in effect, pre-pays part of the premium for later years by paying a higher amount in the early years. Ten years into a twenty-year level premium life or disability insurance policy even the healthiest members of a given cohort will be paying a lower premium than would be available if they started fresh elsewhere. Level premiums and other methods for keeping low risks in the pool facilitate the spreading of what Robert Works has called in another context 'classification risk' - the risk that the risk status of the insured will worsen in the future (Works 1998). Waiting periods before coverage takes effect, pre-existing condition exclusions in health insurance, discounts that apply only after a period of continuous insurance coverage, and penalties for early termination of accumulating value forms of life insurance are all methods for keeping low risks in the pool. For present purposes, this last approach is less significant for its overall effect on the ability of insurance institutions to spread risk than for its demonstration of two important points. First, even some voluntary insurance institutions ordinarily and regularly define people's choices in a way that constrains their ability to realize the full benefits of their lowrisk status. Second, people are willing to accept such constraints to further risk-spreading objectives. These points are a partial response to autonomy objections to government regulation of insurance risk classification. Part II

The first part of this chapter described adverse selection as a dual problem and highlighted control alternatives to insurance risk classification. As this discussion reflects, insurance risk classification is only one possible approach to the problem of adverse selection, and it can create the very depooling that, in theory, it is intended to prevent. Thus, the economics of adverse selection cannot adequately explain either a social decision to leave insurance risk classification in the hands of insurance companies or the decision by insurance companies to classify risks in the

Adverse Selection and Risk Classification 267 way that they do. This second part will describe alternative, moral justifications for the practice of insurance risk classification and explore these justifications in the context of three historical debates. Justifying Risk Classification Like political and economic arguments couched in the language of moral hazard, arguments made in the name of adverse selection draw on a reservoir of respect for science, and they are similarly buttressed by moral appeals. Risk classification is justified not only as necessary and inevitable, but also as a good thing. The leading moral justifications for risk classification are as follows: 1) without risk classification, low risks are unfairly forced to subsidize high risks; 2) risk classification promotes socially beneficial efforts to prevent loss; and 3) risk classification promotes individual responsibility. To illustrate these justifications in action I will briefly describe three public policy debates over risk classification. The first is the nineteenthcentury debate over age-based pricing in British fraternal insurance. The second is the Depression era and continuing debate over experience rating in U.S. unemployment insurance. The third is a recent debate over discrimination by U.S. life, disability, and health insurers against battered women. The age-based pricing debate illustrates the fairness or subsidy justification; the experience rating debate illustrates the loss prevention and responsibility justifications; the insurance for battered women debate illustrates some of the limits of those justifications. Age Rating in Fraternal Life and Sickness Insurance Well into the twentieth century, many fraternal insurance organizations required their members to make equal contributions to their common fund, regardless of age or health status (Emery 1996). In Britain actuaries began challenging this practice by the 'friendly societies' in the early nineteenth century (Ansell 1835; Hardwick 1859). Friendly societies were fraternal organizations that provided life and sickness insurance benefits to their members. The sickness benefits were what we would call today short-term disability benefits. The actuaries complained that the premiums paid to the societies by younger members were used, unfairly, to subsidize the benefits of older members. Charles Ansell, an actuary for the Atlas Insurance Company, set out

268 Tom Baker this fairness argument in his Treatise on Friendly Societies, published under the auspices of the Society for the Diffusion of Useful Knowledge: It has been common heretofore to charge members of Friendly Societies who might enter them, at ages often differing by 20 years, the same rates of contribution; but since the following tables, and the data on which they are founded, show very plainly that for every benefit to which they refer the proper contribution varies with every year of age, the injustice of requiring men of different ages to pay a like rate must be manifest; and as little excuses can be hereafter urged for a continuance of so objectionable a practice, it will, in all probability, be at once abandoned, as being utterly at variance with that feeling of equity and benevolence to which all well-regulated Friendly Societies owe their origin and existence. (Ansell 1835: 107)

Notwithstanding the actuaries' complaints, many friendly societies continued to require equal contributions regardless of age, in keeping with their principles of fellowship and mutuality (cf. Clawson 1989). These friendly societies rejected the underlying assumption of the actuaries' argument: that justice required that members pay for insurance according to their individual chances of loss. A second argument pushed with increasing vigour in the latter half of the nineteenth century combined Ansell's fairness argument with the logic of adverse selection. The actuaries argued that the friendly societies had not set aside sufficient reserves to protect themselves against the ageing of their membership, with the result that the younger workers would not receive the benefits that they were promised in return for their premiums. Thus, the level premium arrangement was a means for the older members to take advantage of the younger. The actuaries reasoned as follows. Because the societies lacked sufficient reserves to pay future claims without the infusion of new members, the current younger members were dependent on the ability of those societies to continue to attract younger members, whose premiums would pay part of their benefits. But other, more actuarially sound, friendly societies were charging age-based premiums. Thus younger people in the future would heed the call of justice and self-interest - as the actuaries defined them - and join these organizations, abandoning the currently young in friendly societies with ageing populations that could not in the long run support the benefits promised. The Quarterly Review summed up one of the conclusions of a series of actuarial reports on friendly societies in 1864 as follows:

Adverse Selection and Risk Classification

269

So long as the societies consisted, for the most part, of young, healthy men, and the average amount of sickness remained low, the payments made seemed ample; the funds accumulated, and many flattered themselves that they were in a prosperous state, when in fact they contained the sure elements of decay. For, as the members grew older, their average liability to sickness was regularly increasing. The effects of increased age upon the solvency of benefit clubs soon becoming known, the young men avoided the old societies, and preferred setting up organisations of their own. The consequence was, that the old men began to draw upon their reserves at the same time that the regular contributions fell off; and when, as was frequently the case, a few constantly ailing members kept pressing upon the society, the funds at length became exhausted, and 'the box' was declared to be closed. (Quarterly Review 1864: 327-8)

In this view, friendly societies that continued charging level premiums were 'swindling the young members for the sake of the older ones' (Chambers 1859: 356). The proposed solution, which many friendly societies continued to resist, was to adopt age-adjusted premiums of the sort pioneered in the commercial life insurance societies. Whether the level premiums in fact explain the gradual demise of the friendly societies is a subject of historical dispute. The traditional view was to this effect, although there have been suggestions that the more significant problem was that the friendly societies began paying 'retirement' benefits under the guise of 'sickness' benefits (Gilbert 1965). Recent and more sophisticated financial analysis of the Canadian counterparts to the friendly societies demonstrates that, notwithstanding the claims of the nineteenth-century actuaries, many level premium fraternal insurance organizations were financially sound well into the 1920s (Emery 1996). For present purposes, the significance of the age-rating debate lies, not in the merits, but rather in the relationship between the subsidy and adverse selection arguments invoked in the debate. The subsidy argument makes it 'fair'; the adverse selection argument makes it 'necessary.' Where there is resistance to the fairness claim, the necessity claim helps to construct the high-risk members of the pool as different from, and hostile to, the low-risk members of the pool. At the same time, the fairness claim legitimates behaviour by insurers that makes risk classification 'necessary' - that is, the efforts of the risk-classifying insurers to poach the low-risk members from the resisting insurers (what I would call insurer-side adverse selection).

270 Tom Baker Even if the mid-nineteenth-century predictions of the actuaries proved true and the level premium friendly societies failed because they could not attract new members, that demonstrates neither that age-based premiums were fair nor that they were necessary. At most, the age-based premiums were necessary only because of the insurer-side adverse selection, which could have been prevented through government policies that reduced poaching. Moreover, while the ability of some members to collect their benefits was certainly unfair, the failure to attract new members was as likely to be attributable to the actuaries' disparagement of the financial soundness of the level premium societies as it was to the level premiums themselves.10 Indeed, as explained above, as long as the premiums are adequate, level premiums can in fact help control insured-side adverse selection by keeping low risks in the insurance pool. Disparaging the financial soundness of a friendly society or other insurer, however, can easily become a self-fulfilling prophecy, much like a run on a bank. Who would want to start a long-term relationship with an insurer that might not be able to pay claims in the future? None of this is to claim that concern with fairness or adverse selection played no role in the shift to risk classification. But stronger explanations lie elsewhere most likely, in my view, in the ascendance of the actuarial vision of insurance in which 'the ideal type of insurance involves premiums paid in advance, guaranteed indemnity in the event of a covered loss, and riskbased premiums based on the best available information regarding the expected losses of the individuals insured' (Baker and Simon 2002: 10). Experience Rating in Unemployment Insurance In contrast to life, health, and disability insurance, there has never been a broad unemployment insurance market. Some individual firms and labour unions have provided short-term unemployment insurance benefits for their employees or members, and some lenders have provided a modest amount of unemployment insurance in the form of debtforgiveness provisions that are contingent upon unemployment. But the problems of adverse selection and moral hazard seem to have made a broad market in unemployment insurance impracticable.11 An additional factor inhibiting the development of an unemployment insurance market is the fact that much of the benefit of unemployment insurance accrues to society at large, and not simply to the individuals or firms that purchase it. Unemployment insurance allows workers time to look for jobs that match their education and training, and it supports

Adverse Selection and Risk Classification 271 families and communities during hard times, all of which lead to a more productive economy. Termed 'positive externalities' by economists, benefits like these cannot be promoted by the market because there is no practical way for individual firms to charge society at large. It is not that positive externalities prevent the development of a market, simply that the market will provide less of the thing that produces the positive externalities than society would be willing to pay for. The positive externalities of, and adverse selection in, unemployment insurance help explain the decision by governments in most developed countries to provide it.12 Providing unemployment insurance through the state solves the positive externality problems, because the state has the means to charge society at large for unemployment insurance benefits. It also solves the adverse selection problem, because the state has the authority to bind risks to the pool. Since states can solve, and in most cases have solved, the adverse selection problem by mandating universal participation in an unemployment insurance system, why have I included a discussion of this type of insurance in an essay on risk classification? With mandatory insurance and a single insurance provider, there should be no need to classify risks to prevent adverse selection. Yet, at least in the United States, there is risk classification in unemployment insurance, in the form of experience rating. Experience rating is the insurance term for charging different prices based on past experience. It is a form of risk classification because past experience is used to predict future risk. So, for example, an employer that has laid off significant numbers of workers in the past is treated as a higher risk for lay-offs in the future and, accordingly, is charged a higher rate. The existence of risk classification in unemployment insurance further demonstrates that the appeal of risk classification goes beyond concerns about adverse selection. One basis for that appeal is the fairness justification addressed in discussing the debate over age rating in friendly societies: people should pay according to the risk they bring to the insurance pool. Although not fundamentally different in the unemployment context, the fairness claim may be stronger, since many employers are likely to have more control over their unemployment risks than most people have over the ageing component of their risk status.13 There is a second, prevention justification for experience rating that relates to this same control: experience rating rewards employers who provide stable employment and punishes those who do not, and thereby encourages employers to stabilize their employment rolls. William

272 Tom Baker Beveridge, one of the architects of the British social welfare system, described the underlying moral hazard concern as follows: Those dangers, in a sentence, lie not so much in the risk of demoralizing recipients of relief, so that they do not look for work, as in the risk of demoralizing governments, employers and trade unions, so that they take less thought for the prevention of unemployment ... The fear of causing unemployment may vanish from the minds of trade-union negotiators and open the way to excessive rigidity of wages and so to the creation of unemployment. Industries practicing casual engagement or perpetual short time may settle down to batten on the taxation of other industries or of the general public in place of reforming their ways. (Beveridge 1930: 43)

As with any other moral hazard situation, the 'solutions' will be a mix of control and coinsurance. From a moral hazard perspective, experience rating is a form of coinsurance in which employers face the threat of paying higher premiums in the future as a consequence of lay-offs. Control options involve restrictions on lay-offs for employers and voluntary quitting for employees.14 Generalizing from this example we can see many instances in which risk classification is justified on the grounds of loss prevention. Reduced property insurance premiums for buildings with sprinkler systems, fire alarms, and night watchmen arguably reduce insurance losses, as do reduced auto insurance premiums for cars with good safety records and anti-theft devices. Whether non-smoking and other lifestyle classifications accomplish the same goal for life and health insurers is more controversial, but these discounts have been justified on the same basis. A final justification for experience rating in unemployment insurance is less pragmatic and more openly ideological. Labour economist Joseph Becker described this justification as follows: '(1) As its main mechanism for the allocation of resources, our society has chosen the free market. (2) The market works more efficiently as market prices more accurately reflect the full costs of production. (3) The costs of production are reflected in market prices the more fully as the unemployment tax is the more completely experience-rated. (4) Thus experience rating accords with society's choice of the market as the main mechanism for the allocation of resources' (Becker 1972: 7). Although one could take issue with these four points,15 they are a useful description of a belief structure that supports experience rating, specifically, and risk classification more generally. The underlying idea

Adverse Selection and Risk Classification 273 here is that actors should be made responsible for the risks that they bring with them to the market. Accordingly, prices for insurance should be tailored to the expected risk of each insured. In the context of unemployment insurance, that means that unemployment premiums should be based on the experience and risk of the individual firm, or perhaps even on the experience and risk of the individual employee, although that has not yet to my knowledge been advocated. What makes this different from the prevention (moral hazard) justification for risk classification is the explicit connection to the governmental rationality - 'governmentality' - of liberalism and responsibility. Risk classification is not simply fair, and it does not depend on the empirical grounds of moral hazard. It is constitutive of an entire approach to governing the self and others.16 The connection between risk classification and individual responsibility helps explain why risk classification seems like a 'natural' and 'essential' aspect of insurance to people brought up within the liberal tradition, and it also helps explain the passionate commitment of many in the insurance industry and actuarial professions to risk classification. Nevertheless, it is important to be clear that, even within a liberal framework, responsibility only justifies risk classification in general. Whether a particular risk classification is justified depends on the nature of the classification and the meaning and purpose of responsibility which is a complex concept, with multiple and not always consistent meanings (Baker 2002). People can be made responsible ('accountable') so that they will become responsible ('trustworthy, loyal ...'). Or people can be held responsible ('accountable') only when they are in fact responsible (have a causal or controlling role with regard to the risk in question). Like any other governmental rationality, 'responsibility' does not provide a single answer to a social question, but simply a conceptual and institutional approach. Prohibiting Discrimination against Battered Women My final example comes from the controversy surrounding the revelation in the late 1980s that some large U.S. insurers were refusing to sell life, health, and disability insurance to battered women on the grounds that they posed an unacceptably high risk (Hellman 1997). In response to this news, states began enacting legislation prohibiting insurers from discriminating against victims of domestic violence, and bills to do the same on a national level were introduced into the U.S. Congress. The

274 Tom Baker insurers' defence was that excluding battered women from the insurance pool was 'actuarially fair.' According to the insurers, the history of abuse meant that these women were much more likely to make life, health, or disability insurance claims in the future than women who were not victims of domestic abuse. Their higher risk meant that it was 'actuarially fair' to exclude battered women from the insurance pool, just as insurers excluded other unacceptably high risks. This actuarial fairness justification is a somewhat more elaborate version of what I earlier described as the simple fairness, or subsidy, justification in the discussion of age rating in fraternal life insurance. The legal philosopher Deborah Hellman used this controversy as an occasion to examine the ethical basis of insurance risk classification. As she described, the actuarial fairness justification for excluding high risks rests on a view that 'fairness requires that low risk insureds be permitted to join together in insurance pools and thereby benefit from their good health' or other low risk status (Hellman 1997: 398). This view rests, in turn, on the existence of a liberty interest - freedom of association - that is qualitatively distinct from the moral hazard and adverse selection justifications for risk classification. Whether the liberty interest justifies any particular risk classification depends, in Hellman's view, on whether the resulting low-risk status is 'deserved' and, if it is not deserved, whether one accepts the link between desert and entitlement made most famously by Rawls (1971: 101, 104, 311-12). If the low-risk status is deserved, the moral claim to benefiting from that status is a strong one. If the low-risk status is not deserved, the moral claim to benefiting from that status is a weaker one that rests on drawing a clear distinction between desert and entitlement - a distinction often associated with the work of Nozick (1974: 225-6) so that people may be morally entitled to the fruits of attributes that they have done nothing to deserve. Hellman concludes that good health is 'morally arbitrary': the victims of domestic violence do not deserve their 'high' risk status, just as most other people usually do not deserve their life or health risk status. She describes the dispute over the significance of her conclusion as follows: Does the moral arbitrariness of that fact mean that low-risk insureds are not entitled to the benefits that follow from good health? Or instead, is it interference with the low-risk insured's ability to benefit from good health that requires justification? Those struck by the moral arbitrariness of good health are likely to believe that risk rating is unjustified in most cases. Since

Adverse Selection and Risk Classification 275 the healthy do not deserve the benefits health makes possible, they have no legitimate claim of entitlement to them. This rationale supports community rating and single payer schemes.17 Those who are struck by the idea that my talents, fortunes, and experiences are my own and thus that the community needs a powerful justification to interfere with my ability to use and enjoy these traits, are likely to believe that any restraint on my ability to join with whomever I wish in insurance schemes constitutes an infringement of liberty requiring a defense. This rationale supports the utilization of risk rating by private insurers. (402) Thus, debates over the legitimacy of particular forms of risk classification invoke classic debates over the nature of distributive justice. As the insurers' position in the battered women controversy demonstrates, actuarial fairness adopts a version of the libertarian position. Actuarial fairness saddles people with all the consequences of their highrisk status, whether deserved or not. Conversely, it entitles other people to all the benefits of their low-risk status, also whether deserved or not. Yet, on close examination, we can see that actuarial fairness is not in fact grounded on the liberty interests of individuals. As Hellman (1997) and others have noted, proponents of actuarial fairness defend the freedom of insurance institutions to classify, but they do not place individual liberty above the interests of those institutions (see also Daniels 1990). Defending individual liberty above the interests of insurance institutions would mean vesting individuals with a right to accurate risk classifications, and the level of accuracy required would not be within the sole purview of the insurance institutions themselves. Put another way, a truly libertarian approach would obligate insurers to do more than classify applicants accurately within the context of their ordinary business arrangements. It would obligate them to conduct research to determine accurate forms of classification. How much research, and at what cost, would have to be determined, but there is no reason to believe that it would correspond to the amount of research insurers presently determine is appropriate. In practice, the fairness argument has been mobilized in public policy debates not to protect the rights of low-risk individuals, but rather to promote the freedom of insurance organizations to classify insureds through any means they wish. While some 'low-risk' individuals may believe that they benefit by risk classification, any particular individual is only one technological innovation away from losing his or her privileged status - the reality that lies behind the widespread concern with genetic

276 Tom Baker testing by insurance companies.18 If there is a fundamental moral principle at work here, it is not liberty, but utility. Once we are in the realm of utility, liberty is only one value among many, and there is no reason to make insurance institutions the arbiter of that utility, especially given the collective action problem created by risk classification. This is not to say that risk classification is always and everywhere bad, but rather that the fairness justification for classification does not carry all the moral force that its proponents assert. Like Hellman, I do not pretend to be able to resolve the debate over the morality of risk classification. My purpose in invoking it is to demonstrate that risk classification involves substantial moral commitments. Classifying insureds according to risk both reflects and creates a moral vision. Risk classification reflects a commitment to individual responsibility, recognizing that what that means is up for debate. Risk classification also creates that commitment by helping to persuade people that the purpose of insurance is individual protection and, accordingly, that the insurance group is a collection of individuals without any responsibility to one another (cf. Baker 2002). Seen in this light, it is hardly surprising that many friendly societies and other fraternal insurance organizations long resisted risk classification, and that strong forces are arrayed against health insurance risk classification today (Stone 1993, 2002). Conclusion These are hardly the last words on either adverse selection or the challenge that it poses to the promise of insurance. My intent, so far only imperfectly realized, is to expose the concept of adverse selection, and the technologies that relate to it, to the kind of genealogical analysis previously accorded to the related concept of moral hazard. As with work on moral hazard, the objective is to demonstrate how the rhetoric of adverse selection disguises ideological or moral commitments. The goal is not to discredit the concept itself, but rather to clear away some of the underbrush that inhibits addressing very real problems in the design of socially responsible insurance institutions. In particular, much of the literature on insurance treats risk classification as an inevitable, essential response to the problem of adverse selection and ignores its role in promoting that problem. The literature often takes for granted the following familiar belief structure. 'Private' insurance is better than 'public' insurance. 'Voluntary' insurance is better

Adverse Selection and Risk Classification 277 than 'mandatory' insurance. The following statement by the then president elect of the American Academy of Actuaries to the U.S. Commission on Civil Rights is typical: 'The basic principle is quite simple to state: The risk classification process is essential to the viability of private, voluntary insurance mechanisms. Where substantive differences in risk of loss exist, they must be recognized in a private, voluntary insurance mechanism to avoid anti-selection by those subject to high risks against those subject to low risks ... It is at best questionable, and more likely, impossible, that broad social cost-spreading objectives can be accomplished through voluntary, private market mechanisms' (U.S. Commission 1979: 141). These are arguments in favour of particular interests, within the context of a particular moral vision. The particular interests are those of insurance institutions seeking maximum autonomy from societal control. The particular moral vision is that of 'actuarial fairness' - a watereddown form of liberalism that privileges individual interests over the common good and that privileges, above all, the interests of insurance institutions organized on its terms. Acknowledgments Thank you to Timothy Alborn, Aaron Doyle, Brian Glenn, Francis J. Mootz, David Moss, and Marianne Sadowski for helpful comments on an earlier draft and to participants at the Risk and Morality Conference for stimulating discussion.

Notes 1 In this paper I will not challenge the empirical claims made in the name of adverse selection, although, as work in progress by Peter Siegelman reflects, there is considerable evidence that adverse selection plays a much smaller role in many insurance pools than theory would suggest. 2 An economist might take exception to this statement because, at least in theory, there are circumstances in which adverse selection by insureds is so strong that a market may not be possible without classification (Rothschild and Stiglitz 1976). In such a circumstance, we might think of classification as 'containing' the promise of insurance in the sense that a container makes it possible for water to be carried from one place to another. 3 A decision to market insurance to a given target audience classifies that

278

Tom Baker

audience as being composed of members with favourable risk characteristics. For example, Medicare HMOs are prohibited from underwriting (i.e., turning away sick applicants), but they are free to design their marketing so that it appeals to healthy, active seniors. Billboards featuring seventy-fiveyear-old men doing gymnastics and free health club memberships are two ways to do this. Compared to risk-based pricing and underwriting, this is not an effective way of separating a population into risk-based groups; nevertheless, it can perform that function. Contract drafting can also serve a risk classification function. A decision to offer a given type of coverage as an 'extra' rather than as a standard coverage provided by a broad form policy can reflect a judgment that the insurance company cannot identify (classify) individuals who are particularly risky with respect to that type of coverage. Thus, requesting the extra coverage amounts to self-classification as being risky in that way. Two examples are offering mental health coverage as an extra in health insurance policies and sexual harassment coverage as an extra in liability insurance policies. Alternatively, offering a given type of coverage - for example, coverage for home business pursuits under a homeowners' policy - as an extra may reflect a judgment by the company that it is less expensive to ask people to self classify. 4 An additional reason pointed out by Kenneth Abraham is path dependence: insurers have collected risk information based on risk categories chosen in the past. While there may be better risk indicators than those relied upon by insurers, that cannot be known without the information that is developed only once insurers decide to classify risk on the basis of those indications (Abraham 1986). See also Norman Daniels (1990). 5 Carol Heimer has described this insurer-side adverse selection as a form of 'moral hazard' (2002). Although her use of the term is consistent with its early 'immorality' meaning (Baker 1996), it is not consistent with its use in the economics literature. There is value in maintaining a consistent meaning for concepts across the fields of economics and sociology. In the economics of insurance, the term 'moral hazard' has been used for changes in incentives that result from an insurance or similar relationship, while the term 'adverse selection' has been used for behaviour that results from differences in risk status. 6 Most health insurance in the United States is received as an employment benefit. Employment-based health plans offer the same benefits to all qualified employees, at the same cost. Although the health benefits are not mandatory, because of U.S. income tax policy and other reasons, they are typically offered on terms so favourable that all employees who do not have insurance from another source (e.g., from the health plan of a spouse),

Adverse Selection and Risk Classification

279

choose to accept them. When employment groups offer employees a choice of health plans, some adverse selection can occur when low-risk employees disproportionately choose particular plans. 7 In the large group market - which is principally employment-provided health insurance - insurance is widely available and comparatively inexpensive. In the individual (and small group) market, insurers practise extensive underwriting and risk rating; yet insurance is comparatively expensive even for the 'low risk' small groups and individuals (General Accounting Office 1996). 8 Note that I am discussing personal lines insurance here, not commercial insurance for large corporations, which provide many alternatives to traditional insurance. 9 Daniels (1990) reports that New York State is one such example. 10 For examples of articles disparaging traditional friendly societies and promoting the 'scientific' friendly societies, see Tuffnell (1842), Chambers (1853,1859). 11 The moral hazard of unemployment insurance results from the partial but substantial control that employers and workers have over unemployment. The existence of unemployment insurance can change the incentives of employers to retain workers, as well as the incentives of workers to be retained. Adverse selection also results in significant part from this same control. Because low risks can prevent loss in at least ordinary economic conditions by limiting the number of lay-offs, they are likely to drop out of an insurance market that does not offer them a low price. Yet, an unemployment insurer would find it very difficult to segregate individual employers by their ability or willingness to control unemployment, with the resulting 'lemons problem' described so aptly by the economist George Akerlof (1970). Akerlof analysed a hypothetical situation in which used car buyers face a market composed of 'peaches' and 'lemons' and are unable to determine the status of any individual car. In that situation, the most that a rational buyer will pay is the average price, which is less than a peach is worth. So, owners of peaches will tend to keep them, and the car market then becomes disproportionately composed of lemons, so that people will pay even less for cars, driving even more peaches out of the market, and so on. In practice, the used car market does not unravel in this way because people do have ways of evaluating whether a car is a peach or lemon. Unemployment insurers could make the same judgment about prospective purchases of unemployment insurance. However, unemployment insurance presents a more difficult lemons problem than used cars, because of the interaction of adverse selection and moral hazard. In the case of unemployment insurance, the sale

280 Tom Baker is only the beginning of the relationship. Purchasers of insurance who have behaved like peaches in the past may start to act like lemons once the insurer picks up the costs of unemployment (moral hazard), making it doubly difficult to know whether a prospective insured is 'really' a peach or a lemon. 12 David Moss's research suggests that, at least in the United States, concerns about moral hazard explain the motivation of U.S. reformers better than concerns about adverse selection. See Moss (1996: 59). 13 See Beider (1987) for the application of the 'fair opportunity' principle of distributive justice to insurance risk classification. This principle 'requires that the allocation of benefits and burdens reflect some relevant characteristics, which all have had the opportunity to acquire, of the individuals involved' (ibid.: 66). 14 It is worth noting that this example illustrates how difficult it can be in practice to separate moral hazard and adverse selection. 15 For me, (3) is the most controversial step in this chain of reasoning, treading on the always difficult question of 'what is a cost of what.' Why unemployment should be regarded as always and everywhere a cost of production is a mystery, particularly given the 'creative destruction' of capitalism. For example, it seems odd to assign to one form of production the costs of unemployment that results from the emergence of a competing form of production. Yet, assigning those costs to the emerging form of production would inhibit the development of superior technology, which is one of the main benefits of capitalism. What makes (1) debatable includes the use of the term 'free market,' since markets are shaped. What makes (2) debatable includes the enormous complexities hidden by the deceptively simple idea of the 'costs of production' as well the limits on the market highlighted by, inter alia, the economics of information and institutions. Finally, as (4) depends on (1), (2), and (3), it is obviously debatable as well. 16 Cf. O'Malley (2002) and Rose (1999). I do not wish to enter here into the debate over whether there is any significant difference between 'liberalism' and 'advanced liberalism.' Since the nineteenth century one of the core characteristics of liberalism has been this emphasis on individual responsibility (Ewald 2002). 17 'Community rating' is the practice of charging all the participants in a health plan the same rate (a 'community' rate); 'single payer' refers to the practice of allowing only a single, monopoly insurer. As discussed above, this is a common approach to the adverse selection problem. Most Western countries go one step further and mandate participation in the national health plan.

Adverse Selection and Risk Classification

281

18 The fairness argument in favour of age rating for sickness insurance would also justify genetic-based risk classification. Indeed, the subsidy claim is stronger in the case of genetic classification than it is for age classification: everyone gets old, while genetic make-up is fixed at conception. Risk classification on the basis of genetic testing is widely regarded as repugnant (witness the many statutes prohibiting that practice), demonstrating the existence of an at least implicit counter-vision of 'fairness.' References Abraham, K. 1986. Distributing Risk. New Haven: Yale University Press Akerlof, G. 1970. 'The Market for "Lemons": Quality Uncertainty and the Market Mechanism.' Quarterly Journal of Economics 84: 488-500 Ansell, C. 1835. A Treatise on Friendly Societies. London: Baldwin and Cradock Baker, T. 1996. 'On the Genealogy of Moral Hazard.' Texas Law Review 75: 237-92 - 2000. 'Insuring Morality.' Economy and Society 29: 559-77 - 2002. 'Risk, Insurance and the Social Construction of Responsibility.' In T. Baker and J. Simon, Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Baker, T, and J. Simon. 2002. 'Embracing Risk.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Becker, J.M. 1972. Experience Rating in Unemployment Insurance: Virtue or Vice. Michigan: W.E. Upjohn Institute Beider, Perry C. 1987. 'Sex Discrimination in Insurance.' Journal of Applied Philosophy 4: 65-75 Beveridge, W. 1930. The Past and Present of Unemployment Insurance. London: Oxford University Press Burchell, G., C. Gordon, and P. Miller, eds. 1991. TheFoucault Effect: Studies in Governmentality. Chicago: University of Chicago Press Chambers. 1853. 'Friendly Societies.' Chambers Journal 19: 199. - 1859. 'Friendly Societies.' Chambers Journal 32: 355-6 Clawson, M.A. 1989. Constructing Brotherhood, Class, Gender, and Fraternalism. Princeton: Princeton University Press Daniels, N. 1990. 'Insurability and the HIV Epidemic: Ethical Issues in Underwriting.' Millbank Quarterly 68: 497-525 Emery, J.C.H. 1996. 'Risky Business? Nonactuarial Pricing Practices and the Financial Viability of Fraternal Sickness Insurers.' Explorations in Economic History 33: 195-226 Ericson, R., D. Barry, and A. Doyle. 2000. 'The Moral Hazards of Neo-Liberal-

282 Tom Baker ism: Lessons from the Private Insurance Industry.' Economy and Society 29: 532-58 Ewald, F. 2002. 'The Return of Descartes's Malicious Demon: An Outline of the Philosophy of Precaution.' In T. Baker andj. Simon, eds., Embracing Risk. Chicago: University of Chicago Press General Accounting Office. 1996. Private Health Insurance. GAO/HEHS 97-8. Washington, DC: General Accounting Office. Gilbert, B. 1965. 'The Decay of Nineteenth Century British Provident Institutions and the Coming of Old Age Pensions in Great Britain.' Economic History Review 17: 550-63 Hardwick, C. 1859. The History, Present Position, and Social Importance of Friendly Societies. London: Routledge Heimer, C. 1985. Reactive Risk and Rational Action: Managing Moral Hazard in Insurance Contracts. Berkeley: University of California Press - 2002. 'Insuring More, Ensuring Less: The Costs and Benefits of Private Regulation Through Insurance.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Hellman, D. 1997. 'Is Actuarially Fair Insurance Pricing Actually Fair?: A Case Study in Insuring Battered Women.' Harvard Civil Rights-Civil Liberties Law Review 21: 355-411 Jacques, D.R. 1849. 'Society on the Basis of Mutual Life Insurance.' Hunt's Merchant Magazine and Commercial Review 16: 152, 153 King, G. 1876. 'On the Mortality Amongst Assured Lives.' Journal of the Institute of Actuaries 19: 381-405 Lippincott, H.C. 1905. 'The Essentials of Life Insurance Administration.' Annals of the American Academy of Political and Social Science 26: 192-208 Moss, D. 1996. Socializing Security: Progressive-Era Economists and the Origins of American Social Policy. Cambridge: Harvard University Press Nozick, R. 1974. Anarchy, State and Utopia. New York: Basic Books O'Malley, P. 2002. 'Imagining Insurance: Risk, Thrift, and Life Insurance in Britain.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Quarterly Review. 1864. 'Workmen's Benefit Societies.' Quarterly Review 116: 318-50 Rawls,J. 1971. A Theory of Justice. Oxford: Oxford University Press Rose, N. 1999. Powers of Freedom. Cambridge: Cambridge University Press Rothschild, M., andj. Stiglitz. 1976. 'Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information.' Quarterly Journal of Economics 90: 629—49

Adverse Selection and Risk Classification 283 Stone, D. 2002. 'The Moral Opportunity of Insurance.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press - 1993. 'The Struggle for the Soul of Health Insurance.' Journal of Health, Politics, Policy and Law 18: 287-317 Tuffnell, E. 1842. 'Improved Friendly Societies.' Penny Magazinell: 387 U.S. Commission on Civil Rights. 1979. Discrimination Against Minorities and Women in Pensions and Health, Life & Disability Insurance. Washington, DC: U.S. G.P.O. van Amringe.J.H. 1873. 'Life Assurance.' The Galaxy 15: 249-54, 396-401, 527-35. Serialized in February, March, and April 1873 issues Weisbrod, C. 2000. 'Insurance and the Utopian Idea.' Connecticut Insurance Law Journal^: 381-422 Works, R. 1998. 'Excusing Nonoccurrence of Insurance Policy Conditions in Order to Avoid a Forfeiture: Claims-made Formats as a Test Case.' Connecticut Insurance Law Journal 5: 506-674

11 Insurers as Moral Actors CAROL A. HEIMER

Scholars disagree sharply about what exactly insurers are accomplishing. Three competing visions about how insurers shape behaviour dominate current scholarly work on insurance. Seen through an economic lens, insurance is about the incentives built into an insurance contract (Arrow 1963, 1968; Pauly 1968, 1974; Shavell 1979; Holmstr0m 1983; Stiglitz 1983; Heimer 1985; for a more extensive review of this literature, see Baker 1996). A key insight of this perspective is that insurance contracts alter incentives, and that insurers, whether they be private insurers or governments, must therefore guard against the deleterious effects of moral hazard. Although both insurers and policyholders are subject to these effects, theorizing has focused almost entirely on incentive effects on clients. Insurance and risk management are also seen as disciplinary systems in which insurers regulate behaviour at the same time as they spread risk (Heimer 1985; Simon 1988; Ewald 1991; O'Malley 1999). In this view, insurance is part of the long, general process, described by Foucault (1977), of substituting discipline for punishment, a process that reduces the economic and social costs of social control while reaping the benefits of extending discipline into schools, factories, military institutions, and the like. Insurers make several distinct contributions to this process. First, they work as the disciplinary agents of the rest of society, curbing our tendencies to take risk by mandating risk-reducing routines and building an infrastructure that reins us in, in all sorts of ways, so that social life becomes safer and more predictable. Second, although they do this in the interest of avoiding liability and keeping claims low, the net effect is that much of the work of governing is displaced from political systems where legislation is debated to administrative ones where regula-

Insurers as Moral Actors 285 tions are issued by insurer fiat. Third, by using actuarial techniques to predict and compensate for accidents and other statistically deviant behaviour, insurers reduce the social costs of non-normative behaviour. By facilitating adaptation to risk, actuarial techniques make the punishment, control, or reshaping of abnormal or unacceptable behaviour less necessary. All three of these innovations - social control engineered into an unobtrusive infrastructure, rule making by administrators or private insurers, and rates and compensation that depend on actuarial predictions - defuse opposition. The techniques of resistance and opposition appropriate to legislative rule making do not work well in contesting insurance rules. By making the distribution of costs and benefits dependent on traits that may or may not be shared by other members of a person's social group, actuarialism undermines people's sense of a shared fate and therefore their capacity for collective opposition. By obscuring targets and undermining alliances, insurers make opposition difficult and so reduce the costs of governance. What might have been political becomes merely administrative - governance. But all of this assumes that the administration of insurance is relatively inexpensive, that insurers and their agents are up to their assigned tasks, and that they themselves can be disciplined. Although insurance tends to displace social control from the political realm to the administrative, questions about system design bring insurance back into the political realm. Insurance may be a technology of discipline, of risk management, but it is also a technology of distribution, of risk sharing. Questions about who is included and who excluded from the risk pools created by insurance inevitably end up in the political arena. The moral opportunities of insurance shape our understanding of who is part of the community and therefore whom we are obligated to help. Under some circumstances our sense of community is enlarged either by including new people in the community or by accepting a communal obligation to help with adverse events that were once conceived as solely individual burdens (Stone 1999; Baker 2002). But whether opportunities to enlarge the community and include previously excluded groups are seized may be another matter. At a minimum, we need to investigate the conditions under which public statements endorsing a more inclusive view in fact translate into changes in policies. These three competing visions of insurance are summarized in Table 11.1. At the same time that scholars are enlarging their understanding of what insurance does, a few have begun to stretch the paradigm in

TABLE 11.1 Insurance and policyholders: Comparisons of economic, disciplinary, and political approaches Insurance as incentive system

Insurance as disciplinary, governmental, and actuarial system

Main insight

Incentives change with insurance contract

Insurance technology is based on governmentality and actuarial techniques that constitute further developments in the sequence of disciplinary systems described by Foucault. The routines of insurance embed risk management and risk reduction in infrastructure. Actuarial techniques emphasize adaptation to risk (through estimation and compensation for losses) more than punishment of deviance or control/ reshaping of abnormal or unacceptable.

Focal actor

Policyholder

Policyholder, but society generally because there are spillover effects

Insurance as political system Insurance is a disciplinary, risk management system. To reduce contention, social control has been removed from political system to administrative sphere. Insurance is a distributive, risksharing system. Insurance creates a moral opportunity to consider who is included in the community and what kinds of risks will be shared.

Policyholders who were eligible for coverage and potential policyholders who have been excluded

TABLE 11.1 (concluded) Insurance and policyholders: Comparisons of economic, disciplinary, and political approaches Insurance as incentive system

Insurance as disciplinary, governmental, and actuarial system

Insurance as political system

Advocated actions

Design contract to eliminate incentives for shirking

1 When the objective is to reduce costs of resistance: Design social control into the insurance infrastructure. 2 When the objective is to protect political consciousness of citizens: Avoid using politically or morally meaningful categories as basis of insurance rating because this de-moralizes them.

1 Bring decisions about insurance back into political system. 2 Enlarge community of those covered; increase scope of coverage (more risks covered and covered more fully).

Focal type of insurance

Commercial insurance, but applies to social insurance as well

Social insurance programs sponsored by governments, but also applies to commercial insurers and even other entities that employ actuarial techniques

Social insurance programs especially, but applies to commercial insurance as well

288 Carol A. Heimer another way. Until very recently scholarly work on insurance has focused almost exclusively on what insurers do to policyholders. The analysis often makes exclusions and oppressive conditions in the contract seem inevitable. Governments, like insurers, also make things seem inevitable, but as political parties are more able to challenge or change those apparent inevitabilities, alternative views are more readily available in the culture at large. And although the disciplinary and political visions of insurance have moved well beyond the academically reputable version of insurer visions represented in the economic perspective, they have retained the focus on policyholders. Because we cannot understand what insurers do and how they do it without a fuller understanding of who insurers are, a deeper examination of insurers themselves is needed. With that objective, a few scholars have begun to ask whether classical analyses of the moral hazard problem should be revised to consider how the insurance contract modifies the incentives and behaviour of insurers as well as of policyholders.1 This chapter tries to bring these two developments together. It offers a portrait of insurance that scrutinizes insurers as well as policyholders and considers them as complex organizations.2 In their interactions with policyholders, other insurers, organizations that set standards and perform inspections, organizations that mandate insurance coverage as a condition of doing business, and governmental bodies such as legislatures and regulatory agencies, insurers perform work as actuaries, underwriters, ratemakers, and claims adjusters. They also help build the institutions of the financial world and shape and respond to regulations. Although insurers are agents who act on behalf of groups of policyholders, they also have interests and perspectives that are partly distinct from those of policyholders. Like policyholders, insurers and their employees have utility functions, are disciplined by themselves and others, and have moral sensibilities. What we need, then, is a theory of the economic, disciplinary, and moral-political formation of the class of 'agents of the group' that insurance schemes of various kinds produce. Insurance and Incentives

At its most basic, insurance is a social arrangement to reduce the effects of losses by employing the resources of the group to cushion individuals. The key tasks of insurers are to organize the insurance pools, turn them into groups with a common fate, and act as agents of these groups. In commercial insurance this is done by charging a premium to each

Insurers as Moral Actors 289 policyholder to create the fund from which those who experience losses are compensated; recipients of compensation are those who contributed to the fund or their designated beneficiaries. In government-run social insurance schemes, the fund is instead created through taxation and the group eligible for compensation may not be the same as the group contributing to the fund; the group of recipients might be smaller or larger than the group of contributors. This rudimentary framework is sufficient to create the shift in incentives that has long worried insurers. Because they no longer bear the full burden of losses, policyholders have a diminished interest in preventing losses. For example, in commercial insurance, owners who have purchased theft policies may be a tad less careful about preventing thefts. A firm whose operations are covered by business interruption insurance may be less motivated to get the enterprise back on its feet. Declines in the value of Thoroughbred horses in the 1990s led to an increase in unexplained deaths of horses; the horses were covered by livestock mortality insurance (Kimball 1992: 195-6). People covered by health insurance, whether it is a private scheme such as Blue Cross or a public one such as Medicare, may 'overuse' the services of physicians. Health insurers worry about how to control 'malingering.' Some policy makers believe that these incentive effects extend even into the realm of childbearing. According to one policy argument much ridiculed by critics (Roberts 1995, 1997: 202-45), women whose welfare payments will be increased for each new addition to the family may be less vigilant in their use of contraceptives. So much for alterations in the incentives of the insureds. The formation of the insurance contract leads to a shift in the incentives of the policyholders, but a second step is required before the incentives of insurers shift as well. As long as the insurer is merely the insurance pool and has no separate corporate existence, it will have no interest separate from that of the policyholders. Once the insurer has a separate existence, however, it has an interest in protecting the funds it has collected, for instance, so that profits can be distributed to stockholders or the jobs, salaries, and bonuses of those employed by the insurer preserved. The gulf between policyholder and insurer interests should be smaller in mutuals than in stock companies; although mutuals will not need to look out for the interests of stockholders, they will have employees to protect. Government-run schemes are not immune to these incentive shifts career civil servants know that their jobs depend on husbanding the government's resources. In the example of health insurance used above,

290 Carol A. Heimer contractual provisions such as co-pays and deductibles allow insurers to avoid paying the full cost of 'losses.' And child exclusion rules may punish careless contraception among welfare recipients by prohibiting coverage of subsequent children, a practice euphemistically referred to as 'family caps' (contraceptive accidents are of course not punished among the employed and married, who still receive tax credits and 'free' public education for subsequent children). Only very rarely does an insurer ask whether policyholders have uncompensated losses or a civil servant worry that some poor families are not getting the assistance to which they may be entitled. These changed and conflicting interests infuse the operation of insurance from start to finish and affect the behaviour of both policyholder and insurer. Insurance, Knights and Vurdubakis (1993) remind us, was not always distinct from wagering. Many of the refinements of insurance technology have had the objective of controlling moral hazard and reducing, if not eliminating, the impulse to gamble with insurance. In Britain, it was only after the 1774 Gaming Act that policyholders were required to demonstrate an insurable interest before purchasing insurance. Before that time, insurers happily profited from selling policies to those who wished to gamble on the lives of strangers, apparently without too much concern about whether insurance contracts encouraged homicide. Both policyholder and insurer benefited from this practice, insurers by making profits on policies they would not otherwise be able to sell and policyholders by receiving insurance payments without any pretence that they were recouping a loss. Analysts of this historical practice have tended to forget that insurer and policyholder conspired together. Instead of regarding the 1774 Gaming Act and similar laws as controlling both insurer and policyholder moral hazard, analysts have commented favourably on the new controls on immoral policyholder behaviour without discussing how the rules prohibited unscrupulous insurers from selling inappropriate insurance policies. Before they can form a contract, the two parties have to find one another. Each has some interest in representing itself as a more attractive partner than it actually is and in avoiding contracts with unattractive partners. Writings on adverse selection remind us that a potential policyholder's interest in gaining insurance coverage varies directly with his or her need for insurance — essentially with the person's assessment of the likelihood of experiencing an accident or other covered loss. People who are in poor health want health insurance because they think they are likely to need medical care. They may then misrepresent them-

Insurers as Moral Actors 291 selves to insurers, concealing previous illness or worrisome symptoms, in order to secure coverage. Insurers then engage in a defensive strategy of trying to uncover evidence that people may be in poor health now or are likely to be in the future.3 But insurers also misrepresent themselves. Because their income depends on selling insurance, they need buyers. Insurers need to create a demand for their products either by convincing potential policyholders that they will be better off if they buy the product (you cannot be a responsible parent unless you have sufficient life insurance to provide for your child's education should you die) or by making the purchase of their product a prerequisite for something else. Insurers do both of these things, creating demand through advertising and institutionalizing insurance as the ticket to homeownership and other elements of a middle-class life. Examining the influence of insurance advertisements on court decisions about what is included in the insurance contract, Baker (1994) argues that in making sweeping claims that a policyholder will 'own a piece of the rock' or 'be in good hands,' insurers are offering something different from what is described in the fine print of the insurance contract. Indeed, the fact that the contract is designed exclusively by insurers rather than being an outcome of a negotiation between insurer and policyholder offers another temptation for insurers. Knowing that policyholders will not be able to see the fine print until after they have signed (see Stone 1994; for a different view, see Kimball 1992: 6-8), insurers are relatively free to describe the policy however they like without being challenged by potential policyholders.4 This problem of ambiguity of contract provisions is sufficiently pervasive that it is recognized in one of the key principles of contract interpretation, the contra proferentum principle. According to this principle, when a contract can reasonably be interpreted in more than one way, the court should prefer the meaning that operates against the interests of the party who supplied the words.5 Analyses of the development of concepts such as insurable interest and adverse selection, or of practices such as requiring deductibles, focus almost exclusively on policyholder incentives, however, overlooking the necessity of manipulating and controlling insurer interests. In fire insurance, for instance, the problem of excessive face values is presented primarily as a matter of controlling the desire of policyholders to reap windfall profits after a fire, rather than controlling the impulse of insurers to bring in large premiums or of agents to garner higher commissions. That an insurance agent and the company collected

292 Carol A. Heimer 'surplus' commissions and premiums, and then paid not the amount for which it was insured but only the lower value of the building lost, is exactly the same kind of moral hazard. In the usual accounts, insurers are innocents who had to learn the hard way about policyholder moral hazard, not morally frail organizations whose greed had also to be controlled. At best, a line is drawn between insurer and insurance agent, and the moral failings of agents selling overvalued policies are acknowledged while those of insurers are minimized. Kimball's (1992) textbook on insurance law illustrates the point. In addition to a concentrated six-page discussion (134-9) of moral hazard, Kimball also touches on the subject directly at nine other points in the text and elsewhere alludes to incentive issues without specifically identifying them as moral hazard. He devotes nearly fifty pages (486-533) to insurer bad faith, discussing bad faith doctrine in first- and third-party cases, punitive damages, and the like, and a smaller section to agent misconduct (103-5). Interestingly, he makes no connection between moral hazard, a problem of policyholder intentions and behaviour, and 'bad faith,' which is a very similar problem of insurer intentions and behaviour. The few scholars thinking about how insurers might be affected by the incentives embedded in the insurance contract have usually assumed that insurer and policyholder incentives were largely mirror images of one another. They have argued that policyholders would be less motivated to prevent covered losses and interested in recovering as much as possible, and that insurers would try to force policyholders to remain vigilant and deny or skimp on reimbursement. As I have shown above, these suggestions seem entirely reasonable, but they do not go far enough. They fail to recognize key differences between insurers and policyholders. I provide a fuller account of insurer incentives below, pointing out discrepancies between the simplistic depiction of insurers in traditional accounts of the insurance contract and the much richer range of roles they play in the real world. Insurers Are Financial Institutions as Well as Risk Managers What do insurers mainly get out of the contract? Insurance is typically described as a contract in which insurers collect premiums from policyholders, promising to compensate them when pre-specified kinds of losses occur. The theory is that insurers make their money from charging just a bit more than the policyholders as a group are expected to

Insurers as Moral Actors 293 lose. Because insurer profits depend on the difference between the premiums collected and the costs of policyholder losses, insurers should be motivated to help policyholders figure out how to keep losses down which is a good thing, given that insurance compensation covers only the monetary losses experienced by policyholders, and often only a portion of those. But insurer incentives are in fact somewhat different than depicted in this account. Commercial insurance is as much a financial institution as a loss-shifting, risk-management one, and investment profits are more important to insurers' bottom line, and therefore to stockholders, than are underwriting profits. Generally speaking, a high premium/high loss regime should be preferable to insurers than a low premium/low loss regime, given that insurers are better off having substantial premiums to invest during the interval between receipt of premium and payment of compensation. The balance in commercial insurance between the business of insuring and the business of investing surely varies from one form or organization to another, with stock companies experiencing more pressure than mutuals to attend to the bottom line. In this regard, the current boom in 'demutualization'6 is significant, as is the dismantling of the barriers between banks and insurance.7 And the balance also surely varies from one time period to another. During stock market booms, the lure of investment profits is undoubtedly more important than it is during periods when returns on investments are more modest. We should be careful, too, not to overstate insurers' capacity to shift the organizational focus - government regulations requiring reserves have, after all, placed limits on the magnitude and type of investments that insurers are permitted to make with policyholders' money. Because the funds for governmental insurance programs come from taxation, and often from current accounts rather than investment income, these arguments apply much less to them than to commercial insurers. Nevertheless, the principle is the same: insurers typically are in some business other than insurance and are beholden to some group other than those covered by insurance. (Of course, policyholders have more to their lives than their relations with insurers, and that helps to explain why they may not be swayed by insurers' attempts to shape behaviour.) It may be investment income and stockholder interests that drive commercial insurers, but the well-being of governmental insurance agencies depends much more on keeping taxpayers happy. In the political climate of the United States, politicians and therefore civil servants seem to believe that taxpayers want to limit the cost of social

294 Carol A. Heimer insurance, particularly when payers and recipients tend not to come from the same social group. Thus American government insurance programs cut costs by limiting the duration of coverage, shrinking the pool of those eligible for assistance, and requiring evidence that the recipient is doing her best to reduce 'losses' by applying for jobs or attending educational programs. When the pool of recipients more nearly matches the pool of contributors (at least in the long run), as is the case with Medicare and Social Security, politicians and civil servants believe that taxpayers will be more generous. I argued above that insurer incentives shift when the insurer acquires a corporate existence that provides it with interests different than those of the pool of people covered by the insurance scheme. This can happen in more than one way. For instance, a separate corporate identity can be created through the formation of a stock company or through the formation of a government agency funded by one group (taxpayers) but benefiting another (welfare recipients). And the interests of the parties can diverge to varying degrees, depending for instance on the state of the stock market or the extent to which payers see themselves as members of the same community as recipients. Insurers Are Complex Organizations, Not Unitary Actors, So It Is Not Clear Who Is a Party to the Contract Additional issues are raised when we recognize the full complexity of insurers. Who exactly is experiencing the incentive? A second deviation from the standard depiction thus arises because insurers are organizations, not people. Because an organization is not a unitary actor, it cannot be expected to have a single utility function or a consistent and coherent reaction to any given situation. Insurers are organizations composed of elements with conflicting interests and staffed by people whose needs may differ from those of the organization. Although in principal/agent theory, insurers would be described as agents acting on behalf of policyholders, we need to acknowledge that insurance is a system of nested agency relations. Insurers may be the agents of the policyholders, but they also are principals employing agents to market their insurance policies and settle claims. As principals, insurers sometimes find that they have agents they cannot fully trust, so they build incentive systems to motivate these agents, discipline them with training sessions, and construct appropriate moral environments for them. And of course these incentive systems sometimes have unintended consequences.

Insurers as Moral Actors 295 'Insurers' can claim that they do not sell homeowners policies with excessive face values, all the while giving commissions to a sales staff, some of whom do exactly that because it brings higher income. Agents naturally have an interest in selling excessively large policies when their commissions vary directly with the size of the premium, and therefore with the face value. Because the company, not the agent, ultimately pays the claim and pays only the amount lost, the agent has little reason to worry. The slippage may even be useful, because the company, which can claim not to have full control over its agents, may get to invest those surplus premiums. Policyholder confusion about the differences among brokers, soliciting agents, and local recording agents, and the degree to which each of them is controllable by insurers, all work to the advantage of insurance companies. The same is true of the hard bargaining of claims adjusters, who can use threats of delay and attorneys' fees to persuade the claimant to accept less in compensation than was actually lost. Insurers are ordinarily free to create incentives, for example, to motivate their insurance adjusters to reduce claims. Some adjuster techniques may be at the edge of legality, if they are not outright illegal. Insurers take advantage of the fact that they never approved illegality, only motivated it. Likewise insurers can gain an additional advantage by postponing payment of agreed or court-ordered claim values until claimants take some further action to force payment. But this too depends on ambiguity about whether postponement results from the sloth of employees in the accounts payable department or from an explicit company policy of indefinite delay. In both of these examples, not being a unitary institution can be turned to advantage. Moreover, insurance and risk management are not always secured solely through formal insurance. Insurance may instead be arranged by insurers working in conjunction with the risk managers of policyholder organizations, who may self-insure, insure through a captive insurer, and attempt to reduce or off-load risk through corporate policies. In such cases, only a few risks may be covered through formal insurance. A selfinsurer is a special kind of agent for the group to which it becomes liable; strategic bankruptcy is a special device for such insurers. This argument about the fragmentation of complex organizations applies to both commercial and governmental insurance. As I discuss below, the division of labour within an organization means that the actions of routine-following subordinates may not correspond very closely to policy statements of CEOs. Policy statements do not typically result in

296 Carol A. Heimer the overhaul of a functioning organization whose work can be done only by carefully orchestrating the activities of a large number of people. Policy statements may be fully decoupled from action or they may result in some tweaking. Only very rarely will they lead to anything more.8 Insurance Is a Multi-Party Contract, Not a Two-Party One Who are the parties to the insurance contract, and who, therefore is arranging the terms that provide the incentives to insurers and policyholders? Often insurance contracts are more properly described as bargains between insurers and third parties such as the employers who arrange for and subsidize health or life insurance, the physicians or dentists who supply the care, or the auto body shops who repair cars after accidents. In principle, policyholders could form coalitions with service providers. In fact, coalitions between insurers and service providers are probably more common, because service providers and insurers are both repeat players. A coalition between them is thus more profitable than a coalition between either of them and individual policyholders. Insurer literature on insurance fraud would suggest otherwise, but that is partly because insurers see nothing amiss in their coalitions with service providers. From the perspective of policyholders, though, there is something wrong with this coalition. A discussion of health insurance illustrates the point. The 'losses' covered by health insurers are the cost of health care - bills for the services of physicians, nurses, and therapists; bills for hospital stays or for the use of hospital rooms and equipment for outpatient procedures; the costs of medications and equipment to be used at home. The policyholders' objective is health, and they have a strong desire to avoid even reimbursable 'losses.' Insurers have conflicting interests here. On the one hand, they want policyholders to keep 'losses' low by consuming relatively little health care. On the other hand, they want people to worry about potential 'losses' enough to ensure that they sign up for insurance and pay their premiums. The interests of health care providers vary with insurance arrangements. With fixed annual premiums, the interests of insurers and health care providers are aligned. Both want to discourage consumption of health care by subscribers. But when the contract retains a fee-for-service element, health care providers will prefer that policyholders consume a great deal of health care. Both health care providers and insurers should be interested in convincing people that they need health care of the sort

Insurers as Moral Actors 297 that can be supplied only by professionals - acute care or expensive preventive care such as colonscopies, mammograms, or EKGs, for instance, rather than inexpensive preventive care like exercise or smoking cessation classes. Reimbursing for acute care but not preventive measures recruits physicians as insurers' allies in the fight to retain control of the market in the face of threats of government regulation and government provision of health care, but it typically increases physician fees rather than health care delivery. Thus physicians and hospitals often become agents of the insurer or government health plan instead of acting simply as agents of the policyholder. Instead of simply using their expertise on behalf of the policyholder, these professionals function as enforcers of 'loss reduction' on behalf of the insurer and follow insurer protocols as well as medical protocols. In effect, then, the insurer becomes a more complex actor in two senses: it extends its reach well into the sacrosanct relation between patient and physician (or analogous relations in other arenas) and the insurer now acts on behalf of physicians, with whom it has formed an alliance, as well as on behalf of its policyholders. Insurance Is an Ongoing Strategic Game At what point do these incentive effects go into force? Insurance contracts give both parties an incentive to cheat, or at least not to comply fully with the original agreement. But these strategic considerations pervade each stage of the unfolding insurance contract - the early application, offer, and negotiation of the contract; the subsequent actions of the policyholder, now protected from some of the financial consequences of a possible loss; the investigation following a reported loss; the bargaining about exactly how much a policyholder will be compensated when a loss occurs. We should therefore think of insurance as an ongoing strategic interaction, a game in which insurers and policyholders both continue to adjust to each others' moves, and in which insurers are complex actors whose components may also engage in strategic interaction. In the insurance literature, claims adjusters are cautioned about insurance fraud. But they are not cautioned that policyholders may be padding claims partly because they anticipate being disadvantaged by the fine print of the insurance contract or by claims adjusters who will find some excuse to pay only a portion of the loss.9 The inclination of policyholders to pad claims seems less reprehensible when interpreted as a realistic response to expectations that insurers will

298 Carol A. Heimer refuse to abide by the spirit of the contract. Seen in this light, padding claims seems more similar to insurers' adjustments of rates to take account of moral hazard than to outright fraud. As we have seen in this section, when we consider insurance as a system of incentives, it matters a great deal that insurers are organizations rather than people. As corporate actors, insurers enjoy many advantages that individuals typically lack (Coleman 1982): they can have longer lifespans than people, they often have more resources, they tend to be repeat players in the fields in which they operate, and they can draw on the talents and labour of many people. Although many of the insights about the effects of incentives on the generic role-playing individual apply to organizations, and especially to the individuals who work in them, the roles occupied by insurers and their policyholders are quite different. Insurers are financial institutions as well as risk managers and so have mixed interests. They are wealthy, long-term participants whose bargaining power enables them to structure the situation to their own advantage. These points are summarized in the first column of table 11.2. Disciplining Insurers (as Well as Policyholders)

As a disciplinary system, insurance works on both insurer and policyholder. Discussions of the disciplining effect of modern risk management are curiously devoid of anyone actually doing anything. Foucault (1977: 195-228) himself is something of an exception, with his detailed descriptions of Bentham's design of the Panopticon and the advantages it would yield in surveillance of prisoners. But even he is unable to tell us much about how any panoptic system actually worked in practice. Bentham's Panopticon was never built (although several prisons were constructed with panoptic elements), but Foucault makes very little of that fact. Insurers discipline policyholders through a series of mechanisms that create communities of fate (e.g., through deductibles), make recovery contingent on attempts to reduce the magnitude of losses (e.g., the marine insurance requirement to 'sue and labor'), employ the services of third-party enforcers (such as fire inspectors) when policyholder interests are especially likely to conflict with those of the insurer (e.g., when a factory owner might turn off sprinklers because low water pressure is interfering with production), and routinize the practices that insurers favour (Heimer 1985). The theory of governance through risk is that governing bodies will shape and control the activities of citizens by

TABLE 11.2 Insurers as organizations: Comparisons of economic, disciplinary, and political approaches Insurance as incentive system

Insurance as disciplinary, governmental, and actuarial system

Insurance as political system

Focal actor

As generic role-playing individuals. Insurance company, its employees, and third parties who shape insurance markets and contracts actors who may be part of a single system, but whose interests are not uniform. All in their capacity as generic roleplaying individuals.

As actors defined and shaped by disciplinary structures. Pairs in a disciplinary chain. Insurers and their agents are governed or disciplined by others so that they in turn will discipline policyholders.

As people with the capacity to make meaningful choices. Groups and individuals with political standing; people in decision-making roles.

Principal actions

Forming insurance pool, marketing insurance, designing rating and underwriting systems, evaluating and compensating losses.

Fighting with government over regulation, negotiating with other members of the environment over mutually beneficial rules.

1 Debating which decisions belong in political realm, which should be left to administrators. 2 Designing systems. 3 Articulating moral visions that define boundaries of groups and expand or contract coverage.

Main insight

1 Insurers are financial institutions as well as risk managers - insurer incentives diverge from those of policyholder as soon as insurer

1 Insurers are disciplined by market pressures, regulatory bodies, and relations with financial institutions and other members of their environments.

1 Moral visions may be decoupled from insurer routines; moral visions that match pre-existing routines more closely will have larger instrumental effects.

TABLE 11.2 (concluded) Insurers as organizations: Comparisons of economic, disciplinary, and political approaches

(Main insight)

Insurance as incentive system

Insurance as disciplinary, governmental, and actuarial system

has separate corporate existence or insurance workers have incentive to husband company resources. 2 Insurers are complex organizations, not unitary actors, and this creates ambiguity about who is authorized to form the contract. 3 Insurance is a multi-party contract between insurers, policy holders, and third parties (e.g., physicians). 4 Insurance is an ongoing strategic game, not a fixed contract.

2 Moral visions may compete; moral 2 Insurers sometimes resist discipline, baulking at the task of visions will change with changes in governing troublesome groups insurers' environments (e.g., legal, (e.g., high-risk drivers). cultural, network). 3 Rather than being simply a con3 For some lines of insurance, insurer resistance to being discisequence, solidarity is more likely plined by governments has led to both a cause and a consequence of discipline through negotiation insurance technology. with other industries (e.g., financial institutions). 4 Compared with negotiated governance, political forms of governance come with more conflict, but they also bring more transparency and more opportunity for public participation.

Insurance as political system

Insurers as Moral Actors 301 defining some activities as normal and acceptable, and whatever damage they cause as acceptable, while other activities are defined as unacceptable, and the damage they do is considered unacceptable and therefore subject to penalty (or at least not reimbursable). These governing bodies include quite disparate entities, ranging from governments, working through their welfare, security, and information collection agencies, to insurers and lending organizations and those that regulate their activities. As notions of what is unacceptable change (e.g., when we reassess activities that create environmental pollution), insurers often fight to reject the new liabilities of their clients. Economic rewards and punishments are allocated depending on which activities people pursue. As automobile ownership has become more common, for instance, regulation of the use of cars has been effected more through risk management than through legislation. Rather than legislating a series of comprehensive rules about how cars are to be driven and by whom, most state governments have limited their legislation to rules of the road and an insurance requirement and have left the rest of the governance of the roads to insurers. Rather than debating and reaching consensus on appropriate behaviour, we have yielded jurisdiction to insurers (O'Malley 1999; Heimer 2002). As insurance has become an accepted mode of governance, it has become increasingly difficult to raise questions in the political realm, for instance about who should drive and how drivers should behave. But the shift in the tasks of governance is not limited to a shift from legislation by politically elected bodies to rule making by unelected insurers; a similar shift has occurred in law enforcement. Judges are reluctant to enforce laws about, say, habitual drunk driving or habitual recklessness, preferring to leave enforcement to the insurers, who will raise an offender's auto insurance rates or perhaps refuse coverage. Even more difficult are questions that shift attention from the attributes of individuals (which insurers use to set rates) to the design of systems. Arguing that drunk driving can as easily be seen as a problem of road or car design as of individual overconsumption of alcohol, Gusfield (1981) is not pointing his finger at insurers as the entities creating the ideological hegemony, although one could well argue that he should be. Likewise, much of the regulation of personal life occurs through welfare rules about household composition (can a mother get welfare if the children's father is present?), income (how much can a mother earn and still get benefits?), and employment (must a mother apply for jobs, enrol in job training, or do community service to be eligible for welfare?).

302 Carol A. Heimer Yet we should be extremely wary of concluding that the existence of such rules means that people are actually disciplined by them. Edin and Lein (1997), for instance, show that welfare mothers routinely supplement their income in ways prohibited by welfare rules. Yes, the rules shape activity (perhaps especially by encouraging people to conceal income), but whether or not they 'discipline' is another matter. Likewise higher rates for young drivers do not remove them from the road, because the punishment is not finely tuned to behaviour. Resistance in the Middle Ranks Insurers are not just part of a disciplinary system that acts on policyholders. They themselves must be governed and questions about how to shape insurer behaviour while provoking as little resistance as possible are just as important as questions about how to discipline the populace. Part of the beauty of the panopticon was that by making the work of the guards easy, it increased the likelihood that they would carry out their assigned tasks. Although Foucault (1977) describes disciplinary systems as hierarchies of surveillance (see especially 170-7), with only a few exceptions he concentrates on the observation and discipline of the lowest ranks rather than on surveillance of intermediate or upper ranks. Foucault supplies some detail about hierarchical observation in his discussion of the measures used to control plague within a disciplined society compared to the control of leprosy by exclusion of lepers from the community (1977: 195-200). Here Foucault lists the titles and tasks of several layers of inspectors (magistrate, syndic, intendant, guard), discussing how discipline relies on a system of spatial segregation and permanent registrations, with new information added to the record at each inspection, and even touching on punishments for syndics (a sentence of death, 195) who leave the streets to which they are assigned. The Panopticon, Foucault observes, 'provide[s] an apparatus for supervising its own mechanisms. In this central tower, the director may spy on all the employees that he has under his order ... and it will even be possible to observe the director himself. An inspector arriving unexpectedly at the centre of the Panopticon will be able to judge at a glance, without anything being concealed from him, how the entire establishment is functioning. And, in any case, enclosed as he is in the middle of this architectural mechanism, is not the director's own fate entirely bound up with it?' (204). Moreover, he argues, the panoptic disciplinary mechanism is unlikely to 'degenerate into tyranny' because it 'enables

Insurers as Moral Actors 303 everyone to come and observe any of the observers' and so is 'constantly accessible to the "great tribunal committee of the world"' (207). Beyond these two passages, though, details about what happens above the lowest level are sparse. In using phrases like the 'swarming of disciplinary mechanisms' (211), the 'spatial "nesting"of hierarchized surveillance' (171-2), or the 'principle of "embedding"' (172), Foucault seems to be emphasizing that this is a capillary system with the delegation of disciplinary tasks from higher to lower level agents, rather than the control of lower level inspectors by higher level ones. This is partly because he sees the whole system as essentially self-enforcing, given its deep embedding in the spatial organization of social life, the arrangement of tasks, and the documentation of performance and progress. In a few telling passages, the possibility of rebellion by subordinates in a punitive or disciplinary system is noted. As mentioned above, syndics who left their streets during periods of quarantine for plague would be condemned to death (195). And unsuccessful executioners were sometimes subject to punishment, such as imprisonment or forfeit of a promised reward (52). In Discipline and Punish, Foucault powerfully demonstrates that spectacular public punishments such as torture and execution were politically dangerous. Based on a graphic and detailed symbolic connection between the horror of the crime and the horror of the punishment, these public spectacles were intended to display the sovereign's power and ally the population with the sovereign in opposition to crime. Too often, however, the ritual backfired because the sympathy of spectators lay with the condemned criminal rather than the sovereign: 'out of the ceremony of the public execution, out of that uncertain festival in which violence was instantaneously reversible, it was this solidarity [with the condemned] much more than the sovereign power that was likely to emerge with redoubled strength' (1977: 63). Because the spectators proved to be unreliable participants, the social control system had to be rethought and reconfigured to make the disciplinary effect less dependent on the vagaries of popular sentiment (50, 61, 112). Why we would expect unreliability to be limited to one group of participants is unclear. If a social control system can fail to do its job because the lowest level participants side with convicts and against the state, it can equally fail because others, such as executioners, syndics, or guards, are for one reason or another unable or unwilling to follow their scripts. As the local agent of the sovereign, an executioner, fearful that the wrath of spectators might be turned on him (Foucault 1977: 64),

304 Carol A. Heimer might show mercy when the script called for cruelty. Worried about catching the plague, a syndic might wish to run away rather than keep order in his assigned street. And there are any number of reasons why a guard in the tower of the Panopticon might do something other than keep watch over his charges. By resisting, rebelling, or simply having purposes and interests of their own, subordinate members of the governing class can do as much damage as can be wrought by those being disciplined or punished. It therefore makes sense to extend Foucault's argument up the hierarchy and ask how the functioning of disciplinary systems depends on an assumption of perfect performance by these participants. In extending this argument to insurance, we should be asking not just how insurance disciplines policyholders, but also what kind of behaviour is required from insurers in order to achieve this disciplinary effect. In short, we must ask how insurers themselves are disciplined so that they consistently carry out the routines needed to govern policyholders. Is the governance of insurers in some way inconsistent with or likely to undermine what insurers allegedly do to and with policyholders? If governments are to some degree governing through risk management and insurance, then either they or someone else has to shape insurers into agents who do the governments' bidding. If insurers do not follow their scripts, the disciplinary effect of insurance is undermined just as surely as the surveillance of a plague-ridden town is sabotaged by the desertion of the syndic. Who Disciplines Insurers ? Who is doing the governing and disciplining? In addition to being disciplined by market pressures, insurers are governed mainly in two ways: by regulatory bodies such as insurance commissions established through ordinary political processes and through their relations with other financial institutions and other members of their environments. The first part of this disciplinary apparatus is anything but the costless kind of governance that Foucault envisions. But because the groups being governed are relatively powerful, their resistance is quite effective. Although insurers have had to accept some compromises, they have won many of their battles with governments. The conflicts that have occurred, one could argue, are about failures or refusals of insurance to govern particular population groups. If auto insurance is made mandatory as a way of managing the risks associated with cars and driving, then

Insurers as Moral Actors 305 someone has to govern the troublesome members of the driving public inexperienced young drivers, reckless or thrill-seeking drivers, drunk drivers, and the like. When American insurers baulked, in effect refusing to govern the ungovernable portions of the population, governments insisted by requiring participation in assigned risk pools. Governance then occurs through a multi-layered process in which a few governmentimposed rules and restrictions on driving are coupled with a requirement of insurance coverage that brings with it additional insurer-imposed rules. Insurance coverage is a regime of compensating losses more than reducing them, but insurers preferred to exclude the worst bets. Governments have insisted on including these high-risk drivers in the pool because the policy of compensating rather than preventing losses would otherwise be incomplete. Insurance, unlike most other major industries, continues to be regulated largely at the state rather than the federal level.10 For instance, threats to nationalize health insurance in the United States have been fought off, often with the assistance of the coalition partners described in the section above. But what has been the result? Starr (1982) argues that in resisting regulation by the government, health professionals have ended up being governed instead by corporations - insurers and health care organizations. For insurers, avoidance of governmental regulation has come at the cost of compromises with third parties (e.g., the physicians who have an interest in increasing the 'losses' of policyholders by increasing their fees) who are in some senses the real customers for insurance. The demand for medical services is driven by what physicians prescribe, not what the patients want or believe they need. What insurers do is shaped very much through their agreeing to help with the risk management of other groups. In the case of health insurance, the risks of health professionals as well as those of consumers are reduced through insurance. For individuals, it is the risk of not being able to afford health care that leads to a purchase of insurance. For physicians and hospitals, insured patients are preferred to uninsured ones because of the guarantee that the bill (or at least most of it) will be paid. But who is it, then, that sets the terms of the contract? Typically, policyholders are in a very weak negotiating position, while health care providers are in a stronger position. Consumers do not bargain individually over the elements of their health insurance contracts; details about what sorts of care will be covered and at what rate are worked out in negotiations between health care professionals, health care organizations, and employers of the consumers.11 Insofar as insurers are disci-

306 Carol A. Heimer plined, then, it is through these negotiations. Similar arguments can be made about other lines of insurance. Homeowners' insurance does decrease risks for homeowners, but it also decreases risks for mortgage companies who insist that borrowers purchase insurance. Discipline on insurers comes more from these financial organizations (who institutionalize the purchase of insurance) than from consumers and regulators. The main discipline on insurance comes, then, from third parties who enter coalitions with insurers to increase the demand for insurance products. But for these third parties to enter and remain in such coalitions, insurers have had to agree to shape insurance institutions and practices to benefit the coalition partners. We see, then, that insurers are governed both through traditional regulatory and political processes but also through the more subtle processes of negotiation that occur in a network of interdependent actors. Although the more political forms of governance almost necessarily come with higher levels of conflict and more open opposition, they also come with greater transparency and increased opportunity for public discussion about the goals of social control systems. When social control occurs mainly through pressure and negotiation in business networks, the moral vision of insurers will likely be less attentive to the perspectives and interests of ordinary policyholders. The second column of table 11.2 presents a summary of these points. Insurance and the Body Politic In analysing insurance as an incentive system, we focus on motivation at the individual level, asking whether policyholders will be encouraged to become reckless with property that is covered by insurance, for instance, or whether insurers will be motivated to sell policyholders more insurance than they need. An analysis of insurance as disciplinary system raises questions about whether the system has been designed appropriately. We might ask whether a system of inspections was in place, but also whether inspectors were in fact reliably performing their tasks. Disciplinary systems can easily break down if the disciplinarians fail to enforce the rules. But insurance rules need not be so intrusive. They can instead be built into the infrastructure of standard operating procedures, as occurs in hospitals. There enforcement takes place as a matter of course, because insurers can count on some of their work being done for them in mortality and morbidity reviews, incident reports, accreditation reviews, and the like. In effect, policyholders and insurers both do what

Insurers as Moral Actors 307 they are supposed to, not because they are inclined by temperament or character to do so, but because they are embedded in an infrastructure, a system of sidebets (to use the terminology from Becker's 1960 paper), that would make it costly to do otherwise. Alternatively, systems can be designed to be more forgiving, and actuarial techniques can be used to anticipate risk taking and losses so that accommodations can be made and losses do not threaten the integrity of the social system. In thinking about insurance as a political system, we turn to questions about who participates in designing the system. In the Foucauldian literature on governmentality, the focus is squarely on how administrative mechanisms (forms, routines, inspections, design of spaces) construct individuals and groups in ways that make them manageable and how those same administrative forms become a taken-for-granted backdrop, a normal part of the social landscape. In emphasizing that most people are robbed of agency, we forget that some people retain their agency and work as system designers. In pointing out that political opposition and debate are reduced, we overlook important political debates about what decisions should be removed from the realm of politics and what questions must remain in the political realm. In an admittedly apathetic populace, one still hears occasional voices insisting that insurers cannot be allowed to make all of the decisions about health care or that census bureaucrats should not be the ones to decide whether Americans are classified only as black, white, Asian, and so forth, or whether people should also be able to classify themselves as mixed race. Oddly, in all of this discussion of insurance as an incentive system and as a disciplinary or risk management system, it has been easy to forget that insurance is primarily a distributive and risk-sharing system. Yes, shifting social control from the political system to the realm of administration reduces contention, but only at the cost of decreased popular debate about system design and therefore about redistribution. If we delegate decisions about health insurance to employers and private insurers, for instance, questions about provision of health insurance for stay-at-home mothers and their children become harder (although not impossible) to address. Questions about distribution and risk sharing return to centre stage when we examine the explicitly political questions about the boundaries of risk-sharing communities and who is eligible to participate in decisions about what we as a society should and should not be doing. Column three of table 11.2 summarizes these points about insurance as a political system in which questions of distribution, system

308 Carol A. Heimer design, and the division of labour between politics and administration are debated. Stone's (1993, 1999) provocative essays on insurance as a moral enterprise show what it means to think of insurance as a political system concerned with boundaries and distribution. Contrasting actuarial fairness and social solidarity as objectives for a good insurance market, Stone (1993) argues that if insurers worry too much about charging people according to their anticipated losses, they might end up undermining the community's capacity to organize to provide mutual aid. In later work, Stone (1999) argues that insurance has often provided a moral opportunity to think about and enlarge our understanding of what we owe other members of our community. If we accept the argument that insurers articulate a moral vision and see insurance as a moral opportunity to reflect on who should be included in the community and what kinds of needs should be met communally, we still need to ask who exactly it is that has the moral vision, what consequences might flow from seeing insurance as a moral opportunity, and what influences shape insurers' understandings of their moral obligations. Here I offer two caveats to Stone's argument, one that highlights the slippage between the articulation and the implementation of moral visions and a second that argues for a different causal direction. Moral positions can be articulated for a number of reasons. Insurers, like other organizations, may make moral pronouncements for symbolic as well as for instrumental reasons. And insurers, like other organizations (and even like people) may make moral pronouncements which, for a variety of reasons, they are unable to implement. Although organizations, including insurers, may make moral pronouncements largely to send signals of reputability to other organizations or potential clients (as the neo-institutionalists in sociology argue), the correspondence between symbolic and instrumental surely varies.12 What affects the tightness or looseness of the relation between symbolic and instrumental? Moral visions may compete. Insurance pronouncements carry two main messages: (a) We are going to be there when you need us. We take care of our members or policyholders; (b) We are not going to make a careful person like you pay for someone else's recklessness. These two messages are not fully consistent - if 'members' is defined expansively then surely some reckless folks will be included in the pool. Moreover, some moral visions may be more fully incorporated into insurer routines than others. Although the 'take care of our members'

Insurers as Moral Actors 309 message is a sales message that is built into marketing routines, the message that insurers only offer compensation to policyholders who have been appropriately careful is incorporated into a whole host of routines for sales, rating, underwriting, and adjustment. To be sure, insurer advertisements emphasize that their policyholders include only 'careful, responsible people like you,' but this concern with loss prevention pervades other insurer functions as well. In constructing rates, insurers look for evidence that some groups are more careful than others; in underwriting, they give discounts to individuals who can provide evidence of risk avoidance or loss prevention; and before paying the claim, loss adjusters investigate whether the policyholder's actions might have caused or increased the magnitude of the loss. This inequality between moral visions thus depends partly on the relation between the vision and the self-interest of the insurer (and its employees) and partly on which visions have been worked into organizational routines. All the sincerity in the world may become irrelevant if the routines that subordinate workers are following are inconsistent with the CEO's policy statement. The CEO's statement may suggest a policy of generosity in adjustment when routines call for stinginess, delay, denying claims whenever possible, and intimating that policyholder claims are padded if not frankly fraudulent. Moral visions may have little effect on the core technologies of insurance. Which moral visions are articulated depends partly on the importance attached to the various elements of insurers' environments and what is fashionable in each of sectors. For instance, insurance redlining, now considered morally inappropriate, likely received a less hostile reception in a pre-civil rights era, when the race, ethnicity, and gender of insurance applicants were regarded as useful and legitimate indicators of character (Heimer 1982). The legal environment and professions may be important sources of uniformity, as the neo-institutionalists contend, but that hardly brings immunity from fads and fashions and it certainly does not mean that all forms of insurance are subject to the same pressures. We should expect, for instance, that social insurance would be subjected to different pressures than commercial insurance and that health insurance would be in a different environment than homeowners' insurance. In arguing that insurance can be a moral opportunity, Stone seems to be suggesting that our moral sensibilities are to some degree shaped by the social technologies we employ. In considering how to organize mutual aid, we think about our responsibilities and our interdependence, perhaps coming to see that it costs each of us only a tad more to include

310 Carol A. Heimer additional people in our circle. But Nichols's (2000) analysis of health insurance reforms in six states suggests the peril of overlooking the importance of social solidarity as a prerequisite for insurance reform. When the political will for mandates and substantial subsidies is absent, insurance reform has in fact led lower-risk policyholders to leave the market as prices rose to cover higher-risk policyholders. In some states (e.g., Kentucky), the result has been a failure of the reform program; in most cases the proportion of the state population covered by insurance has dropped. The exit of policyholders from the market is a rather clear behavioural indicator that insurance reforms have not in fact increased solidarity. Nichols's conclusion is that political consensus must precede reform if reforms that extend benefits to more people and cover more risks have any hope of succeeding. Although Nichols's argument and evidence do not entirely undermine Stone's claim - it seems likely that solidarity could be both cause and effect - he offers a salutary reminder that mutual aid depends on some modicum of mutuality. The Moral World of Insurers

The moral world inhabited by insurers is quite different from the moral world inhabited by their policyholders. Insurers' interests are quite different from those of their policyholders. The pressures to which they are subject are not the same. Though both are subject to disciplinary measures, insurers face those who attempt to govern them from a position of relative strength. When we speak of insurers as 'moral actors,' we mean to present a view of the world from the perspective of these actors. We mean to examine how insurers and their employees and agents understand, interpret, and justify their own acts - what acts are coded as appropriate or inappropriate, noble or reprehensible, wise or ill-considered, which activities should be rewarded and which should be punished. This view from the inside may be quite different from the interpretations of policyholders or regulatory bodies, for instance. Because insurers are not unitary actors, identifying their view is more difficult, yet Bourdieu's (1985) notion of 'fields' suggests that we might expect distinctive cultures and distinctive moral universes in different organizational fields. Thus although sales agents may encourage people to be 'safe' by overinsuring, claims adjusters may reduce insurance compensation to below the real value of the loss, and actuaries may then show that premiums balance the (underestimated) losses perfectly.

Insurers as Moral Actors 311 By examining three perspectives that have been used to assess insurance and its effects on individual policyholders and on modern societies, we have seen how some features of insurance culture depend on turning a blind eye to insurance itself and thinking instead about insurers as complex organizations responding to and constrained by other organizations in their environments. Insurers, like policyholders, find their incentives changed by the contracts they form - but on closer inspection, we discover that insurers are not like individuals and that they are investors as well as insurers. Risk, for insurers, is thus as much about the stock market as about policyholder carelessness or fraud. Insurance may be an important social control mechanism in contemporary societies, but when we look closely we discover that insurance can only work if insurers as well as policyholders are subject to discipline. And when we consider how insurance affects our understanding of our place in a risk-sharing community, we see that insurers may articulate but not implement such moral visions and that in fact social solidarity is probably as much a cause as a consequence of accepting responsibility for others' fates. Acknowledgments I am grateful for stimulating discussions with the participants in the conference on Risk and Morality at Green College, where an earlier version of this chapter was presented, and to Barry Cohen and Arthur Stinchcombe for their comments and assistance as it evolved. As always, I appreciate the warmth of the Lochinvar Society.

Notes 1 Notable recent examples of work focusing on insurers (rather than policyholders) include Baker (1996, 2000) and Ericson, Barry, and Doyle (2000). Writing in the 'law in action' tradition, Ross (1970) anticipated this shift in focus, although he was more interested in contrasting insurance 'law on the books' with 'law in action' than comparing policyholder and insurer reactions to the incentives embedded in an insurance contract. 2 Readers should note that pensions, one of the largest and most important insurance schemes, are omitted from this analysis. Despite their importance, pension schemes differ significantly from other forms of insurance in the kinds of incentive, disciplinary, and political problems they pose. For a useful discussion of pensions, see Stinchcombe (1984).

312 Carol A. Heimer 3 See Andrews and Nelkin (2001) on insurers' use of genetic information to deny insurance coverage. 4 Courts have sometimes made insurers stick to the promises their agents have made, however. For instance, in the case of Tidelands Life Ins. Co. v. Franco, 711 S.W.2nd 728 (Tex. App. 1986), Linda Franco claimed damages for the insurance company's refusal to pay expenses for gall bladder surgery. Franco had bought a health insurance policy after Alvin Choate, an agent of Tidelands, assured her that her pre-existing gall bladder problem (which he listed on the form as completely cured) would be covered. Tidelands contended that Choate could not bind the company because he was a 'soliciting agent' not a 'local recording agent.' The court rejected this argument on the grounds that Choate had 'acted with actual, apparent, and implied authority' in discussing the policy with Franco, introducing himself as a Tidelands agent, using an insurance application that referred to him as a Tidelands agent, and filling out the form for Franco. He misrepresented the situation to Franco to induce her to purchase the insurance and then misrepresented her to the insurer to induce it to accept her application. Interestingly, the court had drawn the same conclusion in a very similar case concerning the same insurer and agent two years before. See Kimball (1992: 102—3) for a discussion of Tidelands Life Ins. Co. v. Franco. 5 This principle is articulated in Restatement (Second) of Contracts section 206 (1981). Not everyone agrees that a principle favouring policyholders is appropriate. See Kimball (1992: 6-8). 6 See, for instance, Landberg, who argues that '[d]emutualization and the subsequent initial public offering require that an organization shift from a largely longterm strategic perspective to a short-term, earnings-driven culture. Many mutuals have been moving slowly toward a bottom-line orientation over the last 10 years. However, the full wrath of stock analysts and major investors has not yet been felt by many executives. Middle management and the broad organization have generally not been held truly accountable for nor rewarded based upon their performance. Mutual insurers' primary stakeholders have been the policyholders, agents, rating agencies and regulators. As stock companies, insurers will add fund managers, professional and retail investors, stock analysts, and the Securities and Exchange Commission to the mix' (1999: 59). Similar sentiments are expressed in Pargeans and Adams (1999). 7 Gardner (1999), for instance, observes that '[a]s the traditional barriers between banking and insurance are dismantled, banks and other financial institutions increasingly will enter the world of insurance for profit-making opportunities' (87-90).

Insurers as Moral Actors

313

8 See Blau (1963) and Lipsky (1980) on how policy gets translated into action by lower level employees in ways that may undermine the original policy objective. 9 Much of the literature on insurance ethics consists either of discussions of how to improve ethics training for insurance personnel (see, e.g., Berger 1995; Wood 1993) or of reports on surveys of insurers' own views about which ethical issues they and their colleagues encounter (see, e.g., Press (1994) and the series of papers by Cooper and Frank (1990, 1991a, 1991b, 1992)). Such reports do not, of course, tell us what policy-holders think about insurers' ethics. 10 Although primary regulation occurs at the state level, federal agencies, including the Department of Justice, the Federal Trade Commission, the Internal Revenue Service, the Department of Labor, and the Securities and Exchange Commission, do play an important role in regulating one or another aspect of insurance. Congressional committees also consider questions about insurance in investigations and hearings. 11 See Stirichcombe (1984) for a discussion of how insurers' incentives are shaped by the practice of negotiating with third-party buyers rather than directly with consumers. 12 The neo-institutionalist perspective in sociology argues that much of what transpires in organizations occurs not because it increases efficiency or reduces transaction costs, but because it confers legitimacy and smoothes interactions among organizations (Meyer and Rowan 1977; DiMaggio and Powell 1983,1991). In adopting a practice that is already employed by other organizations, then, an organization signals its willingness to play by the rules endorsed by other actors in its environment. It claims membership in a world in which organizations doing roughly the same thing agree on a core set of rules that make die game fair and a style of interaction with regulators, suppliers, customers, and other publics. How deeply the adoption of any institutionalized policy shapes organizational activity is another matter, however, and institutionalists differentiate instrumental from symbolic effects.

References Andrews, L., and D. Nelkin. 2001. Body Bazaar: The Market for Human Tissue in the Biotechnology Age. New York: Crown Arrow, KJ. 1963. "Uncertainty and the Welfare Economics of Medical Care.' American Economic Review 53: 941-69 - 1968. 'The Economics of Moral Hazard: Further Comment.' American Economic Review 58: 537-8

314 Carol A. Heimer Baker, T. 1994. 'Constructing the Insurance Relationship: Sales Stories, Claims Stories, and Insurance Contract Damages.' Texas Law Review 72: 1395-1433 - 1996. 'On the Genealogy of Moral Hazard.' Texas Law Review 75(2): 237-92 - 2000. 'Insuring Morality.' Economy and Society 29(4): 559-77 - 2002. 'Risk, Insurance and the Social Construction of Responsibility.' In T. Baker and J. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Baker, T., and J. Simon, eds. 2002. Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Becker, H.S. 1960. 'Notes on the Concept of Commitment.' American Sociological Review 66: 32-40 Berger, L.A. 1995. 'Train all Employees to Solve Ethical Dilemmas.' Best's Review (Life/Health Insurance edition) 95(11): 70-2 Blau, P.M. 1963. Dynamics of Bureaucracy. Rev. ed. Chicago: University of Chicago Press Bourdieu, P. 1985. 'The Social Space and the Genesis of Groups.' Theory and Society 14: 723-44 Coleman.J.S. 1982. The Asymmetric Society. Syracuse, NY: Syracuse University Press Cooper, R.W., and G.L. Frank. 1990. 'Ethics in the Property-Liability Insurance Industry: Perceptions of the Issues.' CPCUJournal (December): 224-38 - 1991a. 'Business Ethics in the Insurance Industry.' Journal of the American Society ofCLUand ChFC45 (May): 74-80 - 1991b. 'Ethics in the Life Insurance Industry: The Issues, Helps, and Hindrances. 'Journal of the American Society ofCLUand ChFC45(5): 54-66 - 1992. 'Key Ethics Issues Facing the Industry: The Views of Other Life Insurance Professionals.' Journal of the American Society ofCLUand ChFC46(5): 76-82 Defert, D. 1991. '"Popular Life" and Insurance Technology.' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality, 211-33. Chicago: University of Chicago Press DiMaggio, P.J., and W.W. Powell. 1983. 'Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields.' American Sociological Review 48: 147-60 - 1991. 'Constructing an Organizational Field as a Professional Project: U.S. Art Museums, 1920-1940.' In W.W. Powell and PJ. DiMaggi, eds., The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press Edin, K., and L. Lein. 1997. Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work. New York: Russell Sage Foundation

Insurers as Moral Actors 315 Ericson, R., D. Barry, and A. Doyle. 2000. 'The Moral Hazards of Neo-liberalism: Lessons from the Private Insurance Industry.' Economy and Society 29(4): 532-58 Ewald, Francois. 1991. 'Insurance and Risk.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 197-210. Chicago: University of Chicago Press Foucault, M. 1977. Discipline and Punish: The Birth of the Prison. Trans. Alan Sheridan. New York: Random House - 1991 (1978). 'Governmentality.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 87-104. Chicago: University of Chicago Press Gardner, M. 1999. 'Pressure to Sell Out.' Best's Review (Life/Health Insurance edition) 100(6): 87-90 Gordon, C. 1991. 'Governmental Rationality: An Introduction.' In G. Burchell, C. Gordon, and P. Miller, eds., The Foucault Effect: Studies in Governmentality, 1-51. Chicago: University of Chicago Press Gusfield.J.R. 1981. The Culture of Public Problems: Drinking-Driving and the Symbolic Order. Chicago: University of Chicago Press Heimer, C.A. 1982. 'The Racial and Organizational Origins of Insurance Redlining.'JournaloflntergroupRelations 10(3): 42-60 - 1985. Reactive Risk and Rational Action: Managing Moral Hazard in Insurance Contracts. Berkeley: University of California Press - 2002. 'Insuring More, Ensuring Less: The Costs and Benefits of Private Regulation through Insurance.' In T. Baker and J. Simon, eds., Embracing Risk. Chicago: University of Chicago Press Holmstr0m, B. 1983. 'Moral Hazard and Observability.' BellJournal of Economics 10: 74-91 Kimball, S.L. 1992. Cases and Materials on Insurance Law. New York: Aspen Publishers Knights, D., and T. Vurdubakis. 1993. 'Calculations of Risk: Towards an Understanding of Insurance as a Moral and Political Technology.' Accounting, Organizations and Society 18(7/8): 729-64 Landberg, S. 1999. 'Driven to Perform.' Best's Review (Life/Health Insurance edition) 100(1): 59-61 Lipsky, M. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation Meyer, J.W., and B. Rowan. 1977. 'Institutionalized Organizations: Formal Structure as Myth and Ceremony.' American Journal of Sociology 83: 340-63 Nichols, L.M. 2000. 'State Regulation: What Have We Learned So Far?' Journal of Health Politics, Policy and Law 25(1): 175-96

316 Carol A. Heimer O'Malley, P. 1999. 'Imagining Insurance: Risk, Thrift and Industrial Life Insurance in Britain.' Connecticut Insurance Law Journal 5(2): 675-705 - 2000. 'Uncertain Subjects: Risks, Liberalism and Contract.' Economy and Society 29 (4): 460-84 Pargeans, W., and R. Adams. 1999. 'A Nation on the Brink of Demutualization.' Best's Review (Life/Health Insurance edition) 100(6): 64-6 Pauly, M.V. 1968. 'The Economics of Moral Hazard: Comment.' American Economic Review 58: 531-7 - 1974. 'Overinsurance and Public Provision of Insurance: The Roles of Moral Hazard and Adverse Selection.' Quarterly Journal of'Economics 88: 44-54 Press, A. 1994. 'Ethically Speaking.' Managers Magazine 69(5): 14-17 Roberts, D.E. 1995. 'Irrationality and Sacrifice in the Welfare Reform Consensus.' Virginia Law Review S I : 2607-24 - 1997. Killing the Black Body: Race, Reproduction, and the Meaning of Liberty. New York: Pantheon Ross, H.L. 1970. Settled Out of Court: The Social Process of Insurance Adjustment. Chicago: Aldine Shavell, S. 1979. 'On Moral Hazard and Insurance.' Quarterly Journal ofEconomics 93:541-62 Simon, J. 1988. 'The Ideological Effects of Actuarial Practices.' Law and Society Review 22(4): 771-800 Starr, P. 1982. The Social Transformation of American Medicine. New York: Basic Books Stiglitz, J.E. 1983. 'Risk, Incentives and Insurance: The Pure Theory of Moral Hazard.' Geneva Papers on Risk and Insurance 8(26): 4-23 Stinchcombe, A.L. 1984. 'Third Party Buying: The Trend and the Consequences.' Social Forces 62 (4): 861-84 Stone, D. 1993. 'The Struggle for the Soul of Health Insurance.' Journal of Health Politics, Policy, and Law 18(2): 287-317 - 1994. 'Promises and Public Trust: Rethinking Insurance Law Through Stories.' Texas Law Review 72: 1435-46 - 1999. 'Beyond Moral Hazard: Insurance as Moral Opportunity.' Connecticut Insurance Law Journal 6 (I): 11-46 Wood, E.F. 1993. 'Developing Ethical Standards for Risk Managers.' Risk Management 40 (8): 31-5

12 The Moral Risks of Private Justice: The Case of Insurance Fraud RICHARD V. ERICSON AND AARON DOYLE

Governance beyond the Law The most pervasive framework in studies of law and society is the relationship between the law in action and the law on the books. Studies focus in particular on the discretion available to various legal agents regarding law enforcement options as well as alternative justice remedies. The law in action approach has shown how a great deal of illegal conduct that comes to the attention of law enforcement agents is diverted back into the institution involved - for example, family, school, or business enterprise - to handle through its own mechanisms of selfgovernance. Indeed, such diversion is explicit in the work of regulatory agencies that use compliance models under administrative law (Hawkins 1984). But it can also be discerned in how the public police operate under criminal law (Ericson and Haggerty 1997, 2002; Ashworth 2001). The police, and the criminal law institution more generally, actively discourage most efforts to deal with crime through criminal prosecution and punishment. Through discourses such as 'community policing,' the criminal justice system promotes crime prevention as the responsibility of each and every member of society. When crime prevention fails, private justice mechanisms - at the level of institutions, organizations, or individuals - provide the primary basis of resolution. The law is usually not even taken into consideration in deciding what to do about conduct that could be considered illegal. The governing mechanisms are simply beyond the law. The law on the books/law in action approach is based on the conventional sociological model of deviance-control-order. Deviance is violation of norms specified in rules, backed by a moral concern that the

318 Richard V. Ericson and Aaron Doyle behaviour in question is not only bad but a violation of moral order. The law and its agents offer deterrent, punitive, and repressive means of controlling those who are deviant. This control is directed at reproducing order, not just in the narrow sense of maintaining security of public space, but also in articulating local standards of civility and providing a barometer of moral order. Moreover, the control mechanisms themselves are to exemplify moral order in the form of procedural propriety, for example, with respect to due process norms in legal proceedings. In governance beyond the law the more salient model is one of moral risk-surveillant assemblages-loss ratio security. The problematic conduct is posed not as illegal behaviour defined by the moral codes of law, but rather as a risk defined by the moral utilitarian criteria of the particular institution involved. The governing mechanisms are not legal controls over unwanted conduct, but rather a network of surveillance systems that form an assemblage to provide knowledge useful in addressing moral risks (Haggerty and Ericson 2000). The concern is not order as imagined through the regularities of law, but rather security of the specific institution's loss ratios: the predictability and assurances it can offer to its constituents that the losses they may suffer through moral risks will not undermine the smooth flow of their social and economic participation in the institution. As Beck (1999: 10) observes, 'risk issues cannot be converted into issues of order, because the latter suffocate, so to speak, from the immanent pluralism of risk issues and metamorphose surreptitiously behind the facades of statistics into moral issues, power issues and pure decisionism.' The Moral Economy of Insurance

The private insurance institution exemplifies governance beyond the law on the basis of moral risk, surveillant assemblage, and loss ratio security.1 Moral Risk The insurance institution is premised on risk logics. Indeed, the concept of risk is itself a neologism of insurance. Insurance risk logics are embedded in sophisticated technologies of surveillance, classification, and quantification. These technologies objectify risks and make them amenable to decisions about what and who to insure and compensate and on what terms. They are moral technologies because they evaluate risks and the

The Moral Risks of Private Justice

319

people associated with them as good or bad, and as subject to inclusion or exclusion. Any effort to establish an objective norm through probability statistics also assumes that what is normal is proper and what ought to be (Hacking 1990: 170). In assessing who are 'normal people' for inclusion in an insurance risk pool, and who should be de-selected and unpooled, insurance technologies create morally based social distinctions, hierarchies, and exclusions. Insurance actuarialism articulates moral risks as a means of defining self and other, communities of interest, and the dangers they are to be protected against (Douglas 1992). It also contributes to the ways in which risk has become the moral discourse of our scientized society (Beck 1992,1999). Moral risk is a concept that pervades the insurance institution, although it is more commonly expressed as moral hazard (Heimer 1985; Baker 1996; Ericson, Barry, and Doyle, 2000; Baker 2000; Ericson, Doyle, and Barry 2003). In this inflection, moral risk is posed by something that is self-serving to a particular party in the insurance relationship to the point of immorally increasing risks to one or more of the other parties. The focus is on the insured in particular. The insured may fail to protect what she has insured because she knows that she will be compensated for a loss. Moreover, she may exaggerate insurance claims when she suffers a loss, or even make false claims when she needs money. These moral risks shift the burden to the insurer, who pays illegitimate claims that have a negative impact on loss ratios. The insurer may in turn shift the burden to other insured in the form of higher premiums and less favourable contract conditions in an effort to improve loss ratios. This transfer results in unpooling, as some insured may no longer be able to afford participation in the insurance pool they originally bought into. Moral risks are not limited to the behaviour of insured persons. Insurance agents can pose substantial moral risks to insured persons if they sell them the wrong product or something they do not need at excessive rates (Clarke 1999; Ericson, Barry, and Doyle 2000; Ericson, Doyle, and Barry 2003). Insurance service providers — for example, lawyers, health care professionals, and auto body shop operators - may pose moral risks to the insurer, the claimant, and the pool of insured if they systematically over-bill and provide unnecessary services as part of the claims process. Surveillant Assemblage The insurance institution operates on the webs of significance spun by its surveillant assemblages. Myriad surveillance technologies - inter-

320 Richard V. Ericson and Aaron Doyle views, questionnaires, inspections, audits, credit checks, data matching exercises, private investigations, and so on - are used to produce knowledge of risk that is in turn employed in the governance of insureds, insurance agents, and service providers to protect loss ratios. The insurance surveillant assemblage is a moral kaleidoscope, constantly scanning for a constellation of risk factors that can form the basis of underwriting and claims decisions. It is the coordinating mechanism that forms insurance risk pools and market segments around acceptable loss ratios. It fosters reflexive knowledge of moral risks useful to all parties in the insurance relationship in the active management of loss prevention and of responsible choice in risk taking. Loss Ratio Security

The security of interest to the insurance relationship is the protection of loss ratios. The insurer's loss ratio consists of premium revenue minus administrative expenses, minus loss prevention expenses, minus claims paid, plus reinsurance claims, plus investment returns. Although there are many complexities and subtleties, the basic elements of loss ratio security are as follows. The insurer competitively markets its policies to attract premium revenue. Expenses must be paid to administer all aspects of the insurance company operation, including marketing, claims processing, and expert services (legal, accounting, actuarial, medical). Expenses are also paid for inspections, surveillance, policing, and educational programs that help to prevent the losses insured against and therefore reduce insurance claims costs. By far the largest call on premium revenue is actual claims, which must be kept to a manageable level in order to retain some money for investment. One crucial mechanism for managing the claims level involves the primary insurer taking out its own insurance with other insurance companies: these reinsurers are paid premiums and assume responsibility for claims above an agreedupon level. The money remaining after all of the above expenses have been addressed is invested across a range of products - bonds, equities, real estate, and so on - in an attempt to increase capital and turn a profit. Ideally controllable loss ratios more or less meet the needs of each party to the insurance relationship: insurers who want maximal dollars to invest for profit; agents and service providers who expect reasonable compensation for their work; and insured who expect the company to protect the risk pool and meet its promise to pay in the event of a loss.

The Moral Risks of Private Justice

321

Minimization of claims costs is crucial for loss ratio security. This loss minimization can be effected by mobilizing the insured to act as agents of prevention through, for example, contract incentives for preventive security efforts. It can also be effected through claims management practices. A few significant losses are expected, and reasonable compensation of them is at the core of what the insurance industry sells. But what constitutes reasonable compensation is perpetually at issue, and this determination lies at the heart of the moral risk management enterprise. Claims staff are trained to be meticulous and scrupulous in vetting the claims of both the insured and service providers. As we shall see, there has been expansion recently in both private policing and deployment of surveillance technologies aimed at reducing the 'social inflation' of unjustified claims. Insurance claims fraud is posed as the problem, but it is also used as a basis for justifying more elaborate surveillant assemblages to govern all aspects of the insurance relationship. Loss ratios can also be improved by generating more premium income. If moral risks increase, it is still possible to create 'substandard' pools and charge higher premiums. As a president and CEO of an insurance company told us during an interview, 'My attitude is a good risk is one that pays high enough premium.' Sales agents may also be given commission and other incentives to aggressively sell policies with higher premiums. This is a routine practice, and has given rise to numerous misselling scandals. However, insurers in highly competitive 'soft' markets are restricted from charging too much because they will not attract new business and may lose their preferred policy-holders to competitors. In such conditions, they are more likely to maintain or even reduce their premium levels, hoping to secure acceptable loss ratios through better claims management. The insurance moral risk-surveillant assemblage-loss ratio security system exists to magnify differences among individuals in order to achieve greater precision in the commodification of insurance products and exclusion of those who do not fit (Simon 1987). An insurance company's mandate is to discriminate on merit, for example, favouring policyholders who foster loss reduction, and to make prudent decisions that keep the company solvent and prosperous. Its goal is not a redistribution of resources among insured persons, or equal treatment, but only equitable treatment within each community of interest it has pooled (Gowri 1997: 151-3). This point was made explicit during an interview with an insurance claims consultant. This person worked with insurers across North America

322 Richard V. Ericson and Aaron Doyle to develop dataveillance systems that would help them identify and reduce sources of excessive claims. He said that one technology he had devised was a 'moral cost analysis' based on an assessment of the ratio of claims awards to personal income of claimants. The technology measured the 'claims award basis relative to need or desire and relative to opportunity ... $100 isn't anything to me ... Why would I lie for $100? I get more gain off the moral cost to me, in a sense it's just self-image ... a street person down there, $100 is a bunch more.' Those who have high moral cost considerations in making a claim are naturally better moral risks. Those assumed to have no reputation to lose, in other words the poor, are obviously not worth the risk. Insurance Fraud

Fraud is a moral risk that pervades all insurance relationships. As we document in subsequent sections, fraud results from the way in which insurance relationships are governed. At the same time, the fraud produced by the insurance system justifies the introduction of additional governance mechanisms that serve a variety of purposes beyond the control of fraud. Fraud is produced through governing, and governing is produced through fraud. The Mismeasure of Fraud

The epidemiology of insurance fraud is elusive. The difficulty stems from problems of definition, as well as the fact that most suspected fraud is not dealt with as such. Where insurance fraud is concerned, criminology's 'dark figure' of crime problem becomes a black box. The incidence and prevalence of fraud is unknown, except to the extent that it is made into a problem for the governance of insurance relationships more generally. Fraud is an artifact of how the insurance industry organizes to deal with it. As a senior official in an insurance private policing operation commented, 'The scary part is, as we continue to increase the fight, we continually uncover more and more fraud. So people say, "You are putting all these dollars in there, but the problem is increasing." You can't say something is fraudulent until you find it. There's been such a growth in "insurance fraud." It has probably always been there, it is just that more has been uncovered than ever before.' Some industry insiders, especially those committed to a crackdown on fraud for more general purposes of governance, offer a strict definition.

The Moral Risks of Private Justice

323

Fraud is intentional deception that exposes the recipient to unjustified losses. For example, in a popular magazine article on insurance fraud, the then executive director of the Canadian Coalition Against Insurance Fraud, who was also vice-president of Public Affairs for the Insurance Bureau of Canada, offered the following definition: 'It's any act or omission resulting in illicit collection of property and casualty insurance benefit... [from] fabricated claims, to inflation or padding of legitimate claims and false statements on applications.' In the context of actually trying to govern fraud in the claims process, the definition loosens. A definition that appears in insurance company manuals for claims adjusters says that fraud occurs when the claimant misrepresents facts with the intention of having the insurer act on them. For example, one such manual declares, 'For fraud to exist there must be a false representation of fact, knowledge of the statement's falsity by the person making it, and intent by the insured that the insurer will act on it.' While this definition is broad and encompassing, the manual immediately notes, 'Slight exaggeration in the amount of the claim itself will not constitute fraud (although this may represent a moral hazard from an underwriting perspective), however falsifying documents to support an exaggerated claim would constitute fraud.' The distinction made between exaggeration and fraud is typical. An auto claims adjuster said in interview that while 'everyone exaggerates,' fraud is clear-cut falsification of a document or material evidence. Her company offered a five-hour training session that tried to specify the fine line between exaggeration and fraud. Another interviewee, involved in organizing this training program, was asked how the line was drawn. He replied, 'If you could give us the answer to this, you'd be a hero.' The answer is that insurance adjusters and investigators take a pragmatic approach. When asked to define fraud, an insurance claims investigator employed by a multinational management consultant firm replied, 'Well, it's easy. You see we're not the cops, so it's easier. We pursue it when we're paid to pursue it, and that's it... We're driven by our clients ... They're driven by fear and self-interest.' A manager of special investigators for an auto insurer said that this pragmatic approach to fraud has three elements. First, it must be determined that the claimant intended to deceive for personal gain. Second, the magnitude of the loss must be great enough to make further investigation worthwhile. Third, further investigation and possible sanctions must be both viable and efficient. If investigative options appear foreclosed, or adequate sanctions are not available, the determination of fraud dissipates.

324 Richard V. Ericson and Aaron Doyle The combination of intentionality, loss magnitude, and investigative pay-offs also leads to distinctions between 'soft core' abuse of the system and 'hard core' criminal fraud. We interviewed a consultant who developed dataveillance systems to detect insurance fraud. He said an example of 'soft core' abuse in auto insurance claims is the claimant who has a minor neck injury that becomes a major injury, usually on the urgings of a lawyer. An 'injury ring' that stages a fake auto accident followed by multiple claims from several 'injured' parties would be considered 'hard core' fraud. Similarly, an investigator for a workers' compensation insurer distinguished between '20 per cent abuse' and '5 per cent criminal fraud' among all claims. He said 'abuse' is largely accounted for by 'malingering individuals ... off on a couple of extra weeks longer than necessary simply because they've gone back to the doctor and said, "I need a couple of extra weeks" ... We're in an injury system, not an adversarial system. On balance of probabilities, we give the nod to the claimant.' The detection of fraud in everyday claims adjusting is routine. In a vehicle hit-and-run insurance claims centre we studied, out of an average of sixty-five claims each day, six were judged fraudulent based on material damage assessment. These claims were denied on the spot. However, an additional thirteen claims were believed to be fraudulent but not investigated further because the company decided it was more expedient to pay the claim. Cases involving personal injuries and disabilities are even more ambiguous and difficult to assess. Studies of bodily injury claims arising from vehicle claims in Massachusetts (Weisberg and Derrig 1991, 1992) judged 31 per cent (N = 597) of one sample and 47 per cent (N = 1, 154) of another to have involved dishonesty. Such data are not routinely produced by insurers themselves to provide systematic evidence of levels of fraud. Where fraud is suspected but not pursued for reasons of expediency, there may be no data collected at all. Even when fraud is detected, it may not be recorded on company records if a resolution is reached with the perpetrator. Industry spokespersons nonetheless make up numbers to sustain a view that fraud is a serious moral risk that threatens the industry. The National Insurance Crime Bureau in the United States, founded in 1992, hired as its first director a former U.S. Army general who had worked in the Pentagon in the Bush administration's war on drugs, and who talked in similar terms about insurance fraud as a threat to the nation. Similarly in Canada, industry associations and spokespersons regularly proclaim that 'insurance fraud is the second leading source of criminal profits

The Moral Risks of Private Justice

325

next to drugs.' A brochure on fraud distributed to policyholders of an auto insurer declares '10 percent to 20 percent of all claims are fraudulent ... That means $160 to $320 million of your premiums are paid to people who don't deserve them.' Another document was more precise, stating that in 1996 fraud cost $308 million, 15 per cent of all claims costs. An anti-fraud investigator for this company used this same figure during an interview, and then commented that it 'is probably being conservative, because where I grew up everybody tries to screw us.' The visualization of insurance fraud facts is also exemplified in the following industry association document: The crime of insurance fraud is growing in Canada; its economic impact is exceeded only by that of tax evasion. Property and casualty insurers in Canada believe that at least 10 to 15 percent of household, automobile and commercial insurance claims are fraudulent - either completely fraudulent or inflated. That's about $1.3 billion a year in fraudulent claims that must be paid from the premiums of all policyholders ... [I]nsurance fraud has costly secondary consequences; staged 'accidents' and deliberately set fires, medical examinations required because of false personal-injury claims and, of course, the wasting of costly and-all-too scarce police resources. Consumers are unaware that even the padding of a legitimate claim - often called 'opportunistic' fraud or claims 'build up' - is illegal. The claim that insurance fraud is growing is not substantiated by onetime figures of fraudulent claims and their costs. These one-time figures are made up to create a dramatic sense of rampant 'social inflation' excessive expectations about compensation through insurance - that justifies anti-fraud sanctions such as higher premiums and more stringent contract conditions. Industry insiders themselves admit that these visualizations of social inflation are themselves fraudulent claims. Thus, in an editorial entitled, 'The Mismeasure of Fraud,' written by the editor of Canadian Insurance magazine, the following assessment is made: Although we have seen in countless media the estimate of fraud in Canadian property and casualty insurance - 10-15 percent of claims or roughly $1.3 billion - few actually know where this measurement came from or, for that matter, how accurate it is. My exhaustive research has turned up only one reference to this number. A judgment in a 1992 New Jersey Supreme Court case stated that 'insurance fraud is a problem of massive proportions that currently results in substantial and unnecessary costs to the general

326 Richard V. Ericson and Aaron Doyle public in the form of increased rates. In fact, approximately ten tofifteenpercent of all insurance claims involve fraud.' No one really knows how much these particular judges knew about fraud, but the estimate lives on. Since the Canadian Coalition Against Insurance Fraud clearly borrowed the name from its U.S. counterpart (Coalition Against Insurance Fraud), it is quite likely that this number was also extrapolated to fit the Canadian environment (Harris 1998: 5).

Fraud through Governing Within the industry, fraud is seen as a product of how insurance relationships are governed. Fraud is embedded in the moral risk-surveillant assemblage-loss ratio security nexus through which insurance governs. As Garland (1997: 186; see also Lianos and Douglas 2000; Garland 2001) observes about criminogenic situations more generally, The everyday conduct of economic and social life supplies countless opportunities for illegitimate transactions. Viewed en masse, criminal events are regular, predictable, systematic, in the way that road accidents are. It follows that action upon crime should cease to be primarily action upon deviant individuals and become instead action designed to govern social and economic routines.' In the insurance relationship, fraud is but a factor in moral risk calculus, something taken into account in assessing acceptable levels of loss. It is only policed vigorously if it is seen to have a substantial impact on loss ratio security. Otherwise it is tolerated, or even deemed acceptable because, ironically, fraud can actually contribute to loss ratio security. In this light, the surveillant assemblage cannot be too intrusive. As Garland (1997: 187) also observes, the criminogenic situation 'has functional ends of its own that are easily disturbed by heavy-handed regulation.' Moral regulation of fraud cannot upset the insurance relationship and its commercial values. Precisely because fraud is institutionalized as a normal component of the insurance relationship, any intervention any act of moral risk and responsibility redistribution - must not adversely affect business as usual. The preferred approach is governanceat-a-distance, trying to mobilize each party involved in the insurance relationship to be self-governing with respect to the moral risks of fraud they pose, and to be vigilant about others who may take advantage of them. 'The "criminogenic situation" poses difficulties for government because it generally has a commercial or social value of its own which sets limits on crime control. Precisely because crime occurs in the course

The Moral Risks of Private Justice

327

of routine social and economic transactions, any crime-reducing intervention must seek to preserve "normal life" and "business as usual." The characteristic modes of intervening involve the implantation of nonintrusive controls in the situation itself, or else attempts to modify the interests and incentives of the actors involved' (ibid). Governing through Fraud While fraud is produced through the governing mechanisms of the insurance industry, it also provides a rationale for elaboration of those mechanisms. The 'shocking numbers and graphic accounts' (Orcutt and Turner 1993; see also Best 2001) of insurance fraud in the public media are used to justify crackdowns on claims processing, including a proliferation of surveillance technologies. This proliferation goes beyond the regulation of fraud per se. The insurance industry governs through fraud (Simon 1997) by using it as a justification for generating surveillant assemblages that serve broader purposes of moral risk management and loss ratio security. In what follows we demonstrate the processes involved in both fraud through governing and governing through fraud, examining in turn sales fraud, claimant fraud, and claims services fraud. Sales Fraud

Sales Fraud through Governance Fraud perpetuated by insurance sales agents is a product of how those agents are governed. The 'misselling' of life insurance is a case in point. Misselling is a euphemism for institutionally induced and endorsed fraud. The traditional system for life insurance sales involves recruitment of neophytes who are required to produce a substantial list of 'prospects' from among their family, friends, and acquaintances. Some salary may be offered in the start-up period, but otherwise the sales staff, including sales managers and even executives, are commission-dependent. As a result, they have a strong incentive to sell products that yield higher premiums for the company and higher commissions for themselves. For example, as is widely acknowledged in the industry, some consumers, particularly those who are on limited budgets, are best off with a simple term life policy that carries only death benefits. However, an effort is

328 Richard V. Ericson and Aaron Doyle made to sell instead investment-based policies, such as whole life or universal life products, because they are more lucrative to the company and sales staff. While these products potentially allow the policy to accumulate value, they are based on investment assumptions that often fail to materialize or that have a modest yield in comparison with other investments the consumer could make. Sales agents are pressured to neutralize the moral bonds they have to their family, friends, and acquaintances in order to coax them into inappropriate purchases. Ostensibly for training purposes, more senior sales personnel often accompany the new agents when they sell to family and friends and play a major role in closing the deal. In addition to structuring commissions so that agents sell more expensive investment policies instead of term policies, life insurance companies offer commission and other incentives to 'churn' or 'twist' existing policyholders into more expensive replacement policies. Commissions for replacement policies are sometimes higher than commissions for new client policies. At the same time as more and more features creep into life insurance products, making them virtually impossible for consumers to understand, more and more incentive features creep into the system of agent compensation so that they will push the new products. High-flying agents are put into the equivalent of frequent flyer programs, with regular, gold, and platinum status depending on sales performance. The higher the tier the higher the rewards to the agent, both in terms of material compensation and discretion to handle clients differentially for even greater rewards. Greater discretion can include things like 'relationship pricing,' whereby top agents are given some flexibility in setting life insurance premium rates for preferred customers who also buy other lines of insurance and other financial products from them. Commission-based churning was a significant component of the life insurance and pension scandals in the United Kingdom (The Economist, 29 October 1994: 95; Sunday Telegraph, 18 December 1993; Clarke 1999). We interviewed a former agent of a Canadian life insurance company who had established his own insurance brokerage. He estimated that companies depend on replacement policies for about half of their new premium revenue, 'so it's not replacement to better the situation [of policyholders]. It's internal replacements, companies that are reworking the book of business.' He also pointed out that particular campaigns to encourage replacement can be related to financing needs within the company or the wider industry. For example, he described a situation in

The Moral Risks of Private Justice

329

which companies wanted 'to change their internal rate of return on their reserves that they have. So one of their only ways of doing that was to replace their old contracts out to get new contracts in there, so they could change some of the structure with the reserves that they had to keep there.' Churning existing clients into more complicated and potentially less prudent life insurance contracts is compounded by the development of universal and variable policies with what industry insiders call 'feature creep' (for parallel problems in the United Kingdom, see Clarke 1999: 49-50). We interviewed the CEO of a major general insurance company that sold its life insurance company in the early 1990s. He said that this decision was taken because it was 'too costly' to do business. By this he meant that it was not only a highly competitive business with insufficient financial returns, but also one which, as a result, had moral risk costs arising from unsavory sales practices: It got to the point where there were so many life insurance companies pushing life insurance products to people that there was an over-saturation of opportunity and not a sufficient market to buy ... Well the business started to entice, if you want to use that word, sales agents to sell more business, and of course they're paid commission. So then they began to start changing products ... There were so many products, new products that came out. It was absolutely, utterly impossible for the consumer to know what they were buying. And with the enticement of companies, commission-wise, to get business, [they] were going in and replacing existing business with new business. And that began a churning process. That's what happens in a mature market. There are just insufficient sales opportunities and there's too much of an abundance of product.

In various North American jurisdictions, confidence in the life insurance industry has been severely shaken by 'vanishing premium' policies that could more accurately be described as vanishing policies that still bore premiums. This misselling scandal transpired in the high interest rate environment in the 1980s. The 'vanishing premiums' policy required substantial premium payments over the initial years, on the assumption that it would be fully paid up early on. The policy would remain in force for the remainder of the policyholder's life without the policyholder having to pay additional premiums. This type of policy was priced on the assumption of high interest rates over the long term. When interest rates declined sharply, the companies increased the pe-

330 Richard V. Ericson and Aaron Doyle riod for collecting premiums to manage their loss ratio security. Policy holders who baulked at these unexpected extended payments found that the premiums were nevertheless deducted from the accumulated value of their policy. For example, the lead plaintiff in a class action suit involving approximately 400,000 Sun Life of Canada policyholders said that 'without notification, the offset date projections were unilaterally altered by the insurer. In my case, from the originally illustrated twelve years to an astounding additional forty-three years.' In a 1997 newsletter article by the PPI Financial Group entitled 'Illustrations: New Generation of Life Insurance Products Demands Full Disclosure,' it is made clear that development of universal life 'feature creep' products since the vanishing premiums scandal has, if anything, compounded misselling (see also the detailed critical analysis 'Life Insurance Illustrations,' Insurance Planning 4(4), 1997). The vanishing premium policy scandals, class action suits, and extraordinary payments in compensation have not reduced moral risks to the consumer. Instead, they have led to ever more imaginative ways in which insurance companies induce their agents to protect loss ratio security. 'With the complexity and confusion surrounding long-term bonuses, "guaranteed" cost of insurance and expense charges, and the fluctuations affecting basic interest rate guarantees, the potential for misleading illustrations of the UL level COI product is perhaps greater than ever before. This is highly ironic, corning as it does during a time when the insurance industry is under severe attack for alleged abusive marketing practices involving illustrations of older versions of UL and participating policies.' The life insurance industry cannot escape this structure of misselling because the product is not consumer driven. Life insurance is expensive, addresses remote future possibilities that people prefer not to dwell upon, and does not have the commodity appeal of material goods. Therefore it must be sold proactively and aggressively. The commission incentives for being aggressive are reinforced by a sales culture saturated with motivational messages to missell. For example, a widely distributed industry association booklet for agents, entitled The Magic of Life Insurance, stresses that 'agents need to assume and work with the belief that everyone is underinsured ... Life insurance will always be a pressure sale.' A preference is stated for 'benign pressure' in which 'the prospect is told the story of life insurance so skillfully, so orderly, so convincingly that he or she is led to self-pressure. And that's where the pressure belongs.' This statement is followed by 145 'life insurance power phrases' that can be used to put the pressure where it belongs, for example: 'If

The Moral Risks of Private Justice 331 you had a machine that turned out money, you'd insure it, wouldn't you? Well, as far as your family is concerned, you are such a machine.' The only thing worse than a home without a parent is a parent without a home.' 'You're betting that you'll outlive your family, and your family is the stakes.' 'Where else can a person purchase peace of mind, dignity, respect, confidence, and happiness all in the same package.' Commission-induced misselling has been institutionalized in life insurance sales from the outset. Zelizer (1979) documents scathing criticism of insurance sales commission structures in nineteenth century America. An 1877 book, entitled Church of the Holy Commissions, offered 'a searing indictment on the commission system of agency compensation' (ibid. 7). In England, Charles Babbage expressed outrage that the commission structure in the life insurance industry was tantamount to bribery and as such 'ought either to be immediately abolished, or else publicly acknowledged' (Babbage 1826: 141). The commission structure remains a major source of moral risk influencing life insurance sales people, providing incentives for selling practices that put the consumer at risk of making a dubious insurance purchase. Perhaps the most telling example of this point, the exception that proves the rule, came from a company we studied that sold group insurance. This company also sold individual life policies to selected clients who were contacts for group sales. The company had adopted a slogan indicating that they were 'people you can trust.' Asked to explain this slogan, a company executive gave the example that the person who sold these individual policies to key clients was paid a straight salary, rather than any commissions. Scrapping the commissions in this case removed the moral risk of encouraging the sales person to missell to these favoured clients, whose goodwill was needed for the group business. As the company executive told us, This guy's not motivated to sell the wrong product... I couldn't afford to have my individual guy selling the president of the hospital a whole life policy which turns out to be the wrong product and too expensive ... because our guy makes 50 per cent commission in the first year.' With a thirst for premium revenue, insurance companies are tolerant of misselling for the same reason they are tolerant of claims fraud: it is part of the calculus of loss ratio securities. The moral utilitarianism of insurance companies in this regard was addressed by an interviewee who had worked as an internal investigator for a life insurance company. This interviewee, a former seasoned police officer and commercial crime specialist, expressed his obvious shock at both the extent of fraudulent

332 Richard V. Ericson and Aaron Doyle practices in the company and the company's seeming inability to change them. He attributed the situation to 'the mood set and the philosophy of a commissioned sales force,' stating that the company also sponsored 'extravagant award things every year. And I kept asking, "Why are you continuing this? It causes people to cheat and lie."' He provided his own answer: 'Let's face it, I mean every company looks at it, "How much is it costing us to live with this problem as opposed to the cost to fix it?" So as long as the cost to fix it is in excess, we're not going to be concerned about it ... You always have to appreciate the business side of it: I know they're not here to do me a favour today, they're here to make money.' Governing through Sales Fraud Life insurance companies typically respond to market misconduct scandals by denying institutional and organizational responsibility. In public statements as well as internal communications, they sometimes use a 'rotten apple' theory to brand individual agents as unethical and irresponsible. They also resort to blaming the victim, accusing policyholders of unfairly attacking insurers and their agents for their own bad choices. We interviewed a former life insurance agent who said that he sold many vanishing premium policies that turned out to be very disadvantageous to his clients. 'There are people now who dealt with me when I first started, and I didn't do a good job for them because I didn't know what I was doing.' He said that when these products first came on the market both he and the company felt they were reasonable value. However, when interest rates declined dramatically the company went into denial, failing to admit its wrongdoing and to provide remedies. 'Five years before this [vanishing premiums scandal] started there was a problem and everybody knew about it except maybe the consumer.' The company funded the problems through further premiums and charges to policyholders, and then blamed both agents for their shoddy work at point of sale and policyholders for their failure to appreciate what they were buying. Life insurance company internal documents we examined advanced the view that bad agents and disgruntled customers, and not the company governance structures, yield market misconduct. For example, a marketing manual for agents that provided guidelines on advertising and print communications in the sales process stated, 'in all companies there are a number of producers whose questionable practices hurt the

The Moral Risks of Private Justice

333

image of everyone. It is these "bad apples" who have instigated the involvement of the government in regulating insurance market conduct.' This manual also warned agents to beware of consumers who second-guess their own product choices and then transfer blame onto agents and companies. After listing several large legal judgments against life insurance companies, and the threat of more to come, the manual states: 'Class actions commonly result from allegations of wrongdoing carried out by insurance agents and the companies they represent. Consumers are being sold more complex products to cover their life insurance needs and they rely on their insurance agent to act in their best interests and give them the best advice. This opens the door for claims made by disgruntled customers who are unhappy with their choices and are looking to blame the agent for poor advice or poor service practices.' Regulators also tend to attribute sales fraud to the inherent make-up of people rather than how those people are made up by the governance structures of the insurance industry. In the wake of massive life insurance and pension sales fraud in the United Kingdom, the chair of the Savings and Investment Board regulatory agency declared that structural change through regulatory mechanisms would not resolve the problems. Individuals are the source of the problems and they will continue to defraud each other where possible. 'We want to avoid overregulation and too great an emphasis on prescription in all areas of financial services. Some hazards and dangers will always be there. Human nature will not change. People will still commit fraud.' The insurance industry might use sales fraud as a justification for more and better training of sales agents. Professional training rituals may be undertaken where there are serious allegations of market misconduct. For example, some life insurance companies involved in the mid-1990s sales fraud scandals in the United Kingdom were fined and subject to negative publicity. One response to these sanctions was to take the entire sales force 'off the road' for retraining over several months (Clarke 1999: 162). It is now standard practice in most jurisdictions for regulators to require agents to obtain a minimum number of educational credits each year in order to retain their sales licence. However, credit is typically given for attendance at marketing association luncheons and meetings where the presentation is by another sales agent promoting a new insurance product or sales method. An executive of a life insurance brokers' association explained in

334 Richard V. Ericson and Aaron Doyle interview that 'the system wouldn't work if it actually required training.' By this he meant that the agent herself is a commissioned seller in a buyer/seller relationship, not a well-trained professional in a client/ professional relationship. The commissioned seller is paid handsomely for sales made, and nothing for sound advice per se. She herself is churned out of the system as soon as she has exhausted her prospects for premium revenue. A more strident effort to govern through fraud is made when the company is looking for a justification to 'rightsize' its sales force. An industry document we examined suggested that companies use market misconduct scandals to rid of unproductive agents and thereby display company integrity. Referred to as the 'damage control cycle,' the recommendation was to introduce 'stringent policyholder conservation measures.' All policyholders should be contacted directly and advised to remain with the current agent. This would give the company time to ascertain whether the agent had churned or otherwise defrauded the policyholder. The company would offer the policyholder a 'hot line' to answer questions and provide 'objective' premium comparisons. The company would then take action against agent wrongdoers through both public (e.g., publicity, lawsuits) and private (e.g., dismissal) mechanisms. The tactic was to blame agents and hold them responsible in the hope that at least some policyholders would continue to believe the company was not at fault and could be trusted. Insurance companies also use fraudulent sales exposure to justify the introduction of new surveillance technologies to govern the practices of sales agents. In some jurisdictions databases have been established by regulators regarding the licensing status, educational credits, and disciplinary record of insurance sales agents. There are also various databases that compare agents' specific underwriting practices. Agents are also being governed through computer-based sales illustration packages. These packages control the information that agents are able to present to sales prospects and record what has been presented. For example, life insurance companies now script the sale of more complex investment-related products by using such computer-based packages. In one package we studied insurance limits are preset and not editable. Compliance standards are embedded in the information presented and assumptions made. Compliance-related information pages appear on the agent's computer screen if she does not access them herself. The agent is not allowed to access fields that would allow her to change assumptions about the investment returns on the product. The

The Moral Risks of Private Justice

335

technology thus 'covers' the agent via compliance features built into the system regarding insurance limits, illustrations, and assumptions. At the same time it provides a company database regarding the agent's interaction with the prospect, indicating, for example, whether cross-selling opportunities were seized. The technology creates an impression of professionalism, even if professionalism in training and standards is minimal. Insurance company auditing of the sales process is even tighter in the case of the direct sales call centre operations that are proliferating. These operations, in keeping with call centres more generally, provide continuous surveillance of every sales act and ongoing performance calculations of the agent's sales record. Sales calls are recorded for playback if problems arise, for spot-check inspections, and for ongoing training in effective sales practices. In an insurance sales call centre we examined, there was a weekly 'call observation tour' involving the sales team coach and her sales agents individually. Each agent went through selected tape-recorded sales calls with the coach to assess the call against a fifty-point checklist, with items such as 'speaks with a smile on their voice' and 'makes at least two attempts to bind the business in the call.' This call centre also posted weekly performance statistics around the office, listing the percentage of sales 'hits' by each agent within the team in baseball batting average fashion, as well as a league table comparing the weekly performance of each team. It is difficult to imagine a more directly surveilled and audited environment. In life insurance sales a follow-up survey of a purchaser is made in the weeks after a sale to ensure that she continues to appreciate what she has bought. Such surveys, governed by regulatory agencies, may be supplemented with a company's own customer satisfaction survey that addresses agents' conduct. A life insurance company we studied had a consumer complaints' registry, and held a quarterly review of agents subject to complaint. It also had a registry of replacement sales, including a 'quarterly report on different forms of churning activity with dividends to pay premium, early conversion, lapse/surrender followed by new issue, loan on old policy to pay premium on new policy.' Insurance companies also conduct undercover surveillance of their sales agents through 'mystery shopping.' An undercover investigator poses as a customer and records the transaction not only for evidence of possible market misconduct but also to monitor professionalism and efforts at cross-selling. Special assignments focus on particular problems of systemic fraud. We interviewed an investigator who worked for a

336 Richard V. Ericson and Aaron Doyle management consulting firm. He described how his unit 'puts together undercover operations where we will pose as unwitting buyers of policies, and check on the suspected agent. This is not random virtue testing; we don't run out and just check any agent. But if you have a suspicion - someone who is selling a wrong service, or just selling to old people, and where they're cashing in on their term insurance -we'll put people out, video of their room, and do what has to be done to see if this person is giving your clients the straight goods.' In the United Kingdom, mystery shopping of sales agents has been undertaken by a number of organizations beyond insurance companies. In response to life insurance and pension fraud scandals, magazine and television investigative reporters, consumer associations, and state regulators all adapted mystery shopping as a way of documenting the systematic nature of sales fraud (Clarke 1999: 79, 99, 135-6). Claimant Fraud

Claimant Fraud through Governance With respect to insurance claimants, fraud is fostered in the context of four related considerations about governance. First, evidence that would justify treating a claim as fraudulent is usually difficult to establish. Evidence problems are especially pronounced in particular fields of insurance, such as personal injury claims. A claimant who says she is suffering from whiplash following an automobile or work-related accident is a case in point. Medical science is notoriously unable to diagnose and treat this problem, and insurers are at a loss to determine where suffering ends and exaggeration begins, or where malingering ends and fraud begins (Malleson 2002). Second, given evidence problems, it is simply better for loss ratio security to make a 'nuisance payment' that closes the claims file on a suspected fraudster. The indemnity principle is sacrificed to expedience (Ross 1970; Heimer 1985; Baker and McElrath 1997). Moreover, the claimant's insurance contract conditions can be legitimately altered after the claim to transfer more of the risk to her - for example, through higher deductibles, lower limits, and more exclusions - and premiums can be raised to recover the claims cost over time. The field of automobile insurance claims for material damage exemplifies the way in which nuisance payments are institutionalized in the process. Body shops routinely report that what was designated as claim-

The Moral Risks of Private Justice 337 able damage is actually old damage, but the insurer decides the cost of sending out an estimator to investigate further is not worthwhile. Claimants who have their stolen vehicle returned frequently use the opportunity to repair pre-existing damage. We interviewed a fraud investigator for an auto insurer who described this scenario, and how claimants are dealt with when confronted with fraud. Whenever someone gets a car stolen, at that point it is an opportunity for that person to get a lot of little things fixed on the car that were damaged already. And if we can't prove it was not damaged before, we give them the benefit of the doubt. So we grind our teeth, bite our tongue, and say we'll fix it. It is like a house insurance policy. You get broken into, your TV gets stolen, your VCR gets stolen, might as well claim the golf clubs at the same time because it is costing you $500 [deductible] anyway - golf clubs you never had ... [When a claim is denied] we give them a way out. Sometimes we get them to try to admit it... give diem a softer approach ... If you tell us you were involved with this, that you made a mistake, we will just leave it at that. We won't process your claim, but you just won't have a claim. We won't charge you with anything ... When we are busy we are trying to get closures ... It is time management. Do I look, or don't I go? Nah, I'm too busy, I'll just sit on my butt and do another claim, pay it.

The claims process is a micro-negotiation of political economy and fiscal responsibility. The insurance adjuster's starting position in claims settlement varies considerably according to the field of insurance, the type of claim, and the market segment of the claimant. However, governed by loss ratio auditing, quotas, and penalties for exceeding specified claims levels, adjusters may try to draw the line for negotiation at minimal levels. This minimalist position is based on the knowledge that fraud is institutionalized in the insurance relationship. The assumption is that first, the claimant makes an inflated claim knowing that the adjuster will try to minimize it. Second, the claimant knows that if she persists to a reasonable degree, she may be offered a nuisance payment because the adjuster needs to close the file efficiently. Third, if she is reasonable in finessing her claim, the claimant may be able to garner extra compensation for her reasonableness. As we shall describe in more detail later, insurance claims service providers are often brought into the negotiation process to vicariously magnify and mollify claims. This aspect is evident in personal injury claims. A former adjuster for an automobile insurer said that his compa-

338 Richard V. Ericson and Aaron Doyle ny's routine practice in personal injury claims was 'to deny until they hire a lawyer. So there's no question they were entitled to benefits, it's just denied them and hope they go away.' In this context, the claimant must hire a lawyer as a nuisance to extract nuisance payments from the insurer. Lawyers are regulars in the insurance claims process, whereas claimants are one-shot players (Galanter 1974). It is often lawyers who take the initiative in encouraging claimants to amplify their claims in pursuit of a more lucrative settlement. Lawyers also draw health professionals into the claims-building process. A bodily injury litigation manager for an automobile insurer said in interview that lawyers 'build claims quite easily if you pick the right professionals and right doctors. And the lawyer sends them to certain doctors, the claim goes up, and the longer it drags out the better the chance there is that we're going to be paying more money.' The insurer counters with experts who will deconstruct what the plaintiffs' professionals have built. The moral risks posed by the insured are entwined with those of their legal representatives and health service providers. The governance of moral risk becomes multifaceted, involving the insured, lawyers, health professionals, and insurance claims operatives in an elastic struggle among truth, trust, tolerance, and expediency. Relationship management is the third component of how claims fraud is fostered and tolerated. In keeping with broader marketing trends with respect to relationship management (Slater 1997; Turow 1997), insurers foster special relationships with lucrative policyholders over time. The special relationship may include turning a blind eye to possible fraud. Policyholders who have falsified their claims will contribute to ongoing loss ratio security by providing additional premium revenue over time. They may also be customers in other lines of insurance that are especially valuable, leading the insurer to overlook a particular instance of a fraudulent claim. A former senior vice-president of a multinational insurance company said in interview that he preferred to hire in-house adjusters than to rely on independents, because in-house adjusters are more sensitive to relationship management when making judgments about claims. He said that in-house adjusters are socialized into 'your culture, your philosophy,' where 'service is the name of the game.' In-house adjusters know how to minimize or deny a claim more diplomatically, including claims that entail fraud. They know when to pay claims to avoid conflict, always half watching in terms of the future premium revenue the claimant will contribute to that side of the loss ratio ledger.

The Moral Risks of Private Justice 339 There is substantial variation across moral risk market segments. A substandard market customer who pads a claim is less likely to be maintained as a policyholder than a superstandard customer. We studied a company that operated exclusively in the superstandard property and casualty market. Part of its marketing strategy, made explicit in its advertisements, was to offer a superior claims service, which included in-house adjusters who projected professionalism and a 'no-hassle' replacement of expensive losses. A claims manager for this company said this approach was possible because of high premiums and relationship management associated with those premiums. Taking claimants more at face value not only justifies the higher face value of insurance policies, but hopefully a view among claimants that, '"Maybe I don't need to overinflate that claim, or maybe insurance companies aren't as bad as I thought they were." And that's another way of combating fraud. I mean obviously we don't intend to pay anything more than we have to, but there are sometimes we do ... When people discover that you're not going to nickel-and-dime them to death, they tend not to do it to you. But ultimately it comes down to the relationship, or the conversation between two individuals: how do you feel about the person you're talking to?' An additional aspect of fraud tolerance as part of relationship management is the possibility that an allegation might prove wrong. Given the equivocality of definitions of fraud and associated evidence problems, allegations of fraud may be deemed false or unfair. It is usually easier not to bite the hand that feeds, even if one's own hand is nibbled at. The fourth component of the process by which claims fraud is fostered and tolerated relates to broader considerations of social and economic relations. As we noted above with reference to Garland (1997), intensified policing of fraud may seriously disrupt the smooth flow of social and economic relations. For example, workers' compensation insurers face sensitive labour and management issues that make them reluctant to police fraud too heavily. An executive of a workers' compensation insurer made this point in interview. He said that while an auto insurer might be successful in initiating a crackdown on some aspects of disability insurance fraud, including publicity campaigns, he was not able to do so. 'Labour would consider it an affront to suggest that workers would be defrauding the system. They do not believe that moral hazard is a significant issue ... and [they do believe] that before you deal with that you should deal with [employers'] claims suppression, for instance.'

340 Richard V. Ericson and Aaron Doyle Governance through Claimant Fraud Precisely because they know that claims fraud is a product of their governance practices, insurance operatives suspect every claimant. The insurance governance system constitutes the private justice category of 'situational man,' who is 'the economic subject of interest ... a moderately rational, self-interested individual ... a consumer who is alert to criminal opportunities and responsive to situational inducements' (Garland 1997: 190). The insurance consumer has paid very substantial premiums for a long period - which have swelled the investment coffers of the insurer and been distributed to other insured in the pool who have suffered losses - and therefore a payback of extra magnitude seems justified in the event of a claim. The sensibility in this regard was articulated by an executive of a property and casualty insurance industry association: If people have a sense of what their responsibility is, then we're going to have fewer disappointed customers ... [A responsible policyholder is] an individual who understands the contract started in June of one year and ended in June of another and provides a peace of mind for that duration. If you didn't use it you get exactly what you paid for. A person who doesn't have a sense of that, after twenty years of paying for something you can't hold and you can't see, and you can't show your neighbours, is a person who is going to come into being an opportunistic fraudster. He will inflate the claim, cover the deductible. He will turn the Timex into a Rolex; he will do something to get his money back because he sees it still as his money. Not a sentiment he has for anything else he purchases. Claims adjusters and investigators we studied took the view that a policyholder is not satisfied with paying thousands of dollars in premiums every year for various insurances, and expecting only peace of mind in return. When the opportunity arises, even the most respectable person will seek extra benefit. The presumption of fraud is therefore institutionalized in the claims relationship. There is a 'decline of innocence' (Ericson 1994), and the burden of proof is shifted in the direction of the claimant. This shift entails a strong blaming the victim approach. The former chief actuary of a property and casualty insurance company said in interview that all claimants are treated as 'fraudsters,' leading insurers to take 'financially the most expedient view of the claim. I will squeeze,

The Moral Risks of Private Justice

341

extract, cajole; I'll do anything I can to keep the costs of the claim down. And I'll squeeze the customer, victim, in the process.' A colleague who worked in the same company as a claims analyst said that at the height of a traumatic experience, the auto accident victim is blamed for not having prevented the accident in the first place, and then for exaggerating her suffering. The number one complaint people have with the claims process is that they're being treated as criminal: that essentially they're not being given the benefit of the doubt and that they're presumed guilty, and it's up to them to prove their innocence ... Especially after they've had a crash, they need some support mechanism, not being interrogated ... That's what it revolves around is the perception that people don't trust us because they feel we don't trust them, and it's the trust issue that's absolutely central to all this insurance business ... They want to feel that we can deliver on the promise. Because people are not actually buying a product, they're buying security and they want to make sure that if something happens to them that in fact we'll be there for them. It's the gap between the expectations and the actual realities of what happens when you actually have a claim that sets off the negative reactions. Adjusters and investigators make subjective and intuitive moral risk judgments of claimants to decide how to apportion blame for their losses and what compensation is deserved. As such, the insurance operative as moral agent is also 'situational man.' A subcultural term for a claimant who bears moral risks of fraud is 'red.' 'Red' can be read as 'communist,' someone who is a source of social inflation and therefore a threat to loss ratio security. Company manuals and training programs identify a broad variety of 'red flag' indicators. While a multifactoral analysis of red flags can be computerassisted, it also requires the experienced and keen eye of the claims agent. An antifraud claims investigator for an auto insurer reported, 'I've never had a claim that is the same. It is like a snowflake really. Never two flakes the same.' The view that no two cases have the same elements suggests that myriad elements can be interpreted as red flags and used to justify further investigation. It also suggests that investigative information can be produced and distributed for a wide range of governance purposes beyond generating evidence of ever-evasive 'fraud.' Baker and McElrath (1997: 147) found that property claims adjusters 'rely on their experi-

342 Richard V. Ericson and Aaron Doyle ence and intuition to assess whether the claims fit their expectations.' One adjuster told them, '"I look at the car they drive, the other things in the house, the way they carry themselves." Other social attributes mentioned were the character of the home or neighborhood, employment status, business or professional background, immigrant status ... perceived wealth, and what another adjuster called, "a life-style that indicates a basic honesty"' (ibid.; see also Glenn 2000; Baker 2000). Our observations, interviews, and scrutiny of adjusters' manuals revealed six interrelated sources of red flag indicators (for elaboration, see Ericson, Doyle, and Barry, 2003). First, the insurance record of the claimant is a key indicator. Second, employment, income, and credit records are all scanned for suggestions of fraud potential. Third, race and ethnicity are used as indicators of fraud, to target groups believed to be prone to 'scams.' Fourth, there may be incongruities in stories claimants tell about their losses. Accounts of what happened may change over time, or there may be incongruity based on technical evidence. For example, an auto insurance claimant might report the theft of equipment from her vehicle, but the vehicle's wiring is not compatible with the equipment reported to have been stolen. Fifth, the claimant's demeanour in encounters with adjusters is scrutinized for signs of deception. In particular, too little or too much cooperation in the claims investigation are read as signs of deception. Sixth, fraud is deemed more likely when the claimant immediately starts to exhibit 'claims consciousness': an exaggerated focus on claims benefits and/or being too legalistic, which is suggestive that the claimant is trying to extract more than she deserves. Insurance companies and their information suppliers also use data systems to automate red flag sorting, forming a collage of shades from pink to ruby. One operation we learned about used an on-line fraud score system, referred to as a 'fraudulent analysis machine.' We also examined a dataveillance system used by the claims department of an auto insurer. The monitoring of the claimant begins when she telephones the claims centre to report her loss. A graphical user interface prompted by the caller's driver's licence number or vehicle licence number immediately provides information: whether the claimant is insured, the type of coverage, any previous claims, whether other parties have called the centre about the same accident, and whether a more intensive examination of the claim through a special algorithm that analyses claims history is likely to be productive. The claimant's telephone number is also recorded in case she becomes abusive. An adjuster based in the tele-

The Moral Risks of Private Justice

343

phone claims centre explained how the system immediately identifies claimants who are more ruby than pink: [It flags those] who have a poor history with us for payment of claims, for criminal activity, all kinds of different things. [The Special Investigation Unit] can enter a 'hard match' on that individual file. They will flag that file ... When that particular customer calls us with a claim, and they offer their licence number ... the system will do a search automatically against that person in the customer database ... It says this is a 'hard match,' this customer is not a desirable customer, please refer to a special investigations unit right away ... 'Soft matches,' too, which is where the system won't lock up, but it'll show a flag and this agent can go and verify because it's matching against the names, so this is the John Smith that is in fact the bad person ... [If flagged] we don't keep the client ... it gets sent out to the claims office ... [It is all aimed] to save money on those ones that are fraudulent potentially.

The effort to 'save money' on the claims end is linked to restrictions on the ability to 'raise money' on the premiums end. When premium revenue is tight, more stringent governance through claims fraud occurs. Changes within the insurance industry and its systems of governance have fostered a surveillance-based crackdown in the name of fraud control. A U.S. industry executive indicated to us that a boom in fraud policing during the 1990s was driven by increased pressure on insurance companies to keep premiums down in a soft market. In the previous hard market claims costs could more easily be passed on to the consumer in higher premiums. The executive also indicated that new laws in California and other states that restrict premium increases forced the companies to resort to stronger 'anti-fraud' measures to reduce claims costs and to bolster their competitive position and profits. A Canadian provincial auto insurance company we studied was operating under a government-mandated freeze in premium increases. This prompted a heavy crackdown on fraud as one way to contain costs. The response of this insurer, in keeping with other disability insurers facing similar problems, has been to expand its own Special Investigation Unit as well as the surveillance functions performed by its adjusters. An interviewee who did fraudulent claims data analyses for this company said that the intensified surveillance was devised because 'they were squeezed for money, so they wanted it. And it looked like ... a cost-free

344 Richard V. Ericson and Aaron Doyle way. You don't have to raise premiums, you don't have to lower your service, you just get cash. And no one is going to stand up and say, "I'm the fraud lobby," and lobby for keeping the money. So it looked good.' The disability insurance system is fraught with moral risks and attendant threats to loss ratio securities (Malleson 2002; Ericson and Doyle forthcoming: chap. 3). Medical diagnoses and treatment regimes are illdefined, and 'disability' is a creation of the insurance industry's mandate to commodify impairment, pain, and suffering in relation to cultural norms about work and productivity. Insured persons with disability claims tend to consume health services and compensation payments to the extent they are led by experts to believe they need them. But what they need is perpetually at issue, because expert knowledge of disabilities is limited and contested. The response of the insurance system is to crack down on disability claims through the elaboration of surveillance. This response is exemplified in what disability insurers refer to as 'work hardening' regimes. Work hardening is a process of returning the claimant to work at the earliest possible stage in order to control the iatrogenic system and thereby enhance loss ratio security. If an insurer wants to improve loss ratio security through a claims crackdown, then a change in the definition of disability through 'work hardening' governance is deemed the way forward. A key source of loss ratio security problems - professional claims service providers — are made part of the solution. In the absence of effective medical knowledge, health professionals are mobilized by the insurance system to develop moral risk knowledge to govern claimants. For example, they develop moral risk knowledge to diagnose and treat 'illness behaviours' such as malingering and failure to fulfil responsibilities. The very term 'illness behaviour signifies that the illness is doubted: the person is behaving as if she is ill, but this behaviour needs to be interrogated. The simplest definition of 'illness behaviour' is whatever behaviour appears 'dysfunctional' to the clinical eye as it is focused by the insurance system. A manual for clinicians dealing with whiplash claimants explicitly defined 'illness behaviour' as any effort to seek insurance compensation that is unwarranted. This undue compensationseeking is described as 'consciously or unconsciously using or maintaining symptoms to obtain psychological, social or economic advantage ... expressed in behaviour that appears to seek sympathy, manipulate others, avoid work or responsibility, or save face when expectations are too high.' In clinical shorthand, the claimant is said to suffer from 'compen-

The Moral Risks of Private Justice 345 sationistis' (Munglani 1999: 14). This 'dysfunctional' behaviour may actually have a relation to physical medicine. For example, a manual for whiplash claimants points to a range of 'somataform disorders' or 'physical symptoms that suggest a general medical condition that is not fully explained by the physical symptoms.' Among the somataform disorders is malingering or 'factitious disorder,' which is defined as 'A conscious or willful invention or distortion of symptoms to achieve secondary gain, usually a financial gain. It should be pointed out that it is not easy to detect malingering, as it can sometimes appear the same as previously discussed syndromes. The difference between malingering and the other psychological issues is that in malingering the person knows that they are faking an illness or injury, and why - usually for money or to avoid traumatization or abuse.' Another syndrome is 'pain disorder,' where it is determined that pain is largely related to 'psychological factors.' In the extreme form it becomes 'chronic pain syndrome,' which is a 'condition of its own' and not just an acute pain that lasts longer. A lot of suffering is seen as illegitimate and the appropriate treatment is to restrict benefits. The American Medical Association (AMA) (1993) guidelines on the evaluation of permanent impairment state that pain is located within the malleable self and is thus manipulable. The key is to suspect the claimant at the outset so that she can be pointed to the work-hardening route immediately. 'Chronic pain is a self-sustaining, self-reinforcing, and selfregulating process ... a destructive illness in its own right... The patient's maladaptive behaviour may have medical, social and economic consequences that greatly outweigh any somatic components of the illness ... Chronic pain may result from inappropriate management of acute pain ... Early detection and prompt, effective intervention require a higher index of suspicion and are essential to effective management' (ibid. 305). The AMA guidelines instruct that chronic pain is a risk that is to be diagnosed in moral terms. Among the diagnostic criteria are 'dramatization,' 'dependency,' and 'dysfunction.' 'Dramatization' is evidenced by 'words used to describe the pain [that] are emotionally charged, affective and exaggerated. Patients may exhibit maladaptive, theatrical behaviour, such as moaning, groaning, gasping, grimacing, posturing or pantomiming.' Their dependency is initially 'on their physicians and [they] demand excessive medical care. They expect passive types of physical therapy over long time periods, but these provide no lasting benefit. They become dependent on their spouses and families and

346 Richard V. Ericson and Aaron Doyle relinquish all domestic and social responsibilities.' 'Dysfunction' is signified when the claimant 'withdrawal from the social milieu ... [and] becomes an invalid in the broadest sense: physical, emotional, social and economic' (ibid. 308-9). Although the AMA declares that pain is 'a destructive illness in its own right' (ibid. 305), pain is ruled inadmissible for insurance compensation unless it is accompanied by a measurable, physical, functional impairment. The moral utilitarianism of chronic pain diagnostics - making the self responsible for both the 'destructive illness' of pain and its cure through work hardening - ensures that on its own, pain brings no gain. 'An individual who complains of constant pain but who has no objectively validated limitations in daily activities has no impairment. The proper test is not, "Does this daily activity cause pain?" but rather, "Can the patient perform this daily activity?"' (ibid. 309). Work hardening is based on the classic moral utilitarianism of Jeremy Bentham: good people work hard. Instead of placing the onus on the insurance and health care systems to produce better knowledge for diagnosis and treatment, ignorance is confessed. This confession allows the focus to become the production and distribution of surveillance knowledge aimed at getting the person back to work quickly. Surveillance knowledge shifts responsibility onto health care professionals to promote moral risk management of fraud, malingering, and irresponsibility, and onto the claimant, who is expected to risk manage her pain at work. Her prudence entails return to work even if it is painful. The discourse of moral risk management highlights the person's 'abilities' rather than her disabilities, and relative 'wellness' and 'best outcomes' (or least worst outcomes) rather than full recovery in all cases. Thus 'pain and other nonfinancial aspects of bodily injury ... can be understood in insurance terms as an ineluctable, nonmonetary deductible or coinsurance "payment"' (Baker 1996: 278). Claims Services Fraud

Claims Services Fraud through Governance Occupations and professions that provide insurance claims services also participate in fraud as a result of the ways in which the insurance system governs them. We previously observed that disability claimant fraud is entwined with the professional practices of lawyers, health service providers, and insurance claims operatives. Disability insurance

The Moral Risks of Private Justice 347 provision is illustrative of how fraud through governing occurs among claims service professionals. Disability is a creation of the insurance system. Like illness more generally (Sontag 1979, 1989), disability involves a moral battle fought via classification schemes, figurative language, and metaphorical reasoning. Insurers frame the meaning of disability within insurance logics over and against the logics of professional service providers. Insurers are able to achieve interinstitutional hegemony over the definition, diagnosis, and treatment of disability not only because they control the money to pay insurance claims, but also because of the weakness of knowledge claims by health professionals in many areas (e.g., whiplash) (cf. Abbott 1988: 251). Disability is defined first and foremost in relation to work and the income derived therefrom: it is the inability to be fully productive. As such it is based on the prognostic criteria of insurance logics. A workers' compensation board manual on permanent functional impairment stated that while physicians alone determine the level of functional impairment, disability 'is not a purely medical condition ... [It is] an administrative and not solely a medical responsibility or function. It is an appraisal of the worker's present and future ability to engage in gainful activity as it is affected by such diverse factors as age, sex, education, economic and social environment, in addition to the definitive medical factor of permanent injury.' Similarly, the medical profession declares that 'a disability arises out of the interaction between impairment and external requirements, especially those of a person's occupation. Disability can be thought of as the gap between what a person can do and what the person needs or wants to do' (AMA 1993: 2). In summary, disability is an insurance construct for deciding how much productive ability a person has left. As stated by a work and health specialist we interviewed, 'whether somebody is disabled in a short term, medium term, or long term, is a matter of how they get insured, not a matter of, really, a significant difference in health.' Disability provides a framework through which insurers rationalize the individual destinies of work careers. The insurance institution provides the 'special circuits of protected mobility ... within the greater game of the social market' (Gordon 1991: 45). The director of corporate planning for a workers' compensation board in Canada described how insurance logics are embedded in medical standards and guidelines themselves. He referred to a published article in which the editor of the AMA (1993) manual described the standard setting process:

348 Richard V. Ericson and Aaron Doyle 'Well, we sat around a table like this and said what do you think an arm is worth? Yeah, okay, 25 per cent for the arm at the elbow, 50 per cent for the arm and shoulder, that's what we'll decide it is' ... The milieu there was that these disability ratings ... [are] paid on the basis of the number of weeks you're entitled to ... the most common rate is five hundred weeks. So when these people are sitting around the table, they're thinking what this is going to work out to in real terms with the individual, the average individual, let's say thirty-four years of age at the top. So that's what they did, they looked at the average wages, what the average person would get and they come up with a percentage that they thought was reasonable ... [Everything] is engineered in meetings ... The problem is ... there's not a lot of comparability between [AMA guide] editions ... what do you do retrospectively, retroactively, with cases? Maybe we're disadvantaged, or over-compensated to use the rate today. And it's a big worry.

The definitions of disability are also negotiated in the pragmatic contexts of claims settlement. A work and health research expert said in interview that soft tissue injuries resulting from traffic accidents and work-related back injuries, both biomechanical and psychosocial, are predictable and subject to risk analysis. '[We can] predict as well as smoking and cancer ... but there is a bunch of serious insurance problems about how you make rules for what gets compensated for disability, or what is considered to be work-related.' The determination of whether an injury is work related 'is all a matter of who says it is. There is no magic independent technology or device that would give you the truth on this. So if somebody says it was caused by work, it was caused by work.' The person at the forefront of this negotiation of causal attribution is the disability claims adjuster. She takes into account the assessments of health care professionals who are diagnosing and treating a claimant's disability in terms of the structure and culture of the insurance company. Finnegan (1999: 6), an expert on claims management for vehicle insurers, describes the ways in which a disability claims adjuster handles a whiplash case in relation to the 'simultaneous strong and unconscious influence that corporate culture has on whiplash settlements.' For example, 'When she is having trouble meeting productivity goals or managing a high number of pending claims, she might have "fire sales," making quick settlements for higher than usual amounts. Similarly, her authority level - the amount she can pay without her supervisor's approval influences her settlements. A surprisingly high percentage of whiplash claims are settled just below adjusters' authority levels. As a result an

The Moral Risks of Private Justice

349

adjuster with high average authority levels has higher payments than a company with lower levels' (ibid.). Disability insurance adjusters we interviewed said they were faced with a 'flavour of the month' problem. New disabilities are invented and magnified by health care professionals, lawyers, and claimants. The claims adjuster's task is to not only assist the claimant and those who serve him, but also to adjust the insurance definitions of disability to help maintain interinstitutional hegemony. We interviewed a member of a special claims team for an automobile insurer that was trying to adjust the definition of disabilities in this regard. 'We call them the description of the month. You'll hear about fibromyalgia claims ... TMK, the TemporaMantibular joint issue where you have an auto accident and it affects your jaw ... Popularity, they're described for awhile and a year later no one is talking about them anymore. We've seen about twenty of those in the past ten years ... We have an insurance company with high liability limits and trends developed that way.' A manager of litigated claims for the same company also reviewed various 'popular' diagnoses used in personal injury cases to argue how 'arbitrary' they are and how he had to fight the 'allegations' to protect company interests. For example, the company had about ten times the North American average for brain injury claims. His job was to hire neurologists and psychiatrists who could counter the neuro-psychologists used by plaintiffs in order to deflate claims of brain injury as a form of depression. He said that claimants 'believe that they have it, they really do ... because they've seen a ton of doctors and stuff.' The problem is accentuated among claimants with a history of 'emotional problems' or 'some abuse' who 'magnify and focus on this one incident as being a major cause and problem that they face on a daily basis.' Given that it is defined and compensated in such elastic and situational ways, disability poses substantial moral risks for insurers at the hands of claims service providers. In particular, the disability insurance system is exposed to the iatrogenic tendencies of professions to expand their jurisdictions, privileges, and material benefits to the point where they become 'disabling professions' (Illich et al. 1977). Professional service providers build claims through questionable diagnoses, unnecessary treatments, and over-billing (Malleson 2002). The insurance industry responds with its own diagnostic labels, including that of 'social inflation.' Social inflation refers to a culture of unreasonably high expectations regarding both what insurers should compensate and what their responsibility should be for loss prevention. The concept gains currency among

350 Richard V. Ericson and Aaron Doyle insurers who experience escalating claims costs and the need to improve their loss ratio security. For example, a consultant's report to the Insurance Corporation of British Columbia (ICBC), the government automobile insurer, uses the concept to address 'the tendency within society to demand greater entitlements for injury or inconvenience and to a growing element of litigious behaviour' (ICBC 1996: 27). Social inflation results in 'the claims process being viewed as an avenue to achieve financial gain as opposed to a mechanism for ensuring the recovery of good health and replacing real economic loss' (ibid. 22-3). The escalation of claims payments is attributed in particular to 'a greater propensity for maximizing claims for non-economic losses (i.e. "pain and suffering") ... increasing expectations for payment after collisions, regardless of actual nature of injury ... [and] increased levels of advertising by lawyers and a greater likelihood of claimants seeking legal representation '(ICBC 1997: 4; see also ICBC 1996: 204). Social inflation and legal inflation are entwined: 'the dynamics of the liability system have made it possible to get relatively large awards for pain and suffering ... [and] led to a growing expectation of such compensation when minor accidents occur' (ICBC 1996: 21). The cumulative result is that '"pain and suffering" awards have increased dramatically resulting in auto premium increases of 138 percent during the past decade. This percentage is nearly twice the consumer price index increase (66 percent) during the same period' (ibid.). The discourse of social inflation is intended to govern professionals' claims about pain and suffering. As gatekeepers of disability, health care professionals sometimes shut the gate behind those who have entered the system, and they may even lock it. Thus the Physical Medicine Research Foundation, funded by ICBC, makes the following declaration in its promotional mission document: The Canadian Centre for Occupational Health and Safety recently reported that the total cost of muscle and joint injuries in Canada has climbed to some eight billion dollars annually. This cost to our society extends beyond the financial aspect to the enormous losses in human terms. Consider the immense welfare and worker's compensation payments made just to people permanently disabled by back pain. Rather than helping people return to normal function, these payouts may be perpetuating their lives of miserable pain ... this lack of sharing, cooperation and coordination in the health care field has served to compound the problems, increasing the medical costs and the pain for millions of people.

The Moral Risks of Private Justice 351 Lawyers are also depicted as socializing claimants into disability for pecuniary reasons. We interviewed a manager of a special investigatio team that was established by an automobile insurer to reduce personal injury claims costs. He focused on lawyers, who will propose,' This is what you've got, and this is what it is worth, but to get it to be worth that you have to be off work for six weeks. And you'd better go to the physiotherapist for ten weeks. And you'd better do this and this.' And then in the psychology part, I would think with claimants it is, 'Okay, I want to get this money, I better make sure it is worse.' And I think they almost talk themselves into this sense of a big circle, and then they get worse and worse and worse. And of course they're not going to get better because it's that, 'Okay, I'm greedy. I want more money, to get there I have to do this.' And then in their own mind they are sick, and they can't get out of that circle. And I think that we perpetuate a lot of that.

An insurance claims control expert with experience in many jurisdictions said that claims inflation for minor injuries is commonplace. She was hired by the attorney general's office in an American state jurisdiction to audit the claims of the top ten vehicle insurance companies operating in the jurisdiction. Eighty-five per cent of our whiplashes were third party ... All the whiplashes claimed on first party were major accidents ... [for which] the average number of treatments was three. The third party ... running thirty or something! We looked at the average cost per treatment, it's 50 per cent higher. We multiplied that out and we concluded 94 per cent of whiplash was 'tort somatic,' the tort system generated it. My number for British Columbia might be 97 per cent... There is such a thing as whiplash, you can get it, but while all objective injuries are going down, whiplash has gone zooming up. The level of treatment has no medical basis whatsoever ... I mean is this organized fraud? ... Whiplash basically is the brain, there can be endless complications with it ... a multi-billion dollar industry ... with no scientific basis at all ... I think it's promoted in a wide variety of ways. [After mentioning lawyers and doctors advertising during televised soap operas]. Anybody who needs to invent injuries to get the money has to make them.

A personal injury lawyer we interviewed was explicit in stating that he did indeed socialize clients into their disabilities for insurance claims purposes. As an insurance-based category used to settle claims, disability

352 Richard V. Ericson and Aaron Doyle requires lawyers to fit clients to the category in order to achieve a just and justifiable outcome. He said that 'brain injury' clients in particular are typically unaware of the ramifications of their condition and thus benefit from his expertise in finding medical formulations that also meet legal requirements for lucrative and just settlements. 'So when I meet some of these people and I say, "How are you?" they say, "Oh, I'm fine." "Any problems?" "No" ... If anything my clients underestimate or don't even recognize that they've got problems.' It is the insurance industry that makes various professional experts part of its system of 'organized fraud' even as it tries to regulate them for loss ratio security. In the field of disability insurance, professionals are compelled to help commodify highly subjective complaints into objectified forms for purposes of compensation. Everything must be treated as if it is objective for the practical purpose of deciding whether, how, and how much to compensate a claimant. When something becomes regarded as objective, it is treated as such: objective knowledge is a warrant for action. Thus it is not surprising that professionals treat subjective disability injuries seriously, authorize the claimant to be off work and seek treatments, and amplify claims to the point where they themselves become moral risk suspects who are accused of fraud and deemed a threat to loss ratio security. Governing through Claims Services Fraud

Just as insurers promote preventive medicine among their insured populations, so they try to prevent medicine among health care providers for fear that too much surplus value will pass into their hands. Governing the moral risks posed by physicians and other health care providers has always been a key component of insurance operations. For example, Porter (1995: 38) reports that in 1849, one William Sanders appeared before a parliamentary select committee in Britain regarding the operation and solvency of insurers known as friendly societies. Sanders emphasized that the solvency of a life and health insurance company depends much more on governing moral risks posed by the insured and the physicians attending her than on the accuracy of actuarial tables. Sanders explained 'how he kept the Birmingham General Provident Institution solvent. Tables giving rates of sickness were important, but the crucial element was strict rules, to define the bounds of appropriate sickness.' As we have already demonstrated in the case of disability fraud, 'strict

The Moral Risks of Private Justice

353

rules to define the bounds of appropriate sickness' are difficult to ground in medical knowledge. The Quebec Task Force on Whiplash Associated Disorder (WAD) reviewed several thousand published studies and concluded that only a handful met a reasonable standard of scientific methodology. Faced with having to make recommendations on the better management of WAD, it adopted a simple philosophy: 'prudence in the absence of evidence' (Spitzer et al. 1995: 40). Prudence meant that both highly variable 'clinical judgment and individuality of patients ... should not be taken as excuses for a laissez-faire, highly variable approach to management' (ibid. 38). The pragmatism of the Quebec Task Force in this regard extended to the point of defining recovery from WAD as occurring when insurance compensation is terminated! As a physical medicine research specialist remarked in interview, since there is no clinical specification of normalcy 'they had to use something.' Disability claims specialists we interviewed pointed to health care providers as the major source of fraudulent claims and attendant social inflation. Typical was a personal injury adjuster for a vehicle insurer, who said his number one problem is 'sympathetic doctors ... a lot of doctors take advantage of us, they know they're gonna be paid as long as there are claimants, as long as their patients keep coming back.' The response is to use health care services fraud as a justification for a more elaborate surveillant assemblage to govern their practices and, by extension, claims building by the insured. The goal is to make disability claims service professionals less sympathetic, to persuade them to offer less treatment, and to enforce work hardening. The first step is to govern more stringently who is authorized to provide particular services. For example, designated assessment centres for disability claims have been established to carefully select service providers and build in tight surveillance mechanisms. A state regulator of vehicle insurance told us that designated assessment centres were established in her jurisdiction to target 'insured doctors': physicians who are willing to diagnose and treat according to the dictates of the insurance company in each case. The comfortable life of an 'insured doctor' is possible because 'there are so many controversial diagnoses that exist out there and especially around soft tissue injury ... but as well around psychological disorders.' We interviewed the CEO of an insurance company that operated in this jurisdiction. He praised the loss security ratio advantages of these centres in providing 'far less costly ways of ensuring that people aren't cheating in the system, then you can take the overhead cost out.' They do so by reducing the need for a large and dis-

354 Richard V. Ericson and Aaron Doyle persed surveillant assemblage, since surveillance can be more concentrated through the centres. A second step in intensifying governance of health service providers is to make them more aware of the iatrogenic effects of their practices, and consequently more self-governing. This approach is evident in efforts to check the proliferation of WAD. A public insurer of vehicles we studied funded a major initiative to educate physicians and other health service professionals in this regard. A curriculum was developed, along with practice guidelines, conferences, and training sessions. A manual to guide practitioners offered various tips on how to avoid over-diagnosis and excessive treatments that might compound rather than correct problems. For example, there was a warning about over-prescription of drugs that might inure claimants to their disability rather than cure them. Pain killers and other drugs should only be prescribed where they support getting the claimant back into working order: • Use time-contingent rather than pain-contingent schedules for medications and reactivation ... Pain-contingent schedules label pain as dangerous as well as aversive ... It is felt that a pain-contingent analgesia schedule has the potential to 'reward' pain, whereas the time-contingent dosing does not. • Drugs should not be used to escape the drudgery of pain, but rather to facilitate functional recovery. • Be cautious of a patient response whereby analgesic is used to facilitate opting out of responsibilities or activities. Be careful of the response: 'I have to take painkillers so I can rest (or feel good). Rather the response should be: 'I use the pain medication so I can get things done.' This example points to the third aspect of governance of and through the health services provided. As discussed above, the physician and other providers are constituted as claims investigators on behalf of the insurer. For example, in the name of better medical knowledge about the injury, the physician is to serve in effect as an accident investigator. 'Since there is a correlation between impact magnitude and the amount of energy applied to the body during a collision, it is worthwhile to ask the degree of impact, the amount of vehicle damage, the type of impact (rear-end, front-end, side-on, etc.), where patient was seated, position of headrest, use of seat belt etc.' Physicians are also to serve as agents of surveillance in the patient's return-to-work plan. For example, they are

The Moral Risks of Private Justice 355 instructed to 'be aware of patient's general home and work responsibilities and to take these into account throughout the work-hardening process.' A fourth step in tightening the governance of health service providers is to change compensation arrangements that affect diagnoses and treatments. For example, regarding WAD, a vehicle insurer we studied recommended a change in compensation of health service providers that would reward them for surveillance that identifies and checks claimants at risk of chronicity. In one jurisdiction we studied, physicians received higher levels of payment if they billed for an injury through the provincial workers' compensation scheme rather than the provincial health care plan. A workers' compensation official told us in interview that this led to some physicians 'saying it's work-related when it is not - they know they'll get paid more.' The situation required remedy through coordination of the service fees across health compensation schemes. A fifth approach to the governance of health service providers is to refine the forms on which they classify and report their medical activities (for parallels in the police world, see Ericson and Haggerty 1997: chap. 15; Ericson and Haggerty 2002). For example, since accounts of WAD and what to do about it vary so widely, one governance mechanism is to develop a new reporting format that will formalize consistency. We interviewed a claims specialist for a vehicle insurer who elaborated on the rationale for this 'epidemiological surveillance' approach. What is important for insurers' management of their loss ratio is consistent classification, regardless of whether it is otherwise valid or reliable from the viewpoint of expert knowledge in the health sciences: Through establishing a common approach to soft tissue injury it's possible for these doctors to say, 'Well, okay, based on the way that we look at this kind of injury, here's how I would classify them.' Once classified, then you could pretty well leave it to anyone else, to debate what that classification is worth ... We reached an agreement with the medical association on the format and style of medical reports for these kinds of injuries, so they complete them in a way that asks them questions about the injury so it can be classified ... It's a centralization and an attempt to describe or classify injuries of the subjective type because if we don't do it, they're all over the map ... We're at the point where we would take any classification system as long as they would all agree to it.

Agreement on classification is central to more stringent governance.

356 Richard V. Ericson and Aaron Doyle In the words of another interviewee from the same insurance operation, it constitutes 'best practice' as simply, 'I do it, and others do it too, so therefore let's do this.' In making health service providers more predictable in this sense, agreement on classification allows the sixth step to be taken in tightening the governance of health service providers: the introduction of new database surveillance and auditing technologies that monitor best practices and prices. In one Canadian province we studied, the public vehicle, workers' compensation, and health care insurers developed an integrated database for surveillance of both health service providers and claimants. For example, the vehicle insurer had the capacity to match its personal injury claims database with the workers' compensation and health care databases to ascertain the treatment pattern of physicians involved in claims, including pharmaceutical practices. All vehicle personal injury and workers' compensation claims were processed by the health care insurer. This arrangement saved the vehicle insurer and workers' compensation insurer about $7 million annually. The health care insurer deemed this to be a reasonable price for centralized control over information that would enhance its capacity to build a more complete patient history for all residents of the province and to monitor the billing practices of health care providers. This 'Billing Integrity Program' was electronic and under development to allow real-time auditing to identify problems immediately. 'Billing integrity' is a major concern of personal injury and health care insurers. This health care insurer had approximately 5,000 rules regarding the processing and adjudication of claims. Some rules were very clear about the extent to which the insurer was willing to compensate a physician. For example, 'aggressive doctors' were governed by a patient quota restriction that paid the full fee for services to forty-seven or less patients per day, only 50 per cent of the fee for services to the next thirteen patients, and nothing at all for patients in excess of sixty. If a physician was an outlier in terms of performing a particular procedure much more frequently than other physicians, an investigation was initiated. Suspicious patterns were also the basis for other types of in-depth investigations: for example, into whether the billed procedure was the one actually executed, or whether the medical procedure executed was really necessary. Other forms of database surveillance pervade the disability claims process. For example, disability insurers try to measure impairment

The Moral Risks of Private Justice 357 levels for compensation as well as to detect fraud by having claimants undergo a series of physical movement tests that measure their ability to squeeze, grip, lift, and so on. There is now a sophisticated computerbased measurement technology that performs reliability checks on each physical movement tested. Each movement is tested three times and cross-checked against both other physical movements tested and selfreported levels of pain. This technology has developed not only as a kind of lie detector of the body, but also to govern unreliable health care professionals who vary widely in assessments employing more primitive means. It addresses limits to expert knowledge in assessing functional impairment and pain by substituting knowledge of moral risks posed by health service professionals charged with quantifying the level of disability and compensation awarded. It promises to make them more objective and reliable in their moral classifications, and therefore more vigilant in their governance of disability claimants. A seventh step in more stringent governance of health service providers involves new insurance arrangements regarding their professional liabilities. All of the managed care mechanisms that we have addressed to this point are creating a new array of errors and omissions and other professional liability exposures for health service professionals and the organizations that employ them. Sensitive to these exposures, insurance companies underwriting this field place additional restrictions on professional practices. According to one major American insurer in this field (Sedgwick 1997: 24-5), there is an improved environment for governing health service professionals through liability insurance, which makes the underwriting conditions more favourable. The improved environment includes the fact that the number of physicians working as independent practitioners in the United States decreased from 75 per cent in 1983 to 50 per cent in 1996, making them more controllable through the centralized systems of managed care operations. Managed care operations have introduced new surveillance procedures, resulting in the 'mandatory nature of risk management and its consequent motivation as a discipline in most health care environments ... increasing effects of peer review pressures ... increasing oversight of both the cost and quality of care that is provided by managed care entities ... [and] increasing oversight by government and industry groups that focus on identifying health care practitioners (e.g., the national practitioners data bank, state and national medical societies, etc.).' The seven steps described illustrate that objectifying experts' judg-

358 Richard V. Ericson and Aaron Doyle ments and making them appear factual through surveillant assemblages is deemed crucial for governance of moral risks and preservation of loss ratio security: Arbitrary power appears to have been toned and liberalized through neutrality and objectivity ... But if experts have, in the process, been rendered governable, this has changed expertise itself: financial vocabularies, grammars and judgments have infiltrated the very terms in which experts calculate and enact their expertise. And the apparent transformation of the subjective into the objective, the esoteric into the factual masks somewhat the weak knowledge base - the uncertain status, inescapably partial erosion, lack of evidential support, history of failure, vulnerability to changes in fashion and convention and much more - of the new forms of expertise granted the power to objectify. (Rose 1999: 153) The Moral Risks of Private Justice

In this chapter we have used the example of insurance fraud to illustrate how potentialities for economic crime are addressed through private justice mechanisms. The private justice system of the insurance industry protects loss ratio security by developing surveillant assemblages that continuously identify and classify moral risks. How the resultant knowledge of moral risk is acted upon varies according to the specific market conditions and contexts of loss ratio security. If loss ratio security is not unduly threatened by particular moral risks, or if it can be addressed through other mechanisms, such as increased prices and insurance contract conditions, then the moral risks will be tolerated in the interest of the smooth flow of social and economic relations. If loss ratio security cannot be addressed through these other means, the surveillant assemblage will be elaborated to yield more precise knowledge of moral risks and to justify a crackdown on selected contributors to them. In this view moral risks are a product of the private justice system and used as well to legitimate adjustments to market conditions affecting loss ratio security. We have shown how sales fraud, claimant fraud, and claims service provider fraud are each a product of the governance mechanisms of the insurance institution itself. Fraud occurs through governance because it is variously made, managed, suppressed, and visualized on the terms and conditions of insurance relationships. At the same time, governance occurs through fraud because fraud is

The Moral Risks of Private Justice 359 used to justify the elaboration of surveillant assemblages that serve an array of governmental purposes beyond regulation of fraud. In the name of fraud, some sales agents are labelled as rotten apples and removed from their position, in the hope that their removal will morally cleanse the rotting barrel. Those still in place are subject to enhanced surveillance which makes them more responsible for loss ratio security, both in the form of hustling for more premium revenues and in being penalized for a poor claims record on insurance policies they have underwritten. In the name of fraud, the insured are subject to fine calibrations of risk which segment them into increasingly specialized insurance pools and result in structured inequality. In the name of fraud, claims service providers are subject to new surveillant assemblages which govern through objectification, factuality, and standardization. At the same time, service providers are embedded in these new surveillant assemblages, charged with greater responsibilities for both self-governance and tighter governance of the claimants they are supposed to serve. This is a private injustice system because it poses other moral risks. There is a differential response to fraud. Fraud by company agents is tolerated to a large degree. Fraud by service providers is tolerated to a considerable degree. Fraud by policyholders is tolerated to some degree, if the client is otherwise respectable and a market segment target for financial services relationship management programs. Policyholders who cannot contribute to loss ratio security in this way are governed through fraud to alter the terms on which they participate in the insurance relationship, or to exclude them altogether. In insurance, differentiation, segmentation, and exclusion are simultaneously processes of marketing and moral risk assessment in underwriting. Preferred moral risks are doubly desirable as insurance clients: they are seen to be both affluent customers and less risky in terms of claims. However, insurers also profit by pooling substandard risks, as insureds in the resulting pool, with little market choice, are compelled to purchase insurance under the most substandard of arrangements. A private insurance company is not in the business of redistributing resources among the insured, but rather of discriminating in favour of those who contribute to the goodness of the pool and the prosperity of the company. It would be acting immorally if it did otherwise. Thus it constitutes communities of interest, user-pay communities into which entrance is purchased with cash, not collective sentiments. Instead of social solidarity and community as 'a moral field binding persons into

360 Richard V. Ericson and Aaron Doyle durable relations' (Rose 1997: 6), insurance increasingly fragments populations into selective moral risk-rated communities with a price tag. Again, this tendency marks the confluence of, on the one hand, niche marketing by insurers to survive in the marketplace and, on the other, intensified efforts to cope with moral risks they themselves produce through ever more finely tuned risk assessments of populations. * As we have documented, the insurance private justice system takes a strong blaming the victim approach. The onus is placed on each and every participant in insurance relationships to contribute to the surveillant assemblage as it fosters loss ratio security. This emphasis on individual responsibility in turn leads to the view that there should be less benefits for those who fail to contribute fully. In this respect all difference, and the inequalities that result from it, is seen as a matter of moral risk taking and management. The person who is not a fully responsible articipant is a 'fraud' even if he or she does not commit it. Meanwhile those who do commit fraud - but are otherwise deemed worthy within the moral utilitarianism of the insurance institution - continue as valued participants. Note 1 The data presented in this paper are derived from our research program on insurance and governance (Ericson, Doyle, and Barry 2003; Ericson and Doyle forthcoming). We have conducted 224 open-focused interviews in Canada and the United States. Interviewees included people with a wide range of responsibilities in the insurance industry, professionals in expert systems that serve the industry, representatives of consumer associations, members of the general public who were consumers, industry regulators, senior civil servants, and members of Parliament. Observational research included attending industry conferences and observing sales operations, loss prevention operations, and claims-related examination practices. The research was funded by the Social Sciences and Humanities Research Council of Canada, the Canada Council Killam Research Fellowship Program, and the Visiting Fellows Program of All Souls College, Oxford.

References Abbott, A. 1988. The System of Professions: An Essay on the Division of Expert Labor. Chicago: University of Chicago Press

The Moral Risks of Private Justice 361 American Medical Association. 1993. Guides to the Evaluation of Permanent Impairment. 4th ed. Chicago: American Medical Association Ashworth, A. 2001. 'The Decline of English Sentencing, and Other Stories.' In M. Tonry and R. Frase, eds., Sentencing and Sanctions in Western Countries. New York: Oxford University Press Babbage, C. 1826. Comparative View of the Various Institutions for the Assurance of Lives. London: J. Mawman Baker, T. 1996. 'On the Genealogy of Moral Hazard.' Texas Law Review 75: 237-92 - 2000. 'Insuring Morality.' Economy and Society 29: 559-77 Baker, T, and K. McElrath. 1997. 'Insurance Claims Discrimination.' In G.D. Squires, ed., Insurance Redlining: Disinvestment, Reinvestment, and the Evolving Role of Financial Institutions, 141-56. Washington, DC: Urban Institute Press Beck, U. 1992. Risk Society: Towards a New Modernity. Trans. M. Ritter. London: Sage - 1999. World Risk Society. Cambridge: Polity Best, J. 2001. Dammed Lies and Statistics: Untangling Numbers from the Media, Politicians and Activists. Berkeley: University of California Press Clarke, M. 1999. Citizens'Financial Futures: The Regulation of Retail Financial Services in Britain. Aldershot: Gower Douglas, M. 1992. Risk and Blame: Essays in Cultural Theory. London: Routledge Ericson, R. 1994. 'The Decline of Innocence.' University of British Columbia Law Review 28: 367-83 Ericson, R.V., and K.D. Haggerty. 1997. Policing the Risk Society. Toronto: University of Toronto Press - 2002. 'The Policing of Risk.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press Ericson, R., D. Barry, and A. Doyle. 2000. 'The Moral Hazards of Neoliberalism: Lessons From the Private Insurance Industry.' Economy and Society 29: 532-58 Ericson, R., and A. Doyle. Forthcoming. Uncertain Business: Insurance and the Limits of Knowledge. Toronto: University of Toronto Press Ericson, R., A. Doyle, and D. Barry. 2003. Insurance as Governance. Toronto: University of Toronto Press Finnegan, D. 1999. 'Shades of Grey.' Recovery 9(4): 6-8 Galanter, M. 1974. 'Why the "Haves" Come out Ahead: Speculations on the Limits of Legal Change.' Law and Society Review^: 95-160 Garland, D. 1997. '"Governmentality" and the Problem of Crime: Foucault, Criminology, Sociology.' Theoretical Criminology 1: 173-214 - 2001. The Culture of Control: Crime and Social Order in Contemporary Society. Chicago: University of Chicago Press

362 Richard V. Ericson and Aaron Doyle Glenn, B. 2000. 'The Shifting Rhetoric of Insurance Denial.' Law and Society Review 34: 779-808 Gordon, C. 1991. 'Government Rationality: An Introduction.' In G. Burchell, C. Gordon, and P. Miller, eds., TheFoucault Effect: Studies in Governmentality, 1fh1-51. Chicago: University of Chicago Press Gowri, A. 1997. 'The Irony of Insurance: Community and Commodity.' PhD dissertation, University of Southern California Hacking, I. 1990. The Taming of Chance. Cambridge: Cambridge University Press Haggerty, K.D., and R.V. Ericson. 2000. 'The Surveillant Assemblage.' British Journal of Sociology 51: 605-22 Harris, C. 1998. 'The Mismeasure of Fraud.' Canadian Insurance, May: 5 Hawkins, K. 1984. Environment and Enforcement: Regulation and the Social Definition of Pollution. Oxford: Oxford University Press Heimer, C. 1985. Reactive Risk and Rational Action: Managing Moral Hazard in Insurance Contracts. Berkeley: University of California Press Illich, I., I. Zola, J. McKnight, J. Caplan, and H. Shaiken. 1977. Disabling Professions. Toronto: Burns and MacEachern Insurance Corporation of British Columbia. 1996. Motor Vehicle Insurance in British Columbia at the Cross Roads: Volume 1: The Case for Change. Report prepared by KPMG, Eckler Insurance Services Inc., Eckler Partners Ltd. 19 December 1996 - 1997. Loss Prevention: Road Safety and Auto Crime Five Year Plan. Vancouver: ICBC Lianos, M., and M. Douglas. 2000. 'Dangerization and the End of Deviance: the Institutional Environment.' In D. Garland and R. Sparks, eds., Criminology and Social Theory, 103-25. Oxford: Oxford University Press Malleson, A. 2002. Whiplash and Other Useful Illnesses. Montreal and Kingston: McGill-Queen's University Press Munglani, R. 1999. 'The Roots of Chronicity.' Recovery 10(2): 14-15 Orcutt, T, andj. Turner. 1993. 'Shocking Numbers of Graphic Accounts: Quantified Images of Drug Problems in the Print Media.' Social Problems 40: 190-206 Porter, T. 1995. Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. Princeton: Princeton University Press Rose, N. 1997. 'Between Authority and Liberty: Governing Virtue in a Free Society.' Paper presented to a conference on Civil Society, Communitarianism, Third Sector, Helsinki, October - 1999. Powers of Freedom: Reframing Political Thought. Cambridge: Cambridge University Press Ross, I. 1970. Settled Out of Court. Chicago: Aldine

The Moral Risks of Private Justice 363 Sedgwick Insurance. 1997. Insurance Market Trends and Development. Simon, J. 1987. 'The Emergence of Risk Society: Insurance, Law and the State.' Socialist Review 95: 61-89 - 'Governing through Crime.' In G. Fisher and L. Friedman, eds., The Crime Conundrum: Essays on Criminal Justice. Boulder: Westview Slater, D. 1997. Consumer Culture and Modernity. Cambridge: Polity Press Sontag, S. 1979. Illness as Metaphor. London: Allen Lane - 1989. AIDS and its Metaphors. London: Allen Lane Spitzer, W.O., M.L. Skovion, L.R. Salmi, DJ. Cassidy, J. Duranceau, S. Suissa, and E. Leiss. 1995. Scientific Monograph of the Quebec Task Force on Whiplash Associated Disorders: Redefining 'Whiplash' and its Management. Spine, Suppl. 20(85): 2-73 Turow, J. 1997. Breaking Up America: Advertisers and the New Media World. Chicago: University of Chicago Press Weisberg, H., and R. Derrig. 1991. 'Fraud and Automobile Insurance: A Report on Bodily Injury Claims in Massachusetts. 'Journal of Insurance Regulation 9: 497-541 - 1992. 'Massachusetts Bodily Injury Tort Reform. 'Journal of Insurance Regulation 10: 384-440 Zelizer, E. 1979. Morals and Markets: The Development of Life Insurance in the United States. New York: Columbia University Press

This page intentionally left blank

PART FOUR Governing Risky Desires

This page intentionally left blank

Introduction

In Part 4, Jonathan Simon, Nikolas Rose, and Mariana Valverde address efforts to govern risky desires. Jonathan Simon examines the extreme sport of high altitude climbing and conundrums about who is responsible for rescue operations and their costs. His research not only furthers the analyses of insurance and moral risk provided in Part 3 but also contributes to debates on collective systems of risk sharing in liberal societies. Nikolas Rose explores the way in which recent advances in the life sciences and biomedicine are transforming the governance of risky desires, in particular those related to mental health problems as well as cravings and addictions. He shows how such problems are being reconfigured as diseases of the brain, amenable to targeted governance with drugs. Mariana Valverde also analyses efforts to develop drugs that act upon risky desires. She focuses on a drug that is intended to reduce cravings for alcohol. She joins Rose in suggesting that there is a waning of conceptions of the normal and the pathological in dealing with risky desires, and an increase in efforts to isolate a biochemical process and to respond through continuous use of a drug that targets that process. Jonathan Simon uses debates about rescue operations in Denali Park (Mount McKinley), Alaska, to address broader issues in the governance of risky desires. He interrogates when efforts to rescue people who have put themselves at risk become viewed as counter-productive. The conundrums of when rescue should be risked demonstrate the interaction of the moral risk and moral opportunity effects of efforts to help others. Moral risk refers to the assumption that people will change their investment in preventive efforts as risk is shifted away from them. This assumption has become pervasive in contemporary liberal regimes. It initially gained currency as a critique of the welfare state, but it is now a

368 Part Four: Governing Risky Desires taken-for-granted view of human behaviour. 'More help is less' because people fail to help themselves and their problems end up being compounded rather than corrected through benign intervention. In extreme sports, people not only take on risks voluntarily but actually seek great pleasure from them. Indeed, part of the pleasure is 'total' responsibility for the risks and perfecting the art of making adjustments to them. As Simon notes, it is not surprising that high altitude climbing images and narratives are used figuratively in other contexts where the thrill and rewards of risk taking are promoted, such as speculative investment in equity markets. The actual story of mountaineering and rescue is more complicated. The sport is not as individualistic and devoted to embracing risk as it first appears. Rather, government regulation, adventure tour operator packaging, and guides with local knowledge combine to make climbing Mount McKinley less risky. Moreover, the duty to rescue is a foundational norm of the mountaineering culture worldwide, affecting many of the social practices of climbing. This norm is underpinned by a sensibility of rationality over desire. For example, an instructional book declares, 'Never let judgment be overruled by desire when choosing the route or deciding whether to turn back.' Specific techniques involved in choosing a climbing team, roping, progressing up slopes, dealing with someone in a precarious position (and therefore threatening the safety of other members of the team), and so on all cultivate special bonds of trust, risk, heroism, fate, and solidarity in the climbing community. The question remains how safe is safe enough, and at whose expense. Simon chronicles how the U.S. National Park Service built a sophisticated rescue infrastructure for Denali Park. However, following a particularly expensive year in terms of deaths and rescue costs in 1992, the service was heavily criticized. The critical discourse focused on the moral risks posed by underqualified climbers. These climbers were said to rely on the prospect of having helicopters and rescue stations a cell telephone call away in case of misfortune. The service responded by introducing a screening procedure, a sixty-day waiting period, educational materials replete with lessons of moral risk and responsibility, and a climbing fee aimed at partial cost recovery. These measures were in turn criticized by the climbing community as undermining the autonomy of climbing culture, including its own means of organizing rescue, and as a de facto introduction of a compulsory rescue insurance premium. Local guides, also fearing loss of autonomy and business, chipped in with declarations such as 'the desire to confront risk and hazard is part of the

Introduction

369

human character.' Another critique emerged from conservative taxpayers' associations, who argued for even more control over who climbs, formal bonding and private insurance requirements, and additional forms of cost recovery. This constituency wanted the government to get out of the business of subsidizing an international counter-culture of people with peculiar risky desires. Simon argues that these debates not only address issues of efficient loss bearing (moral risk distribution) but also create moral investment in the cultural determination of loss relevant behaviour (moral opportunity) . That is, the debates foster reflection on the meaning of loss, the need for mutual aid, the need for standards of loss minimization and response, and the assistance capacities of existing knowledge and technology. They also have practical effects, such as expanding the skill base to address the risky activity and investing in this skill base and associated technologies. These features of moral opportunity are leading to a formalization of a duty to rescue that has always been implicit in the climbing community. Simon's analysis suggests comparisons among contexts of moral risk, moral opportunity, and governance. His basic question is: when are efforts to rescue people viewed as counter-productive? A basic answer might be: when people have voluntarily placed themselves at high risk, suffered the consequences, and helping them would place others at high risk. But a proper answer must be more specific, pertinent to the particular situation of the people and institutions involved in risk taking and management. In contemporary liberal regimes people are encouraged to speculate with their finances, and to both rejoice at the seemingly effortless windfalls and absorb precipitous descents. Nevertheless, in some circumstances there have been spectacular bailouts, for example, the one following the Savings and Loan scandal in the United States, which cost the government hundreds of billions of dollars in deposit insurance claims. The complexion is different again when lives are at risk. As Simon shows in the case of Denali Park, the National Park Service went to extremes to develop emergency services to cushion the effects of extreme risk taking. When life itself is at risk, many would agree that extraordinary efforts at high cost should be expended. But should such efforts be dedicated especially to those who experience misfortune through no choice of their own? Should resources be differentially allocated to protect the vulnerable and the innocent?

370 Part Four: Governing Risky Desires Apparently not in liberal regimes such as the United States, where 44 million people are made vulnerable through not being able to afford health insurance, even while resources are dedicated to making high altitude climbing less risky for a select few members of the international climbing community who enjoy at least the thought of putting their lives on the line. Consider also the contemporary sensibility of war, in which the typical objective is to avoid putting one's own armed forces at high risk, for example, through prolonged ground-based actions. The goal in going to war may be to 'rescue' the unfortunate citizens of an unsavoury regime, but this rescue must not be effected at the expense of death and injury to one's own citizens. The risks of life and death are differentially valued. Nikolas Rose explores the governance of risky desires that relate to mental health problems and addictions. These problems are increasingly being addressed through new developments in the life sciences and biomedicine. They are now regarded, at least in part, as diseases of the brain amenable to treatment through new drugs. In the case of prescribing naltrexone to control alcohol or heroin use, for example, the intention is to act upon desire - to remove it or at least reduce it - as opposed to previous attempts to make the satisfaction of desire unpleasant (e.g., Antabuse treatments for alcoholism) or to substitute more acceptable addictions (e.g., methadone for heroin). The treatment strategy is one of 'targeted governance' in which the focus is on a particular brain function and its chemical processes. This new style of thought in biological psychiatry is creating a 'pharmo-therapeutics of desire' from which emerges a new kind of person, the 'neuro-chemical self.' Rose details a shift in psychiatry towards 'molecularization': seeing mental disorders as discrete illnesses with a specific etiology and prognosis and therefore as amenable to a very specific drug treatment. This shift is entwined with developments in neuroscience that posit a specific, brain anomaly as the basis of each specific anomaly in mood, cognition, affect, or conduct. Broad classifications such as depression, schizophrenia, and neurosis are yielding to molecular classifications related to specific determinations of affect, will, cognition, and mood. Prozac, for example, has advanced as a smart drug that does not act on the person as a whole but is targeted on a specific anomaly believed to underlie undesirable variations in thoughts and actions. People are reinscribed as susceptible somatic individuals who vary in the degree to which they 'suffer from molecular errors that are potentially correctable.' The logic is one of perpetual adjustment of individual biology and circumstances.

Introduction

371

Rose observes that this new logic of biological psychiatry seems finally to overcome the Cartesian dualism of body and soul. The old distinction between organic and functional disorders disappears. Mind becomes what the body and brain can be observed to do. Any problem of the mind/body is now the behavioural consequence of an anomaly in the brain - its neurotransmitters, receptors, and so forth - that is subject to correction through drug technologies. The pharmacological industry is part of this logic as it participates in the invention of new drugs for intervention. Invention and intervention are part of the same process. That is, the neurochemical brain is further dissected and made knowable through the processes of developing therapeutic drug interventions to manipulate its functioning. The truth about psychiatric truth is that it is now embedded in 'biovalue': producing 'surplus value from vitality and its problems.' What was previously devalued as pathology, and seen as a 'welfare' drain on the national economy, is now seen as an engine of profit and economic growth. Using the example of how this logic applies to risky desires for alcohol, Rose observes that the heavy moral account of 'alcoholics' as not fulfilling responsibilities and as posing dangers has lost its weight. Now the person with risky desires is depicted as being subject to an error or mistake in a component of the brain's machinery and its neurochemistry. The error or mistake is correctable via a drug, combined with learning processes about neurochemical selfhood. Rose argues that this new logic marks a transition from normalization to correction. The normalization of the deviant is based on the statistical depiction of the average, the social articulation of the functional, the moral assertion of the good, and the medical determination of the healthy. The new logic of the molecular gaze focuses instead on variation without a norm, anomaly without abnormality. Rose believes that this unravelling of the logic of the norm is accentuated by broader institutional forces - from insurers, medicine, non-governmental organizations, charities, pressure groups - and by subjects and patients themselves, all of whom seek efficiency, equity, and economy in the governance of risky desires. Rose also detects a shift from risk to susceptibility in this new molecular logic. Using the example of the human genome, he begins with the observation that 'the normal is rare' and 'variation is the norm.' In contemporary biological thought inherited predisposition to a pathology no longer holds; instead, everyone is viewed as bearing genomic vulnerabilities to difficult conditions. These vulnerabilities do not mark defects of the person, only the small, discrete, molecular anomaly. The

372 Part Four: Governing Risky Desires molecular anomaly could be called a risk, but it is being alternatively rendered in the language of 'susceptibility.' This language is not probabilistic. Rather, it is aimed at each individual's specific susceptibilities based on DNA sequencing and the patterning of bases in locations on a specific chromosome. In effect, people have, for example, alcohol addiction susceptibility genes in need of correction rather than genes that determine a risk of alcoholism. 'Susceptibility' extends the reach of disease, and that of its professional agents in medicine and the pharmaceutical industry. Because everyone has susceptibilities, the professional agents are empowered to engage people who never previously thought of themselves as ill. Biology is no longer destiny, only something to be continuously corrected. People come to recognize themselves not as ill, but as biogenetic persons who must be ever-vigilant and responsible for preventing that which promotes susceptibilities. This involves the ongoing work of self-correction: drug taking, adjusting activities, purchasing insurance coverage, and so on. The biochemical self must become a protoprofessional, using experts to help her achieve genetic responsibility. Far from being a oneoff process of normalization, the effort must be a continuous one of making responsible choices to enhance self-capacities. This enhancement rests on a three-pronged co-production: a drug, a condition it will target, and a will to target. This co-production forms the core of the new bioeconomics — the politics and economics of life itself— which bears its own risky desires for profit through continuous control. Mariana Valverde also focuses on new drugs aimed at risky desires. She begins by depicting them as 'smart drugs' in two senses. First, they are knowledge-based products the main production costs of which are loaded in research and development rather than in the material product itself. Second, they are supposed to act like smart weapons, hitting 'hot spots' of the body while minimizing collateral damage in the form of side effects and unintended consequences. In this regard, they contribute to the broader model of targeted governance characteristic of liberal regimes. This is a model of decentralized knowledge and power networks deployed to identify threats and splinter them with precision. The military no longer engage in campaigns, only precision bombing. The police no longer claim responsibility for social order, only the targeting of crime hot spots. The social services no longer engineer society through universal welfare programs, but only target specific problem people. The clinical professions no longer rehabilitate the whole person from a condition of pathology to a state of normality, but only isolate a specific bodily process that can be targeted with drugs.

Introduction

373

Valverde advances her analysis through the substantive example of using naltrexone to treat alcohol addiction/dependence. Naltrexone helps people to drink less by reducing their craving for alcohol. It does so by targeting a particular receptor in the brain. The distinct identity of being an alcoholic - required in some treatment programs such as those conducted by Alcoholics Anonymous - gives way to seeing alcoholism as an imbalance of brain chemicals. The dream is governance at the drop of a pill. The historical record indicates that alcoholism has never been fully medicalized. Efforts to address it have always blended the moral with the psychological and physiological. The excessive drinker is depicted as being in disrepair and as having lost control of the self. However, under the influence of genetics and neuroscience, the focus is shifting to a biochemical site. This new focus, combined with the redefinition of alcohol as a drug, gives credence to the view that alcoholism is a drug addiction that can itself be targeted with smart drugs. The clinical use of naltrexone, especially in the United States, is typically combined with cognitive and behavioural therapy. There is a reluctance to rely on naltrexone alone, not because of scientific considerations but rather because of the lingering disease model which associates alcoholism with the moral and medical. Alcoholism seems to be more resistant to de-pathologization than other deviant identities. One reason is the circular reasoning involved in designating people as alcoholics: they are by definition people unable to evaluate and manage their risks, and therefore seemingly driven by some underlying pathology or disease of the will. Thus the traditional approach to treatment is to strengthen willpower to the point of avoiding the first tempting drink. Naltrexone, and the cognitive and behavioural therapies used with it, in contrast, aim to help the person drink socially and not crave more. Indeed, if naltrexone actually works, the traditional field of alcoholism research and treatment will become extinct. In cognitive behavioural therapy, the language of will is replaced by the language of coping, situations, decisions, and contract. Nevertheless, since the goal is to help people to say no to bad things, the regulation of willingness is still evident. Valverde cites evidence that naltrexone also diminishes desires for other pleasures, such as sex, sweets, and exercise. While the user of naltrexone may see these diminishments as collateral damage, at least some developers of the drug see them as a good thing worthy of further promotion. One promoter dreams that the drug might 'extinguish gambling, sexual obsession [and] compulsive thrill-seeking criminal activity.'

374 Part Four: Governing Risky Desires Traditional treatment programs that address willpower alone are able to audit their successes in terms of abstinence. Cognitive behavioural therapy using naltrexone is difficult to audit because managed drinking is the goal. Where smoking-cessation researchers have developed a minicomputing device to measure craving - it beeps at regular intervals to remind the user to input measures of mood and desire - similar devices have not entered the field of alcohol and drug treatment. A more fundamental problem is that it remains difficult to operationalize alcoholism in terms of the brain. No one knows which brain processes actually control 'craving.' Drinking is affected by broader biochemical and environmental processes, rather than a location on a single biochemical site. In spite of the accelerating popularity of the biochemistry and neurology of desire - exemplified in the popping of Prozac or Viagara - 'the alcoholic brain remains a moving target' (454). In contemporary liberal regimes people are told that if they look after their specific, little problems through local knowledge of risk, the general big picture will take care of itself. Of course, governance through local knowledge of risk requires incessant audits, monitoring, prescriptions, and performance assessments. Valverde cites an expert who recommends that alcoholics, and perhaps other pleasure seekers, take naltrexone for life, much like the diabetic takes insulin. This recommendation indicates that targeted governance is not intended to hit the bull's-eye and declare victory, but rather to set in motion a process of perpetual assessment of and adjustment to risky desires. The result is more governance, not less. While this governance may seem to make people more willing, it does not make them more free.

13

Risking Rescue: High Altitude Rescue as Moral Risk and Moral Opportunity JONATHAN SIMON Behind many of our public policy discussions in recent years has lurked a compellingly uncharitable thought, the idea that efforts to rescue people are, in the end, often counter-productive. This thought is of course rarely unqualified and seems to have its greatest force when applied to systems of public rescue run on a bureaucratic basis. The principle often referred to as 'moral hazard' or 'moral risk' holds that, all other things being equal, people will reduce their investment in precaution as risk is lifted from them.1 Give people income support during prolonged periods of unemployment and they will invest less in staying employed, and less in accepting any kind of employment (Murray 1984). Give people full compensation when their homes burn down or flood, and they will accumulate more goods and take fewer precautions to protect them. In the name of taming moral risk, a growing political consensus in the United States and other societies argues for retrenching not simply public welfare and social insurance but also tort liability and private insurance coverage. As Tom Baker (1996) has shown, the concept of moral risk had its origins in the prudential concerns of insurance underwriters in the nineteenth and early twentieth centuries, who sought to avoid taking on 'moral hazards' in reference to specific persons and losses. In the second half of the twentieth century, however, this pragmatic concern with problematic individuals and losses morphed into a powerful generalized critique of welfarism in both social insurance policies of the state - for example, workers' compensation (McCluskey 1998) - and the widespread aggregating of risks through private insurance. The conclusion that more help is less has become widespread among not only economists and policy makers but a broad portion of middle-class voters in

376 Jonathan Simon most liberal societies. As a result social policy, which once favoured aggregating risk through insurance (Baker and Simon 2002), has turned against insurance and now favours greater individual risk bearing. I have argued elsewhere that the rising popularity of extreme sports, including high altitude climbing, is emblematic of the larger shift in the public culture of risk associated with the moral hazard critique. Extreme sports make individuals as completely responsible for the risks of failure as any one can be (since they may easily pay with their lives). In those extreme sports where the individual can actually exercise considerable control over the risk through care taking - for example, mountaineering, as opposed to being shot out of a cannon - this strategy is often cited by participants as leading to important gains in performance and experience, which Stephen Lyng has called 'edgework' (Lyng 1990). Climbers embody a logic seemingly different than the one embodied in insurance and moral risk, one where people take on risk voluntarily and make an art of adjustments entailed by the assumption of total responsibility. It is little wonder that mountaineering images and narratives are frequently borrowed and employed as icons for those who would advocate cutting 'safety nets' in areas like health care, retirement provision, and survival for the poor. In different ways the discourse of mountaineering and other forms of edgework may offer important models for individuals experiencing the retransfer of risk. Consider in this light the status that mountaineering has historically given to the ethical duty to rescue, the very activity moral risk analysis urges us to constrain. While liberal societies generally have avoided a legal duty to rescue, mountaineering in many historical instances has embraced a fundamental duty to rescue those who can be saved at reasonable risk. When we look at practices such as mountaineering, risk transfer and risk embrace are not alternative strategies but intertwined ones. Some climbers have always chosen to climb alone ('soloing', in contemporary mountaineering parlance) in an explicit defiance of any likelihood of rescue (miracles, of course, are always possible). But such soloists are in fact performing against a powerful norm of rescue that has been valorized in various ways by mountaineers since the late eighteenth century (Fleming 2000; Ring 2000). Climbers in this central norm are expected to engage in rescue of self, of climbing mates, and of other parties encountered. Indeed, many of the dominant elements of mountaineering as a social practice are organized in large part around the provision of rescue, including the practice of guiding, the size and composition of climbing

High Altitude Rescue as Moral Risk and Moral Opportunity

377

teams, and the method of belaying a climber with a rope. Climbers in North America have been formally organized for political and recreational purposes primarily in mountain rescue groups (Rjeldsen 1998). In virtually all post-industrial societies, the organization of rescue has become deeply troubling for government. Those parties and politicians that have generally embraced the strategies of neo-liberalism find themselves politically exposed on the problem of rescuing those who seem socially deserving, like senior citizens who have worked most of their adult life but will be poor in old age if generous taxpayer-financed retirement income is cut dramatically. Those parties and politicians that remain committed to the mid-twentieth-century model of rights-based rescue policies (which now includes no major party in the United States, although adherents may be found in Canada, Europe, and other parts of the globe), that is, to an entitlement to social provision against at least the ordinary risks of industrial society, find themselves exposed to the resentment of taxpayers and the possibility that they may only shift risks in unpredictable directions. In this chapter I want to take a closer look at a small frontline in the debate about rescue where the broad question of what forms of rescue ought to be provided to ordinary risk of whole populations intersects with the extraordinary risks of high altitude mountain climbing and the duty to rescue as a norm within mountaineering. A unique forum in which many of these issues are playing out today is Denali National Park and Preserve in the State of Alaska. Denali National Park is the location of a mountain known to native Alaskans (and now more commonly among climbers) as Denali, and officially as Mount McKinley, which at just over 20,000 feet is the highest peak in North America. Denali has always been dangerous to climb: its high altitude and polar location make any problem potentially fatal. Since the 1960s it has also become one of the world's most popular mountaineering objectives. The existence of a relatively easy route, in good weather, that requires no technical rock climbing, and the availability of a range of private services to fly climbers to the mountain and guide them to the top, means that one of the highest mountains in the world is also one of the most accessible. Once on the mountain, however, bad luck and bad weather can lead to one of the greatest survival challenges imaginable. This combination of access and risk means Denali has one of the highest mountaineering death rates in the world. Since mountaineering became popular in the 1970s, there have been three or four fatalities a year out of an average of 1,000 to 1,200 climbers.

378 Jonathan Simon As a national park inviting visitors, Denali's dangers pose a unique challenge of governance. This challenge has only been exacerbated by the growth of a global market in adventure tourism. The mountain draws a wide range of climbers from all over the world. Climbing parties include all types, from guided groups moving up the relatively easy 'West Buttress' route to small, tightly knit climbing parties with years of common experience to groups that form out of the mix of international tourists who gather at the base camp village on the Kahiltna glacier. Governing Denali is a task delegated by the United States Congress to the jurisdiction of the National Park Service (NFS). Whether it chooses to regulate heavily or dissociate itself from mountaineering activity, the NFS has ultimate legal responsibility for who is allowed to climb, who gets rescued, and what costs are collected for both activities. As a federal agency, and by tradition one of the most law-centred, the NFS finds itself exposed to considerable criticism from both Congress and interested members of the public. In recent years this criticism has focused on the NPS's extensive use of specialized helicopters and crews to perform rescues at a higher altitude than is attempted anywhere else in the world. This practice began in the 1960s and has included the military, professional pilots, and expert climbers as well as NFS rangers. In the early days these episodes were erratic and unpredictable and were largely limited to the lower portions of the mountain. In the 1980s, with dramatic growth in the international adventure tourism market, the NFS organized a full-time rescue service utilizing a privately leased high altitude helicopter and crew on full-time alert capable of reaching the upper portions of the mountain. The NFS also runs a rescue station at the base camp from which most climbers depart, at 7,000 feet on the Kahiltna glacier, and a special high altitude base and medical facility at 14,200 feet, about a third of the way down from the summit. These operations were estimated in 1998 to cost about $1.6 million per mountaineering season (Energy and Natural Resources 1998:9). Denali's rescue operations came in for public scrutiny after a disastrous mountaineering season in 1992. Extreme weather led to a wave of rescue calls, not all of which were successful. Twenty-two major rescue efforts were mounted, notwithstanding which eleven climbers died. As with the deaths on Mt Everest four years later (Krakauer 1998), there was an outpouring of media attention on and criticism of the extreme risks being taken by allegedly under-qualified individuals, all seemingly coun-

High Altitude Rescue as Moral Risk and Moral Opportunity 379 tenanced and abetted by the NFS. In 1994 the NFS introduced new policies aimed at reducing both rescues and fatalities to be implemented in the 1995 mountaineering season (April through June): a minimum sixty-day pre-registration period was established to enable the NFS staff to 'provide information to prospective mountaineers in advance of their climb' and a $150 fee was imposed.2 These rules became permanent in 1996.3 In addition, the NFS staff at Denali have developed educational materials to be shown to potential climbers registering in Alaska and a forty-page booklet that they will mail out in five languages and which can be downloaded from the web (NFS 2001). This education program draws heavily on the work and values of climbers and is intended to make potential climbers more aware of the risks of the mountain, and thus potentially more responsible. The situation remains unstable. Many in the organized mountaineering community resent the paternalism of the Denali mountaineering program, which they believe is undermining the autonomy of mountaineering culture and actually encouraging more rescue situations by reducing incentives for self-help. Worse still, climbers fear a government mandate to pick and choose who gets to climb and which routes they are allowed to take, something that works against not only beginners but also experienced climbers attempting isolated and difficult routes. The fee has similarly garnered considerable resentment from climbers, who consider it unfair to be singled out for a user fee. Critics maintain that it is compulsory rescue insurance, an argument bolstered by the fact that the total revenue raised is roughly the equivalent of the contract price for the full-time high altitude helicopter and crew. The NFS insists that the fee is not a rescue insurance premium and pays only for additional staff to manage the mountaineering program at Denali, that it is used primarily to hire rangers with mountaineering specialties. A different stream of criticism has emerged from conservative 'taxpayers' associations on the prowl for examples of wasteful government spending. From this perspective, the NFS has not done enough to discipline climbers, either by limiting access on the front end or collecting rescue costs via a formal bonding or insurance requirement. The expensive and high-risk effort to provide high altitude rescue on Denali has been painted as a kind of misguided welfarist program for an international counter-culture of climbers. The high technology response to emergencies on the mountain is seen as constituting a subsidy from the federal government for engaging in peculiar and dangerous behaviour.4

380 Jonathan Simon One of the most powerful critics is Alaska's Republican Senator Frank Murkowski. In 1998 Senator Murkowski chaired hearings of the U.S. Senate Committee on Energy and Natural Resources, Mount McKinley Rescue Activities (Energy and Natural Resources 1998). Testimony was taken from NFS staff, military personnel, and a number of the private guiding concession holders. Senator Murkowski was highly critical of the NFS approach. In November 2000, Congress enacted a Murkowski proposal for a study of whether NFS at Denali could do more to recover the costs of high altitude rescue and report back by fall 2001. The senator has broadly hinted that he favours some kind of mandatory insurance premium to climb Denali.5 In the remainder of this chapter I want to use the problem of high altitude rescue on Denali as a window into the growing ambivalence about social institutions of rescue in the United States and other liberal societies. Part I briefly describes the tradition of rescue in mountaineering as a counter-tradition to the dominant jurisprudential norm against duties to rescue among strangers. Part II provides a more detailed description of the rescue situation on Denali. Part III examines a version of the moral risk critique that has been mounted against NFS rescue efforts on Denali. Part IV offers a defence of the emerging NFS mountaineering program as an example of how rescue, like insurance, creates a 'moral opportunity' (Stone 2002). As Tom Baker has argued (1996), most moral risk analysis is truncated, revealing only one tail of the possible distribution of incentives, that is how insurance creates moral risks for potential loss victims (usually workers, consumers, and so forth), while ignoring how limits on coverage or liability create moral risks for insurers to under-invest in safety and prevention. More recently Deborah Stone (2002) has argued that the tendency of insurance to generate yet more demand for insurance is less the unravelling of self-reliance depicted by moral risk economists than a process of change in the 'cultural understanding' which transforms norms, expectations, and values as well as behaviour. This is not just a question of identifying the economically efficient loss bearer, but also of governance and political culture. Insurance creates flows of power and knowledge around a risky activity that may increase overall social commitment to reducing losses and make the remaining ones more economically manageable. These effects may mitigate and ultimately exceed the costs associated with moral risk, pointing to policies that expand rather than contract insurance relationships.

High Altitude Rescue as Moral Risk and Moral Opportunity 381 The Duty to Rescue in the Mountains Liberal societies are reluctant to impose any onerous duties of rescue on ordinary citizens, and common law ones like the United States generally refuse to require even non-burdensome efforts at rescue (Murphy 2001; Heyman 1994). Such affirmative obligations are legally enforced in situations where contract (innkeeper, ship's captain) or status (child, spouse) create distinct dependencies. The duty to rescue in mountaineering intersects at times with this jurisprudence but extends far beyond where the law would impose it.6 Independently of any specific legal obligation, the duty to rescue other climbers in trouble is deeply embedded in the practice and values of mountaineering in almost all its forms.7 Since law only intermittently governs mountaineering, the duty to rescue is not always a legally cognizable duty (although it often is, depending on the relationship of the parties). The novice climber is likely to encounter the duty to rescue in the many forms that mountaineering stories take (magazines, books, oral tradition) rescue narratives being among the most common and most valued forms of such stories. The more experienced climber may well be personally familiar with the reputational sanctions within the mountaineering community produced around rescue issues. Dishonour in the anthropological sense can befall a climber for needing rescue herself, for failing to rescue someone else, or for the way she behaved in a rescue situation. At the same time, honour of the highest kind is accorded to those who against all odds rescue themselves and/or others. Indeed, rescue seems to open the pathway into the deepest adventures in mountaineering. While risking dishonour if they fail, rescuers are also given moral exemptions from many of the norms that constrain performance in normal mountaineering. No single object in the mountaineering universe better evokes this potential honour/dishonour or reveals the underlying logic of rescue in mountaineering than the rope. Since the beginning of mountaineering the rope has been the piece of equipment most associated with the sport. Although sometimes used as a direct aid to climbing, its dominant use from the earliest days has been to tie climbers together so that if one should lose her footing, the other or others can attempt to prevent a fall. The development of the iron piton towards the end of the nineteenth century allowed climbers to connect their ropes directly to the mountain. Until then the only thing the rope was connected to was another climber. Since the rope could as easily pull a safe climber off the moun-

382 Jonathan Simon tain as secure a falling one, the decision to rope together was always fraught with great risk (Ring 2000: 218). In the golden age of European mountaineering in the Alps, when modern alpinism was invented and the iconic image of the rope first established in literature, 'roping up' was an act of complete and supreme trust in other people. Despite the risk, roping became standard good practice by the time the Matterhorn was climbed in 1865.8 Modern 'belaying' technology still requires climbers to place enormous trust in each other, but at least it establishes an objective set of techniques which, if properly used, can make rescue easy. Even today, when rope technique relies much more on gear and on anchors to the mountain itself, the rope remains a profound symbol of both solidarity and good form.9 The importance of the duty to rescue is marked by its centrality to the very roles into which mountaineering practice has been divided and defined, from guide and client to expedition team-mates. Where mountaineering has been organized around commercial relationships between guides and climber/tourists, the duty to rescue has helped define the very core of the guide role. In the mid-nineteenth century, when the Alps became the centre of the first great awakening of interest in mountaineering as a sport, almost all mountaineering was done with the assistance of local guides responsible for finding a path, carrying the climbers' belongings (bigger expeditions involved separate porters), and, most importantly, saving a climber if he should slip.10 The English climbers who dominated the early years were mostly Victorian gentlemen interested in the sublimity of ascending perilous alpine cliffs, not hauling food and tents. Most importantly, there were no good maps of the mountains. The only way for an English climber to acquire any ready knowledge of a mountain in the Alps was to take along a knowledgeable local. In the most famous climbing town of the nineteenth century, Chamonix in the French-dominated province of Savoy, the guides were sufficiently organized to control, by and large, the conditions under which the far wealthier English tourists accessed the mountains. From the end of the Napoleonic wars the guides of Chamonix formed their own corporation and made rules for the conduct of tourism, including setting fees and standards for how many guides would be required for different kinds of tours (Clark 1949: 72; Ring 2000:41). At the very least guides were 'path pointers'; at the most they were effective governors of mountaineering in the Alps. But the most valorized aspect of the guide role throughout the nineteenth century and

High Altitude Rescue as Moral Risk and Moral Opportunity 383 beyond was to keep the client safe and rescue him from danger.11 When weather or mishap rendered the guide incapable of rescuing the client, that duty moved on to the collectivity of local guides organized as a rescue squad, which would begin to ascend when it became clear that a party was overdue and not visible on the mountain. The occasional guideless party might draw a rescue response, but one marked by resentment: climbers who had failed to support the local guides now relied upon their tradition of rescue. While both guides and climbers in the pursuit of alpine peaks recognized that guides could not warrantee against fatal falls, guides were expected to secure the climber if followed carefully and to die trying to bring the climber home if something went wrong. Guides who returned from the route without their clients could expect a judicial inquiry (Ring 2000: 184). Charles Goss (1948: 2-3) celebrates both aspects of the guides' role in the preface to his collection of famous nineteenth-century mountaineering incidents:12 'In bringing this preamble to a close, I would like to underline two aspects of these mountain tragedies: the work of the rescue parties, and the affection of the climber for his guide; features which bring light to many a picture otherwise so dark. The heroic work of the rescue parties is appreciated too little. These frequently obscure men who, completely scornful of the danger, obey only their consciences in attempting to rescue parties in distress, unknown to themselves display incomparable spiritual qualities. How many of them have died, victims of their inflexible sense of duty!' Nothing epitomized the great guide more than dying in a successful effort to save his clients. Such a death came to one of the most famous Italian guides of the nineteenth century, Jean-Antoine Carrel, one of the pioneer climbers on the Matterhorn. Carrel died on some rocks near the base of the Matterhorn, decades after narrowly missing out on being the first climber to summit the mountain, having successfully led a party back from an abandoned summit attempt during a severe storm. After ensuring that his clients had an easy roped climb to the valley floor, Carrel became insensible from exhaustion and died upon the rock on which his colleagues had attempted to revive him. Edward Whymper, the most famous English climber of the nineteenth century, and the first European climber to establish his fame not simply through first ascents of storied peaks but by writing about them afterwards, both employed Carrel and competed with him to summit the Matterhorn in the 1860s. Whymper wrote a moving eulogy to his friend that highlighted the duty of rescue: 'The manner of his death strikes a chord in hearts he never

384 Jonathan Simon knew. He recognized to the fullest extent the duties of his position, and in the closing act of his life set a brilliant example of fidelity and devotion. For it cannot be doubted that, enfeebled as he was, he could have saved himself had he given his attention to self preservation. He took a nobler course; and, accepting his responsibility, devoted his whole soul to the welfare of his comrades until, utterly exhausted, he fell staggering on the snow' (Goss 1948: 225). Over a century later, one of the most famous English-language writer/ climbers of our day, John Krakauer, offered a strikingly similar eulogy to Rob Hall, a guide who died high on the summit of Mount Everest during the violent storm in 1996.13 The guiding done by Hall and his competitors differed dramatically from guiding in the Alps during the nineteenth century. Hall was from New Zealand, but he operated in an array of high altitude mountains, mainly in Asia, and competed with guides from around the world to serve an elite global market for high stakes adventure tourism. Very real changes have occurred in the nature of the guiding enterprise (some of which Krakauer (1998) bemoans in his book), but he valorizes just those aspects of Rob Hall's final performance on Everest that would have won the praise of guides in the Chamonix corporation. 'Rob was bringing up the rear - what we call, you know 'pulling the sweep' - sort of looking after the slowest, weakest members. And this is what did him in. He was with a Japanese woman named Yasuko Namba and a Seattle man named Doug Hanson who ... kind of collapsed. And Rob stayed with them, as a good guide should, and that cost him his life, I have no doubt.'14 The power to rescue was, in fact, the key to the guides' power. Any nineteenth-century tourist of reasonable physical fitness might successfully 'scramble' up the visible buttresses of the Matterhorn or the Eiger on a pretty day, but who would get that tourist down (inevitably much harder), especially if a sudden change of weather made it impossible to navigate by landmarks off the mountain? Sometimes the weather would prove too much for even an experienced guide, and rescue would depend on a ground-based team of available guides. Even when they could not save a climber's life, the ground-based rescue team would endeavour mightily to bring back his body, something Victorians were more obsessed with than people of comparable background are today. It was this power to rescue that allowed people of little more than peasant class in one of the poorer parts of Europe to gain a curious equity with well-born English gentlemen. In perhaps the most famous boast in nineteenth-century alpine history, the guide Peter Bohrens is supposed

High Altitude Rescue as Moral Risk and Moral Opportunity

385

to have ordered a meandering client back into the rope saying: 'Herr, you are master in the valley: I am master here' (Clark 1949: 21). In those mountaineering cultures dominated by autonomous selforganized mountaineering groups, such as found in the United States, rescue is no less important in defining important features of the mountaineer role. As with the guide system there are two faces of rescue. One focuses on the immediate group of climbers, the team, including the person or persons in need of rescue, who must be part of any effort. The other focuses on remote rescue efforts that are typically launched only after an accident has become apparent, as when a climbing party comes upon another party in distress, or when the failure of the climbing party to return - or today quite commonly by a cell phone or radio message alerts others off site. While guide systems generally make their duty to rescue readily apparent (it being part of the marketing), that duty is more implicit among autonomous mountaineering cultures. For example, the most widely used mountaineering handbook in the United States, The Freedom of the Hills, includes a 'climbing code.' While no duty to rescue is expressly provided, the logic of rescue pervades most of the elements included. A Climbing Code • A climbing party of three is the minimum, unless adequate prearranged support is available. On glaciers, a minimum of two rope teams is recommended. • Rope up on all exposed places and for all glacier travel. Anchor all belays. • Keep the party together, and obey the leader or majority rule. • Never climb beyond your ability and knowledge. • Never let judgment be overruled by desire when choosing the route or deciding whether to turn back. • Carry the necessary clothing, food, and equipment at all times. • Leave the trip itinerary with a responsible person. • Follow the precepts of sound mountaineering as set forth in textbooks of recognized merit. • Behave at all times in a manner that reflects favorably upon mountaineering, with minimum impact to the environment. (17)

The first three provisions build rescue into the very structure of the climbing group. The minimum recommended number of three is to allow one person to go for additional help while another tends to an injured party. Roping from an anchored belay, the fundamental method

386 Jonathan Simon of protection during modern mountaineering, is all about mutual rescue. When the lead climber falls the belaying partner can break the fall, relying on the anchor to hold them both. The lead climber's fall is limited to twice the distance above the last piece of protection through which the belay rope passed. Maintaining party solidarity for all of the above reasons and more is crucial to survival. Jonathan Waterman's (1991: 101) analysis of mountaineering accidents at Denali found that when parties split up, the risks of needing rescue by others grows exponentially. Yet a duty beyond team-mates is implied by the advice to leave the trip itinerary with a reliable person who can alert rescuers to the need for an outside effort. Mountain rescue has also become a central basis for self-organizing among North American mountaineers. Beginning in Seattle in 1948, local mountaineering clubs were galvanized by the advances in rescue technique and climbing achieved by the Border Patrol on the German Austrian border (Kjeldsen 1998: 103). These voluntary organizations developed cooperative relationships with law enforcement in mountain areas and developed their own network of rescue climbers ready to be sent into the mountains to aid not only climbers but lost hikers and downed airplanes. The clubs, following the European tradition, made a point of carefully documenting and reporting on their rescue activities, and thus on the range of mountaineering risks in the area. They also functioned as the most important vehicle for generally organizing climbers in an area for political and other purposes. Life and Death on Denali The summit of Denali is more than 20,000 feet above sea level in a polar zone. The weather is generally worse than on the summit of Mount Everest, although the latter peak stands nearly 10,000 feet higher in a temperate zone. Basic background temperatures of 20° or 30° below zero Fahrenheit are common during the peak spring and summer climbing months. Exceedingly cold temperatures mean that frostbite is a standard experience and failure to prepare adequate shelter for the night can lead to severe tissue damage and sometimes death. Windstorms near the summit regularly reach hurricane force levels and numerous 'falls' from the top of the mountain probably resulted from climbers quite literally being blown off. Far more common is the routine but still potentially fatal experience of having one's tent, sleeping bag, or gloves blown off the mountain. As one of the NFS mountaineering

High Altitude Rescue as Moral Risk and Moral Opportunity 387 managers at Denali tried to explain when questioned about fatalities on the mountain during the 1998 hearing of the Senate Committee on Energy and Natural Resources: 'Education helps, sir, but predominantly what happens on Denali, like other big mountains, is the weather ...' (1998: 7). In the early days Denali's lethal conditions posed little challenge for government because few people had the resources or stamina to undertake the weeks of hiking required to get to the base of the mountain, and those who did had few illusions that help would be available if an accident were to occur during their summit attempt. The first ascent of the mountain took place in 1913. The sense of isolation and necessary independence had not changed much four years later, when Denali National Park and Preserve opened as the nation's first federal park. While mountaineering was one of the original appeals of recreation at Denali, it accounted for little of the generally modest tourism traffic until after the Second World War, when Alaska tourism picked up. In 1951 Bradford Washburn pioneered a route up the West Buttress of Denali that allowed a climber to ascend a series of moderate to steep snow fields in stages to the top, with places to camp. Equipped with crampons and an ice axe virtually any relatively fit adult with adequate guidance can now reach the summit in good weather. With the West Buttress route available, Denali was ready for the growth of popular interest in mountaineering in the 1960s. By the late 1970s more than 500 hundred climbers a year were attempting the summit. The number of climbers rose again in the 1980s, from 695 in 1984 to 1,277 in 1994, a level that has been maintained ever since (Energy and Natural Resources 1998: 5). This population is surprisingly diverse along a number of dimensions. Approximately half the climbers ascend in groups organized by the commercial guiding services that operate on Denali, the other half consist of autonomous climbing groups. Most are U.S. citizens but one-third are not and come from a wide variety of countries. How People Die on Denali In his survey of Denali climbing accidents from the first ascent through the 1980s, Surviving Denali, Jonathan Waterman identifies six categories of serious risk on the mountain: 'High Altitude Pulmonary Edema; High Altitude Cerebral Edema; Frostbite; Climbing Falls; Crevasse Falls; Avalanches.' High altitude pulmonary edema (HAPE) leads to disabling

388 Jonathan Simon and eventually fatal filling of the lungs with fluids in a pneumonia-like condition caused by the inability of the heart and lungs to cope with the altitude change. High altitude cerebral edema (HACE) is an analogous fluid infusion of the brain tissue leading to eventually fatal cognitive dysfunction. The incidence of both HAPE and HACE can be reduced through adequate adjustment to the altitude change, but both conditions can attack even well-adjusted climbers under exceptional strain or suffering from a triggering illness. Some drugs, mainly diuretics, can help, as can the immediate application of oxygen, but in both cases survival depends on descending immediately to a lower altitude. Since the sufferer may be unable to walk alone or at all, this requires other climbers to bring the stricken climber down or a helicopter evacuation. Because Denali is marketed to adventure travellers as non-technical, many people do not consider falling a serious risk. Indeed, the typical climbing fall at Denali involves a fatigued guided climber on the West Buttress route who missteps on one of the steep snowfields. If roped to others and to able guides such a fall is a non-event. Unroped falls, however, can quickly turn into a high-speed slide of 400 feet or more, resulting in broken limbs, ribs, or worse. These injuries are not usually fatal, but only because evacuation is so routinized. Were such climbers dependent on personal evacuation by a few friends from the top of the mountain to the landing fields at 7,000 feet on the Kahiltna glacier, some fatalities would be inevitable over time. Falls on rock typically occur on the more remote and challenging routes. If properly belayed by a climbing partner they are unlikely to lead to injury. Solo climbers who suffer such falls are very likely to die unless rescued. Crevasses are hidden faults or gaps within the structure of a glacier. They can often extend hundreds of feet down. During snow season accumulation makes these faults invisible until an unfortunate climber or hiker walks over a weak point and falls into a dark abyss. Almost any glacial terrain will have some crevasses and some areas are heavily shot through with them. The standard means of self-protection is for a team of two or preferably three to walk at least one hundred feet apart with rope taut between them. If one of the party should suddenly fall through a snow bridge the others have time to use their ice axes to break their partner's fall and prevent themselves, along with several hundred pounds of attached gear in the worst-case scenario, from being dragged in. Once stabilized, the able climbers can place anchored protection like an ice screw and use the rope to bring the fallen climber back to the surface,

High Altitude Rescue as Moral Risk and Moral Opportunity 389 after perhaps doing a careful roped descent to make sure the party is stable and in good position to be hauled back up. Where injuries from the fall itself are severe, even a successful extraction must be followed by an evacuation of the injured climber. For most climbers who have an accident on the West Buttress route resulting in injury (or who fall prey to HACE or HAPE) the most important rescue issue involves help getting down the mountain to where evacuation by normal small plane is relatively easy. Because of Denali's great height, removing from the upper reaches of the mountain a climber who has become disabled can present a real challenge to a party of fit climbers even on the relatively easy West Buttress. If the accident takes place on the more remote and technical routes to the summit, evacuation may be more complex or impossible for even the best and most committed climbing team. Rescue from traditional groundbased teams is rarely possible. Such rescue can take days and place fastclimbing rescuers at risk of altitude illness. For all these reasons rescue on Denali traditionally relied heavily on self-initiative and team-work. It also meant that serious injury could always turn fatal if the weather became severe or the climbing group fell apart, isolating the weakened climber.15 Against these often desperate limits, high altitude helicopter rescue, organized by the government and drawing on a group of extremely adventuresome rescuers and pilots, has steadily pushed up the height at which evacuations can now be attempted. National Park Service Response From the beginning of the National Park Service, protection of visitors has been a primary mission of the agency. At Denali, in particular, climbing has always been a major attraction and rescuing climbers a mission of the NFS (Energy and Natural Resources 1998: 4). As interest in climbing Denali grew in the 1960s, the NFS became apprehensive about how to manage this potentially dangerous practice. Initially, the service attempted to control who could climb via an appraisal of the climbing team presenting itself to the rangers for permission (Waterman 1991: 6-7). The service also endeavoured to save all those it could by using local pilots, guides, and in some cases helicopters from nearby military bases. In the 1970s, the NFS backed off from direct regulation of who could climb and focused instead on improving emergency response.16 By the late 1980s the NFS had committed itself to providing an aggressive rescue effort aimed at using technology to reduce the risk to climb-

390 Jonathan Simon ers who got into trouble on the upper mountain. The NFS leased the use of a high altitude helicopter and accompanying crew available on a fulltime basis during the mountaineering season. It also set up a heated medical tent at 14,200. Originally a research station, this facility quickly became a de facto clinic and rescue shelter for climbers in distress. The NFS, however, made no efforts to regulate mountaineering other than dispensing informal advice from the ranger station in Talkeetna to climbers paying the same park entrance fees as hikers and other users. Instead, treating the need for rescue as fixed, the service committed itself to an expensive effort to save all who could be saved. As described above, this situation came under scrutiny after a lethal mountaineering season in 1992, when a streak of bad weather, even more severe than usual, resulted in twenty-two life-saving rescues and a sobering eleven fatalities. Extensive media coverage of the fatalities generated much of the same kind of discourse that later emerged around the 1996 storm on Mount Everest (Chessler 1999: 171). Why were so many people dying in pursuit of a novelty adventure? Was somebody luring vulnerable and unprepared people onto the mountain? Who should pay for hundreds of thousands of dollars in annual rescue costs? Was the effort to rescue so many encouraging greater risk taking on the mountain? Three years later, in 1995, the NFS attempted to ameliorate the moral risk of its own rescue operation by imposing the sixty-day waiting period and $150 'climbing management fee' (a policy that became permanent in 1996). These were the first efforts in decades to control climbers, and the logic underlying them had changed. Instead of deciding who could climb or not, the NFS placed a number of burdens on all climbers. The waiting period was designed to permit the NFS to educate climbers as to the risks of Denali and the preparations necessary to meet them adequately. The post-1992 assessment indicated that an important source of problems on Denali was foreign climbers with little money, whose lack of preparation can have fatal consequences if they run into bad weather or an accident (Murphy 1998). While the 50 per cent of climbers ascending with a guided group received some screening and collective provisioning from the guide companies, these low-budget climbers were typically on their own or in small groups. Small accidents or routine bad weather was enough to require rescue for some of these climbers. The waiting period was clearly designed to discourage drop-in tourists who may have responded to short-term opportunities in the travel market. Presumably if a climber has to plan his trip at least two months in advance, there is a better chance that he will prepare for it; in any event, the NFS wants to encourage such preparation.

High Altitude Rescue as Moral Risk and Moral Opportunity

391

The official policy of the NFS, as published on their Denali website, remains in favour of rescuing those who need rescue but discretion is to be exercised as to when and how. 'Denali National Park and Preserve recognizes that a certain number of park visitors each year will become ill, injured, or incapacitated in some way. It is the policy of Denali National Park and Preserve to assist those in need, when, in the opinion of the park personnel appraised of the situation, it is necessary, appropriate, within the reasonable skill and technical ability of park personnel and provides searchers and rescuers with a reasonable margin of safety.' The NPS has also launched a cultural initiative aimed at encouraging self-preparedness for rescue on Denali that invokes the tradition of rescue as a duty in mountaineering. To bolster the educational opportunity created by the waiting period, the NPS used the proceeds of the $150 climbing fee to hire climbing rangers to conduct training workshops and informally interface with prospective climbers, provide advice, and spot high-risk-for-rescue candidates. The NPS also engages in extensive document production aimed at climber education. As mentioned above, a forty-page booklet is available in any of five languages. The booklet, also available on the web, outlines virtually all features of Denali climbing that require advanced planning and provides at least introductory advice on how to proceed. The booklet is intended to raise awareness of the dangers of climbing Denali and invest a sensibility of self-reliance in the reader/climber: 'Those who depend upon rescue efforts of the strength and expertise of others to extricate themselves from difficult positions are inviting disaster. Helicopters and/or acclimatized rescuers are often not available or the weather prohibits their response. In the Alaska Range, travelers should be prepared with knowledge, equipment, strength and common sense to support their own expeditions.' The NPS strategy is to create a sense of responsibility for managing risk in the individual climber and in climbing teams and to provide a flow of information. Towards this end the NPS publishes an annual review of the past mountaineering season, available by mail or free on the web.17 The carefully edited newsletter comes complete with extraordinary photographic images of Denali, and in parts reads like a sophisticated contemporary mountaineering magazine without the advertisements. The Annual Survey describes all recent notable ascents and route developments. A separate section provides a brief but careful description of all the rescue operations mounted. In bold type the newsletter offers a meta-commentary on what lessons can be drawn from each case about climbing parties, their equipment, or their strategies. Its

392 Jonathan Simon economical discussions of the major threats to survival incorporate harrowing and inspiring passages from mountaineering literature set on Denali. Here are two examples: HACE, On May 29th, an American climber descending solo near Denali Pass became disoriented due to High Altitude Cerebral Edema (HACE) and was assisted down to the 17,200-foot camp. Later that same day, the other two members in his party had descended with the group's tent and stove. The sick climber was provided shelter that night at 17,200 feet, then escorted down the ridge the following day where Ranger Gordy Kito and NFS Volunteers assisted the climber to the 14,200-foot camp. Because of a breakdown in group dynamics this near tragedy had to be prevented by fellow mountaineers. HAPE, On May 30th, an American climber was diagnosed with severe Acute Mountain Sickness (AMS) and High Altitude Pulmonary Edema at the 17,200-foot camp. A few hours later he was assisted by his climbing partner and other climbers down to the 14,200-foot camp. Unlike the previous incident, here was a good example of early recognition and a quick descent with the cooperation of expedition members and other climbers.18

The tone of the booklet is friendly and advisory rather than commanding and coercive. For those climbers who have picked up the message of how crucial preparedness is for Denali ascents, the booklet is one of several sources providing a stream of information crucial to that project. Since the new strategy was implemented in 1995 there have been signs that rescues and fatalities are coming down. In 2000 there were no climber fatalities on the mountain. Ironically, the NFS itself suffered the most severe accident of the year, the crash of a private plane carrying mountaineering ranger Gale Shaffer and two volunteer high altitude patrol members, Brian Reagan and Adam Kolff. The crash, a routine feature of Alaska travel, killed all aboard, including veteran bush pilot Don Bowers. This team was flying into the NFS base camp on Denali to take up their regularly scheduled patrol on the upper mountain. Moral Risks and Subsidized Immorality on Denali

Despite the decrease in both rescue operations and fatalities, the NFS is under increasing pressure from political leaders, who view its policies towards climbers as overly permissive and costly. Legislation authored by Alaska's Republican Senator Frank Murkowski in 2000 called for studies

High Altitude Rescue as Moral Risk and Moral Opportunity 393 of privatizing rescue efforts. At the same time, many veteran climbers and guides have adopted a version of the moral risk critique of insurance, arguing that the NFS has made the problem worse by creating expectations of rescue and that these expectations are eroding the traditions of self-reliance that mountaineering has traditionally promoted. These issues came to a head at the hearings called by Senator Murkowski in August 1998. Media attention had been drawn back to Denali earlier in the 1998 season, when a party of British soldiers ran into trouble on the upper reaches of the mountain, necessitating a spectacular and costly rescue. While the British government was fulsome in its praise for the NFS following this successful rescue, the agency found itself under attack for misplaced generosity. In an era of budget cutting, service privatization, and welfare reduction, it was easy to paint the NFS mountaineering strategy as an expensive program to subsidize an international youth counter-culture of climbers. Although at a million dollars a year the approximate government cost for mountaineering at Denali is not hard to lose in the federal budget, Senator Murkowksi and other critics view the mountaineering program as an important symbolic precedent for resource allocation. The growing rescue program on Denali received a scathing moral risk critique from one of its own when Waterman's Surviving Denali was published in 1983. Waterman has been participating in the Denali mountaineering scene since the late 1970s and has worked for both guide services and the NFS. Most of the book is a careful analysis of accident patterns on Denali, and its frequently drawn practical conclusions have made it an important tool for climbers preparing for Denali. However, running throughout the book is the theme that the expansion of rescue resources on Denali has been counter-productive. The first chapter of the book, 'The Self-Sufficient Pioneers,' chronicles mountaineering patterns on Denali before the popularization of mountaineering and the growth of NFS rescue efforts. Accidents happened in the pioneer days, but Waterman argues that they were rare because climbers understood that any accident could become fatal given that no rescue was possible. Then came the fall. 'Perhaps climbers were lulled into a false sense of security because of the possibility of rescue or by the belief that there was safety in numbers. Self-sufficiency was superseded by radios, helicopters and dependence on other climbers. All of these factors had an adverse effect, causing caution to be thrown to the winds; such behavior would have been foolhardy for the radioless, isolated pioneer climbers' (Waterman 1991: 15). Of all the features of mountaineering management that have made

394 Jonathan Simon Denali more attractive to less-prepared climbers and encouraged moreprepared climbers to be less cautious, none is more pernicious in Waterman's view than the helicopters which have increasingly become a crucial part of rescue on Denali. The first helicoper rescue occurred in 1954 (Waterman 1991: 5). A pioneer-style climbing team had attempted the dangerous South Summit. The lead climber slipped, pulling the other roped team-mates down 900 feet. One climber died and a second broke his hip; the other two hiked out to seek help. Six days later a helicopter brought a seven-man rescue team to 6,000 feet, the highest helicopter landing on the mountain at that time. This team succeeded in bringing the injured man out alive to a lower pass for a routine helicopter evacuation. Despite the technological assistance, the 1954 rescue effort still reflected the pioneer values. The team had stayed together and team members had hiked out to get help, there being no percentage in waiting to be rescued at that point. Over the next forty years the extent of helicopter rescue would grow enormously. Tn time, as more climbers came, self-sufficiency was forgotten, helicopters flew higher and higher, and rescues became commonplace' (Waterman 1991: 6). With high altitude helicopters injured climbers can be picked up as high as 17,000 feet, in the right weather, leaving team members little to do but radio or cell phone for a pick up (ibid.: 33). Ironically, while the NFS continues to invest heavily in its technologically oriented rescue program, the post-1992 mountaineering program has sought to promote the mentality of self-sufficiency that Waterman espouses. Indeed, his work is heavily cited and quoted in the NFS brochure and on their website. As I shall argue further below, the NFS strategy seems directed at counterbalancing the incentive effects of organized and financially free rescue with moral effects of identity, by promoting a mountaineering identity in which self-sufficiency and duty to rescue others play prominent roles. This response, however, has thus far won the NFS few friends. The commercial guides and a significant part of the politically organized mountaineering community reject the current NFS approach as excessive, unfair, and prone to moral risk. Dunham Gooding, president of the American Alpine Institute, a commercial guide service based in Bellingham, Washington, that operates a concession on Denali as well as in five other states and eleven other countries, took a strongly critical view of the NFS approach in his written comment to the committee. Gooding related the discussion that took place between a group of

High Altitude Rescue as Moral Risk and Moral Opportunity 395 concession holders and NFS officials in Anchorage in 1994. According to Gooding, the concession holders unanimously supported removing the high altitude 'Lama' helicopter that the NFS maintains by private contract on an 'at ready basis.' The NFS manager was strongly opposed to limiting NFS access to the Lama, noting that it would 'take away the most important tool his staff has, and that to eliminate it would endanger them.' We responded by noting that we understood they would not be able to do everything they are doing now. They would clearly have to function at a new level, one at which they would not endanger themselves. McKinley's location in the Arctic as well as its high altitude and technical challenge make it an extremely difficult climb that involves a high level of hazard and risk. There is neither an expectation nor a desire in the climbing community that the NFS seek to change that. It does appear to the climbing community that the NFS is attempting to change the level of risk on McKinley. (Energy and Natural Resources 1998: 62)

Gooding and other guides appear ready to see virtually the entire NFS effort dismantled. They question the hiring of additional mountaineering skilled rangers, the heated hospital tent/rescue station at 14,000 feet, and especially the high altitude helicopters. Given the fact that rescues launched from the base of the mountain are likely to come too late, this would leave the NFS with little to do about the climbing accident problem, other than explain to the public and bereaved parents that 'the desire to confront risk and hazard is part of the human character' (ibid.). Other testimony at the hearing suggested that critics see the high altitude helicopters as an especially excessive and counter-productive part of the rescue operation. Gary Bocarde, a veteran Denali climber and owner of Mountain Trip, Inc., an Anchorage-based guide service and concession holder on Denali, suggested that even though the NFS used helicopters in the past, the altitude limits on helicopter landings forced climbers on the upper reaches of the mountain to participate in rescue by bringing the injured climber(s) down low enough on the mountain for a safe helicopter landing. Since these informal rescuers would be both on the scene and dedicating their own time and labour with inevitably some risk of weather turning deadly - to bringing the injured party from the very top of the mountain to the 14,000 foot level (Energy and Natural Resources Committee 1998: 42), the call for rescue

396 Jonathan Simon would be more hesitant and circumspect. This might have led to some fatal dithering by reluctant rescuers, but it also assured a close reading of precisely how important evacuation by air was by people close to the events and with a stake in the decision. Under the high altitude helicopter rescue regime currently in place, parties can call the NFS to pull their climbing partners off the very top of the mountain while they concentrate on their summit attempt, unimpeded by decisions about how serious the condition is and whether they could carry out a rescue safely if they abandoned their own summit ambitions. Bocarde and other commercial guides claim to prefer abandonment of the program to more aggressive efforts to capture the costs promoted by Senator Murkowski and others. The latter approach would only worsen the moral risk problem: So having these things on the mountain, even though - I mean, it is great when there is a rescue, and obviously lives are saved, but I think we messed up with the mindset of climbers, and actually people are counting on these things, and if you increase the cost, I would bet you money you are going to increase the number. I mean, if I am a climber, I pay $500, and I sprain my ankle, I am going to call for a rescue. It is just - that I think you need to look at these factors before you automatically increase this, the user fee for the climbers, and nowhere else is there a user fee ... I think by having all these various things that I mentioned, it makes people not be self sufficient, and in the seventies, people, the groups were up there that needed a rescue - they were a lot more self sufficient. (Energy and Natural Resources Committee 1998: 42-3)

In his written statement for the committee, Bocarde suggested that NFS is susceptible to its own kind of moral risk. The availability of the helicopters under their own direct control 'makes it too easy for the NFS to get caught up in the desire to do spectacular rescues or body recoveries' (Energy and Natural Resources Committee 1998: 44).19 Gooding likewise commented, 'it must be very difficult to manage a park where you know people are going to die climbing every year,' suggesting that a cultural clash of values between the NFS and climbers might be the real issue at dispute (ibid.: 62). The guides testifying in the 1998 hearings cannot realistically claim to speak for Denali climbers as a community, but no other sector can make the claim either. Denali, unlike many renowned mountaineering areas,

High Altitude Rescue as Moral Risk and Moral Opportunity

397

has no well-developed local, autonomous mountaineering community. There are professional guides on the mountain who climb with clients for money, but the kind of autonomous mountaineering community that thrives in California, Arizona, Washington, Colorado, and New England, to name only U.S. examples, does not exist. The organized climbers and guides who are willing to do without the high altitude helicopters and the mountaineering rangers do not necessarily represent the interests of less experienced or unorganized climbers. In the absence of the NFS, rescue is likely to depend heavily on the guide services. Presumably if the NFS lost congressional support for high altitude rescue, the commercial guides would make their own deals with helicopter pilots, deciding between air evacuation and rescue from the mountain on the basis of their own cost structure. The big losers in that reorganization will be independent climbers who opt to climb the mountain unguided and run into trouble. As in the nineteenth-century Alps, rescue would be largely up to the guides on the mountain. 20 Commercial companies like those of Gooding and Bocarde make use of the rescue services themselves on behalf of their clients, when it is convenient, and currently at no charge. More importantly, high altitude rescue contributes to an unknown degree to the larger flow of clients that has fed the guide companies. Bocarde did go out of his way to suggest that his guides were extensively involved in aiding NFS rescue operations, a participation characterized by NFS administrators in their testimony as 'voluntary.' As a business owner, Mr Bocarde was no doubt warning the pro-business senator from Alaska that any effort by the NFS to pass costs on to the guide companies might be met by a reciprocal commodifying of their rescue participation. Short of scaling back the cost of rescue, the consensus among even the guides and organized climbers breaks down. Mr Gooding, for example, argued strongly against any demand for rescue payment, a policy he suggested would violate traditional American values of rescue - services like police and fire, for example, rarely if ever charge specific users, as opposed to all taxpayers, for their services. Bocarde, in contrast, supports compulsory insurance that would amount to a rescue charge (Energy and Natural Resources 1998: 45). Congress has yet to consider any definitive legislation on the matter. Senator Murkowski enacted a law creating a commission to study the problem, which had not reported by 2001. But his comments at the hearings suggest a likely direction. First, he clearly believes that the general public as taxpayers is paying too much for rescue at Denali.

398 Jonathan Simon Pressed by all the witnesses to consider the rescue and similar resource demands that all recreational users place on the government through the NFS, Murkowski drew an interesting distinction that could form the basis for a legislative effort to concentrate the costs of rescue on climbers. 'You know climbing Mount McKinley in the eyes of most people is an unusual and very unique adventure, as opposed to a boater going across and getting in trouble, and there is a select few that have the desire or perhaps the capability, or both' (Energy and Natural Resources 1998: 54). Under the Murkowski doctrine, climbers should bear the deep costs because they seek (and, if they succeed, attain) a special and different kind of 'deep benefit' from the mountain. More particularly, Senator Murkowski advocates that rescue costs should be redeemed through private insurance purchased by the climber.21 The NFS itself seems reluctant to move towards mandatory insurance and billing for rescue. They clearly felt the difficulty of responding to the fact that they make no effort to collect for rescue, offering no defence. Speaking for the NFS management, Denali Chief Ranger Ken Kehrer offered a moral risk critique of the mandatory insurance problem. The difficulty that we have with insurance and rescue was exemplified by the French, though, with Mount Blanc, and while they do have an extensive rescue program what they reported was that because people had rescue insurance they were less likely to rescue themselves. The amount of self-reliance was diminished. Consequently, there is an expectation of, I have insurance, therefore, come get me, I am hurt' (Energy and Natural Resources 1998: 12). Clearly, the moral risk critique of mountain rescue has real analytic purchase. It is not rescue alone that has enticed more than a thousand climbers a year to attempt to reach the summit of Denali - one has to credit the revolution in adventure tourism, in cheap travel, and in the media coverage of Denali as well - but rescue has obviously played an integral role in encouraging climbers of a less experienced sort, and more experienced climbers in a more relaxed mindset than they would otherwise have. On the other hand, taking a decision now to withdraw high altitude rescue operations would doom an unknown number of climbers in the near future to deaths that would have been avoided rather easily a year or two before. One can imagine the congressional hearings that might be generated by a well-publicized refusal to rescue someone with the military helicopters available and the death of some photogenic college sophomores as a result. Mandatory insurance would almost surely raise the number of rescues

High Altitude Rescue as Moral Risk and Moral Opportunity

399

in the short term. The leap from a current mountaineering fee of $150, which the NFS constantly reminds climbers is not a rescue insurance premium, to insurance costs of at least several hundred dollars, would inspire a greater willingness to call on rescue.22 If the NFS continued to manage rescue and recovered costs from the insurance companies, those companies might concede that the NFS should continue to do the educating and cultural investing in self-sufficiency that it currently engages in, but that is only one possibility. Certainly, whoever is managing rescues at Denali is going to have to undertake a much more coercive policing of Denali access than is now the case. At a minimum this will shift the role of the NFS rangers, or whoever takes on this function, in a police direction. Fines for climbing the mountain without insurance will have to be severe enough to make people avoid rescue even when it might kill them, or, perhaps more ruthlessly, to send a mountaineering ranger/cop tumbling if one should decide to visit a climbers' tent on some chipped-out ice ledge along the treacherous Cassin Ridge.23 Moral Opportunities on Denali The NFS rescue program on Denali shares with modern forms of insurance as loss spreading the creation of both moral risks and what Deborah Stone has usefully called moral opportunities. Following Stone (2002), I think of moral opportunity not simply as a question of efficient loss bearing but as a question of the potential to effectively govern lossrelevant behaviour itself. Like moral risks, moral opportunities are not linear but spread out in an institution across any practice or discourse in which loss-relevant knowledge or power has strategic value. Pursuing the chain of moral opportunities requires us to consider not just loss but the social meaning of loss, and the social norms regarding mutual aid, standard setting, raising technology levels, and problematizing losses. In this light we can see a range of moral opportunities opened by high altitude rescue on Denali. The availability of rescue no doubt encourages risk takers, but it also encourages investments by those involved with rescue in mechanisms for achieving greater levels of security. To take the most direct example, the capacity to engage in high altitude rescue has progressed continually because of the investment of the NFS in leasing the necessary technology and compensating the pilots. In turn, pilots and rescuers' experiences in extreme conditions have raised their rescue skills to new levels.24 The value of those technologies and skills now and in the future is unknowable; indeed, rescue has become

400 Jonathan Simon its own practice, like mountaineering itself, with a cultural capital whose value is inherently difficult to assess. While these skills and technologies may generate positive externalities of their own, they increase moral risk by encouraging those with less skills to climb the mountain and those with high skills to take greater risks there. Since 1994 the NFS, through its mountaineering program, has implemented countermeasures to the moral risk that represent in some respects a self-conscious strategy to promote the moral opportunity of rescue. One of the most important choices the NFS has made is to invest in the discourse of mountaineering itself as something that sustains a self-sufficiency ethic. The NFS is mobilizing mountaineering talent by hiring rangers, publishing books and newsletters, and organizing volunteers. Rescue in this sense provides an opportunity to revitalize the culture of mountaineering at a time and in a place where that culture is under stress from a number of competitors and operating in a global mountaineering circuit with a large diversity of mountaineering cultures. Stone (2002) points out that one of the most important roles of insurance is creating new leverage to demand care taking by insureds. By fostering rescue, even if it stops short of promising a rescue, the NFS program imposes new duties on climbers, including taking more responsibility for their own and others' rescues. The NFS effort is actually codifying a duty to rescue in North American autonomous mountaineering that has always been implicit but rarely made explicit. Whereas the 'climbing code' in the influential mountaineering handbook, The Freedom of the Hills, contains no explicit duty to rescue, the Denali climbing code does. The NFS booklet states in bold letters that 'Mountaineers will be Expected to Aid Other Climbers in Emergencies' (NFS 2001: 16-17). From a moral opportunity standpoint we should also note the gains created by the very behavioural adjustments that traditional moral risk analysis treats as inherently negative. The creation of a large pool of climbers who attempted Denali in the 1980s and 1990s has contributed in countless ways to the development of mountaineering as a global cultural practice. People who would not have been able to have this experience before because of the unreasonable risk of death now attain the summit, acquiring whatever psychological and social gains edgework (Lyng 1990) produces (which for present purposes I would simply assert is more than zero). There is even a sense in which the latent moral opportunity effects of risk taking may in time directly counteract moral risks. In the last decade

High Altitude Rescue as Moral Risk and Moral Opportunity 401 there has been a growing recognition within the global mountaineering public that climbing Denali is a serious threat to life even with the best rescue efforts and that climbers must prepare in advance for extreme adversity even on the tourist route. The tales of the thousands of those who did climb the mountain have contributed to that recognition just as much as they have invited further visitors. The sheer number of adventurers on Denali has enabled an unprecedented knowledge base about accidents on Denali in the realized form of Jonathan Waterman's Surviving Denali, now in its second edition. That book has clearly affected behaviour on Denali by providing a very clear and chilling portrait of precisely what can go wrong. It has also advanced extreme weather and altitude accident analysis generally by allowing the accumulation of an unprecedented body of human data. It is possible that these cultural changes alone would have reduced the moral risk problem on Denali. Whether the reduction in the number of rescues and fatalities annually on Denali since the disastrous 1994 season can be attributed to growing self-consciousness in the mountaineering community, to the efforts of the NFS mountaineering program, or simply to good weather, is not yet clear. Conclusion The effort of the NFS represents a unique strategy aimed at fostering an existing but weakened system of mountaineering ethics while simultaneously creating technological solutions to the failures of human effort. Interestingly, it is a strategy that deliberately eschews traditional insurance in favour of fostering direct rescue efforts. Yet even insurancebased loss management strategies have much to learn from such experiments in how to harness self-regulating cultures as a mechanism to counteract both moral risk and the cultural devaluation that the moral risk critique has produced. This is particularly true of fields like workers' compensation. Because it provides income for those not working as the result of certain kinds of injuries, politicians have found it easy to characterize the moral risks of the system as a palpable political problem. But the underlying forms of solidarity associated with the workplace provide a potential source of self-correction and revalorization. As the Denali example suggests, the politics of such strategies are inherently risky. The National Park Service mountaineering program at Denali finds itself challenged from the right by neo-liberal attacks on moral risk and subsidization, and from the populist left for its demand

402 Jonathan Simon for more control over climbing behaviour on the mountain. The base on which its success ultimately depends, a vigorous and self-regulating mountaineering community on Denali, is still a work in progress. Notes 1 In this chapter I follow Richard Ericson in using 'moral risk' rather than 'moral hazard.' As Ericson and his colleagues have shown (see chapter 12 this volume; Ericson, Doyle, and Barry 2003) moral hazard talk tends to isolate risks generated by insurance coverage from other kinds of risk. 2 Department of the Interior, National Park Service, Denali National Park and Preserve, Action: Notice - Mountaineering Program, Thursday, 15 December 1994, 59 Federal Register 64696. 3 Department of the Interior, National Park Service, Denali National Park and Preserve, Action: Final Rule, 61 Federal Register 6943. 4 Murkowski has less compunction against subsidizing mining, fishing, and hunting on federal lands. See 'Murkowski Wins Senate OK of Four Alaska Bills,' 19 November 1999, Government Press Releases by Federal Document Clearing House, Inc. (1999), Westlaw locator number: 1999 WL 28846520. 5 An Act to Review the Suitability and Feasibility of Recovering Costs of High Altitude Rescues at Denali National Park and Preserve in the State of Alaska, and for other Purposes, 9 November 2000, 114 Stat. 2201, Public Law 106486 (S698), 106th Congress, Second Session. 6 Courts have often treated roped climbers as an exemplary case of affirmative duty creation. The act of cutting loose a fallen climber who cannot be rescued but who is inevitably pulling the survivor off the mountain is a classic example of the defence of necessity in criminal law. 7 There is a striking parallel in both law and literature between mountaineering in this regard and seafaring. Admiralty law, for example, created far more extensive protections for the labouring sailor than contract law on land recognized in the nineteenth century. 8 Roping climbers and guides together was controversial throughout the period among both climbers and guides. Guides viewed the rope as inherently degrading. They used it originally only to keep tourists together and on the same path, generally with a guide at each end holding the rope (Ring 2000: 183). Many climbers believed as well that ropes impeded climbing and that if a climber slipped at a steep place the rope would only pull down everybody else. During the 1850s, however, roping was promoted by some of the leading Victorian climbers. The new vision, articulated by

High Altitude Rescue as Moral Risk and Moral Opportunity 403 climber and writer John Tyndall, was that climbers and guides roped together about the waist could use the alpenstocks in their hands, freed from holding the rope, to secure each other (Ring 2000: 184). By the time of Whymper's ascent of the Matterhorn in 1865, the party used the newly orthodox rope techniques that included roping all guides and climbers together around the waist, keeping sufficient distance between the climbers to allow time for reacting, and taut rope so that the weight of a falling climber could be spread among all of those above the fall (Fleming 2000: 275). Despite these precautions the rope used on the Matterhorn descent broke, sending the lead guide and three English gentlemen to their death. The eventual consensus on the fall, however, blamed the weakness of the rope, or its being cut, and not the practice of roping per se. 9 While many of the top competition and sponsored climbers do 'solo' climbs where they climb alone and unroped, the practice remains profoundly controversial. Occasional photos of an unroped climber in the major mountaineering magazines almost always attract a raft of critical letters to the editor. 10 Perhaps sparked by the post-1996 market for books on fatal mountaineering accidents, two books have recently been published on the development of the mountaineering scene in the Alps in the mid-nineteenth century. Both Fleming (2000) and Ring (2000) address the dominant role of English climbers and entrepreneurs in turning one of Europe's poverty zones into what remains a fabulous playground for the rich and famous. Both draw on the remarkable work of Ronald W. Clark (1949). 11 Like any client relationship it could also go badly. Edward Whymper, the most celebrated climber of the mid-nineteenth century, infamously characterized guides as 'pointers out of paths and consumers of large quantities of beef and wine' (Whymper 1890). He would develop a more sentimental view after a different guide, Michel Croz, led him to the summit of the Matterhorn and then died trying to protect an even less qualified English climber on the way down in the above mentioned (see note 8) infamous accident that killed three English gentlemen as well. Whymper's book on the accident and his general adventures in the Alps became a Europe-wide bestseller in the 1870s. 12 The stories that fill the book are written as thrilling tales of mountain danger but also provide useful ethnographic detail about the composition of rescue parties, interactions between guides and clients, and post-accident judicial inquiries. 13 Krakauer (1998) witnessed and documented in his book, Into Thin Air (no doubt the best-selling mountaineering book of all time). 14 Noah Adams, 'Survivor of Everest Catastrophe Discusses Climb: Interview

404 Jonathan Simon with Jon Krakauer,' All things Considered, National Public Radio, 16 May 1996, 4:30 p.m., Transcript No. 2215-13. 15 The sheer size and remoteness of the mountain means that even parties that survive an initial accident and have great skill may not be able to escape death. In 1981 Chris Kerrebrock, an expert climber, fell into a crevasse uninjured but was wedged too tightly to be pried out by his equally expert mountaineering partner Jim Wickwire (Waterman 1991: 148-9). Caught in an area in which their line-of-sight C.B. radio technology was blocked by the landscape, the climbers were on their own although only minutes by air from rescue. Kerrebrock finally told Wickwire to abandon him rather than risk becoming hypothermic himself in the deep crevasse. Kerrebrock died while Wickwire waited in stunned horror a hundred vertical feet away. Wickwire barely made it to the agreed-upon pick-up zone with his own life. His account is related in Wickwire (1998: 8-24). 161 have not uncovered the precise timing or reasoning for the change but an important turning point may have been a disastrous accident in 1967, in which two teams forced together by the NFS as a condition for permission to climb a remote and difficult route fragmented and seven of the weaker climbers died of falls and exposure trying to descend from the summit (Waterman 1991: 13). 17 Denali National Park and Preserve, 2000 Mountaineering Summary. 18 Ibid.: 8. 19 Bocarde accused the NFS of wasting money on body recoveries. The modern tradition within autonomous mountaineering has been to leave bodies in the mountains if recovery would require any significant risk. 20 This would come close to the system organized at Chamonix near Mt Blanc, the first town to capture the nineteenth-century tourism in the Alps beginning in the 1820s. In the Chamonix system a corporation of guides controlled the terms and conditions of guided climbing, including prices, the appropriate numbers of guides, and the methods of climbing used. While the corporation had methods of disciplining guides who worked in violation of the regulations, the only real check against autonomous groups of foreigners climbing the mountain on their own terms was the threat of not providing a guide-based rescue if something went wrong. Generally, the basic Christian ethic made it impossible for the guides to utterly refuse rescue - although such public bans would subsequently be used as a protest against extreme risk taking - but they certainly modified the extent of hardship and risk they were prepared to endure. 21 Gary Bocarde, the commercial guide, also endorsed a European model of compulsory insurance (Energy and Natural Resources 1998: 45). The same point was raised in a letter to the committee from the British Honorary

High Altitude Rescue as Moral Risk and Moral Opportunity 405 Counsel for Alaska. The British interest was due to the media storm about the rescue of the British climbers on a remote part of the mountain earlier that season. After offering the gratitude of his government, the Honorary Counsel made the comparative point. 'It would seem to me that, at least in the case of foreign climbers - irrespective of whether they are military or civilian groups -wishing to climb McKinley, the U.S. National Park Service should require in advance that the party have insurance that, in the event of a rescue operation, would reimburse the U.S. agencies for the cost. This insurance should not be by local bond but be covered by one of the major international insurance companies that provide coverage for international adventure tourism. This would preclude the group from feeling that, since they had paid a local bond for a rescue, they could activate its use in less than emergency situations. Requiring such insurance would undoubtedly involve a fairly hefty insurance premium, which would ensure that the foreign climbers would be further reminded of the difficulty and danger of the climb before they undertook it' (Energy and Natural Resources 1998: 59). 22 The head of the National Outdoor Leadership School stated that the school offered rescue insurance at approximately $200, including not only evacuation but the $4,000 tuition refund if an injury causes a client to withdraw from the program. This premium, and similar premiums offered through commercial guide services, surely reflects the fact that at present the NPS does not pass on rescue costs to anyone, including insured climbers. Obviously, any move to selectively charge insured climbers would create a rapid dissolution of the insured pool altogether, as climbers would opt to forgo coverage and take their chances with the NPS (which the commercial companies and the NOLS currently permit them to do). 23 Policing the summit would be one solution but it would miss the majority of climbers and pose both enormous survival challenges for the police themselves and enormous analytic challenges to cyber-surveillance in a place where, at certain times, visibility of a few feet cannot be had. 24 It emerged at the hearing that the high altitude helicopter leased with crew by the NPS from a private contractor was no longer even being made by the military industries. References Abraham, K.S. 1995. Insurance Law and Regulation: Cases and Materials. 2nd ed. Westbury, NY: Foundation Press Baker, T. 1996. 'On the Genealogy of Moral Hazard.' Texas Law Review 7'5: 237-92 Baker, T., andj. Simon, eds. 2002. Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press

406 Jonathan Simon Chessler, M. 1999. 'After Thin Air.' American Alpine Journal 7'3: 171-7 Clark, R. 1949. The Early Alpine Guides. London: Phoenix House Energy and Natural Resources, U.S. Senate Committee. 1998. Hearings: Mount McKinley Rescue Activities, Monday, August 24, 1998, Anchorage Alaska Ericson, R.V., A. Doyle, and D. Barry. 2003. Insurance as Governance. Toronto: University of Toronto Press. Fleming, F. 2000. Killing Dragons: The Conquest of the Alps. New York: Atlantic Monthly Press Goss, C. 1948. Alpine Tragedies. Trans. M. Barnes. London: George Allen and Unwin, Ltd. Heyman, SJ. 1994. 'Foundations of the Duty to Rescue.' Vanderbilt Law Review 47: 673-755 Kjeldsen.Jim. 1998. The Mountaineers: A History. Seattle, WA: The Mountaineers Krakauer.J. 1998. Into Thin Air. New York: Anchor Lyng, S. 1990. 'Edgework: A Social Psychological Analysis of Voluntary Risk Taking.' American Journal of'Sociology 95: 851-86 McCluskey, M. 1998. 'Illusion of Efficiency in Workers' Compensation Reform.' Rutgers Law Review 50: 657-941 Mitchell, R.G.,Jr. 1983. Mountain Experience: The Psychology and Sociology of Adventure. Chicago: University of Chicago Press Murphy, K. 1998. 'To Some, the Height of Lunacy.' Los Angeles Times, Wednesday, 1 April Al Murphy, L. 2001. 'Beneficence, Law, and Liberty: The Case of Required Rescue.' Georgetown Law Journal 89: 605-65 Murray, C. 1984. Losing Ground: American Social Policy, 1950-1980. New York: Basic Books National Park Service. 2001. Mountaineering in Denali National Park, Alaska. www.nps.gov/dena/mountaineering/talkeet.html. Ring, J. 2000. How the English Made the Alps. London: John Murray Simon, J. 2002. 'Taking Risks: Extreme Sports and the Embrace of Risk in Advanced Liberal Societies.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press. Stone, D. 2002. 'Beyond Moral Hazard: Insurance as Moral Opportunity.' In T. Baker andj. Simon, eds., Embracing Risk: The Changing Culture of Insurance and Responsibility. Chicago: University of Chicago Press. Waterman, J. 1991. Surviving Denali: A Study of Accidents on Mount McKinley 1903-1990. 2nd ed. Golden, CO: American Alpine Club Whymper, E. 1890. Scrambles Amongst the Alps. Philadelphia: Lippincott Wickwire,J. 1998. Addicted to Danger. New York: Pocket Books

14

The Neurochemical Self and Its Anomalies NIKOLAS ROSE

What are the implications of recent advances in the life sciences for the government of human conduct? In this chapter, I want to consider this question in relation to one specific area - the contemporary government of disorders of desire.1 Not disordered desire in the sexual sense, which, despite the colonial ambitions of psychoanalysis, was never more than one small corner of the empire of desire. I want to consider those forms of disordered desire known as cravings and addictions, disorders that Mariana Valverde (1998) has termed 'diseases of the will.' I am interested in the mutations that may be occurring in our ways of thinking about and acting upon these pathologies in the context of recent developments in the life sciences and biomedicine - the decoding of the human genome, the new neuroscientific knowledges of the brain and its mechanisms, and the linked developments in the technologies of treatment, notably psychopharmacology. In this shift, 'diseases of the will' have become, at least in part, diseases of the brain. One index of this transition is the emergence of a number of new kinds of drugs that are being used in the treatment of addictions - or as psychiatrists now like to call them, 'dependencies.' The best example here is probably a drug called naltrexone. Naltrexone is used in the treatment of addiction to opiates and of alcoholism, or what is termed in the Diagnostic and Statistical Manual of the American Psychiatric Association (DSM-TV) (1994) alcohol dependency. The interesting thing about this class of drugs is that they treat addictions not by making the satisfaction of desire unpleasant, as Antabuse does, and not by substituting a more acceptable addiction, such as methadone for heroin, but, apparently, by removing or reducing a particular desire. They act upon desire. Take this drug, and the craving of the alcoholic for drink, of the junkie for drugs, will diminish.

408 Nikolas Rose While this is the story reported in the popular press, the arguments in the scientific or professional literature are more complex and nuanced (see chapter 15 this volume). But my aim here is to use these endeavours to reshape pathological desire by pharmaceutical means as a way of trying to understand more general shifts in the government of pathological conduct. On the one hand, I will argue, we can see a shift in the kinds of persons upon which such government will act - the birth, that is to say, of genetic and neurochemical selfhood. On the other hand, we can see the rebirth of an old governmental aspiration that Mariana Valverde terms 'targeted governance.' Targeted government - the direction of regulation towards those specific sites and persons thought most problematic - may seem at odds with the important argument that control, today, is not individual and reformatory but collective and probabilistic, aimed at risk reduction rather than therapy or reformation. But there is no real contradiction. In post-social strategies of government much control is indeed 'designed in,' as suggested by terms such as 'actuarial justice' or 'societies of control' (Feeley and Simon 1992; Deleuze 1995). That is to say, all manner of everyday locales such as amusement parks, and practices such as shopping, have been explicitly shaped to secure specific types of conduct. But such attempts to 'design in control' are coupled with strategies that seek to target specific 'high risk' or 'high need' persons or sites which are thought to require particular attention (Rose 2000a). My specific focus is the pharmaco-therapeutics of desire. This is not quite what Peter Kramer (1994) was thinking of when he coined the term 'cosmetic psychopharmacology,' although he did talk about the way in which Prozac - the cosmetic he was discussing - diminished the compulsion of one of his patients to make use of pornography. Prozac was an SSRI, a serotonin selective reuptake inhibitor; naltrexone is something different, as we shall see. But despite these differences, the argument made here reflects a similar way of thinking. And this way of thinking is exemplary of a broader shift in what I have termed, following Ludwik Fleck and Ian Hacking, the 'style of thought' of biological psychiatry (Rose 2000b; see also Fleck 1979; Hacking 2001). In this style of thought, the old oppositions between mind and brain are starting to be re-posed. The psychological space between organs and conduct that opened up in the twentieth century is beginning to be flattened out.2 Individuals are becoming somatic - in this case genetic and neurochemical - selves. Yet, contrary to many predictions of the critics, this is not a rebirth of essentialism, reductionism, or geneticism. As Carlos Novas

The Neurochemical Self and Its Anomalies 409 and I have argued, plenty of scope remains for autonomy, choice, individuality, and responsibility (Novas and Rose 2000). Molecularization

The fourth edition of the Diagnostic and Statistical Manual of the American Psychiatric Association (1994: 194-204) identifies fifteen kinds of alcohol related disorders divided into two types. The first, alcohol use disorder, includes alcohol dependency and alcohol abuse. The second, alcohol-induced disorders, has thirteen varieties from alcohol intoxication (!) to alcohol-induced sexual dysfunction. Each has its own code, and its own diagnostic criteria. These are just a few of the approximately 350 categories of psychiatric illness which DSM-IV defines. Attempts at the systematic classification of types of mental illness in the United States began as a statistical, not a clinical project. The 1840 census records the frequency of 'idiocy/insanity'; the 1880 census used seven categories (mania, melancholia, monomania, paresis, dementia, dipsomania, and epilepsy); and the first official nosology, adopted in 1918, was an attempt to classify the considerable population of individuals then confined in mental hospitals. Over the course of the twentieth century, governmental and clinical aspirations fused, and classification systems demonstrate the rise and fall of different conceptual models of mental illness. The first DSM, produced in 1952 in the wake of psychiatry's wartime experience, was constructed in terms of Adolf Meyer's notion of reaction types and conceived of mental disorders as reactions of the personality to psychological, social, and biological factors. DSM-II, published in 1968, contained 180 categories framed in the interpretative language of psychoanalysis. DSM-III, published in 1980, is often seen as a response to the crisis in legitimacy of psychiatry over the 1970s, a response that indicated a reshaping of the psychiatric gaze. Where DSM-II was 134 pages long, DSM-III ran to almost 500 pages, and the revised version of 1987 contained 292 categories. This reflected a shift to a new way of viewing mental disorders as discrete illnesses, each marked by a number of supposedly objective criteria, each, in principle at least, having a specific aetiology and prognosis, and being amenable to a specific kind of treatment. Thus the contemporary proliferation of categories may be more than bureaucratic inflation. Despite DSM-IV's recognition, at the outset, that individuals within any diagnostic group are heterogeneous, and that its categories only serve as aids to clinical judgment,

410 Nikolas Rose this refinement and proliferation of categories of pathology is indicative of something more, something we might term 'the molecularization' of psychiatry. The thought style of contemporary psychiatry has become inextricably bound up with developments in neuroscience that posit a specific anomaly in the brain - most frequently in one of the neurotransmitter systems - as the ground of each specific anomaly in mood, cognition, affect, or conduct. Thus, for example, a particular type of depression might now be linked to an anomaly in just one of the many subtypes of one of the seven sub-families of receptors for the neurotransmitter serotonin. As a further stage in this molecularization, the new genomics seeks the precise polymorphism in a particular sequence of bases in a locus in a particular gene that is correlated with a precise type of disorder of thought, emotion, or conduct. For example, attempts are made to link a particular subcategory of DSM-IV schizophrenia to a substitution in a single base in the sequence of DNA of a particular gene localized to a precise place on a particular chromosome, leading to a substitution of one amino acid for another in an enzyme involved in neurotransmission. In this style of thought, the broad categories of the start of the twentieth century - depression, schizophrenia, neurosis - have given way to a desire to dissect pathologies of mood, cognition, will, or affect at a different scale. The psychiatric gaze is no longer molar but molecular, and here this phrase should be taken literally. The same 'molecular' dissection of pathology can be seen in psychopharmacology - the fabrication, trialling, marketing, and prescribing of psychiatric drugs. Prozac, for example, became an iconic drug not because it was more effective than previous antidepressants, but because it purported to be the first drug whose molecule had been deliberately fabricated to disrupt one, and only one, aspect of a single neurotransmitter system. The claim was that this drug, fluoxetine hydrochloride, increased the level of available serotonin in the synapses by disrupting the reuptake of serotonin, and that it did this by locking into the precise sites on the neuronal membrane, the receptors, through which reuptake would otherwise have occurred. Hence all the hype about Prozac being a smart drug, a clean drug, which did not produce all those unwanted effects - shuffling gait, dry mouth, tremors - produced by the earlier clinically derived 'dirty' antidepressants, which hit a number of sites in a number of systems at the same time. The dream, that is, is of drugs that do not act upon the person as a whole, but are targeted precisely to

The Neurochemical Self and Its Anomalies 411 correct a specific anomaly now thought to underlie a specific undesirable variation of mood, emotion, conduct or will. These new ways of thinking have consequences for conceptions of risk and strategies for its management. I have argued elsewhere that the political vocation and professional practices of psychiatry have been restructured in terms of the management of risk. One might say that individuals with inherited molecular anomalies, which presumably they have had since birth, are persons 'at risk' and that interventions directed at them are yet more examples of 'risk government.' This would not be too misleading. But I think we lose something in intelligibility by recoding these endeavours in the language of risk. Risk implies probabilities and calculabilities. What we have here, to use the terms usually employed in the specialist literature, are instead predispositions, vulnerabilities, or susceptibilities. And these susceptibilities are now somatic and molecular. We are seeing the emergence of the concept of the susceptible somatic individual. This idea of susceptibility has very significant consequences for what I will call, for want of a better term, control strategies. These new ways of thinking also have implications for the ways in which mental pathologies are coloured with moral significance. One of the prevailing themes in sociological critiques of psychiatry from the 1960 onwards, evident in the work of Thomas Scheff and Erving Goffman, was that a psychiatric diagnosis acted as a 'master status.' It overrode and effaced all other categories of selfhood: the individual so diagnosed was not 'suffering from' schizophrenia, he was 'a schizophrenic.' And we know too that, from the mid-nineteenth century onwards, the catalogue of such master statuses produced an array of problematic figures: the criminal, the juvenile delinquent, the alcoholic, the prostitute, the degenerate, the homosexual. These designations did not merely refer to persons who engaged in particular forms of undesirable behaviour but designated a particular type of person. Their individuality was suffused by, exhausted by, their particular pathology, from the shape of their ears to their handwriting, their posture, and gestures, from the tone of their voice to their intelligence and capacity for moral judgment. These were 'the abnormal,' and their violation of norms was not a matter of irregularity of conduct but of deviant personhood. What we are now witnessing is a shift in the logic of the norm that underpinned such notions of 'the abnormal.' At the very least, we are moving from dividing practices based on the binary of normality and abnormality to practices based on

412 Nikolas Rose the idea that all individuals vary, and that most, if not all, suffer from molecular errors that are potentially correctible. In this new logic of correction, I shall suggest, what is required is a kind of constant adjustment of biology and circumstances. Biological Psychiatry as a Style of Thought When I say biological psychiatry is a style of thought (Rose 2000b), I mean to imply that it is a way of seeing, a way of explaining, a way in which reasoning is embedded within certain practical and intellectual techniques, conventions about experimentation, about instruments and inscriptions, measurements and model systems. These are conventions about sense-making activities themselves and about the terms under which they may be criticized, corrected, and revised. A style of thought is not just a type of explanation, it is about what counts as an explanation and what there is to explain. That is to say, it is about the very object of explanation, the set of problems, issues, and phenomena that an explanation is attempting to account for. A new style of thought brings, or tries to bring, a new object into existence, or so modifies an old object that it appears in a new light, with new properties, and new relations and distinctions with other objects. Let me examine this particular style of thought in more detail. In the nineteenth century, clinical medicine as we know it began when the gaze of the doctor plunged into the interior of the body, linking the symptoms that could be discerned on the surface with the organic locations and lesions within. From that moment onwards, to diagnose an illness was to interpret symptoms in terms of the internal organic malfunctions that were their cause. The new profession of psychiatry inevitably dealt with people who came to their notice, first of all, because of a breach of the conventions of order, conscience, civility, and personhood. But despite their conviction that they were dealing with a disease of the brain, and despite dissecting innumerable brains of deceased asylum inmates, nineteenth-century psychiatrists failed to ground the symptoms of madness in lesions or abnormalities within this deep corporeal interior. Instead, nineteenth-century doctors of the mad used their educated vision to make their diagnoses. Drawings and later photographs of the varieties of madness filled the textbooks, from Esquirol's Atlas of 1838 through Bucknill and Tuke's Manual of 1874 to Charcot's Iconographie, which was published from 1877 to 1918 (Gilman 1982; Rose 1998). This psychiatric gaze focused upon the surface of the body, on posture, gaze,

The Neurochemical Self and Its Anomalies 413 the colour of the skin of the melancholic, the gestures of the maniac, the movements of the hysteric. In the twentieth century, the eye lost its diagnostic power and gave way to the ear. For Kraepelin the distinctive biography, the aetiological pattern, diagnosed by the case history, provided the diagnostic material; for Freud, it was the voice of the patient, what was said, that was to be interpreted, and which permitted diagnosis. In each case, an interior psychological space opened up between the organs and conduct. Biography, environment and the like deposited their traces in that space. Motives, desires, and aspirations flowed from it. Understanding of pathology must locate its explanations in that same psychological space. The psyche became the 'obligatory passage point' and mental ill health seemed undeniably to be a psychological matter. Today, however, an augmented, prosthetic gaze seems to have opened up the corporeal basis of mental pathology to the gaze of psychiatry, and in so doing, to have supplied it with a new objectivity. The previously inaccessible bodily seat of the living mind now seems to be spread out before the psychiatric gaze. Brain imaging shows the differences in the regions that light up in the 'normal' and 'schizophrenic' brain, or fail to light up in the brain of a depressed person. Molecular neuroscience anatomizes the neurones, the receptors, the neurotransmitters, the cell membranes and ion channels, and allocate a specific pattern of functioning to each normal or abnormal mental state. Genetics seeks the gene sequences that correlate with each diagnostic category - to identify the precise functions of the variations they discover - a deficiency in the metabolism of a neurotransmitter, in the receptors, in a particular channel. Pharmaceutical companies fund most of this research: they compete to engineer and manufacture chemical correctives to molecular errors. In this process, psychiatry claims to have overcome, at last, the Cartesian dualism of body and soul. In its new 'neurochemical' account of personhood, psychiatry no longer distinguishes between organic and functional disorders. It no longer concerns itself with the mind or the psyche. Mind is simply what the body, what the brain, does. And mental pathology is simply the behavioural consequence of an identifiable, and potentially correctable, error or anomaly in some aspect of the brain, in its neurotransmitters or receptors. What is this new self? First and foremost, it is neurochemical. The neurochemical model of the brain started to take shape in the 1970s. It came to be believed that fundamental processes of the brain had to do with neurotransmission - the communication of signals between neurones - across the synapses or points of contact between nerves. Initially,

414 Nikolas Rose it had been thought that although nerves themselves transmitted signals by chemical means, transmission across the synapse was electrical. By the 1960s, largely as a result of work on the new psychiatric drugs - first the antipsychotics, such as chlorpromazine, then the antidepressants, such as imipramine and iproniazid (not forgetting the experiments on lysergic acid diethylamine) - it had been accepted that neurotransmission was carried out by chemicals. The first neurotransmitters identified were the monoamines (dopamine, norepinephrine, epinephrine, acetylcholine, and serotonin); later, certain amino acids were also found to be neurotransmitters (notably gamma aminobutyric acid or GABA), and by the start of the twenty-first century, the number of chemicals believed to be involved had reached the hundreds. These various discoveries were implicated in a rather radical reshaping of the technologies of psychiatric truth. This reshaping had a number of components. A whole experimental system was brought into being for the development of drugs and the testing of hypotheses, involving animal models (notably rats), in-vitro systems, and complex new types of apparatus, idealized objects on which scientific thought would now work. New entities were brought into existence - receptor sites, membrane potentials, ion channels, synaptic vesicles and their migration, docking and discharge, receptor regulation, receptor blockade, receptor binding - first as hypothetical, then as demonstrated in the lab, then as experimental entities in their own right, and finally as realities that can be rendered visible by PET scans or other visualization techniques. New truth technologies were invented, notably the randomized controlled trial. New forces became involved, not just teams of research scientists and a whole host of new specialist journals, but also large grant-giving agencies, and, of course, the pharmaceutical companies. These processes were not merely processes of discovery, but of intervention - the neurochemical brain becomes known in the very same process that invents interventions to manipulate its functioning - that is to say, therapeutic drugs. Psychiatric truth now became intrinsically linked with the maximization of what Catherine Waldby (2000: 19) has termed 'biovalue.' The imperative to produce a kind of surplus value from vitality and its problems drives the new symbiotic relation between knowledge, profit, and products. We have a new capitalization of the treatment capacities of medicine, as pharmaceutical companies are driven by the aim of maximizing shareholder value through the marketing of drugs with a mass, worldwide sale. We have a capitalization of truth itself, as vast sums of money, expensive equipment and computing power, and

The Neurochemical Self and Its Anomalies 415 huge research teams are required to produce biomedical truth. And we have a capitalization of the organs, tissues, and cells of the body, as the human genome has become the territory for a new goldrush for venture capitalists and patent lawyers. Those aspects of life that were previously devalued as pathology, whose humane treatment and welfare was a drain upon a national economy, are now vital opportunities for the creation of private profit and national economic growth. Paul Rabinow's (1996) assessment of the new life sciences is especially apt here: the quest for truth is no longer sufficient to mobilize the production of psychiatric knowledge. Health, or rather the profit to be made from promising health, has become the prime motive force in generating what counts for our knowledge of mental disorders. These developments were each to have their implications for the explanation and treatment of addiction. In the 1970s it began to be argued that addiction might be treated with the new psychotropic drugs that were now thought to act on these neurotransmitter systems - affecting the secretion of neurotransmitters, their uptake, or the receptor sites on the cell membranes to which they bound. Drugs such as lithium, which had been used to treat depression, were now being tried out on alcoholics, although the rationale was rather confused - perhaps lithium had a direct effect, but perhaps its effect might be mediated by the fact that alcoholics were actually depressed. Pharmacological treatments were, in any event, not seen as the primary therapeutic strategy. Rather, they were 'primarily pharmacological adjuncts and should be used only in conjunction with behavioural and psychosocial therapies' (Peachey and Naranjo 1984). Drugs like disulfram (Antabuse) might produce aversive responses, thus leading to deconditioning. Or the benzodiazapines might help, because they reduced the anxiety associated with withdrawal. Some speculated that the use of addictive drugs was itself a kind of selfmedication for a psychiatric disorder, in which case treating the underlying depression might lead to a reduction in alcohol use. Ten years later, however, a new style of thought had taken shape, in which the brain and its mechanisms became the primary target for explanations and treatments of addiction or dependence. Leonard (1992: 189) stated the position baldly: 'Most of the theories of dependence envisage a change in one or more neurotransmitter systems upon which the drug acts. Thus sedatives would be expected to facilitate GABAergic transmission and it may be hypothesised that these receptors are desensitised following chronic administration of these drugs. The withdrawal effects would then be postulated to result from a rebound hypersensitiv-

416 Nikolas Rose ity in receptor function.' Actually, alcohol remained a difficult case. It was in relation to other drugs of abuse that the neurochemical model of dependence first firmed up. Opiate addiction and opiate dependence were now allocated to the mu, delta, and kappa opiate receptors. Benzodiazapines and sedatives were allocated to the GABA-A receptor, where it was claimed that they mimicked the action of indigenous substances on this receptor, boosted chloride conductance through a chloride channel, and thus enhanced inhibitory neurotransmission (turning down the action of the neurones in question) and had anxiolytic effects. Tolerance was thought to arise because, after a while, the number of these receptors in the relevant brain sites decreases and/or the receptors become less sensitive. Either way, they become less effective in modulating neurotransmission in response to the drug. This not only leads to a need for more drug to produce the desired effect, the reduction in the numbers or functional sensitivity of GABA-A receptors also generates panic disorders when the drug is discontinued.3 Along similar lines, cocaine was thought to inhibit the dopamine reuptake system, hence leading to more active dopamine in the synapse. Chronic use was thought to inhibit the dopamine transporter even more, leading to an excess of dopamine in the brain, a constant dopamine high, paranoia, and so forth. The same was true, to a lesser degree, for amphetamines. PCP was thought to be an allosteric modulator of the N-methyle-D-aspartate (NMDA) subtype of the glutamate receptor (glutamate being another neurotransmitter), blocking the receptor and decreasing the flow of calcium into neurone, which is thought to produce sedation, decrease anxiety, and so forth. In this style of thought, nicotine works on the cholinergic receptors, first stimulating them to produce dopamine, then blocking them, preventing the normal stimulation of the nerves in the mesolimbic area that release dopamine hence the mini-rush, slow down, mini-rush at next puff! But in time the receptors respond to being turned off by becoming more sensitive, hence making it even more difficult to give up the drug. Nicorette patches work, it seems, because they release a small, controlled constant dose of nicotine, hence modifying this up and down mechanism and helping the smoker give up cigarettes. Our drug, naltrexone, was first developed to treat addiction to the opiates. It is here that we see its effect as a specific antidote to pleasure. It works by binding to the specific opiate receptors, because it is the same shape as the molecules that normally stimulate them. One needs to think here in terms of keys and locks. Binding of the 'therapeutic' drug

The Neurochemical Self and Its Anomalies 417 to these receptors blocks them, hence preventing the 'drug of abuse' from binding to them. And since the drug of abuse can no longer bind to the receptors, it can no longer cause pleasure. Of course, it takes some time for this message to get through to the user, whose learned habits endure for a while despite the absence of the neuronal rewards that had underpinned the original learning of them. This is why, in the treatment regime, the user must not be forced or persuaded or even allowed to come off the drug immediately. He or she must keep using the illegal drug while taking the anti-pleasure drug. In this way, the links between the use of heroin or whatever and the effects of pleasure will gradually weaken. The learned connection will be extinguished and, hopefully, the addict will no longer seek pleasure in this way. What, then, of alcohol? The books tell us that the pharmacology of alcohol is ill understood, and up until recently the argument was that its effects, though undoubtedly mediated through the neurotransmitters, were relatively non-specific. But it comes as no surprise that some continue to postulate the existence of an 'alcohol receptor.' In this case, the receptor of choice is the GABA-A receptor. Alcohol, it seems, may enhance inhibitory neurotransmission at the GABA-A receptor, by reducing transmission of chlorides through the little ion channel into the cell that the GABA-A receptor controls. This would enhance inhibition and reduce excitation of these receptors. Others think that alcohol may also work on dopamine receptors. Others think the serotonin system may be involved. In any event, molecular precision is required in explanations that are to be candidates for truth: they must work in terms of anomalies in receptors understood in terms of their specific shapes, in terms of the production of particular enzymes, or in terms of ion channels, for example (De Chaderevian and Kamminga 1998). Take this press release from the National Institute on Alcohol Abuse and Alcoholism. It reports the identification of:4 a new cell membrane channel where ethanol, the alcohol found in intoxicating beverages, may act. Neurobiologists from the Waggoner Center for Alcohol and Addiction Research and Section on Neurobiology, and the Department of Pharmacology and Toxicology, College of Pharmacy, University of Texas (UT) at Austin discovered actions of alcohol while studying a subtype of potassium channels, a diverse family of ion channels that perform many central nervous system functions. Identification of the alcohol-sensitive channel has significant implications because of its key role in brain function. 'Molecular analysis of this cell membrane channel ulti-

418 Nikolas Rose mately will increase our knowledge of how alcohol affects the brain and, thereby, the way a person functions,' said Enoch Gordis, M.D., Director, National Institute on Alcohol Abuse and Alcoholism (NIAAA), a component of the National Institutes of Health.

Our drug, naltrexone, seems to reduce the pleasure in drinking in the same way as it reduces the pleasure in opiates. It is known to bind to opiate receptors. Hence, it seems plausible to suggest that alcohol may also cause the release of endogenous opioids, and that naltrexone may block their effect. This would explain how the drug works for drink as well as for opiates. And, as in the treatment for opiate addiction, it will only work if the patient keeps drinking while taking naltrexone. The subjects themselves must find that the alcohol no longer produces pleasure or satisfies craving. Only then will the learned response linking alcohol to pleasure be extinguished (see chapter 15 this volume). Once more, what is involved here is a nice combination of neurochemical and behavioural selves. But this behavioural self is not inhabited by the deep psyche of the psychodynamic era. It is the flattened out psyche of thought and learning that is in play. A two-dimensional web of learned ways of thinking and acting now mediates between neurochemistry and conduct. In this web, aberrant conduct is learned through mechanisms that are, themselves, not aberrant - undesirable habits and thoughts, here, are the outcome of processes that are not pathological in themselves. What is learned may be unlearned, or may be replaced by a new and less damaging set of learned ways of thinking and acting, an array of competences and skills of life management that are more desirable to others, and indeed to the dependent person. From Normalization to Correction Why do I suggest that this new conception of pathology, in this case of alcohol dependence, represents a shift away from the idea of 'the alcoholic' as a deviant personality - from all those arguments about master statuses, degradation ceremonies, and moral careers? The initial diagnosis seems normative enough. The criteria for a diagnosis of alcohol dependence, as given in DSM-TV for example, only appear as such in relation to conventional cultural standards of personhood, conduct, and public order and reproduce all the well-known problems of such diagnoses. Alcohol dependence is located within the general class of substance abuse, whose criteria include failure to fulfil role expectations at

The Neurochemical Self and Its Anomalies 419 home or at work, neglect of conventional obligations, neglect of conventional criteria of safety and danger, and continued use despite the fact that such use causes or exacerbates recurrent social or interpersonal problems (American Psychiatric Association 1994). As in the conventional criticisms of biological reductionism, we seem to be presented with the familiar reading from the social to the vital; these 'maladaptive behaviours' are interpreted as the specific consequences of malfunctioning at the biological level, in this case mediated through excessive consumption of the substance in question. But despite Dr Gordis's reference to 'the way a person functions,' what we appear to have, at this biological level, seems to possess no moral weight. It is merely an error, a mistake in a bit of the machinery of the brain and its neurochemistry. And this error is potentially correctable. If we are dealing merely with a question of a chemical malfunction, we can, potentially, fabricate an artificial alternative, a drug. Once the neurochemical error has been corrected, the erroneous conduct can itself be remodelled in a desirable direction. It, too, can be seen as a correctable error of thought and learning. I think we are witnessing, at the very least, a mutation in 'the logic of the norm.' What do I mean by 'norm'? It is clear that the terms 'normal' and 'abnormal' are used in this new discourse, and I have used them many times in my discussion so far. I am thus not recommending an abolition of the terms as such, but I would recommend a change that relates to the specific sense in which they began to be used in the sciences and technologies of life in the middle of the nineteenth-century. As is well known, the French historian and philosopher of the life sciences, Georges Canguilhem, argued that this nineteenth-century sense of the term was problematic. He posed this problem in terms of a fundamental distinction between vital norms and social norms (Canguilhem 1978). He asserted that life was itself a vital and normative process. Vital norms arose from the very nature of living beings and the constant work of adaptation that they try to engage in to resist death. The vital norms of the living being, as they are used in medicine and biology, are not merely human judgments or statistical averages across a population: they arose from the specific character of their object. He writes, 'It is life itself, and not medical judgement that makes the biological normal a concept of value and not a concept of statistical reality' (ibid.: 43). Health, for the living organism, is not just normality but normativity, the vital force that resists disease. Pathology was a reduction of this vital normativity; illness was fixity, inability to adapt, stasis. Life was organic, systematic, self-regulating, hence the importance of disease for

420 Nikolas Rose knowledge: 'Disease reveals normal functions to us at the precise moment when it deprives us of their existence ... Health is organic innocence. It must be lost, like all innocence, so that knowledge may be possible ... the truly vital wonder is the anguish caused by disease' (ibid.: 52). Canguilhem thought that social norms were very different. Social norms manifest only adaptation to a particular artificial order of society. They arise from the requirements of those in power for normativity, docility, productivity, harmony, and the like. In a widely quoted passage he remarks, 'Between 1759, when the word "normal" appeared, and 1834 when the word "normalised" appeared, a normative class had won the power to identify - a beautiful example of ideological illusion - the function of social norms, whose content it determined, with the use that that class made of them' (ibid.: 151). The norms of the economy, in the school, in the legal system, and so forth thus arose from judgments of desirability made by groups with particular interests. The most intense problems arise where one set of norms are read in terms of the other, for example, when social norms were treated as if they were vital and arose from the nature of the human being rather than from a contingent social judgment of personhood. I think we can see what Canguilhem was trying to get at. But I do not think we can accept his distinction of vital and social. It is true that a new notion of the norm took shape within programs for the government of individual and collective existence in nineteenth-century Europe. These programs created standards of conduct - of dress, manners, punctuality, conduct, performance - that underpinned the norms that enabled judgments of deviation to be made in schools, factories, the army, hospitals, reformatories and processes of normalization to be set in motion. These norms were linked with those of medicine. Violations of norms of conduct were related to the idea of abnormalities in the healthy functioning of bodify processes. So far I seem to be agreeing with Canguilhem. Where I differ is in relation to the norms he terms vital. For these could only arise within specific apparatuses for the government of health. One such apparatus was the enclosed institutional space of the hospital. Large numbers of sick people were observed and their details were recorded, as they were citizens after all. Their symptoms were written down in case notes. Their prognoses were followed. And after they were dead, their corpses were opened up by dissection, and the organic seat of their sad decline was revealed to the medical gaze. Vital norms arose only when the individual

The Neurochemical Self and Its Anomalies 421 gaze of the doctor or the anatomist could be statisticalized in these 'machines for cure.' But vital norms were social too: as Foucault was fond of pointing out, medicine was the first 'social' science. The norms of longevity, morbidity, and reproduction arose in projects for the government of collective health and sickness. They were mapped out through statistics of birth and death, rates and types of morbidity. They emerged in practices ranging from sewage systems to insurance that regulated the population in the name of health. All these processes of biopolitics were conditions for the emergence of the notion of the vital normativity of the living human being and the human population. My point is that social and vital norms of individuals and populations have always been inextricably intertwined at the very heart of medical knowledge. It was in the universal and compulsory practices of schooling that the idea of 'normal development' in the child was formed and the techniques of weighing, measuring, and assessing were invented: they solidified the idea that there were biological norms of height, weight, and development and that deviations (slow development obesity) were biomedical abnormalities. Our norms of procreation, which radically changed the vital lives of women, emerged within domestication of sexuality and reproduction in the family. Our beliefs about which human ills are normal and which are treatable arose from the transformation of the home into a machine for creating and maintaining hygiene. Norms of the labouring body arose from the penetration of the gaze of welfare medicine into the workplace, which also gave us the apparently natural life course - the times when one was too young to labour, the times of labour, and the times of retirement. So perhaps we can say, in medicine as much as in pedagogy, that from the nineteenth century onwards, the idea of the norm involved an overlaying of social and vital ways of thinking. More precisely, the term 'normal' condensed a statistical, a social, a moral, and a medical judgment. The normal was that which was average, socially desirable, good or virtuous, and healthy. The norm combined or aligned the register of the statistical - the central point in the normal distribution that captured the regularities found in populations of numbers - and the register of the social and moral - the judgments of authorities about the desirability of certain types of conduct- and located these twin registers in a medical field of judgments of health and illness. Pathology was abnormality in all these senses. We are very familiar with this move in the case of intelligence. For the statisticians of intelligence, such as Karl Pearson, this was merely one exemplar of the fact

422

Nikolas Rose

that many, if not all, mental and moral qualities were normally distributed in the population, as were other conditions, such as tuberculosis, which manifested an inherited weakness of constitution.5 But it went much wider. Thus, as Ian Hacking has pointed out, at the turn of the twentieth century, many thought that the tendency to commit crime was both biologically based and normally distributed in the population. His example is Charles Goring's The English Criminal (1913), which tried and failed to confirm Lombroso's view that physical stigmata of criminals could be identified, but nonetheless argued that 'the criminal diathesis' was normally distributed in the population. As Hacking (2001: 149) puts it, 'The assumption of normalcy, the bell-shaped curve, the Gaussian distribution, is enormously powerful as a mathematical tool for analysing the tendency to criminal behavior.' But there is no reason to believe that these variations at the molecular level of base pairs within gene sequences will be normally distributed, or be seen to correlate, in the same way, with tendencies to socially desirable or undesirable conduct. It is in this sense that I suggest that, in this molecular gaze, we may be seeing the emergence of a new way of thinking: variation without a norm and perhaps, even, anomaly without abnormality. These, however, are not the only changes contributing to the waning of the logic of the norm. Across the nineteenth and twentieth centuries, ideas about normality arose out of zones gridded by government, that is to say, out of concerns with 'the conduct of conduct.' But, at the start of the twenty-first century, new forces are judging vitality in relation to different objectives. There are the commercial organizations such as the pharmaceutical companies, whose interests are the bottom line - profit. There are professionals ranging from doctors to research scientists, whose humanistic aspirations are linked to mundane concerns for personal and professional advancement. Insurance companies have their own criteria for benefits and so forth. Medical institutions make decisions as to who to treat, at what ages, for what conditions, under what criteria- for example, whether 'treatment' should be made available for infertility, or plastic surgery for children with Downs' syndrome - decisions that are simultaneously concerned with efficacy, equity, and economy. Philanthropic bodies, such as non-governmental organizations, charities, pressure groups, and campaigning organizations now play a key role in demanding rights to health and to treatment for all manner of conditions and persons. And, of course, the subjects and patients themselves now contribute to judgments as to what constitutes 'a suitable case for treatment.'

The Neurochemical Self and Its Anomalies 423 In this new configuration we have examinations galore, but they do not seem to operate in terms of hierarchical observation, distribution of individuals according to their qualities, and normalizing judgment against institutional expectations. There are plenty of charts of normal functioning, and graphs of distributions, but these do not seem to be standards to judge some people as failing and to open them up for control. These new ways of assessing and intervening upon persons seem to operate differently. Perhaps the implications will become clearer if I turn to my second main thesis, the transition from risk to susceptibility, in the context of the new molecular genetics. From Risk to Susceptibility

In the early days of the human genome project, when the term 'geneticization' came into fashion among critics, it was often thought that the sequencing of the human genome would establish a single 'normal' sequence, a composite or 'consensus genome.' It was suggested that this sequence would serve as a norm of health against which all discrepancies would be judged as morbid abnormalities (e.g., Flower and Heath 1993). A new form of molecular surveillance was often predicted, one which would categorize individuals as healthy or pathological on the basis of the sequences of bases on their genome and administer their lives in the light of this implacable biological truth. But the draft sequence of the human genome published on llth February does not produce a single 'normal' sequence. The first surprise was that there were far fewer genes than had been anticipated around 31,000 sequences coding for chains of amino acids as opposed to the 100,000 that had been predicted. This seemed to mark the defeat of that simple reductionism in which pathology is written in the genes in the form of a simple mutation, with one mutated gene equalling one disease. On the publication of the draft sequence of the genome, Craig Venter of Celera Genomics, who once hoped to be the legal owner of the human genome, pronounced, 'The notion that one gene equals one disease, or that one gene produces one protein, is flying out of the window.'6 The post-genomic project is regrouping around the genes that control genes, around the interactions between genes, around functional genomics, or around proteomics. Perhaps as important, the sequencing seemed to show that there were millions of loci on the genome where individuals differed from one another by as little as a single base in the chains of As, Cs, Gs, and Ts that

424 Nikolas Rose make up 'the genetic code' - an A is substituted by a C, for example.7 These variations that comprise such single nucleotide polymorphisms, or SNPs, do not seem to function along the axis of the normal and the pathological or the healthy and the sick. The two genomes (DNA sequences) that each of us carry differ from one another as well as from those carried by other individuals. Every sequence identified as a 'gene' now seems to be marked by such variation. In the human genome, to quote a recent article on pharmacogenomics, 'the normal is rare.'8 Perhaps one might say that there is no normal human genome - variation is the norm. This is a variation so complex and multidimensional that it cannot easily be mapped onto the earlier 'Mendelian' ideas of genes as units of inheritance, single entities which exist in a small number of alleles, some of which are normal, others pathological. We seem to have a geneticization of variability without the reciprocal positing of a norm. In the nineteenth century, arguments about the inheritance of pathology were all made along the same lines. An inherited predisposition was triggered by an exciting cause - loss of fortune, masturbation, and so forth. But in the contemporary style of biological thought, while vulnerabilities are inherited, and genome substitutes for constitution, there is not a single inherited diathesis, or a constitution that can be healthy or degenerate. We all carry genomic vulnerabilities to different conditions, but neither the vulnerabilities nor the conditions are molar, defects of the person. Instead, they are small, discrete, molecular. In the new style of biological thought, the vocabulary of genetic abnormality seems to mutate. There is still a logic of absolute genetic pathology. Disorders such as Huntington's arise inescapably from a specific identified set of repetitions in the base sequences in a particular gene region. But even that fatal logic is not a logic of fatalism (Novas and Rose 2000). In any event, such apparently implacable genetic pathologies are merely the extreme point of a rather different perception. One might call this 'risk,' but this would be to merge it into a family of related but distinct ways of thinking. So let us term it 'susceptibility.' Previously, the clinical assessment of risk in biomedicine was probabilistic. It might operate in terms of family history - on the basis of the pattern of illness among relatives, an individual could be told that he or she had a one in ten chance of developing a disease or giving birth to a child with a particular condition. Or it might operate by locating the individual in a risk table on the basis of measures that the epidemiology had linked to increased probabilities: age, gender, body mass index,

The Neurochemical Self and Its Anomalies 425 blood pressure, lipid levels, or whatever. An individual with a certain combination of these factors could then be told that his or her chance of a heart attack in the next ten years was less than 10 per cent. But the new molecular gaze promises something different, a precise appraisal of one's individual, specific susceptibilities on the basis of the sequencing of one's own DNA and the identification of a particular pattern of bases in particular locations on a specific chromosome. By the close of the twentieth century, an increasing number of medical classifications of illness were being designated with such apparent genetic precision. For example, one condition involving fronto-temporal dementia and Parkinsonism is known as FTDP-17 because it is linked to a number of mutations in a specific region of chromosome 17. Increased susceptibility to breast cancer has been linked to the mutations known as BRCA1 and BRCA2 on chromosome 13. Researchers claim that specific features of personality, such as novelty seeking, or psychiatric disorders, such as particular types of manic depression, arise from the synthesis or nonsynthesis of particular proteins, and that these are controlled by polymorphisms at specific loci on particular genes. The mapping of the human genome, and the identification of increasing numbers of SNPs will merely accelerate these processes. The genetics of alcohol dependency is now imagined in these terms. As we know, in the late nineteenth century and into the age of eugenics, alcoholism had a dual role in the inheritance of pathology. On the one hand, alcoholism was one among many manifestations of an inherited tainted constitution. Its passage down the generations could be visualized in a genealogical table whose blacked out squares and circles marking the affected men and women in the lineage - indicated the defective germ plasm. On the other hand, alcohol abuse had a deleterious effect on the germ plasm, and so was a factor that led those with already weakened constitutions to breakdown.9 Pedigree charts of alcoholism and its links with feeble-mindedness, sexual immorality, criminality, and the whole catalogue of manifestations of degeneracy were common in eugenic advice up to the outbreak of the Second World War. Such images of inherited weakness were instantiated in guidance to those contemplating marriage - who were well advised to check the family history of their prospective spouse for signs of these defects - and for those with such a history, who were to think very carefully of their social duty before procreating. In the post war years, all the usual techniques were deployed to estimate heritability. Notable are the Scandinavian studies of drinking behaviour in identical and fraternal twins and

426 Nikolas Rose adoptees, many based upon the very detailed census information and registers of alcohol abuse maintained in these countries. Most of these studies claimed that, even when adopted, male biological children of alcoholics had a much higher risk of becoming alcoholics - a risk about four times higher than that of biological children of non-alcoholics.10 From 1970s onwards, probabilistic attempts to estimate group risk were supplemented by studies that tried to identify markers - sites on specific chromosomes linked to the 'inheritance' of alcoholism - often based upon isolated 'inbreeding' populations. Blum and Noble (1990) claimed to have gone beyond the detection of markers and to have identified a specific gene linked to alcoholism, a variant of the gene for the D2 dopamine receptor. They later argued - despite other studies failing to replicate their findings - that this was a 'reward gene' that leads to reinforcement for a wide range of compulsive behaviours.11 Such 'gene for' explanations remain the stuff of popular journalism. For example, on 9 April 2001, Metro, the local London free paper, headlined 'Scientists seek alcoholism gene' and went on, 'A gene may be responsible for turning drinkers into alcoholics and British scientists are determined to identify it' (2). But behavioural genetics now takes a rather different line. Alcohol dependence is now conceived of as a matter of polygenetic susceptibilities, in which a range of variations (polymorphisms) on a number of different genes contribute, in different but specific ways, to an increased susceptibility - or sometimes a decreased susceptibility - to alcohol dependency. These are not 'genes for alcoholism,' but genes that control the synthesis of the proteins involved in the production and transportation of neurotransmitters, receptors, enzymes, cell membranes, or control of ion channels regulating neurones. As far as addiction is concerned, the belief that there is a genetic element has not weakened, but it is now posed in terms of risk, or 'vulnerability.' Linkage studies from the 1980s onwards suggested that so-called vulnerability genes might be found on chromosomes 1, 2, and 7, and a 'protective' locus had been found on chromosome 4. This protective locus was near the region of that gene that contained the alcohol dehydrogenase gene cluster, which includes genes encoding isozymes that accelerate the metabolism of ethanol to acetaldehyde. Interest was raised because this was thought to be the basis of the aversive 'flushing reaction' that prevents many Asians from regular or heavy drinking (Reich et al. 1998).12 Research goes on apace in genetically engineered mice to correlate coding sequence polymorphisms with

The Neurochemical Self and Its Anomalies 427 behavioural responses to alcoholism. Hood and Buck (2000) report that variation in the GABA sub A receptor gamma 2 subunit is associated with genetic susceptibility to ethanol-induced responses such as taste aversion and motor incoordination, which may be motivationally negative in relation to alcohol. Many who think that environment may have a role to play still retain this form of argument. Curling and Cook (2000) agree that much genetic research has tended to ignore psychosocial issues but point out that psychosocial research has similarly tended to ignore biological variables. They suggest that once specific genes involved in alcoholism have been isolated, new environmental factors that trigger susceptibilities will be identified. They predict that the comorbidity of alcoholism with anxiety, depression, and antisocial personality may soon be understood in the context of genetic effects from specific genetic susceptibility loci. Commenting on the implications of the publication of the map of the human genome, Nestler and Landsman (2001: 834-5) assert, 'Epidemiological studies indicate that 40-60 percent of an individual's risk for an addiction, whether it is to alcohol, opiates or cocaine, is genetic.' But we are not talking here about the inheritance of a single 'gene for' alcoholism. Rather, what is at stake are the genes that regulate different aspects of the neurotransmitter systems now thought to be involved. Thus Nestler and Landsman discuss the possible involvement of genes whose products regulate the sensitivity of a particular type of receptor, the G-protein coupled receptor, thought to be the initial target for drugs of abuse. Even with the current level of knowledge of the human genome, receptor desensitization potentially involves at least four gene families, each of which contains genes relating to a number of distinct subtypes of particular sorts of protein, such as kinases, arrestins, phosducins, and signalling regulators. Together, it seems, these constitute 'addiction vulnerability genes' - that is to say, susceptibilities. This kind of argument has implications for medical practice and treatment. Enthusiasts claim that genetic screening will soon be able to reveal individual susceptibilities by identifying these polymorphisms in particular locations on specific chromosomes. In the context of the development of cheap, automated methods of gene screening - and the likely increase of genetic screening - of prospective parents, of foetuses, of schoolchildren, of offenders, of psychiatric patients, of customers on a private basis, it is likely that awareness of one's precise inherited susceptibilities prior to any signs of a pathology will spread from the few to the many. Genetic susceptibility will be one of Ian Hacking's looping

428 Nikolas Rose kinds. The aspiration, of doctors and of the vulnerable or susceptible individuals themselves, will be to develop treatments that are precisely targeted at the specific abnormalities that do, or may, generate the disorder. Hence the promise of pharmacogenomics, in which medicine will be tailored to individual susceptibilities and pharmaceutical companies will add niche marketing to their immensely profitable catalogue of products for mass production. In this sense, then, the rise of the idea of susceptibility greatly extends the ambit of disease and the powers of the doctor to engage those who are neither phenomenologically nor experientially ill. Such persons may indeed live healthily for years, never become ill, or they may become incapacitated or die in a hundred other ways. Nonetheless, as everyone has their own susceptibilities, everyone becomes, potentially at least, asymptomatically ill and a suitable case for medical tutelage. Of course, this argument can lead to all manner of coercive practices of control of individuals on the basis of their anomalous biology. But before we rush too quickly to condemn geneticization, reductionism, determinisms, and the like, we need to pause. The idea of risk assessment conjures up images of certainty and calculability and hence, perhaps, a certain fatalism. But in the world of susceptibilities, biology is not destiny. The management of uncertainty is the norm. Molecular presymptomatic diagnosis gives no calculability to the 'when' or the 'how' of illness or death, nor to the manifold life decisions that must be made between the moment of diagnosis and the imagined end point. What forms of self-government arise in this new space of uncertainty? Some suggest that individuals and their families are left in what Robert Proctor (1995: 247) has termed 'enlightened impotence.' Others predict fatalism, anxiety, guilt, fear, stigmatization, and discrimination without hope of remedy (e.g., Nelkin and Tancredi 1989). These possibilities exist. But advocates of behavioural genomics think the implications are different. For we are not dealing with single genes and unitary pathologies, but with variations in multiple loci in multiple gene systems, each of which may itself be modulated by intra- and extra-cellular 'environmental' factors in the course of development and in the mature individual, resulting in a whole variety of different types and levels of susceptibility to particular disorders in specific environments (McGuffin, Riley, and Plomin 2001). In this field of multiple possibilities, pathways, interactions, and dependencies, the consequences of a genetic and neurochemical account of mental illness are not fatalism. On the contrary, most advocates claim that the diagnosis of genetic susceptibilities will

The Neurochemical Self and Its Anomalies 429 help individuals take responsibility for the management of their condition, for the control of their vulnerabilities, for the use of individually tailored behavioural and pharmacological correction of the underlying error or deficiency, thus enabling susceptible individuals to maintain themselves in their everyday life. Of course, there remains the problem of a new and refined form of genetic discrimination, in which insurance eligibility or employment, once linked at a rather gross level to family history of health and illness, will become molecularized. I have argued elsewhere that it is unlikely that the worst-case scenarios of the critics will come about (Novas and Rose 2000). As all diseases and pathologies come to be seen as having a genetic basis, and all such bases are in turn viewed as polygenetic matters of susceptibility, the binary distinction of 'genetic' and 'non-genetic' diseases will be impossible to maintain. As we all become reconfigured as biogenetic persons, to exclude all with susceptibilities for illness from insurance or employment would, indeed, be to exclude all. The more likely outcome is modulation rather than exclusion. This would be the requirement that, if we are to be employed or to receive insurance, we continually monitor our susceptibilities in the light of everything that might provoke them. And that we adjust our forms of life, activities, insurance cover, and financial planning more generally in the light of them. We would be obliged to sculpt our lives in terms of our own particular genetic and biological individuality. And maybe these will soon seem no more strange or reductionist than what have become common sense attributes of our somatic existence - height, weight, body shape, or tastes. We can already see signs of the new forms of activism arising in this complex field. That is to say, the emergence of new ways of crafting one's life and modulation of a lifestyle in the light of personal susceptibilities. Potential sufferers are to become skilled, prudent, and active, allies of the doctor, proto-professionals, and to assume their own share of the responsibility for managing their biogenetic selves. Those who are designated as at genetic risk and their families are not passive elements in the search for remedies. Such persons are increasingly demanding control over the practices linked to their own health, seeking multiple forms of expert and non-expert advice in devising their life strategies, and asking of medics that they act as the servants and not the masters of this process. Persons identified as susceptible to a particular condition have an investment in scientists fulfilling their promises and discovering the basis of, and the cure or treatment for, genetic conditions.

430 Nikolas Rose Contemporary biomedicine, like its predecessors, is one of the key sites for the fabrication of the contemporary self- free yet responsible, enterprising, prudent, encouraging the conduct of life in a calculative manner by acts of choice with an eye to the future and to increasing their well-being. Genetic personhood is one of 'genetic responsibility.' And genetic responsibility induces new forms of biological community, what Paul Rabinow (1996) has termed 'biosociality' (see also Novas and Rose 2000; Rose forthcoming). Consider, for example, the Genetic Alliance, which since the mid-1980s has sought to foster 'a dynamic coalition of consumers and professionals to promote the interests of children, adults and families living with genetic conditions,' bringing together almost three hundred support groups with consumers and health care professionals, 'creating partnership solutions to common concerns about access and availability of quality genetic services.'13 Those with similar genetic susceptibilities, or their families, gather into support groups, establish websites and email discussion lists, raise funds for medical research, engage their own medical experts, and develop and disseminate practices for crafting a life 'at risk' and a new ethics of the susceptible. This shift from implacable abnormalities to manageable susceptibilities is entirely consistent with the wider reshaping in practices for the government of persons. As is well known, Gilles Deleuze (1995) has suggested that contemporary societies are no longer disciplinary, in the sense identified by Foucault - they are societies of control. Where discipline sought to fabricate individuals whose capacities and forms of conduct were indelibly and permanently inscribed into the soul - in home, school, or factory- control is continuous and integral to all activities and practices of existence. We are required to be flexible, to be in continuous training, to participation in life-long learning and perpetual assessment; we are faced with continual incitement to buy and to improve ourselves, urged to engage in constant monitoring of our health and never-ending risk management. In these circuits, the active citizen must engage in a constant process of modulation, adjustment, and improvement in response to the changing requirements of the practices of his or her mode of everyday life. The ambition of these new technologies for governing the self is not merely to return the individual to a fixed norm of civilized conduct as housewife or worker, a one-off program of normalization. Rather, its ambition is the restoration and continual, long-term maintenance of the free, autonomous individual obliged to choose and to take responsibility for his or her life as if it were an outcome of acts of choice. But further,

The Neurochemical Self and Its Anomalies 431 biological psychiatry seems to be offering the possibility of the calculated modification and augmentation of specific aspects of selfhood. In its new neurochemical and psycho-pharmacological guise, it is contributing to the idea of the flexible, manipulable self- manipulable not only in the service of projects of normalization, but manipulable by the person himor herself in the service of enhancement of capacities. These new ways of thinking about genetic and neurochemical selves are thus deeply implicated in the continual process of modulation of capacities that has become the life's work of each active citizen. Desire and the Neurochemical Self

How do these arguments relate to the general question with which I began, the pharmaceutical government of desire? First, I have suggested that the government of addiction has become 'targeted.' Alcoholics Anonymous, as Mariana Valverde (1998; see also chapter 15) argues, works on identity. It is the whole person who is pathological. 'I am Nikolas and I am an alcoholic.' While in many respects we are still living in the age of 'the addict' and 'the alcoholic' in which these features are inscribed within the truth of a 'deviant personality' - we are also beginning to see a more precise and specific form of address. The goal is not to reshape a life or normalize a personality, but instead to isolate a malfunctioning process, and a related set of problematic beliefs, cognitions, and life skills, and to engineer interventions that will address this very specific pathological complex with a minimum of collateral damage. The aim is to enable the individual to reenter the circuits of everyday life, where he or she will re-engage with the cybernetics of control built into education, employment, consumption, and leisure. While many professionals still think of their clients as having damaged, defective, or dangerous identities, I think we are beginning to see signs of a shift from strategies of normalization of the deviant to measures seeking the correction of specific anomalies: interventions which work not by eliminating autonomy, freedom, and choice in individuals so affected, but by locating them in a regime requiring a constant process of adjustment. Of course, the results of such attempts at targeting often prove surprising. Prozac is the most obvious example. On the one hand, this carefully sculpted molecule designed to target a specific condition is now offered as the treatment not merely for mild to moderate depression but for a whole range of other disorders: anorexia, bulimia, obsessive compulsive

432 Nikolas Rose disorder, panic disorder, and many more. Indeed, now that Prozac has gone out of patent, Eli Lilly, the manufacturers, are marketing the same drug - fluoxetine hydrochloride - in a different package for premenstrual dysphoric disorder, a candidate diagnosis in DSM-TV apparently marked by severe mood and bodily disorders around the time of menstruation and affecting 3-5 per cent of menstruating women in the United States. While this development appears to run counter to the molecularization of diagnoses and treatments, or, perhaps, to question the project of DSM-TV for the ever-finer distinction of types of disorder, other, newer drugs claim to target particular receptors even more precisely - the 5HT2 receptor, for example - thus once more claiming to redivide psychiatric disorders in molecular terms. At the same time, one molecule and one target seems insufficient. In relation to depression, for instance, we now have a range of third generation products that act on serotonin and norepinephrine, or serotonin and dopamine. Biovalue, here, seems to demand constant innovation, and the cycling from the specific targeted cure - the magic bullet - to the wonder drug that will cure all seems endemic to marketing and perhaps to the very project of commercial psychopharmacology itself. Yet, despite this cycling, in any of these phases, it is not the pathological individual that is targeted but a molecular anomaly. Second, I have argued that addiction itself is being rewritten in terms of the susceptible neurochemical and genetic self. Having resisted medicalization for so long, addiction becomes, at least in part, not a disease of the person, or of the will, but a disease of the brain. Of course, as Mariana Valverde has shown, this short-circuiting of the will is only partial. For in the new 'flattened' self, in which the organs and bodily processes are in ever closer proximity to thought, emotion, cognition, and action, a role still remains for techniques for the reconstruction of the will. Cognitive and behavioural therapies have a new role: to equip the vulnerable individual with the insight, the skills, and the selfmastering capacities required to live a life in the knowledge of his or her susceptibilities and to take responsibility for avoiding the actions, settings, or relations that might provoke them. Finally, I would like to return to biovalue, and its relation to this process of modulation of the somatic self. The best-selling drugs these days are not those that treat acute illnesses, but those that are prescribed chronically - Premarin, Lipitor, Omeprazole, Prozac, Viagra. The power to reshape life by drugs seems to extend way beyond what we previously understood as illness. Biomedicine has already rewritten the norms of

The Neurochemical Self and Its Anomalies 433 reproduction, its timetables and its kinship relations. Hormone replacement treatment is already rewriting the norms of female ageing. Drugs for 'panic disorder,' such as Alazopram, are rewriting the norms of social interaction. The capitalization of the power to treat intensifies the redefinition of that which is amenable to correction or modification. This process is not simply blurring the borders between normality and pathology, or widening the net of pathology. We are seeing an enhancement in our capacities to adjust and readjust our somatic existence according to the exigencies of the life to which we aspire. These capacities for enhancement involve the co-production of a drug, the condition it will treat, and the desire and demand for it. Where Foucault analysed biopolitics, we now must analyse bioeconomics, for human capital is now to be understood in a rather literal sense, in terms of the new linkages between the politics and economics of life itself. Notes 1 This paper was presented at the Conference on Risk and Morality held at the University of British Columbia in May 2001. It was conceived as a companion piece to that given in the same session by Mariana Valverde and is indebted to her comments and suggestions. I have reframed it here to stand on its own and I am to blame for any mistakes. An earlier, rather different, version of this paper was presented as 'Normality and Pathology in a Biological Age,' given as a public lecture in the Faculty of Humanities, University of Copenhagen, 9 March 2001. 2 I discuss the 'opening up' of this psychological space in Rose (1985). 3 Actually, as Leonard (1992) points out, there is considerable controversy over these speculations. The benzodiazapine complex shows these 'plasticity effects' in animal brains, but there is little evidence of changes in human brains in post-mortem studies, and the experimental changes, in any event, require much higher doses than those that produce functional tolerance when used clinically. What is important for our purposes is that while mechanisms and specifics are disputed, the thought style remains unchallenged and defines the form of possible explanations and criticisms of those explanations. 4 http://www.niaaa.nih.gov/press/1999/discover-text.htm. Lewohl, et al. (1999) and her associates identified the effect of ethanol on the G-proteincoupled inwardly rectifying potassium channel (GIRK). GIRKs are widely distributed in the brain and play a major role in regulating inhibitory

434 Nikolas Rose responses in the central nervous system, and GIRK channels were shown to be sensitive to ethanol at a concentration below the legal level for intoxication (lOmM). They suggested that this high sensitivity to relatively minor changes may influence substantially the capability of neurons to communicate. 5 This is one focus of my 1985 study, The Psychological Complex (esp. chaps. 3-6): 'Individual psychology derives its conception of its object from the statistical normativity of the population. The norms which it proposes are not those of life but those of large numbers. The possibility of a knowledge of the individual, for the psychology of the individual, is not provided through a conception of the psyche, its processes, its homeostatic mechanisms, the laws of its development and the abnormalities to which they can give rise. It is founded upon a metaphysic of the quantification of qualities and the laws of variation in populations ... It was founded through the identification of norm in the register of the statistical with norm in the register of the social. The operation which made individual psychology possible was the identification of statistical norms of variation with social norms of expectation. The abnormality which was so crucial in the founding of a medical notion of bodily norms was a disturbance in its object, the body itself. But the abnormality around which individual psychology was organised was not an abnormality of a life process, or one specifiable in terms of ease and dis-ease. It was an abnormality in terms of a norm of functioning specified by particular social apparatuses. The unease which enabled the normativity of individual psychology to be established was constituted by the objectives of government rather than the vicissitudes of the psyche. It was the school, the courts, the police and the army which provided the psychology of the individual with those whom it would have to be able to construe as abnormal' (229). 6 Quoted in Yamey (2001). 7 The 'public' genome identifies some 1.42 million single nucleotide polymorphisms. Craig Venter's group identify 2.1 million SNPs, very few of which, less than 1 per cent, seem to result in variations of the proteins coded. 8 Tn Pharmacogenetics, Wendell W. Weber (1997) quotes from Somerset Maugham's account of his experiences as a young medical student... "I have always worked from the living model. I remember that once in the dissecting room when I was going over my 'part' with the demonstrator, he asked me what some nerve was and I did not know. He told me; whereupon I remonstrated, for it was in the wrong place. Nevertheless he insisted that it was the nerve I had been looking in vain for. I complained of the abnormal-

The Neurochemical Self and Its Anomalies 435

9 10 11 12 13

ity and he, smiling, said that in anatomy it was the normal that was uncommon. I was annoyed at the time, but the remark sank into my mind and since then it has become forced upon me that it was true of man as well as anatomy. The normal is what you find but rarely. The normal is the ideal. It is a picture that one fabricates of the average characteristics of men, and to find them all in a single man is hardly to be expected." Maugham's observation - that the normal is rare - is at the heart of the challenge and promise of pharmacogenomics' (Norton 2001: 180). Thanks to Oonagh Corrigan for this quotation. For a discussion of this theme in Nazi eugenics, see Proctor (1988: 239-40). See the review of these studies in Grove and Cadoret (1983). These claims and the contrary findings, are reviewed in Holden (1994). Quoted from the NIH News Release of 20.5.98, http://www.niaaa.nih.gov/ press/1998/cogal-text.htm. http://www.geneticalliance.org/allianceinfo.html.

References American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. New York: American Psychiatric Association Blum, K., and E. Noble. 1990. 'Allelic Association of Human Dopamine D2 Receptor Gene in Alcoholism.' Journal of the American Medical Association (April): 2055-60 Canguilhem, G. 1978. On the Normal and the Pathological. Dordrecht: Reidel De Chaderevian, S., and H. Kamminga, eds. 1998. Modularizing Biology and Medicine: New Practices and Alliances, 1910s-1970s. Amsterdam: Harwood Deleuze, G. 1995. 'Postscript on Control Societies.' Negotiations. New York: Columbia University Press Feeley, M., and J. Simon. 1992. 'The New Penology: Notes on the Emerging Strategy of Corrections and its Implications.' Criminology 30: 449-74 Fleck, L. [1935] 1979. Genesis and Development of a Scientific Fact. Trans. F. Bradley and T. Trenn. Chicago: Chicago University Press Flower, M., and D. Heath. 1993. 'Micro-anatamo Politics: Mapping the Human Genome Project.' Culture, Medicine and Psychiatry 17: 27—41 Gilman, S. 1982. Seeing the Insane. New York: Wiley Grove, W.M., and RJ. Cadoret. 1983. 'Genetic Factors in Alcoholism.' In B. Kissin and H. Begleiter, eds., The Pathogenesis of Alcoholism: Biological Factors. New York: Plenum Press Curling, H.M.D., and C.C.H. Cook. 2000. 'Genetic Predisposition to Alcohol Dependence.' Current Opinion in Psychiatry 12: 269-75

436 Nikolas Rose Hacking, I. 2001 a. 'Degeneracy, Criminal Behavior, and Looping.' In D.Wasserman and R. Wachbroit, eds., Genetics and Criminal Behavior. Cambridge: Cambridge University Press - 2002. 'Inaugural Lecture for the Chair of Philosophy at the College de France.' Economy and Society 31 (1): 1-14 Holden, C. 1994. 'A Cautionary Genetic Tale: the Sobering Story of D2'. Science 246: 1696-7 Hood, H.M., and K. Buck. 2000. 'Allelic variation in the GABA sub A receptor gamma2 subunit is associated with genetic susceptibility to ethanol-induced motor incoordination and hypothermia, conditioned taste aversion, and withdrawal in BXD/Ty recombinant inbred mice.' Alcoholism: Clinical and Experimental Research 24: 1327-34 Kramer, P. 1994. Listening to Prozac. London: Fourth Estate Leonard, B.E. 1992. Fundamentals ofPsychopharmacology. Chichester: Wiley Lewohl, J., W. Wilson, R. Mayfield, S. Brzowski, R. Morrisett, and R. Harris. 1999. 'G-protein-coupled Inwardly Rectifying Potassium Channels are Targets of Alcohol Action.' Nature Neuroscience 2 (12): 1084-90 McGuffin, P.B. Riley, and R. Plomin, 2001. Toward Behavioral Genomics.' Science 291: 1232-49 Nelkin, D., and L. Tancredi. 1989. Dangerous Diagnostics: The Social Power of Biological Information. New York: Basic Books Nestler, E.J., and D. Landsman. 2001. 'Learning about Addiction from the Genome.' Nature 409: 834-5 Norton, R.M. 2001. 'Clinical Pharmacogenomics: Applications in Pharmaceutical R&D.' Drug Discovery Today 6: 180-4 Novas, C., and N. Rose. 2000. 'Genetic Risk and the Birth of the Somatic Individual.' Economy and Society 29: 485-513 Peachey, J.E., and C.A. Naranjo. 1984. 'The Role of Drugs in the Treatment of Alcoholism.' Drags 27: 171-82 Proctor, R. 1988. Racial Hygiene. Cambridge: Harvard University Press - 1995. Cancer Wars. New York: Basic Books Rabinow, P. 1996. 'Severing the Ties: Fragmentation and Dignity in Late Modernity.' In Essays on the Anthropology of Reason. Princeton: Princeton University Press Reich, T. et al. 1998. 'Genome Wide Search for Genes Affecting the Risk for Alcohol Dependence.' Neuropsychiatric Genetics (American Journal of Medical Genetics) 81: 207-15 Rose, N. 1985. The Psychological Complex. London: Routledge - 1998. 'The Psychiatric Gaze.' Unpublished paper first given to the London History of the Present Research Network

The Neurochemical Self and Its Anomalies 437 - 2000a. 'Government and Control.' In D. Garland and R. Sparks, eds., Criminology and Social Theory, 183-208. Oxford: Oxford University Press - 2000b. 'Biological Psychiatry as a Style of Thought.' Paper to a symposium on models and cases in scientific thought, at Princeton University - Forthcoming. 'The politics of life itself.' Theory, Culture and Society Valverde, M. 1998. Diseases of the Will Cambridge: Cambridge University Press Waldby, C. 2000. The Visible Human Project. London: Routledge Weber, W. 1997. Pharmacogenetics. Oxford: Oxford University Press Yamey, G. 2001. 'Dispute as Rival Groups Publish Details of Human Genome.' British Medical Journal 332: 381

15

Targeted Governance and the Problem of Desire MARIANA VALVERDE

Like the 'smart bombs' used by the U.S. Air Force, smart drugs are supposed to act selectively on particular targets. The appeal of smart weapons is obvious: if targets can indeed be isolated and acted upon, authorities will be able to unleash powerful new devices upon political or bodily 'hot spots' while minimizing the collateral damage and attendant bad publicity that have marred the forward march of technical progress for both military and pharmaceutical interests. In the sphere of the social, targeted governance has also become extremely popular. Targeted policing, for example, is all the rage in North America. The targeting can be site-specific, namely, focusing on 'flashpoints' to the detriment of regular foot patrols throughout the city. Or it can be risk-factor specific, such as the profiling carried out by immigration officers. Either way, there seems to be a sense that the city or the nation as a whole cannot or should not be policed: instead, there should be a rational selection of high-risk spaces, high-risk people, or risk factors. In social services, targeting has also become popular since large-scale efforts to re-engineer 'society' through universal programs became discredited. The term 'social planning,' once a beacon of Utopian hopes, sounds anachronistic, quaint, frankly embarrassing - on the rare occasions on which it is still used. Instead, we now have a plurality of uncoordinated 'targeted social programs.' In the world of Canadian social service provision, for example, there is an obsession with a group formerly known as children and now designated by the bureaucratic label '0 to 6.' In the international arena, the World Bank is pressuring many countries to adopt a targeting strategy of social service provision, under

Targeted Governance and the Problem of Desire 439 the banner of helping 'the poorest of the poor.' Developing as well as developed nations are thus rushing into the neo-liberal world of targeted governance - the world of decentralized, networked knowledge/ power, a world governed by benchmarks, audits (Power 1994), and performance indicators (Morrison 2000). The popularity of strategies of targeted governance is closely linked to what Nikolas Rose describes as the waning of 'the era of the normal and the pathological' (Rose 2001). As against journalistic claims about the wholesale reform and normalization of humans enabled by genetic science, he points out that scientists themselves now avoid claims about normality. They prefer to talk about specific biochemical and genetic processes. The development of selective serotonin reuptake inhibitors (SSRIs) such as Prozac is here exemplary: instead of making overall judgments about a person, neuroscience takes the more modest route of isolating a process — serotonin reuptake in receptor proteins present in a particular set of neurons - and trying to develop a specific therapy that will not cause the kind of collateral damage that has given psychopharmaceuticals such a bad name among consumers. Nikolas Rose calls the new biomedical logic a logic of'correction' and 'constant adjustment.' In this chapter I will explore the logic of targeted governance, not in abstract theoretical terms but, as befits the topic, in relation to a particular target or object: alcoholism treatments. I will describe how the longstanding quest for a pharmaceutical magic bullet for alcoholism has recendy resulted in a cautious but palpable euphoria about drugs originally developed for purposes of opiate addiction treatment which are now thought to have a targeted effect upon excessive or uncontrolled drinking as well: opiate receptor antagonists. Although there is evidence that nalmefene is the most 'targeted,' least toxic of this group of drugs (Mason et al. 1999), naltrexone has been most widely used and is the only drug whose application to alcoholism has been patented, as well as approved by the U.S. and other governments. The product announcement by DuPont, which patented naltrexone under the brand name of ReVia, describes the process: 'How does ReVia work? It is believed that alcohol causes the release of endogenous opioids. The binding of these opioids to the [opiate] receptors in the brain may be responsible for the positive reinforcing effects of alcohol. ReVia, an opioid receptor antagonist, competitively binds to these receptors, blocking the endogenous opioids at these sites' (cited in Sinclair 2001: 2).

440

Mariana Valverde

This product announcement avoids using the stigmatizing psychiatric term 'dependence,' leaving open the question of whether drinking ought to be classified as an addiction or simply a pleasure that has positive reinforcing effects. Some medical studies do claim that naltrexone acts upon dependence itself (the term that has replaced 'addiction' in medical discourse);1 but, in keeping with the logic of targeted governance, practitioners and scientists can agree that naltrexone works to help people drink less, even if they disagree on the larger issues of ontology and causation. Naltrexone is the first drug approved for alcoholism treatments by the U.S. government since the approval fifty years ago of the now discredited, harshly disciplinary drug disulfiram (better known under its brand name of Antabuse). Naltrexone, prescribed some twenty years ago to help addicts stop using heroine,2 is being hailed as that Holy Grail of alcohol studies, an anti-alcoholism 'targeted medication' (Sinclair 2001; Heinala et al. 2001). While minimizing excessive drinking, this drug, unlike Antabuse, does not demand abstinence. Its target is not drinking per se; by contrast with Alcholics Anonymous (AA) and other abstinence-based programs, the target of naltrexone is not any and all bottles, but rather that venerable social and psychic problem associated with excess, immoderate, uncivilized drinking: alcoholism. The American National Institute for Alcohol Abuse and Alcoholism, which provides 90 per cent of alcoholism treatment research funding in North America, announced the approval of naltrexone in the following words: 'Unlike disulfiram, naltrexone and other potential agents now under NIAAA investigation [nalmefene], directly target hallmark features of alcoholism: abnormal alcohol-seeking behaviour, impaired control over alcohol intake, and physiological dependence manifest in craving when alcohol is removed' (NIAAA 1995, my emphasis). Naltrexone does not work by diminishing the effects of physical withdrawal (benzodiazepines are generally used for this purpose, as they have been for decades). Neither does it create a substitute, more socially acceptable addiction, as methadone does for heroin. Some say that it works by diminishing the physiological reward created by endorphins released by alcohol, thus eventually 'extinguishing' the conditioned response of continuing to drink (Sinclair 1997, 2001); others simply state that it diminishes both the actual drinking and the 'craving' (O'Malley et al. 1992), without any theorizing. Either way, naltrexone, we are told, acts invisibly and automatically to readjust the level of craving. Is naltrexone the neo-liberal magic bullet of the post-normal era? It

Targeted Governance and the Problem of Desire 441 certainly looks like it. The drug targets one kind of receptor in the brain rather than acting systemically. Instead of acting on the person as a whole, on the soul, the body, and the will, naltrexone acts upon one specific invisible process. This cause-effect relationship, while tightly bound to the Pavlovian 'extinction of conditioned response' paradigm by behaviourists, is also compatible with the more rationalist perspective favoured by cognitive behavioural psychology, a much more powerful paradigm in North American treatment circles than behaviourism. Whether the drug will be accepted by those who favour treatment approaches derived from AA is yet to be seen, however. Dr Pearlman, a Boston physician typical of the hybridity of North American treatment approaches in that he has personal experience with AA and runs a program combining cognitive behavioural therapy and naltrexone, counsels alcoholics to become abstinent if they can do so, but sees this drug as very promising for the large number of alcoholics who will not stay in AA or otherwise maintain total sobriety (personal communication, 2 April 2001). But AA members are unlikely to support the introduction of this drug on a more than experimental basis. The drug allows alcoholics to continue to drink, thus breaching AA's belief that alcoholics are a distinct population with a distinct identity whose main feature is the inability to drink at all. Alcoholics Anonymous, like naltrexone, targets alcoholism, but it does so by governing through persons — alcoholics have to self-identify as such and act accordingly. To go to AA one has to believe, however pragmatically, in the existence of alcoholics. To go to a naltrexone clinic may in some cases (in the United States, in particular) require a temporary allegiance to the concept of 'alcoholism,' but alcoholism can be refigured as a limited imbalance of brain chemicals. Using naltrexone as a site upon which to reflect on the appeal - and contradictions — of targeted governance may thus also shed some light on the hypothesis about the waning of strategies that seek to normalize whole persons. Technologies of normalization and pathologization usually deploy, and reproduce, 'persons' and 'disease': if the logic of normalization is giving way to the logic of 'adjustment' and 'correction,' as Nikolas Rose suggests in chapter 14 (this volume), this transition may be linked to the waning of identity-based governance more generally.3 The targeted governance of alcoholism/excessive consumption by means of a smart drug is still in its infancy, but there is sufficient evidence to suggest that like many other Utopias of governance, targeting alcoholism with naltrexone pills may be more satisfying and coherent as a dream than as an auditable reality. And it may well be that the

442 Mariana Valverde contradictions and failures already visible in the biochemical targeting of alcoholism may be linked to more general problems and contradictions in the post-social project of targeted governance. The Alcoholic Brain as a Moving Target

In a historical sociology of alcoholism treatments (Valverde 1998), I found that alcoholism is an unusual condition. First, it has never been successfully medicalized. Second, projects seeking to medicalize alcoholism have never effected the kind of reductionist biologization that has flourished in otherwise similar fields, such as mood disorders. In other words, alcoholism has been a hybrid condition, partly moral and partly psychological and physiological; and even those biomedical authorities who have at various points tried to read it as primarily physiological have shied away from reducing it to any one process or site - in today's paradigm, a particular deficiency or excess of certain brain chemicals. The American National Institute on Alcohol Abuse and Alcoholism has gone a fair way down the biological road in recent years. This may be due to the saturation of the public sphere with messages about illegal drugs, a process that, since the recent redefinition of alcohol as a drug, has had much influence on the way that alcohol is discussed, in health circles at any rate. The drug-alcohol relation is now the reverse of what it once was: while in the late nineteenth century the use of opiates was thought of as a species of 'inebriety,' nowadays drinking is subordinate to the paradigm of drug addiction. In a newsletter on pharmaceutical innovations in alcohol research, the NIAAA has a section entitled 'Alcohol and the Brain,' under which we read the following: 'All brain functions, including addiction, involve communication among nerve cells (neurons) in the brain. Each of the brain's neurons connects with hundreds or thousands of adjacent neurons ... Determining the specific neurotransmitters and receptor subtypes that may be involved in the development and effects of alcoholism is the first step in developing medications to treat alcoholism (NIAAA 1996). In keeping with the 'drugs and the brain' approach, then, alcoholism has come to be read more biologically than in the past in many definitions and programmatic statements. But whereas both legal drugs such as SSRIs and illegal drugs like cocaine and Ecstasy prompt drug companies and prevention professionals to produce and disseminate appealing, brightly coloured pictures of brains and neurons (see for example www.clubdrugs.org), alcohol research and alcohol education are not in

Figure 1 Dr David Sinclair's behaviourist model of 'craving' is used in the network of private ContrAl Clinics that began in Finland and popularized naltrexone use in many countries.

444 Mariana Valverde

Figure 2 The Centre for Addiction and Mental Health is Canada's premier alcohol research centre, but its educational material on alcohol is remarkably unscientific in both form and content.

fact as brain-centred as the 'alcohol and the brain' statement might suggest. The pictures illustrating the educational posters and leaflets on alcohol distributed by Canada's foremost authority on the subject, the Centre for Addiction and Mental Health in Toronto (formerly the Alcoholism Research Foundation of Ontario), show people - not brains, and not neurons - in various stages of disrepair. Falling into gutters, getting ulcers, losing jobs, becoming dishevelled, these dysfunctional individuals give the impression that alcohol affects one's whole life, and that drinking is not a matter of a single brain chemical but a question of the self. In keeping with the long tradition, pioneered by Alcoholics Anonymous (Valverde 1998: chap. 5), of regarding alcoholism as a complex hybrid entity that cannot be solved by pills or other purely biological means, one of the key promoters of naltrexone treatment in the United States, Joseph Volpicelli, openly admits that the quest for a target for

Targeted Governance and the Problem of Desire 445 alcoholism treatments might be a wild goose chase: 'In contrast to opiates, alcohol does not seem to have specific agonist effects in stimulating opiate receptors directly. Rather, it is believed that alcohol has nonspecific effects on the lipid components of cellular membranes' (Berg et al. 1990: 138). Other, less invested researchers have commented that the quest for a single biochemical process that could be targeted in order to moderate excessive drinking has been motivated by nothing but 'the current Zeitgeist': The search for the neurobiological substrates of alcohol's effects on behaviour in general and its positive reinforcing effects in particular now span several decades ... The goal of each of these lines of research has been to determine the involvement of a particular neurotransmitter in regulating this behaviour ... Over the past two decades, a body of data emerged which clearly implicated several neural systems in the mediation of ethanol intake in different ways and different levels of specificity ... It is unlikely that any one single neuronal system would be identified as the primary-unitary system underlying ethanol's effects on behaviour in general and ethanol intake in particular. What is far more likely is the emergence of a body of data pointing at a complex multi-system interaction in the mediation of ethanol's behavioural effects. (Amit and Smith 1990: 161, 162, 172)

Despite repeated suggestions that excessive drinking does not have a single biochemical site, the tremendous cultural capital acquired by genetics and neuroscience in recent years cannot but exercise an influence on the rather low-status and poorly funded pursuit of alcoholism treatment research. The quest for a pharmaceutical magic bullet for alcoholism is thus overdetermined, and may not be fatally affected by the failure to 'discover' a specific biochemical process that can be isolated and manipulated. In the 1970s, some researchers thought that lithium could be effective not only for the significant number of alcoholics who suffered from bipolar disorder but also for 'regular' alcoholics. One such researcher was the Dr Sinclair who has now developed a proprietary naltrexone method, ContrAl (Sinclair 1987). In the 1980s, alcoholism researchers again attempted to find new uses for psychotropic drugs being developed for other conditions, mainly serotonin uptake inhibitors (SUIs). It was these 1980s experiments and trials with SUIs that first broached the possibility that people might be able to continue drinking (thus avoiding the well-known problems of relapse, guilt-induced binging, and so forth)

446 Mariana Valverde but without getting the pleasurable reinforcement thought to be the immediate, psychobiological cause of excessive drinking. A review of SUI-alcohol clinical studies concluded that some decrease in drinking was observed among subjects on SUIs. The Toronto authors who were leaders in this particular field concluded that SIUs 'may act by decreasing desire to drink and the reinforcing properties of alcohol' (Naranjo and Bremner, 1990: 105, see also Naranjo, Bremner, and Poulos 1990). That the effect of anti-addiction drugs on desire may have been broader than would be regarded as desirable by most patients, if not most clinicians, was only hinted at in these studies: Naranjo and Bremner reported some weight loss among their subjects even in very short (2 to 4 weeks) trials, but did not highlight this side effect. Other, less invested researchers concluded, however, that SUIs produce a 'potent anorexic effect' (Amit and Smith 1990: 167) that is anything but specific. This may have contributed to the fact that no SUI was (to my knowledge) approved by governments for use as an anti-alcoholism treatment. Similarly, an interesting finding that was not followed up, but which would surface again in the naltrexone trials, was that SUIs were reported 'to have an effect on taste factors independent of the caloric value of the food' (Amit and Smith 1990: 168). In other words, SUIs might work to decrease alcohol consumption only because they make both food and drink taste funny. Other avenues for biochemically readjusting either the intake of alcohol or the desire to drink or both were explored during the 1980s and early 1990s. Generally, it was found that many psychopharmaceuticals do have a measurable effect on drinking, and studies that attempted to quantify desire rather than (or as well as) measure drinking behaviour also concluded that levels of what some social psychologists call 'craving' could be chemically manipulated to some extent. Chemical Change or Psychological Change?

In various labs around the world, animal and human studies are being conducted to discover which particular neurotransmitters are involved in the behaviour of continuing to drink past the point of intoxication. But it is interesting that even those alcoholism researchers who spent many years studying rats on drugs (e.g., Sinclair, Volpicelli) do not want to question the old consensus about the need for psychotherapy. Volpicelli, a professor at the University of Pennsylvania, now heads up a treatment program using naltrexone that is curiously entitled 'Assisted Recovery.'

Targeted Governance and the Problem of Desire 447 This label incorporates naltrexone to the AA tradition, specifically, AA's belief that alcoholics can only be in recovery with the assistance of an entity that compensates for their deficiency. In the clinics proliferating under the name of 'Pennsylvania model' clinics, naltrexone is marketed as a chemical substitute - or a supplement - for AA's Higher Power (see www.recovery2000.com). In the chain of clinics owned and operated by ContrAl in Finland and other countries, and also in some sites in North America, the term 'recovery' is not mentioned. Along more rationalist lines, cognitive behavioural psychologists devise programs designed to help heavy drinkers acquire 'coping skills,' and naltrexone pills are given only to those who are willing also to attend 'coping' therapy sessions.4 'Coping' therapy is usually contrasted to 'supportive' therapy, which, in North American treatment practice at any rate, implies supporting the client to remain abstinent. Coping therapy avoids the traditional moralism about 'falling off the wagon' through a more pragmatic cognitively oriented program in which clients are given specific instructions about how to reorganize their time and space to minimize temptation, and especially, how to avoid the guilt consequent upon taking one or two drinks that sends them into a tailspin of 'relapse.'5 One of the two main American scientists engaged in testing and promoting naltrexone treatment, Yale's Stephanie O'Malley, explicitly describes naltrexone as a 'pharmacological support' for cognitive behavioural therapy (O'Malley et al. 1992). This is also the way in which Dr Volpicelli's method is promoted (www.recovery2000.com). But while short courses of cognitive behavioural therapy used in relation to drugs may be becoming more popular in relation to alcohol than the lifelong program promoted by AA, the North American dogma that alcoholism is a psychic problem and requires therapy is never questioned. It is very significant that Stephanie O'Malley's seminal study, one of the two cited by government officials as justifying approval for this use of the drug, did not even include a group of clients who were receiving the drug but no therapy. The study was designed to see whether naltrexone worked better with supportive (abstinence-oriented) therapy or coping therapy (O'Malley et al. 1992; see also O'Malley et al. 1990) .6 Similarly, attempting to hold at bay the spectre that has haunted North American alcohol treatment for decades - the notion that alcoholics might be able to drink socially in a controlled manner - Berg and Volpicelli timorously wrote, 'while we certainly do not want to suggest that naltrexone could allow an alcohol-dependent individual to drink socially, the evidence presented

448 Mariana Valverde here does suggest that treatment with naltrexone can give patients a "second chance" after a slip from abstinence' (Berg et al. 1990: 145). The director of the NIAAA, while swept up in the magic-bullet biochemical enthusiasm immediately following government approval for naltrexone, likewise felt obliged to add: 'At the present time clinical research indicates that the best treatment results are achieved with a combination of pharmacotherapy and skilled counselling (NIAAA 1996). That this conclusion was driven by dogma rather than research is obvious from the fact that the NIAAA had to go out of its way to ignore O'Malley's Yale finding that naltrexone plus abstinence-oriented psychotherapy was not a good combination at all, and was indeed worse than being on placebo, a finding replicated in several other studies (see Agosti 1994; Heinala et al. 2001; Sinclair 2001). John David Sinclair, a rat psychologist at the Helsinki alcohol research unit that was until recently an integral part of the Finnish alcohol monopoly ALKO, is undoubtedly the most biologistic of all leading naltrexone researchers. He is a strict Pavlovian in a world that, especially in North America, is characterized by hybrid knowledges of the will promoted by hybrid enterprises such as twelve-step programs run by paraprofessionals defining themselves as alcoholics in recovery. Even he, however, shies away from stating that taking a naltrexone pill when one anticipates being in a situation with a high risk of excessive drinking amounts to a cure. When pressed on this point, he stated that while abstinence-oriented therapy is worse than useless for people on naltrexone (since if they are not actually drinking the pill has no opportunity to adjust to the neural pathways so as to wear away the conditioned response of prolonged drinking), 'coping' therapy should be provided along with the pills.7 The private clinics, mostly in Finland but now also as far away as Israel and Venezuela, using Sinclair's naltrexone method provide eight sessions with a psychologist or psychiatrist along with the pills. Dr Pearlman's clinic in Boston provides five sessions. In Toronto, the Addiction Research Foundation also makes naltrexone pills contingent upon attending therapy. The unease about prescribing pills and nothing else is therefore not driven by scientific considerations but rather by the lingering weight of the disease model of alcoholism. As various identity-producing diseases, from nymphomania to homosexuality, have been depathologized or otherwise reconfigured, the alcoholic is still an extremely successful identity. Harm reduction and risk management are certainly present in the broader field of the governance of alcohol consumption - drinking

Targeted Governance and the Problem of Desire 449 and driving campaigns, most famously, stigmatize the person who refuses to give up the car keys after having had a drink, not drinking per se. But harm reduction perspectives have been very marginal within treatment circles, particularly in North America. That the subpopulation identified as alcoholics cannot learn to evaluate and manage their risks is not only the great dogma of alcoholism treatment, especially in North America and in the Nordic countries, but is even a self-evident, circular truth, since alcoholics are defined precisely as those individuals who cannot manage the risks of drinking. And yet, whatever the theories of the researchers testing it, if while on naltrexone people classified as 'true alcoholics' (Jellinek 1960) become able to drink socially rather than having to follow the old advice of targeting their willpower exclusively on that first tempting drink, this discovery may open the door for the eventual 'extinction' of the whole field of alcoholism. Those formerly defined as alcoholics could be redefined as suffering from an ontologically obscure but nevertheless particular deficiency, whose nature does not have to be elucidated before it is remedied by the ingestion of a pill that somehow lowers the craving/ desire for the next drink. How this happens does not have to be known. AA has always advertised itself as effective 'because it works,' not because it has an ontologically coherent theory of alcoholism. And if naltrexone can prove that it works but at a lower emotional and psychic cost (because the individual does not have to actively refuse all drinks), then it will be regarded as more effective. But while succeeding in targeting 'getting drunk' as opposed to drinking in general, there are indications that naltrexone does not manage to target the desire to get drunk quite as smartly as one might wish. Although evidence is still scarce, it seems that using naltrexone regularly may lower enjoyment of and/or interest in activities such as taking exercise, having sex, and eating sweets. Sinclair, who initially suggested one naltrexone pill every day, has now changed his preferred protocol: Most earlier clinical trials have prescribed taking the antagonist every day. Selective extinction, however, is achieved by having patients take it only when they are drinking. In the optimal embodiment of this feature, patients avoid making other responses that are probably reinforced through the opioid system (e.g. eating highly palatable foods, having sex, jogging) while they are on the antagonist. Then when the craving for alcohol is manageable, they have days when no antagonist is taken, no alcohol is drunk, but these behaviours are now enjoyed. The simpler form of taking

450 Mariana Valverde naltrexone only when [actually before] drinking has also been called 'targeted medication.' (Sinclair 2001: 7)

The language of 'targeting' is here deployed rather misleadingly: taking the pill only on days when one anticipates drinking can of course be described as targeting, but the reason for limiting the pills only to certain days is precisely that naltrexone may be acting not in a targeted manner but more generally on pleasure-seeking behaviour. Clinicians dispensing naltrexone have neither proved nor disproved this generalizing, non-targeted effect. The only study I have found that inquired about sex drive reported that naltrexone does not lower men's sexual desire (Anton et al. 1999). But when pressed, one clinician told me about a patient who spontaneously reported that when playing squash while on naltrexone, he suddenly felt as if it were too much trouble to go and hit the ball, 'as if, "what's the point of squash anyway?"'8 This is only one person's story, but it is perhaps significant that in at least one version of the brain picture used in ContrAl PowerPoint presentations, Dr Sinclair has 'chocolate' marked on one of the icons that are supposed to represent receptors. Sinclair writes as if the broader effects that may or may not accompany the lowering of the urge to drink are in any case desirable: Pharmacological extinction [of the 'drink more' response] is a new form of medicine for treating behaviours that have become too powerful. What other behaviours might be treated? It can stop a rat from taking sweets. Can it control compulsive eating in humans? It can stop a rat from drinking methadone. Can it be used successfully to terminate methadone substitution therapy? Can it extinguish addiction to gambling, sexual obsession, compulsive thrill-seeking criminal activity? We will see. (Sinclair, 1998: 411)

These tentative findings cast doubt not so much on the success of the drug but rather on the governing project to target the behaviour of getting drunk and/or the desire to keep drinking. Will naltrexone pills enable alcoholics to drink socially, by contrast with the old-fashioned prohibitory strategy of AA, but without lowering their life force or animal spirits? Will alcohol science manage to produce a clearly distinguishable target for the programs seeking targeted governance of intoxication and/or alcoholism? And if the targeting is found to be general rather than specific, who will decide whether the 'side

Targeted Governance and the Problem of Desire 451 effects' - decreased interest in gambling, chocolate eating, and so forth - are desirable? The knowledges of desire that are being developed in naltrexone clinics and clinical trials, and more broadly in what is curiously known as 'craving research,' will inevitably be freighted by an older form of power/ knowledge: the dialectic of desire and willpower that has characterized the whole field of alcohol and addiction treatment since the rise of modern medicine. From Maximizing Willpower to Reducing Craving

From the private inebriate asylums for gentlemen of the late nineteenth century to the AA group meetings proliferating from the mid-twentieth century, alcoholism/inebriety treatments have generally sought to act upon the will. Alcoholism has never (except perhaps in France) been characterized as a purely physical disturbance - the ingestion of too much alcohol. It has been and still is regarded as a problematic relation of the self to consumption. How do you feel about your drinking? Do you drink to get away from problems? Can you stop drinking if you decide to stop? Do you hide your drinking from others? These are kinds of questions one finds in determinations of'alcohol dependence.' There is a consensus among experts that the existence of alcohol problems, even problems that are directly linked to drinking, like cirrhosis of the liver, is not a sure sign of the existence of alcoholism or alcohol dependence. Although the gap between alcohol problems and alcoholism is undoubtedly greatest in the United States, even in Europe alcohol dependence, like other forms of dependence, is not defined through physical symptoms or objective risk factors. The World Health Organization concluded in 1989 that 'dependence' is 'a cluster of physiological, behavioural, and cognitive phenomena in which the use of a drug takes on a much higher priority for a given individual than other behaviours that once had higher value. A central descriptive characteristic of the dependence syndrome is the desire (often strong, sometimes overpowering) to take drugs, alcohol, or tobacco' (quoted in Lindstrom 1992: 58). Alcoholism is still characterized, then, not as an imbalance of fluids but rather as the existence of a powerful inner compulsion. 'Loss of control,' the term popularized by alcohol science in the United States in the 1960s, is still found in alcoholism diagnosis and treatment literature, despite its transparently puritanical overtones. Traditionally, it was thought that the road back to control lay in

452 Mariana Valverde strengthening and maximizing the drinker's willpower. Alcoholics Anonymous tells people to put all of their willpower efforts in a single basket, concentrating on saying 'no' to the first drink. Targeted willpower, if you like. The classical AA definition of success is often clouded, today, by the discourse of self-help, self-esteem, dysfunctional families, and the rest. But in its classic form, AA provides a modest and yet powerful targeted tool with excellent audit potential: as long as you are not drinking (and continue to go to meetings), you are sober. You will never reform or be cured; there is no normalization possible for the alcoholic. But you can lead a normal life if you simply focus all your willpower - and that provided by the Higher Power - on a single act: refusing to take the first drink. Other, more rationalistic approaches try to modulate the drinker's willpower so as to regulate rather than prevent drinking. This more ambitious project is enacted through such techniques as learning to space one's drinks farther apart or readjusting the context within which drinks are chosen. People are encouraged to develop strategies such as purchasing attractive substitutes, avoiding certain events altogether, or modifying their daily time and space routines. People following these 'coping' programs have to exercise their will in relation to numerous acts and situations, but the cognitive faculties also get quite a work out: the drinker is in a constant state of self-monitoring. There are thus huge differences between AA and the techniques taught by more rationally oriented cognitive behavioural programs aimed at regulating drinking. What they have in common is that they act on a drinker's capacity and willingness to say 'no' to either all or some drinks. In cognitive behavioural therapy, the language of the will has disappeared, replaced by the language of 'coping,' 'situations,' 'decisions,' and 'contract.' But the ghost of the free will is clearly visible: the goal of therapy is to make it easier to say no to 'bad things.' In keeping with the traditional quasi-religious goal of strengthening one's capacity to say no to temptation, cognitive programs follow a logic of maximizing personal capacities. The opiate antagonists, however, are spoken about not in the language of boosting willpower but rather in the language of deficit reduction - more specifically, craving reduction. It would be as consistent with the scientific evidence to claim that naltrexone is a Popeye-spinach of the will, or of the rational capacity, as to say that it reduces craving. So if we are to understand what underlies the language of 'craving,' we must look elsewhere than to objective or scientific factors.

Targeted Governance and the Problem of Desire 453 The clinical psychology literature on 'craving' is full of semantic inconsistencies. The word 'desire' is sometimes used as a synonym for craving, as in the WHO definition of dependence cited above. At other times, however, 'craving' is defined as an irresistible urge for something specific, as in the cliche of pregnant women craving pickles and ice cream. So in the second usage, 'craving' is targeted desire. But the literature on craving is unusual among behavioural science areas of specialization in that it is full of worries about key concepts expressed in debates that threaten the field as a whole. Can craving be measured, inside or outside the laboratory? Is craving purely subjective? If it cannot be measured and scientifically operationalized, should the concept of craving be retained at all? An ingenious tool developed by smoking-cessation researchers, the Ecological Momentary Assessment mini-computer, seems promising for measuring craving in vivo: this is a Palm Pilot programmed to beep every so often to remind the subject to input a few measures recording mood and desire (Stone and Shiffman 1994). But for whatever reason this machine for recording desire in real life has not been used by alcohol and drug researchers. In any case, some of the people in the field argue that 'craving' is too negative a term, suggesting a lack, when the whole problem of drugs and alcohol is that, unlike food, we do not only take them when we feel a biological urge to consume. So what about pleasure? Is 'pleasure' a useful term? Or is pleasure simply the lay term for the satisfaction and reinforcement of a craving (Volpicelli et al. 1995; Abrams 2000)? And so forth. The latest research in this field tells us that craving is not coterminous with Pavlov-type physical symptoms such as salivation, since, like hunger, it can occur at any time and is not necessarily linked to physiological processes. More significant for our purposes is the finding that craving as measured in self-report scales is not always associated with quantity of drinking (Tiffany and Conklin 2000: 148). Finding that not all craving occasions lead to drinking would have led earlier scientists to reflect in laudatory terms about the strength of human willpower; but, in keeping with current orthodoxy, the authors of this study merely state that 'the regulation of drug use in experienced addicts can function independently of the processes that control craving' (ibid.). The craving researchers have also discovered that craving, far from being an irrational urge, is strongly influenced by cognitive factors. Experimental subjects who know that they will not get actual drugs or alcohol in the course of a study have a tendency to exhibit far fewer physical symptoms of 'craving' and to report far less psychological

454 Mariana Valverde craving (Meyer 2000: 222). So which brain process contains 'craving'? Nobody knows. It is thus clear that although we may now be past the age of 'diseases of the will' and into the age of deficits of brain chemicals, it is nevertheless proving very difficult to operationalize alcoholism in brain terms. Shifting the focus from the ever-elusive will to MRI pictures of the brain has not eliminated the chronic ontological ambiguities that have characterized 'alcoholism' since its inception in the late eighteenth century. There is a consensus that drinking is affected by a host of biochemical (and environmental) processes rather than located on a single biochemical site. To confound targeted governance further, the allegedly smart opiate receptor antagonists are not so smart after all, or are perhaps too smart. Naltrexone also works to alleviate heroin and methadone withdrawal and perhaps even to diminish compulsive gambling.9 The biochemistry and the neurology of desire currently preoccupy not only scientists but the public: we live in the age of Viagra as well as in what Nikolas Rose has called 'the age of Prozac.' And no doubt drugs exist, and others will be developed, that do have a measurable effect on both objective behaviours and 'cravings' insofar as it is possible to operationalize and quantify them. But the alcoholic brain remains a moving target. Even the brain-on-drugs, which is more easily visualized in terms of chemicals and neurons, may prove to be more of a moving target than the MRI scan pictures would suggest. A press release by the National Institute for Drug Abuse entitled 'Cocaine's pleasurable effects may involve multiple chemical sites' makes this point nicely: 'Recent studies with genetically altered mice have suggested that cocaine's euphoric effects may involve not just one, but several, chemical sites in the brain. These studies indicate that medications for treating cocaine addiction may need to target these multiple sites just as cocaine does' (NIDA1998). A Final Note on Temporality

The logic of correction, monitoring, audit, and adjustment that characterize our particular historical moment seems to be compatible with two opposing temporalities. On the one hand, the demands of cost-cutting and efficiency, as well as the general scepticism about scientific cures for the soul, have put pressure on long-term projects like psychotherapy: witness the ubiquitousness of 'brief therapy' and coping-skills training. Identify the little problem, the broken window of the soul or of the

Targeted Governance and the Problem of Desire 455 house, fix it before it gets big, and get on with life - that seems to be the advice we receive at the therapist's office as well as the police department. On the other hand, however, the logic of audit, monitoring, performance assessments, and lifelong learning seems to carry with it a life sentence. Risk assessments are constantly updated; tenure is abolished in favour of yearly assessments; and so on. On the biomedical side, this may mean that adjustments are envisaged as temporally indefinite. What about naltrexone? In the United States, clinical trials lasted a few weeks, twelve at the most, for reasons left unspoken but probably related to DuPont's anxiety to secure the patent and obtain FDA approval for the new drug. The Finnish trials were also very time-limited. But a few years later, the inventor of the 'Sinclair method' tells us that although naltrexone should not be taken every day because of a possible rebound effect, it could or should be taken for life: There is no experimental or theoretical justification for terminating naltrexone treatment after a fixed period of time of any duration. The only real justification is financial: taking naltrexone every day for the rest of one's life would be expensive' (Sinclair 2001: 8). For Sinclair, alcoholics needing to take naltrexone all their lives is no different than diabetics' life-long dependency on insulin. The quest for drugs to enable the targeted governance of specific biochemical processes under the banner of governing more modestly and specifically seems thus to lead us straight into the ominous temporality of continuous assessment and adjustment. Governing through specific flows and processes rather than through persons has become fashionable in many fields, but its modes of temporalization have rarely been queried. Even when there is a specific target entity or process that can indeed be effectively isolated, constantly monitoring and correcting the target is not likely to amount to governing less. Notes 1 Some psychiatrically oriented studies of naltrexone tested the 'hypothesis that some people are predisposed to alcoholism because of a genetic or environmentally induced deficiency in opiate receptor activity. The basic tenet of the Opioid Compensation Hypothesis is that alcohol drinking stimulates opate receptor activity and it is the increase in opiate receptor activity that reinforces alcohol drinking and leads to alcohol dependence' (Berg et al. 1990: 137). But naltrexone's fortunes are by no means linked to

456 Mariana Valverde

2

3

4 5

6

7 8 9

those of the contested entity that used to be known as addiction and is now, in medical texts, called 'dependence.' Naltrexone is currently being used to help people stop using methadone as well as heroin in private clinics in Australia, although there is much controversy about whether the so-called Rapid Detox offered is successful - there are reports of excess suicides, among other criticisms (Australia Courier Mail 6 June 2001). An example of this type of use can be found at members.ozemail.com/au. Elsewhere I explore the waning of the 'homosexual' identity by examining human rights cases that revolve around gay lifestyles - see chapters 4 and 5 in Mariana Valverde, Law's Dream of a Common Knowledge: Legal Inquiries into Habits and Lifestyles (in press, Princeton University Press). Personal communication from Anu Goodman, April 2001. Coping therapy is often practised by psychologists and social workers rather than psychiatrists. One of the reasons for the very limited availability of naltrexone treatment in North America is that Dupont marketed Re Via (naltrexone) only to psychiatrists, rather than to therapists and general practitioners (personal communication from Dr Peter Selby, clinical director, Centre for Addiction and Mental Health, July 2001). The other key study, by Volpicelli's group, involved seventy men who went through the regular inpatient detox program provided in a standard manner to alcoholics by Veterans Administration hospitals in the United States (Volpicelli et al. 1992). The published study is not very informative about the psychological dimension of the detox program, but it is significant that Volpicelli's clinical trial also lacked a group that received naltrexone pills but no therapy of any kind. E-mails from Dr Sinclair to author, April and May 2001. Personal communication, Anu Goodman, April 2001. The Centre for Addiction and Mental Health is beginning work on clinical trials of naltrexone for people who drink too much and are also 'compulsive gamblers' (Dr Peter Selby, personal communication, July 2001).

References Abrams, D. 2000. 'Transdisciplinary Concepts and Measures of Craving: Commentary and Future Directions.' Addiction 95 (Supplement 2): 237-46 Agosti, V. 1994. 'The Efficacy of Treatment in Reducing Alcohol Consumption: A Meta-analysis.' InternationalJournal of the Addictions 30: 1067-77 Amit, Z., and B.R. Smith. 1990. 'Neurotransmitters Regulating Alcohol Intake.'

Targeted Governance and the Problem of Desire 457 In C. Naranjo and E. Sellers, eds. Novel Pharmacological Interventions for Alcoholism, 161-83. New York: Springer Verlag Anton, R.F., D. Moak, L.R. Waid, P.K. Latham, et al. 1999. 'Naltrexone and Cognitive Behavioural Therapy for the Treatment of Outpatient Alcoholics: Results of a Placebo-Controlled Trial.' American Journal of Psychiatry 156(11): 1758-64 Berg, J.B.,J. Volpicelli, D. Herman, and C. O'Brien. 1990. The Relationship of Alcohol Drinking and Endogenous Opiods: The Opioid Compensation Hypothesis.' In C. Naranjo and E. Sellers, eds., Novel Pharmacological Interventions for Alcoholism, 137-47. New York: Springer Verlag Heinala, P., H. Alho, K. Liaianmaa.J. Lonnqvist, and J.D. Sinclair. 2001. 'Targeted use of Naltrexone without Prior Detoxification in the Treatment of Alcohol Dependence: A Factorial Double-Blind Placebo-Controlled Trial.' Journal of Clinical Psychoparmacology 21(3): 287-92 Jellinek, E. 1960. Tlie Disease Concept of Alcoholism. New Haven, CT: Hillhouse Press Lindstrom, L. 1992. Managing Alcoholism: Matching Treatments to Clients. New York: Oxford University Press Mason, B., F. Salvato, L. Williams, E. Ritvo, and R. Cutler. 1999. 'A Doubleblind, Placebo-controlled Study of Oral Nalfemene for Alcohol Dependence.' Archives ofCenmd Psychiatry 56(8): 719-24 Meyer, R. 2000. 'Craving: What Can be Done to Bring the Insights of Neuroscience, Behavioral Science and Clinical Science into Synchrony.' Addiction 95 (Supplement 2): 219-27 Morrison, J. 2000. 'The Government-Voluntary Sector Compacts: Governance, Governmentality, and Civil Society.' Journal oj''Law and Society 27: 98-1 32 Naranjo, C., and K.E. Bremner. 1990. 'Evaluation of the Effects of Serotonin Uptake Inhibitors in Alcoholics: A Review.' In C. Naranjo and E. Sellers, eds., Novel Pharmacological Interventions /or Alcoholism, \ 05—20. New York: Springer Verlag Naranjo, C., K.E. Bremner, and C.X. Poulos. 1990. 'A Human Model for Testing Drug-Induced Concomitant Variations in Alcohol Consumption and the Desire to Drink.' In C. Naranjo and E. Sellers, eds., Novel Pharmacological Interventions for Alcoholism, 288-98. New York: Springer Verlag NIAAA. 1995. Press Release: 'Naltrexone Approved for Alcoholism Treatment.' 17January - 1996, 3 July. Alcohol Alert. NIDA. 1998. 'Cocaine's Pleasurable Effects May Involve Multiple Chemical Sites.' Notes.

458 Mariana Valverde - 1999. News Release: 'Alcohol Researchers Identify New Medication that Lessens Relapse Risk [nalmefene].' 20 August O'Malley, S., et al. 1990. 'Naltrexone in the Treatment of Alcohol Dependence: Preliminary Findings.' In C. Naranjo and E. Sellers, eds., Novel Pharmacological Interventions for Alcoholism, 148-59. New York: Springer-Verlag O'Malley, S., A. Jaffe, G. Chang, R. Schohenfeld, R. Meyer, and B. Rounsaville. 1992. 'Naltrexone and Coping Skills for Alcohol Dependence: A Controlled Study.' Archives of General Psychiatry 49: 881-8 Power, M. 1994. 'The Audit Society.' In A. Hopwood and P. Miller, eds., Accounting as a Social and Institutional Practice, 299-316. Cambridge: Cambridge University Press Rose, N. 2001. 'Normality and Pathology in a Biological Age.' Unpublished ms. Sinclair, J.D. 1987. 'The Feasibility of Effective Psychopharmacological Treatment for Alcoholism.' British Journal of Addiction 82: 1213-23 - 1997. 'Development in Finland of the Extinction Treatment for Alcoholism with Naltrexone.' Psychiatria Fennica 28: 76-97 - 1998. 'New Treatment Options for Substance Abuse from a Public Health Point of View.' Ann Med of the Finnish Medical Society 30: 406-11 - 2001. 'Evidence about the Use of Naltrexone and for Different Ways of Using it in the Treatment of Alcoholism.' Alcohol and Alcoholism 36: 2-10 Stone, A., and S. Shiffman. 1994. 'Ecological Momentary Assessment [EMA].' Annals of Behavioral Medicine 16: 199-203 Tiffany, S., and C. Conklin. 2000. ' A Cognitive Processing Model of Alcohol Craving and Compulsive Alcohol Use.' Addiction^ (Supplement 2): 145-53 Valverde, M. 1998. Diseases of the Will: Alcohol and Dilemmas of Freedom. Cambridge: Cambridge University Press Volpicelli, J., A. Alterman, M. Hayashida, and C. O'Brien. 1992. 'Naltrexone in the Treatment of Alcohol Dependence.' Archives of General Psychiatry 49: 87680 Volpicelli, J., N. Watson, A. King, C. Sherman, and C. O'Brien. 1995. 'Effect of Naltrexone on Alcohol "High" in Alcoholics.' American Journal of Psychiatry 152:613-15

Contributors

John Adams is Professor of Geography, University College London. His most recent book is Risk (University College London Press, 1995). Tom Baker is Connecticut Mutual Professor of Law and Director of the Insurance Law Center, School of Law, University of Connecticut. His most recent books are Embracing Risk: The Changing Culture of Insurance and Responsibility (co-edited with Jonathan Simon, University of Chicago Press, 2002), and Insurance Law and Policy (Aspen, 2003). Aaron Doyle is Assistant Professor of Sociology, Carleton University. His most recent books are Arresting Images: Crime and Policing in Front of the Television Camera (University of Toronto Press, 2003), Insurance as Governance (with Richard Ericson and Dean Barry, University of Toronto Press, 2003) and Uncertain Business: Insurance and the Limits of Knowledge (with Richard Ericson, University of Toronto Press, forthcoming) Richard V. Ericson is Principal of Green College and Professor of Law and Sociology, University of British Columbia. His most recent books are Policing the Risk Society (with Kevin Haggerty, University of Toronto Press and Oxford University Press, 1997), Governing Modern Societies (co-edited with Nico Stehr, University of Toronto Press, 2000), Insurance as Governance (with Aaron Doyle and Dean Barry, University of Toronto Press, 2003) and Uncertain Business: Insurance and the Limits of Knowledge (with Aaron Doyle, University of Toronto Press, forthcoming). David Garland is Professor of Law and Sociology, New York University. His most recent books are Criminology and Social Theory (co-edited with

460

Contributors

Richard Sparks, Oxford University Press, 2000) and The Culture of Control: Crime and Social Order in Contemporary Society (Oxford University Press and University of Chicago Press, 2001). Ian Hacking holds the Chair of Philosophy and History of Scientific Concepts, College de France, and is University Professor of Philosophy, University of Toronto. His most recent books are Rewriting the Soul: Multiple Personality and the Sciences of Memory (Princeton University Press, 1995), The Social Construction of What? (Harvard University Press, 1999), Mad Travelers: Reflections on the Reality of Transient Mental Illnesses (Harvard University Press, 2002), and Historical Ontology (Harvard University Press, 2002). Kevin D. Haggerty is Assistant Professor of Sociology, University of Alberta. His most recent books are Policing the Risk Society (with Richard Ericson, University of Toronto Press and Oxford University Press, 1997) and Making Crime Count (University of Toronto Press, 2001). Carol A. Heimer is Professor of Sociology, Northwestern University, and Senior Research Fellow, American Bar Foundation. Her most recent book is For the Sake of Children: The Social Organization of Responsibility in the Hospital and the Home (with L. Staffen, University of Chicago Press, 1998). Alan Hunt is Professor of Law and Sociology, Carleton University. His most recent book is Governing Morals: A Social History of Moral Regulation (Cambridge University Press, 1999). Pat O'Malley is Canada Research Chair in Law, Carleton University. His most recent book is Crime and the Risk Society (edited, Dartmouth, 1998). Michael Power is P.D. Leake Professor of Accounting and Co-Director, Centre for the Analysis of Risk and Regulation, London School of Economics and Political Science. His most recent books are The Audit Explosion (Demos, 1994) and The Audit Society: Rituals of Verification (Oxford University Press, 1997). Nikolas Rose is Professor and Head of Sociology, London School of Economics and Political Science. His most recent books are Inventing Our Selves: Psychology, Power and Personhood (Cambridge University Press,

Contributors 461 1996) and Powers of Freedom: Reframing Political Thought (University of Cambridge Press, 1999). Violaine Roussel is Assistant Professor of Political Science, Universite Paris VIII. Her most recent book is Les Magistrates dans les Scandales Politiques en France (Editions La Decouverte, 2002). Jonathan Simon is Professor of Law, University of Miami. His most recent book is Embracing Risk: The Changing Culture of Insurance and Responsibility (co-edited with Tom Baker, University of Chicago Press, 2002). Mariana Valverde is Professor of Criminology, University of Toronto. Her most recent books are Diseases of the Will: Alcohol and the Dilemmas of Freedom (Cambridge University Press, 1998) and Law's Dream of a Common Knowledge: Legal Inquiries into Habits and Lifestyles (Princeton University Press, 2003).