The Oxford Handbook of Law and Economics 9780199684267, 9780199684205, 9780199684250, 9780198803737, 019968426X

Table of contents :
Cover
The Oxford Handbook of Law and Economics
Copyright
Contents
List of Figures
List of Tables
List of Abbreviations
List of Contributors
Part I Constitutions, Statutes, and Regulations
1. Optimal Constitutional Structure
2. The Design of Constitutions
3. Democratic Rulemaking
4. Regulatory Decision-​Making and Economic Analysis
5. Economics of Federalism
6. Economics of Tax Law
Part II Civil Process and Litigation
7. Judge-​Made Law and the Common Law Process
8. Economics of Civil Procedure
9. Law and Economics of Evidence
10. Economics of Litigation
11. Settlement and Trial
12. Legal Fees and Lawyers’ Compensation
13. Legal Profession and the Market for Lawyers
14. Law and Economics of Alternative Dispute Resolution
Part III Crime and Punishment
15. Economics of Criminal Law: Crime and Punishment
16. Economics of Criminal Procedure
17. Guns and Crime
18. Prosecutorial Strategies
Part IV Beyond National Legal Systems
19. Economics of Ancient Legal Systems
20. Economics of Legal History
21. Economics of International Law
22. Enforcement of International Law
23. International Finance and Sovereign Debt
24. Economics of International Organizations
25. Choice of Law and Conflict of Laws
Name Index
Subject Index

Citation preview

 

T h e Ox f o r d H a n d b o o k o f

L AW A N D E C ON OM IC S

 

 

The Oxford Handbook of

LAW AND ECONOMICS Volume 3: Public Law and Legal Institutions Edited by

FRANCESCO PARISI

1

 

3 Great Clarendon Street, Oxford, ox2 6dp, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Oxford University Press 2017 The moral rights of the author‌have been asserted Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2017935284 ISBN 978–0–19–968426–7 (Volume 1) ISBN 978–0–19–968420–5 (Volume 2) ISBN 978–0–19–968425–0 (Volume 3) ISBN 978–0–19–880373–7 (Set) Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

 

Contents

List of Figures  List of Tables  List of Abbreviations  List of Contributors 

ix xi xiii xv

PA RT I   C ON ST I T U T ION S , S TAT U T E S , A N D R E G U L AT ION S 1. Optimal Constitutional Structure  Richard A. Epstein

3

2. The Design of Constitutions  Tom Ginsburg

28

3. Democratic Rulemaking  John M. de Figueiredo and Edward H. Stiglitz

37

4. Regulatory Decision-​Making and Economic Analysis  Mariano-​Florentino Cuéllar and Jerry L. Mashaw

59

5. Economics of Federalism  Robert P. Inman and Daniel L. Rubinfeld

84

6. Economics of Tax Law  Daniel Shaviro

106

PA RT I I   C I V I L P RO C E S S A N D L I T IG AT ION 7. Judge-​Made Law and the Common Law Process  Ben Depoorter and Paul H. Rubin

129

8. Economics of Civil Procedure  Robert G. Bone

143

 

vi   Contents

9. Law and Economics of Evidence  Chris William Sanchirico

171

10. Economics of Litigation  Bruce H. Kobayashi

201

11. Settlement and Trial  Andrew F. Daughety and Jennifer F. Reinganum

229

12. Legal Fees and Lawyers’ Compensation  Winand Emons

247

13. The Legal Profession and the Market for Lawyers Albert Yoon

259

14. Law and Economics of Alternative Dispute Resolution  Talia Fisher

280

PA RT I I I   C R I M E A N D P U N I SH M E N T 15. Economics of Criminal Law: Crime and Punishment  Isaac Ehrlich

295

16. Economics of Criminal Procedure  Keith N. Hylton

325

17. Guns and Crime  Anthony A. Braga

344

18. Prosecutorial Strategies  Roger Bowles

370

PA RT I V   B E YON D NAT IONA L L E G A L SYS T E M S 19. Economics of Ancient Legal Systems  Geoffrey Parsons Miller

395

20. Economics of Legal History  Daniel Klerman

409

21. Economics of International Law  Alan Sykes and Andrew Guzman

439

 

Contents   vii

22. Enforcement of International Law  Paul B. Stephan

465

23. International Finance and Sovereign Debt  W. Mark C. Weidemaier and Mitu Gulati

482

24. Economics of International Organizations  Joel P. Trachtman

501

25. Choice of Law and Conflict of Laws  Erin O’Hara O’Connor

527

Name Index  Subject Index 

549 557

 

 

List of Figures

6.1 Individuals with different wage rates 

109

10.1 The structure of litigation—​simple one-​stage model 

203

10.2 Filing, litigation, and settlement—​low-​stakes (200) 

204

10.3 Filing, litigation, and settlement—​high-​stakes (600) 

205

10.4 The effect of two-​way cost-​and fee-​shifting 

207

10.5 The effect of one-​way fee-​shifting 

209

10.6 The effect of a contingency-fee arrangement—​high stakes

210

10.7 One-​stage negative value lawsuit 

212

10.8 Three-​stage model of litigation 

213

10.9 Asymmetric and sequential costs 

214

10.10 Three-​stage litigation with information revelation—​total payoffs 

215

10.11 Three-​stage litigation with information revelation—​marginal payoffs 

216

10.12 Option value 

218

10.13 Option value with settlement 

219

10.14 Consolidated pre-​trial 

219

10.15 Option value with consolidated pre-​trial 

220

10.16 Reversing order and option value 

221

10.17 The effect of prevailing litigant two-​way fee-​shifting 

221

10.18 Prevailing litigant two-​way fee-​shifting and option value 

222

 

 

List of Tables

10.1 Pre-​trial motion outcomes 

216

18.1 Conviction rates by offense type, England and Wales 2012–​13 

386

 

 

List of Abbreviations

ABA

American Bar Association

ADR

alternative dispute resolution

APA

Administrative Procedures Act

ATLA

Association of Trial Lawyers in America

ATF

Alcohol, Tobacco, Firearms, and Explosives, Bureau of

BIT

Bilateral Investment Treaty

BPD

Boston Police Department

CFR

Code for Federal Regulation

CIL

customary international law

CJS

criminal justice system

CWC

Chemical Weapons Convention

EBA

Extreme Bounds Analysis

FIST

Federalism Impact Statement

FRCP

Federal Rule of Civil Procedure

FSIA

Foreign Sovereign Immunities Act (United States)

GATT General Agreement on Tariffs and Trade ICANN Internet Corporation for Assigned Names and Numbers ICCPR International Covenant on Civil and Political Rights IO

international organization

IPD

Indianapolis Police Department

LSAT

Law School Admissions Test

NAFTA North American Free Trade Agreement NAS

National Academy of Sciences

NCVS

National Crime Victimization Survey

NEV

negative expected value

NPRM Notice of Proposed Rulemaking OIRA

Office of Information and Regulatory Affairs

OPCW Organization for the Prohibition of Chemical Weapons

 

xiv   List of Abbreviations PEV

positive expected value

PROMIS

Prosecutor’s Management Information System, Washington, DC

PSN

Project Safe Neighborhoods

PUD

planned unit development

RST

retail sales tax

SIA

State Immunity Act (United Kingdom)

UCR

Uniform Crime Reports

UGPA

undergraduate grade point average

USBJS

US Bureau of Justice Statistics

USC

United States Constitution

VAT

value-​added tax

WTO

World Trade Organization

WTO DSB  World Trade Organization’s Dispute Settlement Body

 

List of Contributors

Robert G. Bone, The University of Texas at Austin School of Law Roger Bowles, University of York Anthony A. Braga, Northeastern University Mariano-​Florentino Cuéllar, Stanford University Andrew F. Daughety, Vanderbilt University John M. de Figueiredo, Duke University Ben Depoorter, University of California, Hastings Isaac Ehrlich, University at Buffalo Winand Emons, University of Bern Richard A. Epstein, The New York University School of Law Talia Fisher, Tel Aviv University Tom Ginsburg, University of Chicago Law School Mitu Gulati, Duke University Andrew Guzman, USC Gould School of Law Keith N. Hylton, Boston University School of Law Robert P. Inman, Wharton School, University of Pennsylvania Daniel Klerman, USC Law School Bruce H. Kobayashi, George Mason Law School Jerry L. Mashaw, Yale University Geoffrey Parsons Miller, New York University Law School Erin O’Hara O’Connor, Florida State University College of Law Jennifer F. Reinganum, Vanderbilt University Paul H. Rubin, Emory University

 

xvi   List of Contributors Daniel L. Rubinfeld, University of California, Berkeley and NYU Chris William Sanchirico, University of Pennsylvania Daniel Shaviro, New York University Law School Paul B. Stephan, University of Virginia School of Law Edward H. Stiglitz, Cornell University Alan Sykes, Stanford Law School Joel P. Trachtman, Tufts University W. Mark C. Weidemaier, University of North Carolina School of Law Albert Yoon, University of Toronto Faculty of Law

 

Pa rt  I

C ON ST I T U T ION S , STAT U T E S , A N D R E G U L AT ION S

 

 

Chapter 1

Op t im al C onsti t u t i ona l Stru ct u re Richard A. Epstein

In dealing with any comparative venture it is usually common to stress the differences that arise across legal and political cultures, and not their similarities. That result is driven in large part by the perception that a complete legal, historical, and economic analysis has to address the unique forces that drive different nations in different directions. Descriptively, these differences really matter. Formally, socialist countries often offer a rich palate of positive rights—​to jobs, education, housing, minimum income—​to which all people are entitled. But they cannot deliver on the rights that they provide. In contrast, an optimal constitution takes a classical liberal form. Formally, it guarantees few direct benefits to its citizens. But it is better able to deliver on its promise through limited government power that expressly delineates spheres of protected individual rights that are presumptively insulated from government control. Any empirical study would have to devote enormous attention to the genuine shipwrecks that happen in the first class of cases—​think of Cuba, the Soviet Union, North Korea, Venezuela. These fail as governments because they cannot generate the resources necessary to support their extensive commitments to positive rights, after locking themselves into state ownership of the means of production.1 The reason that these ambitious governments fail is that they systematically ignore the dangers of faction and self-​interest that plague any organization. The successful constitution has to be able to respond to these forces. And in order to do so, it must have a systematic normative plan for success. So it is that the constitutions of the Western democracies, and their imitators elsewhere, have developed two strategies to deal with these problems. The first is to lower the expectations of what it is that government can do. The second is to limit the ways in which these particular activities can be undertaken, with both structural limitations and the protection of individual rights. 1 

See, e.g., Constitution (Fundamental Law) of the Union of Soviet Socialist Republics (Oct. 7, 1977), http://​www.departments.bucknell.edu/​russian/​const/​77cons01.html#I.

 

4   Richard A. Epstein The great challenge therefore is to find a model of how these collective activities should be undertaken, which I  believe involves the adoption of the worldview that I have elsewhere called the “classical liberal constitution.”2 And in this regard, the best way to attack the problem is to look at the way in which the challenges of governance are met by those private organizations that engage in collective decision-​making. It is of course the case that voluntary organizations differ systematically from political ones, chiefly for the reason that they have no territory to govern and defend. But the best way in which to proceed with the normative analysis is to figure out how these organizations work, and then ask what modifications should be made in dealing with the transition from voluntary organizations to political ones. In dealing with this issue, there are of course many different kinds of voluntary organization, each with its own peculiarities. The most obvious form of comparison is the public corporation with a large number of diffuse shareholders, which for this purpose are roughly analogous to citizens. But of equal importance in this context is the variety of planned unit developments, condominium and cooperative associations (collectively, “PUDs”) that add to the diffuse shareholders two characteristics that are relevant in any analysis of government power. The first of these is that individual members of these organizations have exclusive property rights in their own units, just as they share in the use of the common elements that are necessarily part of these systems. The second is that these systems all rely on a system of taxation or assessment to fund the common activities of the business, which they must do without upsetting the private rights that they hold in their individual units. It is, therefore, to these organizations that I shall devote primary attention, as opposed to charitable organizations, partnerships, labor unions, and friendly organizations.3 In section 1.1, I shall identify the mechanisms that corporate organizers and property developers use to attract and keep outside capital, noting the role of two forms of protection: structural protections and those for individual rights. Thereafter I shall examine how these mechanisms carry over to political institutions, noting how the analysis is organized along two different axes—​one of which deals with the difference between unitary and federalist systems, and the other with the difference between presidential and parliamentary systems. The bottom line on this matter is that, relative to the question of individual rights, it is more difficult to develop any general theory about the structural constitution. Indeed, the optimal structure will depend heavily on the size and shape of the polity, and the ethnic and regional differences within it.4 But the theory should be able to formulate a set of substantive protections of individual rights that, by reducing the scope of government activities, necessarily reduces the possibility of intrigue in the collective decisions that necessarily remain, even where political pressures are severe due to divisions based on region, ethnicity, religion, or wealth. The central theorem is this: the greater the internal

2 

Epstein (2014a). The book covers most of the topics here, to which I do not give separate citation. For an account of these differences, see Epstein (2014b). 4  For a more complete statement of my views, see Epstein (2011). 3 

 

Optimal Constitutional Structure    5 differences within the collectivity, the greater the need to limit the scope of government activities in order to control the latent internal tensions.

1.1  Contracts, Corporations, and PUDs There are few if any private corporations or PUDs that are formed by government diktat. Instead, the law establishes a framework in which private parties can form these associations voluntarily for themselves. To see how this is done, start with corporations. In the early days, corporations needed special charters in order to be formed, so that the group fortunate enough to acquire such a charter could then seek to block entry by a rival firm seeking the privileges of incorporation, chiefly limited liability.5 But limited liability only goes to the protection of individual shareholders from the claims of corporate outsiders. The governance issue, which is vital to the question of constitutional design, involves the relationship of the shareholders to each other. The first point about corporate formation is that it has to satisfy the same imperatives that are needed for any form of simple two-​party contractual venture from the ex ante perspective. As a medium of exchange, two parties will only enter into an agreement if each party thinks itself better off from the exchange than otherwise. Only deals that are win/​win in expectation for both participants can emerge. When both sides agree, there is a positive sum game, from which both benefit, usually in proportions that are not easy for outside observers to detect or measure; even each party to the agreement may find it difficult to understand the net gains to the other side. But the conditions to secure those mutual benefits are not easy to come by, because there is always the assurance problem facing the party who performs first: How can it be sure that the other side will reciprocate when the time comes? The assurance problem is reduced by the simultaneous exchange of goods or services. But even the simplest contract dealing with long-​term investment usually requires the paying party to perform first, and the party will only do so if confident that the gains will be available at the second stage of the transaction. Contract law and state power cannot be the drivers of exchange, but they can provide the infrastructure that controls the inherent risks from sequential performance, which, in turn, expands the potential for gains from trade above those obtainable from simultaneous exchanges. The likelihood of contractual success thus depends on a rich tapestry of private norms, close personal affiliations, third-​party guarantors, and the judicious use of state power. The more well established these elements are, the greater the likelihood of contractual success, and, in turn, the greater the willingness of people to participate in similar ventures. The same logic of requiring gains in expectation for all applies to multiparty ventures, like corporations and PUDs, only now the central task of the founders is more difficult. 5 

For discussion of the process, see Butler (1986).

 

6   Richard A. Epstein They must persuade individuals to commit capital to a firm or a venture, knowing that they will have to commit their money today into a venture over which they will not have direct control tomorrow. At the time of formation, the founders of these organizations have the right incentives to form an organization, both in the delineation of their rights, and in the mechanisms for their enforcement. On the former, the insiders know that each time that they grant a right to some individuals, within the closed universe of the corporation they necessarily impose a duty on someone else. In many situations, when large public issues are at stake, the world is not neatly divided into two discrete groups, one whose members are subject to duties and the second whose members hold the correlative rights. In fact, the standardization of shares means, as a first approximation, that all rights and duties are identical for the shareholders. Clearly the quantity of rights is fixed, but their value is not, so that the promoter of any collective venture has a strong incentive to set matters right in order to maximize the total value that he can receive from all shareholders when they make their initial investments. These voluntary organizations thus have favorable qualities that political organizations find difficult to achieve, because political organizations with territorial bases can never hope to generate the unanimity that is always possible when the founding group sells shares to outsiders—​all of whom invest only if they think that the value of the shares they receive, with its complex set of rights and duties, will be worth more than the amount of the invested capital. It is for this one reason that the corporation forms a useful normative baseline for examining the organization of constitutional structures. There are multiple devices that increase the odds that any given positive sum venture will be carried to a favorable conclusion. Corporate formation is not a simple exchange. It is a long-​term set of relationships in which each person has to have confidence that down the road they will receive a positive return from the collective venture. One risk that the promoter must dispel is that he will divert the assets in the firm to his own personal use. Another is that, in the course of running the business, the promoter will destabilize the initial investments by favoring one group of shareholders over another. The risks in question can be subtle, and need not involve the outright confiscation of all capital. It is enough if the dice are loaded so that some people get a higher rate of return on investments than do others. From the ex ante perspective any corporate wobble adds uncertainty to the venture, which reduces the rate of return for risk-​averse shareholders, especially when the resources spent to obtain benefits or to resist efforts by others to obtain them are deadweight losses that reduce the anticipated return to all parties. Stability of entitlements (which is not the same as low-​risk businesses) is win/​win all the way around, so that the promoters have to have strong incentives to shut down ex post opportunism. But what strategies should be used to deal with the common separation of ownership and control,6 and the agency cost problems to which it gives rise (Jensen, 1976)? Early on, one device was to limit the ends to which corporate funds could be spent. Under this doctrine of ultra vires, shareholders had greater knowledge of the limited set of resource 6 

For one influential account, see Shleifer and Vishny (1997). The classic study on the subject in the American context is Berle and Means (1932).

 

Optimal Constitutional Structure    7 uses, and they could choose their investment in areas that they knew best, where the targeted scope maked matters easier to monitor. In addition to these limited ends, the corporation can adopt a business structure that minimizes the possibility of diversion of assets by dividing governance powers among a variety of individuals. It is thus commonplace for complex structures that feature a CEO, appointed by the Board of Directors, to run the business subject to the Board’s oversight and, if need be, power of removal. Individual investors will also have more confidence in the venture if other ordinary shareholders, whom they neither know nor trust, will not take an active role in running the business, reducing both risk and monitoring costs. The size and independence of the Board are always critical. It must be large enough to enjoy some form of legitimacy, but small enough to make key decisions rapidly and well. Typically, boards function as a committee of the whole on key issues, but the efficient organization of any private association depends on the ability to delegate specific tasks—​budget, management, renovations, external relations—​to smaller committees that can develop the expertise and devote the time to tackle these issues. Ultimately, however, the shareholders have some say in the operation of the business, reflected by their ownership of voting shares, which can generate a market in corporate control if management misbehaves (Manne, 1965). Some of the protections are structural, including the right to pass on corporate transactions, such as mergers, acquisitions, and spinoffs. But even in the day-​to-​day operation of a corporation, the alienability of corporate shares allows people to express their dissatisfaction with current policy by selling. The ability to exit from a venture without the concurrence of other shareholders makes it easier for investors to enter the business early on. The greater coherence in capital structure also makes it easier to borrow capital from lenders who also have less to fear from asset dissipation. In general, these arrangements will be marked by strict debt covenants. For, owing to potential conflicts of interest, the same set of trustees cannot owe generalized fiduciary duties to two classes of investors who have different objectives. The creditors are interested in stable flow sufficient to cover principal and interest. The shareholders are more willing to take large risks given that they get all of the upside, and can externalize part of the downside on creditors—​at least if those creditors do not take sufficient steps to protect their interests. This constant concern with conflicts of interest gives some clues to the capital structure of the firm. The first rule for dividing control is of course, not one person, one vote; it is one share, one vote. Furthermore, in public corporations, there is typically only one class of share. That single class of shares makes valuation problems easier, because common shareholders do not have to figure out the value of various classes of preferred stock in order to make their valuations: they need only deal with debt, for which the valuation is easier. And the large single class creates a thick market that in turn allows for more accurate pricing in cheap and high velocity transactions. The single class of stock also makes it easier to figure out dividends and other distributions. These, when done on a pro rata basis, reduce the possibility that insiders can manipulate accounts in ways that prejudice one group of shareholders at the expense of others. In addition, the single class of stock makes it easier to organize derivative judicial actions brought on behalf of a corporation in the event that there is some diversion of

 

8   Richard A. Epstein corporate funds. These actions in effect use a fair-​value standard in cases of self-​dealing as a way to ensure that the insiders that control the operation do not funnel off assets to themselves by selling corporate assets at bargain basement prices or pay outside assets for an undeserved premium. Yet at the same time, a business judgment rule always applies to officers and directors when they are dealing in nonconflict situations, because it is well understood that no one would take the job of running a corporation if he had to internalize the costs of wrong decisions made in good faith on behalf of the shareholders. But even a good faith standard offers only a qualified and not an absolute form of immunity for any dereliction. In the United States, and to a lesser extent, the United Kingdom, derivative actions brought by one shareholder may be brought on behalf of all shareholders against parties who have received improper benefits from the corporation. If the suit is dismissed, no reimbursement is typically allowed against corporate assets. But if the derivative succeeds against those third parties, then the moving parties receive compensation from the corporation for the funds that have been restored to it. Clearly, no corporation can operate sensibly without some degree of external oversight, and if it is not through individual lawsuits, it must come through direct government oversight. There are, however, several ways that PUDs are closer to governments than corporations (Epstein, 1987a). The key point is that it is never the case in PUDs that all members have only a fractional issue in some common asset pool. Instead, each member has its own separate unit, so it is now necessary to sort out the common from the separate elements. At the very least, this decision requires that there be strong protection in both the exclusive possession and use of the unit. This requires not only that others be kept out, but also that the unit owner has guaranteed access to the unit through the areas under common control. The usual arrangements allow for the sale of individual units, but this process is more complicated than it is for corporate shares because other co-​owners have a standing interest in the identity of the neighbors. Therefore, they must have some say over whether new parties can enter the unit, which leads to difficult tradeoffs. The easier it is for current owners to block new entrants, the harder it is for these owners to sell their individual units. Accordingly, blocking rights become more important when turnover levels are low but not otherwise. The presence of common elements in PUDs also requires a more extensive governance structure that collects taxes from individual owners—​always in proportion to stated value of their individual units. Through its own board, a PUD must then spend the tax revenue to maintain the common elements within the system. These arrangements have to be set out in advance for all to know. Given that it is impossible to foresee all the variations that can take place, the initial articles for these PUDs contain a mix of structural arrangements and individual rights guarantees; it is understood that both internal governance procedures and external review, subject to the same general mix of self-​dealing and business judgment rules, apply to these organizations. There are in many cases fine points of difference in how these organizations are put together. Much depends on the overall wealth of the group, and potential divisions among association members. But here too, the ability of individuals to pick among competitive institutions leads to a form of efficient sorting that reduces the strains on the governance structure.

 

Optimal Constitutional Structure    9 And throughout it all, the strong evidence that these arrangements work is that they were imposed by a single owner on all the original purchasers into the venture. Subsequent purchasers have notice of the various rules, and they too take subject to its restrictions, allowing the form of governance to remain relatively constant even though membership changes. As in all cases, stability is not prized above all other goods. Extrinsic changes in circumstances can trigger the possibility of making structural changes by amendment, but almost always by supermajority vote by shares.

1.2  The Constitutional Extension of the Private Models The question then arises how this voluntarist model can apply to real-world constitutions that are formed, without unanimity, in the rough and tumble of political life. The wrong response is to assume that, since these differences are profound, a fresh start is needed in order to answer the relevant questions. But before one jumps to an alternative format, it should be evident that many of the fixed features of successful constitutions are already adumbrated in these private law models. So why junk it for an unspecified alternative that does not carry with it any warrant of its own soundness? The better response here is to do otherwise, and ask how the salient differences in operation lead to making adjustments in the overall operation of the model. The first of these relevant differences has to do with the exit rights from the organization. In both corporations and PUDs, individual owners may sell their interests if they have disagreements with the course and direction of the venture, allowing other individuals whose preferences are more closely aligned with the collective to take their place. To, be sure, exit is not costless, because the seller of any share or a property interest has to find a buyer, which is relatively easy with publicly traded companies, but less so for individual PUD units. Real-estate transactions are generally harder to organize than the sale of shares, and matters are complicated because of the overlay of the governance structure. The exit becomes far more difficult from nations (but less so from states or cities), for now it is no longer possible to sell one’s citizenship rights or remain in the territory after exit takes place. And nothing compensates for the massive level of personal dislocation that comes from leaving family, friends, and associates behind. These complications do not justify shutting down the exit option, which for all its flaws sends a clear notice that all is not well at home, precisely because that option is so costly to exercise. Nations that limit exit send a clear signal that will make it more difficult for them to attract new entrants, or indeed, new investment. They will also create pockets of discontent at home on the part of those who are forced to stay against their will. The upshot is that, descriptively, these exit rights will gain in importance as the political situation deteriorates at home. Perhaps the most notable example of this was the massive departures from East

 

10   Richard A. Epstein Germany in 1961, chiefly of young and productive people, that led to the erection of the Berlin Wall that remained in place for twenty-​eight years (Klein, 2014). Even in less dramatic situations, the exit right exerts a powerful influence against the climate of government abuse, even though exit can never target any particular practice.7 To be sure, protecting the exit right does not allow people to shield their fixed assets from confiscation, so a well-​designed constitution will simultaneously back the exit rights with explicit property protections against confiscation, especially for immovable assets like land. Even if developers can avoid an oppressive jurisdiction, a landowner cannot. Thus as the stakes get higher, the exit right is less able to provide full protection against confiscation. But the situation is hardly unique to constitutional settings. In ordinary business contracts, parties have exit rights—​e.g., the right to reject defective goods upon tender, or withdraw from partnerships—​that offer cheap and easy, but incomplete protection. Typically, actions for breach of contract are needed in large cases to protect the expectation interest in lost profits. The second clear implication is that where exit rights are weak, substitute protections have to be strengthened. One such right is the right to vote in political elections and participate in public debate. These are of course rights that are protected in corporations and PUDs. In these contexts, the voting rights deviate from those in corporations and PUDs in one key facet. In political settings, voting rights are never calibrated to property interests, but instead follow a one-​person, one-​vote rule that leaves property holders exposed to risk of confiscation by selective regulation under majority rule. A voice in the deliberative process therefore becomes valuable for those who speak but also for those who only listen, which is why the standard rules of order never allow a majority vote to shut down debate.8 But deliberation is always a two-​edged sword, because it can also allow the majority to rally together in order to increase the risk of confiscation by showing how the interests of the many are aligned against the interests of the few. Historically, this risk of confiscation was one reason to impose property qualifications on voting rights. But that rule moves matters too far in the other direction, for persons without substantial property still have huge personal stakes in the outcome of political decisions. It therefore becomes even more imperative in the public setting to supply the same strong property protections routinely afforded property owners in PUDs, whereby property can be taken for public use, only on payment of just compensation, to offset the dangers of majority, or sometimes, even well-​positioned minority factions. The need for these property and procedural safeguards increases as the population becomes more heterogeneous. It is commonly observed that small countries like Denmark are able to have an extensive network of social services given the large commonalities of an ethnically unified population. That optimistic scenario, often upset by 7 

For discussion, see Epstein (1992). For these connections, see Hirschman (1970). Loyalty is not a mechanism. But the higher the level of loyalty, the lower the level of exit, and the less the need for voice. Essentially, the loyalty involves an overlap in interests that eases tensions, which is why “all for one, one for all” is such a key element in tightly knit units, especially in military contexts. 8 

 

Optimal Constitutional Structure    11 high levels of immigration in the best of times, is sorely tested by massive movements in refugees, as is all too evident in the debates over the mass exodus from Iraq, Syria, and other Middle Eastern countries to Europe. The higher level of political instability only strengthens the case to give both process and substantive protections to all citizens in the political order on matters of process, voice, exit, and property. These systematic differences, therefore, do not render irrelevant the forms of voluntary association. But they unambiguously point out the need to bulk up both the structural and individual rights protections. The next section explores how that can be done. The third issue is in a sense more intractable. The question of territory plays no role with respect to large corporations, and the size of even large PUDs is always small. But nations can be of all sizes. In some instances they are landlocked and in others they have access to the sea. In some cases, the smallness of the size precludes the use of territorial subdivisions, and in others they are so large that these divisions are strictly required. The clear lesson from this is that the kinds of variations that we should expect on this organizational dimension are larger, and that the defense of a classical liberal constitution may rule out some extreme alternatives but will on average not dictate the choice of these institutional arrangements. With these caveats, let us turn to the examination of the way in which classical liberal constitutions should emerge from their private law precursors.

1.3  Key Elements of the Classical Liberal Constitution The challenge now is how to translate the structural and substantive limitations into the constitutional context. Specifically, the issue here is how to construct a durable order that can obtain the assent of all its members by leaving them better off than they were before. To make the exposition tractable, I shall start with those safeguards that are needed in all types of government and then turn to those which become relevant only in the context of more complex federalist systems.

1.3.1 Separation of Powers and Checks and Balances The first safeguard in question involves the separation of powers between various branches of government, each of which is often given powers—​e.g., a presidential veto—​ to check the authority of the other branches. The usual tripartite division is among the legislative, executive, and judicial powers. The theory behind the separation is that no single person or group of individuals can be trusted to hold the reins of power, given the risk that they will divert public assets to their private ends. Consistent with the general view that greater procedural safeguards are needed in political settings, the separation

 

12   Richard A. Epstein of power usually involves two houses of the legislature that are to some extent set off against each other, either by the length of term in office or the mode of election. Just that system is found in the United States, where Congress includes both a House of Representatives and a Senate. These representatives are elected for different periods—​ two years for the House and six years for the Senate. The functions are also split up, so that only the Senate has to ratify treaties and give its advice and consent for high-​level appointments like ambassadors, judges, and senior administrative officials. The executive for its part is required to carry out the law, but is not in a position to make law on its own initiative. There is, just as there is in corporations and PUDs, a constant struggle to figure out what forms of administrative action fall within the purview of authority granted the executive under the charter, and there is no easy formula that works in the myriad of cases in which some discretion is necessarily required and those cases which represent a flouting disregard of the applicable law. This problem surely has greater traction in a presidential system than in a parliamentary one. The divisions are more sharply drawn in the former than the latter, and, as with so many questions of this sort, it is highly contested as to which system in principle works the best. But in dealing with matters of long-​term institutional stability the major question is not the proper treatment of modest disputes, but the larger terrain that is settled. “Ambition must be made to counteract ambition” (Federalist No. 51). These words of Madison are somewhat optimistic, as are most of the Federalist papers in their discussion of the curative effects of the “extended republic” (Federalist No. 10) on the dangers of faction. There are times in which ambitious people manage to work all too well together on a wide range of issues. But the art of government is always to minimize risks, because it is beyond the power of any set of legal institutions to eliminate it. Separation of powers is in general a good first step in this direction. And it is instructive that in places like Great Britain, the decline of the House of Lords as a check on the House of Commons has as one of its consequences a general expansion in the overall size and power of the state.

1.3.2 The Administrative State The issue of separation is further clouded by the inevitable rise of the administrative state, which, by way of analogy to corporate and PUD committees, tends to operate in distinctive spheres. The key element in this regard is the ability to constrain the delegation so that the agency does not go beyond the scope of its statutory authorization. It has proved notoriously difficult to develop any hard-​edged rule that constrains the runaway administrative body. Efforts in the United States to place strong prohibition on the creation of independent agencies—​i.e., those whose personnel cannot be removed by the president—​have proved in large measure a failure, as courts have mumbled incoherent words about how these agencies are either quasi-​legislative or quasi-​judicial. When forced to choose, they look quasi-​executive.9 But bodies of this sort must be used to 9 See Humphrey’s Executor v. United States, 295 U.S. 602, 624 (1935).

 

Optimal Constitutional Structure    13 attend to a wide range of functions, so that it is unwise to condemn generally all administrative agencies as unlawful (Hamburger, 2014). The key question in all cases is how they operate, which depends critically on two questions. The first of these is the nature and scope of the mission. It is generally possible to effectively delegate the task of determining the level of interference between two radio frequencies, just as it is possible to determine administratively the rules of the road on public highways. But it is quite a different matter to determine the composition of the traffic under a general mandate that looks only to “the public interest, convenience, and necessity,” a standard which invites administrative agencies to make substantive decisions on what kind of programming is most desirable. The more modest tasks are the only ones allowed under a classical liberal constitution. The latter are staples of the modern progressive diet.10 The second issue concerns the relationship between the agencies and the court. The modern trend in the United States has been to allow for high levels of administrative deference in interpreting the meaning of the terms of statutes under the highly influential decision in Chevron Inc. v. National Resources Defense Council.11 This deference persists in the face of a key provision of the Administrative Procedure Act of 194612 that gives courts the power to decide questions of law—​a section that was neither cited nor discussed in Chevron. The best way to think of the division of power between courts and independent agencies is to view the agency as if it were a combination of a trial judge and jury, whose decisions should be reviewed de novo on all questions of law but should be overturned on matters of fact only in the case of a clearly erroneous decision.13 On this view, the greatest danger is any system that combines the prosecutorial and adjudicative function in the same agency, regardless of its substantive mandate.

1.3.3 The Judicial Role These difficulties with the administrative state point to the thorny issue of the proper role of the judiciary in a system of divided powers. All corporations and PUDs have complex internal grievance procedures, but usually there is at least a review of internal actions that can be obtained in courts that are outside and independent of the entities themselves. On matters of governance, there is no way to have a judiciary that is outside the governance system, and so the question is how it should best be done. The traditional accounts of separation of powers by writers like John Locke (1690) and Baron Montesquieu (1748) did not lay much stress on the judiciary, thinking that its key 10 See National Broadcasting Co. v. United States, 319 U.S. 190, 215–​216 (1942) (stating the Federal

Communications Act “does not restrict the Commission merely to supervision of the traffic. It puts upon the Commission the burden of determining the composition of that traffic.”). For a defense of the opposite position, see Hayek (1944: 82). Needless to say, the Federal Communications Commission could never develop coherent rules for frequency allocation by administrative order. 11  467 U.S. 837 (1984). 12  5 USC § 706. 13  For discussion, see Epstein (2011).

 

14   Richard A. Epstein function was to resolve particular disputes in accordance with the law of the land, given as it was by the legislature. On this view, one major constitutional design feature is that the political branches can never cut people out from access to the courts, a sentiment that is clearly expressed in the Magna Carta in Clauses 39 and 40,14 and one so influential in modern countries that pride themselves on the creation of an independent judiciary to which all shall have access. But access on what terms? Most classical liberal systems start with the simple view that access is open for people to challenge the decisions of the state in ways that force it to follow the letter and spirit of the law in individual cases. In that sense they surely are entitled to receive what is due to them by statute or regulation. But there is the further question of whether these procedures meet some higher law standard (Corwin, 1928), which British writers tellingly called “natural law” (Schwartz, 1953). In all cases, moreover, the well-​run system tries to secure accuracy in adjudication by insisting that disputes be tried before neutral judges, under predetermined norms that are known and knowable to all persons. This set of procedural protections is part of the rule of law,15 which under the standard account is a virtue of all legal systems, not only those that embody classical liberal principles, and are found, for example, in Article 6(1) of the European Convention on Human Rights.

1.3.4 Taxation, Regulation, and Expenditure The next question is what powers of taxation and regulation should be exercised within a classical liberal constitution. On this point, there is a surprising unanimity of writers like Aristotle, Locke, Smith, and Hayek in favor of a flat tax for raising general revenues (Epstein, 2002), and the purposes to which these ought to be put—​i.e., the creation of public goods, which when provided by one must be provided by all (Olsen, 1965). The basic argument for these two propositions starts at root with the greater need to constrain discretion in political organizations than in voluntary ones. Using a flat tax on a broad base of income meets that objective because it militates strongly against selective exclusions for one source of income rather than another. In addition, using the flat tax as the sole source of general revenues eliminates the possibility that ad hoc taxes—​ like transfer taxes—​can be imposed, for example, at the time of death on estates above a certain size, which quickly leads to a regime of gift taxation that is needed to prevent circumvention of the tax. The flat tax is consistent with letting the government raise any level of revenue that it needs for general purposes, and thus induces a kind of political watchfulness as people become more reluctant to impose taxes on others if they know 14 

Magna Carta, Cl. 39 (“No free man is to be arrested, or imprisoned, or disseised, or outlawed, or exiled, or in any other way ruined, nor will we go against him or send against him, except by the lawful judgment of his peers or by the law of the land”); Cl. 40 (“We will not sell, or deny, or delay right or justice to anyone”). 15  For a leading exposition, see Fuller (1964).

 

Optimal Constitutional Structure    15 that they must impose a like tax upon themselves. The flat tax is surely preferable to a head tax in which people are charged a fixed amount for services, regardless of income, which cannot work in cases where the tax exceeds the income of the person taxed. The flat tax, moreover, has a strong, but constrained, built-​in redistributive component, even when the revenues raised may only be spent on the provision of classical public goods. The value that people receive from certain collective goods like clean water or police protection decline with income. So it is the case that conspicuous public improvements financed by the rich usually work to the benefit of the larger population, as was surely the case with public sewers and vaccination programs which pushed up life expectancy in all segments of the population. The classical liberal constitution takes a parallel view on the expenditure side, by limiting these expenditures to the provision of standard public goods. That notion was embedded at the federal level in the United States Constitution by the provision that the proper objects of taxation were limited to “to pay[ing] the debts and provid[ing] for the common defense and general Welfare of the United States.”16 Note that there are no commas in this series, which is organized to supply public goods, not of the individual states, but of the United States as a whole. The clear purpose of these provisions was to limit the objects of taxation to exclude transfer payments among individuals. Indeed, the concern was so great that the arcane rules of direct taxation were intended to make sure that rich states did not have to pay more for public improvements than poor states, from which they derived only equal benefits.17 To use the corporate analogy, no transfer payment between shareholders would advance the general welfare of the corporation at large. By restricting these transfer payments, the conscious effort was to limit taxation to the provision of public goods, for which market mechanisms are in general inadequate. As ever, the pressures on this view arose almost instantaneously, so that the words “of the United States” have commonly been stripped out from the passage in order to give a more flexible definition of general welfare for which transfer payments are in fact permissible, as early as 1791 by Hamilton.18 But within the classical liberal theory the point of collective action is to overcome holdout and coordination problems, not to run a welfare or transfer system. At least under the 1787 understandings of the United States Constitution, such a system was a function of the states, not the federal government. That narrow view of the subject tends to be widely rejected today throughout the world as it was rejected in the United States in 1937 by cases that attached a broad reading to the provision in question, with the consequent uncontrollable expansion of the federal government.19 Leaving welfare to the states constrains 16 

US Const. art. I. § 8, cl. 1. See also US Const. art. I, § 9, cl. 4 (“No Capitation, or other direct, Tax shall be laid, unless in Proportion to the Census or enumeration herein before directed to be taken”). For explication see Meese III, Spalding, Forte (2012). 18  See, e.g., Hamilton (1791). 19 See Steward Machine Co. v. Davis, 301 U.S. 548 (1937) (upholding broad definition of power to tax and spend); National Federation of Independent Business v. Sebelius, 132 S.Ct. 2566 (2012) (using broad definition to sustain the taxes/​penalties under the Affordable Care Act). 17 

 

16   Richard A. Epstein these forces given that exit options are a useful, if imperfect, check on government power. There is a second key correlative to the taxing power, which is that whenever possible, general revenue taxes should not be used to supply local benefits when it is impossible to impose local taxes.20 Thus the construction of a new road in a cul-​de-​sac provides benefits chiefly to the local citizens, at which point the special assessment prevents the externalization of costs on outsiders. And true to the basic condition of governance, these are usually only after some kind of election that requires supermajority support from local residents whose stake is measured by the market value of the land, or when appropriate, its front footage. In these cases too, the general point here is to seek to provide for local collective goods without introducing the political instability that arises whenever transfer payments are allowed. The rule here is capable of extension to other situations, whereby, for example, user fees can be imposed to maintain the costs of roads that might otherwise come from other sources. And by the same logic it is a mistake to turn these taxes over to the Treasury, where they undermine the exclusive use of general taxes to support common expenses.

1.3.5 Standing The next question that arises is how to enforce this system of taxation, and indeed all other limitations on the use of enumerated government powers. The correct answer in this case is to allow individual taxpayers or citizens of the applicable polity to bring actions against the entity if the taxation in question is directed for purposes outside of those allowed for under the constitution. The common issue in this case is whether individual citizens have standing to raise the challenge. The notion of standing as a limitation on judicial power is very old, as the term derives from the Latin locus standi, which asks whether a given person has a sufficient stake in a case to be able to bring suit. The endless confusion of the doctrine stems in large measure from the inability to understand the different logic that applies to ordinary tort or contract actions on the one hand and the actions of individual shareholders or citizens against their private or public government on the other. In the first instance, it is clear that standing for physical injury or breach of contract is sensibly confined to those persons who were either tort victims or contracting parties. The point is that small losses that always happen to other individuals are too numerous and too difficult to value to be worth the public investment to correct them, especially when the one party who bears the brunt of the wrong is able to maintain an action. But the situation is quite different when an individual shareholder, for example, seeks to enjoin corporate actions as ultra vires, for here the correct response starts with the need to address a firm-​specific public goods problem. Someone has to be able to sue or otherwise the structural wrong goes uncorrected. 20 

For discussion, see Diamond, S. (1983). For the breakdown of the system in the United States, see Louisville & Nashville R.R. v. Barber Asphalt Paving Co., 197 U.S. 430 (1905).

 

Optimal Constitutional Structure    17 This tension was made quite vivid in the 1923 case of Frothingham v. Mellon,21 where a unanimous United States Supreme Court held that neither an ordinary citizen nor a state could bring suit to enjoin expenditures made by the federal government under the Maternity Act.22 Similar suits were allowed in England23 and in the various states,24 but the United States Supreme Court held that the capacious definition of the judicial power, which did not contain the term standing in it, imposed, on grounds of separation of powers, limitations on the power of ordinary individuals without some special interest to maintain the suit. The odd point here is that the larger and more diffuse the population, the greater the need to use these mechanisms to prevent breaches of the government power. But once these suits are denied, states like Massachusetts and individuals like Mrs. Frothingham have no choice but to accept the payments themselves, lest they be subject to the tax, the entire benefit of which goes to people in other states. There is little doubt that restricting standing to sue helps to expand the power of government by increasing the returns to transfer payments.

1.3.6 Federalism Complications The efforts to impose structural limits on government power take on special urgency in a federalist system, where it is now necessary to police the boundary between national and subsidiary units, i.e., states or provinces. The constitution has to solve a collective action problem, without becoming the source of that problem itself. It was just this challenge that the founders at the United States 1787 Constitutional Convention confronted and addressed. The founders were trying to form a more durable union—​one with a president, with a Congress with the power to tax, to form an independent judiciary, and to act decisively in foreign affairs—​ out of the ashes of the Articles of Confederation that had been adopted in 1781. The question is how to secure the win/​win proposition when states are different in size, power, and wealth. To see the measure of the disparity it is important to note that the 1787 map did not have two separate states of Virginia and West Virginia, but one large state that encompassed both, and that in a union that occupied a far smaller territory than it does today. In any federalism context, it is in general dangerous to allow for concurrent jurisdiction over a wide range of economic activities, because that approach means that the stricter regulator will in each case be able to set the standard of conduct.25 It is therefore necessary to find ways to partition regulation between sovereigns when it is impossible to draw any physical boundary between the two levels of government. In general, 21 

262 U.S. 447 (1923). 67th Cong., Sess. I, Ch. 135, p. 224–​6. 23  See Wade (1961). Under the traditional English practice, no individual had standing to sue a public body, but had to ask the attorney general for his assistance, which was freely granted. 24 See Crampton v. Zabriskie, 101 U.S. 601 (1880). 25  For discussion, see Gardbaum (2007). 22 

 

18   Richard A. Epstein these jurisdictional boundaries work best when they have the same hard-​edged quality associated with physical boundaries between neighboring parcels of land or neighboring states. The earliest historical example of enumerated powers is found in the United States, but other federal nations, like Australia and Canada, have to address the same problem, albeit on a somewhat smaller canvas. More recently, the rise of the European Union involves a system of enumerated powers, which in many senses is the inverse of that which is found in the United States insofar as foreign affairs powers in the United States are confided in the federal government but are left to individual member states in the European Union. Some sense as to how this approach works is found by the list of enumerated powers in Article I, section 8 of the United States Constitution. Most of these powers are particularistic—​the ability to set naturalization laws, to regulate patents and copyrights, and to raise armies is not capable of great expansion. But the Commerce Clause proved to be the tool that undid the system of enumerated powers, which was put in place to limit the scope of bargaining and haggling in the federal legislative arena. The basic provision is simplicity itself: “Congress shall have power to … regulate commerce with foreign nations, and among the several states and with the Indian tribes.” This clause was an ingenious effort to create a partition by which the federal government could regulate the movement of goods and services across state lines, leaving regulation of activities like manufacture, agriculture, and mining to the individual states.26 This doctrine does not tend to apply to unitary states on the ground that it is simply too dangerous to announce that there are certain subject matter areas that are just outside the scope of government power. But the doctrine of enumerated powers has a long and complex history in dealing with federal systems where it is important to delineate the respective spheres of the national and the respective subunits. The difficulty with the US position, ironically, was that it gave too much power to the federal government. Namely, there was nothing about the articulation of the power that did not allow its use to protect the influence of private cartels whose activities could be bolstered by the official government support that it received under the various agricultural adjustment acts. The more sensible approach in this case is to limit the power of the Congress to keep open the arteries of commerce and trade among the several states so that none could step up barriers to the free movement of goods and services across state lines. Within the United States that objective has been achieved partially by the judicial creation of the dormant commerce clause jurisprudence, under which it is said that the mere grant of a federal power to regulate certain areas necessarily precludes the exercise of state power that blocks competition. That judicial effort to create a national market has had enormous positive consequences in the United States, and the same can be said of the efforts in the European Union to have free movement of goods and services across state lines. But in both the United States and the European Union, the power 26  See, e.g., Gibbons v. Ogden, 24 U.S. 1 (1824); United States v. E.C. Knight, 156 U.S. 1 (1895). For defense of that position, see Epstein (1987). For its repudiation, see National Labor Relations Board v. Jones & Laughlin Steel Co., 301 U.S. 1 (1937); Wickard v. Filburn, 317 U.S. 111 (1942).

 

Optimal Constitutional Structure    19 of the central government to impose direct regulation on business behavior is inconsistent with classical liberal principles to the extent that it bolsters up state cartels. That endorsement of government-​run cartels was explicitly ratified in the United States in the 1943 decision in Parker v. Brown,27 which held that a state-​run raisin cartel was insulated from the Sherman antitrust law under which that same cartel, run privately, would have been per se illegal. The decision was all the more inexplicable because the cartel in question internalized gains to California by imposing costs on the vast majority of buyers outside the state. As noted earlier, there are surely dangers of state abuse of power that the exit right cannot control, and for these activities a federal guarantee of the rights of individuals against state governments could provide the needed corrective. This was done in part under the Fourteenth Amendment, which gave the federal government the power to restrict the states from abridging the privileges or immunities of the citizens of the several states, or to deprive any person of life, liberty, or property, without due process or law, or to deny any person the equal protection of the laws. These guarantees are not limited to economic areas where they point strongly to the protection of competitive over monopoly institutions. There is, however, little doubt that the rise of the progressive ideology in the United States as elsewhere has reduced the effectiveness of these provisions to keep states at bay. It is, to be sure, possible to devise a structural constitution on the key issues of taxation and regulation. Indeed, the American Constitution through the adoption of the Reconstruction Amendments after the Civil War (1861–​5) came close to implementing that scheme before the good work was gutted by a succession of misguided decisions.28

1.4  Protection of Individual Rights 1.4.1 The Centrality of Private Property The last portion of a classical liberal constitution deals with the explicit protection of individual rights, which should be more or less the same across different societies and times. It has often been observed that property is the guardian of every other right, and for good reason (Ely, 2007). Unless people can be secure in their homes, businesses, and possessions, they must spend all their time worrying about their protection from the predations of others, which reduces the time available for them to spend in gainful pursuits. Hence the basic theory of the social contract is that all individuals surrender some portion of their property to a central government that uses these resources to provide protection for the property interest that private citizens still retain. The scope of 27 

317 U.S. 341 (1943). See, e.g., The Slaughter-​House Cases, 83 U.S. 36 (1876) (narrow reading of privileges and immunities); United States v. Darby, 312 U.S. 100 (1941) (sustaining overtime laws against a variety of constitutional challenges). 28 

 

20   Richard A. Epstein the property in governmental affairs is the same as it is for PUDs. It must embrace not only the right to exclude others, but also the right to enter, use, develop, and dispose of property without laboring daily to retain possession of what they own. The benefits of secure property rights are not confined solely to commercial or economic ventures, but are equally important for family, religious, charitable, scientific, and social activities. At the same time, there are important social interests that can be satisfied only through public efforts, for any complex political organization like any PUD needs to have its collective and common elements open to all for the benefit of transportation and communication, which are funded by assessments that are, of course, the direct private analog to government taxation. Even beyond that, the maintenance of public institutions requires that governments acquire land for a variety of government purposes, including administrative, political, military, educational, and other uses standard to all states. The question then is how best to secure the proper division between these public and private components.

1.4.2 Common Property, Private Property, and Government Takings The first step in this inquiry is to recognize, as has been done from Roman times, that certain resources—​rivers, seas, beaches, and air—​are held in common for all persons,29 and thus do not have to be acquired by the state from private individuals because they are not, and never have been, subjected to private ownership. Indeed, the function of the government in these cases is to make sure that these essential facilities are not conveyed to private parties where they could easily lose their public character. Hence the development of public trust doctrines that are intended to keep in public solution those assets that must be held in public hands, while allowing for the distribution of those excess public assets so long as the state does not give them away without receiving fair value in exchange: “Nor shall public property be given to private use, without just compensation,” under a “givings” clause, is one way to put the point.30 The point here is of exceptional importance because under this view it is, for example, inappropriate for any government to give monopoly power to any private party except when strictly necessary for the maintenance of public welfare. It follows therefore that governments exceed their proper powers when they create monopolies in labor or agricultural markets, or private subsidies to some groups that give them a competitive advantage over another. Yet by the same token, there is nothing in the classical liberal constitution that blocks the application of standard competition or antitrust law with respect to private monopolies, especially cartels. Neither is there anything that prevents the rate regulation of 29 

See Justinian Institutes Book II, title I. See Epstein (1987c). For the development of the public trust doctrine, see Cohen (1992). For an expansive reading of the doctrine, see Sax (1970). For a discussion of the key case of Illinois Central Rd. Co. v. Illinois, 146 U.S. 387 (1892), see Kearney and Merrill (2004). 30 

 

Optimal Constitutional Structure    21 natural monopolies such as common carriers and public utilities, so long as they continue to receive a risk-​adjusted competitive rate of return for their efforts.31 The classical liberal constitution accepts the strong proposition found in the competition law of the European Union and the antitrust law of the United States that the welfare implications of pure competition are positive and those of state-​backed monopolies are not. It therefore has to restrain the power of government not only when it regulates, but also when it gives away various property, permits, privileges, and licenses to private parties in ways that upset that competitive balance.32

1.4.3 Public Use, Just Compensation, and the Police Power The converse side of the problem is of even greater importance. The general solution accepted in one form or another everywhere is that the government may take private property for public use, so long as it provides just compensation for what it has taken. The logic of this arrangement is as follows. The purpose of some public use limitation is to confine the ends for which government may operate in order to impose one decisive limit on its operations. That constraint is lost when the “public use” language in this context is given the same broad reading that it receives too often in taxing cases.33 Yet even when the public use limitation is satisfied, just compensation is key. Whether one thinks of the matter in terms of intrinsic fairness or economic efficiency, allowing government simply to take private property is an open invitation to faction and intrigue. That rule is most evident with the outright and permanent transfer of property from one party to another, for which the requirement of compensation places a powerful constraint on the willingness of government to take, when they know that they must raise taxes to finance the venture. The constraint can be weakened, of course, if the government can use general revenues to support takings for particular persons, or, conversely, use taxes targeted on specific persons to provide general benefits. So both sets of constraints operating in tandem are likely to work better. It does not follow, however, that cash compensation need be made in all circumstances, for that rule would gut virtually every sensible system from a general income tax to a statutory recordation requirement for ordinary land transfers. In these key situations, it is important to recognize what Oliver Wendell Holmes, Jr. called “the average reciprocity of advantage,”34 which is a cryptic way of noting that comprehensive forms of regulation impose costs, by taking partial interests in property to which they return benefits by the like limitations on the property use by others. There is, of course, no a priori way of noting whether the burdens per parcel and benefits net out, without doing 31 

For discussion, see Duquesne Light v. Barasch, 488 U.S. 299 (1980). For extensive discussion, see Epstein (1993). 33 See Kelo v. City of New London, 545 U.S. 469 (2005). For a balanced exposé, see Somin (2015). 34 See Pennsylvania Coal v. Mahon, 260 U.S. 393, 415 (1922). 32 

 

22   Richard A. Epstein at least some empirical evaluation, but there are at least some signposts along the way. Where the initial parties are all in parallel positions and the incidence of the regulation or rule is similar for all, each will vote for the public good when he or she votes his private interest. But if there is a disproportionate burden or benefit, it is a good sign that some compensation is owing so that the winners cannot externalize their costs on losers. The best single-​sentence statement of the ideal is: “The Fifth Amendment’s guarantee that private property shall not be taken for a public use without just compensation was designed to bar Government from forcing some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.”35 This proposition, often stated as one of intrinsic fairness, has a desirable economic incentive effect. The common question everywhere is the extent to which this logic applies to government regulations that are short of outright dispossessions, but which impose substantial restrictions on either the use36 or the disposition37 of the property in question. The current legal view draws a strong distinction between these two kinds of government intervention, and allows compensation for legal restrictions on sale or use only in rare circumstances tantamount to a complete wipeout of economic value.38 But why? The actual line between the two classes of cases is obscure. Thus one could describe a landmark preservation statute as one that restricts development or as a taking of air rights, a well-​recognized property interest in every legal system. The political risks that lead to protections of outright taking are active here, and the vaunted separation between takings and regulations is one of the great miscalculations of our time. To be sure, many general regulations are protected by the implicit in-​kind benefit rule, where the benefits have to be empirically shown and not just postulated on the ground that good governments never do bad things. It would, however, be a mistake to think that whenever the government restricts the private use or disposition of property it must pay, even if no compensation, express or implicit, is provided. To be sure these are takings cases, but they may in principle be justified to protect against wrongful conduct, most notably in the form of nuisances, by other individuals. This issue presents the exact same difficulties that are found whenever remedies are needed in private disputes between ordinary citizens. In all cases, it is necessary first to identify those ends that are properly subject to sanctions. That done, the next question is whether the means chosen are appropriate to deal with the peril in question. A complete analysis of the police power requires a discussion of both parts of the problem.39 As is always the case, the key for understanding these two situations is to treat the government (like the PUD board) as though it stands in the shoes of the private 35 

Armstrong v. United States, 364 U.S. 40, 49 (1960). Penn Central Transportation Co. v. City of New York, 438 U.S. 104 (1978). 37  Andrus v. Allard, 444 U.S. 51 (1979) (citing Penn Central and finding no taking for blocking right of sale for eagle feathers). 38  Lucas v. South Carolina Coastal Council, 505 U.S. 1003 (1992). For the international perspective on these regulatory takings, see Alterman (2011). 39  See Epstein (1985), discussing ends in c ­ hapter 9 and means in c­ hapter 10. 36 

 

Optimal Constitutional Structure    23 parties whom it represents in any dispute. If these parties could get some remedy for themselves, the government should be able to do so as well. But when the private party could only act upon payment of compensation, the government cannot under the guise of the police power use its majoritarian powers to effectuate that change without giving compensation to the aggrieved landowner. In order for this system to work, it is clear that there must be some limits on the police power, lest the requirements of just compensation and public use be wholly circumvented. And within the classical theory those limitations were found in the familiar police power “quartet,” which coves health, safety, general welfare, and morals. These carry through the classical liberal theme that the government should not be allowed to use its purported police power justifications for paternalist or anticompetitive ends. This happens with much of progressive or social democratic legislation that is designed to protect or entrench public or private monopolies, including, for example, the use of antidiscrimination laws (a proper counterweight to monopoly) in competitive markets.40 The limited definition does not allow for these actions to take place, and in so doing makes sure that the protection afforded to individual rights dovetails with the structural protections dealing with such matters as separation of powers, limited taxation, and enumerated powers. On the question of means, the first distinction is between harms that were completed and those that are only threatened. In the private law, damages are usually the appropriate remedy, for which the public law analogue is the fine. In these cases the difficulties deal with matters of proof, not high questions of principle. In practice, most cases of discrete or “special” harms are in general best remedied by the aggrieved party personally, so that fines are often used for low-​level damages that affect a large population, where the money in question is usually put into the public treasury (often to deal with the underlying problem) and not distributed to aggrieved individuals.41 It is, however, always more difficult with respect to threatened harms, because the uncertainty in the underlying situation necessarily invites two kinds of error. The first type is that the remedy stops activities, which, if allowed to run their course, would have resulted in no harm. The second type of risk is that the remedy allows activities to take place, after which harm takes place. It is not possible to drive both types of error to zero simultaneously, so the question is how to minimize their joint effect. It is important to note that these two types of loss are not perfectly symmetrical. If the government refuses to permit activities, then the process comes to an end. But if in the alternative it allows the activity, the party in question knows that it is still subject to severe sanctions if the actions taken turn out bad, and in cases of dangerous activities is required to pay damages in the event of harm, or to purchase insurance or permit inspections of ongoing activities. Given this balance, the usual test in most private law systems is to reserve 40 

See Epstein (2014c). For an early statement of the line between public and private nuisances still observed today, see Anon. Y.B. Mich. 27 Hen. 8, f. 27, pl. 10 (1535). 41 

 

24   Richard A. Epstein injunctions for actual or imminent harms.42 But modern permit systems are often far more lax and block activities until the regulated party can negate to a high degree of certainty the possibility of any number of remote risks, which vastly slows down productive activities often for at most marginal gains. The classical liberal system subjects the government to higher standards of review for permitting than the dominant progressive or social democratic systems, especially in environmental or zoning settings. The scheme thus far described looks as though it were limited to property rights, but it covers the full range of economic liberties, dealing with such matters as government drug approval, or occupational licensing, and other social arrangements, including religious and speech rights. Religious and speech rights are also not absolute, for they are always subject to proper police power limitations, but these must be carefully circumscribed so that they do not cover protecting people from criticisms, as with political protests, or exposure to offensive religious practices, such as polygamy, which is routinely made criminal even as gay marriage obtains protected legal status. A classical liberal constitution does not pick and choose favorites, but protects all private actions except to the extent that they pose risks of force, fraud, and monopoly. Modern constitutions go far beyond this model, but their broad acceptance has usually been defended not on substantive grounds, but on the belief that these issues are too difficult for courts to deal with—​except in those cases where they choose to do so. But all this is an illusion. There are of course tricky empirical issues in dealing with various questions of government structure and individual rights. But the hard cases at the margins of the classical liberal constitution have no warrant to switch to a vastly broader conception of state power for which no systematic justification has yet to be offered.

1.4.4 Unconstitutional Conditions The most recent development in dealing with individual rights stems from the growth of the permit system, which asks the following question: To what extent can the government condition a grant of a permit to engage in some activity on the willingness of the permittee to follow certain conditions laid down by the government? The problem gains its urgency because the government enjoys essentially a monopoly power in its ability to issue permits; the party who is aggrieved by the refusal to receive a permit cannot go off to some other agency in order to receive the needed permit. Indeed the problem is even more severe than this because it is commonplace in land use, environmental law, transportation, and many other legal regimes requiring a given party to receive multiple permits—​each with its own blocking position—​before engaging in certain activities. In dealing with this issue, the appropriate private comparison is the common carrier or public utility who since the earliest times did not have the absolute right to accept or

42  For exhaustive accounts, see Laycock (1990); Gergen, Golden, and Smith (2012). Note that in the United States, it has become more difficult to get injunctive relief in patent cases. See eBay, Inc. v. MercExchange, L.L.C, 547 U.S. 388 (2006). This case was criticized in Gergen, Golden, and Smith (2012).

 

Optimal Constitutional Structure    25 turn down business on whatever terms they saw fit but must do so on fair, reasonable, and nondiscriminatory terms, including rates.43 It is just that concern that animates the doctrine of unconstitutional conditions, which says that even when government may take or exclude a given person from a certain activity, it must not do so on impermissible conditions (Epstein, 1993). The early opponents to this doctrine claimed that the greater power (to exclude all) necessarily carried with it the lesser power (to select those who could be admitted), and on what conditions: “The right to absolutely exclude all right to use necessarily includes the authority to determine under what circumstances such use may be availed of, as the greater power contains the lesser.”44 But this argument is a painful non sequitur because the greater power is the ability to select who is subject to a rule and who is not. Indeed, the application of an all-​or-​nothing rule is just a nondiscrimination rule that limits discretion by forcing government officials to apply the same rules to their friends that they do to their enemies. Rightly understood, the ultimate question then is what conditions can be attached to various grants and permits.45 Clearly, it would not be permissible to allow the government to exclude persons from the public road unless they waived their protections against unreasonable searches and seizures, to participate in political activities, or even to act as a common carrier and take all customers in what otherwise would have been a private carrier. All individuals might accept these waivers, and so all political and many economic activities could be shut down. But by the same token, it is surely permissible for the government to condition entry on the public roads on passing a driving test or on agreeing to litigate all disputes arising from road accidents within the jurisdiction. The former conditions look like efforts to deviate from competitive market solutions, while the latter conditions are intended to beef up the security of all individuals under the traditional police power definitions. Although this issue of permissible conditions on public access is complex and heavily litigated, the short answer to the problem is that conditions are permissible to the extent that they target behaviors, which, if they occurred, would be subject to legitimate government sanctions but not otherwise. So accident prevention meets this standard, but anticompetitive conditions do not. But note that if the local constitution expands the class of permissible government activities under the police power, then it will necessarily expand the type of conditions that can be added to zoning permits, for example, so that the builder of a new building could be 43 

Allnut v. Inglis, 104 Eng. Rep. 206 (K.B. 1810). For a more detailed account of this critical case, see Epstein (1998). 44  Davis v. Massachusetts, 167 U.S. 43, 48 (1897). 45  For leading cases, see Nollan v. South Carolina Coastal Commission, 483 U.S. 825 (1987) (rejecting the requirement that a landowner must surrender a lateral easement across beachfront property to get a building permit); Dolan v. City of Tigard, 512 U.S. 374 (1994) (holding that the government’s conditioning permits to provide for offsetting benefits, or to control harmful externalities under tests of legitimate “nexus” for ends, and “rough proportionality” for means); Koontz v. St. Johns River Water Mgmt. Dist., 133 S.Ct. 2586 (2013) (failing to clarify which conditions are applicable in cases of potential environmental harms).

 

26   Richard A. Epstein required to bear the costs of repairing the old train station or school. At this point, the unity of all approaches becomes clear. A classical liberal system of taxation, as noted above, restricts the uses of special assessments to the case of special benefits. The use of conditional permitting undermines that connection, by allowing public projects to be financed exclusively by new arrivals. The new standard thus opens up the permitting process to serious political struggles. There is no doubt that these ad hoc adjustments have become more popular. On the good side, they allow governments to force parties to internalize harms before they occur; on the bad side they encourage rent-​seeking, which the optimal constitutional order seeks to forestall, here as in other contexts.

References Alterman, R. (2011) Takings International: A Comparative Perspective on Land Use Regulation and Compensation Rights. Chicago, IL: American Bar Association Publishing. Berle, A. A. and Means, G. C. (1932) The Modern Corporation & Private Property. New Brunswick, NJ: Transaction Publishers. Butler, H. N. (1986) “General Incorporation in Nineteenth-​Century England: Interaction of Common Law and Legislative Processes,” International Review of Law and Economics, 6: 169–​88. Cohen, L. R. (1992) “The Public Trust Doctrine: An Economic Perspective,” California Western Law Review, 29: 239–​76. Corwin, E. S. (1928) The “Higher Law” Background of American Constitutional Law. Cambridge, MA: The Harvard Law Review Association. Diamond, S. (1983) “The Death and Transfiguration of Benefit Taxation: Special Assessments in Nineteenth-​Century America,” Journal of Legal Studies, 12: 201–​40. Ely, J. W. (2007) Property as the Guardian of Every Other Right: A Constitutional History of Property Rights. Oxford: Oxford University Press. Epstein, R. A. (1985) Takings: Private Property and the Power of Eminent Domain. Cambridge, MA: Harvard University Press. Epstein, R. A. (1987a) “Covenants and Constitutions,” Cornell Law Review, 73: 906–​27. Epstein, R. A. (1987b) “The Proper Scope of the Commerce Power,” Virginia Law Review, 73: 387–​455. Epstein, R. A. (1987c) “The Public Trust,” Cato Journal, 7: 411. Epstein, R. A. (1992) “Exit Rights under Federalism,” Law and Contemporary Problems, 55: 147–​65. Epstein, R. A. (1993) Bargaining with the State. Princeton, NJ: Princeton University Press. Epstein, R. A. (1998) Principles for a Free Society:  Reconciling Individual Liberty with the Common Good. New York: Perseus. Epstein, R. A. (2002) “Can Anyone Beat the Flat Tax?” Society of Philosophy and Policy, 19: 140–​7 1. Epstein, R. A. (2011) Design for Liberty: Private Property, Public Administration, and The Rule of Law. Cambridge, MA: Harvard University Press. Epstein, R. A. (2014a) The Classical Liberal Constitution:  The Uncertain Quest for Limited Government. Cambridge, MA: Harvard University Press.

 

Optimal Constitutional Structure    27 Epstein, R. A. (2014b) “Redistribution within Collective Organizations: What Corporations, Condominiums, and Unions Tell Us about the Proper Use of Government Power,” New York University Journal of Law & Liberty, 8: 280–​315. Epstein, R. A. (2014c) “Public Accommodations under the Civil Rights Act of 1964: Why Freedom of Association Counts as a Human Right,” Stanford Law Review, 66: 1241–​91. Fuller, L. L. (1964) The Morality of Law. New Haven, CT: Yale University Press. Gardbaum, S. (2007) “The Breadth v. the Depth of Congress’s Commerce Power: The Curious History of Preemption during the Lochner Era,” in R. A. Epstein and M. S. Greve (eds) Federal Preemption: States’ Powers, National Interests. Washington, DC: AEI Press, 48–​78. Gergen, M. P., Golden, J. M., and Smith, H. E. (2012) “The Supreme Court’s Accidental Revolution? The Test for Permanent Injunctions,” Columbia Law Review, 112: 203–​49. Hamburger, P. (2014) Is Administrative Law Unlawful? Chicago, IL: University of Chicago Press. Hamilton, A. (1791) Report on Manufactures 1011, 1012, http://​www.constitution.org/​ah/​rpt_​ manufactures.pdf. Hayek, F. (1944) The Road to Serfdom. Chicago, IL: University of Chicago Press (1976 reprint edition). Hirschman, A. O. (1970) Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press. Jensen, M. C. and Meckling, W. H. (1976) “The Theory of the Firm:  Managerial Behavior, Agency Costs, and Ownership Structure,” Journal of Financial Economics, 3: 305–​60. Kearney, J. D. and Merrill, T. W. (2004) “The Origins of the American Public Trust Doctrine:  What Really Happened in Illinois Central,” University of Chicago Law Review, 71: 800–​931. Klein, C. (2014) “10 Things You May Not Know about the Berlin Wall,” History in the Headlines, http://​www.history.com/​news/​10-​things-​you-​may-​not-​know-​about-​the-​berlin-​wall. Laycock, D. (1990) “The Death of the Irreparable Injury Rule,” Harvard Law Review, 103: 687–​771. Locke, J. (1690) The Second Treatise of Government. Manne, H. G. (1965) “Mergers and the Market for Corporate Control,” Journal of Political Economy, 73: 110–​20. Meese III, E., Spalding, M., and Forte, D. eds (2012) The Heritage Guide to the Constitution. Washington, DC: Regnery Publishing. Montesquieu (1748) The Spirit of the Laws. Olsen, M. (1965) The Logic of Collective Action. Cambridge, MA: Harvard University Press. Sax, J. L. (1970) “The Public Trust Doctrine in Natural Resource Law:  Effective Judicial Intervention,” Michigan Law Review, 68: 475–​566. Schwartz, B. (1953) “Administrative Procedure and Natural Law,” Notre Dame Law Review, 28: 169–​98. Shleifer, A. and Vishny, R. W. (1997) “A Survey of Corporate Governance,” The Journal of Finance, 52: 737–​83. Somin, I. (2015) The Grasping Hand: “Kelo v. City of New London” and the Limits of Eminent Domain. Chicago, IL: The University of Chicago Press. Wade, H. W. R. (1961) Administrative Law. Oxford: Oxford University Press.

 

Chapter 2

The Desi g n of C onstit u t i ons Tom Ginsburg

Constitutions have been a central topic for the economic analysis of law since Buchanan and Tullock (1961) introduced the discipline of constitutional political economy.1 Their concern was how to design higher-​order rules so as to optimize “ordinary” processes of decision-​making within groups, a question also of concern to the related discipline of public choice theory. These efforts followed in the footsteps of the founding fathers of the United States, who sought to apply reason to the problem of constitutional design, and created an enduring form that has spread to nearly every nation-​state. How do we create institutions that will prevent domination, enhance freedom, and facilitate growth and investment? These questions of classical political theory have benefitted from rigorous analysis using economic tools. From the outset, economic analysis has been deployed for both positive and normative ends. Buchanan and Tullock’s project was normative, but economic analysis provides a useful set of tools both to critique real-world constitutions as well as to analyze their attributes. Optimal design of constitutions in theory is rarely matched in practice, but this is no hindrance to understanding the form, duration, and impact of actual constitutions. This chapter first reviews the ends of constitutional design. It then offers a positive theory of constitutional bargaining which can be used to inform normative design questions. Ultimately, the question of whether particular institutions ought to be included in a constitution depends on the extent to which such texts make a difference, which itself is an empirical question that has been subjected to some scrutiny (Persson and Tabellini, 2003). The chapter concludes with a brief review of the empirical literature on constitutional design.

1 

For more on constitutional political economy, see Voigt in this volume and Van den Hauwe (2000).

 

The Design of Constitutions    29

2.1  Goals of Design From an economic perspective, the goals of constitutional design are several-​fold. First, constitutions are generally viewed as devices to minimize agency costs on the part of government. They do so by establishing structures and processes to minimize the probability that small minorities or interest groups can capture government on a permanent basis. This anti-​domination rationale is a central concern of political theory, and is one that has merit notwithstanding economic consequences. Yet it is also the case that political domination is likely to lead to economic cronyism, and to undermine broader transitions to thriving open-​access orders in which sustainable growth can occur (North, Wallis, and Weingast, 2009). Constitutional guarantees of freedom are seen as central for economic thriving. A related function of constitutions is to constrain intertemporal choice through precommitment. In this line of thought, constitutions function to constrain the future on behalf of the past, to take certain high-​stakes issues off the table and to channel political energies into ordinary politics. By limiting the stakes of government, in the sense of restricting the costs that winners can impose on losers, constitutions make government possible. An even more basic function of constitutions is coordination, first identified by David Hume (McAdams, 2015; Ordeshook, 1992; Weingast, 1997; Carey, 2000). Constitutional coordination problems range from very simple ones, such as who is to preside over the first session of the legislature, to very complicated ones, such as what constitutes a violation of, say, due process of law. While it may make little difference whether a constitution stipulates the minimum age for head of state to be 34, 35, or 36 years old, it is important that a constitution give some answer so as to avoid disagreement down the road, and whatever answer is chosen is likely to generate relatively little controversy. For any issue, large or small, intersubjective agreement is necessary to avoid social conflict. Weingast (1997) models a game in which a leader can exploit disagreement among subjects to dominate them; but providing a clear definition of the rules on paper can facilitate enforcement of constitutional norms. This coordinating function of constitutions is found in both democracies and dictatorships, as a positive matter. In addition to these broad ends, some economists emphasize particular examples of constitutional policies that are desirable to facilitate growth and economic activity. For example a set of secure property rights, and a system for enforcing contracts, are thought to be central from the perspective of the new institutional economics (North, 1991). Central bank independence and judicial independence are other policies that are often tied to security at the constitutional level. Others emphasize the risks of particular constitutional institutions. In their normative critique of the Constitution of Spain of 1978, for example, Brennan and Pardo (1991) fear the consequences of socio-​economic rights found in the document.

 

30   Tom Ginsburg In short, even setting aside more symbolic or cultural functions that are analyzed by other disciplines, economic theory provides several different goals of a constitution. Some provisions, such as a separation of powers, make sense from the point of view of limiting agency costs. Others, like constitutional rights, can also be explained from a precommitment perspective, in that constitutions take certain core interests off the table of political activity. Finally, there are some provisions that are commonly found in constitutional texts that do not make sense from either of these perspectives, but instead seem better accounted for as a matter of simple coordination. Provisions on a flag or national anthem might fall into this category.

2.2  Bargain Theory and Constitutions Design implies a technocratic paradigm in which a designer, perhaps behind a veil of ignorance, chooses institutions to govern in the future. Alas, this image does not easily fit the messy reality of real-world constitution making. At the most basic level, there is rarely a Solon, a single wise designer, but instead an array of political forces that must work together despite different interests. From an economic perspective, we should assume that constitutional designers are motivated, not so much by a technocratic desire to get the results right, but instead by self-​interest. While early theorists assumed that greater level of entrenchment would ensure more attention to the general interest in constitutional bargaining, others have pointed out that interest groups may be drawn to the relatively more enduring constitutional level precisely because they seek to entrench their policies. Furthermore, real-​world exercises in constitutional design are often undertaken in less than ideal conditions: they are typically concluded under severe time pressures, and often in the aftermath of significant social and political conflict. Thus the conditions seldom obtain for realization of the noble goals, even if self-interest could be overcome. Contract theory is a good place to start to analyze the formation of constitutions (Cooter, 2000; Elkins et al., 2009). Just as with contracts, constitutions are agreements among parties to a mixed game of conflict and cooperation. While all parties may have an incentive to work together to produce an agreement, they often differ over the distribution of costs and benefits. Hence, standard bargaining problems arise, and factors like the reservation price may determine if an agreement is obtained. But unlike in a commercial setting, where a failed bargain simply leads to a search for another partner with whom to trade, constitutional conflicts are often situations of “bilateral monopoly.” This term denotes bargaining situations with only one buyer and one seller, and such bargains are notoriously difficult to conclude. Knowing that the other side likely does not have an alternative country with which to join, each side may hold out for the best possible deal, failing to reveal the information that makes deals possible. Indeed, both sides have the incentive to signal that they will not conclude a deal, so as to extract more from the other side. Holding out might lead the other side to back down, but it may also lead

 

The Design of Constitutions    31 it to being equally stubborn in signaling resolve. When bargaining fails, no agreement can be had, and the result might even be civil war. The probability of concluding an agreement is in part a function of the joint surplus to be obtained from agreement, as well as the opportunity costs that obtain from moving away from current arrangements. Even if there are large gains to be had, however, imperfect and asymmetric information can get in the way of a deal. Perhaps parties disagree on the probabilities of certain contingencies materializing; or perhaps the problem is one of private information about intentions and capabilities of the parties. There may be significant transaction costs in the form of distrust or legacies from past encounters. Even with the best of intentions, many prospective constitutional bargains may fail to materialize. Even if these problems are overcome, constitutional drafting, like other forms of legal drafting, is essentially a problem of constrained optimization. If drafters faced no limits of time and agreement, they would draft a complete constitution that addressed all contingencies and spelled out how they are to be resolved. But in the real world, time and agreement are scarce and costly. We call these kinds of bargaining costs “decision costs.” A further problem is that the state of the world may be unpredictable. Many a deal fails in the throes of unanticipated consequences and unforeseen external events. If parties have different assessments of the future, or even different discount rates with regard to the same assessments of the future, they will find it difficult to agree on the distribution of costs and benefits. It is worth noting that the positive empirical literature on constitutional design has not done a superb job of identifying universally optimal institutions, or even of producing clear normative recommendations as to the relevant contextual factors. Limits on the information available to constitutional decision-makers, and the potential for downstream uncertainties about the optimality of particular constitutional provisions, mean that constitutional decision-​making involves the potential for serious “error costs.” The veil of ignorance about the performance of prospective institutions is a very real one. These features that create decision costs and error costs make bargaining a real challenge. Taking them into account, we move on to different levers of constitutional design that can be mobilized in service of a bargain once achieved.

2.3  Design Dimensions 2.3.1 Scope Constitution-​makers face the immediate question of which topics to include in a constitution and which to leave out. Returning to our discussion of the purposes of constitutions helps to sort out the general categories. Implicitly, constitutional designers are making a decision to entrench certain institutional features at the constitutional level,

 

32   Tom Ginsburg which is presumably difficult to change, rather than the level of ordinary legislation or executive action. Putting something into the constitutional text is a decision to take it off the table of future actors, or else to resolve coordination problems to facilitate their behavior. It follows that certain core features of constitutions, such as rights, are precisely those that will serve to limit agency problems of government or facilitate the precommitment necessary for politics to operate. Of course, if constitution-​makers are ignorant about all the relevant contingencies with regard to a particular policy, they ought to leave it for later decision-​makers. An intermediate approach is what Dixon and Ginsburg (2011) call constitutional deferral, or deciding not to decide everything. To illustrate, a constitutional drafting committee might believe that there should be a constitutional court, but not be confident about the number of members that is optimal. They might then put a clause into the constitution that states that the first legislature shall pass a law concerning the constitutional court. Such an approach helps minimize both decision costs, because the constitution-​makers need not spend scarce time on a relatively detailed matter, and also error costs, if the designers do not have confidence that they can identify the optimal number of judges at the time of drafting. By increasing opportunities for information revelation before the decision is made, such an approach increases the probability that the optimal decision will be made. This strategy of leaving some things out is likely to be most useful for either trivial matters, like the color of the flag, for which there is likely to be little disagreement and hence no reason to spend scarce bargaining resources, or else very high-​stakes items on which there is substantive disagreement. Dixon and Ginsburg (2011) give the example of the decision by South African constitutional negotiators to remain silent on the constitutional status of the death penalty, a decision that was critical to the adoption of the 1994 South African Constitution. The two major parties to the negotiation, the National Party (NP) and African National Congress (ANC) were sharply divided on the issue of the death penalty, and would not have been able to reach a compromise. By remaining silent on the issue, they effectively postponed it to be decided by the Constitutional Court, which eventually outlawed it.

2.3.2 Detail Another design dimension, related to the approach of deciding not to decide, is to use ambiguity. The large literature on rules and standards is relevant here (Kaplow, 1992). To summarize, rules are costlier to spell out up front, but reduce discretion in decision-​makers down the road. In contrast, standards minimize decision costs at the time of drafting, but can increase error costs from the point of view of the constitutional designer, particularly if agency problems lead to deviation from the optimal policy. They empower downstream decision-​makers over upstream ones. There is a long tradition of asserting that constitutions should be framework documents that do the bare minimum. In debates over the US Constitution, Edmund

 

The Design of Constitutions    33 Randolph asserted that it should include “essential principles only; lest the operations of government should be clogged by rendering those provisions permanent and unalterable, which ought to be accommodated to times and events.”2 Yet this does not seem be universally agreed. Constitutional specificity differs dramatically across different contexts. Tuvalu has more words in its Constitution (36,641) than residents (11,992), while India’s Constitution is the world’s longest at more than 140,000 words. One major factor has to do with shared unwritten understandings among the subjects of the constitution, who must cooperate to make it effective. These unwritten understandings form the background against which the text is written. In contexts in which intersubjective agreement is high, written detail is unnecessary to facilitate coordination.

2.3.3 Flexibility Another issue that concerns the intertemporal balance of power between designers and their successors has to do with the flexibility of the constitutional text, particularly as embodied in the amendment rule. Real-world amendment rules vary a good deal, and are in fact difficult to compare across countries. A large literature on amendment argues that it is necessary to adjust to changes in the state of the world, and thus is a crucial device for dealing with error costs in constitutional design. Yet the optimal level of amendment difficulty is not obvious, and probably differs over different substantive provisions of the constitution (Buchanan and Tullock, 1961). Key factors would be the degree of variability in the relevant environmental conditions across time; the level of information available to designers when they create the constitutional bargain; and the confidence they have in downstream decision-​makers. Formal and empirical work in this area have moved the analysis forward. For example, Dixon and Holden (2012) examine an under-​analyzed feature of amendment rules, namely that the difficulty of amendment is determined not only by the threshold required in any voting body (majority, two-​thirds, three-​quarters, etc.), but also by the overall size of the body. They show that as the denominator goes up, the practical difficulty of marshalling any particular threshold increases. This means that, for example, as the population of the United States has risen over two centuries, the absolute difficulty of constitutional amendment has increased as well, perhaps becoming unobtainable. More recently, Ginsburg and Melton (2015) show that the level of difficulty in constitutional amendment is unrelated to the actual observed level of amendments in different countries.

2 

See Supplement to Max Farrand’s The Records of the Federal Convention of 1787, at 183 (J. H. Hutson ed., 1987).

 

34   Tom Ginsburg

2.3.4 Timing and Duration Time is an important dimension of institutional design (Ranchordas, 2014). Constitutions are typically framed as being designed to endure forever, even if in fact they seldom come close to this aspiration. Yet on closer examination, constitution-​ makers frequently manipulate time in interesting ways, and could probably do much more along these lines. For example, they can utilize sunrise clauses that delay the start period of particular rules, or sunset clauses, wherein particular institutions or rules expire after a certain period. They can mandate reviews of performance after a certain amount of time; and ultimately, as considered in the next subsection, they can designate the entire constitution as temporary or interim, embedding the bargaining process into a multi-​stage game. As to when constitutional provisions ought to take effect, designers will often designate an in-​force date for the text as a whole, to allow for smooth transitions from prior regimes. There may also be mandatory periods during which certain choices are off the table. The Southern delegates to the Federal Convention in the United States, for example, obtained a twenty-​year prohibition against federal legislation on the slave trade. Such management of transitional periods can reduce the stakes of bargaining and thus facilitate agreement, especially if parties to the bargain are optimistic that the world will shift in their favor. (This turned out not to be correct for the Southern delegates in the United States, who expected development toward the southwest but did not anticipate the expansion facilitated by the Northwest Ordinance.) Duration of provisions is another design dimension. From a bargaining perspective, any particular constitutional provision should only last as long as the sum of decision costs and error costs for it are minimized. If we think that changes in the state of the world will impact the performance of a particular design choice, then it is advisable to have it expire, or at least be subjected to a reassessment, down the road. We have several examples of sunsetted provisions and mandatory reviews. For example, the 1992 Constitution of Fiji required a review after several years’ time. Many US states require their citizens to vote periodically on whether to call a new constitutional convention. The idea that changes in the state of the world required automatic expiration of laws goes back at least to Plato’s Athenian Stranger and was highly influential on the thought of Thomas Jefferson, who believed that constitutions ought to be replaced every nineteen years. For the most part, however, designers have rejected this position. As Ranchordas (2014: 1) puts it, “in the real world, laws do not expire when they should, but when they can.”

2.3.5 Iteration In recent years, there has been a proliferation of “interim” and “transitional” constitutions, particularly in post-​conflict environments like Iraq, Somalia, and Tunisia.

 

The Design of Constitutions    35 These documents lay out a minimal set of government institutions, and typically will have a timeline for production of a more permanent constitution. One can view this approach as a decision to spread out the bargaining process over multiple iterations, each of which is smaller in scope and stakes than an “all in one” bargain. The logic of interim constitutions is that by limiting their content to a minimal set of institutions, negotiation will be less difficult, and implementation may set in motion a “course-​ of-​dealing,” reducing decision costs over time. This would seem a desirable strategy, particularly in environments of very low trust, but there are countervailing risks. Underlying the trajectory of continued cooperation across iterations is the idea that the joint gains from proceeding to the next round exceed the costs. But the distribution of gains is important. Incrementalism in legal transitions may incentivize strategic behavior by losers in an earlier round to “punish” winners in subsequent rounds (Levmore, 2010). Anticipating this, winners in an early round may wish to freeze the negotiation after the early round.

2.4 Conclusion: Constitutional Design in Practice A small literature examines the structure of constitutional bargaining from an economic perspective. For example, looking at the United States Constitution-​making process, McGuire and Ohsfeldt (1986, 1989) conduct an empirical evaluation of voting behavior of the delegates to the US constitutional convention, and find some support for public choice hypotheses of self-​interest among drafters. But there is little comparative literature in this vein (Brennan and Pardo, 1991). A separate branch of empirical literature considers the impact of constitutions on such dependent variables as economic growth. Persson and Tabelllini (2003) are major figures here. Elkins et al. (2009) take up the question of what factors help constitutions to endure, drawing on a large database that should facilitate more empirical testing of hypotheses generated by law and economics. They argue, among other things, that the optimal level of interest-​group activity is not zero, as some degree of self-​interest is necessary to make constitutions endure and thus provide their purported stabilizing benefits. To conclude, while law and economics has done a decent job of generating hypotheses on constitutional design, there is much more room for empirical examination to see whether the normative and positive hypotheses are supported. There are obviously major challenges to inference in a setting in which experiments are difficult to produce. Yet as empirical research techniques advance, one can be optimistic that the rich theorizing done to date will be subjected to equally impressive and rigorous testing.

 

36   Tom Ginsburg

References Brennan, G. and Pardo, J. C. (1991) “A Reading of the Spanish Constitution (1978),” Constitutional Political Economy, 2: 53–​79. Buchanan, J. and Tullock, T. (1961) The Calculus of Consent. Ann Arbor, MI:  University of Michigan Press. Carey, J. (2000) “Parchment, Equilibria, and Institutions,” Comparative Political Studies, 33: 735–​61. Cooter, R. (2000) The Strategic Constitution. Princeton, NJ: Princeton University Press. Dixon, R. and Ginsburg. T. (2011) “Deciding Not to Decide: Deferral in Constitutional Design,” International Journal of Constitutional Law, 9: 636–​72. Dixon, R. and Holden, R. (2012) “Constitutional Amendment Rules:  The Denominator Problem,” in T. Ginsburg (ed.) Comparative Constitutional Design. New York: Cambridge University Press, 195–​218. Elkins, Z., Ginsburg, T., and Melton, J. (2009) The Endurance of National Constitutions. New York: Cambridge University Press. Ginsburg, T. and Melton, J. (2015) “Does the Constitutional Amendment Rule Matter at all? Amendment Cultures and the Challenge of Measuring Amendment Difficulty,” International Journal of Constitutional Law, 13: 686–​7 13. Hutson, J. H. ed. (1987) Supplement to Max Farrand’s The Records of the Federal Convention of 1787. New Haven, CT: Yale University Press. Kaplow, L. (1992) “Rules versus Standards:  An Economic Analysis,” Duke Law Journal 42: 557–​629. Levmore, S. (2010) “Interest Groups and the Problem with Incrementalism,” Pennsylvania Law Review, 158: 815–​58. McAdams, R. (2015) The Expressive Power of Law. Cambridge, MA: Harvard University Press. McGuire, R. and Ohsfeldt, R. (1986) “An Economic Model of Voting Behavior: The Ratification of the United States Constitution,” Journal of Economic History, 46: 79–​111. McGuire, R. and Ohsfeldt, R. (1989) “Self-​interest, Agency Theory, and Political Voting Behavior: The Ratification of the United States Constitution,” American Economic Review, 79: 219–​34. North, D. (1991). Institutions, Institutional Change, and Economic Performance. New  York: Cambridge University Press. North, D., Wallis, J., and Weingast, B. (2009) Violence and Social Orders. New York: Cambridge University Press Ordeshook, P. C. (1992) “Constitutional Stability,” Constitutional Political Economy, 3: 137–​75. Persson, T. and Tabellini, G. (2003) The Economic Effects of Constitutions. Boston, MA: MIT Press. Ranchordás, S. (2014) Constitutional Sunsets and Experimental Legislation. Chelthenham: Edward Elgar. Van den Hauwe, L. (2000) “Public Choice, Constitutional Political Economy and Law and Economics,” in B. Bouckaert and G. De Geest (eds) Encyclopedia of Law and Economics, Vol. I: The History and Methodology of Law and Economics. Cheltenham: Edward Elgar, 603. Weingast, B. (1997) “The Political Foundations of Democracy and the Rule of Law,” American Politcal Science Review, 91: 245–​63.

 

Chapter 3

De mo cratic Ru l e ma k i ng John M. de Figueiredo and Edward H. Stiglitz

3.1 Introduction One feature common to all modern democracies is the delegation of substantial authority to unelected bureaucrats. For the most part, this regularity is driven by the simple fact that the demands of lawmaking in modern society have outstripped the institutional capacity of legislatures to craft their own rules (Cox, 2006). Legislators do not have the time nor expertise to fully solve the detailed and complex problems that a modern society faces. One important way legislatures have adapted to this shortcoming is to delegate to expert administrative agencies rulemaking authority1—​the authority to make rules that carry the force of law.2 A more complete theoretical explanation for the regularity of delegation also points to incentives created by elections as well as the intertemporal instability of legislative coalitions. Some authors argue legislators delegate “difficult” policy choices to agencies, thereby claiming credit for “solving” the public policy problem, but avoiding blame for any pain that the solution imposes on electorally relevant actors (e.g., Ely, 1980; Fiorina, 1982; Schoenbrod, 2008). More recent theoretical work, confronting the question of why voters do not blame legislators for the act of delegation itself, has stressed that voters, in an incomplete information environment, may not know if the delegation occurs for a “good” reason (such as agency expertise) or for a “bad” reason (such as corruption), creating the space for legislators to delegate for bad reasons without being held accountable (Jordan and Fox, 2011). Yet another strand of literature argues that delegation may 1 

Although we briefly touch on other countries, our focus throughout is on rulemaking in the US context. 2  Delegated authority may also come, for example, in the form of the ability to adjudicate disputes between parties; although we touch on such alternative forms of delegated authority, our focus is on rulemaking.

 

38    John M. de Figueiredo and Edward H. Stiglitz occur to allow the government to credibly commit to a course of action (e.g., Rogoff, 1985) and permit legislators to “insulate” policy from future legislative coalitions that may have divergent policy preferences (e.g., Moe, 1990; Shepsle and Horn, 1989; Shepsle, 1992; R. de Figueiredo, 2002). To give a sense of the scope of agency rulemaking, administrative agencies in the United States finalized over 2,800 rules in 2013, regulating virtually every corner of American life, from air and water quality, to food quality, to automobile and workplace safety.3 By comparison, the US Congress produced only seventy-​two public laws during that same year. These figures cast into relief a fundamental normative question that has occupied scholars since the inception of the administrative state: in what sense can administrative rulemaking be “democratic” if we do not elect the individuals writing the rules? In this chapter, we rule out one possible view of the term “democratic rulemaking” immediately. It cannot require the regulator to be directly elected. Outside of a few narrow circumstances, such as state Public Utility Commissions, citizens do not directly elect regulators. And this is for good reason. At present there are nearly two hundred federal agencies with rulemaking authority;4 it would be exceptionally difficult for voters to determine the quality or competence of each regulator both due to the sheer number of regulators, and the fact that expertise is, by definition, difficult for laypersons (i.e., voters) to assess. This then leaves two main possibilities for democratic rulemaking benchmarks, both of which we consider in this chapter. The first is what we refer to as “legislative matching” or the “representative” or “republican” (with a small “r”) benchmark.5 Under this approach, the relevant question is whether the rulemaking matches what would have been passed as a statute if Congress had the same expertise and time that the agency used to develop the policy. This normative baseline focuses on examining to what extent delegation from the legislature to agencies distorts public policy outcomes from contemporaneous congressional legislative preferences.6 3  During 2013, that is, agencies reported a total of 2,803 actions in the Federal Register under the title “final rule”; this figure probably understates the total level of rulemaking, as agencies sometimes list the action as something other than a “final rule” when publishing in the Federal Register (for instance, “final action”). Of course, this figure masks incredible heterogeneity in the content of rules, some important, some not, some increasing regulatory burdens, some reducing regulatory burdens, and so on. On this last point, see, e.g., Carey (2015). 4  We draw this figure of agencies with rulemaking authority from Farina (2010: 361). 5  We stress that the term “republicanism” refers to our representative institutions, and is quite distinct from notions of “civic republicanism” articulated in Seidenfeld (1992) and elsewhere. 6  Benchmarking by the representative institution makes a certain amount of sense in the context of rulemaking, as it focuses on what is novel about rulemaking relative to lawmaking. Note that for most of the article, when we refer to “congressional” preferences, we have the Congress that is contemporaneous to the rulemaking in mind. One might also imagine a benchmark based on the preferences of the enacting legislature, and, in fact, when courts interpret statutes they at least formally have the enacting and not the current Congress in mind. Our enterprise in this chapter, however, is different from the courts’ enterprise: we wish to establish a normative benchmark, not develop a theory of statutory interpretation.

 

Democratic Rulemaking   39 A second possibility is what we refer to as “electorate matching” or the “democratic” (with a small “d”) benchmark. Under this approach, the relevant question is whether the rulemaking matches what the median voter would have done, if that representative voter had the same expertise and time that the agency used to develop the policy. This normative baseline, therefore, essentially bypasses the legislature as an intermediate institution, asking to what extent rulemaking outcomes are consistent with the preferences of the voters themselves, as if expressed through a referendum where voters had the relevant expertise and time. With those normative baselines in mind, we organize this chapter around the central positive theories that scholars have advanced to address the democratic legitimacy of the administrative state. We divide the remainder of this chapter into three main parts. The first part of the chapter (section 3.2) explores how governmental institutions allow the legislature and the president to control agencies. These institutional structures lead to either republican or democratic outcomes for rulemaking. The second part of the chapter (section 3.3) explores how procedures within the agencies affect the ability of political principals to achieve republican or democratic outcomes. The final section of the chapter (section 3.4) identifies shortcomings in the literature and suggests paths for future research. The overall conclusion from this chapter and the literature is that no existing theory of democratic rulemaking predominates; they all suffer from serious conceptual and empirical faults. In this environment, there are many promising paths for future research by academics to fill gaps in the literature.

3.2  Institutions and Democratic Rulemaking 3.2.1 What Institutional Structures? Institutional structures establish certain core rights and responsibilities of actors in the administrative policymaking process. Often, though not always, these roles derive from interpretations of the Constitution. For example, one critical set of structures relates to control over administrative personnel. At least the boundaries of the question of who controls the appointment and removal of officers, and in what contexts, largely turn on one’s interpretation of the Constitution. Likewise, control over agency budgets is one important structure that influences regulatory policymaking, and congressional preeminence in this area is rooted in the Constitution.7 Other times, working within permissible constitutional boundaries, institutional structures emerge through statute or executive order. The most notable structure in this regard is the Office of Information 7 

Art. I, § 9, cl. 7.

 

40    John M. de Figueiredo and Edward H. Stiglitz and Regulatory Affairs (OIRA), within the Office of Management and Budget, which gives the president the capacity to audit the rulemaking efforts of many agencies.8

3.2.2 Executive Agencies and Presidential Control An influential theory holds that rulemaking is “democratic” by virtue of the President, an individual who is envisioned both to control agencies and to be electorally accountable for agency actions (Hamilton, Federalist 70; Prakash, 2003; Yoo, 2005). This perspective, often alloyed to the so-​called unitary executive theory, advocates for the concentration of administrative authority in the hands of the President, as doing so renders it plain who is responsible for regulatory actions, and provides voters the channel of presidential elections to discipline those actions. The notion that concentrating authority and, runs the theory, responsibility in the hands of a single individual is historically well pedigreed. Attacking proposals for a “plural” executive consisting of several individuals, Alexander Hamilton wrote in Federalist 70 that, “one of the weightiest objections to a plurality in the Executive … is that it tends to conceal faults and destroy responsibility.” In the event of maladministration under a plural executive, he continued, blame “is shifted from one to another with so much dexterity, and under such plausible appearances, that the public opinion is left in suspense about the real author.” By placing administrative authority in the hands of a single individual—​the President—​the public will readily understand the author of any “pernicious measure,” and then may use elections to punish him for these actions. Though the history is not uncontested, courts have often construed the Constitution in ways sympathetic to this theoretical perspective, giving the President the preeminent role in administrative personnel decisions.9 Prominently, it is the President who makes the appointments of principal officers to agency positions, with only the “advice and consent” of the Senate.10 And, as interpreted by the courts, the President likewise generally enjoys a dominant position in the removal of such officers.11 Over time, the President has developed additional tools and structures of control. Most importantly, the President, with help from Congress, created in 1980 the OIRA, which allows the President to audit agencies’ rulemaking efforts (Bubb and Warren, 2014). But presidents have also developed and amplified other tools. For example, they have burrowed their political appointees deeper into the agencies through non-​confirmed and Schedule C appointments (Light, 1995; Lewis, 2008; Bonica, Chen, and Johnson, 2015). 8 

OIRA influence over independent agencies is limited. This chapter is not the place to chronicle the many nuances of doctrine in this area of US constitutional law, but it is still important to note at least that the President’s control over personnel is more limited in independent agencies, see, e.g., Humphrey’s Executor v. United States, 295 U.S. 602 (1935). 10  Note that much turns on whether the appointment refers to an inferior or principal officer. For a case illustrating this fault line, consider Morrison v. Olson, 487 U.S. 654 (1988). 11  Of course, there is no “Removals Clause” of the Constitution; for an early and still-​influential interpretation on removal, see Myers v. United States (1926). 9 

 

Democratic Rulemaking   41 Although this presidential domination perspective promises an elegant mechanism to impose democratic disciplining on rulemaking, we have several reasons to be skeptical. Perhaps most centrally, this perspective relies on presidential elections to endow rulemaking with democratic credentials. However, presidential elections may be poor institutions through which to discipline regulatory behavior. Besley and Coate (2003, 2008) observe that elections force voters to choose among “bundles” consisting of a potentially wide array of policy and regulatory commitments. The fact that voters only choose among bundles, they show, implies that politicians may adopt non-​majoritarian positions on some issues. As Berry and Gersen (2008) and Gersen (2010) argue, this result diminishes some force from the notion that presidential domination of rulemaking necessarily produces “democratic” rules in the relevant sense. Rather, the fact that voters select only among bundles of policies suggests that the President enjoys wide discretion to adopt non-​majoritarian positions on many regulatory policies. A second criticism is that it seems unlikely that voters can process the information necessary to hold the President accountable for regulatory choices (e.g., Farina, 2010; Mendelson, 2011), even if the institutional limitations above did not apply or applied only at the margins. The most rudimentary legal distinctions—​the difference between a statute and a regulation—​elude even highly motivated citizens (Cuéllar 2005), indicating they would have difficulty separating congressional from executive work product.12 These considerations suggest that presidential elections most likely serve as poor institutions for delivering “democratic” rulemaking.13 A third class of criticisms questions the ability of the President to control the bureaucracy—​questions, that is, the fundamental premise, as a positive matter, of executive domination. The seed of this class of criticisms is implicit in the long-​standing debate over who controls administrative agencies (e.g., Weingast and Moran, 1983; Weingast, 1984; Hammond and Knott, 1996), an inquiry that only makes sense in the context of factually ambiguous control over agencies. But more recently, scholars have begun to generate 12 

We discuss this point further in section 3.3.1, where we engage with the deliberative potential of the rulemaking process. 13  Another virtue of so concentrating authority, according to proponents, is that the President is the only individual that all citizens vote for. So being, goes the notion, the President is less prone to parochialism and other pathologies that afflict other representative entities. As Woodrow Wilson put it, “The nation as a whole has chosen him … he is the representative of no constituency, but of the whole people … he speaks for no special interest” (Wilson, 1885). These national credentials, unique to the President, elevate the “democratic” status of his regulations, pulling regulatory outcomes toward voter preferences and away from those of special interests. However, the coalitions that elect a president represent something less than the “nation as a whole.” The Electoral College provides presidential candidates with an incentive to pursue a coalition consisting of a fraction of the voters, plausibly less than a majority, and in this sense presidential elections engender the same type of parochialism associated in Congress (Nzelibe, 2005). Thus, in this view, elections introduce a certain form of disciplining of regulatory behavior, but quite plausibly induce accountability to some sub-​coalition of voters rather than to the median of the electorate, much less to the “nation as a whole.”

 

42    John M. de Figueiredo and Edward H. Stiglitz data directly pertaining to the executive domination thesis. One set of scholars, for instance, surveyed agency officials, and they reported “sporadic” rather than systematic control by the White House over regulatory matters (Bressman and Vandenbergh, 2006). Another recent study, likewise, asked agency officials, “In general, how much influence do the following groups have over policy decisions in your agency?,” then listing, among other entities, “congressional committees” and the “White House.” In a seeming direct refutation of the executive domination thesis, the average response by officials appears to lie between 0 and 1 on a scale between “–​4” and “4”, with “–​4” representing complete congressional (committee) domination, and “4” representing suggesting complete White House domination (Clinton, Lewis, and Selin, 2014). All of this suggests that, in general, control over agencies is as a practical matter shared by Congress and the President. Based on the views of agency officials, the White House does not dominate policymaking to the exclusion of Congress. Even as the President asserts his interests through OIRA review and other means (Kagan, 2001; Croley, 2003), Congress retains many tools of bureaucratic influence, and agencies retain their own policy agendas and means of evading executive oversight (Nou, 2013).14

3.2.3 Temporal Considerations and the Role of Independent Agencies As has long been argued, constitutions represent a technology to allow us to bind our selves to a certain course of action, a form of democratic action “by the body politic … to protect itself against its own predictable tendency to make unwise decisions” (Elster, 2000:  88).15 Following this notion, the US Constitution contains many elements designed to resist responsiveness to transitory shifts in majority preferences: for example, a challenging amendment procedure; a core separation of powers structure; and a bicameral legislature, with the Senate acting as “an anchor against popular fluctuations” (Madison, Federalist No. 63). As many constitutions both embrace democratic 14  This feature of “shared” powers of administration is deeply embedded in the structure of the US Constitution (e.g., Fisher, 1998). The framers designed a system, which though termed a “separation” of powers system, in fact fostered near-​constant interaction and contestation between the branches over shared areas of control. Congress routinely attempts to influence regulatory policymaking by using a wide range of tools, from oversight hearings and the subpoena power to appropriations (for accounts of such varied tools, see generally Beermann, 2006; Chafetz, 2012). These tools tend to confine and condition executive discretion with respect to regulatory policymaking, and in this way undermine the core premises of the executive domination thesis. Given this fundamental constitutional structure of separated yet still shared powers, the idealized vision of this executive domination thesis is unlikely to be realized, through judicial decisions or other means (Huq, 2013; Stiglitz, 2014a, 2014b, 2016). 15  More colorfully, the New Jersey politician John Potter Stockton, who resisted post Civil War amendment to the US Constitution, said, “Constitutions are chains with which men bind themselves in their sane moments that they may not die by a suicidal hand in the day of their frenzy,” quoted in Elster (2000: 89).

 

Democratic Rulemaking   43 ideals and resist popular impulses, so might administrative rulemaking. Most theories of “democratic” rulemaking, however, neglect this temporal aspect of preferences, so central to the study of constitutionalism (but see below; Majone, 1996, 1997). The basic notion of time inconsistency is that the preferences of an individual or institution may change over time. A major source of inconsistency derives from the fact that the individual or group that controls policy at time period t may not control it at time period t + 1. For example, the dominant coalition in the legislature that passes a policy at time t may subsequently lose an election, and the identity of the dominant coalition shifts at time t + 1 (Shepsle and Horn, 1989).16 Or, classically, a legislature’s preferences over actions may change once another actor takes some action, such as a decision to set wage levels or make an investment (e.g., Rogoff, 1985). Although the general notion of time inconsistency in preferences is not foreign to the study of administrative action, its role in democratic rulemaking is perhaps understudied. Scholars have long recognized that designing administrative agencies precisely to thwart majority impulses potentially enhances citizen welfare (e.g., Kydland and Prescott, 1977; Rogoff, 1985; Dixit, 1998). Such studies, however, tend to concentrate on monetary policy, understanding the delegation of that policy to “independent” central banks as a way of committing to a policy of low inflation and defeating the problem of dynamic inconsistency in preferences (Rogoff, 1985). There, the problem is that the legislature cannot commit to a policy of low inflation. The solution, in the standard account, is to delegate monetary policy authority to an independent agency with the ability to so commit. This is the more or less conventional account for central bank independence (e.g., Drazen, 2004), but the same logic might apply any time that the legislature wishes to commit to a course of action when otherwise faced with incentives to renege in some fashion. For instance, after inducing third parties to invest in costly infrastructure, such as railroads or communications networks, the government often faces an incentive to expropriate through “low-​ball” ratemaking or other regulatory means. This is one plausible understanding of why many agencies with jurisdictions over policy areas requiring heavy investment, such as communications (FCC) or, in an earlier era, railroads (ICC) often have institutional features promoting autonomy from the political branches. By delegating policy authority to independent agencies, the legislature is able to induce investment that otherwise would wither in recognition of the legislature’s inability to commit to not expropriate. Thus, as in the classic story of Ulysses and the sirens, in a 16  Temporal inconsistencies may also arise when a new group or coalition gains predominant control over the agency following delegation. One common concern, for instance, is that the regulated entities will “capture” the administrative agency, following the desires of the regulated entities by adopting lax regulatory standards (Kwak, 2013) or, classically, erecting barriers to entry for potential competitors (Stigler, 1971). This also is a form of temporal inconsistency, in the sense that the regulated industries, rather than the legislature, controls agency behavior post-​delegation. Of course, one way in which this might not represent such an inconsistency is if the legislature delegated just so that the agency might be captured. Although we do not want to dismiss this as a relevant possibility, do not focus on this case.

 

44    John M. de Figueiredo and Edward H. Stiglitz range of policy areas, agency independence arguably serves to protect the legislature from itself. Thus, what the Constitution accomplishes through the separation of powers, bicameralism, and other devices, the administrative state might accomplish by structuring administrative agencies to be insulated from the political principals, centrally the President.17 This is achieved, in the standard account, through various structures. For example, making officers of the agencies removable only “for cause,” rather than “at will,”18 but also, for example, by creating agencies with boards rather than a single head, staggering the terms of such members, and providing means of self-​financing, as through licensing fees or other assessments, to the agency (e.g., Barkow, 2010; Datla and Revesz, 2013).19 Providing agencies with “autonomy” or “independence” is likewise one response to the problem of the enabling legislature not controlling agency policy in the future. Moe (1990) and de Figueiredo (2002), for example, observe that due to electoral uncertainty, today’s legislative winners may be tomorrow’s legislative losers, providing the legislature with an incentive to delegate policymaking authority to an “insulated” or independent bureaucracy. Barkow (2010), similarly, argues that one response to the problem of agency capture—​in one interpretation, a source of inconsistent preferences—​might be to insulate the bureaucracy from political and special interest pressure. More recently, scholars have focused on another implication of the legislature’s inability to commit: the development of agency expertise.20 This literature begins with the premise that the agency has the time and ability to develop expertise, but that doing so is costly; the legislature either cannot develop this expertise, or it is much more costly to do so. This difference drives the legislature to delegate policymaking authority to the agency. However, the agency is concerned that the legislature will “expropriate” the expertise for its own ends, and this discourages agency investment in expertise, as in the more general story above.21 Along these lines, Gailmard and Patty (2007) show that an agency’s expertise is sensitive to the level of policy discretion conferred on the agency by 17  It is worth noting, however, that on one view agency “independence” principally serves to insulate agencies from presidential influence—​not “politics” or congressional influence—​and that as a result, Justice Scalia notes, executive control is “simply … replaced by increased subservience to congressional direction,” FCC v. Fox Television Stations, Inc., 129 S. Ct. 1800 (2009). If all that is achieved by making an agency independent is that executive influence is displaced in favor of congressional influence, it is unlikely that many of the benefits we discuss in this part will be realized. 18  For the seminal case on this point, casting such political insulation as constitutional, see Humprey’s Executor v. United States, 295 U.S. 602 (1935). 19  The Federal Reserve is the most commonly noted self-​funding agency, but many other agencies, often regulating the financial sector, have mechanisms of self-​funding (see Barkow, 2010). 20  Note that the notion that agents’ incentives to collect information are sensitive to the degree of independence from the principal is familiar to students of legislative politics, where “independence” derives from restrictions on floor amendments to committee legislative proposals (Gilligan and Krehbiel, 1987). For an excellent assessment of the literature on administrative law and endogenous information acquisition, see Stephenson (2011). 21  One move in response to this problem is to reconceptualize information so as to not be fully expropriatable (Callender, 2008; Hirsch and Shotts, 2012).

 

Democratic Rulemaking   45 the legislature. Cameron, de Figueiredo, and Lewis (2015, 2016) develop integrated models of agency expertise based on specificity of skills, political control, and discretion and show how both shirking and turnover in agencies occurs when there is an inability of politicians to commit to following the advice of expert bureaucrats.22 Bubb and Warren (2014) study the optimal level of agency bias and policy review, as through OIRA, in the context of endogenous information acquisition. In many cases, independent or politically “insulated” agencies should be viewed as advancing democratic objectives. This is plain when agency independence is understood as a response to the possibility of agency capture by special interests. But it should also be understood as advancing democratic values even when independence protects one democratically elected coalition against another democratically elected coalition, or when independence protects the legislature from itself. This follows from the fact that independence enlarges the scope of public policy problems that the legislature can competently address.23 Without these forms of constraint we might not be able effectively to pursue objectives sought by many voters: low inflation monetary policy, for example, or agencies that develop expertise in their policy domains. Depending on the context, these structures might promote either, in our terms, democratic or republican outcomes.

3.3 Procedures 3.3.1 What Procedures? Procedures are similar to institutional structures in that they establish the rights and privileges of actors in policymaking; they differ primarily in that they derive almost exclusively from statute or executive order.24 In the United States, administrative procedures come from three main sources: the Administrative Procedures Act (APA) of 1946 which provides the framework for rulemaking procedures and subsequent judicial review; the enabling or organic statute governing the functioning of a particular agency; and other general legislation that covers specific procedures across most or all agencies, such as the Paperwork Reduction Act of 1980 or the National Environmental Protection 22  Bolton et al. (2016) show that career senior executives in the government are most likely to leave early in a new presidential administration whose policy agenda is opposed to the mission of the agency because the civil servants’ expertise may be undervalued by the incoming politician. 23  Another rationale for insulating decision-​making from political accountability focuses on information asymmetries between voters and politicians; in this context, mechanisms of accountability can induce politicians to knowingly adopt inferior policies in an attempt to appear competent or otherwise congruent with citizens. For an excellent overview of this perspective, see Gersen and Stephenson (2014). 24  The Due Process Clause provides some constitutional boundaries on administrative procedures, but such constraints apply for the most part in the adjudicatory context, which is not our focus in this article.

 

46    John M. de Figueiredo and Edward H. Stiglitz Act (Bressman, 2007).25 The procedures contained in these sources—​importantly, often as glossed by courts—​determine the role of individuals and interest groups in regulatory proceedings and govern the manner of information exchange and required steps before a rule can gain the force of law.26 Most often, agencies make policy through rulemakings or adjudications, along with the much discussed and legally ambiguous “guidance” documents (e.g., Magill, 2004; Epstein, 2015).27 Our focus is on rulemakings. In highly stylized form, a prototypical rulemaking emerges from the following procedures.28 First, the agency publishes a Notice of Proposed Rulemaking (NPRM) in the Federal Register. The NPRM, in principle, sets out at least the basic substance of the proposed rule,29 along with the statutory authority for the rulemaking. Following notice, interested parties may submit comments on the proposed rule, explaining their support of or opposition to its various provisions. The agency then “after consideration of the relevant matter presented,” (5 USC § 553(c)) may drop the rule, or more likely revise it as necessary and publish the final rule in the Federal Register. The rule becomes legally binding no sooner than thirty days after it is published in final form (5 USC § 553(d)).

3.3.2 Deliberative Democracy: Procedures as Achieving Democratic Outcomes In a core innovation to rulemaking procedures, the Administrative Procedures Act (APA) sought to enhance public participation during the development of agency rules.30 The APA codified procedures that endow the rulemaking process with elements of deliberation, and scholars have long argued that participatory procedures 25 

For a discussion of international administrative procedures, see Jensen and McGrath (2011). For example, as discussed later in this chapter, section 553 of the APA calls for agencies to publish notice of proposed rulemakings (NPRM) in the Federal Register, to allow interested parties to submit their views on the proposal, and to publish the rule thirty days before it is to take effect (APA § 4, codified at 5 USC § 553). The general procedures set forth in the APA have been refined and supplemented over the years by the agencies themselves and by the courts (see generally Breyer et al., 2011). One can find agency-​generated procedures in the Code for Federal Regulation (CFR) as well as other agency-​specific documents. Judges, of course, also effectively impose procedures on agencies. See, e.g., Bressman (2007). A full discussion of court-​imposed procedures is beyond the scope of this chapter. 27  An agency “guidance” is an informal statement by the agency that purports to interpret or explain another legally binding rule. The fear is that these statements, in effect, create new rules rather than interpret or explain existing ones. 28  These procedures derive from section 553 of the APA. Note that for the purposes of this part, we abstract from the important role of OIRA in rulemaking, which we discuss above. 29  A significant debate surrounds how detailed the notice of proposed rulemaking must be. See, e.g., Nou and Stiglitz (2016). 30  In one account of the APA’s history, the “bill’s notice and comment rulemaking balanced the interests of agencies in speed and efficiency and the interests of the public in participating in the rulemaking process … it required agencies to solicit and consider public comment on the rules … it [was] the most important change that the APA impose[d]‌on [then-​prevailing] agency practice” 26 

 

Democratic Rulemaking   47 confer legitimacy and various other benefits.31 Most notably, the notice and comment procedures promote a type of democratic dialogue between government agency and regulated entities (Seidenfeld, 1992; Carlitz and Gunn, 2002), which arguably improves policy decisions, for example, by enhancing assessments about the cost-​effectiveness of rules (Gailmard and Patty, 2013; Brandon and Carlitz, 2002; Coglianese, 2004). Recent advances in rulemaking technology have inspired fresh interest in this deliberative rationale for rulemaking. By fusing the traditional rulemaking process to the Internet and related information technologies, some see the “potential to enlarge significantly a genuine public sphere in which individual citizens participate directly to help make government decisions” (Shane, 2004). Under the view that lowered costs of participation may induce citizens to participate in the rulemaking process (Stanley and Weare, 2004; Shulman et al., 2003; Schlossberg et al., 2007), agencies have experimented with various methods to facilitate deliberations, including discussion groups, blogs, and real-​time response capabilities (de Figueiredo, 2006; Farina et al., 2011). The case for deliberative democracy and rulemaking is roughly as follows. To the extent that individuals are involved in a deliberative democratic rulemaking process—​with their voices heard and reflected in regulations—​regulatory policy outcomes may be close to the well-​informed, well-​considered median voter’s preferences. Under the stronger versions of this approach, “deliberative” democracy may in fact be better served through rulemaking than other methods of policymaking, such as legislation. After all, the rulemaking process structures decisions to ensure at least some minimum level of engagement and rational discussion between the government and the public—​a characteristic frequently absent from the legislative process. Moreover, unlike legislation, citizens can argue that a court should set a regulation aside because the agency failed to adequately respond to adverse arguments presented in public comments; courts routinely do so. In this sense, rulemaking by agencies may be the “best hope” of realizing “deliberative decision-​making informed by the values of the entire polity” (Seidenfeld, 1992). While the deliberative democracy argument has gained some traction in the academic literature over the past two decades, we also have reasons for pause. First, agencies have frequently determined the content of the rules and regulations before they issue the NPRM (Wagner et al., 2011; Farber and O’Connell, 2014; Nou and Stiglitz, 2016). By the time open public deliberation supposedly transpires, agencies have, under a common view, largely determined the policy outcome through private, closed-​door meetings (Kerwin and Furlong, 2011). Moreover, not only are the meetings largely closed, but the set of interest groups participating in the crafting of the NPRM are highly unlikely to be representative of the median voter (Krasnow et al., 2001; McGarity and Wagner, 2008). (Shepherd, 1996: 1557). See McNollgast (1999); Shapiro (1986); Strauss (1996) for further discussions of the history of the APA. 31 

For example, such procedures increase bureaucratic legitimacy (Cramton, 1971) and federal government credibility (Beierle, 2004), strengthens individual autonomy and rights of self-​governance (Noveck, 2004) and increases public understanding of rulemaking (Coglianese, 2004). For a more complete review of the literature on participative democracy, with coverage of the public law, see Thompson (2008); Chambers (2003); and Ryfe (2005).

 

48    John M. de Figueiredo and Edward H. Stiglitz Indeed, it may be this opaque character of the pre-​NPRM period that interest groups and agencies find so attractive. Second, even after agencies issue the NPRM, the rulemaking process does not proceed along the lines of a democratic ideal.32 The meaningful public comments tend to come almost exclusively from a small number of highly interested groups (Cuéllar, 2005; Yackee and Yackee, 2006). These seasoned interest groups are not only trying to influence the regulations in the agency, but they are also laying the groundwork for future challenges to the rules in the courts. Those individual citizens who do engage the comment-​and-​reply process tend to be very poorly informed, often submitting form letters supplied by interest groups (Cuéllar, 2005; Mendelson, 2011); the “grass roots” interest groups themselves often become involved for the publicity and fundraising opportunities afforded by the rulemaking (Schlosberg et al., 2007). The comments that individuals submit in response to NPRMs provide a glimpse into the understanding that the most engaged citizens have of regulations. One scholar, for example, examined the comments that individuals—​as opposed to, for example, law firms or lobbyists—​submitted for a proposed Department of Treasury rule on financial privacy (Cuéllar, 2005). He found that such individuals submitted the majority of comments, but that they “proved to be tremendously unsophisticated.” Crucially, “few of them recognized the distinction between the regulation and the statute,” (Cuéllar, 2005), suggesting that these motivated citizens lacked even the most basic awareness necessary to influence regulatory policy. This rule is one rule among thousands issued during an administration that a fully informed voter might consider; it is difficult to read such accounts and remain optimistic about the citizen-​deliberative potential of rulemaking. Most accounts of e-​rulemaking, similarly, indicate that technology has so far not greatly enhanced the participatory aspects of rulemaking (de Figueiredo, 2006; Mendelson, 2011). Indeed, even as technology lowers the cost of mobilization and appears to increase the number of comments, the average quality of individual comments may decrease as uninformed and marginally interested citizens represent a larger share of those participating in the rulemaking. Most observers conclude that the vast majority of citizen comments are not helpful to agencies (Weare et al., 2004; de Figueiredo, 2006). As a result, so far, e-​rulemaking does not appear to act as an effective counterweight to agencies’ otherwise strong incentives to focus on the well-​informed (but non-​median) interest groups most likely to present a litigation risk.33 Third, despite claims of transparency and disclosure through the carefully crafted and monitored notice-​and-​comment procedures, many of the more influential post-​ NPRM contacts occur outside of the formal notice-​and-​comment procedures (Kerwin 32  It is important to note that the empirical results of studies of deliberative democracy are quite mixed (see Thompson, 2008). 33  In response to these problems, researchers are seeking out innovative approaches to engage citizens through technology. For a potential promising approach, see Farina et al. (2011).

 

Democratic Rulemaking   49 and Furlong, 2011). A parallel process in many agencies—​ex parte procedures34—​creates a privileged and more opaque process where considerable horse-​trading occurs (de Figueiredo and Richter, 2014; McGarity and Wagner, 2008). During these ex parte meetings, individual interest groups gain preferential access to regulators to air their concerns, transfer ideas and information, and otherwise attempt to mold the regulations to their liking (Krasnow et al., 2001). These ex parte contacts follow a predictable pattern of both timing and access (de Figueiredo and Kim, 2004). Although agencies generally comply with disclosure requirements with respect to these contacts, analysis of the disclosures shows they reveal very little of the substance of the meetings. The disclosures, instead, generally report only the barest essentials: the fact the meeting occurred and the general topic or agenda for the meeting (de Figueiredo and Tiller, 2001; de Figueiredo and Kim, 2004). Finally, agencies often avoid the rulemaking process altogether. They might do so by strategically relying on other policymaking forms with different procedural requirements, such as adjudications (Mashaw and Harfst, 1986; Magill, 2004), or by channeling what amounts to policy changes into guidance documents, or interpretative rules, formally not legally binding instruments that regulated parties may nonetheless regard as effectively binding (Magill, 2004; Magill and Vermule, 2011; Epstein, 2015). We note in passing that negotiated rule​making, where interested groups work together to foster deliberation and forge a solution that is mutually acceptable to all parties, may encourage negotiation and compromise. However, empirically, negotiated rulemaking is rarely used and is generally viewed as having limited success (Coglianese, 1997). If our best hope for deliberative democracy lies in rulemaking, all of this suggests the hope is nonetheless dim.

3.3.3 Control: Procedures as Achieving Legislative Outcomes An alternative view of administrative procedures is that their main purpose is to enhance congressional control of agencies, and to check the power of the bureaucracy and the President (McNollgast, 1987). In this sense, they reflect more a republican rather than democratic tone.35 This congressional control procedural theory views agency control as a principal–​agent problem (McNollgast, 1987; 1989). The principal, Congress, must insure that the better-​informed bureaucrats will show fidelity to the objectives of Congress in designing and executing policy outcomes. Rather than wait for ex post auditing to 34 

We use quotes to acknowledge that the APA only prohibits ex parte contacts in the context of formal proceedings, and that the term is often applied by analogy to informal proceedings, such as rulemakings under APA § 553. 35  We emphasize that Congress has other tools as well—​it does not only rely on procedures to control agencies.

 

50    John M. de Figueiredo and Edward H. Stiglitz rectify bureaucratic deviations from congressional preferences, positive political theory argues that well-​designed procedures statutorily imposed upon the agency will minimize the zone of discretion for the agency through ex ante mechanisms. When bureaucrats test the boundaries of discretion and drift away from congressional preferences, “fire alarms” will be triggered during the rulemaking process ex ante (McCubbins and Schwartz, 1984; Lupia and McCubbins, 1994b), and judicial and congressional auditing will protect the Congress ex post. A number of classes of procedural mechanisms are available to Congress to mitigate the information asymmetry problem and control agencies (McNollgast, 1987; 1989). A first class of procedures attempts to minimize information asymmetries by flushing the rule into the open before it becomes finalized. Most prominently, the notice-​and-​ comment procedures, at least in idealized form, force an agency to reveal its agenda before it becomes a reality, allowing interest groups and aligned congressional factions the opportunity to intervene, for example, through oversight hearings, budgeting, or scrutiny of executive nominees (McNollgast, 1987, 1989; Ferejohn and Shipan, 1990; Eskridge and Frickey, 1994; de Figueiredo et al., 1999). A second class of procedures is designed to enfranchise interest groups whose views are close to the enacting Congress. For example, requiring agencies to complete environmental impact statements—​part of the National Environmental Protection Act—​provides a procedural avenue for groups to challenge rules on environmental grounds. In so doing, congress assures like-​minded interest groups a place in the regulatory process; often this preferred position is judicially enforceable. Moreover, if the agency slights these voices, the favored groups may alert Congress to this fact, again allowing allied congressional factions to mobilize against the agency (McCubbins and Schwartz, 1984; Lupia and McCubbins, 1994b). Congress delegates enforcement of these procedures, in large part, through the courts. For example, a court may set aside a rule if the agency sidesteps the notice-​and-​comment process or ignores arguments presented in public comments. Together, these procedures limit the discretion of agencies to affect policy and channel policymaking to reflect the preferences of the enacting and/​or current Congress. This approach allows Congress to influence rulemaking cost-​and time-​effectively, as decentralized agents with similar interests commit their time and resources to monitoring agencies (Bressman, 2007; Lupia and McCubbins, 1994a). Under this theory, procedures are designed to reflect a republican form of government—​the procedures attempt to generate rulemaking outcomes similar to those that would obtain if Congress had the time and expertise to consider the issue. Despite the intuitive elegance of this theory of congressional control of agencies, it faces a number of challenges. First, it is unclear how robust the theory is. For example, one rationale for notice-​and-​comment rulemaking is to enable Congress to decentralize monitoring and delegate it to like-​minded interest groups. These same interests, however, have preferences of their own which may cause them to untruthfully report the behavior of agencies to congress (Asimow, 1994). A variety of asymmetric information models exist to describe this behavior (see Grossman and Helpman, 2001, for an overview), but each requires special assumptions. These games generally require additional

 

Democratic Rulemaking   51 apparatus for existence of pure strategy equilibria, such as repeated games and reputation, information verification, costly entry to the game, or additional features that do not always map to administrative procedures. Second, while the McNollgast papers do admirably describe the mechanisms of the procedural control theory through anecdotes and examples, many academics have found it extremely difficult to empirically test the main tenets of theory (Huber and Shipan, 2002).36 Do procedures, in fact, enhance fidelity of administrative actions to legislative intent? There is little direct evidence in support (or against) this proposition, and the effectiveness of administrative procedures is thus unclear (but see Balla, 1998; Holborn and Spiller, 2002). Third, critically, this procedures-​as-​control approach relies on the courts to interpret and enforce the procedures (Bressman, 2007). Yet the courts themselves should be viewed as actors in this effort to control the agency. They have preferences over policy (de Figueiredo, 2005), preferences they can realize through their interpretation of procedures (Tiller, 1998; Spiller and Tiller, 1999). These decisions can act directly counter to congressional preferences. Indeed, courts have effectively amended the APA through common law innumerable times, generally with the interests of the court (rather than the legislature) in mind (Bressman, 2007). For example, the rise of hard look review induced agencies to generate considerably more comprehensive records during rulemaking proceedings, seemingly supporting the general purpose of the APA and mitigating the information asymmetries. However, in doing so the court also plausibly substantially raised the bar for any type of rulemaking and thus arguably crippled efforts to implement a variety of legislative initiatives (McGarity, 1992).37 Thus, even as administrative procedures hold promise to enhance congressional control of agencies (McNollgast, 1987), fostering “republican” rulemaking, many important foundations of the perspective remain largely unsubstantiated and contested.

3.4 Discussion To what extent is rulemaking “democratic”? How can we reconcile policymaking by unelected bureaucrats with our commitment to democratic principles?38

36  Huber and Shipan (2002: 36) have noted that the procedural theory of congressional control of the bureaucracy is too general and thus hard to test and difficult to refute. 37  However, note that recent research questions the extent to which such doctrinal developments, in fact, impaired the ability of agencies to produce rules (e.g., O’Connell, 2008). 38  Those who doubt the possibility of democratic rulemaking often argue in favor of a more robust non-​delegation doctrine, which would in principle force the legislature to develop policies itself. The debate over the wisdom and feasibility of this recommendation is vast and outside the scope of this chapter, which assumes the inevitability of delegation and rulemaking. For a recent skeptical entry, consider Stiglitz (2015).

 

52    John M. de Figueiredo and Edward H. Stiglitz Scholars have made many important contributions to our understanding of the possibility of democratic rulemaking. The current literature provides a rich set of possibilities, and we highlight four: the so-​called unitary executive theory, emphasizing presidential control and accountability; the structure-​and-​process school of thought, emphasizing congressional control; an insulation perspective, holding that the public interest and democratic values are often best advanced by retarding political control over administrative agencies; and a deliberative perspective, arguing that rulemaking is the “best hope” for achieving a vision of deliberative democracy. We thus have a fairly rich sense of possibilities. But our sense of the plausibilities is less secure. And we have virtually no certainties. We believe that it is time for the literature of democratic rulemaking to enter a period of self-​assessment, of pruning, and of dismissing. This means more social scientific inquiry into rulemaking: the combination of traditional social scientific tools, with at least a passing competence in the relevant institutional details and administrative law. The most obvious path forward is empirical. For instance, we have little direct evidence on the core questions of whether institutional design of agencies, in fact, promotes policy outcomes consistent with presidential preferences; we have little evidence on whether administrative procedures promote policy outcomes consistent with congressional preferences. Our sense of whether rulemaking procedures, in fact, promote deliberative values—​relative to, say, legislation—​is nearly a void. Legally and institutionally informed empirical studies promise to help us eliminate the most implausible theories currently in circulation. Moreover, the execution of more advanced empirical designs, involving natural and field experiments, hold promise. But theory also holds much promise. We have at present very few formal models that study the democratic properties of rulemaking. For example, only recently have scholars started to study the theoretical relationship between political control and agencies’ incentives to develop (costly) expertise (Stephenson, 2011; Bubb and Warren, 2014; Cameron et al., 2015, 2016; Gailmard and Patty, 2012; see Bawn, 1995 for an early article in this tradition). These contributions sharpen our understanding of the extent to which a fundamental motivation for having an administrative state—​expertise—​might be in necessary tension with democratic values. However, these models do not provide insights on differences in the tradeoffs that presidential versus congressional control might pose; executive and legislative institutions use different tools of control, for example, which likely carry different implications for the incentives of agencies and their civil servants to develop expertise. Finally, the scholars of democratic rulemaking will likely find traction going forward in integrating behavioral perspectives into the legal and economic frameworks to understand agencies. Advances in both theoretical modeling and empirical testing of behavioral approaches provide a basis for understanding more deeply, and in a more integrated way, the decisions of regulators and the participation of individuals and interest groups in the democratic rulemaking process.

 

Democratic Rulemaking   53

Acknowledgment We thank Alex Bolton and Cynthia Farina for very helpful comments.

References Asimow, M. (1994) “On Pressing McNollgast to the Limits: The Problem of Regulatory Costs,” Law and Contemporary Problems, 57: 128–​37. Balla, S. J. (1998) “Administrative Procedures and Political Control of the Bureaucracy,” American Political Science Review, 92: 663–​73. Barkow, R. E. (2010) “Insulating Agencies: Avoiding Capture through Institutional Design,” Texas Law Review, 8: 15–​80. Bawn, K. (1995) “Political Control versus Expertise: Congressional Choices about Administrative Procedures,” American Political Science Review, 89: 62–​73. Beermann, J. M. (2006) “Congressional Administration,” San Diego Law Review, 43: 61–​181. Beierle, T. C. (2004) “Digital Deliberation:  Engaging the Public through Online Policy Dialogues,” in P. Shane (ed.) Democracy Online: The Prospects for Political Renewal through the Internet. New York: Routledge, 155–​67. Berry, C. R. and Gersen, J. E. (2008) “The Unbundled Executive,” University of Chicago Law Review, 75: 1385–​434. Besley, T. and Coate, S. (2003) “Elected versus Appointed Regulators: Theory and Evidence,” Journal of the European Economic Association, 1: 1176–​206. Besley, T. and Coate, S. (2008) “Issue Unbundling via Citizens’ Initiatives,” Quarterly Journal of Political Science, 3: 379–​97. Bolton, A., de Figueiredo, J. M., and Lewis, D. E. (2016) “Elections, Ideology, and Turnover in the U.S. Federal Government,” National Bureau of Economic Research Working Paper #22932. Bonica, A., Chen, J., and Johnson, T. (2015) “Senate Gate-​Keeping, Presidential Staffing of ‘Inferior Offices’, and the Ideological Composition of Appointments to the Public Bureacracy,” Quarterly Journal of Political Science, 10: 5–​40. Brandon, B. H. and Carlitz, R. D. (2002) “Online Rulemaking and Other Tools for Strengthening our Civil Infrastructure,” Administrative Law Review, 54: 1421–​78. Bressman, L. S. (2007) “Procedures as Politics in Administrative Law,” Columbia Law Review, 107: 1749–​821. Bressman, L. S. and Vandenbergh, M. P. (2006) “Inside the Administrative State: A Critical Look at the Practice of Presidential Control,” Michigan Law Review, 105: 47–​99. Breyer, S. G., Stewart, R. B., Sunstein, C. R., Vermeule, A., and Herz, M. (2011) Administrative Law and Regulatory Policy: Problems Text, and Cases (7th ed.). New York: Aspen Publishers. Bubb, R. and Warren, P. L. (2014) “Optimal Agency Bias and Regulatory Review,” The Journal of Legal Studies, 43: 95–​135. Callander, S. (2008) “A Theory of Policy Expertise,” Quarterly Journal of Political Science, 3: 123–​40. Cameron, C. M., de Figueiredo, J. M., and Lewis, D. E. (2015) “Quitting in Protest: A Theory of Presidential Policy-​Making and Agency Response,” Princeton University Working Paper.

 

54    John M. de Figueiredo and Edward H. Stiglitz Cameron, C. M., de Figueiredo, J. M., and Lewis, D. E. (2016) “Public Sector Personnel Economics: Wages, Promotions, and the Competence-​Control Tradeoff,” National Bureau of Economic Research Working Paper #22966. Carey, M. P. (2015) “Counting Regulations:  An Overview of Rulemaking, Types of Federal Regulations, and Pages in the Federal Register,” Congressional Research Service Report. Carlitz, R. D. and Gunn, R. W. (2002) “Online Rulemaking: A Step toward E-​Governance,” Government Information Quarterly, 19: 389–​405. Chambers, S. (2003) “Deliberative Democratic Theory,” Annual Review of Political Science, 6: 307–​26. Clinton, J. D., Lewis, D. E., and Selin, J. L. (2014) “Influencing the Bureaucracy: The Irony of Congressional Oversight,” American Journal of Political Science, 58: 387–​401. Coglianese, C. (1997) “Assessing Consensus:  The Promise and Performance of Negotiated Rulemaking,” Duke Law Journal, 46: 1255–​349. Coglianese, C. (2004) “E-​Rulemaking: Information Technology and the Regulatory Process,” Administrative Law Review, 56: 353–​402. Cox, G. W. (2006) “The Organization of Democratic Legislatures,” in B. R. Weingast and D. Wittman (eds) The Oxford Handbook of Political Economy. Oxford: Oxford University Press, 141–​61. Cramton, R. C. (1971) “The Why, Where, and How of Broadened Public Participation in the Administrative Process,” Georgetown Law Journal 60: 525–​50. Croley, S. (2003) “White House Review of Agency Rulemaking: An Empirical Investigation,” University of Chicago Law Review, 70: 821–​85. Cuéllar, M.-​F. (2005) “Rethinking Regulatory Democracy,” Administrative Law Review, 57: 411–​99. Datla, K. and Revesz, R. (2013) “Deconstructing Independent Agencies (and Executive Agencies),” Cornell Law Review, 98: 769–844. de Figueiredo, J. M. (2005) “Strategic Plaintiffs and Ideological Judges in Telecommunications Litigation,” Journal of Law, Economics and Organization, 21: 501–​23. de Figueiredo, J. M. (2006) “E-​ Rulemaking: Bringing Data to Theory at the Federal Communication Commission,” Duke Law Journal, 55: 969–​93. de Figueiredo, J. M. and Kim, J. K. (2004) “When Do Firms Hire Lobbyists? The Organization of Lobbying at the Federal Communications Commission,” Industrial and Corporate Change, 13: 883–​900. de Figueiredo, J. M. and Richter, B. (2014) “Advancing the Empirical Literature in Lobbying,” Annual Review of Political Science, 17: 163–​85. de Figueiredo, J. M. and Tiller, E. H. (2001) “The Structure and Conduct of Corporate Lobbying: How Firms Lobby the Federal Communications Commission,” Journal of Economics and Management Strategy, 10: 91–​122. de Figueiredo, R. J. P. (2002) “Electoral Competition, Political Uncertainty, and Policy Insulation,” American Political Science Review, 96: 321–​33. de Figueiredo, R. J. P., Spiller, P. T., and Urbiztondo, S. (1999) “An Informational Perspective on Administrative Procedures,” Journal of Law, Economics, and Organization, 15: 283–​305. Dixit, A. K. (1998) The Making of Economic Policy:  A  Transaction-​cost Politics Perspective. Cambridge, MA: MIT Press. Drazen, A. (2004) Political Economy in Macro Economics. Telangana: Orient Blackswan.

 

Democratic Rulemaking   55 Elster, J. (2000) Ulysses Unbound:  Studies in Rationality, Precommitment, and Constraints. New York: Cambridge University Press. Ely, J. H. (1980) Democracy and Distrust: A Theory of Judicial Review. Cambridge, MA: Harvard University Press. Epstein, R. (2016) “The Role of Guidances in Modern Administrative Procedure: The Case for De Novo Review,” Journal of Legal Analysis, 8: 47–​94. Eskridge, W. N. and Frickey, P. P. (1994) “Foreword: Law as Equilibrium,” Harvard Law Review 108: 26–​108. Farber, D. A. and O’Connell, A. (2014) “The Lost World of Administrative Law,” Texas Law Review, 92: 1137–​90. Farina, C. R. (2010) “False Comfort and Impossible Promises: Uncertainty, Information Overload, and the Unitary Executive,” University of Pennsylvania Journal of Constitutional Law, 12: 357–​424. Farina, C. R., Newhart, M. J., Cardie, C., and Cosley, D. (2011) “Rulemaking 2.0,” University of Miami Law Review, 65: 395–​448. Ferejohn, J. and Shipan, C. (1990) “Congressional Influence on Bureaucracy,” Journal of Law, Economics, and Organization, 6: 1–​20. Fiorina, M. P. (1982) “Legislative Choice of Regulatory Forms: Legal Process or Administrative Process?” Public Choice, 39: 33–​66. Fisher, L. (1998) The Politics of Shared Power: Congress and the Executive, Vol. 1. College Station, TX: Texas A&M University Press. Fox, J. and Jordan, S. V. (2011) “Delegation and Accountability,” The Journal of Politics, 73: 831–​44. Gailmard, S. and Patty, J. W. (2007) “Slackers and Zealots: Civil Service, Policy Discretion, and Bureaucratic Expertise,” American Journal of Political Science, 51: 873–​89. Gailmard, S. and Patty, J. W. (2012) “Information and Contestation: A Formal Model of Notice and Comment,” unpublished manuscript. University of California, Berkeley. Gailmard, S. and Patty, J. W. (2013) Learning While Governing: Expertise and Accountability in the Executive Branch. Chicago, IL: The University of Chicago Press. Gersen, J. E. (2010) “Unbundled Powers,” Virginia Law Review, 96: 301–​58. Gersen, J. E. and Stephenson, M. C. (2014) “Over-​Accountability,” Journal of Legal Analysis, 6: 185–​243. Gilligan, T. W. and Krehbiel, K. (1987) “Collective Decisionmaking and Standing Committees: An Informational Rationale for Restrictive Amendment Procedures,” Journal of Law, Economics, and Organization, 3: 287–​335. Grossman, G. M. and Helpman, E. (2001) Special Interest Politics. Cambridge, MA: MIT Press. Hamilton, A. (1982) “Federalist 70,” The Federalist Papers. Hammond, T. H. and Knott, J. H. (1996) “Who Controls the Bureaucracy? Presidential Power, Congressional Dominance, Legal Constraints, and Bureaucratic Autonomy in a Model of Multi-​Institutional Policy-​Making,” Journal of Law, Economics, and Organization, 12: 119–​66. Hirsch, A. V. and Shotts, K. W. (2012) “Policy-​Specific Information and Informal Agenda Power,” American Journal of Political Science, 56: 67–​83. Holburn, G. L. F. and Spiller, P. T. (2002) “Interest Group Representation in Administrative Institutions:  The Impact of Consumer Advocates and Elected Commissioners on Regulatory Policy in the United States,” University of California Energy Institute, University of California, Berkeley.

 

56    John M. de Figueiredo and Edward H. Stiglitz Horn, M. J. and Shepsle, K. A. (1989) “Commentary on ‘Administrative Arrangements and the Political Control of Agencies’: Administrative Process and Organizational Form as Legislative Responses to Agency Costs,” Virginia Law Review, 75: 499–​508. Huber, J. D. and Shipan, C. R. (2002) Deliberate Discretion? The Institutional Foundations of Bureaucratic Autonomy. New York: Cambridge University Press. Huq, A. (2013) “Removal as a Political Question,” Stanford Law Review, 65: 1–​76. Jensen, C. B. and McGrath, R. J. (2011) “Making Rules about Rulemaking: A Comparison of Presidential and Parliamentary Systems,” Political Research Quarterly, 64: 656–​67. Kagan, E. (2001) “Presidential Administration,” Harvard Law Review, 114: 2245–​385. Kerwin, C. M. and Furlong, S. R. (2011) Rulemaking: How Government Agencies Write Law and Make Policy. Washington, DC: CQ Press. Krasnow, E. G., Sidall, D. R., and Berg, M. D. (2001) FCC Lobbying: A Handbook of Insider Tips and Practical Advice. New York: Aspen Publishers. Kwak, J. (2013) “Cultural Capture and the Financial Crisis,” in D. Carpenter and D. A. Moss (eds) Preventing Regulatory Capture. Cambridge: Cambridge University Press, 71–​98. Kydland, F. E. and Prescott, E. C. (1977) “Rules rather than Discretion: The Inconsistency of Optimal Plans,” The Journal of Political Economy, 85: 473–​92. Lewis, D. (2008) The Politics of Presidential Appointments. Princeton, NJ:  Princeton University Press. Light, P. C. (1995) Thickening Government: Federal Hierarchy and the Diffusion of Accountability. Washington, DC: Brookings Institution Press. Lupia, A. and McCubbins, M. D. (1994a) “Designing Bureaucratic Accountability,” Law and Contemporary Problems, 57: 91–​126. Lupia, A. and McCubbins, M. D. (1994b) “Learning from Oversight: Fire Alarms and Police Patrols Reconstructed,” Journal of Law, Economics, and Organization, 10: 96–​107. Madison, J. (1788) “Federalist No. 63,” Hamilton, Madison, Jay, The Federalist Papers (1788): 387–​8. Magill, M. E. (2004) “Agency Choice of Policymaking Form,” The University of Chicago Law Review, 71: 1383–​447. Magill, E. and Vermeule, A. (2011) “Allocating Power within Agencies,” The Yale Law Journal, 120: 1032–​83. Majone, G. (1996) “Temporal Consistency and Policy Credibility:  Why Democracies Need Non-​Majoritarian Institutions,” manuscript, European University Institute. Majone, G. (1997) “From the Positive to the Regulatory State: Causes and Consequences of Changes in the Mode of Governance,” Journal of Public Policy, 17: 139–​67. Mashaw, J. L. and Harfst, D. L. (1986) “Regulation and Legal Culture: The Case of Motor Vehicle Safety,” Yale Journal on Regulation, 4: 257–​316. McCubbins, M. D., Noll, R. G., and Weingast, B. R. (1987) “Administrative Procedures as Instruments of Political Control,” Journal of Law, Economics, and Organization, 3: 243–​77. McCubbins, M. D., Noll, R. G., and Weingast, B. R. (1989) “Structure and Process, Politics and Policy: Administrative Arrangements and the Political Control of Agencies,” Virginia Law Review, 75: 431–​82. McCubbins, M. D., Noll, R. G., and Weingast, B. R. (1999) “The Political Origins of the Administrative Procedure Act,” Journal of Law, Economics, and Organization, 15: 180–​217. McCubbins, M. D. and Schwartz, T. (1984) “Congressional Oversight Overlooked: Police Patrols versus Fire Alarms,” American Journal of Political Science, 28: 165–​79.

 

Democratic Rulemaking   57 McGarity, T. O. (1992) “Some Thoughts on ‘Deossifying’ the Administrative Law Process,” Duke Law Journal, 41: 1358–​462. McGarity, T. O. and Wagner, W. E. (2008) Bending Science: How Special Interests Corrupt Public Health Research. Cambridge, MA: Harvard University Press. Mendelson, N. A. (2011) “Foreword: Rulemaking, Democracy, and Torrents of E-​Mail,” George Washington Law Review, 79: 1343–​80. Moe, T. M. (1990) “Political Institutions:  The Neglected Side of the Story,” Journal of Law, Economics, and Organization, 6: 213–​53. Nou, J. (2013) “Agency Self-​Insulation under Presidential Review,” Harvard Law Review, 126: 1755–​2479. Nou, J. and Stiglitz, E. H. (2016) “Strategic Rulemaking Disclosure,” Southern California Law Review, 89: 733–​92. Noveck, B. S. (2004) “The Electronic Revolution in Rulemaking,” Emory Law Journal 53: 433–​94. Nzelibe, J. (2005) “The Fable of the Nationalist President and Parochial Congress,” UCLA Law Review, 53: 1217–​74. O’Connell, A. J. (2008) “Political Cycles of Rulemaking: An Empirical Portrait of the Modern Administrative State,” Virginia Law Review, 94: 889–​986. Prakash, S. (2003) “The Essential Meaning of Executive Power,” University of Illinois Law Review, 2003: 701–​820. Rogoff, K. (1985) “The Optimal Degree of Commitment to an Intermediate Monetary Target,” The Quarterly Journal of Economics, 100: 1169–​89. Ryfe, D. M. (2005) “Does Deliberative Democracy Work?” Annual Review of Political Science, 8: 49–​7 1. Schlosberg, D., Zavestoski, S., and Shulman, S. W. (2007) “Democracy and E-​Rulemaking: Web-​Based Technologies, Participation, and the Potential for Deliberation,” Journal of Information Technology and Politics, 4: 37–​55. Schoenbrod, D. (2008) Power without Responsibility: How Congress Abuses the people through Delegation. New Haven, CT: Yale University Press. Seidenfeld, M. (1992) “A Civic Republican Justification for the Bureaucratic State,” Harvard Law Review, 105: 1511–​76. Shapiro, M. (1986) “APA: Past, Present, Future,” Virginia Law Review, 72: 447–​92. Shepherd, G. B. (1996) “Fierce Compromise:  The Administrative Procedures Act Emerges from New Deal Politics,” Northwestern University Law Review, 90: 1557–​601. Shepsle, K. A. (1992) “Bureaucratic Drift, Coalitional Drift, and Time Consistency: A Comment on Macey,” Journal of Law, Economics, and Organization, 8: 111–​18. Shulman, S. W., Schlosberg, D., Zavestoski, S., and Courard-​Hauri, D. (2003) “Electronic Rulemaking: A Public Participation Research Agenda for the Social Sciences,” Social Science Computer Review, 21: 162–​78. Spiller, P. and Tiller, E. (1997) “Decision Costs and the Strategic Design of Administrative Processes and Judicial Review,” Journal of Legal Studies, 26: 347–​70. Stanley, J. W. and Weare, C. (2004) “The Effects of Internet Use on Political Participation Evidence from an Agency Online Discussion Forum,” Administration and Society, 36: 503–​27. Stanley, J. W., Weare, C., and Musso, J. (2004) “Participation, Deliberative Democracy, and the Internet: Lessons from a National Forum on Commercial Vehicle Safety,” in P. Shane (ed.) Democracy Online: The Prospects for Democratic Renewal through the Internet. New York and London: Routledge, 167–​80.

 

58    John M. de Figueiredo and Edward H. Stiglitz Stephenson, M. C. (2011) “Information Acquisition and Institutional Design,” Harvard Law Review, 124: 1422–​83. Stigler, G. J. (1971) “The Theory of Economic Regulation,” Bell Journal of Economics and Management Science, 2: 3–​21. Stiglitz, E. H. (2014a) “Unaccountable Midnight Rulemaking:  A  Normatively Informative Assessment,” NYU Journal of Legislation and Public Policy, 17: 137–​92. Stiglitz, E. H. (2014b) “Unitary Innovations and Political Accountability,” Cornell Law Review, 99: 1133–​84. Stiglitz, E. H. (2015) “The Limits of Judicial Control and the Nondelegation Doctrine,” manuscript, Cornell University. Stiglitz, E. H. (2016) “Folk Theories, Dynamic Pluralism, and Democratic Values,” manuscript, Cornell University. Strauss, P. L. (1996) “Changing Times: The APA at Fifty,” The University of Chicago Law Review, 63: 1389–​422. Thompson, D. F. (2008) “Deliberative Democratic Theory and Empirical Political Science,” Annual Review of Political Science, 11: 497–​520. Tiller, E. (1998) “Controlling Policy by Controlling Process: Judicial Influence on Regulatory Decision Making,” Journal of Law, Economics, and Organization, 14: 114–​35. Wagner, W., Barnes, K., and Peters, L. (2011) “Rulemaking in the Shade: An Empirical Study of EPA’s Air Toxic Emission Standards,” Administrative Law Review, 63: 99–​158. Weingast, B. R. (1984) “The Congressional-​ Bureaucratic System: A Principal–​ Agent Perspective (with Applications to the SEC),” Public Choice, 44: 147–​91. Weingast, B. R. and Moran, M. J. (1983) “Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission,” The Journal of Political Economy, 91: 765–​800. Wilson, W. (1885) Congressional Government:  A  Study in American Politics. Boston, MA: Houghton Mifflin. Yackee, J. W., and Yackee, S. W. (2006) “A Bias towards Business? Assessing Interest Group Influence on the US Bureaucracy,” Journal of Politics, 68: 128–​39. Yackee, J. W., and Yackee, S. W. (2011) “Testing the Ossification Thesis: An Empirical Examination of Federal Regulatory Volume and Speed, 1950–​1990,” George Washington Law Review, 80: 1414–​92. Yoo, C. S., Calabresi, S. G., and Colangelo, A. J. (2005) “The Unitary Executive in the Modern Era, 1945–​2004,” Iowa Law Review, 90: 601–​732.

 

Chapter 4

Regu l atory De cision-​M a k i ng a nd Ec onom ic A na lysi s Mariano-​F lorentino Cuéllar and Jerry L. Mashaw

4.1 Introduction The economic analysis of regulation is a vast topic, with implications for environmental protection, communications and technology policy, public health, immigration, national security and other areas affecting risk and welfare in society. In this brief chapter we can cover only a portion of the relevant ground. We plan to narrow the focus of our discussion accordingly, by limiting ourselves to the following essential topics: First, what do we mean by “economic analysis” and what do we mean by “regulation”? Second, why has this topic become an important one, not only the United States, but in most advanced democracies? Third, why is economic analysis and regulation a contested, even contentious, aspect of modern regulatory activity? Finally, and most important, how is economic analysis structured into regulatory decision-​making, and how might existing arrangements evolve over time? This last question is essentially one of institutional design. It encompasses at least the following more specific issues that cut across substantive domains: (1) In what circumstances should economic analysis be applied to regulatory decision-​making? (2) Who, that is, what governmental instrumentalities, should specify the content of the economic analysis that is required? (3) When economic analysis is required as a part of regulatory decision-​making, what role should it play? (4) What sort of accountability or oversight structures should be provided to insure that required economic analyses are done, take the appropriate form and have the specified effects on regulatory decision-​ making. Finally, (5) what are the effects of using economic analysis of regulation on

 

60    Mariano-Florentino Cuéllar and Jerry L. Mashaw other public values, such as, the transparency of public policymaking and democratic participation in the making of public policy? To be sure, even this more narrowly drawn specification of our topic remains large. We will be painting with rather broad brushstrokes. The subtleties of texture and hue, and even some of the details of the images, will be supplied by the literature and cases to which we will refer. As we develop this picture, we hope to elucidate why a familiarity with both the promise and the limits of economic analysis is crucial to any serious understanding of regulation—​whether prescriptive or descriptive—​in advanced economies. In practice, the aspirations of economic analysis in the regulatory state must be reconciled with various administrative and legal arrangements necessary to advance values, such as adjudicatory fairness or political responsiveness, that are often in tension with the technical rigor to which economic analysis aspires. Accordingly, the picture we paint is one where the resolution of some of the most difficult issues associated with economic analysis of regulation depend on institutional factors that define how analyses are conducted, whose interests are weighed in the process, and what standards are used by pivotal decision-​makers responding to economic analyses in an uncertain world.

4.2  Some Definitional Limitations The law regulates the conduct of individuals and organizations in a variety of ways. The criminal law and much of the private law of torts, contracts, and property regulate human behavior either explicitly or implicitly. Economic analysis clearly can be applied to these domains of law, as other chapters in this handbook will attest. Our concern here, however, is with forms of bureaucratic regulation that are characteristic of advanced economies. To oversimplify a bit, we are concerned with the standard model of regulatory activity that is initiated by legislation, implemented largely by administrative institutions with jurisdiction (and, presumably, specialization) relevant to a particular regulatory task, and supervised by various forms of legal, political, and managerial accountability. Moreover, most of our examples will be taken from what is generally characterized as “risk regulation.” Modern administrative states seek to protect their citizens from many pervasive risks to health, safety, financial security, and a host of other values or interests. The economic analysis of regulation that we will be discussing is addressed to the economic welfare consequences of risk regulation broadly construed.1 Economic analysis itself can take a number of forms as applied to regulatory decision-​ making. Regulators might be asked to analyze the effects of their regulations on competition, on job creation, or simply costs of compliance. Analysis of compliance costs might further specify that the administrators should be particularly concerned with

1 

Croley (2007), Skrzycki (2003), Kerwin (2003).

 

Regulatory Decision-Making and Economic Analysis    61 effects on individuals or small entities,2 or with particular types of costs, such as paperwork burdens.3 The ambitions of economic analysis might also be limited in various ways. An agency might be instructed merely to determine whether its proposed regulations are economically feasible in the sense that regulated parties can comply without going bankrupt.4 Agencies might be asked to engage in some form of “cost effectiveness” analysis. Economic analysis of this type asks questions such as, given the goal of reducing X risk by 50 percent, what is the cheapest method of achieving that goal? Or, alternatively, given that we are willing to spend X on risk reduction, what method of risk reduction will give us the largest bang for the specified bucks? Our concern here, however, will be with perhaps the most ambitious form of economic analysis of regulatory action—​cost–​ benefit analysis. We call this form of economic analysis ambitious because it tends to ask a more comprehensive question: Tell us all of the potential costs and all of the potential benefits from your particular regulatory policy, and then compare those costs and benefits by reducing them to a single metric, US dollars or some other local currency.5 Simply to state the demands of cost–​benefits analysis in this way is to reveal its potential information demands and perhaps to hint at some of the conceptual conundra that are likely to arise in implementing this ambitious, and some would say reductionist, program.6 Indeed, regulatory analysis under Executive Order 13563 in the United States is more ambitious still. In a crucial paragraph it provides: As stated in [a previous executive order] and to the extent permitted by law, each agency must, among other things: (1) propose or adopt a regulation only upon a reasoned determination that its benefits justify its costs (recognizing that some benefits are difficult to quantify); (2) tailor its regulations to impose the least burden on society, consistent with obtaining regulatory objectives, taking into account, among other things, and to the extent practicable, the costs of cumulative regulations; (3) select, in choosing among alternative regulatory approaches, those approaches that maximize net benefits (including potential economic, environmental, public health and safety, and other advantages; distributive impacts; and equity); (4) to the extent feasible, specify performance objectives, rather than specifying the behavior or manner of compliance that regulated entities must adopt; and (5)  identify and assess available alternatives to direct regulation, including providing economic incentives to encourage the desired behavior, such as user fees or marketable permits, or providing information upon which choices can be made by the public.7

2  The Regulatory Flexibility Act, Pub. L. No. 96-​354, 94 Stat. 1164 (1980). (Codified at 5 USC Sec. 601 at seq.) 3  The Paperwork Reduction Act, Pub. L. No. 96-​511, 94 Stat. 2812 (1980). (Codified at 44 USC Sec. 3501 at seq.) 4  Industrial Union Dept. v. American Petroleum Inst., 448 U.S. 607 (1980). (Interpreting the Occupational Safety and Health Act). 5  Mishan (1994). 6  Adler and Posner (2006); McGarity (1991). 7  Executive Office of the President (2011).

 

62    Mariano-Florentino Cuéllar and Jerry L. Mashaw

4.3  Why Cost–​Benefit Analysis? Even the relatively complex formulation embodied in the preceding executive order begins with a basic consequentialist idea that is both intuitively and analytically appealing. Indeed, in one sense, cost–​benefit calculations seem simply a part of a ubiquitous human decision-​making practice. We constantly compare the benefits that we expect from consumer goods against the prices that sellers are asking for them. And, we often do more extended cost–​benefit calculations when we face important and complicated issues in our lives. Anyone who has made a list of reasons for and against taking a particular course of action has engaged in a rudimentary form of cost–​benefit analysis. We cannot have everything; scarcity is our fate. Even Bill Gates’ time is scarce; we must make trade-​offs. But, obviously we can engage in cost–​benefit calculations without actually doing mathematics or monetizing the value of the costs and benefits that we perceive. Whether we should engage in this particular form of economically oriented cost–​benefit analysis remains contested in many areas. Yet there appear to be at least two important reasons for monetization. One is conceptual, the other historical and political. The conceptual case is pretty straightforward: It is hard, as the Americans say, to compare apples and oranges. (“Chalk and cheese” seems to be the British locution and academic economists have for many years seemed to prefer “guns and butter.”) If we want to know if the costs are greater than the benefits, or vice versa, it is difficult to do so unless we express them in the same metric. Regulatory agencies engage in similar analyses, but tend to do so in a rather more complicated and contentious environment than individuals making ordinary consumption decisions. Consider the following example: Industrial and commercial boilers provide considerable value to companies and individuals because they produce heat and are useful in industrial processes. But they also release particulate matter into the air that can cause individuals to suffer adverse health effects increasing mortality and morbidity. Presented with data to this effect, an agency with statutory authority to regulate these boilers undertakes an economic analysis of the costs and benefits of a particular regulatory rule. In the course of analyzing a variety of statutorily permissible alternatives, the agency weighs factors such as the costs to industry of replacing a variety of boilers, the kinds of particulate matter released by different boilers, the health consequences to the population of the increased pollution, and potential benefits in reduced adverse health consequences. Because these benefits and costs arise over time, the agency must consider how much to weigh benefits that arise in the future relative to costs that arise immediately from industry’s purchase of the new boilers. In order to render these different factors comparable, the agency seeks (to the greatest extent possible) to convert these different costs and benefits into a comparable unit of analysis—​money. The resulting analysis could conceivably help not only reduce (without eliminating) uncertainty about the consequences of the rule, but perhaps also mitigate slightly the extent of political conflict over the policy in question through the provision of relatively objective information.

 

Regulatory Decision-Making and Economic Analysis    63 Economic analysis of regulatory policies bearing some similarity to this example can bring a measure of clarity to real-​world regulatory decisions, particularly given the risks of biases and limitations in conventional individual decision-​making.8 The impact of a particular regulatory policy on social welfare is the core normative criterion in many prescriptive theories of regulation,9 and even scholars who question whether regulatory policies should be evaluated relative to multiple criteria would be hard-​pressed to argue against some consideration of social welfare effects.10 Standard definitions of cost–​benefit analysis in the regulatory context presume that both costs and benefits are expressed in monetary units, so that these can be compared.11 Yet the value and limits of monetization as an analytical technique sparks a measure of controversy among some observers of the regulatory process.12 On the one hand, monetization of regulatory benefits seems to provide a means of escaping the question of how to define the context of regulatory analysis involving particular rules. Money’s fungibility, at least in principle, allows for the monetized benefits of toxic spill prevention to be compared to equivalent figures involving air safety improvements, counter-​terrorism initiatives at the border, securities market improvements, and better prescription drugs. Yet the technical choices involved in monetization (including valuation techniques where markets do not exist and determining discount rates on the valuation of statistical lives) are often elided by the apparent simplicity of the comparison. Nonetheless, it would be difficult to argue that an agency should make a decision about whether (or how) to regulate industrial and commercial boilers without some effort to account for the economic costs and benefits. To the extent that this analytical technique is assigned substantial importance in the government’s ultimate policy decision (by some combination of statute, executive practice, court review, or informal norm), it becomes particularly important to understand how costs and benefits would be identified, who would be involved in the process, and what broader consequences would arise from different approaches to these questions. To better understand how this process might play out in practice, consider a real-​ world example involving a new federal regulatory rule governing boilers in the United States. In December 2012, the Environmental Protection Agency (EPA) finalized its revised hazardous air pollutant standards for industrial, commercial, and institutional boilers and process heaters under the Clean Air Act. The rule prescribes numerical emission limits for carbon monoxide, hydrochloric acid, mercury, and particulate or filterable particulate matter.13 As one of the Obama Administration’s most expensive regulations, the rule received close scrutiny of its projected costs and benefits. EPA calculated total costs at $2.2 billion to $2.4 billion per year, and benefits that would be 8 

Sunstein (2000). Adler and Posner (2006). 10  Richardson (2000). 11  Mishan (1994). 12  Adler and Posner (2006) and Ackerman and Heinzerling (2004). 13  Summary of Economic Analysis in the Federal Register, http://​www.gpo.gov/​fdsys/​pkg/​FR-​2013-​01-​ 31/​pdf/​2012-​31646.pdf. Additional details at: http://​www.epa.gov/​airquality/​combustion/​actions.html. 9 

 

64    Mariano-Florentino Cuéllar and Jerry L. Mashaw 13 to 29 times that amount. The total monetized benefits reflected the health benefits associated with reducing exposure to air toxins. The models used assumed that all fine particles, regardless of their chemical composition, were equally potent in causing premature mortality.14 EPA explained that the scientific evidence was not yet sufficient to support the development of differential effects estimates by particle type. EPA did not quantify the benefits from reducing ecosystem impairment, visibility impairment, and exposure to hazardous air pollutants. The EPA’s analysis did not garner universal agreement. Capital costs are the most expensive element of the rule, although there will also be testing and monitoring costs, and other annual operating costs. While EPA projected $5 billion in total capital costs, a study by HIS Global Insight on behalf of the Congress of Industrial Boiler Owners (CIBO) calculated capital costs of $14 billion. Also, EPA predicted that the regulation would result in a small net increase in jobs (2,600 to 5,400 jobs), while the American Forest and Paper Association claimed that the regulation might cost 40,000–​60,000 jobs. Including the effects on suppliers and other industries, CIBO claimed 230,000 jobs may be at risk. The Congressional Research Service, for its part, contended that the industry cost estimates were exaggerated.15 A researcher at the W. E. Upjohn Institute for Employment Research noted that “the industry’s figures ignore the possibility that the consumer demand and capital involved in these regulated industries will go somewhere else in the economy.” EPA acknowledged that potential control device cost-​savings and increased record-​keeping and reporting costs associated with the emissions averaging provisions in the final rule were not accounted for in either the capital or annualized cost estimates. The distinctions between the CIBO estimates and those of the EPA highlight some of the recurring analytical issues that arise in understanding the economic consequences of regulatory policies. While EPA calculated cost for compliance values derived using the OAQPS Cost Control Manual Algorithms developed in the 1990s, CIBO calculated estimated costs based on updated vendor estimates. Also, whereas EPA made assumptions that adequate control could be achieved through the use of low-​NOx burners and tune-​ups for CO, fabric filters for Hg, and packed bed scrubbing for HCl, “CIBO analyses made more realistic assumptions for controls needed to comply.”16 Industry groups and other governmental agencies were not the only commentators on these issues. The National Resources Defense Council (NRDC) challenged CIBO’s calculations.17 First, it contended that CIBO did not offer a sufficient justification for its price elasticity calculation regarding changes in demand for regulated goods in response to increased compliance costs (“instead of using standard economic methodology, the authors create an arbitrary metric that bears no relationship to economic theory”). Environmental advocates also noted that the analysis of costs and benefits was too narrowly focused on the boilers, and failed to account for the economic benefits of money 14  Summary of Economic Analysis in the Federal Register, http://​www.gpo.gov/​fdsys/​pkg/​FR-​2013-​01-​ 31/​pdf/​2012-​31646.pdf. 15  McCarthy (2012). 16  See generally. CIBO materials: http://​www.cibo.org/​issues.htm. 2012 CIBO comment on proposed rule: http://​www.cibo.org/​pubs/​0790_​feb21.pdf. 17  http://​switchboard.nrdc.org/​blogs/​ljohnson/​boiler_​industrys_​junk_​analysis.html.

 

Regulatory Decision-Making and Economic Analysis    65 that might not be spent on boilers but could be spent on other goods and services. The NRDC also questioned the industry’s assumption that abatement would occur in the first year rather than being amortized, and failed to explain the assumption that output losses were equally likely at any regulated facility. For now, notice three important aspects of this real-​world example. First, while the conceptual case for doing cost–​benefit analysis, is straightforward, doing cost–​benefit analysis is not, even as a technical matter. Second, participants in this process argued their cases in highly technocratic terms. Basic questions about valuing lives saved or the appropriateness of monetizing the benefits illnesses avoided and comparing these figures with the costs of factors are submerged. Finally, and most obviously, cost–​benefit analysis plays out in a highly political context. How technical controversies and analytic uncertainties are ultimately resolved will depend on how the decisional structure is organized.

4.3.1 The Political Origins of Cost–​Benefit Analysis The historical rationale for cost–​benefit analysis is more complicated, and to some degree elusive. The story seems to go something like this: Much of regulation in the postNew Deal era was explicitly economic regulation. Various commissions and boards regulated the rates of common carriers, attempted to promote harmonious bargaining between labor and capital, and sought to provide transparency to both consumer and financial markets. In the 1960s and 1970s our regulatory attention shifted. The economy had performed magnificently in the post-​WWII period. The public’s and political leaders’ concerns from making markets work shifted to questions of the equitable distribution of economic output, equal access to economic opportunity, and protections against the risks that accompanied economic growth. Legislation in the 1960s and 1970s, therefore, featured, first, equal opportunity legislation focused on groups that seemed to have been denied equal participation in the post-​war boom—​racial minorities, women, older Americans, and persons with disabilities. And second, Congress adopted far-​reaching regulatory legislation dealing with newly perceived risks from pollution, foods and drugs, work place hazards, and transportation. In its scope and reach regulatory legislation in the so called “Great Society” period vastly exceeded the ambitions of Progressive or New Deal regulatory legislation.18 By the 1980s, indeed beginning in the mid 1970s, the heyday of economic expansion and ever-​rising standards of living seemed to be over. Public discourse featured words like “stagflation” and “the misery index.” Both referred to the undeniable fact that both inflation and unemployment seemed to be rising at the same time, a macro-​economic rarity. What had happened? Without attempting to sort out its many sources and their respective influences, some combination of academic economic analysis, conservative think-​tank 18 

On the general development of federal regulation in the United States, see Rabin (1986).

 

66    Mariano-Florentino Cuéllar and Jerry L. Mashaw publicizing, and political discourse combined to promote a relatively simple idea:19 We had tried to do too much too fast. Public spending by legislatures had a built-​in counterweight, the requirement to raise revenues such that budgets remained in sufficient long-​term balance that interest rates for public borrowing did not become prohibitive. Regulators had budgets for how much they could spend on personnel, laboratories, research, and the like, but they had no budgets constraining the costs they could impose on regulated parties—​and thus on the economy as a whole. We were attempting to reap the social benefits of greater equality of opportunity, and to reduce risks from everyday hazards to health and well-​being, without reckoning the costs of those social gains. To be sure no one knew definitively whether the costs of regulatory action were a major contributor to economic slowdown. What was clear was (1)  that students of regulatory policy could find some relatively obvious examples of expensive regulatory requirements that seemed to have modest payoffs, and (2)  that the legislative charters under which regulators plied their trades rarely asked them to consider the costs or to compare those costs with the potential benefits from their regulatory actions. Requiring them to do so seemed almost a self-​evidently good idea. Requirements that regulatory agencies consider cost and benefits when adopting regulatory policy was in some sense modeled on a pre-​existing government-​ wide analytic mandate. The National Environmental Policy Act of 1970 (NEPA)20 required every federal agency to prepare an analysis of the environmental impacts of its actions unless the agency could certify that the action contemplated had no significant environmental effects. The NEPA sought to counter the tunnel vision or mission orientation of federal agencies whose jobs were developmental, particularly infrastructure agencies like the Department of Transportation or the Corps of Engineers. Requirements for cost–​benefit analysis are essentially the mirror image of the NEPA’s environmental impact requirements. Agencies that were attempting to curb the health, safety, and environmental fallout from economic activity would be asked to consider how those curbs impacted the economy. With cost–​benefit analyses in hand we could seriously ask the question whether the benefits justified the costs. Moreover, to the extent that this analysis was meant to constrain agency decision-​ making, quantitative analysis seemed to be important. If agencies could merely speculate in a qualitative fashion about costs and what the benefits might be, they could easily view the benefits from carrying out their mission as worth costs they identified, but left unquantified. If analysis was to be rendered “objective” and therefore subject to oversight by political or legal controllers, or the public at large, agencies could not be left to talk about apples and oranges. The question needed to be put in terms of dollars and cents.

19 

20 

See generally McGarity (2013). Pub. L. No. 91-​190, 83 Stat. 852 (1970) (codify at 42 USC §4321 at sec).

 

Regulatory Decision-Making and Economic Analysis    67

4.4  Designing a Cost–​Benefit Regime Perhaps the first question is where cost–​benefit analysis is appropriate. Anywhere and everywhere is surely not the answer. Some questions simply do not lend themselves to this particular analytic technique. Or perhaps we should say, some questions do not lend themselves to the sort of quantitative and monetized cost–​benefit analysis under discussion here.21 We are not likely to try to assign monetary values to protecting the rights of minorities to cast their votes in elections. Nor do we think it morally appropriate to tote up the costs and benefits in monetary terms of requiring hospital emergency rooms to stabilize patients who appear at their doorsteps with life-​threatening conditions. In a host of circumstances, normative claims sounding in democracy, equality, or equity will simply trump normative claims based on considerations of economic welfare. The political process, constitutional considerations, and practical problems of implementation will also complicate any answer to whether cost–​benefit analysis should be required of a particular regulatory action. American presidents reaching back to Jimmy Carter have imposed increasingly elaborate cost–​benefit requirements on executive agencies, subject to centralized review by some entity in the Executive Office of the President. Centralized review took something like its current form in 1980. Shortly after entering the White House, President Reagan signed Executive Order 12291 requiring agencies to weigh the costs and benefits of proposed regulations, subject to review by an Office of Information and Regulatory Affairs (OIRA) within the Office of Management and Budget (OMB). Subsequently, Presidents Clinton, George W.  Bush, and Obama each updated the Order to reflect their administration’s priorities. For example, within the broad framework of an economic analysis of the costs and benefits of regulatory rules, President Clinton (Executive Order 12866) and President Obama (Executive Order 13563) allowed for the consideration of “equity” and “fairness,” values acknowledged to be difficult to quantify, while President Bush (Executive Order 13422) focused agencies’ analysis on variables more readily susceptible to numerical valuation. Administrative cost considerations have limited those requirements to “major federal actions.” And, although there is no conceptual distinction of moment between many of the determinations of the Federal Communications Commission (FCC) and the Department of Transportation, the FCC and a number of other so-​called independent agencies are exempt from these presidentially imposed cost–​benefit requirements because of separation-​of-​powers concerns. Nevertheless, a 2013 report found that independent agencies do often quantify at least some of the costs (and, less frequently, the benefits) that result from the major rules they adopt.22

21  Some commentators believe that very few regulatory questions actually lend themselves to resolution by cost–​benefit analysis. See, e.g., Ackerman and Heinzerling (2004). 22  Copeland (2013).

 

68    Mariano-Florentino Cuéllar and Jerry L. Mashaw Moreover, congressional legislation trumps executive requirements placed on regulatory agencies save in those rare instances that the president has an independent constitutional authority not subject to congressional control or limitation. Hence, modern executive orders demanding cost–​benefit analysis by executive agencies exempt those functions for which such a requirement would be inconsistent with governing legislation. The Occupational Safety and Health Administration (OSHA), for example, does not do cost–​benefit analyses because its statute requires “feasibility analysis” instead. And, in setting ambient air quality standards the EPA is enjoined to consider simply the necessities of public health, a limitation that the Supreme Court has ruled excludes cost–​benefit analysis.23 As suggested earlier, legislatures may prefer to require special forms of economic analysis, such as the costs imposed on individuals or small entitles, or the costs imposed by record-​keeping and information provision requirements, rather than full-blown cost–​benefit analyses. Legislation may also demand cost–​benefit analysis, or something very much like it, in circumstances where quantification is virtually impossible. As we shall see in the discussion below of judicial review of regulatory cost–​benefit analysis, the Securities and Exchange Commission (SEC) has had particular difficulty in recent years in satisfying reviewing courts that it has done an adequate economic analysis because its statute demands, of rules that might more properly be understood to be as aimed at fairness or shareholder democracy than at economic efficiency.

4.4.1 Who Decides? Where cost–​benefit analysis is to be applied is obviously a political question—​a question which raises yet another design issue: Who decides? As the previous discussion suggests the answer to that question depends importantly on constitutional arrangements in a particular state, as well as the more micro-political concerns surrounding particular regulatory issues. Although few American observers can imagine the US Congress imposing cost–​benefit analysis requirements on itself, those requirements apply to legislation as well as to agency regulatory actions in some jurisdictions. And, in some sense regulators always get to decide for themselves whether cost–​benefit analysis requirements apply to particular regulatory actions, at least as an initial matter. Whether those agency determinations are decisive depends upon the accountability regime that is constructed to oversee agency compliance. Other jurisdictions face many of the same choices that American institutions have faced regarding how to establish the institutional and analytical foundations for cost–​ benefit analysis. Australia, for example, has required Regulatory Impact Assessments since 1985 and has earned praise for the expertise it has developed and the quality of its analyses.24 In contrast to Australia, Canada has a somewhat more flexible framework 23  24 

Whitman v. American Trucking Assoc., Inc., 531 U.S. 457 (2001). OECD (2010).

 

Regulatory Decision-Making and Economic Analysis    69 that leaves more of the implementing details of the analysis to agencies and requires an analysis of the equity implications of regulatory policies. Canada’s standard for the evaluation for all significant regulatory proposals does not purport to prescribe every aspect of the analysis.25 Regulators must nonetheless demonstrate that economic, environmental, and social benefits to Canadians, business, and government over time outweigh the costs, and that net benefits are maximized.26 In South Korea, the Program for Reduction of Administrative Burden on Businesses established the principle of cost–​benefit analysis and required agencies to conduct regulatory impact analysis for major regulations.27 The regulatory agency is in charge of the assessment, but regulations are reviewed by the Regulatory Reform Commission, co-​chaired by the Prime Minister.28 Mexico began requiring cost–​benefit analysis for some regulations in 1992, although early quantification efforts were widely criticized.29 Subsequent reforms in 1996, 2000, and 2010 broadened and strengthened the requirements. Meanwhile, the United Kingdom’s regulatory assessment regime is among the most far-​reaching and rigorous, but took longer to assume its current institutional form.30 UK policy clearly states that regulatory proposals should only be accepted when the benefits justify the costs.31 In contrast to the American system’s focus on regulatory rules rather than statutes, the British system also applies to legislation. These and other jurisdictions have deployed economic analysis to enhance the fit between regulatory policies and societal needs. Nonetheless, the merits of each arrangement will likely remain subject to continuing debate regarding the range of acceptable techniques that can be used, the basis for evaluating the use of different time horizons, the potential risks that some techniques will run by creating a false illusion of objectivity, thereby reducing transparency and concealing the contingent and judgmental nature of some elements in a cost–​benefit analysis.32

4.4.2 What Is Required? Nestled in the question of who decides whether or where cost–​benefit analysis is to be required is the question of what the content of a cost–​benefit analysis should be. To put the question slightly differently: precisely what methodology of cost–​benefit analysis should be deployed. How are costs or benefits to be identified, quantified, and compared? In the United States presidential executive orders have over time become increasingly specific about the contents of regulatory cost–​benefit analyses. And the executive branch oversight bureau, OIRA in OMB has issued increasingly specific guidelines 25 

Treasury Bd. of Canada Secretariat (2007a). Treasury Bd. of Canada Secretariat (2007b). 27  Jeong (2013). 28  Korean Stat. No. 5368, Ch. 4 §22 (Enacted Aug. 22, 1997). 29  OECD (1999). 30  OECD (2002). 31  Taoiseach (2004). 32  Cf. Heinzerling (1998). 26 

 

70    Mariano-Florentino Cuéllar and Jerry L. Mashaw about how cost–​benefit analyses are to be done. At one level these questions of content or methodology are highly technical issues of economic analysis. At another level they are deeply political. To illustrate this point we will content ourselves with a few prominent examples. What value to put on lives saved by risk regulation is one of the most perplexing and contentious issues in cost–​benefit analysis. To be sure there are purely technical issues. If we were to decide that the appropriate measure of the value of lives saved was the risk premium that workers demand who are engaging in risky occupations, there are still a number of difficult technical, methodological issues about how to produce those estimates. But, the deeper and more contentious issues have to do with what might constitute a good measure of the value of a human life. Some would say that there is no such measure, all lives are equally and infinitely valuable. Were cost–​benefit analysis to accept that position it would be out of business wherever regulation had life-​saving potential. The infinite value position has been rejected, perhaps sensibly. Our ordinary practices belie a consistent belief in the infinite value of all lives. We engage in a host of individual activities and collective programs that we know will produce some, often statistically predictable, loss of human life. Hence the risk premium in risky employment is but one of myriad implicit valuations that might be attached to human life. We could, for example, calculate the implicit value of lives predictably lost in implementing infrastructure projects, or divide the known costs of pre-​existing regulatory or public health programs by the number of lives saved. The problem is that these implicit valuations do not just vary a little; they may vary by factors of thousands. Which numbers should be chosen? The question gets murkier still when we begin to ask whether we should take into account whose lives are being saved. Are the lives of low-​wage workers less valuable than the lives of the better compensated because less lifetime income is lost when their lives are cut short? Should we be looking at lives saved or numbers of years of life saved? Does it matter whether the lives saved are those of the healthy and fully functional as against those of the ill and disabled? In short should the length and quality of life be included in our estimation of its value? These are obviously not questions of economic method. They are questions of moral, political, and even religious belief. Those who do cost–​ benefit analysis must somehow arrive at a working definition or means of evaluation, but whatever they decide they inevitably paper over a host of deep issues about which reasonable people continue to disagree. The question of how to discount future benefits to present value has the same dual technical/​normative character. At one level this is just a question of what interest rate should be applied when benefits that will be realized tomorrow must be paid for today. Obviously in our economic lives we do not treat dollars available today as the equivalent of dollars available tomorrow or next year. If you want to borrow money and pay it back later, you will have to pay interest. Assuming that we can agree on a value of a life saved today what is the appropriate interest rate to apply when the lives saved by regulatory actions will occur not today, but next year, in the next decade, or even in the next generation? If the benefits extend into the indefinite future, it is a mathematical certainty that any positive interest rate that discounts the value of future lives will treat some of them

 

Regulatory Decision-Making and Economic Analysis    71 as having essentially no value at all. Can that be correct? Indeed, could it be viewed as correct with respect to any future benefit whether those benefits are life-​saving or simply other aspects of human well-​being? And yet, to treat benefits far in the future as if they were as valuable as benefits available to us today seems transparently inconsistent with our ordinary practices in our everyday lives. Over time, OMB has provided increasingly detailed instructions to agencies regarding how to undertake economic analyses of regulatory policies. Consider, for instance, the details specified in OMB Circular A-​4: As a default position [OMB states] that a real discount rate of 7 percent should be used as a base-​case for regulatory analysis. The 7 percent rate is an estimate of the average before-​tax rate of return to private capital in the U.S. economy. It is a broad measure that reflects the returns to real estate and small business capital as well as corporate capital, and it is the appropriate discount rate whenever the main effect of a regulation is to displace or alter the use of capital in the private sector.33

This particular discount rate is no doubt defensible on the basis of one analytical technique (the before-​tax average rate of return to private capital in the United States, during a given time period), and as a strategy for reducing uncertainty and limiting variation in the techniques used by regulatory agencies. Yet valid alternatives also exist (indeed, the same circular notes that “when regulation primarily and directly affects private consumption … a lower discount rate is appropriate”).34 The 7 percent discount rate nonetheless provides an example of how analytical presumptions driving the outcomes of economic analysis may fade from view once the calculation yields a result. And in many cases, the instructions to agencies in the OMB circular leave the agencies in a difficult position, making it far from clear how the agency should comply with such imperatives. These are hardly the only conundra surrounding the practice of cost–​benefit analysis. And, they fuel endless debates concerning whether cost–​benefit analysis is a manageable and useful tool, a technique of obfuscation that not only hides or papers over serious questions of value but submerges considerations of non-​economic normative commitments and social values beneath an esoteric methodology that can be a cover for ideological resistance to government regulation. These latter concerns put front and center two additional issues of institutional design.

4.4.3 Cost–​Benefit Analysis and Political Economy The first engages questions of political economy. Although there is reason to doubt the idea that regulatory policy is pervasively “captured” by industry,35 there is no question 33 

OMB Circular A-​4, p. 33 (2003). OMB Circular A-​4, p. 33 (2003). Of course, a vast array of regulatory rules affects both private capital and consumption. 35  Cuéllar (2014). 34 

 

72    Mariano-Florentino Cuéllar and Jerry L. Mashaw that regulatory policy is nonetheless constrained by factors such as legislative coalitions, organized interests, organizational learning challenges for agencies, and public reactions. These constraints exist in large measure because political actors understand the high stakes of regulation for their particular goals. Many observers might be wary of incorporating factors such as political resistance into the agency’s calculus of costs and benefits. Still, other things being equal, it is not obvious that the scope of the cost–​benefit analysis on a particular proposed regulation should exclude factors such as the impact of that rule on an agency’s ability to pursue some other legislative goal that would enhance social welfare, the consequences of the rule for the country’s international negotiations,36 or the agency’s ability to implement rules in a manner that will likely increase the agency’s ability to learn from experience.37 Yet in its current incarnations, cost–​benefit analysis and comparable techniques of economic analysis routinely exclude institutional factors reflecting the broader institutional consequences of regulatory policies.38 Other political economy concerns are also relevant to the institutional design of regulatory analysis. Suppose that certain requirements for regulatory analysis, as they operate in the real world, allow organized interests to delay implementation of regulatory policies that are likely to deliver great value to society. Even if one believed that such requirements were sound for an individual rule, they could be problematic in the aggregate if organized interests could more easily delay a large number of valuable rules. The key point here is that any analytic requirement needs to be scrutinized in terms of how different political actors with specific views about the regulatory process will use it. Political economy considerations might also arise in a distinct way if maximizing social welfare in a few domains through new regulatory rules provokes a broader political (e.g., legislative or popular) response that reduces bureaucratic autonomy, changes statutory rules, or otherwise interferes with regulation in different settings.

4.4.4 Cost–​Benefit Analysis and Value Formation The second and more intractable issue revolves around the relationship between economic analysis and value formation. In many advanced democracies the regulatory process depends as a descriptive matter not merely on the analysis of costs and benefits as derived by insulated bureaucracies, but on the definition of those costs and benefits. A core prescriptive issue for democracies is how to reconcile the rigorous assessment of regulatory policies with the role of pluralist or popular mechanisms for defining what society values. Perhaps it is not surprising that the governments of advanced democracies (including that of the United States) have failed to create policymaking institutions that insulate the economic analysis of regulatory policies from various kinds of interest-​group and public pressures.39 As a normative matter, it may be entirely defensible 36

 See Mass. v. EPA, 549 U.S. 497 (2007).   Contra Sunstein (2013). 38   Cf. Kornhauser (2000). 39  See Adler and Posner (2000). 37

 

Regulatory Decision-Making and Economic Analysis    73 for at least some of these reactions to be taken into account in policy decisions along with economic analyses of regulatory policy.40 If tempered by the social welfare concerns noted above, then perhaps public discussion, education, and experimentation may play a material (if nonetheless modest) role in helping society understand how to successfully inform the economic analysis of regulatory policies, whether the specific issue is whether a particular government function should be outsourced,41 how to allocate migration opportunities,42 or what to do about tobacco addiction.43 Moreover the link between regulatory policy and the advancement of social welfare becomes more complicated once we consider several points concerning the malleability, instability, or normative inconclusiveness of many human preferences or goals. Individual preferences obviously matter in a regulatory system that cares about social welfare. Hence, one might accept the value of greater executive capacity to compare the merits of distinct risk regulation priorities across strikingly different contexts without assuming that all scarcity constraints in regulatory policymaking can be managed by a cadre of experts ensconced deep in the Executive Office of the President.44 Even if we examine what should happen within particular regulatory domains such as pharmaceutical regulation or campaign finance, recognizing the value of technical competence is simply not the same as assuming that all regulatory questions can be resolved by such competence. Such an assumption is untenable as a matter of practical politics, but there is also a normative dimension that makes the public’s preferences (and not just hermetically sealed technical evaluations) potentially important. The interdependence between economic analysis and public attitudes is perhaps particularly salient in regulatory domains such as immigration, where only a vapid description can remove the value questions from the choice of how large to make a guest-​worker program and whether it should include a path to citizenship.45 As noted earlier, in the discussions of valuing lives and determining discount rates, these ineluctable value questions also loom in health and safety domains, and inhabit even the most abstract issues, such as social welfare measures.46 At a minimum, these descriptive and normative observations counsel against a regulatory system that meets someone’s definition of maximally promoting 40 

Cf. Richardson (2000); Cuéllar (2005). See Verkuil (2007). 42  See Tichenor (2002). 43  Croley (2007). 44  Cf. Heinzerling (1995), discussing some of the practical, political, and normative difficulties associated with Breyer’s proposal for greater centralized review of risk regulation policies across contexts. 45  Policymakers would no doubt benefit even from the more conventional economic analysis, focused purely on the consequences of a certain immigration policy for government revenue or per capita GDP. The point is that it is both descriptively useful for policymakers and normatively defensible to seek the means of reconciling that more conventional analysis with the dynamic of broader public responses to changes in immigration policy. That goal could be achieved either by expanding the scope of a more conventional economic analysis to include a broader range of costs and benefits, or by accepting the narrower analysis but limiting its relevance to the ultimate policy decision. 46   For an interesting discussion of the stakes and analytical territory associated with the so-​called “time slice” problem in welfare determinations, see Adler (2007). 41 

 

74    Mariano-Florentino Cuéllar and Jerry L. Mashaw social welfare while ignoring the views (rightly or wrongly) held by large chunks of the mass public. A second complication, however, is that those views are subject to change. At least two well-​established, decades-​long research programs—​prospect theory and its variants in the decision sciences, and the study of social cognition in psychology have long shown choices and even attitudes to be highly context dependent.47 Given that the bare requirements of procedural legitimacy necessary to produce regulations or regulatory statutes rarely squelch the enormous territory of regulating discretion, context-​dependence in some sense leaves thoughtful policymakers with genuine quandaries about what views or hedonic experiences to honor. In formulating regulatory priorities should OSHA regulators respond to the safety concerns of workers who have just seen a terrible, but low-​ probability workplace accident take place, or to the cost concerns of a union that has just delivered a long list of concessions to avoid having a Midwestern factory shuttered? As workers and other members of the electorate move through their lives, their views may also change—​sometimes creating highly time-​inconsistent preferences.48 Both lifespan changes as well as broader contextual effects may be explained by reference to the idea that people generate views (whether about policies or their own lives) from utility functions with multiple dimensions.49 But in many ways this observation merely restates the problem: without an entirely persuasive technical means of understanding all these dimensions, policymakers trying to satisfy public concerns must grapple with the fact that the concerns expressed are partly a function of the conditions experienced. A similar dilemma arises when it comes to making sense of the divergence between expected utility calculations and individuals’ risk perceptions. It is not surprising that people engage in “probability neglect.” The challenge comes in deciding what constitutes a mere example of a person’s mind misfiring, and what constitutes a deeper value (or at least preference) worth heeding. There is little doubt that probability errors among the public (but also among lawmakers) are real, and greatly complicate the work of the regulatory state. By the same token, it is not obvious that dread associated with nuclear risks is simply an error. Dying from a nuclear accident may be a more remote possibility than dying from a risk considered less dreadful. But, it is premature to treat that response as a mere glitch of mental accounting.50 People have complicated inner lives. Their perceptions of nuclear dangers may represent guesses about the suffering that would be involved in dying from a nuclear accident versus a car crash. Some views reflect partially unarticulated anxieties about the disposal of nuclear waste, or responses to the idea that the fortunes of nuclear power signal something about the political success of despised political coalitions.51 Despite some occasional easier cases, it often remains an unacknowledged challenge in regulatory law to judge whether public views should be 47 

See, e.g., Kelman, Rottenstreich, and Tversky (1996); Quattrone and Tversky (1988); Allen (1982). SeeVisser and Krosnick (1998). 49  See Druckman and McDermott (2007); Jones (1994). 50  Cf., e.g., Kahan et al. with Sunstein (2005). 51  Cf. Tanaka (2004). 48 

 

Regulatory Decision-Making and Economic Analysis    75 treated as mistaken simply because public responses diverge from what a more technically focused economic analysis might yield.52 Regardless of which conceptions of risk or value are prioritized at the outset of an economic analysis, it may be unsatisfying as a prescriptive matter for regulatory policy to claim “responsiveness” to public views without acknowledging its role in surreptitiously shaping those views.53 Changes in the regulatory agenda, for example, can trigger distinct priorities among individual members of the public.54 The Food and Drug Administration’s (FDA) tobacco regulation initiative, for instance, triggered spikes in public concern about the dangers of smoking.55 Automobile safety regulation seems to have radically shifted the value consumers place on vehicle safety. Regulatory programs can also bolster their own constituencies, as when Interstate Commerce Commission rate regulation of surface transportation spurred companies and unions to solve collective action problems and demand more favorable treatment.56 And whether intentionally or not, regulators’ decisions impact not only the public agenda but the context in which individuals decide what they seek from themselves, each other, and their government. Even as the FDA tobacco regulation initiative appeared to spur growing concern about the dangers of tobacco, its regulatory program was also designed to affect individual desires to use tobacco products.57 If democratic societies accept that individual preferences matter in regulatory policy, but that not all of them are stable, meaningful, or worthy of the same attention, they may legitimately seek to temper the focus on economic analysis of regulatory policy with techniques to galvanize some public awareness of—​and involvement in—​the process of managing regulatory policy. At a minimum, such a goal would seem to call for devoting attention to public discussion of regulatory policies among a variety of organized interests and perhaps even members of the lay public, elucidating choices about how to educate the public about those policies and the regulatory state in general, and experimentation capable of enriching what’s known about how to make regulatory policy succeed in a factually uncertain, politically fragmented environment.

4.4.5 Consideration or Constraint? How should cost–​benefit analysis factor into regulatory decision-​making? As was mentioned earlier certain statutes have been interpreted not to permit the use of cost–​benefit analysis as an element in regulatory decision-​making. This perhaps defines one end of 52 

For one effort to address this issue, see Adler (2006). Cf. Willis (2013). 54  See, e.g., Jones (2001). 55  See Kessler (2000). 56  See Rothenberg. 57  See Kessler (2000), discussing the underlying public education and behavior change goals of the FDA tobacco regulatory initiative; Croley (2007), 194, “The FDA’s efforts had a profound effect on public awareness of and sentiment towards cigarette advertising and distribution practices.” 53 

 

76    Mariano-Florentino Cuéllar and Jerry L. Mashaw the spectrum of design choices. At the other terminus lie requirements that cost–​benefit analysis be undertaken, and that regulatory action be withheld unless the cost–​benefit ratio demonstrates that benefits outweigh the cost. In the United States regulatory legislation is often silent or ambiguous concerning the role of cost–​benefit analysis, which means that judicial review is a crucial factor in the role that cost–​benefit analysis will play in particular regality regime. In Whitman v. American Trucking Associations, 531 U.S. 457 (2001), for instance, the Supreme Court required a clear textual commitment of authority to the EPA in order to allow consideration of costs in setting air standards. The relevant Clean Air Act section, though, only instructed the EPA to set standards which “are requisite to protect the public health” with “an adequate margin of safety.”58 Were it not for the lengthy briefing submitted on the issue, Justice Scalia wrote for the Court, “One would have thought it fairly clear that this text does not permit the EPA to consider costs in setting the standards.”59 The Court was influenced by its recognition that the Act explicitly called for weighing costs in other provisions; Section 7408(b)(l) requires the EPA to provide States with information “relating to the cost of installation and operation.” And §109(d)(2)(C)(iv) requires a Scientific Committee to “advise the Administrator of any adverse public health, welfare, social, economic, or energy effects which may result from various strategies for attainment and maintenance” of air standards. This language, the Court said, requires consideration of costs; requiring an “adequate margin of safety” does not.60 Similarly, in Tennessee Valley Authority v. Hill, 437 U.S. 153 (1978), the Court read the Endangered Species Act to prohibit weighing the financial costs of regulation against the potential loss of an endangered snail darter. That Act required all federal agencies “to insure that actions authorized, funded, or carried out by them do not jeopardize the continued existence” of an endangered species or “result in the destruction or modification of habitat of such species.”61 The Court interpreted the plain language of the Act to regard the value of endangered species as “incalculable,” thus obviating attempted calculations of costs and benefits.62 Second, statutes may be interpreted to neither require nor prohibit cost–​benefit analysis. Courts will reach this conclusion if the Act’s language and structure is found to be silent on the matter. In Entergy v. Riverkeeper, 556 U.S. 208 (2009), for example, the Court held that a Clean Water Act provision mandating use of the “best technology available for minimizing adverse environmental impact” neither prohibited nor required the use of cost–​benefit analysis. This statutory language was 58 

42 USC §7409(b)(1). 531 U.S. at 464. 60  531 U.S. at 471. 61  16 USC §1536 (1976 ed.). 62  The Court also forbade cost–​benefit analysis in American Textile Manufacturers v. Donovan, 452 U.S. 490 (1981). There, the Occupational Health and Safety Act required agencies, in regulating toxic materials, to “set the standard which most adequately assures, to the extent feasible, that no employee will suffer material impairment of health…. 29 USC 3,656(b)(5).” This formulation, the Court decided, focusing on the phrase “to the extent feasible,” reflects a weighing of costs and benefits that was already performed by Congress: worker health is to be prioritized, even if it is costly. 452 U.S. at 509. 59 

 

Regulatory Decision-Making and Economic Analysis    77 read as “nothing more than a refusal to tie an agency’s hands as to whether cost–​benefit analysis should be used, and if so to what degree” (556 U.S. at 222). Finally a statute may be read to require the agency’s proper consideration of costs and benefits. In Corrosion Proof Fittings v.  EPA, 947 F.2d 1201 (5th Cir. 1991), EPA’s ban on asbestos was struck down because it did not pass a cost–​benefit test. The Toxic Substances Control Act (TSCA) permits only the “least burdensome” regulation of chemicals, and requires the EPA to “consider the environmental, economic, and social impact of any action” taken.63 The court faulted EPA for failing to measure the costs and benefits of less burdensome alternatives, for discounting costs but not benefits to present real terms, and for wielding unquantifiable benefits as a “trump card” to justify all costs.64 By requiring the government to show that every less-​restrictive alternative would pose an inferior costs-​to-​benefits outcome, despite the logistical hurdles to collecting such information, the ruling effectively disabled EPA from regulating asbestos under the TSCA.65 Although economic analysis of regulatory policies understandably raises contested questions about statutory interpretation and valuation of costs and benefits in the environmental context, disagreements about the scope of agency responsibilities also arise in other contexts. One recent example involves agency responsibilities involving financial regulation. In Business Roundtable v. SEC, 647 F.3d 1144 (D.C. Cir. 2011), the D.C. Circuit invalidated an SEC rule because it found the government’s cost–​benefit analysis to be inadequate. According to the court, cost–​benefit analysis was required by Section 3(f) of the Exchange Act and Section 2(c) of the Investment Company Act of 1940, even if the statute stopped short of using that precise language.66 Those sections oblige the SEC, when it is engaged in rulemaking, to “consider, in addition to the protection of investors, whether the action will promote efficiency, competition, and capital formation.” Meanwhile, in Chamber of Commerce v. SEC, 412 F.3d 133, 143 (D.C. Cir. 2005), the court interpreted these sections to impose upon the SEC a “statutory obligation to determine as best it can the economic implications of the rule” and struck down a rule regulating mutual fund transactions because of the SEC’s failure to consider costs and alternatives. The SEC had concluded that the rule at issue, which required public companies to provide shareholders with information about shareholder-​nominated candidates for the board of directors, provided “potential benefits of improved board and company performance and shareholder value” sufficient to “justify [its] potential costs.” But according to the court, “the Commission inconsistently and opportunistically framed the costs and benefits of the rule; failed adequately to quantify the certain costs or to explain why

63 

15 USCA §2605(a) and §2601(c). 947 F.2d at 1218. 65  See Morgenstern (1997), noting that EPA never regulated anything but PCB’s under section 6 after the reversal of the asbestos rule. In 2010, a district court cabined the Corrosive Proof ruling to the TSCA context, and upheld an FDA ban on the sale of baby turtles under less rigorous scrutiny. Independent Turtle Farmers of America, Inc. v. U.S., 703 F.Supp. 2d 604 (W.D. La. 2010). 66  15 USC §§78c(f), 80a–​2(c). 64 

 

78    Mariano-Florentino Cuéllar and Jerry L. Mashaw those costs could not be quantified; neglected to support its predictive judgments; contradicted itself; and failed to respond to substantial problems raised by commenters.”67 As with the Corrosion Proof ruling, critics fear Business Roundtable has imposed “an onerous, and possibly insurmountable procedural burden,” and that “the decision threatens to paralyze rulemaking by the SEC and other administrative agencies.”68 Cases such as these emphasize the extent to which courts and lawyers have become pivotal actors in the economic analysis of regulation, at least in the United States. But, the full story of the practice of the use of cost–​benefit analysis and regulatory decision-​making can hardly be captured by the explicit textual demands or prohibitions found in statutes or executive orders. Agencies like the EPA who have been held to operate under statutes forbidding the use of cost–​benefit analysis in decision-​making concerning particular rules or regulations may nevertheless use some form of this technique in priority setting. The setting of regulatory priorities is hardly a pure exercise in policy analysis, but agencies would be foolish if they did not consider how they might use their limited resources to produce the greatest net benefits. Rules producing significant public health or worker safety benefits, particularly if they can be adopted in a form that does not impose huge costs and therefore stimulate fierce opposition, should certainly be high on an agency’s regulatory agenda. Moreover, where agencies are told by statute or executive order merely to consider cost and benefits, and to do so only to the extent permitted by their statutes and in connection with other values, cost–​benefit considerations might nevertheless be virtually determinative of agency action. Whether it is or is not depends importantly on the accountability structure within which cost–​benefit considerations are reviewed by either legal or political overseers. Concern with the legal and political environment within which cost–​benefit analysis is practiced leads to our final institutional design consideration.

4.4.6 How and to Whom Are Agencies Accountable for their Cost–​Benefit Analysis? In the United States systematic oversight of agency regulatory action, including cost–​ benefit analyses, occurs in two fora: the federal courts and the OIRA at OMB. (Obviously these public documents are also available to and may be critiqued by Congress, the affected public and academic commentators.) Neither the courts nor OMB have formal authority to prevent agency action because of disagreement with agencies’ analysis of the costs and benefits of their proposals. The substance of agency regulatory or cost–​benefit analyses are not independently reviewable by the federal courts. And while agencies are required to submit their cost–​benefit analyses concerning major federal actions to OIRA, that office is empowered neither by statute nor by executive order to override agency regulatory judgment based on its analysis of the costs and benefits of the agency’s proposed regulations. But, once again, this is far from the whole story. 67 

68 

647 F.3d at 1148. Fisch (2012); Kraus and Raso (2013).

 

Regulatory Decision-Making and Economic Analysis    79 Executive agencies, unless explicitly exempted by statute (which is rare) must submit their budget proposals, any proposals for amendments to their governing legislation, and even their congressional testimony for review by OMB. Unless OMB agrees, agency proposals for budgetary expenditures and for amendatory legislation will not become a part of the “President’s Program” that is presented to Congress. To be sure, agencies can make end runs around these requirements, and are sometimes encouraged by Congress to do so, but this clearance process gives OMB significant leverage in dealing with regulatory agencies. Hence, to go forward with regulatory action in the face of OMB objections to an agency’s cost–​benefit analysis contains serious political risks. If OMB suggests that an agency rethink its analysis, it is highly likely to do so. Multiple rounds of review at the OMB will at the very least delay regulatory action, and it may cause agencies to abandon their projects.69 OMB review also provides opportunities for political intervention. Regulated parties who have been unsuccessful in convincing agencies of the error of their ways may get a hearing before OMB that is more sympathetic to their concerns. Regulatory “hot potatoes” may be referred to the White House where the president, or those acting for him, have an opportunity to persuade agency political appointees that their plans should be modified or abandoned. In short, the apparently modest requirement that agencies do cost–​benefit analyses and submit them for review to specialists in OMB tends to centralize important regulatory decision-​making by providing an important opportunity for both technical and political interventions. A similar story can be told about judicial review of agency cost–​benefit analyses. While these analyses are generally not subject to independent judicial review, they form a part of the record upon which the judiciary will determine whether agency action is arbitrary or unreasonable. We cannot here provide a complete guide to “reasonableness review” in the American administrative system, but two critical aspects of that review bear notice. First, to determine whether an agency has behaved reasonably courts look to see whether the agency has considered all of the required aspects of a particular problem in making its final decision. Where a statute provides, as is the case, for example, with the SEC, that the Commission must consider the effects of its proposed rules on the economy and on competition, parties seeking judicial review of the agency’s determination can and do argue that the agency has failed adequately to consider those elements of its decision. Here formal consideration is not enough. Courts must at the very least look to see whether the agency has been serious in its consideration. This so-​called hard look at an agency’s analysis leads inevitably to a consideration of the adequacy of the agency’s methodology, fact-​finding, and analysis. As we have seen, failure to satisfy the court that it has done a good job on these matters can lead to a finding of arbitrariness and remand to the agency for reconsideration. The same situation obtains to a somewhat lesser degree with respect to agencies whose cost–​benefit analyses are provided only because of the demands of executive orders. Once again, the agency analysis will be a part of the record for review and can be 69 

Olson (1984); Croley (2003).

 

80    Mariano-Florentino Cuéllar and Jerry L. Mashaw challenged for arbitrariness. And, although the agency may by statute not be required to consider costs and benefits at all, a court that believes the agency’s analysis is competent cannot know whether, on a competent analysis, the agency would have decided to adopt the regulation under review. Once again a remand to the agency may be in order. And, while a remand does not require that an agency change its mind, only that it provide a more persuasive rationale for its action, change of mind, or perhaps abandonment of the regulatory initiative, may nevertheless result. In short, while neither OMB nor judicial review provides a veto over regulatory agency action based on objections to the agency’s cost–​benefit analysis, the practical effects of accountability to legal and political controllers are significant.

4.5 Conclusion Like government itself, regulation has the potential to burden or benefit society. Over the years, lawmakers have enacted a series of statutes reflecting long-​standing social commitments to the idea that regulatory initiatives can enhance social welfare. In countries like the United States, those statutes necessarily serve as a point of departure for economic analyses of the costs and benefits of regulatory policies. Statutes and the regulatory policies that implement them also have the potential for dynamic, path-​dependent effects on how people think about what they expect government to provide, and perhaps even how the public understands its own well-​being. The economic analysis of regulation has the potential to inform society’s understanding of how regulation can affect social welfare. It implicates the use of techniques such as the analysis of wage premiums in risky occupations to estimate the cost of a statistical life, consideration of the costs of new technologies for complying with regulatory rules, and explicit discounting of future economic benefits to assess their present value. When the relevant assumptions are carefully specified, such analyses can help provide a clearer picture (for decision-​makers and the public) about the likely consequences of regulation, and perhaps even about the subtle ways in which regulatory initiatives can misfire. Scholars and practitioners continue to refine the techniques involved, including the strategies used to monetize costs and benefits. What the economic analysis of regulation cannot do, by itself, is resolve the question of what value society should assign to every particular attribute of life that could be affected by regulatory policy: the distribution of material resources, for example, the aesthetic consequences of building a dam or a highway, or the value of trust in public institutions. That enterprise depends, as a descriptive matter, primarily on public attitudes and deliberation, politics, and the institutions that structure how a society makes collective decisions. Indeed, when agencies undertake economic analyses of costs and benefits to inform those decisions, they confront a certain tension that arises at the intersection of social welfare analysis and the politics that arise under any given institutional arrangement: there is no clear way of thinking about social welfare in

 

Regulatory Decision-Making and Economic Analysis    81 the regulatory process without making some political choices about what kind of welfare we care about and how it should be measured. At the same time, there is no sensible way to assess those political choices without in turn applying some standard, implicit or otherwise, to measure social welfare. For reasons of both political reality and prescriptive ideals, many of the key decision-​ makers in society naturally welcome a straightforward, unambiguous technique to assess the trade-​offs inherent in regulatory policy. The economic analysis of regulation aspires to fulfill that need. Still, an intellectually honest depiction of that enterprise must contend with the interdependence between how a society measures social welfare and how it defines social values.

References Ackerman, F. and Heinzerling, L. (2004) Priceless: On Knowing the Price of Everything and the Value of Nothing. New York: The New Press. Adler, M. (2006) “Welfare Polls: A Synthesis,” New York University Law Review, 81: 1875–​970. Adler, M. D. (2007) “Well-​Being, Inequality and Time: The Time-​Slice Problem and its Policy Implications,” University of Pennsylvania Law School, Public Law Research Paper No. 07-​30. Adler, M. D. and Posner, E. A. (2000) “Implementing Cost–​Benefit Analysis When Preferences Are Distorted,” Journal of Legal Studies, 29: 1105–​47. Adler, M. D. and Posner, E. A. (2006) New Foundations of Cost–​Benefit Analysis. Cambridge, MA: Harvard University Press. Allen, C. T. (1982) “Self-​Perception Based Strategies for Stimulating Energy Conservation,” Journal of Consumer Research, 8: 381–​90. Copeland, C. W. (2013) “Economic Analysis and Independent Regulatory Agencies,” http://​ acus.gov/​sites/​default/​files/​documents/​Copeland%20CBA%20Report%203-​29-​13.pdf. Croley, S. P. (2003) “White House Review of Agency Rulemaking: An Empirical Investigation,” University of Chicago Law Review, 70: 821–​85. Croley, S. P. (2007) Regulation and the Public Interests:  The Possibility of Good Regulatory Government. Princeton, NJ: Princeton University Press. Cuéllar, M.-​ F. (2005) “Rethinking Regulatory Democracy,” Administrative Law Review, 57: 411–​99. Cuéllar, M.-​F. (2014) “Modeling Partial Autonomy in Public Health Policymaking,” Theoretical Inquiries in Law, 15: 471–​505. Druckman, J. N. and McDermott, R. (2007) “Emotion and the Framing of Risky Choice,” Working Paper: Department of Political Science, Northwestern University. Executive Office of the President (2011) “Executive Order 13563—​Improving Regulation and Regulatory Review” (January 18). Fisch, J. (2012) “Business Roundtable and the Future of SEC Rulemaking,” Harvard Law School Forum on Corporate Governance and Financial Regulation (November 15), http://​blogs.law. harvard.edu/​corpgov/​2012/​11/​15/​business-​roundtable-​and-​the-​future-​of-​sec-​rulemaking/​ #more-​36274. Heinzerling, L. (1995) “Political Science,” University of Chicago Law Review, 62: Article 10. Heinzerling, L. (1998) “Regulatory Costs of Mythic Proportions,” The Yale Law Journal, 107: 1981–​2070.

 

82    Mariano-Florentino Cuéllar and Jerry L. Mashaw Jeong, H. (2013) “The Role of Administrative Law in Economic Development and Democracy in Korea,” http://​link.springer.com/​chapter/​10.1007/​978-​3-​642-​31689-​0_​3/​fulltext.html. Jones, B. D. (1994) “A Change of Mind or Change of Focus? A Theory of Choice Reversals in Politics,” Journal of Public Administration Research and Theory, 4: 141–​78. Jones, B. D. (2001) Politics and the Architecture of Choice. Chicago, IL:  University of Chicago Press. Kahan, D. M., Slovic, P., Braman, D., and Gastil, J. (2006) “Fear of Democracy: A Cultural Evaluation of Sunstein on Risk,” The Harvard Law Review, 119: 1071–​109. Kelman, M., Rottenstreich, Y., and Tversky, A. (1996) “Context-​Dependence in Legal Decision Making, Journal of Legal Studies, 25: 287–​318. Kerwin, C. M. (2003) Rulemaking:  How Government Agencies Write Law and Make Policy. Washington, DC: CQ Press. Kessler, D. (2000) A Question of Intent. New York: Publicaffairs. Kornhauser, L. A. (2000) “On Justifying Cost–​Benefit Analysis,” Journal of Legal Studies, 29: 1037–​57. Kraus B. R. and Raso, C. (2012) “Rational Boundaries for SEC Cost–​Benefit Analysis,” Yale Journal on Regulation, 30: 289–​342. Mishan, E. J. (1994) Cost–​Benefit Analysis (4th ed.). London: Allen & Unwin. McCarthy, J. E. (2012) “EPA’s Boiler MACT:  Controlling Emissions of Hazardous Air Pollutants,” CRS Report for Congress, No. 41459 (March 9). McGarity, T. O. (1991) Re-​inventing Rationality: The Role of Regulatory Analysis in the Federal Bureaucracy. Cambridge: Cambridge University Press. McGarity, T. O. (2013) Freedom to Harm: The Lasting Legacy of the Laissez Faire Revival. New Haven, CT: Yale University Press. Morgenstern, R. D. ed. (1997) Economic Analyses at EPA:  Assessing Regulatory Impact. New York: Resources for the Future. OECD (1999) “Regulatory Reform in Mexico,” http://​www.oecd-​ilibrary.org/​governance/​ oecd-​reviews-​of-​regulatory-​reform-​regulatory-​reform-​in-​mexico-​1999_​9789164180315-​en. OECD (2002) “United Kingdom:  Challenges at the Cutting Edge,” http://​www.oecd-​ ilibrary.org/​ governance/​o ecd-​reviews-​of-​regulatory-​ reform-​ united-​ k ingdom-​ 2002_​ 9789264199255-​en. OECD (2010) “Australia 2010: Towards a Seamless National Economy,” http://​www.finance. gov.au/​sites/​default/​files/​australia_​report_​final.pdf. Olson, E. D. (1984) “The Quiet Shift of Power,” Virginia Journal of Natural Resources Law, 4: 1–​80. Quattrone, G. A. and Tversky, A. (1988) “Contrasting Rational and Psychological Analyses of Political Choice,” American Political Science Review, 82: 719–​36. Rabin, R. L. (1986) “Federal Regulation in Historical Perspective,” Stanford Law Review, 38: 1189–​326. Richardson, H. S. (2000) “The Stupidity of the Cost–​Benefit Standard,” Journal of Legal Studies, 29: 971–​1003. Skrzycki, C. (2003) The Regulators: Anonymous Power Brokers in American Politics. Lanham, MD: Rowman & Littlefield. Sunstein, C. R. (2000) “Cognition and Cost–​Benefit Analysis,” Journal of Legal Studies, 29: 1059–​103. Sunstein, C. R. (2005) The Laws of Fear:  Beyond the Precautionary Principle. Cambridge: Cambridge University Press.

 

Regulatory Decision-Making and Economic Analysis    83 Sunstein, C. R. (2013) Simpler: The Future of Government. New York: Simon & Schuster. Tanaka, Y. (2004) “Major Psychological Factors Determining Public Acceptance to the Siting of Nuclear Facilities,” Journal of Applied Social Psychology, 34: 1147–​65. Taoiseach(2004)“AComparativeAnalysisofRegulatoryImpactAssessmentinTenEUCountries,” http://​www.taoiseach.gov.ie/​eng/​Publications/​Publications_​Archive/​Publications_​2011/​ A_​Comparative_​Analysis_​of_​Regulatory_​Impact_​Assessment_​in_​Ten_​EU_​Countries. pdf. Tichenor, D. J. (2002) Dividing Lines: The Politics of Immigration Control in America. Princeton, NJ: Princeton University Press. Treasury Bd. of Canada Secretariat (2007) “Canadian Cost–​Benefit Analysis Guide 1,” http://​ publications.gc.ca/​site/​eng/​456648/​publication.html. Verkuil, P. R. (2007) Outsourcing Sovereignty:  Why Privatization of Government Functions Threatens Democracy and What We Can Do about It. Cambridge: Cambridge University Press. Visser, P. S. and Krosnick, J. A. (1998) “Development of Attitude Strength over the Life Cycle: Surge and Decline,” Journal of Personality and Social Psychology, 75: 1389–​410. Wiener, J. B. and Alemanno, A. (2010) “Comparing Regulatory Oversight Bodies across the Atlantic: The Office of Information and Regulatory Affairs in the US and the Impact Assessment Board in the EU,” in S. Rose-​Ackerman and P. L. Lindseth (eds) Comparative Administrative Law. Cheltenham: Edward Elgar, 309–​35. Willis, L. E. (2013) “When Nudges Fail: Slippery Defaults,” University of Chicago Law Review, 80: 1155–​229.

 

Chapter 5

Ec onomics of Fe de ra l i sm ROBERT P. INMAN AND Daniel L. Rubinfeld

5.1 Introduction The economic theory of federalism has become the starting point for most scholarly and policy discussions as to how best to organize the federal state. The goal is to correct market failures through government action, and its guiding principle is to assign policy responsibility to the smallest level of government that can accomplish the task. The principle has been formalized by Oates (1972) as the decentralization theorem and with respect to the European Community as the principle of subsidiarity. The economics of federalism views the primary task of government as solving the failures of private markets to satisfy the demands of citizens for goods and services. These failures arise whenever cooperative action is needed to ensure the provision of goods or services at the lowest cost. Examples include all goods where sharing a fixed resource is efficient, as with the theory of public goods (Samuelson, 1954), or where one person’s actions generates external benefits or imposes costs on others. What is needed in both cases is a means of discovering citizens’ willingness to pay for each publicly provided good and service and a means to collect sufficient payments to cover production costs. The economic task for federalism is to provide a means to solve these market failures. For goods and services for which congestion becomes evident within relatively small populations, and where the spatial reach of any externalities is modest, economic federalism recommends using small local governments. Likely local services are K-​12 education, police and fire protection, trash pickup and street sanitation, libraries, parks and recreation, and local roadways. However, for those goods and services shared by large populations and for those instances in which externalities are substantial, economic federalism emphasizes production and distribution by a more centralized governmental unit. With a nation-​state, candidate services might be higher education and research, national defense, courts and jails, protecting property rights and market competition, and providing air and water quality, regional and national highways, and airports and airways.

 

Economics of Federalism    85 From the perspective of economic efficiency, determining which level of government is best suited to manage governmental functions requires balancing the benefits and costs of decentralized and centralized political structures.1 We begin by defining governments as “decentralized” (or local) or “centralized” (or national) according to the combination of three constitutional decisions. First, the partition decision divides the single national citizenry into states and/​or localities. Second, a national government may be established with elected representation from each of the state or local governments as well. If so, the representation decision defines how each state or local government will be so represented. The constitution may allow for only a single, nationally elected representative, or president, or it may allow each local government or state its own elected representative to a national assembly, or legislature.2 Third, and of particular interest to US Courts, the assignment decision allocates the final political responsibility for policy choices to either the president or legislature of the central government or to the individual local governments alone. Together, these three constitutional decisions define the essential federalist structure of the government. We offer three alternatives for consideration:  economic federalism, cooperative federalism, and democratic federalism. In this chapter we review the efficiency arguments for each. Not surprisingly, no one form of governance is clearly dominant in its ability to efficiently allocate national resources. Further, there are other values that must be weighed in the balance too: political participation and democracy, economic fairness, and the protection of personal rights and liberties. Which federal constitution is finally chosen will reflect a balance of economic concerns and these possibly competing values.

5.2  Economic Federalism 5.2.1 The Pros and Cons of Institutional Decentralization In a federal system there are two important dimensions to economic efficiency. Inter-​ jurisdictional efficiency involves the appropriate allocation of individuals and other resources among different jurisdictions. Inter-​jurisdictional efficiency is achieved when the public activities of these interacting governments satisfy the collective demands of individuals at minimum cost. The cost of public provision includes the production costs of producing public services plus the transactions costs of selecting legislators, the decision-​making costs of setting policies, and the monitoring costs of ensuring that 1  This chapter builds on our prior writings on the political economy of federalism; see Inman and Rubinfeld (1996, 1997a, 1997b, 1997c, 2000). 2  Adding another legislative chamber or an executive veto has the potential to improve the efficiency of the federal system. For an analysis of the effects of bicameral legislatures on government performance, see Levmore (1992). For an analysis of the effects of a president with veto powers on government performance, see Fitts and Inman (1992).

 

86    Robert P. Inman and Daniel L. Rubinfeld policies are put in place.3 Intra-​jurisdictional efficiency requires a choice of public activities which satisfies the collective demand within a given jurisdiction, measured as the willingness to pay for those activities for all individuals within that jurisdiction. In federal systems, inter-​jurisdictional inefficiencies arise when the decisions and actions of individuals within one jurisdiction have effects on individuals located in other jurisdictions that are not accurately reflected in the marketplace. Intra-​jurisdictional inefficiencies arise when the political process generates outcomes within a jurisdiction which do not maximize the well-​being of all residents. Ultimately, the choice of the appropriate jurisdiction to be responsible for a government activity involves a trade-​ off—​the larger the jurisdiction the less likely that there will be inter-​jurisdictional inefficiencies, but the more likely that the political process will lead to intra-​jurisdictional inefficiencies. The analysis of the economics of federalism begins with an evaluation of the arguments for a decentralized system. The linchpin is the “Tiebout model” first proposed by Charles Tiebout (1956) as a description of fiscal competition between many independent local governments. The important constitutional decision that defines the Tiebout economy is the partition decision that divides the national population into many political jurisdictions in such a way that each jurisdiction is a low-​cost provider of public services within its boundaries. The Tiebout model is formally analogous to the purely competitive market model with complete information. The assumptions needed to ensure that governmental competition is fully analogous to efficient private market competition are strong and may not hold in practice. In the section that follows we describe how the decentralized system of economic federalism is likely to work.4 The decentralized structure of strict economic federalism may be economically inefficient, however, for reasons we note below. If so, we must look to additional institutions—​intergovernmental bargaining and centralized government decision-​making—​to see if they might yield improved efficiency performance for the public sector.

5.2.2 The Tiebout Economy From the Tiebout perspective, the world of decentralized, competitive governments will allow for a marketplace in which individuals move among local jurisdictions to select the public good, public service combination that they most desire. Just as a private market place creates an incentive for suppliers to produce their goods at the lowest cost and to provide the goods most desired by consumers, so too will the pressure of household shopping, through exit and relocation, force competitive governments to be efficient in the provision of publicly supplied goods and services.

3 

4 

Breton and Scott (1978). For a more complete discussion of the Tiebout model, see Rubinfeld (1987).

 

Economics of Federalism    87 According to this view, public activities should be decentralized, except when there are significant economies of scale or economic spillovers from one jurisdiction to its neighbors. When there are spillovers, decentralized governments are likely to ignore economic costs or benefits which will lead them to over-​or underprovide the activity. In both cases it is natural, then, to look for a more centralized solution. Here economic federalism will turn to a strong, nationally elected president. The Tiebout framework assumes that each household and business is highly mobile and resides or does business within the jurisdiction that offers the tax-​expenditure-​ regulation package that it most prefers. As typically specified, the competitive Tiebout public economy is defined by five conditions: 1. Publicly provided goods, services, and regulatory activities are provided at minimum average cost and the scale of population required to reach this minimum is a relatively small share of the national population. When this assumption holds, many small (i.e., “local”) governments can provide the public services. If so, these governments can compete for residents and firms by offering preferred packages of public services at low tax rates. A failure to match the performance of its neighbors will mean that the inefficient government will lose residents and firms. Goods which can be allocated by the Tiebout competitive process are congested goods, where more users reduce, proportionally, the benefits enjoyed by current users, and congestion occurs at small populations. Existing evidence suggests that this constraint can be met for many government services.5 Examples include elementary and secondary education, local roadways, curative health care, police and fire protection, recreation, and cultural activities. 2. There is a perfectly elastic supply of these local political jurisdictions, each capable of replicating all attractive economic features of its competitors. This ensures that there are a sufficient number of alternative locations for households to choose among. Competition is provided by politicians (mayors, members of city councils, etc.) who run against, and sometimes defeat, inefficient incumbents. Alternatively, public sector entrepreneurs may be enterprising real estate developers who build new jurisdictions analogous to “new entrants” in a private market—​to draw away dissatisfied residents from inefficient governments.6 3. Mobility of households and businesses among jurisdictions is (almost) costless. 4. Households and businesses are fully informed about the fiscal and regulatory policies of each jurisdiction. Assumptions 3 and 4 ensure that when possible inefficiencies arise, households and businesses can move to otherwise similar jurisdictions without those inefficiencies. Finally,

5 

6 

See, for example, Ladd and Yinger (1989: 83, 85). See Epple and Zelenitz (1981).

 

88    Robert P. Inman and Daniel L. Rubinfeld 5. There are no inter-​jurisdictional externalities or spillovers. This assumption ensures that all public regulatory activities can be provided within these efficient jurisdictions—​larger governments or intergovernmental cooperation are not required. In the end, a political economy satisfying assumptions 1 to 5, and organized as a fully decentralized network of competing jurisdictions, will maximize economic efficiency. As in any market economy, citizens and businesses can consume their preferred levels of the public goods and regulatory activities with a minimum expenditure of production and transactions costs. And as stressed by Justice Brandeis with his famous plea for “states as laboratories,” “yardstick competition” between local governments for residents and firms will encourage policy and productive innovations.7

5.2.3 The Performance of the Tiebout Economy While a useful starting point, the Tiebout model does not provide a complete theory of federalism. The five assumptions may arguably apply to local goods and services, where communities provide a wide range of alternative bundles. However, the assumptions are less likely to hold for more centralized activities, where choices are more limited and moving is more costly. If assumptions 1 through 5 do not hold, we must move beyond complete decentralization as the specification for a federal constitution. The model of perfectly competitive private markets has long been an appealing paradigm for economists, in part because some real-​world markets do actually seem to operate in that manner, and in part because if they do, the efficiency consequences seem clear. However, neither of these reasons applies without qualification to the Tiebout framework for competitive governments. Assumption 1 is an assumption about the technology of public goods and services. While there are many services where relatively small populations are sufficient for the efficient provision of the service, there are other important government services, often called “pure” public goods, where the required technological assumption does not hold. Examples include national defense, foreign affairs, telecommunication networks and the distribution of television and radio, ports and airports, interstate highways, public health, scientific research, free trade and competition policy, and ensuring full employment. Because of economies of scale, these services can only be efficiently provided to large populations. If so, there can only be a few efficient providers, perhaps only one, and Tiebout competition cannot guarantee the efficient allocation of the service. Consider assumption 2—​a perfectly elastic supply of jurisdictions. The model of perfectly competitive markets yields an efficient allocation of resources in part because 7  Research by Besley and Case (1995) suggests that residents do use the economic performance of neighboring states as a “yardstick” against which to measure the performance of their own local or state government.

 

Economics of Federalism    89 firms are free to make the efficient choices of inputs and outputs, and in part because entry into and exit from an industry are free. In the Tiebout model, competitive governments fill the role of firms. But governments may be constrained by internal politics when making their input and output choices, and free entry of new governments, say because of land scarcity, is by no means assured in actual public goods economies. (See generally, Epple and Zelenitz (1981), arguing that limited entry permits competitive governments to maximize public “monopoly profits” from the provision of public goods and public regulations where the level and allocation of those profits will be determined politically within the jurisdiction.) If assumption 2 does not hold, there are four potentially adverse consequences for economic efficiency. First, with a limited supply of jurisdictions, each government is likely to contain citizens with different views of what constitutes the government’s best policy. If so, local politics becomes important. If local decisions are made by majority rule, the outcome will not necessarily be economically efficient (see Bergstrom et al., 1988). Second, when the economy can no longer replicate attractive local jurisdictions, there will be economic rents. The presence of these publicly created rents associated with the desirability of some locations over others can in turn create a tension between the fiscal entrepreneurs who wish to develop new communities and the owners of the desirable locations. As a result, there is no guarantee that the market for land and for other scarce resources will operate efficiently. Third, when private sector job opportunities are not identical in all locations, productive private sector employment may be sacrificed as labor is attracted to less efficient locations for fiscal reasons. The consequence is a less efficient private economy.8 Fourth, once the supply of jurisdictions is fixed, the analogy of the decentralized model of federalism to anonymous free-​market competition is no longer valid. Each local government may be aware of the potential effect of its action on other jurisdictions. It will then act strategically in making its decisions on the basis of its best guess as to how other jurisdictions will respond. In this framework the actions of each local jurisdiction, even if small, can have a substantial effect on the welfare of the rest of the decentralized world. In essence, a self-​interested decision by those in power in one jurisdiction can result in a series of distortions that arise throughout the effected network of other jurisdictions. For example, a decision by one jurisdiction to offer a tax break to attract new business could lead to competition in which many jurisdictions offer a variety of tax or business benefits. While it is difficult to predict the exact consequence of the rather complex behavioral patterns that can arise when these strategic interactions are taken into account, inefficiency seems likely. States may, for example, underprovide welfare services, or more generally underprovide public services in which they have particular production advantages and therefore should provide in abundance.9 This could occur because a state 8  See Buchanan and Goetz (1972) for the initial presentation of this argument and the summaries in Boadway and Flatters (1982) and Wildasin (1986). 9  See, for example, Gramlich and Laren (1984) on welfare competition among states.

 

90    Robert P. Inman and Daniel L. Rubinfeld realizes that by providing a high level of welfare benefits it is likely to attract new low-​ income residents from nearby states, thus adding substantially to its cost of providing such benefits. States might also provide substantial tax benefits to attract business, or at least to avoid losing business to other competing states, and in the process undertax business capital and overtax the residential base.10 An equivalent phenomenon can occur with respect to regulatory policies. States might relax their environmental controls to encourage business migration, or simply to forestall the loss of business because other states had or were about to relax their environmental regulations. This argument was developed by Cumberland (1981) and more generally, by Oates and Schwab (1986). Note that the competitive process which leads to a uniformity of regulation among states is not consistent with the advantages of the diversity associated with the Tiebout model, nor is it likely to be efficient. The result could be a “race to the bottom” resulting in an equilibrium in which environmental regulations are “undersupplied.”11 Assumption 3, that citizens and businesses are costlessly mobile between political jurisdictions, is also problematic. The assumption applies best to suburban jurisdictions among which individuals and firms have a good deal of mobility. When the number of cities and towns within a metropolitan area is substantial, families and firms have significant choices with respect to public goods and public regulations. Empirical evidence supporting the Tiebout hypothesis has focused entirely on this local public sector. With respect to household relocation, Gramlich and Rubinfeld (1982) analyzed survey data for Michigan, concluding that their results were “reasonably consistent with a Tiebout interpretation: in large metropolitan areas there is quite extensive grouping; in smaller areas, there is some grouping.”12 Oates (1972) studied the metropolitan New Jersey region, using tests of the capitalization of public goods differentials into property values, and concluded that public goods preferences could explain migration among neighboring cities and towns.13 Activities of local governments in metropolitan areas—​save those with significant spillovers—​are likely to meet the competitive test (see Courant and Rubinfeld, 1981). Yet when economic activity occurs at the level of large governments, assumption 3 is unlikely to hold. Studies of migration show that most household moves are within metropolitan areas, not between states. Those who do move between states are typically job motivated, and are not primarily seeking a more attractive public service/​public regulation environment (Easterbrook, 1983).14 If assumption 3 does not apply, then the model cannot guarantee overall (i.e., market plus public sector) efficiency. 10 

See, for example, Break (1967: 23–​4); Oates (1972: 142–​3); Inman and Rubinfeld (1996). For a more general argument that the regulatory actions of individual states may not result in an optimal federal system, given the incentives that determine political choice, see Rose-​Ackerman (1981); see also Revesz (1992). 12  pp. 554–​5. See Banzhaf and Walsh (2008). 13  For evidence that businesses are also sensitive to local taxes and services when relocating within metropolitan areas, see Bartik (1988) and McConnell and Schwab (1990). 14  See also the detailed evidence on interstate and interregional location of households reviewed in Greenwood (1986). 11 

 

Economics of Federalism    91 Assumption 4, that households and businesses are fully informed, presents another problem for the Tiebout argument. Households and firms must have reasonably accurate information about the tax, spending, and regulatory policies of the communities in which they will reside. Yet, it is often difficult for individuals and firms to ascertain the implications of regulations within their own jurisdiction, let alone make a reasonable judgment about the public activities that will exist in other jurisdictions at a future point in time after which they have relocated. As Stigler (1971: 130) has argued, “the voter’s expenditure to learn the merits of individual policy proposals and to express his preferences … are determined by expected costs and returns” and that, in the case of regulatory policy, this cost–​benefit calculus dissuades voters from discovering very much. When we view the location decision as involving a choice of public sector packages, and we take into account the fact that other jurisdictions are likely to be changing those packages over time, the informational demands that must be satisfied for the competitive system of governments to work effectively seem very great. The difficulty faced by citizens in evaluating the costs and benefits of government fiscal packages is no more evident than when local governments keep taxes low by using local debt or pension underfunding to finance current government services. When assumption 4 no longer holds, there is incomplete information about the costs and levels of governmental activity and an important principal–​agent problem is created. Households and firms have two options: go it alone and risk being exploited by better informed fiscal entrepreneurs and local factors of production, or hire an agent to protect household and firm interests. Complications arise when information about agent performance is costly and some deceitful agents survive, or when the expertise provided by agents to households and firms cannot itself be kept private. To the extent the market for “agents” works perfectly—​all inefficient or exploitive providers are exposed, and only truthful agents survive—​the final outcome will be efficient, but some fiscal surplus will be allocated to the agents as payment for services performed. Such payments can be made as a contingent fee conditional on the performance of the agent in managing the purchase of the public good, or as a competitively determined lump-​sum payment to a previously identified—​by reputation or certification—​“truthful” agent. When information about agent performance or abilities is available at no cost, then agent efforts can be fully monitored (avoiding all problems of moral hazard) and agent abilities will be fully known (avoiding all problems of adverse selection). The agent market can then work perfectly. When information about agents is costly, however, there is no guarantee that the efficient agent market will arise.15 Similarly, when the information provided by the agents to households cannot be kept private, then there will be externality-​caused inefficiency in the information market.16 The final assumption 5, requiring no spillovers, is an important real-​world violation of the Tiebout framework that may require central government intervention. The spillover literature has been critically reviewed and extended by Pauly (1970) among others.17 15 

See Holmstrom (1985). See Grossman and Stiglitz (1980). 17  For a discussion of tax externalities see the survey in Inman and Rubinfeld (1996). 16 

 

92    Robert P. Inman and Daniel L. Rubinfeld When a spillover offers a positive benefit to a neighboring community, for example, through the regulation of socially damaging air pollution emitted by a locally valuable firm, local governments may provide too little of the activity. The converse case provides a different example. Here a socially productive firm produces only local pollution, for example, a garbage or nuclear waste site. The incentive is for the local community to impose stringent (not in my backyard, or NIMBY) regulations on such a firm so as to force its relocation to a neighboring community, with the result being a negative regulatory spillover. If the neighboring communities respond with similar regulations, the firm may have no place to go, even though it is in the interests of all jurisdictions to locate the firm somewhere. With negative spillovers, there may be too much state and local activity.18 In these circumstances, a uniform national standard may be preferred by all. Uniform national regulation not only can take advantage of economies of scale in administration, it can also produce scale economies in production and distribution for firms, while at the same time reducing location costs.19 Augmentation of national standards by states or localities can be allowed, but only if the additional regulations do not generate spillovers by benefitting local households and businesses at the expenses of nonresidents. The general point is clear: when assumption 5 is violated, there is no guarantee that decentralized competition will produce an efficient allocation of public and private resources. Our review of the assumptions and evidence for Tiebout competition leads us to conclude that while many local governments can be an important element in the efficient federal constitution, this institutional structure alone will not be sufficient to ensure full economic efficiency.20 The central weakness is a failure to adequately allocate public goods and services with significant economies of scale—​thus losing assumption 1—​or to internalize the consequences of wide-​reaching positive or negative spillovers—​thus losing assumption 5. In these two instances, additional governmental structures will be required for economic efficiency. Economic federalism looks to a strong president elected by all citizens as a solution.

5.2.4 Presidential Governance Economic federalism adopts as its representation decision a single representative elected by all citizens to make economic policy at the level of the national government—​the president. If the institution of presidential governance is to have a chance at efficient performance, five assumptions must hold:21 18 

For an illustration of this point with respect to the regulation of solid wastes, see David and Lester (1989). 19  See, for example, Rose-​Ackerman (1994). 20  For a relatively recent review and update, see Oates (2005). 21 This political economy approach to managing the central government differs from the usual social welfare or mechanism design approach of public finance. The political economy approach designs an election process that reveals the leader’s preferences, as in Besley and Coate (1997).

 

Economics of Federalism    93 1. Open elections and voting: Any citizen may choose to run for office. There are no restrictions on the ability of individuals to vote for the candidate of their choice. 2. All candidates have equal abilities to deliver on their promised policies: Information regarding the likely success of a policy is common knowledge among all candidates. The executive adopts the best available incentive mechanism for the execution of the policy.22 3. Voters are informed of candidate preferences for policies: Newspapers, radio, and television coverage of candidates, and campaign advertising and debates are also important sources of information about candidates’ preferences. A free and competitive press, policy debates, and campaign advertising can all contribute to economic efficiency.23 4. Policy commitment: To ensure that dynamically efficient policies will be adopted, the executive must be able to commit to the continuation of an efficient policy, even after she has left office. Institutional checks and balances provide commitment devices for current presidents. Examples include independent courts, independent monetary authorities, and independent administrative agencies. The current president appoints the judges, bankers, and bureaucrats whose long-​run preferences tend to agree with his or her own. If the executive abolishes or ignores such agencies, the ability to use these institutions to ensure favored long-​run policies will be compromised.24 5. Majority rule selects the president in two candidate elections: Presidents will be elected in majority-rule elections. There will be no third candidate, since no voter will risk switching their vote unilaterally to a third candidate, thereby enhancing the chances that his second choice becomes a winner. Together these assumptions describe a world of “citizen candidates” where anyone can run for president, seek campaign financing, present her case to voters, and if achieving a majority of votes, serve as president. Once elected, executives can implement their policies. In a pair of important papers, Besley and Coate (1997, 1998) prove that such open two-​candidate elections with informed voters will elect a president whose policies are efficient in the sense that no other feasible candidate is preferred by another majority of citizens. Clearly, this result is one of only “constrained efficiency” because the incentive of citizens to participate needs to be respected and because majority rule may leave some citizens worse off. That said, presidential governance holds the promise of improved economic performance through an assignment decision that allocates responsibilities for congestible 22 

For analyses of the design of policies recognizing the incentives of policy recipients, see Myles (1995: Parts II and III) and for the literature on the design of incentives for public bureaucrats who provide policies, see Dixit (1996, particularly pp. 94–​107). 23  For a review of the role of media in promoting truthful information about candidates, see Gentzkow and Shapiro (2008). On the role of campaign financing in revealing each candidate’s true ability to deliver efficient policies, see Coate (2004). 24  Dixit (1996: 62–​85) provides an overview.

 

94    Robert P. Inman and Daniel L. Rubinfeld public goods to the many local governments of the Tiebout economy and the task of providing pure public goods and local services with significant inter-​jurisdictional spillovers to the nationally elected president. How might the president do? The evidence is supportive of improved economic efficiency. Persson and Tabellini (2003, 2004) find in cross-​country samples that countries with presidential democracies tend to adopt more efficient public budgets and as a result enjoy higher rates of economic growth than do parliamentary democracies. Further, presidential democracies are less likely to run large fiscal deficits over many years, consistent with more efficient intertemporal fiscal policies. In their study of the US historical record, Inman and Fitts (1990) find institutionally “strong” presidents—​as evaluated by historians—​outperform “weak” presidents in the management of domestic fiscal policies. They do so through the use of their administrative powers to fashion majority coalitions valuing aggregate economic efficiency over narrow special interests.25 As a result, domestic spending is reduced, and the savings are returned to citizens in the form of reduced taxation. The resulting public budget is more efficient. The economic successes of Presidents Theodore Roosevelt, Wilson, Franklin Roosevelt, Eisenhower, Johnson, and Reagan attest to the potential efficiency advantages of strong presidential governance. But with presidential governance, there is no guarantee that all citizens will benefit from improved national economic performance. The president needs only a majority for her re-​election. Political minorities may be asked to contribute more in taxes than the national public goods provide in benefits, or made to bear the burden of adverse national spillovers, say by the location of the nation’s nuclear waste site. Fairness aside, the favored president’s majority will be subsidized by the disfavored minority, leading to too much of the national public good or an oversupply of local goods that produce negative externalities. We look therefore for an alternative form of governance—​cooperative federalism—​which protects minorities and has the potential to ensure a more efficient overall allocation of public resources.

5.3  Cooperative Federalism Cooperative federalism begins, as did economic federalism, with the partition decision of the Tiebout economy creating many local governments for the allocation of congestible public goods. But as under economic federalism, we must provide a governance structure that might solve the problems of pure public goods and inter-​jurisdictional spillovers. Cooperative federalism provides for a bargaining environment that encourages all local jurisdictions, or subsets of these jurisdictions, to come together to collectively provide public goods with significant economies or to draft agreements that internalize the economic consequences of spillovers. Under the rules of this federal constitution, the representation decision grants the citizens of each local jurisdiction full 25 

See Fitts and Inman (1992).

 

Economics of Federalism    95 representation in any cooperative agreement—​called a “Coasian (1960) bargain”—​to provide public goods or control spillovers. Minorities can be excluded from an agreement, but that will only occur when the benefits of participating are less than the costs. Minorities cannot be forced to contribute without their consent. The assignment decision allocates all public services to local governments. Successful Coasian bargains require five assumptions to be met: 1. There are no, or very small, resource costs associated with the bargaining process. 2. Preferences over bargaining outcomes and the resources of households are common knowledge. 3. Bargaining agents perfectly represent the economic interests of their constituents. 4. All bargaining agreements are easy to enforce. 5. The parties can agree to a division of the bargaining surplus. Only when assumptions 1 through 5 hold will inter-​jurisdictional efficiency be achieved. How likely are these assumptions to be valid? On its face, assumption 1 seems defensible. The institutions for bargaining—​elected local officials—​are already in place, and the incremental costs of applying these institutions to an additional political agreement appear to be small. However, local agendas are limited, and at some point it simply does not pay to bring additional policies for consideration. One hopes that excluded policies are likely to be marginal, having only a small effect on the economic well-​being of voters. Even if 1 holds, assumptions 2 to 5 may fail. If the preferences of the participants to the bargain are not common knowledge (2 no longer holds), there may be a strategic advantage to concealing costs and benefits, demanding greater compensation from the agreement for the “winning” jurisdiction and expecting less for one of the “losing” localities. If both parties choose to “conceal” their strategic positions, no agreement may be forthcoming. In choosing a bargaining strategy, a jurisdiction must assess the threat points of the other parties—​the minimum offer the other party is willing to accept in order to consummate the bargain. In addition, a party will try to estimate the extent to which the other party will be willing to make concessions prior to reaching its threat point. If each party makes a poor estimate of the other’s threat point, or miscalculates the chances that the other party will accept a compromise, the bargaining process could break down. Clearly, Coasian bargainers are not immune to such possibilities.26 Indeed, attempts at free-​riding are likely to undo bargains when preferences are not common knowledge and the number of jurisdictions with affected consumers is large.27

26 

See Polinsky (1980), Crawford (1982), and Kreps and Wilson (1982). For the first treatment of the effects of group size on voluntary agreements see Olsen (1965). The effect of group size and other variables on bargaining has been studied extensively through the use of experiments (see, e.g., Hoffman and Spitzer, 1982; also Bagnoli and McKee, 1991). 27 

 

96    Robert P. Inman and Daniel L. Rubinfeld Even if consumer preferences and firm profits are common knowledge, there is no guarantee that elected local officials will choose to make the constituents affected by the Coasian bargain better off (3 will no longer hold). In some cases, favored industries might be well represented, but consumers may not be so fortunate. In contrast to private-​market agreements, the affected parties are heard only if elected representatives choose to make their case. Assumption 4 is perhaps the least problematic of the five Coasian assumptions. If jurisdictions do reach an agreement and sign a compact, the agreement will typically be legally enforceable. Ellickson (1979) discusses the idea of creating a series of public property rights and duties to encourage bargaining solutions to spillovers. Note that monitoring costs, a serious problem with respect to many contracts, are not likely to be a problem here. For example, jurisdictions which have agreed to curtail a negative spillover will know if they have been paid compensation. Previously harmed jurisdictions which pay compensation will be able to observe if the “polluting” jurisdiction has curtailed its damaging activity.28 Enforcement problems arise when important contingencies which might affect the agreement cannot be foreseen in advance. For example, if a future event occurs which alters the benefits and costs of a particular policy, the jurisdiction choosing the activity may decide to break the previous agreement. Arguably the most important and perhaps least appreciated source of bargaining failure is the loss of assumption 5. Even if 1 through 4 hold, the parties may not be able to agree as to how the economic surplus generated by the bargaining process can be divided. Each bargain does two things: it establishes an efficient exchange thereby creating economic surplus and it distributes that surplus among the bargaining parties. Proponents of Coase emphasize the first outcome, but tend to ignore the important difficulties of the second.29 Yet, the division of any economic pie is a bargaining problem which may have no solution. Whether such redistribution games have equilibrium outcomes has been hotly debated.30 For example, in “divide-​the-​pie” games bargaining often breaks down because the proposed division is not seen as fair by one or all of the negotiating parties. In such situations, one party rejects an economically beneficial offer from another because the proposed offer violates the first jurisdiction’s, possibly politically motivated, sense of economic fairness. Further, what is seen as an exogenous norm of fairness in a one-​time game may be a rational endogenously chosen strategy in a repeated political game. For example, a local official seeking re-​election cannot appear “weak” when bargaining with another jurisdiction. If the money involved is modest relative to the overall local budget, then the political symbolism of the “share” from the divide-​the-​pie game may be far more important for re-​election prospects than the actual dollars involved. If so, and if officials from 28  For a discussion of the importance of monitoring costs to contract enforcement, see Williamson (1985). 29  For the first to stress this important point, see Cooter (1982). 30  See, e.g., Rubinstein (1982) and Binmore, Rubinstein, and Wolinsky (1986).

 

Economics of Federalism    97 both jurisdictions demand more than half, agreement will not be possible.31 Strategic interplay becomes even more complicated and agreement less likely as the number of bargaining jurisdictions increases beyond two. This is true even if preferences are fully known and the free-​rider problem is not at issue.32 Our analysis of local political institutions reveals strong pressures towards favoring narrow, geographically concentrated interests.33 The hope that voluntary inter-​ jurisdictional compacts might control these inefficiencies is optimistic; the assumptions required for such Coasian agreements are demanding and unlikely to hold in practice. Difficulties are only compounded when local officials treat the division of proceeds from the bargain as a signal of their political “toughness.”34 To sum up, when any one of the assumptions 1 to 5 fail to hold, Coasian agreements will not arise and locally created inefficiencies will remain. Cooperative federalism, while promising in theory, is not likely to be the needed institutional solution to the efficiency weaknesses of strict economic federalism. We turn now to a third alternative, democratic federalism.

5.4  Democratic Federalism Democratic federalism, like economic and cooperative federalism, begins with the partition decision of the Tiebout economy using many local governments for the provision of congestible public goods, but offers an institutional compromise for the representation decision between those of economic and cooperative federalism. Rather than a single elected president for deciding pure public goods or allowed spillovers as in economic federalism or total representation of all jurisdictions as in cooperative federalism, democratic federalism allows for partial representation of the citizens of local governments through multi-​jurisdiction elections of representatives to the national government. Those representatives will then make decisions within a legislature by a majority (not unanimity) voting rule. When elected representatives join together in a national legislature, they are free to form coalitions and set policies, subject to the constraint that decisions receive the approval of a majority of legislators. Majority-​rule legislatures making decisions over multiple public policies must solve a fundamental structural problem: the propensity of the majority-​rule process to cycle from one policy outcome to another. The possibility of cycling has been well recognized at least since the work of De Condorcet (1785): even 31 

For evidence from the experimental literature that a sense of fairness is crucial to the ability of parties to reach agreement, see Roth (1985). 32  See, e.g., Haller (1986). 33  For evidence that state regulations are politically motivated, and in particular, are more likely for geographically concentrated industries and occupations, see Stigler (1971). 34  For evidence of the difficulties states have in fashioning Coasian agreements, even when the number of states is few, see Kolstad and Wolak (1983) and Coates and Munger (1995).

 

98    Robert P. Inman and Daniel L. Rubinfeld though majority rule leads a legislature to vote for policy A over policy B, and for policy B over policy C, it is certainly possible, and often likely, that policy C will defeat policy A. Arrow (1963) has demonstrated that cycling is a symptom of a more fundamental problem with all democratic processes: there is no guarantee that such processes can find a best outcome or be made immune to manipulation. Thus, the design of political institutions will almost inevitably be a search among the second-​best.35 When no winning coalition is capable of holding its majority against small policy variations offered by a losing minority, then either no decision will be made or final policy outcomes will be uncertain. If legislatures are to reach decisions, additional institutions are needed in order to overcome the inherent instability of the majority-​rule process. Two approaches are commonly used by legislatures. The first assigns agenda-​setting powers to a small subset of members, say the speaker of the house or a key legislative committee. Other members in the legislature then simply vote yes or no on the items in the approved agenda. Most likely, policies will be approved by a bare majority—​a minimal winning coalition—​in this strong agenda-​setter legislature.36 There is an extensive literature which tests for the direct influence of agenda-​setters—​typically congressional committees, on policy outcomes.37 A second strategy shares agenda-​setting powers among all members, giving each legislator a right to select his most preferred policy in that policy area most germane to the legislator’s constituents. This second approach to legislative decision-​ making involves each legislator deferring to the preferred policies of all other legislators, provided the other legislators defer to the legislator’s own policy requests. The guiding principle here is a norm of deference—​“You scratch my back, I’ll scratch yours”—​and it typically results in legislative proposals which are approved unanimously. For this reason such legislatures are often called “universalistic.”38 Note, however, that a norm of deference is not an appropriate model for predicting how legislatures will decide redistributive policies such as welfare or progressive taxation, politics where one group of constituents must lose to another.39 Uncertain as to whether he will be one of the winners in the bare majority legislature, an individual legislator will typically favor the more open rules of the universalistic legislature. The universalistic legislature is preferred because under universalism activities that benefit the legislator’s district are nearly certain to be chosen, and a substantial portion of the costs will be borne by other districts. If the benefits exceed average costs in a majority of districts—​a likely case—​then a legislature run by a norm of deference will be

35 

For a survey of this literature in the context of the new political economy, see Inman (1987). For articulation of the general theory of how this process overcomes cycling, see Baron and Ferejohn (1989). For a careful description of what legislative institutions are used to assign agenda control, see Shepsle (1979). 37  See Krehbiel (1992) for a critical but balanced review. 38  The guiding theory can be found in Weingast (1979) and Niou and Ordeshook (1985). 39  Details regarding the legislative institutions needed to enforce the norm of deference can be found in Weingast and Marshall (1988). 36 

 

Economics of Federalism    99 the preferred legislative institution.40 The conclusion that legislators prefer universalism to bare-​majority rule extends to this more general case. Universalistic legislatures operating under a norm of deference run a significant risk that their activities will be economically inefficient, however. The problem is much like that faced by a large group of friends who go to lunch together. The efficient thing to do is to charge each diner the true costs of their meal, including an apportioned share of all appetizers and desserts. But that is complicated, time-​consuming, and how do we know exactly who ate what? What typically happens is that all friends agree to share the check equally. If so, the incentive is for all diners to buy the largest and most expensive meal on the menu. They get the full benefit of the meal, but pay only a fraction of the costs. To not buy the expensive meal is to subsidize the consumption of all one’s friends, without being subsidized in return. Everyone eats too much, and total spending on lunch is inefficiently too much. So too it will be for legislators. Each asks for largest possible project valued by their constituents, with the costs of the project shared by all other legislators. If each legislator’s project benefits only her constituents, the result will be an inefficiently too large public budget. This temptation to overspend applies equally to the granting of tax favors and to government regulation.41 If so, how might we control such inefficiencies? One answer is to not allow legislators to eat lunch in big groups! More concretely, do not allow elected representatives to the national legislature to buy goods and services, grant tax favors, or regulate the economy when the benefits of those activities accrue only to narrow interest groups or to residents of each representative’s jurisdiction alone. The agenda for the national legislature should be restricted to only those goods, services, and regulations that benefit residents in multiple jurisdictions, or ideally, the nation as a whole. Democratic federalism’s assignment decision draws these lines. In the European Union, the guiding principle for assignment is called subsidiarity: Each government activity is to be assigned to that level of governance that most efficiently provides the service with the least economic spillovers, positive or negative, for its neighboring jurisdictions. In application, this means national defense, foreign affairs, telecommunication networks, ports and airports, interstate highways, public health, scientific research, free trade and competition policy, income redistribution, macroeconomic stabilization, and the taxation of mobile factors of production should be assigned to the national government. Alternatively, education, police and fire protection, health care, roads and transit, sanitation, recreation, cultural activities, and the taxation of immobile factors of production should be assigned to smaller local jurisdictions. This assignment of policy responsibilities gives each level of government a chance to do what it does best. The problem, of course, is enforcing such assignments.42 40  Weingast (1979) and Niou and Ordeshook (1985) prove exactly when this outcome holds. See also Inman and Fitts (1990), which extends the analysis to legislative projects of variable size. 41  See Inman and Rubinfeld (1996) and Inman and Rubinfeld (1997c). 42  Inman and Rubinfeld (1997c) have proposed one such process for enforcing assignment in universalistic legislatures called a Federalism Impact Statement, or FIST, modeled after the use of regulatory impact statements to control excessive central government regulations.

 

100    Robert P. Inman and Daniel L. Rubinfeld We should look to non-​constitutional protections against legislative inefficiencies too. Strong political parties capable of disciplining elected party members provide one possibility. Consider first the ability of strong political parties to control policy. Members who might be tempted to deviate from the party’s platform can be disciplined. Government policy will therefore be the policy which is in the interests of the majority party. In this stylized framework, only those producers in legislative districts belonging to the majority party will be favored. And when proposing an activity for their constituents, representatives in the majority party will be required by the party to consider the effects of their proposed activity on constituents in all districts controlled by the party. From the perspective of the majority, therefore, an activity in one district will still provide benefits to party constituents, but will impose only a fraction of the total cost of the activity on party members. From the majority party’s perspective, the preferred activity level in a party-​controlled jurisdiction will be efficient—​party marginal benefits will equal party marginal costs. Because the party-​preferred level internalizes the costs to party members, the inefficiency must be less than the inefficiency created by the universalistic legislature. Put simply, by internalizing at least half of the costs of government activity, strong political parties can improve the efficiency of national policymaking.43 Moreover, there is a possibility for full efficiency if the political party controls all representatives in the legislature. All districts will have an incentive to elect a representative to the stable majority party, since this is their only way to join the controlling coalition and to receive constituent benefits. With a single stable majority party, therefore, there is strong pressure towards achieving economic efficiency.44 Extra-​legislative resources to hold party discipline in place—​whether money or ideology—​will be required, however. Lacking such resources, representatives will gravitate back to universalistic behavior. Just as for economic federalism and cooperative federalism, democratic federalism cannot guarantee full economic efficiency. Legislative politics may encourage inefficient national governmental activities, but legislatures managed by strong political parties hold promise. In the end, democratic federalism may prove the most efficient of our possible forms of federal governance.

5.5  Choosing the Federal Constitution While the economic performance of a federal constitution is important, it is not the only value that will matter when selecting the governance structure for the economy. Also important, and in the end perhaps decisive, will be the values of democratic 43 

For empirical evidence that strong political parties may have controlled legislative inefficiencies, see Inman and Fitts (1990: pp. 111–​15). 44  Wittman (1989) makes this argument.

 

Economics of Federalism    101 participation, economic fairness, and the protection of personal rights and liberties. Both the empirical evidence and political theory suggest democratic participation will be enhanced by giving more weight to local governments in the assignment decision, thus favoring the decentralized institutions of economic or cooperative federalism.45 The value of economic fairness pushes in the opposite direction, leading us to favor democratic federalism and to assign a number of activities to the central government, particularly redistributive policies. For the protection of personal rights and liberties, matters are less clear. On the one hand, what Berlin (1969) calls “negative rights” which cannot be transgressed, are perhaps best protected by local jurisdictions coupled with the free choice of citizens as to their place of residence.46 For that choice to be meaningful, local governments will need to be assigned responsibility for providing important public services as well, particularly police protection. On the other hand, “positive rights” such as equal access to achieve one’s full potential will be best protected within democratic federalism with a strong central government assigned responsibility for equal access to education, health care, and perhaps a safe physical environment.47 In the end, the choice of an “optimal” federal constitution depends upon the relative importance one places upon the potentially competing values of efficiency, democracy, fairness, and personal rights and liberties. Trade-​offs seem inevitable.

References Arrow, A. K. (1963) Social Choice and Individual Values. New York: Wiley. Bagnoli, M. and McKee, M. (1991) “Voluntary Contribution Games:  Efficient Provision of Public Goods,” Economic Inquiry, 29: 351–​66. Banzhaf, S. and Walsh, R. P. (2008) “Do People Vote with their Feet? An Empirical Test of Tiebout’s Mechanism,” American Economic Review, 98: 843–​63. Baron, D. and Ferejohn, J. (1989) “Bargaining in Legislatures,” American Political Science Review, 83: 1182–​1206. Bartik, T. (1988) “The Effects of Environmental Regulation on Business Location in the United States,” Growth and Change, 19: 22–​44. Bergstrom, T., Rubinfeld, D., Roberts, J., and Shapiro, P. (1988) “A Test for Efficiency in the Supply of Public Education, Journal of Public Economics, 35: 289–​307. Berlin, I. (1969) Four Essays on Liberty. Oxford: Oxford University Press. Besley, T. and Case, A. (1995) “Incumbent Behavior: Vote-​Seeking, Tax-​Setting, and Yardstick Competition,” American Economic Review, 85: 25–​45. Besley, T. and Coate, S. (1997) “An Economic Model of Representative Democracy,” Quarterly Journal of Economics, 112: 85–​116. Besley, T. and Coate, S. (1998) “Sources of Inefficiency in a Representative Democracy: A Dynamic Analysis,” American Economic Review, 88: 139–​56. 45 

See Dahl and Tufte (1973), Finifter (1970), Pateman (1970), and Frug (1980). See, e.g., Madison (1982, Federalist 51) and Rapaczynski (1985). 47  See, e.g., Sen (1999). 46 

 

102    Robert P. Inman and Daniel L. Rubinfeld Binmore, K., Rubinstein, A., and Wolinsky, A. (1986) “The Nash Bargaining Solution in Economic Modelling,” RAND Journal of Economics, 17: 176–​88. Boadway, R. and Flatters, F. (1982) “Efficiency and Equalization Payments in a Federal System of Government,” Canadian Journal of Economics, 15: 613–​33. Break, G. (1967) Intergovernmental Fiscal Relations in the United States. Washington, DC: Brookings Institution. Breton, A. and Scott, A. (1978) The Economic Constitution of Federal States. Toronto: University of Toronto Press. Buchanan, J. M. and Goetz, C. (1972) “Efficiency Limits of Fiscal Mobility: An Assessment of the Tiebout Model,” Journal of Public Economics, 1: 25–​43. Coase, R. (1960) “The Problem of Social Cost,” Journal of Law and Economics, 3: 1–​44. Coate, S. (2004) “Pareto-​Improving Campaign Finance Policy,” American Economic Review, 94 (June): 628–​55. Coates, D. and Munger, M. (1995) “Strategizing in Small Group Decision-​Making: Host State Identification for Radioactive Waste Disposal among Eight Southern States,” Public Choice, 82: 1–​15. Cooter, R. (1982) “The Cost of Coase,” Journal of Legal Studies, 11: 1–​34. Courant, P. and Rubinfeld, D. (1981) “On the Welfare Effects of Tax Limitation,” Journal of Public Economics, 16: 289–​316. Crawford, V. (1982) “A Theory of Disagreement in Bargaining,” Econometrica, 50: 607–​38. Cumberland, J. (1981) “Efficiency and Equity in Interregional Environmental Management,” Review of Regional Studies, 2: 1–​9. Dahl, R. and Tufte, E. (1973) Size and Democracy. Stanford, CA: Stanford University Press. David, C. E. and Lester, J. P. (1989) “Federalism and Environmental Policy,” in J. P. Lester (ed.) Environmental Politics and Policy: Theories and Evidence. Durham, NC: Duke University Press, 57–​59. De Condorcet, M. (1785) “Essai sur l’Application de L’Analyse à la Probabilité des Decisions Rendues à la Pluraliste des Voix”, Paris. Dixit, A. (1996) The Making of Economic Policy. Cambridge, MA: MIT Press. Easterbrook, F. (1983) “Antitrust and the Economics of Federalism,” Journal of Law and Economics, 26: 23–​50. Ellickson, R. (1979) “Public Property Rights:  Vicarious Intergovernmental Rights and Liabilities as a Technique for Correcting Intergovernmental Spillovers”, in D. Rubinfeld (ed.) Essays on the Law and Economics of Local Governments. Washington, DC: Urban Institute, 51–​101. Epple, D. and Zelenitz, A. (1981) “The Implications of Competition among Jurisdictions: Does Tiebout Need Politics?” Journal of Public Economics, 89: 1197–​217. Finifter, A. W. (1970) “Dimensions of Political Alienation,” American Political Science Review, 30: 389–​410. Fitts, M. and Inman, R. P. (1992) “Controlling Congress: Presidential Influence in Domestic Fiscal Policy,” Georgetown Law Journal, 80: 1737–​86. Frug, G. (1980) “The City as a Legal Concept,” Harvard Law Review, 93: 1057–​154. Gentzkow, M. and Shapiro, J. (2008) “Competition and Truth in Market for News,” Journal of Economic Perspectives, 22: 133–​54. Gramlich, E. M. and Laren, D. S. (1984) “Migration and Income Redistribution Responsibilities,” Journal of Human Resources, 19: 489–​511.

 

Economics of Federalism    103 Gramlich, E. M. and Rubinfeld, D. L. (1982) “Micro Estimates of Public Spending Demand Functions and Tests of the Tiebout and Median Voter Hypotheses,” Journal of Political Economy, 90: 536–​60. Greenwood, M. J. (1986) “The Factor Market Consequences of U.S. Immigration,” Journal of Economic Literature, 24: 1738–​72. Grossman, S. and Stiglitz, J. (1980) “On the Impossibility of Informationally Efficient Markets,” American Economic Review, 70: 393–​408. Haller, H. (1986) “Noncooperative Bargaining of N 3 Players”, Economic Letters, 22: 11–​13. Hoffman, E. and Spitzer, M. L. (1982) “The Coase Theorem: Some Experimental Tests,” Journal of Law and Economics, 25: 73–​98. Holmstrom, B. (1985) “The Provision of Services in a Market Economy,” in R. Inman (ed.) Managing the Service Economy: Prospects and Problems. Cambridge: Cambridge University Press, 183–​213. Inman, R. (1987) “Markets, Government, and the New Political Economy,” in A. Auerbach and M. Feldstein (eds) Handbook of Public Economics, Vol. II. Cambridge, MA: MIT Press, 647–​777. Inman, R. P. and Fitts, M. (1990) “Political Institutions and Fiscal Policy: Evidence from the U.S. Historical Record,” Journal of Law, Economics and Organization, 6: 79–​132. Inman, R. P. and Rubinfeld, D. L. (1996) “Designing Tax Policies in Federalist Economies: An Overview,” Journal of Public Economics, 60: 307–​34. Inman, R. P. and Rubinfeld, D. L. (1997a) “Rethinking Federalism,” Journal of Economic Perspectives, 11: 43–​64. Inman, R. P. and Rubinfeld, D. L. (1997b) “The Political Economy of Federalism,” in D. Mueller (ed.) Perspectives on Public Choice. New York: Cambridge University Press, 73–​105. Inman, R. P. and Rubinfeld, D. L. (1997c) “Making Sense of the Antitrust State Action Doctrine:  Balancing Political Participation and Economic Efficiency in Regulatory Federalism,” Texas Law Review, 75: 1203–​99. Inman, R. P. and Rubinfeld, D. L. (2000) “Federalism,” in B. Bouchaert and G. DeGeest (eds) The Encyclopedia of Law and Economics, Vol. V [online] http://​www.e-​elgar.com/​shop/​encyclopedia-​of-​law-​and-​economics-​volumes-​i-​v, 661–​91. Kolstad, C. and Wolak, F. (1983) “Competition in Interregional Taxation: The Case of Western Coal,” Journal of Political Economy, 91: 443–​60. Krehbiel, K. (1992) Information and Legislative Organization. Ann Arbor, MI: University of Michigan Press. Kreps, D. and Wilson, R. (1982) “Reputation and Imperfect Information,” Journal of Economic Theory, 28: 253–​79. Ladd, H. F. and Yinger, J. (1989) America’s Ailing Cities. Baltimore, MD:  Johns Hopkins University Press. Levmore, S. (1992) “Bicameralism: When Are Two Decisions Better than One,” International Review of Law and Economics, 12: 145–​62. McConnell, K. and Schwab, R. (1990) “The Impact of Environmental Regulation on Industry Location Decisions: The Motor Vehicle Industry,” Land Economics, 66: 67–​81. Madison, J. (1982) The Federalist Papers by Alexander Hamilton, James Madison, and John Jay. New York: Bantam Classic Edition. Myles, G. D. (1995) Public Economics. Cambridge: Cambridge University Press.

 

104    Robert P. Inman and Daniel L. Rubinfeld Niou, E. M. S. and Ordeshook, P. C. (1985) “Universalism in Congress,” American Journal of Political Science, 29: 246–​58. Oates, W. E. (1972) Fiscal Federalism. London: Harcourt Brace Jovanovich. Oates, W. E. (2005) “The Many Faces of the Tiebout Model,” in W. A. Fischel (ed.) The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates. Cambridge, MA: Lincoln Institute of Land Policy, 21–​45. Oates, W. E. and Schwab, R. M. (1986) “Economic Competition among Jurisdictions: Efficiency Enhancing or Distortion Inducing?” Journal of Public Economics, 35: 333–​54. Olsen, M. (1965) The Logic of Collective Action. Cambridge, MA: Harvard University Press. Pateman, C. (1970) Participation and Democratic Theory. Cambridge:  Cambridge University Press. Pauly, M. (1970) “Optimality, ‘Public’ Goods, and Local Governments: A General Theoretical Analysis,” Journal of Political Economy, 78: 572–​86. Persson, T. and Talellini, G. (2003) Economic Effects of Constitutions. Cambridge, MA: MIT Press. Persson, T. and Tabellini, G. (2004) “Constitutional Rules and Fiscal Policy Outcomes,” The American Economic Review, 94 (1): 25–​45. Polinsky, A. M. (1980) “Resolving Nuisance Disputes: The Simple Economics of Injunctive and Damage Remedies,” Stanford Law Review, 32: 1075–​112. Rapacczynski, A. (1985) “From Sovereignty to Process: The Jurisprudence of Federalism after Garcia,” The Supreme Court Review, 1985: 341–​419. Revesz, R. (1992) “Rehabilitating Interstate Competition:  Rethinking the ‘Race-​ to-​ the-​ Bottom’ Rationale for Federal Environmental Regulation,” New York University Law Review, 67: 1210–​55. Rose-​Ackerman, S. (1981) “Does Federalism Matter? Political Choice in a Federal Republic,” Journal of Political Economy, 89: 152–​65. Rose-​Ackerman, S. (1994) “Environmental Policy and Federal Structure: A Comparison of the United States and Germany,” Vanderbilt Law Review, 47: 1587–​622. Roth, A. (1985) “Towards a Focal Point Theory of Bargaining,” in A. E. Roth (ed.) Game Theoretic Models of Bargaining. Cambridge and New  York:  Cambridge University Press, 259–​68. Rubinfeld, D. L. (1987) “The Economics of the Local Public Sector,” in A. Auerbach and M. Feldstein (eds) Handbook of Public Economics, Vol. 2. Amsterdam: North-​Holland, 571–​645. Rubinstein, A. (1982) “Perfect Equilibrium in a Bargaining Model,” Econometrica, 50: 97–​110. Samuelson, P. (1954) “The Pure Theory of Public Expenditure,” Review of Economics and Statistics, 36: 386–​9. Sen, A. (1999) Development as Freedom. New York: Anchor Books. Shepsle, K. (1979) “Institutional Arrangements and Equilibrium in Multidimensional Voting Models,” American Journal of Political Science, 28: 27–​59. Stigler, G. (1971) “The Theory of Economic Regulation,” Bell Journal of Economics and Management Science, 2: 3–​21. Tiebout, C. (1956) “A Pure Theory of Local Expenditures,” Journal of Political Economy, 64: 416–​24. Weingast, B. R. (1979) “A Rational Choice Perspective on Congressional Norms,” American Journal of Political Science, 23: 245–​62.

 

Economics of Federalism    105 Weingast, B. R. and Marshall, W. (1988) “The Industrial Organization of Congress,” Journal of Political Economy, 96: 132–​63. Wildasin, D. E. (1986) Urban Public Finance. Chicago, IL: Academic Publishers, Harcourt Press. Williamson, O. E. (1985) The Economic Institutions of Capitalism. New York: The Free Press. Wittman, D. (1989) “Why Democracies Produce Efficient Results,” Journal of Political Economy, 97: 1395–​424.

 

Chapter 6

E c onomics of Tax  L aw Daniel Shaviro

6.1 Introduction Many volumes could be, and have been, devoted to the question of how the law of, say, torts, contracts, corporations, civil procedure, or intellectual property could be designed with an eye to maximizing economic efficiency. In tax law—​leaving aside Pigovian taxes, which address negative externalities—​the question of how to maximize efficiency is almost childishly easy to answer. Unfortunately, however, the answer has very little direct relevance to understanding either existing law or potentially desirable tax reform. Suppose the government has a fixed cost of providing public goods, and that tax revenues will be used to meet this cost. Or suppose instead that whatever tax revenues are collected will be handed out again to various people, in amounts that are unrelated to how much tax they individually paid. In either case, taxation involves an externality problem. When people make choices that would affect their tax liabilities, they only have reason to care about their after-​tax positions. But the tax payments go to someone,1 creating a positive revenue externality. Thus, suppose there is a 30 percent wage tax and that, by working for an hour, I could earn $100 before tax. In deciding whether to take the job, I will only consider the $70 that I would get to keep, not the full $100 pre-tax wage that presumably reflects the value of my production to someone else. If my reservation wage, reflecting my disutility of working, is between $70 and $100 and I thus turn down the job due to the tax, then not only is no tax revenue generated in this instance, but the wage tax prevents the creation of surplus (from the excess of $100 over my reservation wage). It thereby gives rise to deadweight loss. Given revenue externalities and the consequent prospect of deadweight loss, maximizing the efficiency of non-​Pigovian taxes is straightforward. They should be 1 

Tax revenues that pay for providing public goods can be thought of as going to whomever would otherwise have borne the cost of such provision.

 

Economics of Tax Law    107 lump-​sum—​that is, invariant to taxpayer decisions. (Or they might be effectively lump-​ sum, by reason of depending on decisional margins that are wholly inelastic.) If I will owe the government the same amount of tax no matter what I do, tax-​induced deadweight loss will not arise The best-​known example of a lump-​sum tax is a uniform head tax—​assuming that one truly cannot avoid it, such as by exiting the jurisdiction or being too poor to pay. Likewise, consider a height tax, if people cannot affect their height (see Mankiw and Weinzierl, 2010). As we will see below, the optimal income tax (OIT) literature, founded by Mirrlees (1971), posits the theoretical possibility of a lump-​sum tax based on “ability,” taken to be an entirely fixed trait, but assumed not to be directly observable, making it unavailable as a tax instrument. From the standpoint of efficiency, no lump-​sum tax is better than any other—​by definition all succeed equally in avoiding the creation of deadweight loss.2 This leads directly to two main questions. First, why are lump-​sum taxes, or instruments that come as close to them as possible, so absent, not just in actual practice but even in theoretical debate about tax policy? The answer turns on the importance of distributional issues. Second, how do considerations of efficiency operate once we have accepted, for distributional reasons, the need for tax instruments that (like the wage tax noted above) have the unfortunate side-​effect of discouraging productive activity?

6.2 Distributional Issues A longstanding tradition in tax policy thinking holds that tax burdens should vary based on “ability” or “ability to pay” (Murphy and Nagel, 2002: 20). Under this view, Bill Gates should pay more tax than a middle-​class worker, who should pay more tax than a homeless person. The question of whether this involves “redistribution” from high earners to low earners depends on one’s baseline for assessing the “re” in “redistribution.”3 However, subject to questions of tax incidence,4 a system that does this clearly is

2 

Lump-​sum taxes may affect behavior, without creating deadweight loss, by reason of income effects. For example, work decisions may be influenced by lump-​sum taxes that affect the resources one has available to meet basic needs. My reservation wage might be higher, for example, when lump-​sum taxes are high than when they are low. However, lump-​sum taxes will not induce me to reject jobs that offer more than my reservation wage, or to accept those that offer less. 3  Redistribution by the tax system implies that there is a pre-​t ax or nontax baseline in which wealth was differently distributed, but what that means, and how it should be interpreted, is unclear. 4  The economic incidence of a tax, or how it changes the distribution of resources in the society, may differ from the nominal incidence, or who literally makes the tax payment. For example, the effects of a tax on prices may cause its incidence to be shifted from the nominal taxpayer to someone else. While I do not discuss incidence at any length in this chapter, useful quick introductions can be found, for example, in Rosen and Gayer (2008: 304–​7) and Bradford (1986: 133–​9).

 

108   Daniel Shaviro progressive relative to levying a uniform head tax, which presumably would be acceptable in the absence of distributional concerns.5 There are numerous potential motivations for seeking greater after-​tax distributional equality than that which would result from levying a uniform head tax. For example, one’s social welfare function might treat greater equality as good in itself. Or one might believe that inequality yields negative externalities, such as from its aiding capture of the political system by the rich that leads to pervasive rent-​seeking. Or one might view it as limiting economic opportunity for children born into the lower strata, or as yielding a more hierarchical and stratified society that generates net unhappiness. In a conventional welfare economics framework, the chief rationale for requiring high earners to pay more tax than low earners is the paired set of assumptions holding that people’s utility functions (a) are similar and (b) feature declining marginal utility (DMU) for wealth. Of these two assumptions, while (a) might be viewed in the main as a non-​falsifiable default stance, (b) arguably is supported empirically by the prevalence of risk aversion, and explained theoretically on the ground that we generally satisfy our most urgent needs first, moving to the ever more psychically dispensable as our budget lines move outward. Acceptance of the two above assumptions suggests that, all else equal, the disutility from bearing a dollar of tax burden steadily declines as one’s material circumstances improve. Suppose that one wholly ignores incentive effects, assumes that people are relevantly identical other than in how well-​off they are, and treats DMU as the sole relevant consideration in assessing their welfare losses from bearing tax burdens. The conclusion that would follow, under a welfarist approach to tax policy,6 is that the government should completely equalize after-​tax distribution, such as by expropriating all wealth (i.e., taxing it at a 100 percent rate) and then distributing it pro rata (Edgeworth 1897: 553). However, ignoring incentive effects is no more plausible, from a welfare standpoint, than ignoring distributional effects and thus endorsing a uniform head tax. The central task that both a 100 percent tax and a uniform head tax ignore is how to handle the trade-​off between efficiency and distributional concerns. The OIT literature squarely addresses this trade-​off.7 A  very basic model, derived from Mirrlees (1971), might proceed as follows. Suppose that people derive utility from market consumption and from leisure, each of which has DMU. Market consumption is funded through work (this being a one-​period model, and hence without saving). 5 

The aim of pursuing distributional aims through the tax system rests on the views, not only that distributional concerns are relevant, but that the tax system is a suitable place to address them. In the law and economics literature, this is generally accepted, with the main point in contention being whether other parts of the legal system should also play a role. See, e.g., Kaplow and Shavell (1994); Sanchirico (2000). 6  A welfarist or welfare economics approach can be defined as one in which “social welfare is posited to be an increasing function of individuals’ wellbeing and to depend on no other factors” (Kaplow and Shavell 2002: 24). 7  For further background concerning the OIT literature, see, e.g., Slemrod (1990); Mankiw, Weinzierl, and Yagan (2009).

 

Economics of Tax Law    109

Leisure (hours)

A Clarissa Brian

Earnings ($)

Figure 6.1  Individuals with different wage rates. Source: Figure 6.1 is adapted from Shaviro (2007: 755).

People differ only in ability or wage rate, defined as the amount that one can earn per unit of time or effort—​inputs that are not themselves directly observable. Ability is innate, but cannot directly be observed. By contrast, earnings, which are the joint product of ability and time or effort, can be observed. People can work as much or as little as they like, up to a maximum (such as the number of waking hours in a day). Each individual chooses the level of labor supply that would place her on the highest indifference curve that is tangent to the budget line resulting from the available time and the wage rate—​as in Figure 6.1, which features two individuals with different wage rates. Not surprisingly, in light of their identical utility functions, Clarissa, who has the higher wage rate of the two, ends up with both more earnings (and thus more market consumption) than Brian, and more leisure. Given their identical utility functions, featuring DMU for market consumption (as well as leisure), taking money from Clarissa and giving it to Brian would, up to a point, increase his utility more than it would reduce hers. The mechanism employed is a tax on earnings that funds a uniform cash grant—​ resulting in redistribution since Clarissa pays more tax than the amount she gets back, while Brian pays less. The income effect of this transfer, considered in isolation, would induce Brian to work less than if the tax-​plus-​grant program did not exist, and Clarissa to work more.8 This, however, merely reflects re-​optimizing by both parties.9 More problematic is the program’s substitution effect, inducing both parties to work less, given that they would lose a portion of their earnings to the tax collector. The program therefore discourages labor supply, creating deadweight loss and reducing the amount of redistribution that is desirable. 8 

Having more wealth reduces the marginal utility of the last dollar earned, so Brian would be expected to shift towards greater consumption of leisure. By contrast, having less wealth induces Clarissa to re-​optimize by working more. 9  It is true, however, that the income effect on Brian has a negative revenue externality, while that on Clarissa has a positive revenue externality.

 

110   Daniel Shaviro Unsurprisingly, the OIT literature finds that tax rates and the size of the grant are negatively correlated with labor supply responsiveness. They are positively correlated with (1) the slope of the DMU curve (and thus the welfare gain from each dollar that is redistributed), (2) the degree of dispersion in ability levels, and (3) any pro-​egalitarian weighting in one’s social welfare function (Slemrod 1990:  165). They also should be positively correlated with the size of any negative externalities that result from wealth inequality. An apparent key finding of the OIT literature, accepted for many decades though more recently challenged, was that tax rates should be relatively flat (Slemrod 1990: 165). The underlying reasoning is straightforward. On the one hand, DMU supports rate graduation, because the last dollar earned by a high earner purchases less utility than the last dollar earned by a low earner. On the other hand, however, lower-​rate brackets are effectively inframarginal for those who know with certainty that they will earn enough to place out of these brackets.10 This causes such brackets generally to offer more revenue bang for the buck, in efficiency terms, than higher-​rate brackets, which tend to be inframarginal for a smaller percentage of the individuals who are subject to them. This consideration pushes back against rate graduation (Shaviro 2011: 832), and indeed offers theoretical support for the conclusion that the marginal rate at the very top of the income distribution should be zero.11 In many OIT models, these offsetting considerations have yielded a marginal rate structure that is relatively flat. However, Diamond and Saez (2011) argue, within the OIT framework, for steep rate graduation at the top. Suppose that “the social marginal value of consumption for top-​bracket tax filers” is so low that “as a first approximation [it] can be ignored” (p. 168)—​i.e., treated as indistinguishable from zero.12 Then the socially optimal marginal tax rate for such high-​end income is the revenue-​maximizing rate, which Diamond and Saez argue may exceed 70 percent (p. 171).13 What is more, if one assigns close-​to-​zero weight to marginal welfare losses incurred by people at the top of the income distribution, “there is little lost if we adopt a policy of increased marginal tax rates that ends up losing a small amount of revenue” (Auerbach 2012: 703). If the exact revenue-​maximizing rate at the high end is uncertain, one might even view as symmetric 10  Thus, consider the Social Security portion of the payroll tax, which imposes an aggregate 12.4 percent tax rate on wages up to its dollar cap ($117,000 for 2014). This tax is projected to raise more than $600 billion annually, yet it has no effect on the marginal work incentives of people who are certain to earn more than the taxable amount. 11  As Slemrod (1990: 164) notes, “raising the marginal tax at the [very] top above zero distorts the labor supply decision of the highest earner but raises no revenue.” However, because this result only holds “precisely at the top,” he concludes that it “should not be taken seriously as a practical guide to tax policy.” 12  This might reflect additional consumption’s lack of marginal utility to the high earner herself, who might already have far more wealth than she could ever expect to spend. Or it might reflect egalitarian social weighting, or the view that any positive marginal utility to consumption at the top of the income scale is likely to be offset by negative externalities associated with high-​end wealth concentration. 13  The assumptions underlying this result include the existence of a fat tail at the top of the income distribution, and the absence of a finite maximum income level.

 

Economics of Tax Law    111 the social cost of mistakenly setting the top rate below the revenue-​maximizing level, and above it.

6.3 Efficiency Issues Once we are taxing productive economic activity for distributional reasons, rather than imposing a lump-​sum levy of any kind, the aim of efficiency analysis of taxation is to minimize the harm done, rather than to achieve a first-​best optimum. Here two alternative approaches—​directly inconsistent with each other—​may apply. The first is neutrality, while the second is differential taxation in response to any of several distinct rationales that include varying elasticities.

6.3.1 Tax Neutrality In a simple one-​period OIT model, the tax base is earnings, or equivalently market consumption (since savings would require adding the dimension of time). Once we make things more complicated, however, tax base design becomes an issue. Even leaving aside the tax treatment of savings (discussed below in relation to the choice between income taxation and consumption taxation), there is the question of whether all items that are potentially in the base—​for example, all earnings or all consumer commodities—​should be treated the same by the tax system. The efficiency case for tax neutrality, as between types of work and/​or alternative commodities, is straightforward. Given the point that paying taxes gives rise to a positive revenue externality, deadweight loss results when taxpayers switch from higher-​taxed to lower-​taxed choices by reason of the tax disparity. Only through tax neutrality, in the sense of taxing equally valuable items the same, can one avoid tax distortions with respect to the choice between taxable items. Moreover, even the fact that one is generally discouraging labor supply and market consumption via the tax does not change this analysis, absent further considerations that are discussed in section 6.3.2. This analysis generally holds even where tax disparities would have progressive distributional effects. Thus suppose one is considering imposing an extra tax on luxury goods, such as yachts, where the alternative would be to raise tax rates on the rich. At a first approximation, the luxury tax does nothing to ease the adverse incentive effect on labor supply, since high earners will still face a higher tax burden if they work more in order to finance a yacht purchase. In addition, however, one is distorting their consumption choices between yachts and other items. A similar analysis applies to tax-​favoring necessities (which may form a higher percentage of low-​earners’ than high-​earners’ consumption bundles) for distributional reasons.

 

112   Daniel Shaviro

6.3.2 Deliberately Non-​Neutral Taxation The efficiency case for tax neutrality within the realm of productive economic activity and market consumption assumes the inapplicability of a number of considerations that might require modifying the analysis. For example, choices that have positive or negative externalities should be encouraged or discouraged, respectively, whether or not the Pigovian element is administratively integrated with the distributionally minded tax instrument. Moreover, if people have cognitive difficulties in optimizing their pre-​ tax choices, then, whether or not tax incentives and disincentives (as distinct from “nudges”) can be counted on to improve things, the actual payoff to neutrality may be reduced or put into question. Also, taxing particular choices that serve as a signal or tag of ability or wage rate can potentially reduce overall distortion even though it inefficiently discourages those particular choices.14 There is also, however, a potential rationale for differential taxation that responds to the basic setup of the optimal tax problem—​that is, the fact that only earnings and/​ or market consumption generally fall within the scope of the tax. Consider the optimal commodity tax literature (derived from Ramsey 1927), which addresses efficient tax rate-​setting for different consumer goods, and finds that varying rather than uniform rates may be optimal. The basic setup here is as follows. Suppose the government is trying to raise revenue through taxes on specific consumer items or commodities. What is more, suppose that commodity taxation’s sole purpose is to raise revenue for the government at the lowest possible efficiency cost—​rather than, say, either to address externalities like a pollution tax, or to address distributional goals. Finally, suppose that lump-​sum taxes, such as uniform head taxes, are unavailable, as are comprehensive income or consumption taxes. Against this background, one might initially suppose that it would be most efficient for all of the taxable commodities to face the same tax rate. Ramsey showed, however, that the goal of minimizing overall deadweight loss, by equalizing each particular commodity tax’s marginal deadweight loss, had instead a very different implication. In particular, suppose that some commodities are more tax-​elastic than others, meaning that taxpayers would substitute away from them more if all commodities were taxed at the same rate. In general, the deadweight loss per dollar of revenue raised generally rises with the tax rate. Thus, each commodity tax would be expected to feature a rising ratio of deadweight loss to revenue as one increases its rate. Under these conditions, the way to minimize deadweight loss, while keeping overall revenue constant, is to have a higher tax rate on commodities that are relatively inelastic, and a lower one on those that are relatively tax-​elastic (Gruber 2011: 602–​3). 14  Suppose, e.g., that one reduced high-​end marginal tax rates while imposing a tax burden of some kind on the decision to attend elite universities. If attending such universities is a signal or tag that is associated with high ability (as distinct from being a way to increase one’s earning potential), then it is theoretically possible that this would permit reducing the efficiency cost of redistribution.

 

Economics of Tax Law    113 With an OIT-​style income tax in place, the relevance of arguments drawn from optimal commodity taxation narrows significantly. When one adds distributional considerations to the analysis and assumes, unlike Ramsey, that one can uniformly tax all consumer goods, uniformity is generally best after all (Atkinson and Stiglitz, 1976). The key underlying idea here is that one needs to embed commodity choice in a broader framework, in which people are deciding how much to work and earn. After all, without earnings one cannot pay for consumer goods. In this framework, it turns out that the only reason for differentially taxing particular commodities relates to the underlying labor supply choice that both the income tax and more limited commodity taxes distort, by discouraging market consumption (and therefore work). In light of this efficiency problem, differential commodity taxation can reduce the labor–​ leisure distortion if it involves tax-​favoring leisure substitutes—​that is, commodities that tend to be consumed more when leisure declines—​and/​or tax-​penalizing leisure complements—​that is, goods that tend to be consumed jointly with leisure (Corlett and Hague, 1953). Suppose, for example, that restaurant and take-​out meals are leisure substitutes, due to their appeal to people who are working long hours, while groceries are a leisure complement, because one needs free time to cook at home. Then it may be efficient to tax groceries and subsidize restaurant meals, as a way of lowering the labor–​leisure distortion (Kaplow, 2008: 138 n. 16). Otherwise, however, “results derived in the original Ramsey framework, in which no income tax is available, fail to provide proper guidance in a world with an income tax” (p. 148).15

6.3.3 Political Economy As the previous discussion shows, despite the revenue externality problem that pushes in favor of tax neutrality, there are a number of reasons why well-​tailored departures therefrom may increase efficiency without undermining the achievement of distributional aims. There is also, however, a political economy argument for generally treating tax neutrality as a core principle. Suppose that, in political and administrative practice, tax neutrality is easier to evaluate and apply than the variegated set of arguments that may support treating particular items more favorably than others. Then, in a political environment rife with agency costs, interest group influence, and low information on the part of voters, generally steering for greater rather than lesser tax neutrality may be likely to result in the achievement of greater tax efficiency in practice.

15 

But see Shaviro (2014: 164–​5) for an example in which the conclusion from Kaplow 2008 does not apply because the requisite income tax is not sufficiently available (e.g., because resident individuals cannot be taxed on foreign-​source income earned through foreign corporations).

 

114   Daniel Shaviro

6.3.4 Relevance of Administrative Constraints The discussion thus far has mainly been at a high level of abstraction. For example, it generally assumes the feasibility of a comprehensive tax on people’s earnings and/​ or on all market commodities (though not leisure), with behavioral responses being limited to those that involve labor supply or commodity choice—​as distinct from, say, tax evasion or avoidance. For both political and administrative reasons, actual income taxes are generally quite far from having so simple and straightforward a character. Some departures from the idealized model involve preferential treatment for particular items or activities—​for example, in the United States, owner-​occupied housing and employer-​provided health insurance. Often, such tax preferences are widely viewed by tax policy experts as having little merit. Even where arguably meritorious, they may have design problems that reflect their having taken the form of “tax expenditures”—​ that is, spending or allocative policy that is funneled through the tax system—​in large part for optical reasons, so that officially measured “taxes” and “spending” will appear to be lower (see Shaviro, 2004a). Real-​world income tax systems also commonly reflect administrative constraints that can cause them to depart greatly in practice from the idealized models. Examples include the realization requirement, under which gains and losses from changes in asset value are not accounted for by the tax system until a transaction such as a sale occurs, and the primary use of entity-​level, rather than owner-​level, taxation of corporate income. These rules can cause huge swathes of economic income to be taxed very differently than close economic substitutes, and can also induce substantial deadweight loss within their operating realms. Further, behavioral responses to taxation can involve, not just labor supply and commodity choice, but a spectrum of tax avoidance and evasion behaviors, in which legal formalities and/​or enforcement efforts play paramount roles in determining what is paid. The issues thereby raised greatly complicate optimal tax design and our ability to understand the practical consequences of a given set of tax rules (Slemrod and Yitzhaki, 2002). These features of real-​world tax systems create complicated second-​best problems for policymakers. For example, they may support the enactment of rules that deliberately mismeasure the taxpayer’s income in one respect, as an offset to what is at least suspected mismeasurement in other respects. Two illustrations, drawn from the US federal income tax system, are the following: (a) Indirect partial repeal of the exclusion for imputed rental income: Homeowners generally can exclude the imputed rental income that they derive by living rent-​ free on their own properties. The exclusion distorts choices between home ownership and rental, as well as between home and non-​home consumption. If we take its existence as given for political or administrative reasons, but wish to

 

Economics of Tax Law    115 address its distortionary effects, one way to proceed would be by denying deductions for home mortgage interest. Suppose a homeowner has a mortgage loan, generating annual interest expense. Interest deductions generally should be allowed in a comprehensive, well-​functioning income tax, as they are (negative) returns to (dis-​)saving, and thus the mirror image of positive returns to saving that an income tax reaches. However, denying home mortgage interest deductions can function as the equivalent of partial, indirect repeal of the exclusion for imputed rental income (Shaviro 2000: 144). In fact, the US rules allow home mortgage interest deductions, subject to certain limitations,16 but this is widely viewed as a deliberate tax preference that, while strongly entrenched politically, attracts little normative support in academic tax policy debate. (b) Loss limits and nonrefundability: In a comprehensive, well-​functioning system, just as economic gains would be fully taxed, so economic losses would be fully deductible. This would include the provision of overall refundability (i.e., the allowance of net refunds). If one instead has loss limits and nonrefundability, one discourages risk-​taking, by causing investments with greater variance to face higher expected tax burdens than those with the same expected pretax returns but less variance.17 Nonetheless, the US tax system not only makes overall losses nonrefundable (although they may be allowed against net taxable income in other years), but disallows net losses in particular income categories. For example, net capital losses generally cannot be deducted in a given year.18 The rationale for such rules is that, given tax planning opportunities, one cannot be confident that taxpayers claiming losses have actually suffered a net loss economically. For example, despite a wide range of anti-​tax shelter rules, taxpayers may be able to arrange transactions in which, due to the realization requirement and the heterogeneity of tax accounting rules, circular cash flows generate enormous 16 

Home mortgage interest is allowable only as an itemized deduction, thus denying it to some individuals who claim the standard deduction in lieu of itemized deductions. In addition, the mortgage loan principal that can generate allowable interest deductions is capped at $1 million for acquisition indebtedness, plus an additional $100,000 of other indebtedness that is secured by the home. See US Internal Revenue Code section 163(h)(3). These amounts are not indexed to inflation, and thus may gradually decline in real terms. 17  Graduated marginal rates generally have this effect, reflecting that nonrefundability can be viewed as an example of rate graduation (i.e., from zero to a positive rate as net income crosses the zero threshold). 18  See US Internal Revenue Code section 1211. Individuals can deduct net capital losses of up to $3,000 per year. Disallowed net capital losses can generally be used in other tax years to offset net capital gains. It also is common for net capital gains to face a lower tax rate than ordinary income. See, e.g., US Internal Revenue Code section 1(h). This, however, may be viewed as relating to the fact that realization elasticity may tend to be higher for capital gains than ordinary income. For example, I may find it easier to continue holding appreciated stocks in my portfolio for tax reasons, than to adjust my labor supply.

 

116   Daniel Shaviro claimed losses. To explain this generically: suppose that I both pay and receive $1 billion from the same counterparty (which can be accomplished simply by swapping pieces of paper). If I am able to claim, for some technical reason, that the $1 billion I paid is currently deductible, while the $1 billion that I received is not currently includable, then I have a net $1 billion loss deduction. This amount can easily be increased to $10 billion or $100 billion, by simply adding zeroes to the swapped pieces of paper. Nonrefundability limits the potential economic benefit from engaging in such transactions, admittedly at an arbitrary point (i.e., when taxable income reaches zero), thus in effect capping the potential fiscal harm. Disallowance for net capital losses responds to an even easier tax-​planning idea, which is to hold a well-​diversified portfolio of individually risky financial assets, and then to sell the losers while holding the winners, thus creating a tax loss even if there is no overall economic loss. The incentive to engage in such strategic trading is, of course, inevitable in a realization-​based system. Once again, denying net capital losses limits the net benefit available from engaging in this strategy, admittedly at an arbitrarily determined point. Rules of this kind raise complicated efficiency issues. For example, they increase tax-​ planning costs in some instances, by inducing taxpayers not to bother generating loss deductions, but increase such costs in others, by motivating efforts to circumvent the rules. Such issues lie at a considerable remove from the world of clean, analytically tractable economic models such as those from the OIT and optimal commodity tax literatures, but they cannot be ignored if one wishes to understand the main battlegrounds in a real-​world income tax system.

6.4  Income Taxation and Consumption Taxation For forty years, one theoretical issue has had pride of place in the US academic tax policy literature. This is the question of whether the main tax instrument used in distribution policy should be an income tax or a consumption tax.19 The great majority of countries use both income taxes and value-​added taxes (VATs) that generally treat current-​year market consumption as the tax base. In the United States, however, with general consumption taxes being limited to relatively small levies (retail sales taxes or RSTs) that are imposed at the state or local government level, much attention has focused on the idea of replacing the existing income tax with a progressive consumption tax. Such a tax would be partly or wholly collected from individuals or households, rather than from

19 

The US academic debate concerning income and consumption taxation was effectively launched by Andrews (1974).

 

Economics of Tax Law    117 businesses like a VAT or RST, thus permitting the application of a graduated rate structure based on overall personal circumstances.20 Perhaps the main virtue of this debate, despite its narrow focus and apparently dim prospect of ever leading to the enactment of major policy changes, is that it has served as a powerful instrument for advancing theoretical understanding of intertemporal issues. The big difference between an income tax and a consumption tax is that the former reaches, while the latter excludes, normal returns to saving (which may be thought of as funding future consumption). By examining the significance of this difference, the income versus consumption tax debate has helped to focus attention on the significance of the time dimension in evaluating both efficiency and distribution, as well as on a variety of administrative and other tax base design issues. In briefly reviewing the main contours of this vast, intricate, and multifaceted debate, it is useful to start with the famous Haig-​Simons definition of economic income (set forth, e.g., by Simons, 1938: 50). If Y denotes income, C consumption, and ΔW change in net worth (i.e., the return to saving), then one’s economic income for a given accounting period, such as a year, can be expressed as follows:

Y = C + ∆W

A consumption tax, in contrast to an income tax, only reaches C. For example, a VAT or RST applies to consumer purchases.21 However, one can also directly tax consumption at the household or individual level, without requiring people to keep track of all their consumer purchases, by subtracting ΔW from each side, and then rearranging the terms in the Haig-​Simons equation as follows:

C = Y − ∆W

In practice, this means that an individual or household-​level measure of income that is modified to allow deductions for all saving, while also requiring inclusion of all dissaving, can reach the same consumption base as a VAT or RST. And with measurement taking place at the individual or household level, graduated rates can be imposed that reflect the taxpayer’s overall circumstances (such as her overall consumption for the year, or over a longer period).22 20   In practice, existing VATs and RSTs typically leave a lot of consumption out of the tax base. In addition, in practice RSTs often reach business-​to-​business transactions, which do not involve consumption as properly defined. 21  In a typical VAT, business-​to-​business sales are taxable to the seller, but the buyer gets an offsetting credit or deduction. The net tax imposed on business-​to-​business sales is therefore zero, just as under an RST, where they are disregarded. 22  Consumption taxes that use this methodology are known as cash-​flow or consumed income taxes. The other main vehicle, within a consumption framework, for using individual or household-​level information to guide the imposition of graduated marginal rates, is the flat tax or X-​tax. This, in turn, is basically a flat-​rate business-​level VAT, modified to provide a business-​level deduction for wages. The wages are then taxed to workers under a graduated rate structure. See Hall and Rabushka (1995); Bradford (2003); Carroll and Viard (2012).

 

118   Daniel Shaviro What are the consequences of excluding ΔW from the tax base? If implemented naively, such as by excluding capital gains from the tax base even if they actually reflect labor income,23 people such as Bill Gates and Mark Zuckerberg would be able largely to exempt the huge fortunes that they have reaped via stock appreciation. However, the relevant tax policy literature shows that—​at least with proper implementation and complete capital markets, and if one uses an unlimited time horizon to make the assessment—​the difference between an income tax and a consumption tax is limited to the normal risk-​free return to saving, which the former, but not the latter, subjects to tax. What do we mean by the “normal risk-​free return?” This is what remains, out of the economic returns that people derive from or that are associated with their investment outlays, once one excludes the following three items, which the literature shows that an income tax and a consumption tax should in principle treat the same: (1) risk premia, (2) compensation for expected inflation, and (3) extraordinary returns, such as Bill Gates’ big score from founding Microsoft. One might think of the remaining residual as akin to the interest rate that a bank would offer on a secure savings account (allowing demand withdrawals, to eliminate interest rate risk) in the absence of expected inflation, and if the bank was not bundling its provision of financial services into the terms of deal. While the full analysis that is needed to reach (or contest) this conclusion is beyond my scope here,24 a simple example helps to make the distinct but related point that, with limited time and just a single rate of return (well below that which a Gates or Zuckerberg historically received on their main business ventures), a consumption tax, by reason of its excluding that return, is neutral as between current and future consumption. An income tax, by contrast, penalizes future consumption relative to current consumption. To this end, suppose that the normal, risk-​free rate of return is 10 percent. A taxpayer earns $100 in Year 1, and can consume what she retains after-​tax either immediately or in a year. Suppose that the tax rate, under either an income tax or a consumption tax, is 40 percent. Finally, suppose that the income tax is levied when the income is earned, while the consumption tax is levied at the time of consumption.25 Consider first the income tax. After paying $40 of tax on her earnings, the taxpayer can consume the remaining $60 right away, in which case she will face no further income tax liability. Or she can save it for a year, earning $6 of interest on which she pays a further $2.40 of income tax. A decision to save would therefore increase her total income tax liability, not just nominally but in present value terms as measured up front. With a 10 percent interest rate, $2.40 of income tax that is paid in a year has an up-​front 23  Suppose that I found a new company, make it extremely valuable through my efforts, and then sell it for an enormous capital gain. From an economic standpoint, my profit may be viewed as reflecting labor income that should not be excluded from either an income tax or a consumption tax base. 24  For an overview of the logic that supports this analysis, see Shaviro (2004b: 100–​3). Among the important contributions to the extensive literature that reached this conclusion are Domar and Musgrave (1944); Bankman and Griffith (1992); Kaplow (1994); and Bradford (1996). 25  At least in principle (whether or not feasibly in practice), it should be possible to change this timing method under either system without altering the present value of tax liabilities, and thus the incentive effects, of either system. See Shaviro (2000: 54–​5).

 

Economics of Tax Law    119 present value of $2.18. Accordingly, adding this to the $40 of tax that the taxpayer incurs up-​front on her $100 of earnings, one could view the income tax as imposing a 40 percent tax rate on immediate consumption, and a 42.18 percent tax rate on consumption that is deferred for a year. Now consider the consumption tax. Just as under the income tax, consuming right away would generate $40 of tax liability, permitting $60 to be consumed. Suppose, however, that one instead waits for a year. At a 10 percent interest rate, one now has $110, permitting the consumption of $66 after paying $44 of tax. Given the 10 percent interest rate, this amount has the same present value as $40 paid immediately. Thus, current and future consumption face the same effective tax rate of 40 percent. This point can be illustrated algebraically as follows. Suppose that X is the amount one earns, t is the tax rate, and r is the interest rate. Under a consumption tax, if one consumes immediately, then in the next period one has X (1 − t )(1 + r ). Consuming at the end of the next period leaves one with X (1 + r )((1 − t ). These terms are algebraically equivalent. Adding more years of deferral to the analysis does nothing to change the bottom-​line result. The effective income tax rate on deferred consumption continues to increase, as one adds further tax payments on continued saving. By contrast, the effective consumption tax rate remains the same no matter how long the taxpayer waits, since the amount that will ultimately be due grows at the discount rate. To regard this as a pertinent description of how a consumption tax would actually bear on, say, Bill Gates or Mark Zuckerberg, one has to be willing to apply this reasoning to indefinite deferral. We don’t know when (or perhaps even if) extremely rich people and their heirs will actually consume their huge fortunes. Yet the deferred tax burden does indeed keep growing at a market interest rate. Moreover, the deferred tax burden, even as it remains unpaid, does indeed affect the purchasing power of the wealth (which is, after all, what defines its value). Accordingly, under a standard view of economic incidence, the consumption tax—​so long as it will permanently remain in place—​is indeed currently borne by wealth-​holders, even if actual tax payments remain indefinitely deferred. However, not everyone finds this reasoning entirely satisfactory—​whether because they feel a need for a finite “due date,” or based on the option value that savers may have to wait for a tax rate reduction. Accordingly, the fact that consumption tax payments, under a standard method of implementation, may remain indefinitely deferred stands, for many people, as a reason for preferring income taxation. With or without such concerns, however, the fact that an income tax, unlike a consumption tax, discourages saving relative to immediate consumption has important efficiency implications under a tax neutrality framework. Given that an income tax effectively imposes higher tax rates on later-​consumed than earlier-​consumed items, the principle that, in general, one should tax all consumer goods uniformly provides direct support for preferring consumption taxation to income taxation (Bankman and Weisbach, 2006). Other considerations could complicate the analysis, however. For example, suppose that an income tax is likely to be more progressive in practice, even though, in

 

120   Daniel Shaviro principle, a consumption tax could match it in progressivity if given more steeply graduated rates.26 One who preferred greater progressivity might count this as an argument in favor of the income tax. In addition, purely from an efficiency standpoint, suppose that savings behavior is a signal or tag, associated high ability that cannot otherwise be observed (Saez, 2002: 228). There also is literature suggesting that a positive tax on savings may be desirable because, “with high savings, one can more easily afford to under-​ utilize one’s earning ability by working less … [thereby] impos[ing] a negative revenue externality” (Shaviro, 2007, discussing the “new dynamic public finance literature,” such as Golosov et al., 2003). A final set of issues raised by the choice between income taxation and consumption taxation is administrative. In the words of William Andrews (1983: 280), a realization requirement may be the “Achilles’ heel” of any practically feasible income tax, including for the reasons that were briefly discussed in section 6.3.4 above. Consumption taxes can avoid the problems associated with realization, albeit facing various administrative, compliance, and enforcement challenges of their own.27

6.5  Entity-​L evel Taxation of Corporations Among the most consequential implications of having a realization-​based income tax pertains to income earned by corporations. Obviously, in the case of a pure Haig-​Simons income tax, all income would be taxed directly to individuals, and it would make no difference whether a corporation was used to conduct particular business operations. However, once we have a realization-​based income tax system, the following considerations may support treating corporations as distinct taxpayers: 1. Corporations often have complex business operations, with management and record-​keeping being centralized at the entity level, and with public reporting requirements that result in the preparation of company-​level measures of financial accounting income. Under these circumstances, entity-​level income taxation has large administrative advantages. 2. In principle, one could require corporations merely to report their income, with the entire amount being “flowed through” for tax purposes to individuals such as shareholders. (This is indeed how partnerships generally are taxed, in the United States and elsewhere.) However, several factors make this difficult to accomplish 26 

Since high earners tend to save a larger percentage of their earnings than low earners, shifting from an income tax to a consumption tax with similar rate graduation would reduce the system’s progressivity if all else (including the degree of rate graduation) remained the same. 27  On the administrative, compliance, and enforcement issues posed by the design of a progressive consumption tax, see, e.g., Weisbach (2009).

 

Economics of Tax Law    121 in a satisfying way. First, for companies with complex capital structures, it may be hard to divide the income between individuals in a reasonably accurate and not overly manipulable fashion. Second, financial interests that people hold outside of the companies’ internal capital structures—​for example, options and other derivative financial positions—​may add to these difficulties. Third, to the extent that a corporation’s taxable income differs from its economic income, the question of whose income it is lacks any well-​defined economic answer (Shaviro, 2014: 181). Fourth, there may be political objections to requiring individuals to pay current taxes with respect to entity-​level income that they cannot directly access absent control of the company. 3. While one could respond to these problems by imposing no tax on corporate income until it is directly realized by shareholders (such as through the receipt of dividends), this would allow corporate investment to enjoy much broader income tax deferral than that which is generally available to prospective shareholders. Taxing corporate income at the entity level almost inevitably has anomalous tax planning effects, such as encouraging or discouraging the use of corporate entities depending on whether the use of a corporation would lower or raise the applicable marginal tax rate. The resulting efficiency problems are exacerbated by longstanding features of “classical” corporate tax systems in the United States and elsewhere. In particular, such systems distinguish between financial instruments that are classified as corporate debt and equity, respectively. Corporate interest payments or accruals on instruments classified as debt are deductible, but dividends paid on equity are not.28 This can both lead to double taxation of equity-​financed corporate income and to discouraging dividend payouts.29 Entity-​level corporate income taxation also has important effects on the taxation of cross-​border investment and capital flows. If all income earned through the use of corporations were taxed directly to the ultimate individual owners, it is plausible that countries would respond by taxing all outbound investment. After all, if income is being used as a proxy for ability or ability to pay, then presumably one should not care where a given individual’s earnings arose. It also is plausible that, absent the need for entity-​level collection of the corporate income tax, inbound capital investment frequently would be exempted from source country taxation, given that inbound investors generally would not bear the incidence of a source-​based tax if they could respond by investing elsewhere and getting the same after-​tax return (see Shaviro 2009: 67–​8). However, with corporate taxpayers’ residence being determined at the entity level but in the face of cross-​border shareholding, the available policy choices grow considerably 28  The US tax rate imposed on the receipt of a dividend generally is 20 percent or lower, whereas the top rate for individuals’ ordinary income is 39.6 percent. 29  Double taxation does not occur if the corporation can avoid entity-​level taxes. In addition, if dividend taxation at a fixed rate is inevitable at some point, then, as the “new view” of dividend taxation makes clear, making payouts sooner rather than later is not discouraged by the tax. See Shaviro (2009: 73–​7).

 

122   Daniel Shaviro murkier. Resident individuals’ foreign source income cannot necessarily be taxed currently merely by imposing worldwide taxation on resident entities, since such individuals can invest abroad through foreign entities. In addition, source-​based taxation of foreign entities is made costlier to forgo by the concern that resident individuals could hide behind such entities when earning income at home. All this contributes to a state of affairs in which (1) source-​based corporate income taxation is nearly universal (other than in small countries that specialize in serving as tax havens), (2) the proper treatment of outbound investment by resident corporations is hotly contested, and (3) large multinational companies have opportunities to engage in large-​scale tax avoidance that imposes high distortionary and tax planning costs relative to the taxes that they end up paying (see generally Shaviro, 2014).

6.6 Tax Reform Few topics in Washington politics are more perennially popular in the abstract than fundamental or large-​scale tax reform. While in principle “tax reform” could simply mean whatever changes to the tax system one happens to like, for several decades it has had a relatively fixed meaning in tax policy discourse. It has tended to denote something along the lines of the Tax Reform Act of 1986, in which Congress enacted an approximately budget-​neutral (within the officially estimated period) and distributionally neutral package that combined base-​broadening with rate reductions. In recent years, the idea of attempting an at least modified 1986 replay has periodically attracted significant political attention. The big differences from 1986 are that some proposals would either (a) follow the 1986 script just for corporations, in response to other countries having lowered their corporate rates, or else (b) include converting the income tax into a consumption tax, or alternatively scaling it back and adding a VAT to replace the lost revenue. Enthusiasm for 1986-​style tax reform often wanes once people learn more about the tax benefits that would have to be repealed in order to finance significant rate cuts. In the United States, for example, base-​broadening for individuals might have to involve the repeal or curtailment of such popular items as home mortgage interest deductions, charitable deductions, and the exclusion for employer-​provided health insurance. Base-​ broadening to pay for corporate rate cuts would place favorable tax rules for new investment on the chopping block, as well as raising awkward issues concerning the relative taxation of corporate and non-​corporate businesses. In part, the enduring conceptual appeal of 1986-​style tax reform reflects aspirations concerning the achievability of political compromise. For example, distributional neutrality as between the pre-​and post-​reform tax code can facilitate legislative cooperation between those who disagree about distributional issues but have other areas of agreement about tax policy. Likewise, the enactment of a budget-​neutral reform package at least ostensibly permits one to avoid enmeshing the tax reform debate with that

 

Economics of Tax Law    123 concerning deficit reduction. Yet an overlap between the two debates is hard to avoid if base-​broadening is being used to fund lower rates, rather than for some other purpose, such as reducing the need for future changes on the spending side of the federal budget. The main efficiency advantage of 1986-​style tax reform arises from base-​broadening, in cases where this promotes neutrality with respect to rival consumption or investment choices that previously were taxed disparately. For example, repealing home mortgage interest deductions may reduce the income tax system’s encouragement of home ownership relative to both rental occupancy and non-​home consumption. However, the effect of cutting marginal rates in the context of a 1986-​style reform is more complicated. Rate reduction that was adopted as a stand-​alone (and revenue-​losing) change generally would be expected to reduce the deadweight loss resulting from the operation of the tax system. However, if this change is accompanied by base-​broadening, such that overall income tax burdens remain about the same as previously, then the tax system’s overall discouragement of work and saving should also remain about the same.30 While 1986-​style tax reform treats base-​broadening and rate reduction as complements (based on assuming a fixed budgetary goal), there is a sense in which base-​broadening and rate increases are the true complements. In general, the less avoidable a given tax instrument is, the less deadweight loss one would expect it to generate per dollar of revenue raised. At the limit, a lump-​sum tax, by reason of being entirely unavoidable, generates zero deadweight loss. The empirical literature concerning responses to income taxation suggests that labor supply tends to be less responsive to the tax rate, at least in the short run, than various other margins that reflect tax planning effort. Thus, base-​broadening has the potential to reduce the efficiency cost of higher rates, despite its potentially adverse impact on labor supply. For example, the 70 percent top rate that Diamond and Saez (2011) advocate is more likely to be at or below the revenue-​maximizing rate level if the tax base is otherwise broad than if it is narrow.

6.7 Conclusion Within law and economics, tax law is relatively unusual in having distributional aims play a central role, alongside efficiency. Indeed, there is prominent literature asserting, although also contesting, the proposition that distributional issues should be left entirely to the tax and transfer system, instead of having any influence on policy in other 30 

Suppose, e.g., that a revenue-​neutral reform cuts rates and repeals the home mortgage interest deduction, on balance increasing homeowners’ tax burdens while reducing those of non-​homeowners. The latter group has stronger work incentives than previously, but the former group may have weaker such incentives if, at the margin, they were earning money to finance home as well as non-​home consumption.

 

124   Daniel Shaviro areas of law.31 One consequence of this greater scope is that, at least within a conventional economics framework, tax policy issues may generally be both more complex and harder to resolve satisfyingly than those in many other areas of law. After all, tax requires both trading off efficiency concerns against distributional ones and evaluating efficiency in a very complex second-​best framework. To those who like to analyze issues that are relatively straightforward and have clear-​ cut answers, this greater complexity and indeterminacy may make tax policy an area to avoid. To others, they are part of its core appeal.

Acknowledgments I am grateful to Alan Auerbach and Mitchell Kane for their comments on an earlier draft of this chapter.

References Andrews, W. (1974) “A Consumption-​Type or Cash-​Flow Personal Income Tax,” Harvard Law Review, 87: 1113–​88. Andrews, W. (1983) “The Achilles’ Heel of the Comprehensive Income Tax,” in C. Walker and M. Bloomfield (eds) New Directions in Federal Tax Policy for the 1980s. Pensacola, FL: Ballinger Publishing, 278–​85. Atkinson, A. B. and Stiglitz, J. E. (1976) “The Design of Tax Structure: Direct Versus Indirect Taxation,” Journal of Public Economics, 6: 55–​75. Auerbach, A. (2012) “The Mirrlees Review:  A  U.S. Perspective,” National Tax Journal, 65: 685–​708. Bankman, J. and Griffith, T. (1992) “Is the Debate between an Income Tax and a Consumption Tax a Debate about Risk? Does It Matter?” Tax Law Review, 47: 377–​433. Bankman, J. and Weisbach, D. (2006) “The Superiority of an Ideal Consumption Tax over an Ideal Income Tax,” Stanford Law Review, 58: 1413–​56. Bradford, D. (1986) Untangling the Income Tax. Cambridge, MA: Harvard University Press. Bradford, D. (1996) “Consumption Taxes: Some Fundamental Transition Issues,” in M. Boskin (ed.) Frontiers of Tax Reform. Stanford, CA: Hoover Institution Press, 123–​50. Bradford, D. (2003) The X-​Tax in the World Economy. Princeton, NJ: Princeton University Press. Carroll, R. and Viard, A. (2012) Progressive Consumption Taxation:  The X Tax Revisited. Washington: AEI Press. Corlett, W. J. and Hague, D. C. (1953) “Complementarity and the Excess Burden of Taxation,” Review of Economic Studies, 21: 21–​30. Diamond, P. and Saez, E. (2011) “The Case for a Progressive Tax: From Basic Research to Policy Recommendations,” Journal of Economic Perspectives, 25: 165–​90. 31 

This literature includes, for example, Kaplow and Shavell (1994) (asserting that the legal system is less efficient than the tax system at redistributing income) and Sanchirico (2000) (contesting this proposition).

 

Economics of Tax Law    125 Domar, E. and Musgrave, R. (1944) “Proportional Income Taxation and Risk-​ Taking,” Quarterly Journal of Economics, 58: 388–​422. Edgeworth, F. Y. (1897) “The Pure Theory of Taxation: III,” Economic Journal, 7: 550–​7 1. Golosov, M., Kocherlakota, N., and Tsyvinski, A. (2001) “Optimal Indirect and Direct Capital Taxation,” Review of Economic Studies, 70: 569–​87. Gruber, J. (2011) Public Finance and Public Policy (3rd ed.). New York: Worth Publishers. Hall, R. and Rabushka, A. (1995) The Flat Tax. Stanford, CA: Hoover Institution Press. Kaplow, L. (1994) “Taxation and Risk Taking: A General Equilibrium Perspective,” National Tax Journal, 47: 789–​98. Kaplow, L. (2008) The Theory of Taxation and Public Economics. Princeton, NJ:  Princeton University Press. Kaplow, L. and Shavell, S. (1994) “Why the Legal System Is Less Efficient Than the Income Tax in Redistributing Income,” Journal of Legal Studies, 23: 667–​81. Kaplow, L. and Shavell, S. (2002) Fairness versus Welfare. Cambridge, MA:  Harvard University Press. Mankiw, N. G. and Weinzierl, M. (2010) “The Optimal Taxation of Height: A Case Study of Utilitarian Income Redistribution,” American Economic Journal: Economic Policy, 2: 155–​76. Mankiw, N. G., Weinzierl, M., and Yagan, D. (2009) “Optimal Taxation in Theory and Practice,” Journal of Economic Perspectives, 23: 147–​74. Mirrlees, J. (1971) “An Exploration in the Theory of Optimum Income Taxation,” Review of Economic Studies, 38: 175–​208. Murphy. L. and Nagel, T. (2002) The Myth of Ownership: Taxes and Justice. Oxford: Oxford University Press. Ramsey, F. P. (1927) “A Contribution to the Theory of Taxation,” Economic Journal, 37: 47–​61. Rosen, H. and Gayer, T. (2008) Public Finance (8th ed.). New York: McGraw-​Hill. Saez, E. (2002) “The Desirability of Commodity Taxation under Non-​Linear Income Taxation and Heterogeneous Tastes,” Journal of Public Economics, 83: 217–​30. Sanchirico, C. (2000) “Taxes versus Legal Rules as Instruments for Equity: A More Equitable View,” Journal of Legal Studies, 29: 797–​820. Shaviro, D. (2000) When Rules Change: An Economic and Political Analysis of Transition Relief and Retroactivity. Chicago, IL: University of Chicago Press. Shaviro, D. (2004a) “Rethinking Tax Expenditures and Fiscal Language,” Tax Law Review, 57: 187–​231. Shaviro, D. (2004b) “Replacing the Income Tax with a Progressive Consumption Tax,” Tax Notes, 103: 91–​113. Shaviro, D. (2007) “Beyond the Pro-​Consumption Tax Consensus,” Stanford Law Review, 60: 745–​88. Shaviro, D. (2009) Decoding the U.S. Corporate Tax. Washington, DC:  Urban Institute Press. Shaviro, D. (2011) “1986-​Style Tax Reform: A Good Idea Whose Time Has Passed,” Tax Notes, 131: 817–​42. Shaviro, D. (2014) Fixing U.S. International Taxation. New York: Oxford University Press. Simons, H. (1938) Personal Income Taxation. Chicago, IL: University of Chicago Press. Slemrod, J. (1990) “Optimal Taxation and Optimal Tax Systems,” Journal of Economic Perspectives, 4: 157–​78.

 

126   Daniel Shaviro Slemrod, J. and Yitzhaki, S. (2002) “Tax Avoidance, Evasion, and Administration,” in A. Auerbach and M. Feldstein (eds) Handbook of Public Economics, Vol. 3. Amsterdam: Elsevier Science BV, 1423–​70. Weisbach, D. (2009) “Implementing Income and Consumption Taxes,” in A. Auerbach and D. Shaviro (eds) Institutional Foundations of Public Finance. Cambridge, MA:  Harvard University Press, 59–​93.

 

Pa rt  I I

C I V I L P RO C E S S A N D L I T IG AT ION

 

 

Chapter 7

J ud ge-​M ade L aw a nd t h e C om mon L aw Pro c e s s Ben Depoorter and Paul H. Rubin

One of the most illustrious normative claims in the law and economics literature, originating with Posner and supported by models of evolutionary legal change, posits that a system of judge-​made law offers efficiency advantages over statute-​based systems. In recent years, however, scholarship has identified aspects of common law systems that undermine the optimism about judge-​made efficiency. This chapter reviews the original economic literature on the efficiency of the common law and describes supply- and demand-​side obstacles to convergence to efficient legal precedents.

7.1  The Efficiency of the Common Law Hypothesis One of the most well-​known normative claims in law and economics literature is that a system of judge-​made law is superior to legislation in producing efficient rules.1 In his pioneering work Posner analyzed various legal common law doctrines and concluded, on the basis of an intuitive measuring of benefits and costs, that the common law is efficient (Ehrlich and Posner, 1974; 1992; first edition, 1973). These writings helped spark the

1 

The original claim (Hayek, 1960) predates the law and economics movement. For an argument to the contrary, see Tullock (1980, 1997). On the value of a mixture of case law and statute-​based legal rules, see Ponzetto and Fernandez (2008). Empirical work has provided some support for the Hayekian argument that common law systems perform better economically by giving less power to the state than civil law systems. See, e.g., Mahoney (2001); Glaeser and Shleifer (2002); Beck, Demirgüç-​Kunt, and Levine (2003). But see Klerman et al. (2011).

 

130    Ben Depoorter and Paul H. Rubin law and economics movement, inspiring a generation of scholars to apply the economic toolkit to examine doctrines across various areas of law.2 Several foundational writings have further substantiated the efficiency thesis. First, a strand of literature turned to supply-​side aspects of case law, focusing on utility maximizing behavior by judges. Posner (1973) claims that judges derive utility from writing efficient decisions since they are insulated from other motives such as personal interest and interest group pressures3 and do not see enough cases to engage in widespread social distribution. Second, a number of pioneering works identified processes within systems of judge-​ made law that induce efficient outcomes over time. Rubin (1977) posits that inefficient doctrines are under greater evolutionary pressure than efficient precedents. While efficient precedents generate settlements, inefficient decisions create losses to one party that are greater than the gains to the opposing party that cannot be bargained away in the settlement process. For parties that have an ongoing interest in the dispute, the prospect of these deadweight losses in future cases, creates an incentive to challenge the inefficient precedent. As a result, inefficient decisions are subject to greater selection pressure, and subject to litigation more often than efficient decisions. Over time, this selection effect filters out inefficient precedents. Priest (1977) extended the model by Rubin, showing that, by imposing greater costs on the parties than efficient rules, inefficient precedents increase the stakes in a dispute. By increasing the stakes, inefficient precedents are more prone to litigation and increased selection pressure—​even if parties don’t have an ongoing stake in the underlying issue. As inefficient rules continue to be challenged in court, the review process increases the opportunities to discard inefficient precedent in favor of more efficient variants that are less likely to be reviewed. Overall, the legal system perpetuates selection of increasingly efficient legal rules. Goodman (1979) argues that efficient precedents are more valuable to potential beneficiaries than inefficient decisions and hence more likely to be subject to investments in litigation.4 These foundational writings prompted a number of refinements to the efficiency thesis. Landes and Posner (1979) point out that the selection effect pressure is weakened by the basic fact that precedents are not binary and existing precedents might be weakened or strengthened (and sometimes entrenched) rather than overturned.5 Other, more complicated models of litigation present a possibility of multiple equilibria in 2 

For an overview of the field of law and economics, see Bouckaert and De Geest (2000); Newman (1998); Posner and Parisi (1997). 3  This presumed judicial taste for efficiency is questioned by subsequent work on judicial attitudes and voting preferences, see section 7.2.2. 4  See also Katz (1988) with an extension considering the role of litigation expenditures. Stake (1991) distinguishes different areas of law, finding that property law is more likely more susceptible to efficient evolutionary forces than tort law. For a similar argument pointing to contract law’s superiority over tort law, see Barzel (2000). Terrebonne (1981) presents a general model. 5  Heiner (1986) explains stare decisis and other features of civil procedure as mechanisms that minimize the costs of error in human decision-​making.

 

Judge-Made Law and the Common Law Process    131 which efficient and inefficient decisions may co-​exist (Cooter and Kornhauser, 1980; Kornhauser, 1989; Von Wangenheim, 1993). Although a common law system might have a propensity towards efficiency, this outcome is by no means guaranteed. Borrowing from chaos theory, Roe (1996) describes how initial conditions, path dependence, and evolutionary accidents can alter legal evolution. The common law efficiency thesis raised interest in the field of law and economics and helped set a research agenda for years to come (Aranson, 1986). Several decades later however, a rich body of literature challenges the efficiency determinism of the earlier law and economics contributions on judge-​made law. We first turn to supply-​side complications in section 7.2. Section 7.3 provides an overview of demand-​side impediments to efficient judge-​made law.

7.2  Supply-​Side Complications While much of the original literature on the efficiency of the common law focused on the role of litigants in the convergence towards efficiency, in recent years scholars have focused attention on the supply-​side of legal change. A rich body of literature on judicial decision-​making and judicial attitudes casts doubt on the ability as well as the motivations of courts to bring about efficient precedent.

7.2.1 Judicial Attitudes and Preferences In a classic paper, Cooter et al. (1979) presented a model that showed that a movement towards precedent efficiency necessitates that judges deliberately desire efficient outcomes. A vast body of literature has examined judicial preferences during the past three decades. Early contributions by Miceli and Cosgel (1994) and Rasmusen (1994) state that judges are motivated by a combination of individual preferences, on the one hand, and a concern with reversal by higher courts or future judges, on the other hand.6 A growing body of empirical research indicates that political and ideological preferences influence judicial voting patterns (Brudney, Schiavoni, and Merritt, 1999; Sunstein, Schkade, and Ellman, 2004; for an overview, see Epstein et al., 2013). A strong relationship is observed between the ideology of justices and voting behavior in the Supreme Court (Segal and Cover, 1989; Segal and Spaeth, 2002). Empirical evidence also 6  Higgins and Rubin (1980) posit that since promotion is a likely motivation of judges, judges seek to avoid reversal and the value of promotion reduces with age. In their sample, the authors found only a weak relation between reversal and promotion and no relation obtained between age or any other factor and reversal, however. Other early work includes Landes and Posner (1980) on differences between federal and state judges given salary differences and life tenure and Ashenfelter, Eisenberg, and Schwab (1995) on influence on judicial outcomes of the political party appointments.

 

132    Ben Depoorter and Paul H. Rubin confirms a relationship between political attitudes and voting behavior on the federal appellate level (George, 1998; Songer and Davis, 1990; Songer and Haire, 1992; Sunstein, Schkade, and Ellman, 2004). Goldman (1966, 1975) uses political party affiliation and other variables to explain judicial decisions on circuit courts. Panel compositions and political affiliations also help explain voting outcomes, as has been demonstrated by Revesz (1997) and Sunstein, Schkade, and Ellman (2004).7 Carp and Rowland (1983) and Rowland and Carp (1996) provide evidence of the impact of political preferences of trial court judges on judicial outcomes.8 The available evidence suggests that ideological bias is present especially on salient issues, such as civil rights and liberties (Sunstein, Schkade, and Ellman, 2004) and less pronounced in less polarized areas of law such as tax and securities (Schneider, 2001, 2005; Grundfest and Pritchard, 2002).9 Finally, a growing literature in law as well as political science considers forces that may constrain the ability of judges to vote in accordance with their preferences altogether—​ efficiency oriented or not. These external constraints include the preferences and likely actions of politicians (Eskridge, 1991; Bergara, Richman, and Spiller, 2003; Kang and Shepherd, 2015) and of voters in judicial elections (Brace and Boyea, 2008; Huber and Gordon, 2004; Shepherd and Kang, 2014).10

7.2.2 Internal Constraints Even if judges were determined to maximize efficiency in case law, several obstacles remain. First, courts see only a small sample of potential disputes (Hadfield, 1992). Moreover, court proceedings do not provide judges with the type of information necessary to engage in the type of planning that would be necessary to set about efficiency within legal systems (Aranson, 1992; Rizzo, 1980a).11 Second, judges are prone to the systematic errors in judgments that have been highlighted in the literature on psychological and cognitive biases (e.g., Gilovich, Griffin, and Kahneman, 2002; Kahneman, Slovic, and Tversky, 1982). Experimental studies involving federal magistrate judges, find that judges appeared susceptible to many of 7 

See also Hettinger, Lindquist, and Martinek (2006) on ideological positions as an explanation for reversals of lower court rulings and the writing of dissenting opinions. 8  Studies on state courts find lower and sometimes no relationship between judges’ attitudes and outcomes (cf. Pinello, 1999; Gibson, 1978; Nardulli, Flemming, and Eisenstein, 1984). Because measures of judicial attitudes are more difficult for state court judges, partisanship is often used as a proxy for ideology in these studies, however (Brace, Langer, and Hall, 2000). 9  Generally, political bias might have deferring effects across areas of law. 10  Measuring television advertising for state supreme court elections, Shepherd and Kang (2014) observe how judicial election pressures impact voting against appeals by criminal defendants. Kang and Shepherd (2015) find that contributions from political parties are associated with partisan judicial voting. 11  Similarly based on insights from Austrian economics Rizzo (1980b) notes that information shortcomings impede judicially efficient outcomes. Rubin (1980a) notes that the Austrian information cost argument applies to efficiency determination in all economics, not merely law and economics.

 

Judge-Made Law and the Common Law Process    133 the biases observed in lay decision-​makers, including anchoring, framing, hindsight bias, the representativeness heuristics, egocentric biases (Landsman and Rakos, 1994; Guthrie, Rachlinksi, and Wistrich, 2001; Englich et al., 2006), and risk mistreatments (Viscusi, 1999).12 For instance, demands made at pre-​hearing settlement conferences anchored judges’ assessments of the appropriate amount of damages to award in the case (Wistrich, Guthrie, and Rachlinski, 2005). Although the federal and state appellate review system provides an opportunity for correction (Shavell, 1995), these findings increase the risk of judicial error in ways that may impede efficient outcomes.

7.2.3 Structural Variables O’Hara (1993) provides a game theoretical perspective to describe how appellate judges are motivated to follow precedent in order to prevent punishment in the form of ignorance of their decisions by other judges on the bench. Such repeat game interaction provides horizontal adherence to prior decisions even in the absence of formal stare decisis. Zywicki (2003) describes several institutional changes that have rendered the common law more vulnerable to rent-​seeking pressures and inefficient outcomes in recent decades. By contrast, in the old common law system judges and courts competed to supply efficient rules to get the business of disputants.13 Whitman (2000) provides a model of self-​interested decision-​making where judges decide whether to follow the precedent or to announce an alternative rule, “preference satisfaction,” and reap a reputational gain if the next judge upholds their decision or a reputational loss otherwise. Whitman finds that if division of opinion among judges is low relative to the strength of judges’ activist tendencies, a common law system will converge on a single rule. The reason is that even activist judges who disagree with a precedent will follow the precedent anyway, for fear of rejection by subsequent judges. If division of opinion is high however, oscillation between the two rules will take place and legal indeterminacy will undermine the efficiency of the law. Gennaioli and Schleifer (2007a) consider a model of legal change of partial judicial bias in a system of precedent. Even if judges primarily seek efficiency they may possess a bias for one or the other side in a legal dispute that can cause the law to diverge from the efficient path. While policy-​biased judges distort the law away from efficiency, a diversity of judicial views also improves the quality of the law because applying new facts to existing precedent increases the precision of legal rules.14 12  Although some of the results are less pronounced than the findings with regard to laymen and jurors (Chapman and Bornstein, 1996; Hastie, Schkade, and Payne, 1999; Hinsz and Indahl, 1995; Malouff and Schutte, 1989; Raitz, et al., 1990; Hans and Vidmar, 1986; Robbennolt and Studebaker, 1999), the basic finding is that even trained and experienced judges are prone to systematic errors due to cognitive limitation, biases, and heuristics. 13  Klerman (2007) describes how competition among courts infused pro-​plaintiff bias in the common law. 14  In a related paper, Gennaioli and Schleifer posit that the possibility of overruling leads to unstable legal rules that rarely converge to efficiency (2007b).

 

134    Ben Depoorter and Paul H. Rubin Miceli (2009) provides an analysis that considers both the impact of judicial bias as well as the impact of selective litigation. The two main conclusions are that (1) judicial bias can impede the efficient evolution of the law by case selection if the fraction of judges biased against the efficient law is larger than the conditional probability that a case being litigated involves an inefficient law, and (2) precedent affects the rate of legal evolution but not the direction of the law; specifically, weaker precedent speeds the rate while stronger precedent slows it.

7.3  Demand-​Side Complications 7.3.1 Interest Groups Effects Hirshleifer (1982) highlighted that evolutionary forces are shaped not just by the ongoing interests and stakes of respective parties in litigation, but also by the capabilities to organize efficiently and effectively. Judicial outcomes might favor parties that are best equipped to mobilize resources for litigation. Repeat-​play, scale effects, and reduced learning costs, provide an advantage to parties with larger stakes (Galanter, 1974) in common law systems as well as in statute-​or code-​based legal systems. In this regard, the common law’s superior isolation from interest group pressure was only temporary since the common law developed when collective action by interest groups was still expensive (Rubin, 1982).15

7.3.2 Plaintiff Selection Effects Rubin and Bailey (1994) explore the role of tort lawyers as a successful interest group with a stake in the expansion of tort law. Plaintiff lawyers have an ongoing stake in the rise of tort awards and remedies over time, as well as increased uncertainty and litigation costs, in the absence of any potentially offsetting pressure from other stakeholders. Defendants are interested only in the case at hand and face several difficulties (including antitrust) to organize effectively, even when situated in the same industry. Defense attorneys have an agency conflict with their clients since more expansive and costly tort law increases the value of defense lawyers. Consumers are dispersed and have a harder time organizing. Finally, insurance companies have an interest in minimizing liability in the instant but in the long run also gain business when the liability system is more costly (see also in the context of product liability, Viscusi, 1991). Collective action among 15  Extensions include Crew and Twight (1990) (finding common law less subject to rent-​seeking than statute-​law-​based systems) and Rowley and Brough (1987) (finding that contract law and property law are less subject to rent-​seeking in a common law system but not tort law). For a formal model on rent-​ seeking on the law, see Bailey and Rubin (1994).

 

Judge-Made Law and the Common Law Process    135 trial lawyers, including through the Association of Trial Lawyers in America (ATLA), provides information pooling that is conducive to winning favorable outcomes and improving precedent to other lawyers active in the same area of law (Rubin, 2005). Fon and Parisi (2003) examine selective litigation decisions by plaintiffs facing judges with varying ideologies. Their model acknowledges that judges recognize the binding force of prior decisions, but they have different policy perspectives and propensities to expand the scope of legal remedies and cause of action, especially in marginal or borderline cases where a gap in the law or novel ambiguity must be addressed. Plaintiffs make rational estimates of likely case outcome, given the sitting judges and or composition of the court panel. Given the opportunity of plaintiffs to engage in forum-​shopping and to file marginal cases in pro-​plaintiff jurisdictions, liberal judges more frequently get the opportunity to create pro-​plaintiff precedents than conservatives have opportunities to set pro-​defendant precedents. In this process, marginal cases that are unlikely to find support in current case law are unlikely to be brought before a conservative court. Once adjudicated, this pro-​plaintiff biased flow of precedents gradually becomes an interpretative benchmark for all future judges. So, even if parties have symmetric interests in future similar cases, and litigation is exclusively driven by the attempt to maximize returns from the case, case selection might create a strong bias toward pro-​plaintiff evolution of the path of the law. In an extension, Luppi and Parisi (2010) observe that this plaintiff selection effect is muted somewhat in legal systems that anonymize judicial decisions. Conversely, plaintiffs are better positioned to make out the ideologies of judges when judges individually sign majority decisions, concurrences, and dissents. For this follows that, all else held equal, the greater anonymity in continental legal systems make these systems less vulnerable to plaintiff selection effects than common law systems. Depoorter (2011) provides a twist on plaintiff selection effects, predicting that under certain circumstances, plaintiffs with an ongoing interest in the issue will strategically pursue losing litigation in order to obtain the potentially countervailing benefits of public and political mobilization. The precedent costs of a loss in trial might be offset by the expected benefits of mobilizing public and political support. Especially when carefully selecting sympathetic plaintiffs, media attention to the losing party, the underlying cause, and the adverse judicial precedent may create pressure on legislative bodies to change the legal status quo.

7.3.3 Information Selection Effects Hylton (2006) provides a general model that shows how judge-​made law can trend over time in favor of litigants with meritorious cases that possess superior information (e.g., Bebchuk, 1984), as relating to the prediction of the case outcome. Since the Priest–​Klein (1984) model predicts that (1) a dispute is more likely to go to judgment if the standard’s application to the particular practice is highly uncertain, (2) the rational components of the litigant’s predictions are identical, plaintiffs and defendants

 

136    Ben Depoorter and Paul H. Rubin are equally likely to win (litigation likelihood ratio = 1) and the path of the law is not biased (Priest, 1980). Hylton points out, however, that when there is asymmetric information, the relative frequency of litigation favors more informed parties with meritorious case. While informed and non-​meritorious disputes (for instance, guilty physicians facing medical malpractice claims) tend to settle, informed and meritorious litigants litigate to judgment. In the presence of asymmetric information then, the path of the law is disproportionally shaped by the information provided by innocent litigants in court. The disproportionate presence of innocent defendants influences the path of the law by the “litigation likelihood ratio” of innocent to guilty litigants but Hylton posits also that case law becomes “informationally biased.” A legal standard will move from ambiguity to higher degrees of precision on the basis of information provided by innocent defendants; for instance by identifying specific types of non-​ negligent conduct. Over time, this informational bias could lead to “rule evolution” as the information embodied in legal rules alters the nature of the rule itself. Although Hylton’s model does not suggest an unambiguous trend toward efficient judge-​made law, the information-​biasing process aligns well with the common law efficiency hypothesis since informed and meritorious parties are more likely to produce efficient outcomes.

7.3.4 Settlement Selection Effects Depoorter (2010) highlights the influence of civil settlements on legal change. Given the impact of non-​legal factors (ability to bear litigation costs, diverging risk attitudes, asymmetric information, the principal–​agent issues, cognitive limitations, headline sensitivity) on negotiated outcomes,16 settlement trends diverge from available legal remedies. Settlement trends are biased since information on novel settlements featuring high awards and novel remedies circulate more widely in specialized reporters, professional interest organizations, mass media coverage and the oral culture in legal communities. Such novel civil settlements may have a one-​directional ratchet effect on the path of the law as individual settlement concessions make it more difficult for similarly situated defendants to deflect future claims. Ambitious trial lawyers use innovative settlements as benchmarks, making plaintiffs in future cases more reluctant to accept settlements below what others have agreed to in prior, analogous cases. Also, the non-​coercive nature of settlements may frame the normative outlook on novel legal claims and they are less likely to be perceived as outrageous by judges and juries if such claims have been gratified by a prior settlement concession. Overall, prior settlement concessions may thus create sustained pressure towards higher awards and novel remedies, resulting in a gradual expansion of the legal system.

16 

See, e.g., Mnookin and Kornhauser (1979).

 

Judge-Made Law and the Common Law Process    137

7.3.5 Procedural Factors Fon, Parisi, and Depoorter (2005) examine the effect of judicial path dependence on the consolidation and contraction of liability rules and legal remedies when judicial decisions do not become a source of law until they mature into a prevailing line of precedents. In such a system the increase in the scope of remedies does not necessitate a selection effect. Instead, asymmetric stakes and judicial path dependence are sufficient to lead to gradual consolidation or contraction of legal remedies. Due to asymmetric stakes in litigation cases can be rationally pursued also when the probability of success is fairly small. In this process, a large number of negative precedents that deny a new cause of action or restrict the scope of an existing remedy may come about. For instance, in this dynamic process of case selection, an increase in the win/​loss ratio and a decrease in litigation costs, increase the scope for gradual contraction of remedies, since they render smaller probability cases worthy of pursuit. Conversely, in other instances an initial judicial innovation may be followed by a small fraction of early favorable decisions could lead to wider acceptance and eventually consolidate into a binding doctrine. Luppi and Parisi (2012) examine the differential effect of the American and the British or Continental allocation of litigation costs on the path of judge-​made law. Under the American rule, where each party is responsible for his or her own legal expenses, some cases with small probability of success are filed that would not be filed under the British rule where the loser pays part of the costs of the winner (Shavell, 1982). Over time the American rule’s low probability cases may accumulate negative precedents, decreasing the likelihood of success of similar cases in the future.

7.4 Conclusion One of the most illustrious normative claims in the law and economics literature, originating with Posner and supported by models of evolutionary legal change, posits that a system of judge-​made law offers efficiency advantages over statute-​based systems. This strand of scholarship helped spark the law and economics movement, inspiring a generation of scholars to apply the economic toolkit to examine doctrines across various areas of law. In recent years, scholarship has greatly expanded and refined economic models of legal evolution. The overview of the literature in this chapter illustrates that these advances undermine some of the initial optimism about judge-​made efficiency.

References Aranson, P. H. (1992) “The Common Law as Central Economic Planning,” Constitutional Political Economy, 3: 281–​319.

 

138    Ben Depoorter and Paul H. Rubin Ashenfelter, O., Eisenberg, T., and Schwab, S. (1995) “Politics and the Judiciary: The Influence of Judicial Background on Case Outcomes,” Journal of Legal Studies, 24: 257–​81. Bailey, M. and Rubin, P. (1994) “A Positive Theory of Legal Change,” International Review of Law and Economics, 14: 467–​77. Barzel, Y. (2000) “Dispute and its Resolution: Delineating the Economic Role of the Common Law,” American Law and Economics Review, 2: 238–​58. Bebchuk, L. A. (1984) “Litigation and Settlement under Imperfect Information,” RAND Journal of Economics, 15: 404–​15. Beck, T., Demirgüç-​Kunt, A., and Levine, R. (2003) “Law and Finance: Why Does Legal Origin Matter?” Journal of Comparative Economics, 31: 653–​75. Bergara, M., Richman, B. D., and Spiller, P. T. (2003) “Modeling Supreme Court Strategic Decision Making: The Congressional Constraint,” Legislative Studies Quarterly, 28: 247–​80. Bouckaert, B. and De Geest, G. eds (2000) Encyclopedia of Law and Economics. Cheltenham: Edward Elgar. Brace, P. and Boyea, B. D. (2008) “State Public Opinion, the Death Penalty, and the Practice of Electing Judges,” American Journal of Political Science, 52: 360–​72. Brace, P., Langer, L., and Hall, M. G. (2000) “Measuring the Preferences of State Supreme Court Judges,” Journal of Politics, 62: 387–​413. Brudney, J. J., Schiavoni, S., and Merritt, D. J. (1999) “Judicial Hostility toward Labor Unions? Applying the Social Background Model to a Celebrated Concern,” Ohio State Law Journal 60: 1675–​765. Carp, R. A. and Rowland, C. K. (1983) “Policymaking and Politics in the Federal District Courts,” Duke Law Journal, 33: 621–​4. Chapman, G. B. and Bornstein, B. H. (1996) “The More You Ask for, the More You Get: Anchoring in Personal Injury Verdicts,” Applied Cognitive Psychology, 10: 519–​40. Cooter, R. D. and Kornhauser, L. A. (1980) “Can Litigation Improve the Law without the Help of Judges?” Journal of Legal Studies, 9: 139–​63. Cooter, R. D., Kornhauser, L. A., and Lane, D. (1979) “Diability Rules, hoo! Information, and the Role of Precedent,” Bell Journal of Economics, 10: 366–​73. Crew, M. A. and Twight, C. (1990) “On the Efficiency of Law: A Public Choice Perspective,” Public Choice, 66 (1): 15–​36. Depoorter, B. (2010) “Law in the Shadow of Bargaining: The Feedback Effect of Civil Settlements,” Cornell Law Review, 95: 957–​87. Depoorter, B. (2011) “The Upside of Losing,” Columbia Law Review, 113: 817–​62. Ehrlich, I. and Posner, R. A. (1974) “An Economic Analysis of Legal Rulemaking,” Journal of Legal Studies, 3: 257–​86. Englich, B., Mussweiler, T., and Strack, F. (2006) Playing Dice with Criminal Sentences: The Influence of Irrelevant Anchors on Experts’ Judicial Decision Making,” Personality and Social Psychology Bulletin, 32: 188–​200. Epstein, L., Martin, A. D., Quinn, K. M., and Segal, J. A. (2013) “Ideology and the Study of Judicial Behavior,” in J. Hanson and J. Jost (eds) Ideology, Psychology, and Law. Oxford: Oxford University Press, 705–​28. Eskridge, W. N., Jr. (1991) “Overriding Supreme Court Statutory Interpretation Decisions,” Yale Law Journal, 101: 331–​417. Fon, V. and Parisi, F. (2003) “Litigation and the Evolution of Legal Remedies: A Dynamic Model,” Public Choice, 116: 419–​33.

 

Judge-Made Law and the Common Law Process    139 Fon, V., Parisi, F., and Depoorter, B. (2005) “Litigation, Judicial Path Dependence, and Legal Change,” European Journal of Law and Economics, 20: 43–​56. Galanter, M. (1974) “Why the ‘Haves’ Come out Ahead: Speculation on the Limits of Legal Change,” Law and Society Review, 9: 95–​160. Gennaioli, N. and Shleifer, A. (2007a) “The Evolution of the Common Law,” Journal of Political Economy, 115: 43–​68. Gennaioli, N. and Shleifer, A. (2007b) “Overruling and the Instability of Law,” Journal of Comparative Economics, 35: 309–​28. George, T. E. (1998) “Developing a Positive Theory of Decisionmaking on U.S. Courts of Appeals,” Ohio State Law Journal, 58: 1635–​96. Gibson, J. L. (1978) “Judges’ Role Orientations, Attitudes, and Decisions:  An Interactive Model,” American Political Science Review, 72: 911–​24. Gilovich, T., Griffin, D., and Kahneman, D. eds (2002) Heuristics and Biases: The Psychology of Intuitive Judgement. Cambridge: Cambridge University Press. Glaeser, E. L. and Shleifer, A. (2002) “Legal Origins,” Quarterly Journal of Economics, 117: 1193–​229. Goldman, S. (1966) “Voting Behavior on the United States Courts of Appeals, 1961–​1964,” American Political Science Review, 60: 374–​83. Goldman, S. (1975) “Voting Behavior on the United States Courts of Appeals Revisited,” American Political Science Review, 69: 491–​506. Goodman, J. (1979) “An Economic Theory of the Evolution of the Common Law,” Journal of Legal Studies, 7: 393–​406. Grundfest, J. A. and Pritchard, A. C. (2002) “Statutes with Multiple Personality Disorders: The Value of Ambiguity in Statutory Interpretation and Design,” Stanford Law Review, 54: 627–​736. Guthrie, C., Rachlinski, J. J., and Wistrich, A. J. (2001) “Inside the Judicial Mind,” Cornell Law Review, 86: 778–​830. Guthrie, C., Rachlinski, J. J., and Wistrich, A. J. (2007) “Blinking on the Bench: How Judges Decide Cases,” Cornell Law Review, 93: 1–​44. Hadfield, G. (1992) “Biases in the Evolution of Legal Rules,” Georgetown Law Journal, 80: 583–​616. Hans, V. P. and Vidmar, N. (1986), Judging the Jury. New York: Perseus Books. Hastie, R., Schkade, D. A., and Payne, J. W. (1999) “Juror Judgments in Civil Cases: Effects of Plaintiff ’s Requests and Plaintiff ’s Identity on Punitive Damage Awards,” Law and Human Behavior, 23: 445–​70. Hathaway, O. A. (2001) “Path Dependence in the Law: The Course and Pattern of Legal Change in a Common Law System,” Iowa Law Review, 86: 601–​65. Hayek, F. (1960) The Constitution of Liberty. Chicago, IL: University of Chicago Press. Heiner, R. (1986) “Imperfect Decisions and the Law: On the Evolution of Legal Precedent and Rules,” Journal of Legal Studies, 15: 227–​61. Hettinger, V. A., Lindquist, S. A., and Martinek, W. L. (2006) Judging on a Collegial Court:  Influences on Appellate Decision Making. Charlottesville, VA:  University of Virginia Press. Higgins, R. S. and Rubin P. H. (1980) “Judicial Discretion,” Journal of Legal Studies, 9: 129–​38. Hinsz, V. B. and Indahl, K. E. (1995) “Assimilation to Anchors for Damage Awards in a Mock Civil Trial,” Journal of Applied Social Psychology, 25: 991–​1026.

 

140    Ben Depoorter and Paul H. Rubin Hirshleifer, J. (1982) “Evolutionary Models in Economics and Law,” in P. Rubin and R. Zerbe (eds) Research in Law and Economics, Vol. 4. Greenwich, CT: JAI Press, 1–​60. Huber, G. A. and Gordon, S. C. (2004) “Accountability and Coercion: Is Justice Blind When It Runs for Office?” American Journal of Political Science, 48: 247–​63. Hylton, K. (2006) “Information, Litigation, and Common Law Evolution,” American Law and Economics Review, 8: 33–​61. Kahneman, D., Slovic, P., and Tversky, A. eds (1982) Judgment under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press. Kang, M. and Shepherd, J. (2015) “Partisanship in State Supreme Courts: The Empirical Relationship between Party Campaign Contributions and Judicial Decisionmaking,” Journal of Legal Studies 44: 161–​85. Katz, A. (1988) “Judicial Decisionmaking and Litigation Expenditure,” International Review of Law and Economics, 8: 127–​43. Klerman, D. (2007) “Jurisdictional Competition and the Evolution of the Common Law,” University of Chicago Law Review, 74: 1179–​226. Kornhauser, L. (1989) “An Economic Perspective of Stare Decisis,” Chicago-​Kent Law Review, 65: 63–​92. Landes, W. M. and Posner, R. A. (1979) “Adjudication as a Private Good,” Journal of Legal Studies, 8: 235–​84. Landes, W. M. and Posner, R. A. (1980) “Legal Change, Judicial Behavior, and the Diversity Jurisdiction,” Journal of Legal Studies, 9: 367–​86. Landsman, S. and Rakos, R. F. (1994) “A Preliminary Inquiry into the Effect of Potentially Biasing Information on Judges and Jurors in Civil Litigation,” Behavioral Sciences and Law, 12: 113–​26. Luppi, B. and Parisi, F. (2010) “Judicial Creativity and Judicial Errors:  An Organizational Perspective,” Journal of Institutional Economics, 6: 91–​8. Luppi, B. and Parisi, F. (2012) “Litigation and Legal Evolution: Does Procedure Matter?” Public Choice, 152: 181–​201. Malouff, J. and Schutte, N. S. (1989) “Shaping Juror Attitudes: Effects of Requesting Different Damage Amounts in Personal Injury Trials,” Journal of Social Psychology, 129: 491–​7. Miceli, T. (2009) “Legal Change: Selective Litigation, Judicial Bias, and Precedent,” Journal of Legal Studies, 38: 157–​68. Miceli, T. and Cosgel, M. (1994) “Reputation and Judicial Decision-​Making,” Journal of Economic Behavior and Organization, 23: 31–​51. Mnookin, R. H. and Kornhauser, L. (1979) “Bargaining in the Shadow of the Law: The Case of Divorce,” Yale Law Journal, 88: 950–​97. Nardulli, P. F., Flemming, R. B., and Eisenstein, J. (1984) “Unraveling the Complexities of Decision-​ Making in Face-​ to-​ Face Groups:  A  Contextual Analysis of Plea-​ Bargained Sentences,” American Political Science Review, 78: 912–​28. Newman, P. ed. (1998) The New Palgrave Dictionary of Economics and the Law. Basingstoke: Palgrave Macmillan. O’Hara, E. (1993) “Social Constraint or Implicit Collusion? Toward a Game Theoretic Analysis of Stare Decisis,” Seton Hall Law Review, 24: 736–​78. Parisi, F. and Posner R. A. (1997), Law and Economics, Vol. 3. Cheltenham: Edward Elgar. Pinello, D. R. (1999) “Linking Party to Judicial Ideology in American Courts: A Meta-​analysis,” Justice System Journal, 20: 219–​54.

 

Judge-Made Law and the Common Law Process    141 Ponzetto, G. A. M. and Fernandez, P. A. (2008) “Case Law versus Statute Law: An Evolutionary Comparison,” Journal of Legal Studies, 37: 379–​430. Posner, R. A. (1973) Economic Analysis of Law (1st ed.). Boston: Little, Brown. Posner, R. A. (2003) Economic Analysis of Law (5th ed.). New York: Aspen. Priest, G. (1977) “The Common Law Process and the Selection of Efficient Rules,” Journal of Legal Studies, 6: 65–​82. Priest, G. (1980) “Selective Characteristics of Litigation,” Journal of Legal Studies, 9: 399–​421. Priest, G. and Klein, B. (1984) “The Selection of Disputes for Litigation,” Journal of Legal Studies, 13: 1–​56. Raitz, A., Greene, E., Goodman, J., and Loftus, E. F. (1990) “Determining Damages: The Influence of Expert Testimony on Jurors’ Decision Making,” Law and Human Behavior, 14: 385–​95. Rasmusen, E. (1994) “Judicial Legitimacy as a Repeated Game,” Journal of Law, Economics and Organization, 10: 63–​83. Revesz, R. L. (1997) “Environmental Regulation, Ideology, and the D.C. Circuit,” Virginia Law Review, 83: 1717–​72. Rizzo, M. J. (1980a) “The Mirage of Efficiency,” Hofstra Law Review, 8: 641–​58. Rizzo, M. J. (1980b) “Law amid Flux: The Economics of Negligence and Strict Liability in Tort,” Journal of Legal Studies, 9: 291–​318. Robbennolt, J. K. and Studebaker, C. A. (1999) “Anchoring in the Courtroom: The Effects of Caps on Punitive Damages,” Law and Human Behavior, 23: 353–​73. Roe, M. J. (1996) “Chaos and Evolution in Law and Economics,” Harvard Law Review, 109: 641–​68. Rowland, C. K. and Carp, R. A. (1996) Politics and Judgment in Federal District Courts. Lawrence, KS: University Press of Kansas. Rowley, C. K. and Brough, W. (1987) “The Efficiency of the Common Law: A New Institutional Economics Perspective,” in R. Pethig and U. Schlieper (eds) Efficiency, Institutions, and Economic Policy: Proceedings of a Workshop Held by the Sonderforschungsbereich 5 at the University of Mannheim June 1986. Berlin: Springer, 103–​22. Rubin, P. H. (1977) “Why Is the Common Law Efficient?” Journal of Legal Studies, 6: 51–​63. Rubin, P. H. (1980) “Predictability and the Economic Approach to Law: A Comment on Rizzo,” Journal of Legal Studies, 9: 319–​34. Rubin, P. H. (1982) “Common Law and Statute Law,” Journal of Legal Studies, 11: 205–​23. Rubin, P. H. (2005) “Public Choice and Tort Reform,” Public Choice, 124: 223–​36. Rubin, P. H. and Bailey, M. J. (1994) “The Role of Lawyers in Changing the Law,” Journal of Legal Studies, 23: 807–​31. Schneider, D. M. (2001) “Empirical Research on Judicial Reasoning: Statutory Interpretation in Federal Tax Cases,” New Mexico Law Review, 31: 325–​52. Schneider, D. M. (2005) “Using the Social Background Model to Explain Who Wins Federal Appellate Tax Decisions:  Do Less Traditional Judges Favor the Taxpayer?” Virginia Tax Review, 25: 201–​49. Segal, J. A. and Cover, A. D. (1989) “Ideological Values and the Votes of U.S. Supreme Court Justices,” American Political Science Review, 83: 557–​65. Segal, J. A. and Spaeth, H. J. (2002) The Supreme Court and the Attitudinal Model Revisited. New York: Cambridge University Press. Shavell, S. (1982) “Suit, Settlement, and Trial: A Theoretical Analysis under Alternative Methods for the Allocation of Legal Costs,” Journal of Legal Studies, 11: 55–​81.

 

142    Ben Depoorter and Paul H. Rubin Shavell, S. (1995) “The Appeals Process as a Means of Error Correction,” Journal of Legal Studies, 24: 379–​426. Shepherd, J. and Kang, M. (2014) “Skewed Justice: Citizens United, Television Advertising, and State Supreme Court Justices’ Decisions in Criminal Cases,” American Constitution Society for Law and Policy, Washington, DC. Songer, D. R. and Davis, S. (1990) “The Impact of Party and Region on Voting Decisions in the United States Courts of Appeals, 1955–​1986,” Western Political Quarterly, 43: 317–​34. Songer, D. R. and Haire, S. (1992) “Integrating Alternative Approaches to the Study of Judicial Voting: Obscenity Cases in the U.S. Courts of Appeals,” American Journal of Political Science, 36: 963–​82. Stake, J. E. (1991) “Status and Incentive Aspects of Judicial Decisions,” Georgetown Law Journal, 79: 1447–​97. Sunstein, C. R., Schkade, D., and Ellman, L. M. (2004) “Ideological Voting on Federal Courts of Appeals: A Preliminary Investigation,” Virginia Law Review, 90: 301–​54. Terrebonne, R. (1981) “A Strictly Evolutionary Model of Common Law,” Journal of Legal Studies, 10: 397–​408. Tullock, G. (1997). The Case against the Common Law. Durham, NC: Carolina Academic Press. Tullock, G. (1980) “Efficient Rent Seeking,” in J. Buchanan, R. Tollison, and G. Tullock (eds) Toward a Theory of the Rent-​Seeking Society. College Station, TX: Texas A & M University Press, 97–​112. Viscusi, W. K. (1991) “The Dimensions of the Product Liability Crisis,” Journal of Legal Studies, 20: 147–​77. Viscusi, W. K. (1999) “How Do Judges Think about Risk?” American Law and Economics Review, 1: 26–​62. von Wangenheim, G. (1993) “The Evolution of Judge-​Made Law,” International Review of Law and Economics, 13: 381–​411. Whitman, D. (2000) “Evolution of the Common Law and the Emergence of Compromise,” Journal of Legal Studies, 29: 753–​81. Wistrich, A. J., Guthrie, C., and Rachlinski, J. J. (2005) “Can Judges Ignore Inadmissible Information? The Difficulty of Deliberately Disregarding,” University of Pennsylvania Law Review, 153: 1251–​345. Zywicki, T. (2003) “The Rise and Fall of Efficiency in the Common Law: A Supply-​Side Analysis,” Northwestern Law Review, 97: 1551–​613.

 

Chapter 8

E c onom ics of C i v i l Pro cedu re Robert G. Bone

8.1 Introduction This chapter examines the law and economics approach to civil procedure. A properly designed procedural system is of paramount importance to any well-​functioning legal system. While it is possible for the substantive law to affect primary conduct directly when people internalize norms and comply with them voluntarily, it is probably more common for substantive law to shape behavior through the credible threat of sanction. And the credibility of the threat depends critically on the efficacy of the procedures used for enforcement. Law and economics has much to say about the proper design of a procedural system. From a normative standpoint, the goal, roughly speaking, is to fashion rules that efficiently enforce the substantive law, in the sense of minimizing the social costs associated with enforcement. These costs include the cost of error and the cost of procedures used to reduce error.1 Procedure reduces the risk of error by enabling the discovery of relevant facts and evidence, testing the accuracy of factual and legal propositions, preventing undesirable settlements, and so on. Accuracy-​enhancing procedure, however, creates litigation and other process costs when it is applied (sometimes also referred to as direct or administrative costs). Adding a new procedure is efficient, therefore, if it reduces error costs enough to justify the additional process costs it creates. More generally, a procedural system is optimal from an economic point of view if any change in the existing procedures adds more in the way of process cost than it reduces in the way of error cost. 1  This same goal can be stated in terms of social benefits and costs. The social benefit of a procedure consists in the error-​risk reduction it makes possible and the social cost consists in the cost of applying the procedure when it needs to be applied.

 

144   Robert G. Bone This is the standard economic account of procedure. It makes an important assumption, that the benefit of procedure lies exclusively in reducing the risk of outcome error. This assumption, while useful much of the time, is not entirely accurate. From an economic perspective, the ultimate goal of procedure is to generate efficient compliance with the substantive law—​or even more generally, to generate efficient incentives in the real world. It turns out that procedure can affect incentives in ways other than reducing the risk of outcome error, and a more complete efficiency analysis should take account of these effects as well. Section 8.3.2.3 below elaborates on this idea. One point stands out clearly. To evaluate benefits and costs, one must have some way to predict the effects of a procedural rule on the behavior of litigants and their lawyers. Predicting effects, however, is a very tricky business, especially in the highly strategic environment of adversarial litigation. For example, expanding discovery might increase outcome accuracy by providing additional opportunities for parties to ferret out relevant facts and evidence. But it also might reduce accuracy if one side employs expanded discovery strategically to impose costs on its opponent and pressure a favorable settlement. Prediction is the domain of positive economics, and as we shall see, positive economics offers powerful tools for analyzing this type of strategic interaction. It is important to bear in mind, however, that economic efficiency is not the only metric for evaluating procedure. One might instead focus on rights enforcement or the justice of the error-​risk distribution across cases and litigants (Dworkin, 1985; Bone, 2003). These alternatives pay attention to outcome accuracy, but they value accuracy in moral rather than economic terms.2 The conventional economic account does not include these alternative approaches. In a rights-​based theory, for example, outcome errors implicate infringements of individual right that trump or substantially constrain utilitarian justifications based on social welfare gains.3 Although these alternative metrics differ from efficiency, they still require some way to predict the effects of procedural rules on error risk. This means that positive economics has an important role to play. Indeed, positive economics is valuable for any moral theory that cares about consequences. Far too many mainstream civil procedure scholars miss this important point. The thesis of this chapter is that law and economics offers tools and insights that are relevant, indeed critical, to designing an optimal procedural system no matter what normative metric one applies. Section 8.2 begins the discussion by defining more precisely 2  A third alternative is completely independent of outcome quality. This third approach focuses on individual participation and values participation in deontological terms, for the way it respects the dignity and autonomy of litigants. To be sure, an efficiency metric also values individual participation but for different reasons—​participation counts only if it produces better outcomes or if it generates private utility for litigants on top of whatever utility they gain from a favorable outcome. The deontological theory, by contrast, values participation in terms of dignity without regard to utility gains and conceives of participation as a right that trumps the social cost–​benefit balance (Bone, 2003). 3  Some scholars have argued that an economic analysis can account for the main features of these alternative theories by adjusting the social welfare function (Kaplow and Shavell, 2002). But other scholars strongly disagree.

 

Economics of Civil Procedure    145 what is included in the category of “civil procedure.” Section 8.3 then describes the law and economics approach in more detail and discusses some of its strengths and limitations. Section 8.4 illustrates the approach by applying it to three specific topics: court-​ annexed mediation, pleading, and discovery. Section 8.5 concludes by summarizing the key contributions that law and economics has made to civil procedure in general and outlining directions for future work.

8.2  What Is “Civil Procedure”? A Closer Look 8.2.1 Scope It is common to define civil procedure as the processes used to resolve civil disputes. This is a useful definition, but it also requires further specification. In particular, it is important to be clear what it means to “resolve” disputes and which “processes” count as part of civil procedure.

8.2.1.1 “Resolve” Disputes Resolving a dispute is not the same thing as settling or ending a dispute. Parties often continue disputing after a formal resolution. Furthermore, resolving a dispute is about much more than satisfying the private interests of the disputing parties. The point of civil procedure is to enforce the substantive law and the point of enforcement is to further the social policies that the substantive law serves. This is important because it means that procedural design is a matter of public concern and should not be left entirely to the disputing parties without considering the social consequences. Moreover, resolving a dispute is not just about applying existing law; it is also about making new law. In common law countries, for example, procedure frames the judge’s common-​law-​making function—​such as, by regulating the information that the judge receives and by structuring legal argument—​and as a result it can affect the quality of the common law rules that judges craft (see generally Fuller, 1978). Even in civil law countries, the law evolves to some extent through the influence of previous decisions (Fon and Parisi, 2006: 521–​3). It follows that the design of a procedural system should take account of effects on law-​making as well as on law-​applying.

8.2.1.2 “Processes” Until relatively recently, it was common to associate civil procedure exclusively with formal court adjudication. Today, the category embraces much more. For example, court procedures in the United States, and in many other countries, feature mediation, arbitration, and other forms of alternative dispute resolution (ADR). Also, settlement has

 

146   Robert G. Bone become more salient as judges in the United States and other countries have become more actively involved in settlement promotion. There is a deep normative reason to include ADR and settlement. Both affect the quality of dispute resolution outcomes, which, in turn, affect how the substantive law is enforced. For example, a civil procedure system that makes mediation a mandatory precondition to litigation should pay attention to the quality of the settlements that mandatory mediation produces as well as the quality of trial outcomes. Moreover, when procedural rules designed for adjudication create costs that push parties toward arbitration, rulemakers should consider arbitration outcomes in addition to adjudication outcomes since both affect primary activity incentives. Still, there are limits. A process must be sufficiently related to formal adjudication to count as part of what lawmakers should consider when designing a procedural system for adjudication. This condition holds in our arbitration example above because we assumed a strong causal link between adjudicative procedure and incentives to use arbitration. Also, there are certain structural and technical aspects of court administration that are normally treated separately. For example, civil procedure is not ordinarily concerned with the number of courts or appellate tiers or with the variety of specialized forums. There is no reason in theory to exclude these topics. While court structure and administration involve special considerations such as federalism and comity, they also affect optimal enforcement of the substantive law. Nevertheless, I shall ignore issues of structure and administration in this chapter. Put differently, I shall assume that a court system is already in place and focus on the procedural rules for resolving particular disputes within that system.

8.2.2 Content A system of procedural rules has three main components. First, the system must specify the entity or entities that have the power to make procedural rules. This power might be assigned to the legislature, an expert committee, judges using the common law process, or litigating parties choosing procedures to govern their own disputes. Each option has advantages and disadvantages. The legislature, for instance, is more politically accountable than an expert committee, but legislators are less likely to be attentive to procedural problems and less likely to be knowledgeable about the technicalities of procedural design. The second component involves the form of procedural rules. The more flexible those rules are—​for example, the greater the use of open-​ended standards like reasonableness—​the more latitude judges have to adapt procedures to the needs of particular cases. Thus, the optimal degree of rule specificity and flexibility depends in part on the capacity of trial judges to make good case-​specific procedures. The third and most obvious component consists of the particular rules themselves. For example, there must be rules about which courts or other dispute resolution bodies have the power to resolve which cases, how strict the filing requirements should be, how

 

Economics of Civil Procedure    147 large the litigating units can be, how broad the scope of information exchange should be, what role private settlement should play, and so on. These three components—​the rulemaking entity, the form of rules, and the content of specific rules—​interrelate. Decisions about optimal form affect choices about specific content, and vice versa. So too, the choice of rulemaking authority affects decisions about optimal form. Different procedural systems make different choices in each of these three areas. Civil law countries, for example, rely mostly on the legislature to make procedural rules. Federal courts in the United States rely primarily on expert committees subject to judicial oversight and weak legislative review. Different systems also make different choices about the form and the content of the rules. Until relatively recently, for example, the United States stood out internationally for its relatively lax pleading rules, and it continues to authorize some of the broadest discovery in the world.

8.3  The Analytic Framework Most law and economics work in civil procedure is positive economics, but there is some significant normative work as well. The following discussion describes some of the most important positive and normative insights.

8.3.1 Positive Theory Much of the scholarship applying positive economics (theoretical not empirical) to civil procedure is written for readers well versed in game theory. The models tend to be mathematically sophisticated and the analysis technically complex. Most articles include a description of the intuitions behind their conclusions, but even these descriptions can be difficult for a nontechnical reader to grasp. This high degree of technicality compounds the already negative attitude of mainstream civil procedure scholars toward economics. The result is a large gap between economists doing civil procedure work and mainstream legal academics specializing in civil procedure. This gap is extremely unfortunate. Positive economics has much to offer civil procedure scholarship. One of its main contributions is to correct facile assumptions found in the mainstream civil procedure literature. For example, many civil procedure scholars assume that stricter pleading requirements harm plaintiffs by screening meritorious suits. This is true, but economic models also identify a potential settlement benefit for meritorious plaintiffs. When defendants cannot tell whether suits are meritorious or meritless, they will tend to be reluctant to settle because a settlement might pay off a meritless plaintiff and attract more meritless filings in the future. By screening meritless suits, stricter pleading increases the defendant’s confidence that the plaintiff ’s suit is in fact meritorious, which improves the chance of obtaining a welfare-​enhancing

 

148   Robert G. Bone settlement. More generally, positive work in law and economics highlights the complexity of predicting consequences in a strategic environment that includes both trial and settlement. The following discussion summarizes the basic tools and some of the most important findings.

8.3.1.1 The Basic Model The basic positive model of litigation—​the so-​called Landes-​Posner-​Gould model—​was developed in the early 1970s (Landes, 1971; Posner, 1973; Gould, 1973: 284–​93). The simplest form of this model assumes risk-​neutral parties and loyal attorneys paid on an hourly fee basis. In this model, parties make litigation choices that maximize their expected value. Expected value is a function of the likelihood of success on the merits (P), the expected trial award conditional on success (W), and the expected cost to the party of litigating the case through trial (C). To illustrate, consider a plaintiff ’s original filing decision. If the plaintiff must pay her own attorney win or lose—​as the American Rule requires—​she will file suit if PW − Cπ > 0, where Cπ is the plaintiff ’s expected cost of litigating the case through trial.4 If this condition holds, the suit is known as “positive expected value” (PEV); if not, it is known as “negative expected value” (NEV).5 This simple model is useful not only for predicting filing decisions, but also for predicting other litigation decisions preparatory to trial. And it is also useful for predicting settlement (Posner, 1973: 417–​20). A party will settle only if she does better (or just as well) settling than trying her case. This sets upper and lower bounds to the range of feasible settlements. For the plaintiff to be willing to accept the defendant’s settlement offer, the offer must be greater than or equal to the plaintiff ’s net expected gain from going to trial (calculated at the point that settlement discussions take place). Otherwise, the plaintiff would be better off (in expected value terms) rejecting the offer and trying the case. By the same logic, for the defendant to be willing to accept the plaintiff ’s settlement demand, the demand must be less than or equal to the defendant’s net expected loss from going to trial. Stated mathematically, settlement is feasible if and only if:  PW + C∆ ≥ PW − Cπ , where Cπ is the plaintiff ’s cost and CΔ is the defendant’s cost of litigating the case from that point all the way through trial. This feasibility condition assumes that both sides have complete information about P, W, Cπ, and CΔ, so they agree on the same estimates. I relax this assumption in the following section. A more complete analysis would also include the costs of negotiating a settlement, but it is common in the literature to ignore these costs in order to make the analysis more tractable.6 Doing so usually does not have a significant impact on the results because 4 

This assumes that each side must also pay its non-​fee-​related costs. Although not precisely correct, this is a useful simplifying assumption. 5  If, instead, the losing party must pay the winning party’s fees—​as the British Rule requires—​the condition for a lawsuit to be PEV is: PW − (1 − P )(Cπ + C∆ ) > 0, where CΔ is the plaintiff ’s estimate of the expected cost to the defendant of litigating the case through trial. 6  The anticipated transaction costs of continuing to negotiate can affect the settlement range, since parties know they can save negotiation costs by agreeing to a settlement rather than prolonging negotiations.

 

Economics of Civil Procedure    149 expected litigation costs almost always swamp settlement transaction costs. Ignoring bargaining costs, the mathematical condition for settlement feasibility stated above implies that settlement is feasible in every case because PW + CΔ always exceeds PW –​Cπ. Simply put, parties are better off settling than going to trial because settlement saves the costs of actually litigating the case (Cπ + CΔ). Under these circumstances, the suit can settle for any amount between PW –​Cπ and PW + CΔ. This range of possible settlements is known as the “settlement range.” Moreover, the difference between the maximum that the defendant is willing to pay (PW + CΔ) and the minimum that the plaintiff is willing to accept (PW –​Cπ) is known as the “settlement surplus.” In our simple example, the settlement surplus is Cπ + CΔ. If a settlement range exists—​or, equivalently, if the settlement surplus is not negative—​the parties can settle. Two additional points. First, it is important to bear in mind that PW + C∆ ≥ PW − Cπ is just a feasibility condition. Satisfying it does not necessarily mean that the suit will settle; it only means that the suit can settle. Whether a settlement takes place depends on how the bargaining process unfolds. For example, each side might bargain hard in an effort to capture most of the settlement surplus, and hard bargaining can scuttle settlements even when both sides know they would be better off settling. Second, the feasibility condition does not tell us how much the suit will settle for, assuming that it does settle. The settlement will lie somewhere in the settlement range, but precisely where depends on the relative bargaining power of the parties.7

8.3.1.2 Refinements and Expansions Starting in the 1980s, scholars refined and expanded the simple model to take account of more realistic features of the litigation environment. Four of these developments are particularly noteworthy: (1) a more sophisticated analysis of strategic interaction, (2) the recognition of divergent expectations, (3) the combination of asymmetric information with strategic interaction, and (4) the addition of attorney–​client agency costs. Before discussing each of these, it is important to mention another variation on the basic model that has not received as much attention in the literature. This variation involves the application of real options theory (Grundfest and Huang, 2006; Cornell, 1990). The real options model analogizes a lawsuit to a research and development project. Just as a firm conducts R&D in stages and re-​evaluates its progress based on information obtained at each stage, so too a plaintiff conducts her lawsuit in stages and re-​evaluates as she proceeds. At the completion of each stage, the plaintiff can continue to the next stage or drop the suit, and she exercises this option in light of the information

7  Cooperative game theory offers techniques for predicting the settlement amount. The most well known is the Nash Bargaining Solution. If the plaintiff ’s lawyer is a better negotiator than the defendant’s lawyer, the settlement is likely to end up closer to the high end of the range—​and vice versa. If the parties have relatively equal bargaining power, it is reasonable to suppose that the settlement amount will end up near the midpoint of the range. In the simple model, and assuming the American Rule on fees, this midpoint is PW + (C ∆ − Cπ ) / 2.

 

150   Robert G. Bone she has acquired to that point. The availability of these options increases the value of the suit for the plaintiff because it enables more effective use of information. This additional value is known as option value, and option theory teaches that the option value of a lawsuit increases with its variance. One application of this model helps explain the filing of NEV suits: a lawsuit that is NEV in the simple model can become PEV when its option value is added to its net expected value. While option theory is promising in some respects, it is not entirely clear what a real options model adds to strategic models more generally. A normal game-​theoretic model can include options to drop the suit at various stages without the need for options theory. In any event, this chapter is not the place to engage option theory in any greater detail, so the following discussion examines other refinements.

8.3.1.2.1 Strategic Interaction The simplest version of the basic litigation model assumes that the litigating parties respond to an exogenously fixed environment. To illustrate, consider the problem of predicting the frequency of strike suits in the absence of regulation (strike suits are meritless suits filed by plaintiffs who know their suits lack merit). One way to analyze this problem is to specify the defendant’s response exogenously and then determine whether the plaintiff would file given the fixed response. For example, suppose we assume that the defendant will answer, conduct discovery, and go to trial if the plaintiff files a strike suit. Given this assumption, it is easy to see that rational plaintiffs will choose not to file strike suits. After all, if they do, they will almost certainly lose at trial or drop the suit before getting to trial, in which case they will incur litigation costs without recovering anything (in other words, the suit is NEV unless the expected trial award is unusually large). The problem with this analysis is that the defendant’s response is not fixed in this way. What the defendant will do depends on what the plaintiff will do, and what the plaintiff will do depends on what the plaintiff anticipates the defendant will do in response. In other words, the rules of civil procedure operate in a strategic environment. How strategic the environment is depends, of course, on how much the particular system relies on adversarial interaction. Still, strategy is always important, and the strategic nature of disputing makes a big difference to predictions. In our strike-​suit filing example, the defendant has several ways to respond to the plaintiff ’s decision to sue. The defendant might choose to do nothing, or file an answer and litigate further if necessary, or make a settlement offer. If he does nothing, the defendant risks a possibly large default judgment. If he files an answer, he avoids the default judgment, but still incurs the cost of preparing and filing the answer (at least assuming the American Rule on fees). Thus, he does best of all, if he offers a settlement in an amount less than the cost of answering.8 As a result, the defendant should be willing to settle for any amount up to the cost of preparing and filing an 8 

Unless the defendant is a repeat-​player who can benefit in the future from establishing a reputation for fighting strike suits.

 

Economics of Civil Procedure    151 answer. The plaintiff anticipates this response and files a meritless suit as long as the expected settlement exceeds the cost of filing a complaint (Rosenberg and Shavell, 1985).9 This is a very simple example, but it illustrates the central point. In a strategic environment, it is a mistake to predict litigation choices by assuming a particular response. Since parties make their choices by anticipating what other parties will do, a party’s choice depends on its prediction of the other party’s response—​and vice versa. Game theory provides useful tools for analyzing this type of interaction. The solution is to find a so-​called Nash Equilibrium. A Nash Equilibrium is a set of strategies that are mutual best responses to one another, in the sense that no party has an incentive to deviate from her specified strategy as long as all the other parties stick to theirs.

8.3.1.2.2 Divergent Expectations The simplest version of the basic model assumes that all parties have identical information about the dispute and estimate the same values for P, W, and C. This is an unrealistic assumption. Parties often have different subjective beliefs about one or more of these variables. These differences can be caused by excessive optimism or excessive pessimism. For example, the empirical literature suggests that individuals tend to interpret the same information in self-​serving ways that reinforce subjective perceptions of their own prospects of success (Lowenstein and Moore, 2004). If this is true of litigating parties, those parties might formulate different estimates of P based on a common pool of information. Also, informational asymmetry can cause divergent beliefs. For example, a plaintiff might have private information about the severity of her injuries or defendant might have private information about its liability-​producing actions. Discovery is supposed to cure these asymmetries, but many important litigation decisions are made before discovery. Different subjective beliefs can produce divergent expectations about the likely trial outcome, and these divergent expectations can scuttle settlement. This section focuses on these settlement effects, and the next section discusses the additional problems introduced by strategic interaction. Even the earliest literature describing the basic model recognized that divergent expectations can block settlements (Posner, 1973: 418–​20). To illustrate, assume that the plaintiff, for whatever reason, believes she has a high probability of winning at trial while the defendant believes she has a low probability. Let Pπ be the plaintiff ’s estimate of her 9 

This model illustrates the importance of a strategic analysis, but it has only weak implications for frivolous filings. The reason is easy to see. The amount of any settlement predicted by this model is limited to the cost of answering the complaint, so the cost of answering must exceed the cost of filing for a plaintiff to be willing to file suit. There is no reason, however, to assume that this condition holds for most lawsuits. Preparing and filing an answer is probably no more costly than preparing and filing a complaint—​and might even be less costly. For a generalization of this model that assumes multiple litigation stages, ignores sunk costs, and predicts substantial settlements in NEV suits, see Bebchuk (1996).

 

152   Robert G. Bone own likelihood of success, and let PΔ be the defendant’s (different) estimate of the plaintiff ’s likelihood of success. Settlement is now feasible only if P∆W + C∆ ≥ PπW − Cπ (assuming that settlement bargaining is costless). But this condition fails whenever Pπ is sufficiently larger than PΔ. In other words, settlement fails in cases of mutual optimism, where the plaintiff is optimistic about her success at trial (so Pπ is high) and the defendant is optimistic about his success at trial (so PΔ is low). In such cases, the minimum that the plaintiff will accept exceeds the maximum that the defendant is willing to pay. So settlement is not feasible.10

8.3.1.2.3 Asymmetric Information with Strategic Interaction When asymmetric information is combined with strategic interaction, it can create even more serious obstacles to settlement. Some of the most important work in this area, beginning roughly in the mid-​1980s (e.g., Bebchuk, 1984), models litigation by assuming that one of the parties has private information. These models—​technically, extensive form games that are solved for perfect Bayesian equilibria—​generate predictions that are sometimes surprising and often insightful. In general, asymmetric information creates opportunities for two different types of strategic behavior: signaling and screening. In a signaling strategy, the privately informed party moves first. If her private information is unfavorable to her, she will try to trick her uninformed opponent into believing that it is favorable. She does this by masquerading as a party with favorable information, or more precisely, by behaving in the same way that a party with favorable information would behave. This sends a false signal that her private information is favorable, which then pays off if her opponent believes it. For example, if an informed plaintiff with strongly favorable information and thus a strong case would demand a high settlement, then a plaintiff with unfavorable information and a weak case will make the same demand, hoping that the defendant will (mistakenly) believe that she has a strong case too. The uninformed defendant anticipates this strategy and tries to discourage it by rejecting high settlement demands. While this response succeeds in reducing the rate of false signaling, it does so by rejecting settlements in some strong cases. This results in strong cases being litigated even when they could have settled but for the informational asymmetry.11

10 



With some simple algebra, we can derive the following feasibility condition for settlement: Pπ − P∆ ≤ (Cπ + C∆ ) / W

This condition is violated whenPπ − P∆ is large enough; that is, when Pπ is large and PΔ is small. 11  For those readers with a technical bent: This signaling game can produce several different equilibria depending on the rules of the game and the game’s parameters. One particularly interesting equilibrium arises when the plaintiff has private information about the trial award and makes a settlement demand that the defendant either accepts or rejects. There is a separating equilibrium in which the plaintiff ’s demand perfectly reveals her type (i.e., her expected trial award), and the defendant rejects some of the time in order to prevent other plaintiffs from pooling (Reinganum and Wilde, 1986). In this situation, cases go to trial when the defendant rejects an offer.

 

Economics of Civil Procedure    153 In a screening strategy, the party without private information makes the first move. He knows that his opponent has private information, but he does not know whether that information is favorable or unfavorable to the opponent. He expects that parties with unfavorable information will try to mimic those with favorable information in order to receive the same beneficial treatment. His goal in making the first move, therefore, is to try to screen the unfavorable cases. For example, he might offer low settlements some of the time and high settlements the rest of the time. This way he saves the expense of paying high settlements to all the weak cases, but at the cost of litigating strong cases whenever his low offer is rejected. As a result, just as in the signaling game, asymmetric information leads to litigation even when a mutually desirable settlement exists in principle. To illustrate with a concrete example, return to our strike-​suit filing problem. In a strike suit, the plaintiff knows that her suit is meritless. Let us assume that the defendant does not know but that he can estimate probabilities. Thus, the asymmetric information is about the type of suit—​whether a specific suit is meritless or meritorious. To simplify, also assume that all meritorious suits have the same likelihood of success and that all are PEV. With information asymmetrically distributed in this way, a plaintiff choosing whether to file a strike suit plays a signaling game. She knows that if she files and mimics a meritorious plaintiff, she might be able to trick the uninformed defendant into believing that her suit is meritorious, and receive a substantial settlement. The defendant anticipates this strategy so he expects that the suit he faces might be meritless with some probability, and thus responds with a screening strategy. If the probability of a meritless suit is high enough, the defendant will choose to settle some of the time and litigate the rest of the time in an effort to deter the filing of strike suits. Strike-​suit plaintiffs anticipate this response and adjust their filing decisions accordingly. In equilibrium, plaintiffs file strike suits some of the time; defendants settle many of these suits for the same amount they settle meritorious suits, and defendants litigate meritorious suits that they would have settled but for the presence of strike suits (Katz, 1990; see also Bebchuk, 1988).12 As this example shows, the predictions of the strategic models based on asymmetric information are more nuanced and in some ways more realistic than the predictions generated by models that ignore strategic interaction. Nevertheless, these strategic models simplify reality by abstracting from the complex details of real cases. Simplification, however, is not necessarily a problem in itself. People regularly predict future behavior in their ordinary lives, and their predictions are based on simplifications of reality. Still, any simplification must include the most salient factors; otherwise, a model’s predictions are likely to miss the mark. 12 

In the Katz model, this equilibrium exists when the background fraction of meritless cases is large enough. There is a different equilibrium when the fraction of meritless cases is relatively small (Katz, 1990). When this condition holds, the defendant is better off settling with everyone, so he always offers the expected trial value of a meritorious suit (which is the only amount both a meritorious and meritless plaintiff would accept). In this equilibrium, all meritless cases are filed, all meritorious and meritless suits settle for the expected value of a meritorious suit, and no cases go to trial. The problem, however, is that strike suits receive large settlements.

 

154   Robert G. Bone These models also make rather strong assumptions about agent rationality, in particular the ability of parties and lawyers to process information and to predict and adjust to strategic responses, and they also assume purely self-​interested choice. The literature on bounded rationality and behavioral economics challenges these assumptions; it describes a variety of biases and heuristics that systematically skew decisions, as well as fairness considerations that can affect settlement behavior (see, e.g., Korobkin and Ulen, 2000; Pecorino and Van Boening, 2015). But these insights, as valuable as they are, do not call for jettisoning the formal models. Until we have a systematic predictive theory that incorporates these factors, it makes sense to generate initial predictions by using formal models and then modify those predictions in light of relevant constraints. In the end, a model is only as good as its predictions, and this makes it critical to test a model’s results for robustness and whenever possible compare those results with reliable empirical evidence. However, one aspect of the asymmetric information models might well be problematic. These models posit limited and unrealistic bargaining conditions. For example, it is customary to select one of the two parties to make the settlement offer and assume that the offer is made on a take-​it-​or-​leave-​it basis. While this approach renders the analysis more tractable, it also abstracts from the give-​and-​take of actual bargaining. Some scholars have used more complicated bargaining models, but much more work needs to be done in this area. In the end, I believe that formal models are more useful for their qualitative than their quantitative results. A model is best employed as a heuristic, a tool for revealing strategic effects and interactions that are difficult to perceive in other ways. Sometimes, models just dress up in formal garb what is otherwise obvious without them. But the best models reveal aspects of an interaction that at first seem counterintuitive but then become much more plausible as one thinks more carefully about the model’s predictions. In the litigation setting, for example, the formal models help to explain why settlement occurs so often and why it sometimes fails even when the parties would be better off settling. In particular, it focuses attention on the role of divergent expectations, which can arise for different reasons, and also on the various ways in which informational asymmetry can be exploited strategically.

8.3.1.2.4 Agency Costs The basic model assumes that attorneys are perfectly loyal agents. This assumption is unrealistic. Consider the simple contingency-​fee arrangement, which is a very common way plaintiffs finance litigation in the United States. In a contingency-​fee arrangement, the plaintiff ’s attorney bears the cost of suit but receives only a fraction of the recovery. This mismatch between cost and benefit gives the attorney a strong incentive to settle the case early if she can, and early settlements often fail to maximize the joint value of the claim (see Miller, 1987: 198–​202).13 Settlements based 13  In the United States at least, a lawyer is required to obtain her client’s approval of a settlement. However, lawyers have a huge informational advantage in this relationship, and there are lots of ways that a lawyer can persuade a client to accept a settlement.

 

Economics of Civil Procedure    155 on less than the full value of a claim shortchange plaintiffs and systematically favor defendants. There are also benefits to contingency-​fee arrangements, such as reducing risk-​ bearing costs, enabling the financing of socially beneficial litigation, and, depending on the attorney’s information, possibly screening meritless suits (Helland and Tabarrok, 2003; Dana and Spier, 1993). Moreover, other fee arrangements can also suffer from agency problems. Hence, the optimal solution is not necessarily to ban contingency fees outright. Instead, one should look for ways to align lawyer and client incentives more closely, and it is possible to design contingency-​fee contracts that improve this alignment (Wang, 2008; Hay, 1996). Most of the concern about agency costs in the procedure literature, however, does not involve contingency fees. Instead, it involves class actions and specifically the relationship between class counsel and the class. The class action is particularly susceptible to lawyer self-​dealing because class members have, at best, only weak incentives to monitor class counsel. As a result, class attorneys sometimes collude with defendants to obtain settlements that maximize their own fees at the expense of the class (Macey and Miller, 1991; Coffee, 1987). Several scholars have proposed solutions that aim to regulate fee awards in ways that align attorney and class interests more closely (Hay and Rosenberg, 2000: 1394–​402; Hay, 1997).

8.3.1.3 Other Considerations The foregoing analysis applies mainly to the content of specific procedural rules. Additional tools are needed to predict effects at the other two levels of a procedural system: the rulemaking process itself, and the choice of optimal rule form. Error risk and process costs matter at these levels just as they do for the content of specific rules. But they matter in a more indirect way. The choice of rulemaker, for example, affects error risk and process cost by influencing the types of rules that might be adopted. So too, choices about optimal rule form work their effects indirectly by, for example, leaving more or less of the procedural design to trial judge discretion. First, consider the choice of rulemaker. The positive law and economics literature is relatively sparse on this subject. There is, to be sure, a rich economics literature applying public choice theory and positive political theory to the legislative process, and much of this work is useful for predicting what might happen if the legislature had sole authority to make procedural rules. In general, legislators are likely to ignore procedural problems that do not command the attention of powerful interest groups, and the legislative response, when it does occur, is likely to favor organized interests over dispersed interests held by individuals who have difficulty organizing (Bone, 1999: 922–​3). Some scholars have also applied public choice and positive political theory to analyze committee-​based rulemaking. Their predictions are quite sensitive to the structure of the committee-​based process. At least one scholar predicts that a committee heavily populated with judge members will tend to produce rules that give trial judges broad discretion to manage litigation in ways that satisfy their own personal preferences

 

156   Robert G. Bone (Macey, 1994). Another scholar notes that a committee-​based rulemaking process is likely to produce open-​ended rules that delegate broad discretion to trial judges in order to dodge interest-​group conflict at the rulemaking stage (Bone, 1999). Other work shows that rulemakers can implement procedural policies that differ from the legislature’s preferences even when the rulemaking process is subject to legislative oversight—​ a result that implicates democratic values (Stancil, 2010). Finally, there is some law and economics work that analyzes procedural rulemaking by party agreement (Bone, 2012; Kapeliuk and Klement, 2008). Party rulemaking raises a host of complicated issues. The law and economics literature on the subject tends to favor giving more power to parties as long as the externalities are not too serious and consent is adequate. As for the optimal form of procedural rules, the economics literature on rules versus standards is instructive (e.g., Kaplow, 1992). A specific rule, for example, is more costly to make than an open-​ended and flexible standard, but it is also less costly to apply if it is followed strictly. As a result, strict rules can be superior to standards when information is hard to obtain ex ante and when the rule is applied frequently ex post, two conditions that are likely to hold for many procedural rules. Moreover, relying on open-​ ended standards that leave it to trial judges to craft optimal rules for specific cases might generate substantial costs, given a trial judge’s limited ability to access information, the strategic dynamics of litigation, and the adverse effects of bounded rationality (Bone, 2007: 1986–​2001).

8.3.2 Normative Theory For evaluation purposes, law and economics uses an efficiency metric that aims to minimize social costs (or, equivalently, maximize social welfare). From an efficiency perspective, procedure has value insofar as it helps to create socially desirable incentives in the real world. Scholars often assume that procedure accomplishes this goal by improving the accuracy of outcomes (including trial judgments, settlements, dismissals, and so on). This standard approach supposes that actors adjust their behavior in response to litigation outcomes, that accurate outcomes incentivize desirable behavior, and that erroneous outcomes incentivize undesirable behavior. Thus, someone designing a procedural system must pay attention to the social costs of outcome error as well as the social costs of the procedures used to reduce that error (Posner, 1973: 401). This standard account captures an important aspect of procedure, but it is also incomplete. A more comprehensive, and more complex, approach recognizes that litigation procedure can affect real-​world incentives in other ways, such as by threatening high process costs independent of outcome. The following discussion first describes the basic efficiency framework as laid out in the standard account. It then identifies a fundamental result implied by that framework. Finally, it discusses the more comprehensive approach and some refinements.

 

Economics of Civil Procedure    157

8.3.2.1 The Basic Framework In the standard account there are two main variables—​expected error cost and expected process cost—​and these variables depend on two factors: the probability of the cost-​ generating event and the social cost of the event when it occurs. In particular, expected error cost is the social cost of an outcome error discounted by the probability that an outcome error will occur. Similarly, expected process cost is the social cost incurred when a procedural rule is applied, discounted by the probability that the rule will actually be applied. The cost of process includes the social costs of litigating and deciding a case as well as the costs of operating the court system. The cost of error includes the social costs of noncompliance with the substantive law and the costs created by suboptimal primary-​ activity incentives. For example, when individuals believe that, because of litigation error, they can sometimes escape sanction when they fail to comply with the substantive law and will sometimes face a sanction even when they do comply, they will be less inclined to comply. The analysis is complicated by the fact that there are two types of error, and these errors often produce different costs. One type is a false positive (or Type I error) and the other is a false negative (or Type II error). A false positive occurs when an innocent defendant is mistakenly found liable, and a false negative occurs when a liable defendant is mistakenly found innocent.14 If individuals think that they might be held liable for doing perfectly lawful activities—​because of the risk of false positives—​they might choose not to engage in those activities, and this will create social costs. So too, if individuals think they can sometimes escape liability when they engage in unlawful activity—​because of the risk of false negatives—​they might be more inclined to engage in unlawful activities, which will also create social costs. Ordinarily, switching procedural rules reduces one type of error but increases the other. To illustrate, consider the burden of proof. Switching from a preponderance standard to a clear-​and-​convincing standard makes it less likely that innocent defendants will face liability (thus reducing false positives) but more likely that guilty defendants will escape liability (thus increasing false negatives). If the two types of error produce different costs, the errors must be weighted differently in the analysis. For example, many people believe that the social cost of failing to vindicate a meritorious civil rights suit (a false negative) is much more serious than the social cost of mistakenly imposing liability on an innocent party (a false positive). Given this belief, switching to a stricter burden might not be desirable, all other things equal, if it increases the risk of false negatives by too much—​since false negatives are the more costly type of error. In sum, this basic framework aims to minimize the total of (1) expected false negative error costs (i.e., the probability of a false negative multiplied by the cost of a false 14  There is no natural assignment of the “false positive” and “false negative” labels. The way I have assigned the labels in the text is fairly standard in the literature. An additional point: These definitions focus on errors in finding liability, but errors can also occur in determining damages and in mistakenly dismissing suits.

 

158   Robert G. Bone negative), plus (2) expected false positive error costs (i.e., the probability of a false positive multiplied by the cost of a false positive), plus (3) expected process costs (i.e., the probability that the rule will be applied multiplied by the cost of applying it).15 This analysis is further complicated by the fact that the different factors interact. For example, when false negatives decline, false positives usually increase. When procedure reduces error costs, it does so by adding process costs. Settlement introduces even more complexity. To illustrate, suppose that the burden of proof is made stricter in order to reduce false positives. Process costs should increase because plaintiffs, faced with a tougher burden, will conduct more discovery and invest more in presenting evidence at trial. However, the prospect of a costlier trial should create greater pressure to settle, and more settlements mean fewer trials and lower process costs. Moreover, if plaintiffs with weak cases choose not to file because of a concern about incurring high process costs, the rule will have achieved some of its benefit in reducing false positives without adding process costs to the mix.

8.3.2.2  Private Versus Social Incentives This framework has one very important implication. Private incentives to litigate are almost certain to diverge from socially optimal incentives (Shavell, 1997). This insight is important because it means that parties cannot be trusted to make efficient litigation decisions on their own. It also means that it can be difficult to identify the appropriate targets of regulation. The divergence between privately and socially optimal incentives to litigate follows from the existence of positive and negative externalities. Parties make litigation decisions to maximize their own private gain, and in doing so, they fail to internalize all the social costs and benefits of their decisions. Under the American Rule on fees, for example, a party has no reason to take account of the fees and costs of her opponent. The same is true, though to a different extent, in a system governed by the British Rule, where parties internalize their opponents’ fees only if they lose. Moreover, in neither system do parties count the public costs of litigation. The same is true for benefits. Parties usually make their litigation decisions by focusing on the private benefits of a remedy without regard to the social benefits of deterrence and precedent creation. Even when parties do concern themselves with deterrence or precedent creation, they do so only partially, since they are able to internalize only a small fraction of the total benefit. When costs are externalized, parties tend to overinvest in litigation, and when benefits are externalized, they tend to underinvest. It follows that private investment in 15 



Mathematically, we can express the problem in the following way:

{

Minimize rFN × cFN + rFP × cFP + q × cP

}

Where the variables have the following meanings: rFN and rFP are the probability of a false negative and a false positive, respectively; cFN and cFP are the average cost of a false negative and a false positive, respectively; q is the probability that the procedural rule will be applied; and cP is the average process cost of applying the rule.

 

Economics of Civil Procedure    159 litigation might be too large or too small relative to the social optimum, depending on the mix of negative and positive externalities (Shavell, 1997: 611–​2). This fact implies a need for regulation in some situations, but it also makes it difficult to identify the situations warranting regulation and to determine an appropriate regulatory response.

8.3.2.3 Expanding the Framework In the basic efficiency framework described so far, procedure creates social value by reducing the risk of outcome error. This is useful as far as it goes, but it is also incomplete. The reason to reduce outcome error is to improve compliance with the substantive law. Compliance, however, is affected not just by litigation outcomes, but also by litigation procedure distinct from outcome (Kaplow, 2013). This insight has important implications that are often ignored in the literature. For one thing, it means that process costs are not necessarily a negative. Someone deciding whether to comply with the substantive law takes account of all the costs of noncompliance (Kaplow, 2013:  1194–​5). This includes expected liability, but it also includes the expected costs of litigating future cases. This means that increasing process costs can strengthen compliance, and it also means that reducing process costs can weaken compliance. Still, it is important not to increase process costs to the point where socially desirable conduct is chilled. In an economic analysis, compliance with the substantive law is not valuable in itself; it is valuable because it leads to efficient behavior in the real world. If process costs are too high, actors might worry too much about being sued and overinvest in complying with the substantive law. They might, for example, forego socially desirable activity out of concern for the process-​cost risk. The fact that this overinvestment does not manifest itself as a failure to comply with the law does not mean it is any less of a concern. That procedure can affect compliance incentives independent of outcome should not be surprising. There is no fundamental distinction between substantive and procedural law. From an economic perspective, both aim to maximize social welfare (Kaplow, 2013: 1235). In theory at least, one can imagine a judge with the power to modify or ignore substantive statutory rule if procedure would do a better job of deterrence than applying the rule.16 This is not how it is done in practice, however, and for good reason. Deferring to the legislature or other lawmaking body respects democratic values and exploits the relative competence of different lawmaking institutions. Still, there is nothing inherent in the nature of procedure or substance itself that bars a more unified approach. There are specific situations, however, where the case for altering the substantive law can seem quite compelling. For example, a few judges have experimented with the use of 16  This would not be appropriate if the legislature were able to adjust liability and remedy to take account of the costs of enforcement. In that case, altering enforcement procedures would risk upsetting the statutory scheme. However, actual legislation is never so finely calibrated, and given the informational and public choice constraints, it is difficult to imagine that it ever could be.

 

160   Robert G. Bone sampling to determine individual liability and damages in large class actions. The goal is to avoid the high costs of conducting separate trials on individual liability and damages issues after collective resolution of the common issues.17 One sampling approach involves trying a random sample of cases and giving class members the average of the sample verdicts. Average recovery can adequately serve deterrence and compensation goals, understood in economic terms, even though it undercompensates for strong cases and overcompensates for weak cases. Giving everyone average recovery can promote deterrence when the average matches the expected liability that actors calculate ex ante (Kaplow, 1994), and it can compensate even those with severe injuries when it reimburses for the cost of insurance purchased before the injury (Rosenberg, 1987: 591–​3; see also Shavell, 1997: 594).18 The problem is, of course, that average recovery is not what the substantive law guarantees. Trial judges using sampling procedures do so in order to promote substantive goals more effectively, but the impact on substantive rights makes the practice extremely controversial.19 To sum up, procedural design should take account of more than error-​risk reduction. It should also focus on the direct effects procedure has on primary activity incentives. One general point emerges clearly from this analysis. Procedure is not value-​neutral. Rulemakers cannot craft optimal procedures without consulting substantive values. This is so because procedural design requires trade-​offs between the social costs of process and the social benefits of compliance measured in terms of the values that the substantive law is meant to promote.

8.4  Three Applications The following discussion applies this normative framework to three specific procedures: court-​annexed mediation, pleading, and discovery. The discussion of each topic is necessarily brief, given space constraints. My purpose is to illustrate the power of an economic approach, and I have chosen these three topics because of their general interest as well as the insights they provide.

17   Perhaps the most famous example is Cimino v. Raymark Inds., Inc., 751 F. Supp. 649 (E.D. Tex. 1990), Judge Parker’s attempt to handle the huge backlog of asbestos cases in the Eastern District of Texas. Judge Parker’s use of a sampling procedure was later reversed by the Fifth Circuit Court of Appeals in Cimino v. Raymark Indus., Inc., 151 F.3d 297 (5th Cir. 1998). See also Hilao v. Estate of Marcos, 103 F.3d 767 (9th Cir. 1996). 18  A more complete analysis would consider a number of additional factors, such as risk-​aversion, agency costs, and litigation investment incentives, and it would also consider effects on filing incentives as well as the risk of overdeterrence. 19  The legal issues are complex. Whether sampling is permissible depends, in part, on the purpose for which it is used and the nature of the legal claim. See Tyson Foods, Inc. v. Bouaphakeo, 136 S.Ct. 1036 (2016). It also matters whether sampling deprives the defendant of an opportunity to present individual defenses. See Wal-​Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011).

 

Economics of Civil Procedure    161

8.4.1 Court-​Annexed Mediation The general topic of alternative dispute resolution (ADR) is analyzed in another chapter of this book. There is, however, one type of ADR that warrants special consideration in a chapter devoted to civil procedure: mandatory court-​annexed mediation. Over the past thirty years, court-​annexed mediation has become a popular device to promote settlement. Courts in the United States have embraced it, and other countries are experimenting with it as part of their own court systems. The law and economics literature on mediation focuses primarily on its value for facilitating information exchange during settlement negotiations (Doornik, 2014; Brown and Ayres, 1994). For example, the mediator, by meeting with each side independently (known as “sequential caucusing” in the literature), is able to learn the private information each side possesses, use it to estimate a value for the case, and then communicate that value to both sides in a credible way without disclosing any private information. Knowing that their private information is safe from disclosure, parties should be more willing to reveal it to the mediator. And if they trust the mediator’s neutrality, they should also be willing to accept his estimate of case value. With common estimates of value, the parties are more likely to settle. Despite these settlement benefits, however, mandatory mediation is not necessarily justified on efficiency grounds. Even if mediators are completely neutral and even if sequential caucusing ameliorates the adverse effects of asymmetric information on settlement, it is not clear that the social benefits of mandatory court-​annexed mediation exceed the social costs. The benefits might not be as large as some of mediation’s supporters claim, and the costs are likely to be greater than commonly supposed (see generally Shavell, 1995: 16–​21, analyzing mandatory court-​annexed ADR). Let us start by considering mediation’s settlement benefits. Since the vast majority of cases settle anyway—​at least in the United States—​mediation is likely to reap substantial benefits only if it shifts the timing of settlement to an earlier point in the litigation. There are, however, limits to a mediator’s ability to affect settlement timing. For one thing, a party who believes that he would enjoy more settlement leverage in formal litigation might refuse to cooperate, or possibly even misrepresent his private information, in order to scuttle the mediation and force the case into litigation. Furthermore, some cases that would have settled on their own at an early stage might go through mediation instead, thereby adding process costs.20 Finally, even if mediation is successful in producing earlier settlements, the reduction in litigation backlog and delay might simply invite more case filings, which would increase litigation costs and reduce mediation’s cost-​saving benefits. On the cost side of the balance, mandatory mediation obviously creates its own process costs, and these costs increase with the time and effort devoted to the mediation. 20  This can occur if one of the parties believes she will do better in mediation and the probability of successful mediation is high enough (so she places a low probability on incurring the costs of regular litigation in addition to the costs of a failed mediation).

 

162   Robert G. Bone Indeed, a lawsuit that is difficult to settle without a mediator’s help might also be very costly to mediate. Moreover, strategic efforts to trick the mediator about the merits can produce systematically skewed settlements that distort compliance incentives. Also, a mediator who applies pressure to encourage compromise runs the risk of inviting weak and frivolous suits. And early settlements following mediation can reduce the storehouse of legal precedent. These benefits and costs are likely to vary with the type of case, making it difficult to draw any general conclusions. Nevertheless, some tentative points are worth mentioning. For one thing, court-​annexed mediation might not be a good idea where there is a robust private market for ADR.21 Moreover, even if court-​annexed mediation makes sense in general, one should be selective about which cases to subject to it, given the variance in mediation’s costs and benefits across different case types. In particular, one might consider limiting mediation to cases that are likely to have serious asymmetric-​ information obstacles to settlement. In addition, high-​stakes cases with hostile and uncooperative parties might not be appropriate for mediation if they are unlikely to settle early even with a mediator’s assistance. And one should be wary of court-​annexed mediation in cases where the precedent-​generating benefits of litigation are substantial.

8.4.2 Pleading The pleading rules in most countries require that a plaintiff allege specific facts in her complaint. Until 2007, the rules were different in the United States. American federal courts required only minimal factual allegations, just enough to notify the defendant of the general nature of the claim. However, in two recently decided cases, Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007), and Aschroft v. Iqbal, 556 U.S. 662 (2009), the United States Supreme Court adopted a stricter pleading standard, one that requires sufficient factual detail to make the claim “plausible.” This change has triggered a storm of controversy. Critics focus on the risk that meritorious plaintiffs will be denied access to court. Supporters focus on the benefits of screening meritless suits and preventing unjustified settlements. Much of this debate generates more heat than light. An economic analysis provides greater rigor and a more balanced perspective. The social benefits and social costs of stricter pleading are more complex than either side of the debate recognizes. The following discussion briefly surveys these benefits and costs. Before proceeding, however, it is useful to provide some background on how pleading and discovery work in federal courts. The plaintiff must file a complaint in order to begin a lawsuit, and her complaint must include factual allegations sufficient to state a 21  If the private value of settlement equals the social value, it makes sense to rely on the parties to choose from the private ADR market, assuming transaction costs are not too high. But courtannexed mediation might still make sense if private and social values diverge or parties have difficulty cooperating.

 

Economics of Civil Procedure    163 legal claim. If the complaint does not contain enough factual allegations, the defendant can get it dismissed. While the plaintiff must have some support for her allegations, it is not necessary that she state evidence or that she have enough evidence at the pleading stage to be able to prove her allegations at trial. After filing, the plaintiff can use discovery to gather the evidence she needs. However, the plaintiff has no access to discovery before she files her lawsuit (with some very limited exceptions). Ordinarily she must find factual support for her allegations by ordinary investigation and cannot force the defendant to reveal its private information. Hence the problem. If the defendant has exclusive possession of the information the plaintiff needs to make plausible allegations, the plaintiff will not be able to prepare a legally sufficient complaint even if she has a meritorious claim. The seriousness of this problem depends on the strictness of the pleading standard. Before Twombly and Iqbal, the pleading standard for most cases filed in federal court was very lax, and plaintiffs had little trouble meeting it. After Twombly and Iqbal, the standard has become stricter, although it is not exactly clear how much stricter it is (Gelbach, 2012). With this background in place, let us now turn to an analysis of the benefits and costs of the pleading standard. On the benefit side, requiring greater factual specificity reduces false positive errors by screening meritless suits. The effectiveness of this screen, however, depends on the information structure of the lawsuit. In particular, strike suits, those in which the plaintiff knows her suit is meritless, are not likely to be deterred. If the plaintiff and her attorney are willing to file a strike suit, they are already bad actors and unlikely to worry much about fabricating the necessary allegations. Nor need they be concerned about exposure at trial since they will drop the suit if it does not settle. Thus, a strike suit plaintiff should have little difficulty meeting a stricter pleading standard (suggesting the superiority of a penalty approach for strike suits). This means that strict pleading is likely to work only for plaintiffs who do not know whether their suits are meritless. Sometimes these plaintiffs have no reasonable way to know because the defendant has the necessary information and investigating is too costly. But, sometimes they could know by doing a reasonable pre-​filing investigation, yet choose not to investigate and end up filing a meritless suit. Stricter pleading can prevent these meritless filings by bolstering incentives to investigate. If an investigation reveals that the suit lacks merit, lawyers in these cases might well refrain from filing (since they are not necessarily bad actors). Moreover, stricter pleading also gives judges a better opportunity to screen suits after they are filed. A sloppy but honest attorney will probably include honest allegations rather than fabricated ones and a judge should be able to use those allegations to evaluate the case. Thus, the primary benefit of a stricter pleading standard is the reduction of false positive error costs associated with those meritless cases in which the plaintiff files without first investigating the merits. However, strict pleading also creates false negative error costs. It screens meritorious suits in which uninformed plaintiffs lack the information they need to plead and cannot obtain that information at reasonable expense by investigating before filing. The cost of these false negative errors, like the cost of the false positive errors that strict pleading avoids, depends on the nature of the case.

 

164   Robert G. Bone In addition to these error-​cost effects, strict pleading also affects process costs, although its impact on process costs is difficult to predict. On the one hand, stricter pleading saves the process costs of litigating suits that would have been filed under a weaker standard. On the other hand, a stricter pleading standard is likely to increase the process costs of applying the standard because judges must examine allegations more carefully when the standard is more demanding. But one must also consider the likelihood that defendants will file motions to dismiss invoking the stricter standard, since courts incur the process costs of applying the standard only when a motion is filed and the standard is actually applied. The fact that dismissal is more likely under a stricter pleading standard should encourage more motions, all other things equal. However, the more successful strict pleading is in deterring plaintiffs from filing deficient complaints, the less likely it is that defendants will file motions to dismiss. It is difficult to draw any clear conclusions, but the analysis suggests some useful points. Stricter pleading probably reaps benefits only in cases with asymmetric information and uninformed plaintiffs. Even in these cases, the efficiency of a pleading standard depends on a number of factors, including the background probability that a suit is meritless, the cost of investigating before filing, the social value of vindicating the policies behind the substantive law, the process costs of litigating and deciding motions to dismiss, and the likelihood that motions will be filed under different pleading standards. Moreover, the standard’s direct effects on compliance and primary activity incentives must also be factored into the mix, and this includes the direct incentive effects of reducing or increasing process costs. In fact, a complete analysis is even more complicated. So far, we have focused only on deterring meritless suits. The goal, however, is to create efficient filing incentives and achieving that goal might also require deterring some meritorious suits. This might be the case for meritorious suits that are so weak on the merits that the deterrence benefits of litigating them fall short of the social costs (Hylton, 2008). To be sure, screening meritorious suits is problematic under a rights-​based or other noneconomic metric, but it is a serious possibility under an economic approach. As complex as this analysis is, one thing is clear. The multiple factors affecting the costs and benefits of a pleading standard are likely to vary so much across different types of cases that pleading standards should vary as well. For example, if civil rights are highly valued so that the social cost of failing to vindicate a meritorious civil rights suit is much greater than the cost of tolerating a meritless filing, it might make sense to avoid applying a strict pleading standard to civil rights suits. Moreover, in some of the other types of cases, the optimal approach might be limited pre-dismissal discovery, or even a staged merits review in which parties conduct limited discovery followed by a judicial estimate of plaintiff ’s likelihood of success, and then additional discovery if the estimate exceeds some threshold level (Bone, 2009). However, staged processes like this raise a host of complicated issues (Kaplow, 2013).

 

Economics of Civil Procedure    165

8.4.3 Discovery The notoriously broad scope of discovery in the United States has long been a target of litigation critics. Broad discovery is blamed for high litigation costs, frivolous suits, and other perceived litigation ills. But these criticisms are too often cast in superficial terms. A law and economics analysis paints a much more complicated picture (Bone, 2012b). Most importantly, it identifies factors that should be considered when designing optimal discovery rules. The following discussion highlights some of these factors. One benefit of discovery is that it reduces the mutual optimism obstacle to settlement by forcing disclosure of private information, thereby bringing the parties’ estimates of likely success, trial award, and costs closer together.22 Indeed, the threat of discovery alone can improve prospects for settlement when information is asymmetric (Farmer and Pecorino, 2005). In addition to increasing the likelihood of settlement, discovery also improves the quality of settlements by aligning settlement amount more closely with expected trial outcome (Cooter and Rubinfeld, 1994). And it improves trial outcomes, too, by forcing the disclosure of relevant evidence that would not otherwise come out at trial. However, discovery also generates substantial social costs, some of which run directly counter to these benefits. First, discovery adds process costs when parties take depositions, propound and respond to document requests, and otherwise make use of discovery tools. Second, the high cost of discovery can encourage parties to settle before much, if any, discovery actually takes place. When this happens, not much process cost is incurred, but the informational asymmetry can interact with the threat of costly discovery to generate settlements that undermine deterrence (Hay, 1994: 513). Third, discovery can reduce the likelihood of settlement by forcing a party to disclose information unfavorable to its position. The disclosure of unfavorable information by one side increases the other side’s optimism relative to the baseline of no discovery, and mutual optimism tends to scuttle settlement. This result depends, however, on what information parties believe would remain private in the absence of discovery (Shavell, 1989). Fourth, the availability of broad discovery invites excessive and abusive use of discovery tools. One reason for excessive discovery has to do with cost externalization. In the United States, for example, the party taking discovery does not have to pay the response costs of its opponent, and this creates incentives to engage in discovery beyond the point where the marginal cost equals the marginal information value (Cooter and Rubinfeld, 1994).23 Another reason for excessive—and abusive—discovery has to do with strategic interaction. 22

  It is not clear, however, to what extent discovery actually improves the prospects for settlement when parties have the ability to disclose information to one another voluntarily (Shavell, 1989). In addition, some of the work on bounded rationality points to possible limits on discovery’s settlement potential. Insofar as parties construe ambiguous information with a self-​serving bias, information exchange through discovery might actually increase the divergence between the parties’ expectations (Lowenstein and Moore, 2004). 23  Still, because parties do not internalize all the deterrence benefits, discovery beyond the private marginal cost–​benefit breakpoint can sometimes be socially desirable (Hay, 1994).

 

166   Robert G. Bone Some scholars model the discovery process as a Prisoners’ Dilemma game (Setear, 1989; Gilson and Mnookin, 1994). In the Prisoners’ Dilemma, each party engages in excessive (and abusive) discovery for fear that her opponent will do so and she will be left worse off as a result. Abusive discovery occurs when parties strategically threaten to impose high discovery costs on their opponents just to pressure favorable settlements. In a products liability suit, for example, the plaintiff ordinarily needs much more discovery from the defendant than the defendant does from the plaintiff. As a result, the plaintiff is in a superior position to credibly threaten high discovery costs. This is a problem when plaintiffs use their threat advantage to leverage settlements in excess of the expected trial award. Discovery need not actually be conducted for this strategy to be effective; all that is needed is a credible threat (although discovery must be conducted sometimes to keep the threat credible). Thus, high process costs are not necessarily the most serious problem for abusive discovery. Rather, the serious problem is skewed settlements that can adversely affect compliance incentives. This brief analysis shows just how complicated it is to design an optimal discovery system. Optimal discovery depends on a number of factors that are likely to vary across case types. This suggests the desirability of designing different sets of discovery rules for different types of cases, assuming, at least, that the administrative costs of applying different rules are not too high.

8.5  Concluding Thoughts Law and economics contributes greatly to our understanding of civil procedure. It supplies a set of positive tools for predicting consequences as well as a useful normative framework for evaluating those consequences. Moreover, it makes an important contribution at a more general level. The economic approach teaches four critical lessons about procedural design that are often ignored in the mainstream civil procedure literature. The first lesson has to do with the importance of viewing procedure through the lens of strategic interaction. All too often, predictions that seem intuitively obvious at first glance fall apart when subjected to a careful strategic analysis. The second lesson has to do with the importance of viewing civil procedure as a self-​ contained and tightly integrated system. Lawyers and parties respond to procedural rules as interrelated parts of a unitary whole, so rulemakers should view them in that way as well and take account of the global effects of the rules they adopt. The third lesson is about the close relationship between procedure and substance. From an economic perspective, the two are unified in a particularly strong way. They both serve the same ultimate purpose of shaping primary activity incentives. This means that the trade-​offs required for optimal procedural design directly implicate the values that underlie the substantive law.

 

Economics of Civil Procedure    167 The fourth lesson is perhaps the most significant. An economic analysis highlights the critical importance of settlement to procedural design. It is often said that settlement takes place in the shadow of trial, but it is also true that trial takes place in the shadow of settlement. As Friedman and Wickelgren have shown, settlements influence trial outcomes by affecting the amount parties are willing to invest in trying cases that do not settle, and the resulting interplay between settlement and trial can weaken the deterrent effect of the substantive law compared to a system with trial alone (Friedman and Wickelgren, 2010). Moreover, an analysis that considers both settlement and trial points to different rules than an analysis that focuses on trial alone. To illustrate, consider the way that optimal discovery rules can change when settlement is taken into account. With only trial in the picture, broad discovery is relatively easy to justify from an outcome quality perspective (ignoring process costs) because broad discovery forces disclosure of more relevant evidence. However, outcome effects change radically when settlement is added to the mix. As we have seen, expanding discovery increases costs and risks, which makes it more likely that parties will settle before discovery takes place and settle for amounts that systematically deviate from the expected trial award. These four insights—​the necessity of strategic analysis, the importance of viewing procedure as an integrated system, the connection between procedure and substance, and the centrality of settlement to procedural design—​transcend any particular normative approach. They should be heeded by anyone concerned about designing an optimal system of civil procedure no matter what standard of optimality the person applies. On a more concrete level, there are a number of useful directions for future work in the field. For one thing, positive models should pay greater attention to settlement amount. Much of the current literature focuses on the rate of settlement. The implicit assumption seems to be that settlements are socially desirable because they make parties better off and save process costs. But this assumption fails to take account of the fundamental divergence between private and social incentives to litigate. In particular, settlements that make parties better off can be undesirable from a social point of view if the settlement amount deviates systematically from the expected trial award. In addition, there is a need for more law and economics work that explicitly analyzes procedure’s effects on compliance and primary activity incentives (see Hylton, 2002). The current literature tends to analyze effects internal to the litigation process, such as filing, investment, and settlement incentives. These are certainly important, but compliance and primary activity incentives lie at the core of what procedure is about. Finally, positive models should be expanded to include more players in the litigation game.24 Most of the current models assume that the party or her lawyer exercises complete control over the litigation. But in a world of moderate agency costs, parties and lawyers share control. This means that the litigation game is more accurately modeled with four players: the plaintiff, the plaintiff ’s lawyer, the defendant, and the defendant’s

24 

I am indebted to my colleague, Abe Wickelgren, for suggesting this point.

 

168   Robert G. Bone lawyer. Moreover, in real litigation, the judge is more than just a source of decisions; she is another strategic player in the game and should be modeled as such. As more players are added, however, models become more complex and more difficult to solve. This is the inevitable trade-​off between realism and tractability. Still, it is important to make an effort to incorporate other litigation actors because doing so can alter the results significantly. This is especially true for the judge. At least in the United States, judges have a great deal of discretion to manage litigation and can use their discretion in strategic ways. If parties expect strategic judging, they will make their litigation choices in anticipation of the judicial response. As we have seen, law and economics offers tools to analyze this type of strategic interaction. Indeed, the limited work that has been done shows some interesting results (e.g., Kockesen and Usman, 2012). In the end, the mainstream civil procedure community has much to learn from the law and economics literature. Moreover, law and economics scholars can benefit from a deeper understanding of the legal and practical dimensions of procedural institutions. Bridging the current gap between these two camps will take effort from both sides. Civil procedure scholars must be more receptive, and law and economics scholars must do more to reach out. The civil procedure field—​and possibly the civil procedure system itself—​will be much the better for it.

Acknowledgments I am grateful to Steve Shavell and my colleague Abe Wickelgren for extremely helpful comments on an early draft.

References Bebchuk, L. A. (1984) “Litigation and Settlement under Imperfect Information,” RAND Journal of Economics, 15: 404–​15. Bebchuk, L. A. (1988) “Suing Solely to Extract a Settlement Offer,” Journal of Legal Studies, 17: 437–​50. Bebchuk, L. A. (1996) “A New Theory Concerning the Credibility and Success of Threats to Sue,” Journal of Legal Studies, 25: 1–​25. Bone, R. G. (1999) “The Process of Making Process:  Court Rulemaking, Democratic Legitimacy, and Procedural Efficacy,” Georgetown Law Journal, 87: 887–​955. Bone, R. G. (2003) “Agreeing to Fair Process: The Problem with Contractarian Theories of Procedural Fairness,” Boston University Law Review, 83: 485–​552. Bone, R. G. (2007) “Who Decides? A Critical Look at Procedural Discretion,” Cardozo Law Review, 28: 1961–​2023. Bone, R. G. (2009) “Twombly, Pleading Rules, and the Regulation of Court Access,” Iowa Law Review, 94: 873–​936. Bone, R. G. (2012a) “Party Rulemaking: Making Procedural Rules through Party Choice,” Texas Law Review, 90: 1329–​98.

 

Economics of Civil Procedure    169 Bone, R. G. (2012b) “Discovery,” in C. W. Sanchirico (ed.) Procedural Law and Economics. Cheltenham: Edward Elgar Publishing, 188–​202. Brown, J. G. and Ayres, I. (1994) “Economic Rationales for Mediation,” Virginia Law Review, 80: 323–​402. Coffee, J. C., Jr. (1987) “The Regulation of Entrepreneurial Litigation: Balancing Fairness and Efficiency in the Large Class Action,” University of Chicago Law Review, 54: 877–​937. Cooter, R. and Rubinfeld, D. (1994) “An Economic Model of Legal Discovery,” Journal of Legal Studies, 23: 435–​63. Cornell, B. (1990) “The Incentive to Sue: An Option-​Pricing Approach,” Journal of Legal Studies, 19: 173–​87. Dana, J. D., Jr. and Spier, K. E. (1993) “Expertise and Contingent Fees: The Role of Asymmetric Information in Attorney Compensation,” Journal of Law, Economics, and Organization, 9: 349–​67. Doornik, K. (2014) A Rationale for Mediation and its Optimal Use,” International Review of Law and Economics, 38: 1–​10. Dworkin, R. A. (1985) “Principle, Policy, Procedure,” in A Matter of Principle. Cambridge, MA: Harvard University Press, 72–​103. Farmer, A. and Pecorino, P. (2005) “Civil Litigation with Mandatory Discovery and Voluntary Transmission of Private Information,” Journal of Legal Studies, 34: 137–​59. Fon, V. and Parisi, F. (2006) “Judicial Precedents in Civil Law Systems: A Dynamic Analysis,” International Review of Law and Economics, 26: 519–​35. Friedman, E. and Wickelgren, A. L. (2010) “Chilling, Settlement, and the Accuracy of the Legal Process,” Journal of Law, Economics, and Organization, 26: 144–​57. Fuller, L. L. (1978) “The Forms and Limits of Adjudication,” Harvard Law Review, 92: 353–​409. Gelbach, J. B. (2012) “Locking the Doors to Discovery? Assessing the Effects of Twombly and Iqbal on Access to Discovery,” Yale Law Journal, 121: 2270–​345. Gilson, R. J. and Mnookin, R. H. (1994) “Disputing through Agents: Cooperation and Conflict between Lawyers in Litigation,” Columbia Law Review, 94: 509–​66. Gould, J. P. (1973) “The Economics of Legal Conflicts,” Journal of Legal Studies, 2: 279–​300. Grundfest, J. A. and Huang, P. H. (2006) “The Unexpected Value of Litigation: A Real Options Perspective,” Stanford Law Review, 58: 1267–​336. Hay, B. L. (1994) “Civil Discovery: Its Effects and Optimal Scope,” Journal of Legal Studies, 23: 481–​515. Hay, B. L. (1996) “Contingent Fees and Agency Costs,” Journal of Legal Studies, 25: 503–​33. Hay, B. L. (1997) “The Theory of Fee Regulation in Class Action Settlements,” American University Law Review, 46: 1429–​82. Hay, B. L. and Rosenberg, D. (2000) “Sweetheart and Blackmail Settlements in Class Actions: Reality and Remedy,” Notre Dame Law Review, 75: 1377–​408. Helland, E. and Tabarrok, A. (2003) “Contingency Fees, Settlement Delay, and Low-​Quality Litigation:  Empirical Evidence from Two Datasets,” The Journal of Law, Economics, and Organization, 19: 517–​42. Hylton, K. N. (2002) “An Asymmetric-​Information Model of Litigation,” International Review of Law and Economics, 22: 153–​75. Hylton, K. N. (2008) “When Should a Case Be Dismissed? The Economics of Pleading and Summary Judgment Standards,” Supreme Court Economic Review, 16: 39–​66. Kapeliuk, D. and Klement, A. (2008) “Contractualizing Procedure,” http://​ssrn.com/​ abstract=1323056.

 

170   Robert G. Bone Kaplow, L. (1992) “Rules versus Standards: An Economic Analysis,” Duke Law Journal, 42: 557–​629. Kaplow, L. (1994) “The Value of Accuracy in Adjudication: An Economic Analysis,” Journal of Legal Studies, 23: 307–​401. Kaplow, L. (2013) “Multistage Adjudication,” Harvard Law Review, 126: 1179–​298. Kaplow, L. and Shavell, S. (2002) Fairness versus Welfare. Cambridge, MA:  Harvard University Press. Katz, A. (1990) “The Effect of Frivolous Lawsuits on the Settlement of Litigation,” International Review of Law and Economics, 10: 3–​27. Kockesen, L. and Usman, M. (2012) “Litigation and Settlement under Judicial Agency,” International Review of Law and Economics, 32: 300–​8. Korobkin, R. and Ulen, T. S. (2000) “Law and Behavioral Science: Removing the Rationality Assumption from Law and Economics,” California Law Review, 88: 1051–​144. Landes, W. M. (1971) “An Economic Analysis of the Courts,” Journal of Law and Economics, 14: 61–​107. Lowenstein, G. and Moore, D. A. (2004) “When Ignorance Is Bliss: Information Exchange and the Inefficiency of Bargaining,” Journal of Legal Studies, 33: 37–​58. Macey, J. R. (1994) “Judicial Preferences, Public Choice, and the Rules of Procedure,” Journal of Legal Studies, 23: 627–​46. Macey, J. R. and Miller, G. P. (1991) “The Plaintiff ’s Attorney’s Role in Class Action and Derivative Litigation: Economic Analysis and Recommendations for Reform,” University of Chicago Law Review, 58: 1–​118. Miller, G. P. (1987) “Some Agency Problems in Settlement,” Journal of Legal Studies, 16: 189–​215. Pecorino, P. and Van Boening, M. (2015) “Costly Voluntary Disclosure in a Screening Game,” International Review of Law and Economics, 44: 16–​28. Posner, R. A. (1973) “An Economic Approach to Legal Procedure and Judicial Administration,” Journal of Legal Studies, 2: 399–​458. Reinganum, J. F. and Wilde, L. L. (1986) “Settlement, Litigation, and the Allocation of Litigation Costs,” RAND Journal of Economics, 17: 557–​66. Rosenberg, D. (1987) “Class Actions for Mass Torts: Doing Individualized Justice by Collective Means,” Indiana Law Journal, 62: 561–​96. Rosenberg, D. and Shavell, S. (1985) “A Model in Which Suits Are Brought for their Nuisance Value,” International Review of Law and Economics, 5: 3–​13. Setear, J. K. (1989) “The Barrister and the Bomb:  The Dynamics of Cooperation, Nuclear Deterrence, and Discovery Abuse,” Boston University Law Review, 69: 569–​633. Shavell, S. (1989) “Sharing Information Prior to Settlement or Litigation,” RAND Journal of Economics, 20: 183–​95. Shavell, S. (1995) “Alternative Dispute Resolution: An Economic Analysis,” Journal of Legal Studies, 24: 1–​28. Shavell, S. (1997) “The Fundamental Divergence between the Private and the Social Motive to Use the Legal System,” Journal of Legal Studies, 26: 575–​612. Stancil, P. J. (2010) “Close Enough for Government Work: The Committee Rulemaking Game,” Virginia Law Review, 96: 69–​133. Wang, S. (2008) “The Optimality of Contingent Fees in the Agency Problem of Litigation,” International Review of Law and Economics, 28: 23–​31.

 

Chapter 9

L aw and Ec onomi c s of Evide nc e Chris William Sanchirico

9.1 Introduction The judicial function of government is said to consist of two broad tasks: interpreting the law, whose dictates are often ambiguous; and determining, in specific cases concerning specific individuals, what happened and under what circumstances, in order to apply the law as interpreted. The law of evidence concerns the procedural rules governing how courts (and other governmental actors) make the latter “factual” determinations, as opposed to the former “legal” determinations. In many jurisdictions evidentiary rules are chiefly concerned with situations in which certain forms of evidence—​for example, “character evidence” or “hearsay”—​are “inadmissible,” meaning that they may not be considered at all by the factfinder, whether judge or jury. The mathematical/​economic analysis of legal evidence, the subject of this chapter, deals to some extent with the effect and propriety of inadmissibility under specific rules of evidence. More often, however, it takes a broader view. Focused also on the substantial remaining stretch of distance between admitting evidence and deciding what it means, the mathematical/​economic analysis of legal evidence studies how deductions regarding acts and circumstances are or should be made from testimony, documents, and other evidence, whether proffered by interested parties or uncovered by the factfinder’s own investigation. Even more broadly, it attempts to divine or prescribe the objectives of legal fact-​finding, often by viewing evidentiary process instrumentally within the context of law’s overall purpose. The organization of this chapter reflects the structure of this literature. The first part of the chapter, the largest part, reviews the main mathematical/​economic approaches to legal evidence focusing on what they take to be the object of fact-​finding and how they deal with the strategic interests of the parties involved in the legal action. The second part of the chapter discusses the mathematical/​economic analysis of several specific

 

172   Chris William Sanchirico evidentiary rules, including those concerning character evidence, hearsay, privileges, proof burdens, and penalties for evidentiary misconduct. According to the parameters of the volume in which this chapter appears, the chapter is specifically restricted to those contributions to evidence scholarship that employ mathematical modeling—​whether probabilistic, game theoretic, or classically economic. This subset of contributions is certainly not exhaustive of those deserving attention. Furthermore, even within the subset of contributions employing mathematical models, many worthwhile papers could not be reviewed in this summary. Hopefully, the reviews that are included will facilitate the location and assimilation of other useful contributions, both those that employ mathematical modeling and those that do not.

9.2  Four Mathematical/​E conomic Approaches to Legal Evidence This part of the chapter reviews the four main mathematical/​economic approaches to fact-​finding. These approaches differ mainly along three dimensions: 1. the extent to which the approach accounts for strategic interaction as between interested parties and the factfinder, including attempts by parties to influence the factfinder, properly or improperly, to such parties’ benefit; 2. the extent to which the approach explicitly views fact-​finding instrumentally, within the larger context of law’s purpose, rather than as an end in itself; 3. the extent to which the approach accounts for strategic interaction among litigating parties themselves, including the possibility of exploiting such interplay in system design. The four main approaches are discussed in sections 9.2.2 through 9.2.5. Each section opens with a general discussion of the approach, moves on to describe several applications or implications, and closes with a review of criticisms. Before turning to the first approach, however, section 9.2.1 describes what it is that distinguishes models of evidence in particular from the far more populous species of procedural models.

9.2.1 Evidence Models versus Procedural Models In most models of legal process the focus of analysis is the incentive to file suit and the incentive to settle out of court. Accordingly, the description of the fact-​finding process in these accounts is typically skeletal.

 

Law and Economics of Evidence    173 In many such models evidentiary process is represented by a single exogenous parameter, the probability p that the court finds the defendant guilty or, in civil cases, liable. Neither the positive determinants nor the optimal generation of p is modeled. Seminal examples of this form of model include Gould (1973), Posner (1973), and Shavell (1982). Other models of legal process step somewhat beyond the single parameter approach to evidence. In these models, which are a species of “contest models,”1 evidentiary procedure is represented not by an exogenous parameter, but by an exogenous function p(x,y). This function determines the probability of the defendant’s guilt/​liability based on the prosecutor/​plaintiff ’s litigation effort, x and the defendant’s litigation effort, y. In these models, both x and y are numbers: each party’s litigation effort is measured along a single dimension, representing something like “intensity.” The function p(x,y) is usually not derived within the model, but rather chosen by the researcher. Often the choice of a particular functional form is not explicitly justified; according to at least one scholar, “analytical convenience” is usually the chief criterion (Skaperdas, 1996: 283). Important examples of this kind of model include Posner (1973: 452–​5), Braeutigam, Owen, and Panzar (1984), Rubinfeld and Sappington (1987),2 Katz (1988), and Bernardo, Talley, and Welch (2000).3 Evidentiary models, by contrast, generally attempt to flesh out what x, y, and p(x,y) are or ought to be. The price of this shift of focus is generally that less attention can be paid to filing and settlement behavior.

9.2.2 Approach 1: Pure Probabilistic Deduction The first mathematical/​economic approach to legal evidence is the “pure probabilistic deduction” approach.4 This approach, which is perhaps best described as purely mathematical rather than mathematical/​economic, appears mainly in the general legal literature on evidence. The focus is on Bayesian analysis and its probabilistic arithmetic. With regard to the three modeling dimensions laid out above: There is generally little discussion of the strategic considerations that naturally arise when knowledge resides with interested parties; fact-​finding is most often viewed as an end in itself; and the interplay between parties is generally not considered.

9.2.2.1 Bayes’ Rule The pure probabilistic deduction approach centers on Bayes’ rule and its implications. A quick review is in order.

1 

On contest models generally see Skaperdas (1996) and Corchón (2007). In this model, the productivity of the defendant’s trial effort depends on whether the defendant is guilty or innocent, which is exogenous. 3  In this model, the productivity of the party’s litigation effort depends on what action the party chose in the primary activity, as in Sanchirico (1995, 2000, 2001b). 4  Finkelstein and Fairly (1970) and Lempert (1977) exemplify this approach. 2 

 

174   Chris William Sanchirico Imagine a criminal action, in which the defendant may be either guilty G, or innocent I. The factfinder, judge or jury, starts with “prior beliefs” regarding guilt or innocence, represented by complementary probabilities, P(G) = 1 –​P(I).5 The factfinder then observes evidence E. This observation causes the factfinder to change, or “update” its6 beliefs regarding guilt or innocence. How the factfinders (hereinafter APM) beliefs are rationally updated depends on the relative size of two ancillary probabilities: the probability that the factfinder would have observed E were the defendant truly guilty, P(E|G) and probability that the factfinder would have observed E were the defendant truly innocent, P(E|I). It makes some intuitive sense that the rational factfinder takes E to be more of an indication of guilt the larger is the former probability relative to the latter. Bayes’ rule gives this intuition precise content. The notion of “relative size” for the two probabilities is captured by the ratio P(E|G)/​ P(E|I), the “likelihood ratio” of evidence E with regard to guilt G. Bayes’ rule then provides a simple relationship between the evidence’s likelihood ratio and rational evidence-​induced change in the ratio of the probability of guilt to the probability of innocence—that is, in the “odds” of guilt. To state this relationship, write P(G|E) and P(I|E) for the probability of guilt and innocence, respectively, after seeing the evidence, the “posterior” probabilities. Bayes’ rules then says that the posterior odds of guilt equal the prior odds of guilt scaled up or down by the likelihood ratio of the evidence with regard to guilt: P (G | E ) P (E | G) P (G) = . P (I | E ) P (E | I ) P (I )



(1)

The formula is easily derived from the definition of conditional probability:  P ( X | Y ) = P ( X ∩ Y ) / P (Y ), where “∩” means “intersection” or “and.”

9.2.2.2 Applications of the Pure Probabilistic Deduction Approach The following three applications of the pure probabilistic deduction approach are examples of how probabilistic arithmetic has been used to clarify practical mistakes in probabilistic reasoning.

9.2.2.2.1 Base Rate Neglect Experimental evidence suggests that factfinders often make the mistake of equating the probability of the evidence given guilt, P(E|G) with the probability of guilt given the evidence, P(G|E). Thus, learning or believing that the evidence they are seeing would be 90 percent likely to be observed were the defendant truly guilty, they deduce that the defendant is 90 percent likely to be guilty. Kahneman and Tversky (1973) and Nisbett 5 

P(G) and P(I) are numbers lying between 0 and 1 inclusive. The closer P(G) is to 1 (and so the closer P(I) is to 0), the greater the factfinder’s assessment of the likelihood of guilt, the number 1 itself representing certainty. 6  I will use the pronoun “it” for the noun “factfinder,” which may denote an individual (e.g., a judge) or a body of individuals operating under collective decision rules (e.g., a jury).

 

Law and Economics of Evidence    175 et al. (1976) identify the genus of errors in which this is one species. Guthrie, Rachlinski, and Wistrich (2007) provide a review of legal applications. This error is often referred to as “base rate neglect.” What is meant by “base rate,” and in what sense is it neglected? Bayesian arithmetic provides a precise answer. From the definition of conditional probability, P (G | E ) = P (E | G) P (G). P (E )



(2)

Thus to equate P(G|E) with P(E|G) is to “neglect” both P(G) and P(E) (or more precisely, their ratio). P(G) is the prior probability of guilt, or the “base rate” of guilt. P(E) is the prior probability of the evidence, or the base rate of the evidence. This second is the probability of seeing the evidence across all defendants, guilty or innocent, taking account of the relative probability of each type: P(E) = P(E|G)P(G) + P(E|I)P(I). Equation (2) says that the probability of guilt given the evidence is less than the probability of the evidence given guilt if the unconditional probability of guilt, P(G), is less than the unconditional probability of seeing the evidence, P(E). That is, even if the evidence is very likely to appear when the defendant is guilty, one cannot deduce, upon seeing the evidence, that the defendant is very likely to be guilty if the evidence itself is far more common than true guilt.

9.2.2.2.2 Trial Selection Bias Lempert, Gross, and Liebman (2000) argue that the tendency of police to “round up the usual suspects” (along with other features of pre-​trial process) reduces, and perhaps reverses, the probative value of evidence that the defendant committed past crimes.7 Even though past crime evidence may be probative of guilt in the general population, they argue, it is less probative among the subset of individuals whose cases reach trial. They point to several causes for this, including the tendency of police to “round up the usual suspects.” The trial selection bias argument may be formalized as an iterative application of Bayes’ rule (Sanchirico, 2001a, 2010a). Doing so lends some insight into its range of applicability. Thus, suppose we have before us an individual who has already reached trial in the capacity of defendant. Accordingly, our original prior odds of guilt P(G)/​P(I) have already been updated to P(G|T)/​P(I|T). We may then regard these updated odds as a new prior and update again on revelation of a criminal record, R. This iterative process will generate the odds of guilt conditional on both reaching trial and having a record: P(G|T) P (T | G) P (G) = . P(I|T) P (T | I ) P (I )



7 

The argument originates in an earlier edition of the same text, Lempert and Saltzburg (1982).

(3)

 

176   Chris William Sanchirico



P (G | R ∩ T ) P (R | G ∩ T ) P(G|T) = . P (I | R ∩ T ) P (R | I ∩ T ) P(I|T)

(4)

Were the individual not at trial, the odds of guilt upon learning of a criminal record would update simply as follows:

P (G | R) P (R | G) P (G) = . P ( I | R) P (R | I ) P ( I )

(5)

To grant that possession of a criminal record is informative of guilt in the general population is to say the likelihood ratio in the last equation, P(R|G)/​P(R|I) is greater than one. To assert that a criminal record is less informative of guilt, or even informative of innocence, within the subset of defendants standing trial is to say that the likelihood ratio in the second to last equation, P(R|G∩T)/​ P(R|I∩T) is smaller than that in the last, P(R|G)/​ P(R|I), or even less than one. With some algebraic manipulation, it can be shown that P(R|G∩T)/​ P(R|I∩T) is less than P(R|G)/​P(R|I), if and only if a past record increases an innocent defendant’s chance of standing trial by a greater proportion than it increases a guilty defendant’s:

P (T | I ∩ R) P (T | G ∩ R) > . P (T | I ) P (T | G)

(6)

Thus, trial bias would not arise merely because a criminal record increases an individual’s chance of standing trial. Trial bias requires that possession of a criminal record sends innocent suspects to trial more frequently than guilty suspects. It seems plausible that a criminal record increases an individual’s chance of standing trial—​as when police have a tendency to “round up the usual suspects.” It is far less plausible that possession of a criminal record sends the innocent to trial more frequently than the guilty. Certainly, a tendency to “round up the usual suspects,” by itself, says nothing about the differential impact of doing so on innocent versus guilty defendants.

9.2.2.2.3 The Irrationality of Hindsight Bias Experimental evidence suggests that factfinders are too ready to deduce from the bare fact that an accident occurred that the defendant knew it would occur (Fischhoff and Baruch, 1982; Kamin and Rachlinski, 1995). Such “(across person) hindsight bias” is important in law because liability often turns on the defendant’s foreknowledge or intent. A complication arises, however, because it is not always irrational to deduce foresight from occurrence (Baron and Hershey, 1988; Hershey and Baron, 1992, 1995; Sanchirico, 2004c, 2010a). Bayes’ rule is useful in identifying (and delimiting) the rational use of outcome information in judging the defendant’s foreknowledge. Such rational use arises when the factfinder believes that the defendant received private information regarding the accident’s likelihood, the factfinder has no way of directly determining the content

 

Law and Economics of Evidence    177 of that information, and there is common agreement regarding the interpretation of information possibly received by the defendant. Thus, suppose that the defendant may have seen either information I or information I’. If the defendant saw information I, he would have determined that the accident was likely—​likely enough to render his state of mind culpable. Conversely, his state of mind would not be culpable if he saw I’. Suppose, further, that the factfinder observes only whether an accident occurred and not, directly, whether the defendant saw I or I’. Suppose further that both the factfinder and the defendant agree that seeing I raises the odds of an accident more than seeing I’: that is, from Bayes’ rule, the factfinder and defendant agree that the likelihood ratio for I with respect to an accident is greater than the likelihood ratio for I’. It can then be shown (Sanchirico, 2010a) that the factfinder should rationally increase its assessment of the odds that the defendant’s state of mind was culpable—​that is, that the defendant saw I, not I’—​upon observing that an accident actually occurred. On the other hand, if the factfinder could directly observe whether the defendant saw I and I’, then it would indeed be irrational for the factfinder to use outcome information to determine the defendant’s foreknowledge.

9.2.2.3 Critiques of the Pure Probabilistic Approach The pure probabilistic approach has been subject to at least three kinds of criticisms. The first kind of criticism manifests as a train of “paradoxes”—​specific queries, examples, or experiments—​that call into question both the normative and descriptive validity of Bayesian analysis. These include the “Allais’ paradox” (Allais, 1953), the “Ellsberg paradox” (Ellsberg, 1961), the “conjunction problem” (Allen, 1986; Levmore, 2001; Stein, 2001; Allen and Jehl, 2003), and the “gatecrasher paradox” (Kaye, 1979; Allen, 1986).8 For example, the Ellsberg paradox—​a matter of balls, urns, and introspection—​calls into question whether it is valid to model “subjective” beliefs regarding likelihoods as if they were “objective” probabilities, an analytical move that is fundamental to the pure probabilistic approach and Bayesianism more generally (Savage, 1954). An objection more specific to legal evidence concerns the neglect of strategic considerations. Arguably, legal evidence differs fundamentally from other forms of evidence, such as that obtained from laboratory experiments, in that the sources of evidence—​ litigating parties, as opposed to cells or molecules—​are not only animate but also strongly interested in the observer’s deductions. The pure probabilistic approach to legal evidence appears to take little or no account of this seemingly material distinction. Another objection specific to legal evidence is that the pure probabilistic approach generally presents fact-​finding as a goal in and of itself, as if trials were a kind of micro-​ historical exercise. The implications of fact-​finding for the regulation of behavior outside the courtroom—​arguably, the vast majority of human behavior—​is at best a background consideration. Other approaches to evidence, considered below, indicate 8 

For more on this first kind of criticism see the symposium papers collected in Boston University Law Review (1986).

 

178   Chris William Sanchirico that while the objectives of uncovering truth and regulating primary activity behavior are related, they are far from coincident.

9.2.3 Approach 2: Omission Models The second approach, “omission models,” is the dominant approach in the economics literature (excluding the hybrid field of law and economics).9 This approach allows for some degree of strategic interaction between those presenting evidence and those interpreting it. However, the analysis of strategic considerations is limited by the assumption that litigants cannot lie or falsify, but only omit to reveal all that they know. Moreover, in many applications of this approach, fact-​finding is conducted for its own sake or with unspecified relation to unstated larger goals.10 The exploitation of strategic interaction among interested parties, however, is an important component of this approach.

9.2.3.1 Basic Structure of Omission Models Milgrom and Roberts (1986) is a seminal and still central example of the omission model approach. This subsection runs through simplified versions of the two main variants of that paper’s model.

9.2.3.1.1 Single-​Agent Model The basic insight of the single-agent omission model is that a “principal,” such as a court or agency, can learn the truth from an interested party by pre-​announcing to the party that any ambiguities in its testimony will be resolved in the party’s disfavor.11 The idea thus invokes the ancient evidentiary device of the spoliation inference, whereby the judge instructs the jury to assume that the facts go against a party who is known to have evidence regarding such facts but refuses to supply it. The truth-​revelation result in single-​agent omission models relies on the assumption that however much information the party supplies or withholds, the party may not contradict the truth. The factfinder thus knows with certainty that the truth is always among the fact patterns that are consistent with the party’s evidence. The only issue then is getting the party to be sufficiently specific. And this is accomplished by the factfinder’s announcement that it will assume the worst for the party if the evidence is more ambiguous than it need be.12 9 

See, e.g., Sobel (1985), Milgrom and Roberts (1986), Shavell (1989), Okuno-​Fujiwara, Postlewaite and Suzumura (1990), Lipman and Seppi (1995), Froeb and Kobayashi (1996), Shin (1998), Daughety and Reinganum (2000), Bull and Watson (2004), and Demougin and Fluet (2005). 10  Bull and Watson (2004) is an important exception. These authors adopt the no-​lying assumption, but like Sanchirico (1995, 2000, 2001b), posit that a party’s evidentiary choice problem is endogenous to the party’s primary activity behavior. 11  In the economics literature, this idea—​or at least its mathematical formalization—​is generally attributed to Grossman (1981) and Milgrom (1981). An early critical analysis appears in Farrell (1986). 12  As this discussion reveals, another important assumption is that the factfinder can tell whether the party is withholding evidence.

 

Law and Economics of Evidence    179 Consider a simple example. Suppose the truth, known to the party, is a number between 1 and 10. The party reports information about the true number to the factfinder, who, based on what it hears, makes a determination regarding what the true number is. The party may report subsets of numbers, rather than a single number, thus leaving the truth ambiguous to the factfinder.13 But whatever subset the party reports, however broad or narrow, it must—​by assumption—​contain the true number. Thus, the party can say that the number is between 4 and 7 when she knows it is 5. But the party cannot say the number is between 7 and 9. If the factfinder announces that it will treat as the truth the party’s least favorite number within the subset of numbers that the party reports, the party will never report a subset containing—​along with the true number, as it must—​a number which the party likes less than the true number. Thus, the true number will be the party’s least favorite number within the party’s reported subset. The result evaporates if the party is allowed to lie. Under the factfinder’s announced decision rule, the party would be induced to report a single number, the one she most highly favors, even if that is not the true number.

9.2.3.1.2 Multiple Agents with Conflicting Interests In economic theory, the introduction of more than one agent often works in favor of the principal. The presence of multiple agents opens up the possibility of playing them off against each other. Consistent with this, Milgrom and Roberts (1986) emphasize that when there are two or more parties, whose interests are sufficiently in conflict, the truth will be revealed even to a factfinder that is not sophisticated enough to fashion the kind of spoliation inference considered in the preceding subsection. Whatever rule the factfinder uses, so long as the rule leads to determinations that are consistent with all evidence produced by all parties, the factfinder will end up choosing the true outcome. Thus, suppose the simple 1-​to-​10 example described above is augmented by adding a second party that also knows the true number and also reports a subset that must contain the true number. Suppose further that the two reporters’ interests are opposed. Imagine, for example, that the first agent wants the factfinder to think that the true number is as high as possible, while the second wants the factfinder to think it is as low as possible. Now consider any rule that the factfinder might pre-announce and deploy to choose among points in the intersection of the two parties’ reported subsets. For example, the factfinder might choose the minimum, maximum, or average of points in the intersection of reports. A Nash equilibrium of the interparty reporting game induced by any such factfinder decision rule is a pair of reported subsets, one for each party, such that 13  Sanchirico (2010a) discusses several interpretations of evidence production under this no-​ lying assumption that appear in the literature. The discussion here centers on the “subset reporting” interpretation, which is the most direct and is also the interpretation used in Milgrom and Roberts (1986).

 

180   Chris William Sanchirico each party’s reported subset maximizes her payoffs taking as given (and known) the other party’s reported subset. In any Nash equilibrium of any interparty reporting game so induced, the factfinder will be choosing the true number. To see that this is true, suppose provisionally that it is not. Then, given that the factfinder chooses within the intersection of reports, that intersection must contain another number beside the true number. By assumption, one party prefers the true number to this other number. But this party could have reported a singleton subset consisting solely of the true number, thus forcing the factfinder to choose the true number. Therefore, the posited outcome cannot be a Nash equilibrium. Once again, this result relies on the assumption that neither party can report a subset not containing the true number. If the parties were so permitted, then their reported subsets would not even necessarily intersect. Each party would report the single number maximizing her payoffs given the rule adopted by the factfinder and the other party’s similarly singleton report. If the factfinder’s rule was to take the average of the reported numbers, for example, one party would always report 1, the other 10, and the factfinder would always choose 5 no matter what the true number.

9.2.3.1.3 The Sufficiency of “Partial Provability” Both the single-​party and two-​party version of the truth-​finding result in omission models operate by inducing the parties—​whether through the factfinder’s decision rule or the parties’ conflicting interests—​to remove ambiguity from their evidentiary presentation. It has been emphasized that the truth-​finding outcome relies on the no-​lying assumption, which effectively equates specificity with truth-​revelation. However, the result also appears to rely—​at least to some extent—​on the ability of the parties to be in fact specific. The question is: To what extent? Certainly some degree of evidentiary precision is required. If in the single-​agent 1-​ to-​10 example above, the party were unable to report subsets that did not contain 2, an assume-​the-​worst rule would not generally reveal the truth. Assuming the party preferred greater numbers to lesser, the factfinder would never choose a number greater than 2, regardless of the true number. On the other hand, it is also clear that complete specificity is not required. Again considering the single-​agent 1-​to-​10 example, suppose the party was only able to report subsets containing the truth and all numbers greater than the truth. For example, were the truth 7, the party could report only {7, 8, 9, 10}. Suppose, further, that the party prefers higher numbers to lower. Then an assume-​the-​worst rule would still reveal the truth, despite the party’s inability to fully specify the state. No matter what the true number it would always be possible for the party to report a subset containing no number that the party regards as worse than the truth. Thus, given the factfinder’s rule and the party’s preferences, it would in effect be possible for the party to be specific. Were the party to say “{7, 8, 9, 10},” she would in fact be saying “7” in the factfinder’s assume-​the-​worst language. Lipman and Seppi (1995) study partial provability in a multi-​party omission model with conflict of interest. In their model, unlike in Milgrom and Roberts’ (1986)

 

Law and Economics of Evidence    181 multi-​party model, the parties present evidence sequentially: first one party, then the other. Lipman and Seppi lay out conditions on the structure of evidence that allow for the possibility that parties are not always able to be fully specific, but that are nevertheless sufficient for truth-​finding.14

9.2.3.2 Applications of Omission Models 9.2.3.2.1 Litigation Bias? Froeb and Kobayashi (1996) add to Milgrom and Roberts (1986) multi-​party model the realism-​enhancing feature that the evidence presented by parties is costly for them to collect. Specifically, the authors model evidence collection as strategic search. Each party pays, per draw, to sample from a single exogenous distribution emitting pieces of evidence. Any given piece of evidence may help or hurt the party’s case given the factfinder’s decision rule. The party decides when to stop sampling, as well as what portion of its accumulated sample to report to the court. The party may present only part of its accumulated sample, but the party may not falsify. Froeb and Kobayashi show that adding evidence costs to the model in this way leads to unbiased trial outcomes, even if the factfinder is itself biased. The exogenous distribution emitting evidence favors the party with the better case. As a result, this party presents more self-​favoring evidence than the other party. Any factfinder bias is counteracted by the fact that the party favored by the bias reacts, in this model, by choosing to sample less often. Daughety and Reinganum (2000) examine a more general strategic search model that contains Froeb and Kobayashi’s (1996) model. They argue that Froeb and Kobayashi’s no-​bias result turns on symmetries built into their model. They find that in general the adversarial system favors defendants.

9.2.3.2.2 Adversarial versus Inquisitorial Process Broadly characterized, the English system of evidentiary process (as inherited by the United States and other jurisdictions) is “adversarial” in the sense that the parties themselves are responsible for evidence generation and presentation, while the factfinder’s role is mainly evaluative. The Continental (European) system is said to differ in that the factfinder actively investigates and the parties are restrained. The Continental system is said to be “inquisitorial” rather than adversarial. Omission models have been fruitfully deployed to illuminate trade-​offs between these two systems. Shin (1998) considers the trade-​off between the fact that the adversarial system allows more sources of information than the inquisitorial, and the fact that adversarial sources are biased. He argues that the biasedness of adversarial process is mitigated by the kind

14  A detailed discussion of partial provability, including Lipman and Seppi’s (1995) complicated model, appears in Sanchirico (2010b). Bull and Watson (2007) employ a modeling structure similar to that in Lipman and Seppi (1995) with the object of understanding the relationship between the mechanism design framework and their conception of “hard evidence.”

 

182   Chris William Sanchirico of assume-​the-​worst-​of-​omission rule studied by Milgrom and Roberts (1986)—​under the same no-​lying assumption. Consequently, he favors the inquisitorial approach. Froeb and Kobayashi (2001) analyze a different trade-​off and have a more agnostic view. In a strategic search model similar to Froeb and Kobayashi (1996), parties present their most favorable draw from the evidentiary distribution and the factfinder averages the reports. Again, the parties may not report a sample element that they did not actually draw. In ex ante expectation, the factfinder’s averaging of sampling extremes (adversarial process) produces the same outcomes as if the factfinder itself drew from the distribution and averaged its full sample (inquisitorial process). The difference between adversarial and inquisitorial process lies rather in the variance of factfinder decision-​ making, which the authors take to be a proxy for error. The relative sizes of each system’s variance depends on model parameters.15

9.2.3.3 Critiques of the Omission Model Approach Omission models rest variously on an array of assumptions. However, the assumption that defines this class of models is also the most controversial: parties can only omit and not lie. The most common defense of this assumption is that perjury and falsification are illegal. This defense is problematic in several respects. First, many of the forms of omission that agents are explicitly allowed to perpetrate in omission models are as illegal as perjury. Witnesses and documents can be subpoenaed, discovery can be compelled, and nonfeasance in either case invites sanction. “Obstruction of justice,” a catchall crime for litigation-​oriented misconduct including document destruction and other forms of omission, is punished with penalties on a par with those for perjury. The (rhetorical) question thus arises: if the lying problem is, could be, or should be solved by making it illegal, why not also the omission problem? Second, the fact that behavior is illegal does not mean that it does not occur. The probability of detection and the severity of sanctioning may be insufficient to deter all potential perpetrators. In the particular case of perjury, the limited data available seem to indicate that falsification and lying are regular occurrences in actual legal process (Sanchirico 2004b). Therefore, perjury law, at least in its current state, does not seem to justify proceeding with the rest of system design as if the lying problem were pre-​eradicated. Of course, detection effort and sanctions for perjury could be increased. And perhaps it would even be possible to effectively eliminate falsification in this way. But would that be the right way to design the system? What would be the social cost of eliminating falsification through detecting and sanctioning perjury? Are there alternative means of dealing with the incentive to falsify besides penalizing perjury?16 How does criminalizing perjury compare to these alternatives in terms of cost effectiveness? 15 

Other analyses of adversarial process include Landsman (1984) and Dewatripont and Tirole (1999). Cooter and Emmons (2003, 2004) propose that, in lieu of sanctioning perjury and obstruction, litigants be required to post bonds whose proceeds they forfeit if their testimony is later determined to have been false. 16 

 

Law and Economics of Evidence    183 Thus, the third problem with the perjury defense of omissions models, is that it exogenizes an important part of the design problem for evidentiary procedure. Practicing litigators are likely to list falsification incentives as one of the chief problems to be solved by any fact-​finding system. By assuming the pre-​existence of completely deterrent perjury law, omission models neglect one of the most important parts of the problem. Why then do so many economic models of evidence assume away the falsification problem in their analysis of legal evidence? One candidate answer is that the falsification problem is intractable given the present state of game theory and (its most relevant subfield) mechanism design. Yet this answer seems unlikely. There is no analogous no-​lying assumption in models of optimal taxation, and in many models of optimal regulation. These are also models of mechanism design. And these are also models in which the government is the principal. Indeed, these are also settings in which perjury and related crimes apply, as tax forms themselves often indicate. The optimal labor income tax model would be quite different were it assumed that individuals could not lie, only omit, in reporting their underlying labor productivity to the government. Instead, the optimal tax model regards the incentive to lie as a problem-​ defining constraint. It then devises a means of using actual labor earnings as a signal of ability, albeit at a cost to efficiency. If the no-​lying assumption is acceptable in litigation models, why not also in tax models? Conversely, if lying is allowed in tax models, why not also in litigation models?

9.2.4 Approach 3: Endogenous Cost Signaling The third approach to legal evidence is “endogenous cost signaling.” This approach appears mainly in the law and economics literature on evidence, though it borrows and combines adverse selection and moral hazard models from the mechanism design literature in game theory. This approach attempts to allow for both lying and omission incentives. It views fact-​finding not as an end in itself but, explicitly, as instrumental to the broader goal of efficient law enforcement. However, it generally does not account for the strategic interaction among multiple parties to a dispute. In attempting to address the full strategic problem as between interested evidence producers and factfinders endogenous cost-​signaling models view evidence as a form of costly signaling, much like labor earnings is a costly signal in optimal tax models such as Mirrlees (1971), and degree attainment is a costly signal in Spence’s (1974) model of education. In cost-signaling models, agents differ from each other according to an unobservable parameter of interest that also determines their marginal cost of a particular observable action (i.e., signal) such as generating another dollar of labor earnings or earning a college degree. The principal in the problem, such as the government, sets out a schedule of rewards that vary according to these observable actions. If the schedule is “separating” at a particular point, then only agents with low marginal signaling costs find it worthwhile to send a particular level of the signal. In this manner, the principal is

 

184   Chris William Sanchirico effectively able to observe—​though not without cost—​the underlying unobservable parameter of interest. As their label suggests, however, endogenous cost-​signaling models of evidence are not merely costly signaling models. This is related to the fact that such models do not view fact-​finding as an end in itself. Rather, such models view evidence production as part of an overall system whose ultimate purpose is to regulate primary activity behavior—​which is to say, the day-​to-​day personal and business behavior that forms the subject of contract, property, tort, and any other substantive law area. As such, endogenous cost evidence models do not, as do costly signaling models, take as given agents’ underlying unobservable parameters—​like labor productivity or educational intelligence—​that determine signaling costs. Rather, in an endogenous cost-​signaling model of evidence, signaling costs—​which is to say evidence costs—​are endogenous to agents’ primary activity action choices.

9.2.4.1 Basic Mechanism Sanchirico (1995, 2000, 2001b) first proposes viewing legal evidence as endogenous cost signaling. The basic mechanism of such models is illustrated in the following simple, two-​part example.

9.2.4.1.1 Part I: Evidence as Exogenous Cost Signaling Let us begin with the hypothesis that the individual standing before the factfinder has some underlying characteristic (or “type”) c. Characteristic c may take the value “high” or “low.” Later we will give some content to c, but for now we take its existence as given. The factfinder cannot directly observe whether the individual’s c is high or low. Suppose, however, that there is some evidentiary presentation E that the individual can make to the factfinder whose cost to the individual differs according to whether the individual’s c is high or low. Imagine specifically that if c is high, the cost is $100 and if c is low the cost is $30. Lastly, suppose that the factfinder announces ahead of time that it will fine the individual $50 unless she presents evidence E. Because $50 is less than $100, the individual will not present the evidence if c is high. Because $50 is greater than $30, the individual will present the evidence if c is low. Therefore, by observing whether the individual presents the evidence, the factfinder will learn whether c is high or low. In the process, the factfinder will be punishing the individual for having a low c rather than a high c. The punishment is $20, which is the difference between the fine paid if the individual’s c is high ($50) and the cost of exonerating evidence if the individual’s c is low ($30). So far we have described an exogenous cost-​signaling model. Its mechanism is the basic mechanism behind all such models including Mirrlees (1971) and Spence (1974). Generalized from the preceding example, that basic mechanism is this: given an array of differentially costly signals whose cost correlation with unobservable characteristics is known, it is possible to varying degrees to structure a signal-​reward function (a “mechanism”) inducing a correlation between unobservable characteristics and observed

 

Law and Economics of Evidence    185 signals, so that deductions about unobservable characteristics may be made on the basis of signals observed. The deductions are a familiar and intuitive form of hypothetical reasoning about human behavior: the individual would not be doing what she is observed to be doing were her true situation otherwise. It should be noted that this costly signaling dynamic is more than merely a clever trick for truth-​revelation. It is fundamental. Given lying incentives, the presence of a differentially costly signal—​and the structuring of a reward function to exploit it—​is essentially the only way to generate evidentiary observations that vary with underlying truths,17 and so the only way to effectively observe such truths.

9.2.4.1.2 Part 2: evidence as endogenous cost signaling Returning to the example: so far the factfinder has learned whether c is high or low. If that was the factfinder’s objective, then the analysis would move on to the next issue, whether it is possible to obtain that objective at lower social cost, taking into account the cost of costly signaling. Suppose, however, that the factfinder is merely one arm of a government whose ultimate objective is not to learn about underlying characteristics, but to induce individuals to follow some law L. The government does not directly observe individuals’ behavior in relation to law L. But can it somehow make use of evidentiary presentation E? Clearly, if there is no relationship between whether c is high or low—​that is, whether presenting E is of high or low cost—​and whether the individual followed law L, then E is of no use in incentivizing compliance. Incentivizing compliance requires creating a differential benefit to the individual from compliance (Sanchirico, 2001b: 344–​5). This requires that the individual’s best attainable payoffs under the legal system are greater when she complies than when she does not. In the system set out here, the individual’s greatest attainable payoffs are greater, $20 greater, when her evidence costs are low rather than high. If there is no link between such evidence costs and compliance, there is no link between her payoffs and her actions, and so no incentive to comply. On the other hand, compliance incentives do become possible when the cost of evidence E is appropriately correlated with compliance. Suppose, for example, that evidence is always of low cost when the individual complies and always of high cost when she does not. Then under the system as it currently stands, we have provided the individual with a $20 incentive to comply with the law. If it otherwise costs the individual less than $20 to comply, she will choose compliance. The notion that evidence production—​interpreted as costly signaling—​can be used to set primary activity incentives when signal costs are endogenous to primary activity behavior is the basic logic of the endogenous cost-​signaling approach to evidence.

17  More precisely, choices must differ according to underlying truth and this can come from either differential signaling costs or the differential impact of reward functions. Differences in reward-​function impact are probably less relevant in the analysis of legal evidence.

 

186   Chris William Sanchirico

9.2.4.2 Applications of Endogenous Cost Signaling 9.2.4.2.1 Finding Facts versus Inducing Compliance Much legal scholarship on evidence regards truth-​finding as the primary object of evidentiary process. Legal decisions and rule revisions generally take this as given. One could defend this approach on its own terms. But perhaps more fruitfully, one might argue that truth-​finding is implicitly a proxy for more defensible less process-​centric goals like primary activity incentive setting. But how good a proxy is it? The endogenous cost-​signaling approach to evidence explicitly regards the fact-​ finding process not as an end in itself but as an instrument for setting primary activity incentives. Viewed in this light, the fact-​finding process is inaptly named. Finding facts—​that is, determining what actually happened—​is neither necessary nor sufficient for setting primary activity incentives. Revelation of the true state of affairs requires that the individual choose different evidentiary presentations when their primary activity actions are different. By contrast, setting primary activity incentives requires that the individual’s payoffs from evidentiary process are different. The latter is not the same as the former (Sanchirico, 2001b: 332–​6, 345).18 The example in section 9.2.4.1 can be used to show that truth-​finding is neither necessary nor sufficient for setting primary activity incentives. As the example is there cast, the factfinder learns whether the individual has been compliant. At the same time the individual’s payoffs from evidentiary process are $20 greater if the individual has complied. A $20 difference in payoffs may be insufficient to induce the agent to comply. The cost of compliance, not counting evidentiary payoffs, may exceed $20. Thus, truth-​ revelation is insufficient for incentive setting. Conversely, a modified version of the example can be used to explain the less obvious point that truth-​finding is unnecessary. Suppose the factfinder changed the rule so that the punishment for not presenting evidence E was $110 rather than $50. Recall that the individual’s cost for presenting E was either $100 or $30. Given the new rule, the individual will present evidence E in both cases. Therefore, the factfinder will not learn whether the individual complied with the law. However, the individual will be provided with an incentive to comply that is greater than in the original, truth-​revealing version of the example. If the individual does not comply, she incurs evidence costs of $100. If she does comply, her evidence costs are only $30. Therefore, the individual is effectively rewarded $70 for complying. This is greater than the effective award of $20 in the original version of the example.

9.2.4.2.2 Verifiability and Contract Design Contract theorists sometimes assert that optimal contracts condition contractual obligations only on contingencies that can be verified in court (Hart and Moore, 1988; Schwartz, 1992). However, Sanchirico and Triantis (2008) provide an explicit account 18 

Other, different reasons for a divergence between truth-​revelation and incentive setting appear in Schrag and Scotchmer (1994), Kaplow and Shavell (1996), and Sanchirico (2001a).

 

Law and Economics of Evidence    187 of evidence production—​within the endogenous cost-​signaling approach—​in which it may be optimal for contracting parties to design a contract knowing that one party will be induced to fabricate evidence, and be successful in doing so, in certain states.19 In Sanchirico and Triantis (2008), the link between primary activity actions (there: contractual performance) and evidence costs is probabilistic. The authors identify a trade-​off between two factors in contract design. First, contracting parties will wish to minimize the cost of evidence production in any given state. All else the same, this factor causes the parties to avoid the more costly presentation of fabricated evidence. Second, certain evidentiary states are more “probabilistically efficient” than others. A relatively probabilistically efficient evidentiary state is one that exhibits a large difference in probabilities across primary activity actions—​so as to maximize incentive creation—​but is overall relatively unlikely to occur—​so as to minimize expected evidence costs. Sanchirico and Triantis (2008) show that if evidentiary states differ sufficiently in their probabilistic efficiency, the desire to load rewards and punishments into efficient states will cause the parties to design their contract so that evidence production is exaggerated beyond the bounds of truth in these states.

9.2.4.2.3 Presumptions Bernardo, Talley, and Welsh [BTW] (2000) use a kind of endogenous cost-​signaling model to study the effect of legal presumptions. In their model a legal presumption favoring one party is modeled as a parametric tilt in favor of that party in the contest function employed to model evidence production. But for the parameter defining the presumption, the functional form mapping evidence onto litigation outcomes is exogenous (see section 9.2.1). Parties’ evidence costs, however, are endogenous to individuals’ primary activity behavior (which in BTW’s model is “shirking”). BTW make several theoretical points and discuss their model’s applicability to several substantive law areas. Their main point concerns a kind of counterintuitive feedback effect on primary activity behavior from adjusting presumptions. Shifting the presumption to favor the defendant may actually induce plaintiffs to file suit more often and, indeed may increase the frequency of plaintiff victory among filed cases. The key dynamic is a matter of endogenous evidence cost differentials. Given the form of BTW’s litigation contest function, shifting the presumption in favor of the defendant helps shirking defendants more than non-​shirking defendants. Therefore, shirking increases, which in turn leads to more suits and more plaintiff victories.

9.2.4.2.4 Allocation of Proof Burdens Many commentators suggest that the burden of proof—​a procedural device discussed in more detail in section 9.3.1—​should be assigned to the party with better information. Yet, in practice the burden of proof is often assigned to the opponent of the party whose primary activity incentives are being set by the law in question—​as when the plaintiff 19 

Bull (2008) also studies contracting, verifiability, and costly evidence production, but in the context of an extended omission model of evidence, such as was discussed in section 9.2.3.

 

188   Chris William Sanchirico has the burden of proving the defendant’s negligence and the defendant the burden of proving the plaintiff ’s contributory negligence (Sanchirico, 2008). The actor’s opponent is unlikely to have better information than the actor herself regarding the actor’s own actions. Using an endogenous cost-​evidence model, Sanchirico (2008) finds some justification for this regularity. The model is generalized in Sanchirico (2005), which studies how litigation design affects the strategic complementarities between parties’ evidentiary choices. The assignment of proof burdens, an aspect of litigation design, arranges these strategic complementarities. Placing the burden on the incentive target’s opponent is shown to efficiently arrange these complementarities.

9.2.4.3 Critiques of the Endogenous Cost Evidence Approach Is it really possible to find evidence whose cost differs according to individuals’ primary activity actions? The endogenous cost-​signaling approach appears to suppose so. But is this plausible? Two responses may be offered. The first is that the endogenous cost-​signaling approach does not actually assume the existence of such evidence. Rather, it asserts that the existence of such evidence is necessary (and sufficient) for evidence production to have utility in setting primary activity incentives. That is, it allows that such evidence may not exist, but insists that if one believes this, one must also believe that setting rewards and punishments based on evidence production is ineffective in inducing compliance with the substantive law. The second response is to attempt to provide some examples of endogenous cost evidence in actual practice. Sanchirico (2004a) describes how cognitive limitations are a source of endogenous cost evidence. For example, telling a detailed and consistent story that is robust to cross-​examination is less costly if the story conforms to reality than if it is fabricated. Another objection to the endogenous cost-​signaling approach to evidence is that it does not consider the possibility of exploiting strategic interaction among competing parties. Endogenous cost signaling can, however, act as a component in a broader model that does. This is the subject of the next section.

9.2.5 Approach 4: Correlated Private Information The fourth approach to legal evidence is “correlated private information.” Like endogenous cost signaling it also appears in the law and economics literature. The basic idea of correlated private information models is to rely on one party’s information to set another party’s payoffs rather than her own. The advantage is that the reporter has no positive incentive to falsify or omit: her own payoffs are not affected by her report. A necessary condition is that the reporter’s information is sufficiently detailed with regard to what is relevant about this other individual’s situation or actions. No doubt an ancient notion, correlated private information has been studied formally in the mechanism design sub-​literature of game theory, most notably in the context of

 

Law and Economics of Evidence    189 auction design. Examples include Crémer and McLean (1985, 1988) and Hermalin and Katz (1991). Sanchirico (2000) studies the use of correlated private information in the context of legal evidence. Sanchirico’s model combines correlated private information with endogenous cost signaling (as discussed in section 9.2.4). With regard to the three dimensions mentioned in the introduction, the model accounts for the strategic interaction between interested parties and the factfinder—​including incentives to omit and lie; regards evidence production instrumentally as a device for incentivizing primary activity behavior; and analyzes the possibility of exploiting the strategic interaction among interested parties. Sanchirico (2000) uses this model to identify and study a trade-​off between the two devices of correlated private information and endogenous cost evidence. The paper suggests that historical shifts in this trade-​off may lend some insight into the historical transformation of evidentiary process. The key trade-​off between these devices derives from the differing nature of the costs. The cost of endogenous cost signaling is, as the phrase suggests, the cost of the costly signaling. These costs—​or differences therein—​are what give the evidence meaning, but they are after all costs. The costs of exploiting correlated private information are fundamentally different. These costs are related to the number of circumstances that trigger hearings and the number of individuals participating in each hearing. The more often suit occurs, and the greater the number of individuals reporting therein, the more information is revealed and the finer are the incentives that can be created. However, the more costly is the system. If, for instance, suit occurs in only one state of the world and there is only one third-​party participant who makes a report of such state, then the signal received by the court will be effectively binary. It will thus be of limited use in creating an incentive for the defendant to take a middling action—​such as reasonable precaution, as opposed to negligence or extreme caution. On the other hand, if suit occurs in two states of the world, the signal received by the court will be effectively tripartite: no suit, suit under one state, suit under the other. Similarly, if the number of individuals reporting during the suit is increased from one to two, the number of possible reports at trial increases to the number of pairwise combinations of such individuals’ reports. Thus the cost of correlated private information accrues primarily in terms of the “fixed costs” of holding hearings—​as opposed to the “variable costs” of the evidence therein “produced,” which may be attributed to endogenous cost signaling. The key trade-​off is thus as follows: obtaining information from interested parties requires costly evidence production. Relying instead on less interested observers necessitates more frequent hearings and/​or greater attendance at each. Sanchirico (2000: 358–​61) suggests that this trade-​off may have some relevance in explaining broad historical trends in legal process, including the transformation of the Anglo-​American jury from its medieval manifestation as a bank of witness/​investigators to its modern incarnation as a blank-​slate panel.

 

190   Chris William Sanchirico

9.3  Specific Rules of Evidence The first part of this chapter described mathematical/​economic approaches to the general problem of legal evidence. This second part discusses the application of mathematical/​economic modeling to specific evidentiary rules.20

9.3.1 The Burden(s) of Proof Several papers have examined the burden of proof in its various legal manifestations. Most of this research has focused on interpreting the relevant legal doctrine in terms of the mathematical/​economic theories of decision-​making under uncertainty and incentive setting. This has turned out to be surprisingly difficult despite the centrality of proof burdens in actual process. The “burden of persuasion,” one species of proof burden, usually lies with the party instigating the claim. In civil cases, for example, this burden requires that the plaintiff, in order to win on an issue, must convince the factfinder with a “preponderance of the evidence,” which is usually taken to mean that the evidence must leave the factfinder thinking that the plaintiff ’s assertion regarding the issue is “more likely than not” to be true. Interestingly, there is generally no sliding scale: plaintiff ’s litigation award is not adjusted for the factfinder’s degree of certainty. Kaye (1982) provides the classic justification for this all-​or-​nothing 50 percent rule. He shows that such a rule minimizes trial error if the two possible errors, wrongful liability and wrongful exoneration, are equally weighted. Sanchirico (1997) suggests that the persuasion burden plays a role in filtering out, at the filing stage, cases with inconclusive evidence. The “burden of production” is a different kind of proof burden in legal doctrine. Roughly speaking, when a party has the production burden with regard to a particular issue, that party loses, even if her opponent remains silent, if she fails to come forward with evidence that would support her position in the mind of a reasonable factfinder. Hay and Spier (1997) argue that the production burden should rest on the party least likely to have a valid claim in the class of cases under consideration. As discussed in section 9.2.4.2.4, Sanchirico (2008) finds some justification for allocating the production burden to the opponent of the incentive target—​a regularity in actual law, as when the plaintiff has the production burden with respect to the defendant’s negligence.

20  Despite the fact that they concern evidentiary rules, not all of the models discussed in the section would strictly qualify as “evidentiary” per the definition in section 9.2.1.

 

Law and Economics of Evidence    191

9.3.2 Detection Avoidance and Sanctions for Evidentiary Malfeasance Evidentiary misconduct—​such as lying, falsifying, forging, refusing to produce requested evidence, or destroying evidence—​is often punished under the law.21 Remedies may be procedural—​as when facts unfavorable to the perpetrator are taken as given, or the jury is instructed to assume that withheld evidence would weaken the perpetrator’s case. Remedies may also involve fines or imprisonment, as when the behavior rises to the level of perjury or obstruction of justice.

9.3.2.1 Models without Sanctions for Detection Avoidance Several papers study detection avoidance activities in the context of Becker’s (1968) classic enforcement model. Such papers do not consider the possibility of sanctioning detection avoidance, but rather focus on how accounting for detection avoidance activity alters the prescriptions of Becker’s basic model. Malik (1990) and Snyder (1990) recognize that raising the sanction on the underlying activity inspires the expenditure of additional private resources on detection avoidance. Malik, whose central example is expenditure on radar detectors, argues that this dynamic effectively adds an additional component to the cost of sanctioning undesirable behavior. This, he shows, adds (yet another) qualification to Becker’s (1968) central prescription—​crafted in a model without detection avoidance—​that it is generally best to generate any desired level of deterrence with large monetary fines and small detection probabilities. Several subsequent papers then provide conditions under which Becker’s prescription is restored. Innes (2001) shows that Becker’s prescription is generally valid—​even in the presence of costly detection avoidance—​if the government invites violators to self-​report and charges them a self-​reporting sanction equal to the effective expected sanction for non-​self-​reporting. Langlais (2008) considers the case in which public detection effort and private detection avoidance effort are “strategic complements”: that is, when the authorities increase the level of detection, violators are inspired to increase the level of their detection avoidance. In this situation the simultaneous reduction in detection effort that accompanies the increased sanction in Becker’s argument inspires violators to reduce detection avoidance. If the reduction in detection avoidance from this source exceeds the increase in detection avoidance caused, per Malik (1990) by the increase in the fine, then the beneficial character of Becker’s maneuver is restored. 21  A more detailed account of the literature on detection avoidance appears in Sanchirico (2010b). For an account of the legal doctrine and empirical evidence surrounding evidentiary misconduct, see Sanchirico (2004b).

 

192   Chris William Sanchirico Langlais (2009) makes the point that not all types of avoidance activities are the same, and that some do not lead to the effects highlighted by Malik (1990). In Langlais’ (2009) model violators who are caught must disgorge their apparent illicit benefits, but may choose to expend effort hiding such benefits. Increasing the separate, punitive portion of the fine—​ which is not a function of illicit benefit—​does not induce additional effort hiding benefits. Accordingly, Langlais finds, in his central case, that the punitive fine should be maximal. Tabbach (2010) studies the case in which the social cost of imposing ex post punishment exceeds the punishment itself, which he argues might occur when the punishment is imprisonment. In Malik (1990), by contrast, the social cost of imposing punishment is zero. Tabbach specifically considers the subcase in which the social cost of imposing sanctions is linear in the level of the sanction with a coefficient greater than 1. In this case, Becker’s prescription to maximize the sanction is restored despite detection avoidance costs. Nussim and Tabbach (2009) focus on a different issue. They argue that in a model with detection avoidance, increasing the underlying sanction can actually induce an increase in the level of underlying violations due to responsive increases in detection avoidance.

9.3.2.2 Sanctioning Detection Avoidance and Evidentiary Misconduct As has been noted, a central question in the design of evidentiary process is whether and to what extent sanctioning evidentiary misconduct is an efficient and effective way to deal with incentives to engage in such misconduct. Sanchirico (2006, 2010b) describes one reason why it may not be. He emphasizes the recursive nature of sanctioning detection avoidance—​the fact that detection of detection avoidance can itself be avoided. Because of this recursivity, sanctioning detection avoidance has contradictory effects on detection avoidance activity: for example, sanctioning “first-​order” detection avoidance discourages such activity but at the same time encourages “second-​order” detection avoidance activity, which is to say, efforts to avoid detection of first-​order detection avoidance. Sanchirico shows that properly accounting for this recursivity would likely require—​impossibly—​sanctioning higher orders of detection avoidance with ever-​ increasing severity. Uniform sanctioning across orders of detection avoidance might well increase detection avoidance all told. Sanchirico suggests that this fundamental difficulty points toward the relative efficacy of dealing with lying and other forms of evidentiary misconduct within the deep structure of evidentiary process, what he refers to as the “technological approach.” This approach focuses on reducing the productivity of avoidance activities for violators. One way to do this is to exploit cognitive limits, à la Sanchirico (2004a). Another way is to exploit the fragility of cooperation among law violators, whose agreements regarding illicit behavior would not be legally enforceable.

9.3.3 Character Evidence Though many exceptions apply, evidentiary rules generally prohibit a party from attempting to prove that an individual acted in a particular way on a particular occasion

 

Law and Economics of Evidence    193 by presenting evidence of the individual’s conforming past acts or general propensity to take such actions. The puzzle is that such “character evidence” is usually informative in this respect (Davies, 1991). Schrag and Scotchmer (1994) advocate admitting character evidence as a way to overcome endemic jury leniency. Their model has two key features. First, the optimal threshold quantum of evidence for guilt is systematically greater when that threshold is chosen with the object of minimizing trial error, as compared to when it is chosen with the object of maximizing deterrence. This feature, which Schrag and Scotchmer derive rather than assume, applies to all kinds of evidence, not just evidence of conduct, and not just character evidence. Second, in Schrag and Scotchmer’s model the jury itself—​not the court or the rules of procedure—​sets the evidentiary threshold for guilt. Moreover, the jury specifically sets this threshold with the object of minimizing trial error. It follows that juries will set too high a threshold for guilt relative to the objective of maximizing deterrence, and thus be too lenient. Schrag and Scotchmer show that admitting character evidence mitigates this leniency with respect to high-​propensity individuals, but exacerbates it with respect to low-​propensity individuals. The net effect on crime is uncertain. But Schrag and Scotchmer emphasize that if the effect on high-​ propensity individuals is more important, admitting character evidence will increase deterrence. Like Schrag and Scotchmer (1994), Sanchirico (2001) analyzes character evidence in light of the distinction between the social objectives of truth-​finding and primary activity incentives. However, Sanchirico highlights a different dynamic and comes to the opposite conclusion. Sanchirico focuses on the distinction between “trace” evidence and “predictive” evidence—​a distinction that is merely academic when truth-​finding is taken to be the object of trial,22 but fundamental when primary activity incentive setting is the objective. “Trace” evidence such as fingerprints and eyewitness recollections are both informative of conduct and are byproducts of conduct. In contrast, character evidence and other forms of predictive evidence, though they may rationally change one’s assessment of the likelihood that particular conduct occurred, are the same whether or not the conduct did actually occur. If the object is to guess whether the defendant or victim engaged in particular conduct, then both types of evidence would be appropriate. But if the object is to affect whether the defendant or victim engages in the conduct, only trace evidence of that conduct is appropriate. Only trace evidence changes with conduct. Thus, keying penalties and rewards to the production of trace evidence is the only way to make penalties and rewards change with conduct. And making penalties and rewards change with conduct is the only way to create incentives. With this basic framework in place Sanchirico (2001) analyzes the ancillary role of character evidence in “tailoring” incentives to the intensity of individual propensities. He argues that this role is best played in the context of admitting other act evidence for

22 

Uviller (1982) proposes this distinction but, favoring a truth-​seeking approach to trial, argues that it is irrelevant.

 

194   Chris William Sanchirico sentencing and punitive damages and not for proving conduct, an allocation that conforms to current law. Sanchirico (2001) also discusses the liberal admission of character evidence for impeachment of witnesses, arguing that conduct “on the stand” is fundamentally different from the primary conduct that is the target of the substantive law. The law does not merely want to prevent lying on the stand, it also wants to know whether a witness has in fact lied. Both trace and predictive evidence are appropriate in this context. Finally, Sanchirico (2001) suggests that the law’s treatment of character evidence, being more in line with the objective of primary activity incentive setting than of truth-​ finding, may itself be evidence that the actual underlying purpose of evidentiary process bends more toward the former.

9.3.4 Hearsay With many exceptions, evidence that an individual asserted that something happened or that something was in a particular state is generally not admissible as evidence that the thing happened or, respectively, was in such state. The inadmissibility of such “hearsay” evidence is the subject of much academic writing within legal scholarship. The mathematical/​economic treatment of hearsay, however, is relatively scarce. Friedman (1992) provides a game-​theoretic analysis of restrictions on hearsay evidence. He takes account of the effect of prohibiting or allowing hearsay evidence on both parties’ incentives to produce, as a witness at the hearing, the declarant whose out-​ of-​court statement is at issue. Friedman views hearsay admissibility as a matter of allocating the burden of bringing forward the hearsay declarant for live testimony. Because, he argues, hearsay evidence will generally be more probative than prejudicial, courts should ordinarily (but not always) place the burden of producing the declarant on the party objecting to the hearsay evidence.

9.3.5 Privileges Various privileges allow parties to refuse to respond to evidentiary requests without sanction. Most privileges excuse parties from producing or answering questions about certain out-​of-​court communications. Others allow parties to refuse to take the stand. Such privileges are said to have three kinds of purposes: to protect allegedly socially beneficial relationships (such as between attorney and client, physician and patient, psychotherapist and patient, husband and wife, or clergy and penitent), to insure certain Constitutional rights (such as the right against self-​incrimination or freedom of the press), or to protect against disclosure of special kinds of information (such as state secrets, trade secrets, or an informer’s identity).

 

Law and Economics of Evidence    195

9.3.5.1 Attorney–​Client Privilege Attorney–​client privilege allows a party to resist disclosure of communications with her lawyer that were made for the purpose of obtaining legal advice. Shavell (1988), and Kaplow and Shavell (1989, 1990) draw a distinction between legal advice offered to parties planning their primary activity conduct and legal advice offered to parties ex post primary activity conduct, during later litigation. The former, it is argued, is valuable in effecting the incentives that the law is meant to provide, as legal advice may better inform parties of the consequences of their choices. Legal advice during litigation, however, is regarded as problematic because it may well lower the quantum of information reaching the factfinder. Accordingly, the authors propose, among other things, restricting confidentiality with respect to advice provided during litigation. Bundy and Elhauge (1991, 1993) question a number of Kaplow and Shavell’s conclusions. Bundy and Elhauge (1991), for example, argue that even litigation advice is likely to be socially beneficial. Litigation advice, they say, helps parties to identify information that is relevant and favorable, and thus can, in this respect, increase the information that a party provides. Moreover, even if litigation advice prevents a party from revealing information unfavorable to her case, the party’s opponent, if also in possession of the information, will present this evidence anyway. Thus, the suppressive effects of litigation advice are mitigated by party competition, while the information-​increasing effects are unimpeded. Correspondingly, the authors find that litigation advice is likely to have negative effects when only one party possesses the information and is able to avoid presenting it.

9.3.5.2 The Privilege against Self-​Incrimination The privilege against self-​incrimination—​often referred to as the “right to silence”—​ allows criminal defendants to refuse to testify or submit to questioning. It also prevents judges and prosecutors, in criminal cases, from encouraging adverse inferences by the jury based on the defendant’s silence. Bentham (1825) and others disfavor the right to silence arguing that it benefits only the guilty—​the innocent having nothing to hide—​and thus distorts trial outcomes. Several recent papers, including Seidmann and Stein (2000), Seidmann (2005), Mialon (2005), and Leshem (2010), use game-​theoretical modeling to question this classical view. In Seidmann’s (2005) model, when a right to silence is allowed, waiving the right may act as a signal of innocence. Thus, contrary to Bentham, the right to silence may benefit the innocent even if the innocent do not themselves exercise it. Leshem (2010) provides a different reason that the right to silence may benefit the innocent. He allows for the possibility that an innocent suspect’s statements might still be contradicted at trial, and no exonerating evidence may come forward. Anticipating this, the innocent might choose to remain silent if given the opportunity to do so with impunity.

 

196   Chris William Sanchirico

9.4 Conclusion The first part of this chapter discussed the four main mathematical/​economic approaches to the general problem of legal evidence:  the pure probabilistic approach, omission models, endogenous cost signaling, and correlated private information. The discussion in each case included applications and criticisms. The approaches were compared along three dimensions: their account of strategic interaction as between interested parties and the factfinder; the extent to which they regarded fact-​finding as instrumental as opposed to inherently valuable; and their account of the strategic interaction among litigating parties. The second part of the chapter considered the application of mathematical modeling to several specific evidentiary rules: burdens of proof, detection avoidance and sanctions for evidentiary misconduct, character evidence, hearsay, and privileges.

References Allais, M. (1953) “Le Comportement de l’Homme Rationnel devant le Risque: Critique des Postulats et Axiomes de l’Ecole Américaine,” Econometrica, 21: 503–​46. Allen, R. (1986) “A Reconceptualization of Civil Trials,” Boston University Law Review, 66: 401–​37. Allen, R. and Jehl, S. A. (2003) “Burdens of Persuasion in Civil Cases:  Algorithms v. Explanations,” Michigan State Law Review, 2003: 893–​944. Baron, J. and Hershey, J. C. (1988) “Outcome Bias in Decision Evaluation,” Journal of Personality and Social Psychology, 54: 569–​70. Becker, G. (1968) “Crime and Punishment:  An Economic Approach,” Journal of Political Economy, 76: 169–​217. Bentham J. (1825) “A Treatise on Judicial Evidence,” in M. Dumont (ed.) The Works of Jeremy Bentham. New York: Russell and Russell. Bernardo, A. E., Talley, E., and Welch, I. (2000) “A Theory of Legal Presumptions,” Journal of Law, Economics, and Organization, 16: 1–​49. Boston University Law Review (1986) “Symposium: Probability and Inference in the Law of Evidence,” 66: 377–​952. Braeutigam, R., Owen, B., and Panzar, J. (1984) “An Economic Analysis of Alternative Fee Shifting Systems,” Law and Contemporary Problems, 47: 173–​85. Bull, J. (2008) “Costly Evidence Production and the Limits of Verifiability,” The B.E. Journal of Theoretical Economics, 8: Article 18. Bull, J. and Watson, J. (2004) “Evidence Disclosure and Verifiability,” Journal of Economic Theory, 118: 1–​31. Bull, J. and Watson, J. (2007) “Hard Evidence and Mechanism Design,” Games and Economic Behavior, 58: 75–​93. Bundy, S. M. and Elhauge, E. R. (1991) “Do Lawyers Improve the Adversary System? A General Theory of Litigation Advice and its Regulation,” California Law Review, 79: 313–​420. Bundy, S. M. and Elhauge, E. R. (1993) “Knowledge about Legal Sanctions,” Michigan Law Review, 92: 261–​335.

 

Law and Economics of Evidence    197 Cooter, R. and Emmons, W. (2003) “Truth-​revealing Mechanisms for Courts,” Journal of Institutional and Theoretical Economics, 159: 259–​79. Cooter, R. and Emmons, W. (2004) “Truth-​bonding and Other Truth-​revealing Mechanisms for Courts,” European Journal of Law and Economics, 17: 307–​27. Corchón, L. C. (2007) “The Theory of Contests:  A  Survey,” Review of Economic Design, 11: 69–​100. Crémer, J. and McLean, R. P. (1985) “Optimal Selling Strategies under Uncertainty for a Discriminating Monopolist when Demands are Interdependent,” Econometrica, 53: 345–​61. Crémer, J. and McLean, R. P. (1988) “Full Extraction of the Surplus in Bayesian and Dominant Strategy Auctions,” Econometrica, 56: 1247–​57. Daughety, A. F. and Reinganum, J. F. (2000) “On the Economics of Trials: Adversarial Process, Evidence, and Equilibrium Bias,” Journal of Law, Economics, and Organization, 16: 365–​94. Davies, S. M. (1991) “Evidence of Character to Prove Conduct: A Reassessment of Relevancy,” Criminal Law Bulletin, 27: 504–​37. Demougin, D. and Fluet, C. (2005) “Deterrence versus Judicial Error: A Comparative View of Standards of Proof,” Journal of Institutional and Theoretical Economics, 161: 193–​206. Dewatripont, M. and Tirole, J. (1999) “Advocates,” Journal of Political Economy, 107: 1–​39. Ellsberg, D. (1961) “Risk, Ambiguity, and the Savage Axioms,” Quarterly Journal of Economics, 75: 643–​69. Farrell, J. (1986) “Voluntary Disclosure:  Robustness of the Unraveling Result, and Comments on its Importance,” in R. E. Grieson (ed.) Antitrust and Regulation. Lexington, MA: Lexington Books. Finkelstein, M. and Fairly, W. (1970) “A Bayesian Approach to Identification Evidence,” Harvard Law Review, 83: 489–​517. Fischhoff, B. (1982) “For Those Condemned to Study the Past:  Heuristics and Biases in Hindsight,” in D. Kahneman, P. Slovic, and A. Tversky (eds) Judgment under Uncertainty: Heuristics and Biases. Cambridge, UK and New York: Cambridge University Press, 335–​53. Friedman, R. D. (1992) “Toward a Partial Economic, Game-​Theoretic Analysis of Hearsay,” Minnesota Law Review, 76: 723–​96. Froeb, L. M. and Kobayashi, B. H. (1996) “Naive, Biased, yet Bayesian: Can Juries Interpret Selectively Produced Evidence?” Journal of Law, Economics, and Organization, 12: 257–​76. Froeb, L. M. and Kobayashi, B. H. (2001) “Evidence Production in Adversarial vs. Inquisitorial Regimes,” Economics Letters, 70: 267–​72. Gould, J. P. (1973) “The Economics of Legal Conflicts,” Journal of Legal Studies 2: 279–​300. Grossman, S. J. (1981) “The Informational Role of Warranties and Private Disclosure about Product Quality,” Journal of Law and Economics, 24: 461–​89. Guthrie, C., Rachlinski, J. J., and Wistrich, A. J. (2007) “Blinking on the Bench: How Judges Decide Cases,” Cornell Law Review, 93: 1–​43. Hart, O. and Moore, J. (1988) “Incomplete Contracts and Renegotiation,” Econometrica, 56: 755–​85. Hay, B. L. and Spier, K. E. (1997) “Burdens of Proof in Civil Litigation:  An Economic Perspective,” Journal of Legal Studies, 26: 413–​31. Hermalin, B. and Katz, M. (1991) “Moral Hazard and Verifiability: The Effects of Renegotiation in Agency,” Econometrica, 59: 1735–​53. Hershey, J. C. and Baron, J. (1992) “Judgment by Outcomes:  When Is It Justified?” Organizational Behavior and Human Decision Processes, 53: 89–​93.

 

198   Chris William Sanchirico Hershey, J. C. and Baron, J. (1995) “Judgment by Outcomes: When Is It Warranted?” Organizational Behavior and Human Decision Processes, 62: 126–​7. Innes, R. (2001) “Violator Avoidance Activities and Self-​ reporting in Optimal Law Enforcement,” Journal of Law, Economics, and Organization, 17: 239–​56. Kahneman, D. and Tversky, A. (1973) “On the Psychology of Prediction,” Psychological Review, 80: 237–​51. Kamin, K. A. and Rachlinski, J. J. (1995) “Ex Post ≠ Ex Ante:  Determining Liability in Hindsight,” Law and Human Behavior, 19: 89–​104. Kaplow, L. and Shavell, S. (1989) “Legal Advice about Information to Present in Litigation: Its Effects and Social Desirability,” Harvard Law Review, 102: 565–​615. Kaplow, L. and Shavell, S. (1990) “Legal Advice about Acts already Committed,” International Review of Law and Economics, 10: 149–​59. Kaplow, L. and Shavell, S. (1996) “Accuracy in the Assessment of Damages,” Journal of Law and Economics, 39: 191–​210. Katz, A. (1988) “Judicial Decisionmaking and Litigation Expenditure,” International Review of Law and Economics, 8: 127–​43. Kaye, D. (1979) “The Paradox of the Gatecrasher and Other Stories,” Arizona State Law Journal, 1979: 101–​9. Kaye, D. (1982) “The Limits of the Preponderance of the Evidence Standard: Justifiably Naked Statistical Evidence and Multiple Causation,” American Bar Foundation Research Journal, 1982: 487–​516. Landsman, S. (1984) “The Adversary System:  A  Description and Defense” (AEI Studies in Legal Policy). American Enterprise Institute for Public Policy Research, Washington, DC. Langlais, E. (2008) “Detection Avoidance and Deterrence:  Some Paradoxical Arithmetic,” Journal of Pubic Economic Theory, 10: 371–​82. Langlais, E. (2009) “On the Ambiguous Effects of Repression,” Annals of Economics and Statistics, 93/​94: 349–​62. Lempert, R. O. (1977) “Modeling Relevance,” Michigan Law Review, 75: 1021–​57. Lempert, R. O., Gross, S. R., and Liebman, J. S. (2000) A Modern Approach to Evidence: Text, Problems, Transcripts and Cases (3rd ed.). St. Paul, MN: West Publishing. Lempert, R. O. and Saltzburg, S. A. (1982) A Modern Approach to Evidence: Text, Problems, Transcripts and Cases (2nd ed.). St. Paul, MN: West Publishing. Leshem, S. (2010) “The Benefits of a Right to Silence for the Innocent,” RAND Journal of Economics, 41: 398–​416. Levmore, S. (2001) “Conjunction and Aggregation,” Michigan Law Review, 99: 723–​56. Lipman, B. and Seppi, D. (1995) “Robust Inference in Communication Games with Partial Provability,” Journal of Economic Theory, 66: 370–​405. Malik, A. S. (1990) “Avoidance, Screening and Optimum Enforcement,” RAND Journal of Economics, 21: 341–​53. Mialon, H. (2005) “An Economic Theory of the Fifth Amendment,” RAND Journal of Economics, 36: 833–​48. Milgrom, P. R. (1981) “Good News and Bad News: Representation Theorems and Applications,” Bell Journal of Economics, 12: 380–​91. Milgrom, P. R. and Roberts, J. (1986) “Relying on the Information of Interested Parties,” RAND Journal of Economics, 17: 18–​32. Mirrlees, J. A. (1971) “An Exploration in the Theory of Optimum Income Taxation,” Review of Economic Studies, 38: 175–​208.

 

Law and Economics of Evidence    199 Nisbett, R. E., Borgida, E., Crandall, R., and Reed, H. (1976) “Popular Induction: Information Is Not always Informative,” in J. S. Carroll and J. W. Payne (eds) Cognition and Social Behavior. Hillsdale, NJ: Erlbaum Associates. Nussim, J. and Tabbach, A. D. (2009) “Deterrence and Avoidance,” International Review of Law and Economics, 29: 314–​23. Okuno-​ Fujiwara, M., Postlewaite, A., and Suzumura, K. (1990) “Strategic Information Revelation,” Review of Economic Studies, 57: 25–​47. Posner, R. A. (1973) “An Economic Approach to Legal Procedure and Judicial Administration,” Journal of Legal Studies, 2: 399–​458. Rubinfeld, D. and Sappington, D. (1987) “Efficient Awards and Standards of Proof in Judicial Proceedings,” RAND Journal of Economics, 18: 308–​15. Sanchirico, C. W. (1995) “Enforcement by Hearing:  How the Civil Law Sets Incentives,” Columbia Economics Department Discussion Paper, No 95-​9603. Sanchirico, C. W. (1997) “The Burden of Proof in Civil Litigation:  A  Simple Model of Mechanism Design,” International Review of Law and Economics, 17: 431–​47. Sanchirico, C. W. (2000) “Games, Information and Evidence Production: With Application to English Legal History,” American Law and Economics Review, 2: 342–​80. Sanchirico, C. W. (2001a) “Character Evidence and the Object of Trial,” Columbia Law Review, 101: 1227–​311. Sanchirico, C. W. (2001b) “Relying on the Information of Interested—​ and Potentially Dishonest—​Parties,” American Law and Economics Review, 3: 320–​57. Sanchirico, C. W. (2004a) “Evidence, Procedure, and the Upside of Cognitive Error,” Stanford Law Review, 57: 291–​365. Sanchirico, C. W. (2004b) “Evidence Tampering,” Duke Law Journal, 53: 1215–​336. Sanchirico, C. W. (2004c) “Finding Error,” Michigan State Law Review, 2003: 1189–​203. Sanchirico, C. W. (2005) “Harnessing Adversarial Process:  Proof Burdens, Affirmative Defenses, and the Complementarity Principle,” University of Pennsylvania Institute for Law and Economics Research Paper Series, No. 05-​01. Sanchirico, C. W. (2006) “Detection Avoidance,” New York University Law Review, 81: 1331–​99. Sanchirico, C. W. (2008) “A Primary-​Activity Approach to Proof Burdens,” Journal of Legal Studies, 37: 273–​313. Sanchirico, C. W. (2010a) “Models of Evidence:  Survey and Assessment,” University of Pennsylvania Institute for Law & Economics Research Paper, http://​www.ssrn.com. Sanchirico, C. W. (2010b) “Detection Avoidance and Enforcement Theory:  Survey and Assessment,” University of Pennsylvania Institute for Law & Economics Research Paper Series No. 10–​29, http://​www.ssrn.com. Sanchirico, C. W. and Triantis, G. (2008) “Evidentiary Arbitrage: The Fabrication of Evidence and the Verifiability of Contract Performance,” Journal of Law, Economics, and Organization, 24: 72–​94. Savage, L. J. (1954) The Foundations of Statistics. New York: Wiley (2nd ed. 1972). New York: Dover. Schrag, J., and Scotchmer, S. (1994) “Crime and Prejudice: The Use of Character Evidence in Criminal Trials,” Journal of Law, Economics, and Organization, 10: 319–​42. Schwartz, A. (1992) “Relational Contracts in the Courts:  An Analysis of Incomplete Agreements and Judicial Strategies,” Journal of Legal Studies, 21: 271–​318. Seidmann, D. (2005) “The Effects of the Right to Silence,” Review of Economic Studies, 72: 593–​614.

 

200   Chris William Sanchirico Seidmann, D. and Stein, A. (2000) “The Right to Silence Helps the Innocent:  A  Game-​ Theoretic Analysis of the Fifth Amendment Privilege,” Harvard Law Review, 114: 430–​510. Shavell, S. (1982) “Suit, Settlement, and Trial:  A  Theoretical Analysis under Alternative Methods for the Allocation of Legal Costs,” Journal of Legal Studies, 11: 55–​81. Shavell, S. (1988) “Legal Advice about Contemplated Acts: The Decisions to Obtain Advice, its Social Desirability, and Protection of Confidentiality,” Journal of Legal Studies, 17: 123–​50. Shavell, S. (1989) “Optimal Sanctions and the Incentive to Provide Evidence to Legal Tribunals,” International Review of Law and Economics, 9: 3–​11. Shin, H. S. (1998) “Adversarial and Inquisitorial Procedures in Arbitration,” RAND Journal of Economics, 29: 378–​405. Skaperdas, S. (1996) “Contest Success Functions,” Economic Theory, 7: 283–​90. Snyder, E. A. (1990) “The Effect of Higher Criminal Penalties on Antitrust Enforcement,” Journal of Law and Economics, 33: 439–​62. Sobel, J. (1985) “Disclosure of Evidence and Resolution of Disputes,” in A. Roth (ed.) Game-​ Theoretic Models of Bargaining. Cambridge: Cambridge University Press, 341–​62. Spence, A. M. (1974) Market Signaling. Cambridge, MA: Harvard University Press. Stein, A. (2001) “Of Two Wrongs that Make a Right: Two Paradoxes of the Evidence Law and their Combined Economic Justification,” Texas Law Review, 79: 1199–​234. Tabbach, A. D. (2010) “The Social Desirability of Punishment Avoidance,” Journal of Law, Economics, and Organization, 26: 265–​89. Tribe, L. (1971) “Trial by Mathematics: Precision and Ritual in the Legal Process,” Harvard Law Review, 84: 1329–​93. Uviller, H. R. (1982) “Evidence of Character to Prove Conduct: Illusion, Illogic, and Injustice in the Courtroom,” University of Pennsylvania Law Review, 130: 845–​91.

 

chapter 10

E c onomics of L i t i g at i on Bruce H. Kobayashi

10.1 Introduction This chapter examines the basic model of the law and economics of litigation. Because the Rules of Civil Procedure and the economics of the litigation/​settlement decision are covered in separate chapters of this volume, this chapter will focus on private civil litigation, in particular the litigation value of a lawsuit and the incentives generated for filing a suit. The chapter begins with the simple one-​stage single plaintiff/​single defendant investment model of litigation, and sets out the conditions for filing, default, settlement, and litigation. The analysis then examines the effects of litigation cost-​and fee-​shifting as well as the effects of percentage contingency fee arrangements within the standard one-​stage model. The model is then modified to take into account sequencing and option value. We show how litigation with multiple stages and the revelation of information alter the investment value of litigation, as well as the effects of litigation reform proposals such as fee-​shifting. Finally, the chapter discusses third-​party or external effects. We examine how these additional complications affect the outcome of litigation, the viability of a lawsuit and the predictions of the standard model of litigation. This chapter is not intended to be a comprehensive overview of economic analyses of litigation and civil procedure (see Cooter and Rubinfeld, 1989; Kobayashi and Parker, 2000; Spier, 2007). Rather, the aim of this chapter is to set out the basic mechanics of the law and economics of private civil litigation, and examine how a more robust examination of sequential decision-​making in litigation alters some of the basic predictions of the simple model.

 

202   Bruce H. Kobayashi

10.2  The Basic One-​Stage Model of Litigation 10.2.1 One-​Stage Models and NEV Suits 10.2.1.1 Litigation Payoffs in a Single-​Stage Model We begin the analysis with the standard one-​stage expected-​value model of litigation (Landes, 1971; Gould, 1973). In this model, a risk neutral plaintiff (P) has a claim against a defendant. The plaintiff estimates that the suit will result in a favorable judgment with probability pP. When judgment is entered for the plaintiff, the defendant is ordered to pay the plaintiff an amount J. The plaintiff ’s cost of litigating a case to judgment equals CP.1 The standard model assumes that the plaintiff seeks to maximize the payoff at trial TP equal to the expected judgment net of litigation costs: TP = pP J − CP . (1)



The risk neutral defendant (D) estimates that the probability of a judgment for the plaintiff equals pD, and seeks to minimize his exposure at trial TD equal to the expected judgment plus litigation costs CD:2 TD = pD J + CD (2)



10.2.1.2 The NEV/​PEV Threshold In the standard model of litigation, the condition that a suit has positive expected value (PEV) is used as a threshold condition for filing a lawsuit:3 1 

In this chapter, the costs of litigation are assumed to be exogenous. Thus we abstract away from the issue of endogenous litigation expenditures, including signaling and screening models based on litigation expenditures. For a recent in-​depth review of this literature, see Spier (2007). 2  The defendant is better off paying the judgment (or defaulting) rather than litigating if: T > J , or D C pD > pD* = 1 − D . However, under the simple structure assumed in Figure 10.1, default is not an option. J For a discussion of the strategic use of default, see Hunter (2002: 172–​3). 3  One US Appellate Court suggested use of the NEV condition to determine whether a patent lawsuit could be subject to the sham exception to Noerr Pennington antitrust immunity, thus subjecting the filing of a lawsuit to scrutiny under the US antitrust laws. Posner, writing for the Seventh Circuit, would deny immunity for the pursuit of valid but negative expected value claim: “But we are not prepared to rule that the difficulty of distinguishing lawful from unlawful purpose in litigation between competitors is so acute that such litigation can never be considered an actionable restraint of trade, provided it has some, though perhaps only threadbare, basis in law. Many claims not wholly groundless would never be sued on for their own sake; the stakes, discounted by the probability of winning, would be too low to repay the investment in litigation” (emphasis added) Grip-​Pak Inc. v. Illinois Tool Works, 694 F.2d 466 (1982), cert

 

Economics of Litigation    203

Plaintiff Offers S

Defendant Accepts Offer (S, -S)

Defendant Rejects Offer

Plaintiff Litigates (TP, TD)

Plaintiff Does not File (0, 0)

Figure 10.1  The structure of litigation—​simple one-​stage model.

TP > 0, or C pP > pP* = P J



(3)

Use of the NEV/​PEV threshold as a threshold condition for filing a lawsuit requires some additional assumptions about the structure of litigation. Figure 10.1 illustrates a simple structure that is consistent with use of the NEV/​PEV threshold as a filing threshold. In Figure 10.1, the plaintiff offers a settlement S, which is either accepted or rejected by the defendant. If the offer is rejected, the plaintiff then must choose to file and litigate the case to judgment, or drop the case. The payoff to file and litigate is higher than the payoff from dropping the case when TP > 0. Thus, with the simple structure of litigation illustrated in Figure 10.1, and under the assumptions that: (i) the plaintiff is maximizing the expected judgment net of litigation costs; and (ii) all costs are incurred simultaneously at the trial stage, the plaintiff will rationally choose to file PEV claims, and will not file NEV claims.4

10.2.1.3 Litigation and Settlement The expected litigation payoffs determine the litigants’ threat points at settlement. That is, the plaintiff ’s litigation payoff determines the minimum acceptable settlement, and the defendant’s litigation exposure determines the maximum he will offer. Thus, any feasible settlement S will satisfy: TP ≤ S ≤ TD



(4)

denied, 461 U.S. 958 (1983), overruled by Professional Real Estate Investors, Inc., v. Columbia Pictures 113 S.Ct. 1920 (1993). 4 

See, e.g., Bone (2003), Chapter 1, p. 34. The effects of altering the structure of litigation are examined in section 10.3.

 

204   Bruce H. Kobayashi pP 1 Litigated

0.9 0.8 0.7

Settled

0.6 pP* 0.5 0.4 0.3 0.2 NEV–Not Filed 0.1 0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

pD

Figure 10.2  Filing, litigation, and settlement—​low-​stakes (200).

The absence of a feasible settlement is a sufficient but not necessary condition for litigation, and occurs when: TP > TD , or

(CP + CD ) , or J (C + CD ) pP > pD + P . J

(pP − pD ) >

(5)

Condition (5) is known as the optimism condition for litigation, as the plaintiff ’s estimate of the probability of a judgment for the plaintiff, pP, must be greater than the defendant’s estimate of the same probability, pD. Figure 10.2 depicts in pP, pD space the conditions that give rise to cases being filed, defaulted, settled, and litigated.5 In the figure, J = 200, and CP = CD = 70. Along the 45-​ degree line are cases in which the two sides agree on the probability of a judgment for the plaintiff. For these cases, a mutually beneficial settlement exists, as condition (5) is never satisfied. The absence of a bargaining range requires an optimistic difference pP − pD >. 7. Thus, condition (5) is satisfied when the parties divergent and optimistic expectations 5 

This useful diagram was introduced by Gelbach (2016) to examine the selection of cases for litigation.

 

Economics of Litigation    205 pP 1 0.9 0.8

Litigated

0.7 0.6 pP*

Settled

0.5 0.4 0.3 0.2 0.1 0

NEV–Not Filed 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

pD

Figure 10.3  Filing, litigation, and settlement—​high-​stakes (600).

lie in the “Litigated” triangle in the upper left hand corner of Figure 10.2. The NEV/​PEV threshold occurs at pP* = .35. The optimism model of litigation predicts that higher stakes relative to costs increase both the probability a suit will be filed and the probability that filed suits will be litigated. Figure 10.3 shows the conditions when the size of the judgment J is tripled to 600, holding all of the other parameters constant. The higher stakes relative to costs of litigation increases the expected value of the suit, and shifts the NEV/​PEV threshold to pP* = .1167. The higher stakes also moves the litigation/​settlement line toward the 45-​degree line, increasing the set of cases that will be litigated due to the absence of a bargaining range relative to the low-​stakes case.

10.2.2 Effect of Fee-​Shifting The standard model of litigation assumes that each side bears its own costs and fees. However, the award of costs and/​or fees to the prevailing party is provided for in numerous statutes, and is a popular litigation reform proposal. This section examines the effects of fee-​shifting on the incentives of the parties to file, litigate, and settle.6 6  Other analyses have examined the effect of fee-​shifting on broader measures of activity, such as compliance with legal standards. See, e.g., Hylton (1993).

 

206   Bruce H. Kobayashi

10.2.2.1 Two-​Way Fee-​ and Cost-​Shifting Use of a system of “loser pays” or a two-​way cost-​and fee-​shifting system is a popular litigation reform proposal. The basic intuition for its use is that use of a loser-​pays system raises (lowers) the expected cost of litigation to parties whose probability of prevailing is less than (greater than) .5. Thus, use of such cost-​and fee-​shifting systems are designed to deter the filing of lawsuits with low probabilities of obtaining a judgment, and to encourage lawsuits with high probabilities of obtaining a judgment (Shavell, 1982; Polinsky and Rubinfeld, 1998). To see this, note that under two-​way cost-​and fee-​shifting:

TP = pP J − (1 − pP )(CP + CD )

(1′)

Comparing (1′) to (1), the expected judgment is unaffected, and the plaintiff ’s expected litigation costs are greater under two-​way fee-​and cost-​shifting when (1 − pP )(CP + CD ) > CP, CD or when pP < . If litigation costs are symmetric, so that CP = CD, the plaintiff ’s CP + CD expected litigation costs rise (fall) for pP < (>).5. The NEV/​PEV threshold increases, and is given by: CP + CD pP > pP* = (1′*) J + CP + CD Similarly, the defendant’s litigation exposure under two-​way fee-​and cost-​shifting is:

TD = pD ( J + CD + CP ). (2′)

Comparing (2′) and (2), the defendant’s expected litigation costs are greater under two-​way CD fee-​and cost-​shifting when pD < . If litigation costs are symmetric, two-​way fee-​ CP + CD and cost-​shifting increases (decreases) the expected litigation costs of the defendant when pD > .5. Finally, the litigation settlement condition is given by:





( pP − pD ) >

(CP + CD ) ( J + CP + CD ), or

pP > pD +

(5′)

(CP + CD ) ( J + C P + C D )

Comparing (5′) to (5), the effect of two-​way fee-​and cost-​shifting effectively increases the stakes relative to the costs of litigation. Specifically, with two-​way cost-​and fee-​shifting, the effective stakes for the litigants now include both the judgment J and the sum of the parties’ litigation costs. Figure 10.4 shows the effect of two-​way fee-​and cost-​shifting using the same parameters as Figure 10.2. Because the expected litigation costs are increased for

 

Economics of Litigation    207 pP 1

Litigated

0.9 0.8 0.7

Settled

0.6 pP* 0.5 0.4 0.3 0.2

NEV–Not Filed

0.1 0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

pD

Figure 10.4  The effect of two-​way cost-​and fee-​shifting.

suits where pP < .5, two way cost-​and fee-​shifting increases pP* to .411 and expands the set of cases that result in NEV suits. Under the assumption that NEV suits will not be filed, two-​way fee-​shifting will reduce the number of suits filed. However, two-​way fee-​shifting increases the set of filed cases that are litigated. Similar to the effect depicted in Figure 10.3, the increase in the effective stakes shifts the litigation settlement line toward the 45-​degree line resulting in an expanded set of cases that result in litigation due to the absence of a bargaining range.

10.2.2.2 One-​Way and Rule 68 Many existing fee-​shifting statutes operate asymmetrically. For example, the antitrust laws7 and civil rights laws8 feature statutory fee-​shifting that shift fees and costs to prevailing plaintiffs. In contrast to the proposals for two-​way fee-​shifting, which are intended to deter low probability lawsuits, the intended effect of one-​way fee-​shifting to prevailing plaintiffs is to yield stronger incentives for plaintiffs and their lawyers to file lawsuits with low net value. Ironically, even fee-​shifting systems designed as two-​ way cost-​and fee-​shifting systems may operate as de facto one-​way systems because of 7 

15 USC § 15(a). 42 USC § 1988(b), Hensley v. Eckerhart, 461 U.S. 424 (interpreting 42 USC § 1988(b) as creating a presumption in favor of awarding fees to a prevailing plaintiff, but not allowing a prevailing defendant to get his attorneys’ fees reimbursed unless the suit was frivolous); 42 USC § 2000(k). 8 

 

208   Bruce H. Kobayashi systematic wealth constraints, and encourage rather than discourage plaintiffs on the margin.9 Thus, rather than operating as designed as a mechanism to deter low value lawsuits, the actual effect may be to serve as a litigation subsidy. Under one-​way cost-​and fee-​shifting, the litigation payoffs are:

TP = pP J − (1 − pP ) CP

(1′′)



TD = pD ( J + CP ) +� CD

(2′′)

Figure 10.5 illustrates the effect of one-​way cost-​and fee-​shifting in favor of prevailing plaintiffs, under the parameters set out in Figure 10.1. The NEV/​PEV threshold probability is given by:

pP* =



CP J + CP

(3′′)

Unsurprisingly, given the goals of one-​way cost-​and fee-​shifting, the pro-​plaintiff one-​way cost-​and fee-​shifting lowers the NEV threshold and induces litigation on the margin. One-​way cost-​and fee-​shifting also expands the set of litigated claims, but not as much as two-​way cost-​and fee-​shifting. In the latter case, the effective states were equal to the judgment plus both litigants’ costs and fees. In the case of one-​way cost-​and fee-​ shifting, the litigants now face effective stakes equal to the judgment plus the plaintiff ’s costs and fees. The litigation/​settlement threshold is given by:

pP > pD +



(CP + CD ) ( J + CP )

(5′′)

It is also possible to have one-​way or asymmetric cost-​or fee-​shifting rules that benefit defendants. One example is Federal Rule of Civil Procedure (FRCP) 68. This rule shifts the defendant’s post-​offer costs when the defendant makes a Rule 68 offer of judgment, and the judgment for the plaintiff that is less favorable than the rejected offer. Notably, Rule 68 shifts only post-​offer costs, and is designed as a mechanism to encourage settlement rather than as a simple litigation tax/​subsidy (but see Miller, 1986). Thus pre-​ offer costs and pre-​and post-​offer attorney’s fees are not shifted through the rule.10 9 

See Snyder and Hughes (1990, 1995) examining the use of two-​way fee-​and cost-​shifting for Florida medical malpractice claims between 1980 and 1985. Shepherd (2010) examines a broader set of fee-​ shifting and offer of judgment policies in Florida over a longer time period. 10  Under FRCP 54(d), “costs other than attorneys’ fees shall be allowed as of course to the prevailing party unless the court otherwise directs.” Thus, the costs, which are often a small fraction of litigation costs, follow the “English Rule” in US federal civil litigation. Attorneys’ fees are distinct from costs under the rule. For an example of the interplay between Rule 68 and one-​way fee-​shifting statutes that shift

 

Economics of Litigation    209 pP 1

Litigated

0.9 0.8 0.7

Settled

0.6 pP* 0.5 0.4 0.3 0.2

NEV–Not Filed

0.1 0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

pD

Figure 10.5  The effect of one-​way fee-​shifting.

Moreover, it applies only when the party favored by the cost-​shifting rule (the defendant) loses.11

10.2.3 The Effect of Contingency Fees The standard model of litigation can be modified to address the use of legal professionals and the generation of lawyer–​client agency costs (Miller, 1967). For example, suppose the plaintiff hires a lawyer using a standard contingency-​fee arrangement with a contingency percentage α = .3. Under such an arrangement, the lawyer bears the costs of litigation, and is given 30 percent of any judgment or settlement as compensation. The client keeps the remaining 70 percent of any judgment or settlement. The litigation payoff for the defendant is unchanged. The litigation payoffs for the contingent plaintiff and his lawyer are given by:

TP = (1 − α ) pP J

(1cc)

costs including attorneys’ fees to prevailing plaintiffs, see Marek v. Chesny, 720 F.2d 474 (1985), Rev’d 473 U.S. 1 (1985). 11  Delta Airlines v. August, 450 U.S. 346 (1981) (finding that Rule 68 offer was not triggered when defendant prevailed).

 

210   Bruce H. Kobayashi

TCL = α pP J − CP



(1cl)

Under the assumption that the lawyer bears all of the costs of litigation (including the opportunity cost of his time), the NEV/​PEV thresholds for the lawyer and client are:

pP* = 0



* pCL =



(3cc)

CP αJ

(3cl)

Under the assumption that the client does not bear any of the costs of litigation, the client will rationally pursue all suits. However, because the contingent lawyer only receives 30 percent of any judgment or settlement, his threshold for agreeing to represent a client is higher than the NEV/​PEV threshold of a vertically integrated client/​lawyer. Note that when J = 200, the lawsuit is an NEV suit for the contingent lawyer even when recovery is certain. Under these conditions, the lawyer will recover 60, but must spend 70 to obtain the judgment. Figure 10.6 shows the effect of a standard percentage contingency contract on the incentives to file and litigate a lawsuit. Because the conditions depicted in Figure 10.2 will generate a corner solution for the lawyer, we use the underlying parameters from Figure 10.3 (i.e., where J = 600) to illustrate the effects of

pP

1

Contingent Lawyer Litigation/ Settlement Threshold

0.9 0.8 0.7

Contingent Client Litigation/ Settlement Threshold

0.6 0.5 pCL*

0.4 0.3 0.2 0.1

pP* = 0

0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Figure 10.6  The effect of a contingency-fee arrangement—​high stakes.

0.9

1

pD

 

Economics of Litigation    211 contingency contracts. While contingent clients will rationally pursue all cases where recovery is possible, contingent lawyers will only represent clients when pCL* > .39 (Miceli, 1994). There is also a divergence in the lawyer’s and client’s decisions to litigate or settle the case. The litigation and settlement condition for the contingent client and the contingent lawyer are given by:





pP > pD +

pP > pD +

(CD )

(5cc)

J

(CP / α + C D ) J



(5cl)

The contingent client does not bear the marginal costs of litigating, and is therefore more willing to choose to litigate rather than settle. Likewise, the contingent lawyer bears the marginal costs of litigation, yet only receives a fraction (α) of the marginal benefits from litigating. This creates an agency cost with respect to the decision to settle (Miller, 1986). Because the model in this chapter treats lawyer effort and litigation expenditures as exogenous, the model does not address lawyer–​client agency costs involving the level of lawyer effort. For analyses of this aspect of the problem, see Smith and Cox (1985), Hay (1997), Polinsky and Rubinfeld (2003), Wickelgren (2004).

10.3 Multistage Models 10.3.1 Effect on the Decision to File: Symmetric Costs and Credible NEV Suits In this section, changes to the structure of litigation and their effect on the predictions/​ implications of the one-​stage model of litigation are considered. These models will show conditions under which a plaintiff will choose to rationally file a lawsuit that has NEV if litigated to judgment. The analysis focuses on the incentives of private litigants in multistage litigation and does not examine broader issues relating to the broader effects of litigation or the divergence between the private and social value of litigation (Kaplow, 2013). We begin by altering the simple one-​stage model to include multiple stages. To begin the analysis, consider the case of a negative expected value suit, where E(J) = 100, and CP = CD = 210. Under the simple litigation structure set out in Figure 10.1, the threat to litigate such a case is not credible (Nalebuff). The plaintiff will not rationally choose to litigate the case, as such a decision will result in a loss of –​110 compared to the decision to drop and not file the case, which results in a net payoff of zero. The defendant,

 

212   Bruce H. Kobayashi

Plaintiff Offers S =100 P = 0, D = 0

Defendant Rejects Offer P = 0, D = 0

Plaintiff Files & Litigates Case P = –110, D = –310

Defendant Accepts Offer P = 100, D = –100

Plaintiff Does Not File P = 0, D = 0

Figure 10.7  One-​stage negative value lawsuit.

anticipating that the plaintiff will not file the case, will reject any offer of settlement S made at the outset. The payoffs for this example are illustrated in Figure 10.7. However, the simple litigation structure illustrated in Figure 10.1 and Figure 10.7 is highly unrealistic. Litigation proceeds in a series of sequential stages, and the costs of litigation are incurred at each stage. Suppose that one alters the litigation structure so that the litigation proceeds in three, equal-​cost stages, allowing the plaintiff to offer settlements before each stage (Bebchuk, 1996). Everything else is unchanged. Thus the expected judgment E(J) = 100, and each of the three stages costs each side 70, for a three-​ stage total of 210. If the case is litigated to judgment, the payoffs are identical to those in the one-​stage game depicted in Figure 10.7. Figure 10.8 depicts this slightly more complex litigation structure. In contrast to the predictions of the one-​stage model where the threat to litigate is not credible, the threat to litigate in the three-​stage model is credible. To see this, consider first the plaintiff ’s choice at stage 3. The plaintiff can litigate, and receive a payout of –​110 (an expected judgment of 100 minus 210 in total litigation costs). If he drops at stage 3, his total payoff is –​140, the sunk litigation costs incurred during stages 1 and 2. Thus, his total payoff is higher if he chooses to litigate. Intuitively, the plaintiff should rationally ignore the –​140 in sunk costs, as proceeding to trial yields an expected net benefit of 30 on the margin (an expected judgment of 100 minus 70 in marginal litigation costs). Given that the plaintiff will choose to litigate at stage 3, the defendant’s payoff if he rejects the plaintiff ’s offer of S = 100 made after stage 2 is –​310 (100 in expected judgment and 210 in litigation costs). Thus, it is rational for the defendant to accept the plaintiff ’s settlement offer S = 100 made after stage 2 to save the marginal litigation costs (70) that would be incurred at stage 3. Given that the defendant will rationally accept the offer made after stage 2, it is rational for the plaintiff to continue rather than drop at stage 2. And knowing that the plaintiff will choose to continue at stage 2, the defendant will rationally accept the offer of S = 100 made after stage 1. Working backwards, the three-​stage model predicts that the plaintiff ’s

 

Economics of Litigation    213

Offer S = 100

Reject Offer P = 30, D = –170

Stage 1

File (C = 70) Offer S = 100 P = 30, D = –170

Reject Offer P = –40, D = –240

Stage 2

Pretrial (C=70) Offer S = 100 P = –40, D = –240

Reject Offer P = –110, D = –310

Stage 3

Litigate (C = 70) P = –110, D = –310

Accept Offer P = 100, D = –100

Do Not File P = 0, D = 0

Accept offer P = 30, D = –170

Drop P = –70, D = –70

Accept Offer P = –40, D = –240

Drop P = –140, D = –140

Figure 10.8  Three-​stage model of litigation.

assurance that he will litigate the case is credible. As a result, the plaintiff ’s initial offer will be accepted even though the case has NEV if litigated to judgment.

10.3.2 Asymmetric and Sequential Costs In this section, we consider the effect of asymmetric and sequential costs (Rosenberg and Shavell, 1985). In their model, they show how asymmetric and sequentially incurred costs can allow a plaintiff to extract a settlement even when a case has zero judgment value. To see this, consider the litigation structure in Figure 10.9, where a plaintiff makes a discovery request at a cost CP. At the time of the request, the plaintiff also offers to settle the case for S. Note that the structure of litigation in Figure 10.9 alters the timing of when costs are incurred relative to the litigation structure contained in Figure 10.1. In

 

214   Bruce H. Kobayashi

Continue to Discovery?

Plaintiff Makes Discovery Request, Offer S

Defendant Responds to Discovery Request P = –CP, D = –CD

Plaintiff Drops Case P = 0, D = 0

Defendant Accepts Offer P = S – CP, D = -S

Figure 10.9  Asymmetric and sequential costs.

particular, the structure depicted in Figure 10.9 does not permit the plaintiff to avoid the costs of litigation if the offer S is rejected.12 The defendant can accept the offer S, or respond to the discovery request at a cost CD. The payoffs if the offer S is accepted are S –​ CP for the plaintiff and –​S for the defendant. If the defendant chooses to reject the offer and respond to the request, the payoff of the plaintiff and defendant are (–​CP, –​CD). Under the assumption that the litigation has no judgment value, the case is dropped after discovery. Examining the defendant’s decision to accept the offer to settle or respond to the request, the defendant will choose to settle if S ≤ CD. Conditional upon an offer S being accepted, it will be rational for the plaintiff to make the offer S if S ≥ CP. If an acceptable offer S ≤ CP, or if S ≥ CD so that the offer will not be accepted by the defendant, the plaintiff will rationally choose to drop the case without making and incurring the cost of a discovery request. Thus, an offer S that will be accepted by the defendant and that is also profitable for the plaintiff requires that CP ≤ S ≤ CD. Thus, costs must be asymmetric and higher for the party responding to the discovery request, i.e., CD ≥ CP.13 12  This assumption is critical. If the plaintiff could withdraw his discovery request and avoid C , P the defendant could reject the offer knowing that the plaintiff rationally will choose to withdraw the request. Altering the sequence of when costs are incurred makes the plaintiff ’s costs sunk and non-​avoidable. 13  The existence of asymmetric costs and their ability to contribute to large settlements that reflect the cost of litigation rather than the merits of the case are at the center of the controversy over discovery costs in litigation and proposals for discovery reform. For a discussion of these issues, see Kobayashi (2014a).

 

Economics of Litigation    215

[–110, –310]

.7 [–208.75, –211.25] .75 [–209, –211] .9 [–210, –210]

.1 [–200, –220]

.3 [120.42, –540.42]

.25 [–208, –212] .8 [–210, –210]

.2 [–200, –220]

.75 [–205, –215]] .75 [–210, –210]

.25 [–190, –230]

Stage 1

.25 [1096.67, –1516.67] .333 [–210, –210]

.667 [1750, –2170]

Stage 2

Stage 3

Figure 10.10  Three-​stage litigation with information revelation—​total payoffs.

10.3.3 Multistage Models and Option Value 10.3.3.1 Effect on Filing In this section, we further expand the sequential model to include the revelation of information. Figure 10.10 depicts a structure that has the same cost structure and expected value as the three-​stage model used in Section 10.3.1. Each of the outcome boxes in stages 1 through 3 lists the probability of reaching the box, followed by the payoffs for the plaintiff and defendant when the box is reached. The expected judgment value of the case if unconditionally litigated to judgment is 100. Each stage costs 70, so the total cost of litigation for each side is 210. The expected value of filing this suit if it is unconditionally litigated to judgment is –​110 for the plaintiff, and the expected exposure for the defendant is –​310. As was shown above, given these payoffs, the threat to sue is not credible in the simple one-​stage model depicted in Figures 10.1 and 10.7. The example in Figure 10.10 has two pre-​trial motions (stage 1 and stage 2) and a trial stage (stage 3). In contrast to the multistage examples presented above, both the probability and the magnitude of the trial payoffs are conditional on the outcomes of earlier pre-trial motions. In particular, there are four conditional trial outcomes, which are listed in Table 10.1. For simplicity, we assume that there are two defense motions, one a motion in limine to exclude the plaintiff ’s expert witness (stage 1 in Figure 10.10) and a motion to limit the scope of the plaintiff ’s supplemental discovery request (stage 2 in Figure 10.10). Figure 10.10 depicts the total payoff if a stage is reached. Figure 10.11 shows the marginal payoffs of the litigation depicted in Figure 10.10. Looking at Table 10.1, both of the defendant’s motions are granted over half of the time (52.5 percent). This results in the far-​left outcome during stage 3. Without their expert and with limited discovery, the plaintiff has a 1-​in-​10 chance of prevailing, with limited damages of 10. Ignoring the 140 in sunk costs incurred during stage 1 and stage 2, the expected value of going to trial conditional upon both of the defendants motions being granted is 1 –​70 = –​69.

 

216   Bruce H. Kobayashi Table 10.1 Pre-​trial motion outcomes Motion to limit discovery granted (p = 3/​4)

Motion to limit discovery denied (p = 1/​4)

Motion in limine granted (p = 7/​10)

1. both motions granted (p = .525)

3. d isc. motion denied, evid. motion granted (p = .175)

Motion in limine denied (p = 3/​10)

2. disc. motion granted, evid. motion denied (p = .15)

4. both motions denied. (p = .075)

[–110, –310]

.7 [–138.75, –141.25] .75 [–69, –71]

.9 [0, 0]

.1 [10, –10]

.3 [190.42, –470.42] .25 [–68, –72]

.8 [0, 0]

.2 [10, –10]

.75 [–65, –75]

.75 [0, 0]

.25 [20, –20]

Stage 1

.25 [1236.67, 1376.67]

.333 [0, 0]

.667 [1960, –1960]

Stage 2

Stage 3

Figure 10.11  Three-​stage litigation with information revelation—​marginal payoffs.

In contrast, the best outcome for the plaintiff, when both of the defendant’s motions are denied, occurs 7.5 percent of the time. This results in the far-​right outcome during stage 3. When the plaintiff goes to trial with his expert and after extensive discovery, he has a 2/​ 3 chance of prevailing, with damages equal to 1960. Conditional on both of the defendant’s motions being denied, the expected value of going to trial, again ignoring sunk costs, is 1237. In addition, there are two mixed outcomes. The first is where the defendant’s evidentiary motion is granted, but the discovery motion is denied. Thus the defendant successfully excludes the plaintiff ’s expert, but is unsuccessful in his attempt to limit discovery. This results in the expected stage 3 trial payoffs listed in the second to the left stage 2 outcome. The probability of recovery is 1 in 5, but damages are still limited to 10. The plaintiffs expected payoff, net of litigation costs at trial, is –​68. The other mixed outcome occurs when the defendant does not exclude the expert, but limits discovery. Here the probability of recovery is 1 in 4, and damages are 20. The expected payoff, net of litigation costs at trial equals –​65. At stage 1, the expected litigation payoff net of costs equals –​138.75 if the plaintiff ’s expert is excluded. Twenty-​five percent of the time the plaintiff faces a net trial payoff of –​68 when he defeats the defendant’s discovery motions, and 75 percent of the time he faces a net trial payoff of –​69 when the defendant’s discovery motion is successful. The other three stage 1 payoffs are derived in a similar manner. Taking the expected value of

 

Economics of Litigation    217 the stage 1 payoffs yields the expected value of this lawsuit when it is taken to judgment unconditionally, which is –​110. The defendant’s exposure under the same circumstances is –​310.

10.3.3.2 Option Value As noted by Cornell (1990), the expected value of the lawsuit depicted in Figures 10.10 and 10.11, based on the assumption that the litigants will unconditionally take the case to judgment, is not the value of the suit if the litigants can avoid negative payoff stages by ending the litigation. Given the parties’ option to end the litigation and avoid the cost of litigation, the option value of a lawsuit will be greater than the expected value of the lawsuit when the parties unconditionally proceed to litigate the case to judgment (Grundfest and Huang, 2006). Examining Figure 10.11, it is in the mutual interest of both parties to end the litigation at stage 3 unless the defendant loses both of his pre-​trial motions.14 Given that both parties would agree to end the litigation at this point, the marginal payoffs if any of the shaded in stage 2 boxes would be zero for both parties. The parties would also agree to dismiss the case if they reached the left outcome in stage 1 (i.e., when the defendant successfully excludes the plaintiff ’s expert). Thus, the parties would agree to dismiss the case as soon as one of the defendant’s motions were granted. Figure 10.12 shows the truncated decision tree that incorporates the parties’ decisions to dismiss the case. The plaintiff ’s option value of the lawsuit is higher than the expected value of the lawsuit when the parties unconditionally proceed to litigate the case to judgment. Moreover, the plaintiff ’s threat to file the lawsuit is credible, as the plaintiff ’s option value is positive. Note that early dismissal also benefits the defendant by reducing his expected litigation exposure. The source of the higher option value for both litigants is the ability of the litigants to avoid the costs of litigation when the payoffs are negative. When the defendant’s stage 1 motion is granted and the plaintiff ’s expert is excluded, the parties will rationally agree to end the litigation and save both the costs of the discovery motion and the costs of the trial (saving 140 relative to the costs of unconditional litigation). When the defendant’s stage 1 motion is denied, but his stage 2 motion to limit discovery is granted, the parties will also rationally agree to end the litigation. This saves the cost of the trial (70 relative to the costs of unconditional litigation).

10.3.3.3 Settlement and Option Value Figure 10.12 assumes that the case will be litigated and does not consider settlement. In the one-​stage model, the defendant will not rationally agree to a settlement because the plaintiff ’s threat to sue is not credible. But the threat to sue is credible when option value 14  FRCP 41(a) governs dismissals of actions. Under FRCP 41(a)(1) an action can be dismissed by the plaintiff without order of court (i) by filing a notice of dismissal at any time before service by the adverse party of an answer or of a motion for summary judgment, whichever first occurs, or (ii) by filing a stipulation of dismissal signed by all parties who have appeared in the action.

 

218   Bruce H. Kobayashi

[1.75, –194.25]

.3 [239.17, –414.17]

.7 [0, 0]

.25 [1236.67, –1376.67]

.75 [0, 0]

.333 [0, 0]

.667 [1960, –1960]

Figure 10.12 Option value.

is considered. Because settlement allows the parties to avoid the costs of litigation in those stages where litigation proceeds, the effects of settlement complements the effects of the option to exit and can further increase the option value of the lawsuit for both parties. The effect of settlement is illustrated in Figure 10.13. The example assumes that the parties agree to a midpoint settlement prior to stage 3 that imposes the expected judgment at trial. Thus, the expected stage 3 payoffs are increased by 70 for each party (equal to the saved trial costs). A similar effect occurs at stage 2 when the defendant’s motion to exclude the plaintiff ’s expert is denied. As a result of the anticipated cost savings, the settlement value of the case increases to 98, which raises the plaintiff ’s recovery and reduces the defendant’s litigation exposure relative to the option value litigation payoffs.15

10.3.3.4 Separate versus Consolidated Motions The decision to consolidate or separate motions affects the option value of a lawsuit.16 Figure 10.14 shows the marginal payoffs for the lawsuit depicted in Figures 10.10–​10.14 when the two pre-trial motions are consolidated. That is, instead of first litigating the evidentiary motion in stage 1 and then separately litigating the discovery motion in stage 2, both motions are litigated (and the costs of litigating both motions are incurred) in a 15  See generally Polinsky and Rubinfeld (1988) (examining the deterrent effects of settlements versus trials). 16  The consolidation or separation of issues is controlled by FRCP 42.

 

Economics of Litigation    219

[S = 98, –S = –98]

.7 [0, 0]

.3 [S = 326.67, –S]

.75 [0, 0]

.25 [S = 1306.67, –S]

.333 [0, 0]

.667 [1960, –1960]

Figure 10.13  Option value with settlement.

[–110, –310]

.525 [–69, –71]

.9 [0, 0]

.1 [10, –10]

.175 [–68, –72]

.8 [0, 0]

.2 [10, –10]

0225 [–65, –75]

.75 [0, 0]

.25 [20, –20]

.075 [1236.67, –1376.67]

.33 [0, 0]

.67 [1960, –1960]

Stage 1 & 2

Stage 3

Figure 10.14  Consolidated pre-​trial.

consolidated stage 1 and 2.17 Figure 10.14 shows the expected value of the lawsuit when the parties unconditionally proceed to judgment is unaffected by the consolidation of stages 1 and 2. Figure 10.15 depicts the option value of the lawsuit with consolidated pre-​trial. Compared to the option value of the litigation illustrated in Figure 10.12, consolidation of stages 1 and 2 reduces the option value of the lawsuit. Intuitively, the option to mutually exit the lawsuit allows the parties to save the costs of trial (70) 92.5 percent of the time. In contrast, taking the pre-​trial motions sequentially also allows saving the costs of trial 92.5 percent of the time, but also allows saving the costs of the second pre-​trial 17 

The example assumes that the cost of litigating the two motions together is the same as litigating the two motions sequentially. For an analysis of the effect of economies of scale in consolidating motions or trials, see Landes (1993, 1995).

 

220   Bruce H. Kobayashi

[–47.25, –243.25]

.525 [0, 0]

.175 [0, 0]

0225 [0, 0]

.075 [1236.67, –1376.67]

.33 [0, 0]

.67 [1960, –1960]

Figure 10.15  Option value with consolidated pre-​trial.

stage (an additional 70) 70 percent of the time. Indeed, consolidation results in negative option value for the plaintiff, again making the filing of the lawsuit non-​credible.

10.3.3.5 Order of Motions The option value of the lawsuit can also be affected by the order of motions.18 It is easy to confirm that the expected value of the lawsuit when the parties unconditionally take the lawsuit to judgment is unaffected by switching the order of the motions. However, such a change can alter the option value of the litigation. Figure 10.16 shows the effect of switching the sequential order of the two pre-​trial motions. That is, the discovery motion (formerly stage 2) is taken first, and the evidentiary motion (formerly stage 1) is taken second. Reversing the order of the pre-​trial motions increases the option value for the plaintiff and lowers the litigation exposure for the defendant. Intuitively, taking the defendant’s stronger motion first increases the probability that the lawsuit will be voluntarily dismissed after one motion by 5 percent, and thus increases option value by increasing the probability that the parties will save the cost of the second motion. Thus, even though the discovery motion is a stronger motion for the defendant, once litigation has commenced, the plaintiff will voluntarily agree to take that motion first.

10.3.3.6 Effect of Fee-​shifting Figure 10.17 illustrates the effect of two-​way cost-​and fee-​shifting in the sequential multi​stage model when the parties unconditionally litigate the lawsuit to judgment. The figure assumes that cost-​and fee-​shifting is imposed at judgment to the prevailing party. Because the probability that the plaintiff is the prevailing party is less than ½,19 a system of cost-​and fee-​shifting lowers the expected value to the plaintiff and reduces 18 

For an analysis of the order of motions, including taking issues of damages before issues of liability, see Stevenson (2006). See also Beckner and Salop (1999). 19  The probability of judgment for the plaintiff is .193.

 

Economics of Litigation    221

[5.25, –190.75]

.75 [0, 0]

.25 [301, –483]

.3 [1236.67, –1376.67]

.7 [0, 0]

.333 [0, 0]

.667 [1960, –1960]

Figure 10.16  Reversing order and option value. [–238.625, –181.375]

.7 [–296.25K, 16.25K] .75 [–237, 97] .9 [–210, 210]

.1 [220, –220]

.3 [129.17, –409.17] .25 [–194, 54]

.8 [–210, 210]

.2 [220, –220]

.75 [–170, 30] .75 [–210, 210]

.25 [230, –230]

Stage 1

.25 [1306.67, –1446.67] .333 [–210, 210]

.667 [2170, –2170]

Stage 2

Stage 3

Figure 10.17  The effect of prevailing litigant two-​way fee-​shifting.

the defendant’s litigation exposure. Thus, if the parties unconditionally litigate the case to judgment, the effect of two-​way cost-​and fee-​shifting is similar to its effect in the one-​ stage model. Figure 10.18 illustrates the option value of the lawsuit with two-​way cost-​and fee-​ shifting. Note that the figure assumes that the plaintiff can unilaterally choose to dismiss the lawsuit. In contrast to its effect on the net present value of the case that is litigated unconditionally to judgment, the option value of the case is higher with two-​way cost-​ and fee-​shifting. This is because prevailing party cost-​and fee-​shifting is not imposed when the case is dismissed prior to judgment. Rather, in the example illustrated in Figure 10.18, it is imposed only when both of the defendant’s motions are defeated and the case is litigated to judgment. When the case is litigated to judgment, the probability of judgment for the plaintiff is 2/​3, and two-​way cost-​and fee-​shifting increases the plaintiff ’s option value of the lawsuit by reducing his expected litigation costs.

 

222   Bruce H. Kobayashi

[7, –199.5]

.7 [0, 0]

.3 [256.67, –431.67]

.25 [1306.67, –1446.67]

.75 [0, 0]

.333 [–210, 210]

.667 [2170, –2170]

Figure 10.18  Prevailing litigant two-​way fee-​shifting and option value.

The actual effect of two-​way fee-​shifting will depend crucially on the plaintiff ’s ability to unilaterally dismiss the lawsuit. Note that in the examples of option value used in prior sections, it was in the interest of both parties to dismiss the action. However examining Figure 10.17, it is in the plaintiff ’s interest, but not in the defendant’s interest, to dismiss the case when one of the defendant’s motions is granted when any of the shaded boxes are reached. Rather, as soon as one of the defendant’s motions is granted, the defendant has an interest of continuing the litigation to judgment, where he has a high probability of prevailing, in order to trigger the shifting of costs and fees. Thus, if the plaintiff is unable to unilaterally dismiss the action, the option value model will not apply.20

10.4  External Effects and the Public-​ versus-​P rivate Return to Litigation In this section, we return to the non-​sequential model to set out a model of litigants’ incentives to litigate and settle cases which takes into account external effects of 20  Indeed, the unilateral right to dismiss a case under FRCP 41(a) ends after the complaint is answered. In addition, the Court has ruled that voluntary dismissal of an action does not result in the dismissal of pending Rule 11 motions. See Cooter and Gell v. Hartmarx, 496 U.S. 384 (1990); Kobayashi and Parker (1993).

 

Economics of Litigation    223 judgment (Kobayashi, 1996) and the value of non-​monetary or injunctive relief imposed via settlement, Kobayashi (2014b), Hylton and Cho (2010), Shavell (1993). To derive the settlement range in such a setting, we must examine again the litigant’s trial payoffs. The minimum the plaintiff is willing to accept in settlement is determined by his trial payoff:

TP = pP ( J + GP ) − (1 − pP ) LP − CP

(1E)

The trial payoff in (1E) contains two new terms, GP, which are the plaintiff ’s external gains of a judgment not included in J, and LP, which are the external losses imposed on the plaintiff when there is a judgment for the defendant. For example, LP included the persuasive or preclusive effects of a judgment against a plaintiff that is a repeat litigant. One common example is an IP rights holder who is facing multiple infringers. If the defendant’s intellectual property rights are invalidated the first time they are litigated, the defendant will be estopped from relitigating the issue of validity.21 Thus, LP represents the loss of the patent due to preclusion (Kobayashi, 2014b). Even in the absence of preclusion, an adverse judgment in one case may reduce the probability that the plaintiff will be successful in later cases. In this case LP represents the loss in litigation value of subsequent cases due to the persuasive effects of the earlier case. Note that LP generally does not include the value of precedent, as the effects of precedent that are separate from the preclusive or persuasive effects of litigation are generally spillover effects that are not internalized by the plaintiff (Shavell, 1977, 1982; Kaplow, 1986). Gp would include, for example, the competitive benefits of deterred or blocked entry that would have occurred absent a judgment upholding the validity of the intellectual property right, or the persuasive effect of the judgment on later cases. In the case of a repeat plaintiff, it does not generally include the effects of precedent for the reasons stated above, or the effects of preclusion.22 The maximum defendant i is willing to pay is determined by his expected cost of going to trial:

Ti = pi ( J + Li ) − (1 − pi ) Gi + Ci ,

(2E)

The new terms Gi and Li are the net external gains and losses for litigant i imposed upon judgment that are not included in Ji. In intellectual property litigation, for example, Gi can include the competitive benefits of early entry that would result from the invalidation of the plaintiff ’s intellectual property rights. Li can consist of the lost competitive benefits that results from entry that would have happened in the absence of litigation.

21  Blonder-​Tongue Laboratories, Inc. v. University of Illinois Foundation, et al. 402 U.S. 313 (1971), Kobayashi (1996). 22  Kobayashi (1996).

 

224   Bruce H. Kobayashi An acceptable settlement would be any offer in which the monetary payment M plus the “injunctive” costs of a settlement, Ii, are less than the defendant’s cost of going to trial:23 S = M + Ii ≤ Ti



(4Ei)

so acceptable settlements for the plaintiff would include settlements where the monetary settlement M plus the non-​monetary benefits of a settlement relative to the status quo, Ip,24 are greater than the benefits of going to trial: S = M + I P ≥ TP .



(4Ep)

A sufficient, but not necessary, condition for litigation is the absence of a bargaining range. A bargaining range will be absent when TP − I P > Ti − Ii , or when: pP (DP + GP ) − (1 − pP )LP − CP − I P > pi (Di + Li ) − (1 − pi )Gi − Ii + Ci



(5E)

or equivalently: pP (DP + GP ) − pi (Di + Li ) − (1 − pP )L p + (1 − pi )Gi + Ii − I P > Ci + C p .



Equation (5E) shows that the sufficient condition for litigation is more likely to be satisfied, ceteris paribus, if the litigants are relatively optimistic (pp > pi), if the plaintiff has relatively large damages or high external benefits from a favorable judgment, if the defendant has relatively large external gains from a judgment favorable to him, or if the defendant’s non-​monetary costs from a settlement are larger than the patentee’s non-​monetary gains. Note that optimism is not required to generate litigation, as is the case in the standard model of litigation and settlement. Intuitively, judgments can have intrinsic value to the litigants, thus creating a demand for a judgment when these effects are positive, and litigation aversion when they are not. Examples of such judgment-​specific benefits would be the precedential or persuasive effects of a favorable judgment. Such effects produce external gains for the patentee by lowering the probability that the patent will be successfully challenged in the future. Another example of a judgment-​specific external effect is the negative effect of non-​party preclusion, which will prevent or stop the patentee from relitigating the validity of the patent held invalid in the first case. To see this, consider the case where there is no disagreement on the probability the plaintiff will prevail, so that pP = pi = p. This yields the following sufficient condition for litigation: p(GP − Li ) + (1 − p)(Gi − LP ) + Ii − I P > Ci + CP

23 

(5E1′)

Note that if the settlement generates benefits to the firm, Ii