Constitutional Economics: A Primer 9781108708395, 9781108486880

Table of contents :
Brief Contents
Detailed Contents
Figures
Tables
Preface
Introduction
1 Conceptual Foundations
1.1 Constitutional Political Economy As Part of the New Institutional Economics
1.2 Definitions
1.3 Concepts of the Constitution
1.3.1 The Constitution As a Social Contract
1.3.2 The Constitutional Contract As Principal–Agent Relationship
1.3.3 The Constitutional Contract As a Precommitment Device
1.3.4 The Constitution As a Bundle of Conventions
1.3.5 The Constitution As the Result of an Evolutionary Process
1.3.6 Summary
1.4 The Tools
1.4.1 Theory
1.4.2 Empirics
1.5 More Potentially Useful Tools
1.5.1 Experiments
1.5.2 Possible Consequences of Constitutional Rules
1.6 Consequences of Constitutional Rules
1.7 Measuring Constitutions
2 Democracy vs. Autocracy
2.1 How Best to Separate Democracy from Autocracy?
2.2 What Role Do Constitutions Play in Autocracy?
2.3 Do Autocracies and Democracies Perform Differently?
2.3.1 Theory
2.3.2 Empirical Evidence
2.3.2.1 Use Adequate Estimation Techniques
2.3.2.2 Ask More Precise Questions
2.4 Do We Have Good Reasons to Assume That Different Kinds of Autocracies Perform Differently?
2.4.1 Theory
2.4.2 Empirical Evidence
2.5 What Are the (Economic) Effects of Extending the Franchise?
2.6 What Factors Make a Country Switch from One Regime to Another?
2.6.1 Theory
2.6.2 Empirical Evidence
2.7 Summary and Outlook
3 Positive Constitutional Economics
3.1 Introductory Remarks
3.2 Basic Rights
3.2.1 Defining Rights
3.2.2 Effects of Basic Rights
3.2.2.1 Theory
3.2.2.2 Empirics
3.2.3 Determinants of Basic Rights
3.2.3.1 Theory
3.3 Horizontal Separation of Powers
3.3.1 Introductory Remarks
3.3.2 Branches of Government
3.3.3 Form of Government: Presidential vs. Parliamentary Regimes
3.3.3.1 Effects of Forms of Government
3.3.3.2 Determinants of Forms of Government
3.3.3.2.1 Theory
3.3.3.2.2 Empirics
3.3.4 The Structure of the Legislature: Bicameralism
3.3.4.1 Effects of Bicameralism
3.3.4.1.1 Theory
3.3.4.1.2 Empirics
3.3.4.2 Determinants of Bicameralism
3.3.5 A Forgotten Branch: The Judiciary
3.3.5.1 The (Economic) Effects of the Judiciary
3.3.5.1.1 Theory
3.3.5.1.2 Empirics
3.3.5.2 Determinants of Judicial Independence
3.3.5.2.1 Theory
3.3.5.2.2 Empirics
3.4 Vertical Separation of Powers: Federalism
3.4.1 Definitions
3.4.2 Effects of Federalism
3.4.2.1 Theory
3.4.2.2 Empirics
3.4.3 Determinants of Federalism
3.4.3.1 Theory
3.4.3.2 Empirics
3.5 The New Separation of Powers: Increasing the Number of Independent Decision-Makers
3.6 Representative vs. Direct Democracy
3.6.1 Introductory Remarks
3.6.2 Effects of Direct Democracy
3.6.2.1 Theory
3.6.2.2 Empirics
3.6.2.2.1 First-Generation Research: Focusing on the United States and Switzerland
3.6.2.2.2 Second-Generation Research: Cross-Country Studies
3.6.2.2.3 Third-Generation Research: Taking Endogeneity Seriously
3.6.3 Determinants of Direct Democracy
3.7 Electoral Systems
3.7.1 Effects of Electoral Systems
3.7.1.1 Theory
3.7.1.2 Empirics
3.7.2 Determinants of Electoral Systems
3.8 Summary
4 Conclusions and Possible Future Issues
4.1 Introductory Remarks
4.2 Conclusions
4.3 Rules for Choosing and Amending Constitutions
4.3.1 Introductory Remarks
4.3.2 Rules for Choosing Constitutions
4.3.2.1 Choice of Members
4.3.2.2 Size of Assembly
4.3.2.3 Time Limits
4.3.2.4 Public or Secret Meetings
4.3.2.5 Voting Rules
4.3.2.6 Public Participation
4.3.3 Procedural Rules for Changing Rules
4.4 Mind the Gap: Analyzing the Divergence between Constitutional Text and Constitutional Reality
4.4.1 Preliminary Remarks
4.4.2 How to Define and Measure the Gap?
4.4.3 Possible Determinants of the Gap: Theory
4.4.4 Challenges for Establishing Causality
4.4.5 Attempts to Analyze the Gap: Empirics
4.5 Emergency Constitutions
4.5.1 Introductory Remarks
4.5.2 Measurement Issues
4.5.3 Who Uses Emergency Constitutions?
4.5.4 Are Emergency Constitutions Effective?
4.5.5 Summary and Possible Implications for Future Research
Appendix 1 Coding Countries according to Two Governance Scores
Appendix 2 Empirical Results at a Glance: Constitutional Rules As Explanatory Variables and Cross-Country Results Unless Otherwise Noted
References
Index

Citation preview

Constitutional Economics

Constitutional political economy has emerged as an indispensable part of political economy. This book offers a concise survey of the questions, methods, and empirical findings central to this topic. What effects – if any – do constitutions have within autocracies? Can small electoral districts help reduce corruption? Does a country’s leadership affect the size of its government? Can direct democratic institutions increase politicians’ accountability to citizens? Stefan Voigt, a pioneer in the field, explores these questions and more throughout the course of this cutting-edge primer. As the number of courses in constitutional economics continues to grow, this book fills an important gap in the literature. This highly original project maintains curiosity about the questions it generates, identifying potential new areas of research whilst successfully demonstrating the impact constitutional rules have on political economy. Stefan Voigt is a professor at the University of Hamburg, director of its Institute of Law & Economics, and editor of Constitutional Political Economy.

Constitutional Economics A Primer Stefan Voigt University of Hamburg

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314 321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi

110025, India

79 Anson Road, #06 04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108486880 DOI: 10.1017/9781108764445 © Stefan Voigt 2020 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2020 Printed in the United Kingdom by TJ International Ltd, Padstow Cornwall A catalogue record for this publication is available from the British Library. Library of Congress Cataloging in Publication Data Names: Voigt, Stefan, author. Title: Constitutional economics : a primer / Stefan Voigt, University of Hamburg. Description: New York, NY : Cambridge University Press, 2020. | Includes bibliographical references and index. Identifiers: LCCN 2019056215 (print) | LCCN 2019056216 (ebook) | ISBN 9781108486880 (hardback) | ISBN 9781108708395 (paperback) | ISBN 9781108764445 (epub) Subjects: LCSH: Economics Political aspects. | Economic policy. | Constitutional law Economic aspects. Classification: LCC HB171 .V59 2020 (print) | LCC HB171 (ebook) | DDC 330 dc23 LC record available at https://lccn.loc.gov/2019056215 LC ebook record available at https://lccn.loc.gov/2019056216 ISBN 978 1 108 48688 0 Hardback ISBN 978 1 108 70839 5 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Brief Contents List of Figures x List of Tables xi Preface xiii Introduction 1 1

Conceptual Foundations 4

2

Democracy vs. Autocracy

3

Positive Constitutional Economics 43

4

Conclusions and Possible Future Issues 89

25

Appendix 1 Coding Countries according to Two Governance Scores 107 Appendix 2 Empirical Results at a Glance: Constitutional Rules As Explanatory Variables and Cross Country Results 114 References 120 Index 136

Detailed Contents List of Figures x List of Tables xi Preface xiii Introduction 1 1 Conceptual Foundations 4 1.1 Constitutional Political Economy As Part of the New Institutional Economics 1.2 Definitions 5 1.3 Concepts of the Constitution 6 1.3.1 The Constitution As a Social Contract 7 1.3.2 The Constitutional Contract As Principal–Agent Relationship 7 1.3.3 The Constitutional Contract As a Precommitment Device 8 1.3.4 The Constitution As a Bundle of Conventions 9 1.3.5 The Constitution As the Result of an Evolutionary Process 10 1.3.6 Summary 11 1.4 The Tools 12 1.4.1 Theory 12 1.4.2 Empirics 13 1.5 More Potentially Useful Tools 15 1.5.1 Experiments 15 1.5.2 Possible Consequences of Constitutional Rules 17 1.6 Consequences of Constitutional Rules 18 1.7 Measuring Constitutions 21

2 Democracy vs. Autocracy

4

25

2.1 How Best to Separate Democracy from Autocracy? 26 2.2 What Role Do Constitutions Play in Autocracy? 28 2.3 Do Autocracies and Democracies Perform Differently? 29 2.3.1 Theory 29 2.3.2 Empirical Evidence 30 2.4 Do We Have Good Reasons to Assume That Different Kinds of Autocracies Perform Differently? 33 2.4.1 Theory 33 2.4.2 Empirical Evidence 34 2.5 What Are the (Economic) Effects of Extending the Franchise? 36 2.6 What Factors Make a Country Switch from One Regime to Another? 36

viii

Detailed Contents

2.6.1 Theory 36 2.6.2 Empirical Evidence 40 2.7 Summary and Outlook 42

3

Positive Constitutional Economics 43 3.1 Introductory Remarks 43 3.2 Basic Rights 44 3.2.1 Defining Rights 44 3.2.2 Effects of Basic Rights 45 3.2.3 Determinants of Basic Rights 47 3.3 Horizontal Separation of Powers 48 3.3.1 Introductory Remarks 48 3.3.2 Branches of Government 49 3.3.3 Form of Government: Presidential vs. Parliamentary Regimes 50 3.3.4 The Structure of the Legislature: Bicameralism 56 3.3.5 A Forgotten Branch: The Judiciary 60 3.4 Vertical Separation of Powers: Federalism 65 3.4.1 Definitions 65 3.4.2 Effects of Federalism 66 3.4.3 Determinants of Federalism 70 3.5 The New Separation of Powers: Increasing the Number of Independent DecisionMakers 71 3.6 Representative vs. Direct Democracy 73 3.6.1 Introductory Remarks 73 3.6.2 Effects of Direct Democracy 74 3.6.3 Determinants of Direct Democracy 79 3.7 Electoral Systems 80 3.7.1 Effects of Electoral Systems 80 3.7.2 Determinants of Electoral Systems 85 3.8 Summary 88

4

Conclusions and Possible Future Issues

89

4.1 Introductory Remarks 89 4.2 Conclusions 89 4.3 Rules for Choosing and Amending Constitutions 92 4.3.1 Introductory Remarks 92 4.3.2 Rules for Choosing Constitutions 93 4.3.3 Procedural Rules for Changing Rules 98 4.4 Mind the Gap: Analyzing the Divergence between Constitutional Text and Constitutional Reality 99 4.4.1 Preliminary Remarks 99 4.4.2 How to Define and Measure the Gap? 100 4.4.3 Possible Determinants of the Gap: Theory 101 4.4.4 Challenges for Establishing Causality 101 4.4.5 Attempts to Analyze the Gap: Empirics 102

ix

Detailed Contents

4.5 Emergency Constitutions 102 4.5.1 Introductory Remarks 102 4.5.2 Measurement Issues 103 4.5.3 Who Uses Emergency Constitutions? 104 4.5.4 Are Emergency Constitutions Effective? 105 4.5.5 Summary and Possible Implications for Future Research

105

Appendix 1 Coding Countries according to Two Governance Scores 107 Appendix 2 Empirical Results at a Glance: Constitutional Rules As Explanatory Variables and Cross Country Results 114 References 120 Index 136

Figures 2.1

Autocratic transitions and democratic transitions

2.2

Autocratic regimes across time

41

40

Tables 1.1 Overview of outcome variables 20 4.1 Conceptualizing constitutional change

99

Preface In one sense, the economic analysis of constitutions has been around for a long time. Charles Beard published his economic interpretation of the US Constitution more than a century ago (in 1913). In another sense, the economic analysis of constitutions is very young: Analyzing the effects of constitutions by drawing on modern econometric tools experienced a boost only after 2000. Over the last 20 years, constitutional economics has been very successful and contributions to this rapidly growing field can be found in the top economics as well as political science journals. Constitutional political economy, or constitutional economics, draws on questions, insights, and tools from at least three disciplines, namely law, political science, and economics. If normative analysis occupies center stage, philosophy comes in as the fourth discipline. This book is aimed at familiarizing anybody interested in constitutional economics with its most basic concepts. In addition – and employing the core of this primer – it contains a survey of the most interesting results brought about within the field to date. Since the idea was to write a primer, no attempt was made to cover all the studies belonging to this field but to concentrate on the more relevant ones. This is, obviously, a very subjective enterprise and others might have come up with different choices. I have been lucky to write this primer at the Institute of Law and Economics at the University of Hamburg. Here, a culture of open and critical discussion has developed over the years. In writing this primer, I profited from critical comments as well as constructive suggestions from my colleagues. Thanks to Marek Endrich, Dr. Jerg Gutmann, Dr. Stephan Michel, and Konstantinos Pilpilidis for reading and discussing every single page of this primer.

Introduction

The economic analysis of constitutions, also known as “constitutional economics” or “constitutional political economy,” is a relatively young research program. Standard economics generally focuses on the analysis of choices within rules, thus assuming rules to be exogenously given and fixed. Constitutional economics broadens this research program by analyzing the choice of rules, using the workhorse method of economics, i.e., rational choice. Constitutional political economy (CPE) seems like a pretty unhandy term – but it makes sense because it comprises the term “political economy” and thus insists that all actors – even those writing constitutions – are modeled as rationally maximizing their own utility. As a short hand, I will, however, frequently refer to “constitutional economics.” Two broad avenues can be distinguished in the economic analysis of constitutions: the normative branch and the positive branch. The normative branch is interested not only in legitimizing the state and its most basic rules but also in identifying Paretoimproving rule changes. Thus, this approach is dominated by adherents of social contract theory. According to Buchanan (1987: 249), the purpose of this contractarian approach is justificatory in the sense that “it offers a basis for normative evaluation. Could the observed rules that constrain the activity of ordinary politics have emerged from agreement in constitutional contract? To the extent that this question can be affirmatively answered, we have established a legitimating linkage between the individual and the state.” The positive branch is interested in explaining (a) the (economic) effects of alternative constitutional rules, and (b) the emergence and modification of constitutional rules. A generation ago, two public choice scholars questioned the relevance of normative theory, stating that “it is extremely doubtful whether such analysis will make a perceptible difference in the character of prevailing institutions. Normative theory is useful in helping us clarify our norms, but it is another question whether such analysis will impact on the pattern of real institutional development” (McCormick and Tollison 1981: 126). These scholars clearly preferred a positive approach, but were rather critical in their evaluation of the state-of-the-art of positive theory: “There is thus a sharp contrast in the two approaches to constitutional-institutional analysis. It is also fair to say, however, that at this point in time the normative approach to constitutional analysis is much better developed in the literature than the positive approach.”

2

Introduction

The above critique was written a generation ago but is no longer true. Ever since the beginning of the third millennium, the landscape has changed completely and CPE has been discovered by mainstream economics. Before that, most of the important contributions to the emerging research program appeared in Public Choice or in Constitutional Political Economy, a journal expressly founded in 1990 to give the research program a push. Today, many important contributions to the field can be found in journals such as the American Economic Review, the Quarterly Journal of Economics, and the Journal of Political Economy, to name three of the most prominent economic journals. It seems that a number of factors played a role in the rapid development of positive constitutional economics. Perhaps, the most important was the demise of the socialist countries in Central and Eastern Europe. The apparent failure of socialism made the dichotomy between socialism and capitalism largely irrelevant, leading to a more focused look at institutional variation within market-oriented countries. Based precisely on this new focus, Djankov et al. (2003) called for a “new comparative economics.” The growing need for well-informed policy advice made it clear that economists had unduly neglected the relevance of institutions, and during this transition, the importance of institutions simply could not be ignored any longer. Solid and empirically tested knowledge is a necessary prerequisite to sound policy advice. It is no accident that constitutional economics and the new institutional economics (NIE) appeared on the scene almost simultaneously. Another factor that contributed to the rise of positive constitutional economics and has been a definite boon for crosscountry analyses is the availability of data sets covering a wide range of variables. Most of the contributions to the literature of the positive branch are somewhat unbalanced because they deal with the economic effects of constitutions, and only a handful of them address the emergence and modification of constitutional rules. Because constitutional economics scholars claim to be interested in the choice of rules as opposed to the choice within rules, endogenizing constitutions might be considered to be the core of their research programs. It is, however, entirely natural for scholars who advocate the positive approach to begin their research program by first analyzing the effects of constitutions. After all, if constitutional rules did not have any significant effect on economic outcomes, expending effort determining the emergence of these rules would be pointless. Fortunately, there is now substantial evidence that constitutional rules do have important economic effects, and it is now time to take the next step of endogenizing constitutional rules. Of course, this might be easier said than done. Although constitutions are frequently rewritten,1 there is still a limited number of observations on which to base econometric evidence. Constitutional choice systematically takes place 1

Elkins et al. (2009) count 736 constitutions written and promulgated since 1789, implying an average rate of 3.5 per year. According to the Asian Development Bank (2005), a total of 122 constitutions have been adopted between 1990 and 2004, most of them in Africa (54) and Asia (38).

Introduction

3

under very special circumstances and there are dozens of potentially relevant determinants for choosing specific constitutional rules. Although quite a few remarkable case studies have been produced, we are still a far cry away from a “general” theory that could be tested empirically. This primer is divided into four main chapters. Chapter 1 lays the conceptual foundations. It discusses various concepts of a constitution, describes a number of conceptual challenges that need to be dealt with in the economic analysis of constitutions, and presents the toolbox of the constitutional economist. Essentially, Chapter 2 does three things. It enquires into the functions of constitutions for autocratic regimes. It asks whether economic outcomes achieved in democratically constituted states systematically differ from those reached by autocracies. And it finally deals with the question of why some countries manage to be democratic while others do not. Chapter 3 constitutes the core of this primer. It summarizes the current state of knowledge in positive constitutional economics proper. The most relevant theoretical conjectures and empirical studies regarding the effects of individual constitution rules are briefly presented, as is the state of knowledge regarding their determinants. In Chapter 4, I offer an outlook where I leave the grounds of established knowledge and speculate about the future of CPE. Constitutional political economy is at the intersection of at least three disciplines: political science, jurisprudence, and economics. More and more, it is economists who are conducting studies and writing treatises on constitutional economics. This trend, also visible in other branches of Law & Economics, implies the danger of “losing the lawyers” due to relying on economic jargon. To prevent this from happening, I have tried to use non-technical language as much as possible. After all, those who write constitutions are still primarily lawyers. Over the last twenty years, research in constitutional political economy has been rapidly expanding. Precisely due to this fast growth, it has become more difficult to obtain an overview of this field. This primer is an attempt to help with that endeavor. It summarizes the current state of knowledge in constitutional economics, and strives to be a helpful guide for those who are not experts in the field.

1

Conceptual Foundations

Before addressing the core issues of this primer, a number of preliminaries need to be dealt with. These include the following: (1) showing that constitutional political economy (CPE) is part of a broader research program, namely, the New Institutional Economics (NIE); (2) defining some key terms; (3) summarizing the various concepts of the constitution that have been used; (4) discussing some of the methodological challenges constitutional economics faces; (5) presenting some of the tools that are used to gain insights; (6) briefly discussing the most relevant outcome variables; and (7) addressing several issues in measuring constitutional as well as other variables.

1.1 Constitutional Political Economy As Part of the New Institutional Economics Scholars working in the NIE analyze (i) the (economic) effects of alternative institutions and (ii) the determinants of institutional change (see, e.g., North 1981, 1990; Voigt 2019 is an introductory textbook for the NIE). Institutions can be defined as commonly known rules used to structure recurrent interaction situations with a sanctioning mechanism that can be employed in cases of noncompliance. North (1990) distinguished between formal and informal institutions, referring to the formality of the rule as the distinguishing criterion. Voigt and Kiwit (1998) propose distinguishing between external and internal institutions. The distinction between external and internal institutions lies in who does the sanctioning. The state carries out the sanctioning for external institutions, whereas internal institutions are based on customary practices and the sanctioning is done by the members of the society. As long as constitutional rules are viewed as a specific kind of institution, the NIE can be interpreted as the more inclusive research program. Surveys or textbooks on the NIE might reasonably be expected to include the economics of constitutions, but they usually do not. It is argued here that constitutional political economy could profit immensely from positioning itself within the broader NIE. The normative branch of constitutional economics has been dominated by contractarian approaches (e.g., Buchanan 1975) in

1.2 Definitions

5

which a very broad capacity to deliberately set and modify constitutional rules is at least implicitly assumed. Additionally, it is often assumed that the actors choosing the constitution have at their disposal complete and perfect knowledge concerning the consequences of alternative constitutional arrangements. These two assumptions are neither realistic nor adequate for understanding how constitutions actually evolve. If these assumptions are abandoned, an important issue can be dealt with explicitly, namely that the enforcement of constitutional rules is precarious and depends on a number of conditions that ought to be identified. Actual enforcement could, for example, depend on “pre-constitutional” internal institutions. These internal institutions could thus constitute hard constraints on the enforceability of constitutional rules.

1.2 Definitions Economists have analyzed constitutions from many different angles, including, but certainly not limited to, as a social contract, as an incomplete contract, as a principal–agent relationship, as a precommitment device, as the result of cultural evolution, and as a bundle of conventions. In this primer, I draw on results from different perspectives, so it seems appropriate to choose a wide definition of the term “constitution.” An early definition within the research program of constitutional economics was provided by Buchanan and Tullock (1962: vii): “We shall mean by this term (constitution) a set of rules that is agreed upon in advance and within which subsequent action will be conducted.” This is a very broad definition. I propose to distinguish between three areas to which it can, in principle, be applied: (1) constraints that an individual imposes on herself; (2) constraints that individuals impose on each other in social settings that can range from family constitutions to constitutions of soccer clubs and firms; and (3) constraints that are imposed on the agents representing the state. In this primer, we are exclusively concerned with the last approach. Constitutions are concerned with procedures for the production of public goods. By writing and adopting constitutions, societies are not deciding with any meaningful detail on what sort of public goods they want to provide for themselves.1 Rather, constitutions contain provisions that are intended to be used in making those decisions. They, therefore, constrain the representatives of society in the ways they choose what public goods should be produced. At the same time, they enable 1

Note that, literally, it is always individuals never societies at large who choose. We, thus, are not departing from methodological individualism but simply saving space when we write that “societies choose.”

6

Conceptual Foundations

politicians to make choices. If one is interested in analyzing a society’s choice of a constitution, one is really interested in analyzing a meta-choice; that is, how a society chooses (on the constitutional level) how to choose later on (on the postconstitutional one). Or, in the words of Peter Ordeshook (1993: 231f.): “A constitution is not a piece of legislation; it is the mechanism people use to guide the formulation of legislation and law.” A constitution can be defined as a system of rules used for the provision of public goods within the collectivity called the state. Constitutions will regularly (1) define the organs that make up the state; (2) delineate competences of these organs; (3) outline procedures to be followed to modify constitutional rules and pass subconstitutional rules; and (4) specify individual rights. Any country that produces a minimum amount of public goods can thus be said to have a de facto constitution. This implies that constitutions do not always need to be written documents, as the production of public goods could simply be following generally accepted conventions. And indeed, the English Constitution is largely based on what Dicey (1885) has termed “constitutional conventions.” But knowing with any degree of certainty which rules have acquired constitutional status is no mean feat (Marshall and Trotter 1984, discusses the theory of conventions following Dicey). Furthermore, having a de facto constitution is not the same as having a constitution in the sense of constitutionalism. Constitutionalism is concerned with constraining government action to certain areas and restricting the means government can employ in the production of public goods.2 Societies in which the government stays within the limits laid down by the constitution are said to have an effective constitution, or, in other words, there is congruence between the society’s de jure and de facto constitutions.

1.3 Concepts of the Constitution In this section, we briefly summarize different ways to analyze constitutions. Depending on the questions that the analysis focuses on, one is led to rely upon different concepts of a constitution. Since the analysis focuses on different questions,

2

Within the (Western) tradition of constitutionalism, much emphasis has been placed on the con straining side of constitutions. Ever since Thomas Hobbes, the state has been portrayed as Leviathan. However, compared to anarchy, the state can also make everyone better off, a view that should lead to equal emphasis on the enabling parts of the constitution.

1.3 Concepts of the Constitution

7

the concepts are not necessarily mutually exclusive. All of them allow one to draw implications for real-world constitution-making.

1.3.1 The Constitution As a Social Contract First-generation constitutional economists were primarily interested in legitimizing the state. Many of them drew extensively on social contract philosophers such as Hobbes, Locke, and others. Buchanan (1975), for example, describes the situation under which the members of a group are likely to pin such a social contract as an “equilibrium of anarchy” in which marginal costs and returns for producing, stealing, and protecting goods are equally high. The individuals realize that they could all be better off, if they could agree on a disarmament contract, which would allow them to reduce the resources used for protecting – and stealing – goods. Since the individuals find themselves in a prisoner’s dilemma situation, they all have an incentive to sign a disarmament contract – and subsequently not comply with it. As all individuals foresee this, they create a protective state to protect their private spheres. Additionally, they create a productive state which is to provide society with those (public) goods whose private production would not be profitable. The idea that individuals create a state by way of contract is not meant to be a historically correct description but simply a heuristic. Buchanan believes it to be helpful not only in explaining existing institutions but also hopes to be able to derive from it some criteria for their evaluation (1975: 50ff.). Probably the most relevant shortcoming of this concept is the incentives of those endowed with the competence to use force. Why should they abide by the rules if they could make themselves better off by reneging?

1.3.2 The Constitutional Contract As Principal–Agent Relationship Principal–agent theory is developed out of the observation that contracting partners often dispose of differing information. A principal entrusts an agent with a number of tasks but cannot observe the actions of the agent costlessly. Additionally, the agent might have to act in situations too complex to be amenable for a clear-cut evaluation with regard to the respective aim. The agent thus disposes of a certain degree of discretion in her actions that she will use to maximize her own – and not the principal’s – utility. The main focus of principal–agent theory therefore lies on possibilities to constrain the agent and to design an optimal contract under the assumption of asymmetrically distributed information. The members of society are thought of as the principals and the government as the agent. The problem consists of designing a contract – in this case, the constitution – in such a way that the agents will be led to maximize the utility of the principals. At the same time, the agents must still have incentives to not only actually do the job

8

Conceptual Foundations

(the participation constraint) but also do a good job. The Leviathan model introduced into constitutional economics by Brennan and Buchanan (1980) can be interpreted as a principal–agent model. The basic assumption of the model is that the government has incentives to maximize (tax) revenue beyond a level that is in the interest of the citizens. Their advice regarding the design of the tax constitution, therefore, is to limit the possibility of generating government revenue via taxes. They also recommend, for example, to explicitly limit the base on which taxes can be levied.3 Direct democracy institutions can also be interpreted as an attempt to alleviate the principal–agent problem between citizens and politicians by enabling citizens to make some decisions on their own instead of entrusting them to the representatives (their agents). The hope that one could be able to restrict politicians via an adequately designed contract has been evaluated quite critically by Tullock: “The view that the government can be bound by specific provisions is naive. Something must enforce those provisions, and whatever enforces them is itself unbounded” (1987: 317f.).

1.3.3 The Constitutional Contract As a Precommitment Device Precommitment can be used as an instrument against one’s own (short-term) weakness of will. The most famous example of using precommitment in this way stems from antiquity: Ulysses had himself bound against the mast in order not to succumb to the singing of the Sirens. In economic parlance, the problem of akrasia, that is, lack of self-control, is often discussed under the heading of time-inconsistent preferences. While principal–agent theory contains a modification of traditional theory by taking account of the fact of asymmetrical information, the precommitment approach contains a modification with regard to the imputed rationality of the actors. If individuals as well as entire groups can be subject to time-inconsistent preferences, one could ask whether societies are capable of protecting themselves from those inconsistencies by precommitment devices. This would presuppose that (a) the members of the respective society know their own weaknesses, (b) the members dispose of an adequate technology for precommitting themselves, and (c) a vast majority of the society's members – including its politicians – are willing to restrain themselves by such mechanisms. From a normative point of view, one could ask why a society at one point in time should be able to bind the society at a later point in time. Since it could be made up of different individuals, one could even question whether one would still be dealing with the same society. Additionally, 3

In a follow up to the original Leviathan model, Brennan and Kliemt (2019) point out an important gap in the original model. If governments are constrained in levying taxes, they are likely to turn to regulation instead. If they are able to “sell” it, supplying regulation can make themselves better off. For citizens, the costs of regulation might, however, even be higher than those of taxation.

1.3 Concepts of the Constitution

9

preferences might also shift. In those cases, flexibility would be preferable to precommitment from an ex ante perspective. (Elster (1984) discusses the Ulysses procedure and presents various devices used throughout history; the work by Holmes (1988) contains a discussion of the normative issues at stake). Elster (2000) observes that, in reality, people regularly want to bind others, but rarely themselves. This is a partial withdrawal from his earlier position that societies do, in fact, use constitutions to commit themselves and that this is normatively desirable. Since constitutions are often written in times of crises, it is rather questionable whether constitution-makers are sufficiently dispassionate to have the long-term interests of their societies in mind. Moreover, even if feasible, precommitment might not be desirable since it can easily clash with efficiency and/or democracy.

1.3.4 The Constitution As a Bundle of Conventions Some constitutional scholars argue that the constitution cannot meaningfully be conceptualized as a contract, rather that it is comparable to social norms which emerge unintentionally and which are accepted by most members of society in a general and unconscious way (e.g., Hardin 1989; Ordeshook 1992). Constitutionmaking can then be thought of as an attempt (i) to hasten the process of the emergence of conventions and (ii) to guide the process in a certain direction. Hardin writes: “Establishing a constitution is a massive act of coordination that creates a convention that depends for its maintenance on its self-generating incentives and expectations” (1989: 119). Since the concept is developed out of dissatisfaction with and in explicit contradistinction from the “constitution as contract” notion, here are four differences between the two concepts. First, a contract serves to solve a prisoner’s dilemma, whereas a constitution (thought of as a bundle of conventions) serves to solve a coordination game that can, however, include a considerable degree of conflict. Second, for a contract to be valid, the explicit consent of the contracting parties is necessary, whereas a constitution can be viable without the explicit consent of a majority as long as there is no serious opposition. Third, the enforcement of a contract is usually secured by the availability of external sanctions, whereas a constitution is secured by the immense difficulty of establishing an alternative constitution. Fourth, contracting parties frequently try to take as many contingencies as possible into consideration, which often gives contracts a static character, whereas constitutions are more easily amenable to evolution. Recently, this concept has been pushed forward by Hadfield and Weingast (2014), who start by observing that constitutional rules need to be enforced somehow. Since there is no level of law that is more basic than the constitution, they assume that the most important actors in the enforcement of constitutional rules are the citizens. A citizen who observes that somebody else’s constitutional rights are being infringed upon by the government could oppose that government. But such action comes at a

10

Conceptual Foundations

cost. Hadfield and Weingast argue that the likelihood that such opposition will ever come about is mainly driven by the beliefs about other people’s likely actions. The beliefs, in turn, are a function of the law relying on distinctive reasoning and on a set of legal attributes. The distinctive reasoning needs to be as clear as possible about what is “constitutional” and what is not. People need to know whether the government stays within the confines of the constitution or not, and such distinctive reasoning can facilitate the establishment of such a clearly identifiable line. According to Hadfield and Weingast (2014), many constitutional traits that have been identified by legal philosophers as normatively desirable traits of law can be better understood if their function in facilitating the solution of coordination problems between the citizens is given center stage. Lon Fuller (1964), for example, famously argued that law should be (i) general, (ii) publicly promulgated, (iii) prospective (i.e., not retroactive), (iv) clear, (v) consistent (i.e., not contain any contradictions), (vi) practicable (i.e., not demand the impossible), (vii) constant over time, and (viii) congruent with the actions of officials. Hadfield and Weingast (2014) now argue that most of these traits are a necessary condition for citizens to form homogenous beliefs about whether government behavior is constitutional or not, and whether they themselves should invest in costly opposition. If rules are not clear – to take only one concrete example – the likelihood of broad opposition against the government is low because it might not be clear whether the government has reneged against the constitution or not.

1.3.5 The Constitution As the Result of an Evolutionary Process Buchanan, as an adherent of the social contract notion of constitutions, conceptualizes individuals as rational actors who have enough foresight to recognize that rules and their implementation have advantages. Because of this foresight, rational actors are willing to consciously constrain their possible range of actions. Hayek, as the best-known representative of an evolutionary notion of constitutions, turns this concept upside down: Man became rational only after he began following some rules. According to this notion, the emergence of (constitutional) rules is not the result of an explicit process of rational choice but the unintended result of human action, rather than of human design. Hayek extends this notion by suggesting that the actors are unable to establish a comprehensive theoretical knowledge about how rules function. Instead, he repeatedly stresses that actors dispose of subjective knowledge that can only partially be communicated and, therefore, cannot be aggregated (e.g., 1945; he developed his notion of the constitution in 1973, 1976, and 1979). Although both Buchanan and Hayek share the assumption of methodological individualism, Hayek stresses that individual actors rely on the surrounding group. In his eyes, trying to rationalize the existence of order in society by drawing

1.3 Concepts of the Constitution

11

on the Hobbesian war of everybody against everybody is a nonstarter, because survival can only be thought of as occurring in groups (1988: 12). The concept of the constitution as the result of social evolution does not rest on “as if” explanations but on descriptions that claim historical accuracy. Constitutional rules are seen as the result of a trial-and-error process. The constitutional rules that survived the evolutionary selection process must have done a better job regulating group interactions than competing rules that did not survive. Part of the selection process, which takes place simultaneously, is groups observing each other. Those groups that are perceived to be more successful are imitated, supposedly implying that their rules will also be imitated. Hayek does not offer an explicit measure of the efficiency of a constitution, but he seems to be willing to derive the adequacy of an existing rule-set from the (relative) capability of a group to grow in size and feed its members. The group observational part of the selection process just mentioned has received the attention of economists under various guises and names. Some have called it “institutional competition,” while others have called it “systems competition.” More recently, the idea that institutions could spread via “spatial diffusion” has received some attention. These guises and ideas are broadly in line with the notion of the constitution as the result of an evolutionary process. As mentioned above, the English Constitution can be thought of as the outcome of an evolutionary process. It is tempting to speculate that in developing this theory, Hayek was inspired by English history, whereas Buchanan was inspired by US history. Congleton (2011) describes the constitutional development of a number of Western parliamentary democracies as an evolutionary process from the point of view of constitutional economics. Interestingly, he also shows that the United States, rather than having created a constitution from scratch, fits the evolutionary model rather well.

1.3.6 Summary Some of the concepts briefly discussed are quite complement, rather than being mutually exclusive. The first three approaches are united by their (implicit) assumption that a constitution can be created from scratch. It is in some of the central assumptions that these approaches differ from each other. Social contract theory not only assumes far-reaching rationality of all actors involved but also a capacity of all members of society to make credible promises, implying the capacity to bind oneself. The principal–agent approach introduces the possibility of asymmetrical information between principals and agents. The notion that a constitution is a precommitment device introduces the possibility that actors might suffer from time-inconsistent preferences or lack of self-control. In game theory, a convention is a solution to a pure coordination game which has more than one equilibrium in pure strategies. Once reached, those interacting are

12

Conceptual Foundations

unlikely to move away from a convention and it can become self-enforcing. Analyzing constitutions as conventions thus focuses on the necessity that they need to be self-enforcing. Thus, conceptualizing constitutions as convention puts the most important shortcoming of social contract theory center stage.

1.4 The Tools 1.4.1 Theory An assumption entirely undisputed in modern economics is methodological individualism. If all the processes and outcomes on the constitutional and postconstitutional level are the product of human action, they must – in principle – be explainable by pointing towards individual behavior that brought them about. From this assumption, it follows straightforwardly that the behavioral model focuses only on the behavior of individuals. The behavioral model used in constitutional economics should be general enough to be applicable to all those actors who are potentially relevant for the creation, modification, interpretation, and implementation of the constitution. This set of actors includes a wide range of individuals. There are scattered and unorganized individuals, representatives of organized interest groups, voters, politicians representing political parties and various branches of government, as well as judges who are assigned the task of interpreting the constitution. In economics, the behavior of the actors is usually modeled as the result of their utility function and the means available to maximize this function. Preferences are assumed to be stable over time. The “means” component not only includes their monetary resources, but also the theoretical, technological, and situational information that the actors are assumed to have, as well as the capability to process this information (their “rationality”). If these elements are sufficiently specified, one possesses a behavioral model. Homo economicus has been criticized at least as much as it has been hailed by economists. Homo economicus, as conventionally portrayed, is endowed with a utility function that is maximized rationally. Under these assumptions, it is not only irrational to vote in large number settings (because the likelihood of casting the decisive vote is very close to zero but the costs of participating are larger than zero), but it is equally irrational to acquire information about the working properties of alternative constitutional rules. Chapter 4 of Brennan and Buchanan (1985) contains an ardent defense of the assumption that all individuals, including politicians and constitution makers, strive to maximize their utility. The authors aim at convincing their critics that this assumption is not only adequate for market contexts, but also for the behavior in non-market ones. Their main argument is concerned with

1.4 The Tools

13

methodological consistency. Since it is the same individuals who act in market and non-market settings, it would be inconsistent to operate with two (or more) different behavioral assumptions. This, however, is only an argument for always using the same model, but not necessarily for using homo economicus. For their substantive defense, they claim that a behavioral model that is based on average behavior overestimates the benefits produced by altruists and, correspondingly, underestimates the costs produced by egoists. The individuals who are evaluating alternative institutional arrangements would, thus, be smart if they chose a worst-case scenario. In other words, Brennan and Buchanan explicitly opt for an assumption that might well be descriptively wrong. The imputed asymmetry would lead the individuals in the constitutional choice-situation to behave risk-averse. Although they might actually very well be risk-neutral, the purported asymmetry would lead them to behave as if they were risk-averse. In subsequent writings, either with different co-authors or on their own, Brennan and Buchanan have both somewhat modified their positions. Brennan explicitly rejected the central notion of the 1985 argument, behavioral neutrality between market and political behavior (Brennan and Lomasky 1993: 14 and passim). At the end of the day, behavioral models have a straightforward purpose, namely to reduce complexity and to put all the emphasis on a limited number of core assumptions. It is the purpose of a model to abstract from many traits; otherwise, it would simply be a replication of reality. Spelling out one’s assumptions explicitly is a precondition for producing theoretical conjectures regarding the effects of some constitutional rules. Suppose we have produced a theoretical model that allows us to generate well-specified predictions. The next step then consists in testing these theoretically derived claims empirically.

1.4.2 Empirics When we are interested in the effects – economic or other – of constitutional rules, we hope to be able to identify some constitutional rule x causing some effect y. Unfortunately, establishing causation is particularly difficult in constitutional economics because the variable hypothesized to cause some effect is never truly exogenous. All constitutional rules are the result of some more or less explicit deliberation, bargaining and collective decision-making process. To establish causation, econometricians have established randomized experiments as a kind of gold standard. This would imply “treating” some countries randomly with a particular constitutional rule (or an entire constitution), whereas other countries remain untreated. If countries are chosen randomly, other factors (like their

14

Conceptual Foundations

geographical location, their resource endowment, and so on) are unlikely to distort the findings. Econometricians would argue that under randomization, selection bias is negligible or even non-existent. Now, in constitutional economics, randomized experiments appear next to impossible for two reasons: first, sample size in cross-country studies is limited to around 200 observations, simply because there are no more than 200 countries in the world. Second, it is unlikely that some 100 countries will pass a particular constitutional rule, simply because some constitutional economist is trying to establish causation via a randomized experiment. If randomized experiments as the first-best strategy are unavailable, one can try to mimic them as closely as possible. One option would be natural experiments, i.e., instances in which constitutional rules were implemented, thanks to “mother nature.” Some exogenous event might take place that can be used to make causal inferences. Again, the availability of natural experiments regarding constitutional provisions is very limited, if they exist at all. In such cases, most researchers proceed by relying on observational data. These come with many problems, but econometricians have developed lots of useful tools to mitigate these problems. Selection bias was already mentioned; omitted variable bias and reversed causality are other examples of such problems. The most traditional way to deal with endogeneity is to rely on instrumental variables. Finding convincing instruments is always a problem. In constitutional economics, complying with the exclusion restriction seems to be particularly difficult, or unlikely. After all, the instrument is required to affect the dependent variable exclusively via the variable potentially causing the effect. It is unlikely, however, that this is ever the case. Until recently, tools, such as matching or difference-in-differences, seemed difficult to apply in constitutional economics. The idea in matching is to match two countries that are completely alike except that only one has been treated with the constitutional rule under scrutiny. Among less than 200 candidates (the number of countries worldwide), it is not very likely to ever find any perfect match. A technique that has been developed more recently might come to the rescue here, namely, the synthetic control approach. Here, the treated country is not matched with another country existing in reality, but with a synthetically created country that has been created such that it mimics the treated country as closely as possible. Metelska-Szaniawska (2016) applied the synthetic control approach to identify the economic effects of post-socialist constitutions in Central and Eastern Europe. Another technique that recently found its way into constitutional economics is regression discontinuity. At times, the application of constitutional rules depends on rather exogenous factors such as city size. For example, signature requirements for kicking off initiatives might decrease in the number of registered citizens, i.e., a

1.5 More Potentially Useful Tools

15

population threshold. If city size is assumed exogenous, admittedly a debatable assumption, then some aspects of direct democracy institutions can be analyzed more reliably. However, regression discontinuity is usually not used in cross-country settings but in within-country studies.4 Given all of these difficulties, and more could be named, Spamann (2015: 138) came to the rather sobering conclusion that “comparative evidence alone will hardly ever be sufficient to establish a causal claim and that statistical methods that purport to do so are likely to do more harm than good in comparative settings.” Fortunately, his overall evaluation did not stop there. The next sentence is: “Even if comparative data cannot identify any single causal theory, however, they are extremely important in narrowing down the set of plausible theories. A thoughtful pursuit of this more modest agenda seems to me the most fruitful avenue.”5 In other words, comparing constitutions across countries can be an important endeavor, even after explicitly taking into consideration the severe methodological challenges that this enterprise is confronted with. In the next section, I discuss possible tools of this enterprise in a more general fashion.

1.5 More Potentially Useful Tools The implicit assumption underlying the last section is that econometrics is the standard tool of positive constitutional economics. We discussed a number of the particular challenges that researchers are confronted with when inquiring into the effects of constitutions. It is also important to note that many details are necessarily ignored when we rely on econometrics. In this section, I discuss a number of tools not suffering from these pitfalls. But alas, they suffer from other shortcomings. Over the last decade or two, experimental economics has made huge progress. In this section, we discuss the possibility of using experiments and case studies to gain new insights for constitutional economics.

1.5.1 Experiments For a long time, experiments were primarily conducted in the laboratory. The lab has the advantage of providing conditions where interactions that take place can be precisely controlled. From one treatment to the other, it is possible to modify a single dimension. This ability greatly increases our capacity to identify causal mechanisms. 4

5

A number of studies relying on regression discontinuity will be cited in Chapter 4 of this primer. Eggers et al. (2018) not only provide an overview of studies relying on population threshold, but also discuss pitfalls and possible solutions of this approach. Spamann (2015) makes these claims not with regard to constitutional economics, but with what he calls “empirical comparative law.” If constitutional economics is empirical and relies on cross country studies, it could be considered part of empirical comparative law.

16

Conceptual Foundations

However, the validity of taking the results from experiments conducted in the lab and applying them to the real world has been an issue for many years. Of late, field experiments have, therefore, gained ground (see, e.g., List 2011). The idea is “to treat” a number of randomly selected entities with the variable of interest and then compare the effects of the treated with the non-treated entities. If the number of entities taken into consideration is large enough, their heterogeneity should not be a problem. “Entities” can be school districts, villages, and so on. In principle, both lab and field experiments can be used in constitutional economics. Both types of experiments can be used to ascertain the effects of constitutional rules, as well as the types of rules people are willing to agree upon. Normative arguments often rest on strong assumptions about the basic constitutional principles people can agree on. For example, Rawls (1971) elaborates two principles, the first one being the principle of equal liberty according to which each person should have an equal right to the most extensive liberties compatible with the same liberties for everyone else. The second principle is the much-acclaimed difference principle according to which rules should make sure that social and economic inequalities are reduced in such a way that the least advantaged person benefits most. Rawls argues that rational persons would choose these two principles if they were behind a veil of ignorance and the principles would be a manifestation of the people's concept of fairness. The deduction of the difference principle is logically derived from and based on precisely specified initial conditions. The derivation itself can, therefore, claim positive status and can be formulated as follows: if certain initial conditions are given, then people will agree on the difference principle as an expression of their value judgments concerning fairness. If one is able to approximate the initial conditions in a laboratory setting, the hypothesis becomes testable. Rawls’ difference principle has, indeed, been subject to such a test. It was found that people overwhelmingly preferred another principle, namely, maximizing the average income with a floor constraint, i.e., a minimum income level that everybody was entitled to (Frohlich et al. 1987).6 More recent studies include the attempt to recognize “constitutional” choice by Sutter et al. (2010), in whose game individuals decide on the kind of institution under which they want to play before actually playing. Kocher et al. (2006) incorporate the choice between individual and collective decision-making in their game. This can be interpreted as the choice between direct and representative democracy. Laboratory experiments can also be used to ascertain the effects of constitutional rules. One recent example is an experiment that studies the effect of the vote of no confidence procedure on the level of public good provision (Tergiman 2015). This type of experiment can be particularly useful when the introduction of a new 6

Voigt (2015) provides a section summarizing experiments that have been conducted in conjunction with the notion of a veil of ignorance or uncertainty.

1.5 More Potentially Useful Tools

17

constitutional rule that has never been used in a similar context is considered. Such experiments are really tests of rules hitherto untested in reality. They might not only improve our capacity to predict the degree to which the intended effects will materialize, but might also draw attention to hitherto neglected side effects. To date, not many field experiments relating to constitutional economics have been conducted. This is not surprising as experimenting with school vouchers, contraceptives, and other devices seems to be easier than experimenting with constitutional rules. The work of Beath et al. (2017) is a rare exception. They convinced an Afghanistan government agency with the authority to allocate resources to villages based on the preferences communicated by those villages, to participate in a field experiment. The experiment consisted of using an alternative method of eliciting citizen preferences. The time-honored method of eliciting preferences in Afghanistan is to rely on consultations conducted with members of the village elite. The treated villages, however, gave all adult residents a voice and the right to vote by using a referendum. In the experiment, half of 250 villages were randomly assigned the referendum procedure, and the remaining half the traditional one. In a sense, the field experiment is an empirical test of the frequently published conjecture that direct democracy devices reduce the principal–agent problem between political decision-makers and citizens by moving political decisions closer to the preferences of the median citizen. The authors find that referendums do, in fact, reduce the influence of local elites over the types of project chosen, as well as their actual location in the village.

1.5.2 Possible Consequences of Constitutional Rules Economists are often suspected of lightheartedly ignoring potentially important details. These might refer to the details of constitutional rules, but also the detailed context in which a rule is created, intended to operate, and so on. Case studies are hailed by many as an important investigative tool. The promise of using “mixed methods,” of gaining insights by mixing quantitative with qualitative tools, seems almost too good to be true. Although there is some established wisdom on how to do case study research (see, e.g., the excellent introduction by Gerring 2006), little is known on how cross-country and case study research can best complement each other. One possibility is to analyze those countries identified as outliers in crosscountry research more closely. What is it that makes them perform so much better or worse than we would expect according to the rest of the sample? In the analysis of constitutions, case study research has been something like the norm. After all, most publications analyzing the constitutional history of a particular country are case studies. But there are also case studies that are, more or less explicitly, informed by the economic approach, and we mention a small selection here. Beard (1913/1986) analyzes the birth of the US Constitution. Elster (2000)

18

Conceptual Foundations

compares the process of constitution-making in the United States to that of the revolutionary constitutions in France. A more recent example of such pairwise comparison is provided by Raudla (2010), who compares the preferences of the constitution-makers in the United States (1787) with those in Estonia (1991–2). A small number of collected volumes contain case studies describing constitutionmaking experiences in various countries at different times (see, e.g., Goldwin and Kaufman 1988; Reynolds 2002; Miller 2010). Recently, more advanced techniques such as “qualitative comparative analysis” have been promoted by Wagemann and Schneider (2010). Although the precise epistemic value of these approaches is unclear, applying them to constitutional questions promises to be exciting.

1.6 Consequences of Constitutional Rules When constitutional rules are assumed to be exogenously given, they serve to explain different (economic) outcomes. But what specific outcomes most interest us? For economists, it seems obvious to analyze the relationship between the constitutional rule system and per capita income or its growth. It has been argued (e.g., by Hall and Jones 1999) that growth fluctuates widely over time within countries and that the results of any study are, thus, heavily dependent on the time period analyzed. One way to circumvent this problem is to use per capita income as an outcome variable as it reflects aggregate growth over a very long period. Another way to circumvent the problem is to use labor and/or total factor productivity as an outcome variable. Conceptually, this makes a great deal of sense since we are not interested in output increases achieved by augmenting capital or labor input, but in increases in output that are achieved by using constant amounts of these factors in a more productive way. Institutions, including constitutional rules, should, thus, display their effects primarily via increases in productivity.7 In reality, however, these variables are full of problems, some of which are discussed in Section 1.7. Growth, per capita income, labor, and total factor productivity are very broad measures. Assuming that constitutional rules have a significant impact on these outcome variables, we are very interested in the transmission channels through which these impacts occur. This means that we are interested in some variables that are less encompassing. A first take could be to ask if the readiness of individuals to invest into specific human capital is a function of the institutional systems, as suspected by Iversen and Soskice (2006). They hypothesize that proportional representation could serve as insurance against the vicissitudes of life.

7

This is not to deny that adequate institutions will also increase the propensity to invest in both physical and human capital.

1.6 Consequences of Constitutional Rules

19

It would, hence, be interesting to use the degree of specific human capital as a dependent variable. Fiscal policy variables are a natural starting point for trying to establish a transmission channel between constitutional rules and macroeconomic outcomes. The state has the power to tax, but it also has the power to provide public goods. The extent and efficiency of both taxation and the provision of public goods might be affected by constitutional rules. It, therefore, seems appropriate to add (1) tax revenue, (2) government spending, and (3) the structure of government spending (reflecting the kinds of public goods provided) to our list of variables of interest. Up until now, the focus has been on the size of social and welfare spending, but there are many additional aspects that warrant inquiry. These include the question of whether social welfare spending actually reduces overall inequality. The distinction between general and targeted public goods is also of interest to see if different types of electoral systems actually display the hypothesized effects. Government output, in terms of public goods provided, is a direct consequence of fiscal policies. We propose looking at some very generally accepted policies, such as fighting poverty and providing for a public education system and a clean environment. Inquiring into the structure of state spending could also imply comparing the percentage of expenditure spent on investment. It also seems worth remembering that governments can affect economic development in ways other than budgetary. Regulation is a case in point. It would be interesting to see if some constitutional institutions are systematically correlated not just with regulation per se, but also with different kinds of regulation. Government efficiency in the provision of public goods should also be reflected in low levels of corruption, few political rents handed out via regulation, and so forth. If constitutional rules have a significant impact on these variables, it implies the existence of an indirect channel through which the rules have an effect on income and total factor productivity. Finally, there are a number of variables usually treated with some suspicion by economists, including fuzzy concepts such as the legitimacy of the constitution or its “stability.” Nevertheless, we plead for their inclusion as outcome variables. If it can be shown that some constitutions – or the procedure by which they are generated – are systematically perceived as more legitimate than others, then this might cause important economic effects. For example, the state’s monitoring costs would be less, which could mean lower taxes and higher income. Constitutions perceived as more legitimate could lead to a lower propensity of citizens to cheat on taxes. Likewise, if the stability of a political regime is significantly influenced by its constitution, this would amount to a hitherto neglected transmission channel (from constitutional rules via stability and investment to income). A further outcome of interest could be the degree to which constitutions enable citizens to form expectations that have a high chance of being correct. Institutional economists stress the uncertainty-

20

Conceptual Foundations

reducing function of institutions. Constitutional rules that increase predictability might have important economic consequences, such as citizens acting on longer time horizons and having a greater propensity to make long-term investments. Another variable that promises to be of interest is the level of reported life satisfaction (also known as “happiness”). Over the last two decades, quite a few economists have become more skeptical of the strict focus on income and its growth, and have argued that happiness would not only be a more encompassing but also a more useful concept (Frey and Stutzer 2018 is a recent survey). Table 1.1 identifies and summarizes four groups of potentially interesting outcome variables. Two measurement issues will be discussed in the next section. First, we discuss issues with measuring the four groups of outcome variables just discussed. Table 1.1 Overview of outcome variables Dependent Variables I 1 2 3 3a 4 II 1 2 3 4

Macroeconomic Variables Economic growth Per capita income Labor productivity Degree of specific human capital Total factor productivity Fiscal Policy Variables Government revenue Government expenditures Budget deficit Composition of government spending

5

Public goods provided

5a 5b

Absence of poverty Quality of public education Quality of the environment Regulation

5c 6

Potential Problems with Most Frequently Used Proxies

Often very volatile over time; period of analysis chosen crucial Both labor and total factor productivity are calculated on a uniform Cobb–Douglas production function for all countries; the measurement of “specific human capital” is a serious challenge

Focus is often on central government, which might considerably underestimate overall state activity Focus on social and welfare spending; measures that distinguish between general public goods and others highly desirable Delineation between general and targeted public goods often not clear-cut Dealing with omitted variable bias a serious challenge

Regulation as alternative to fiscal policy

21

1.7 Measuring Constitutions

Table 1.1 (cont.) Dependent Variables III 1

Governance Variables Corruption

2

Efficiency of government

IV 1

Others Legitimacy of constitution/state Propensity to cheat on taxes Reported life satisfaction (“Happiness”)

2 3

Potential Problems with Most Frequently Used Proxies Subjective value; all the measures available for a large number of countries measure perceptions and are highly subjective Subjective value; measure based on a variable number of surveys; highly subjective and drivers difficult to determine

Subjective indicator; uncertain whether people really act on declared beliefs Subjective indicator; uncertain whether people really act on declared beliefs Subjective indicator; unclear how systematic collective sentiments (pessimism/optimism) can be controlled for

Second, and more importantly, we turn to the issues related to the difficulty of measuring the constitutional rules themselves.

1.7 Measuring Constitutions Empirical research interested in ascertaining the effects of constitutions presupposes that both constitutional rules and outcome variables are measurable in a satisfactory fashion. Research interested in identifying the factors that lead different societies to choose different constitutional rules also depends on the reliable measurability of constitutional rules. In addition, potential determinants now need to be measurable as well. In this subsection, I deal with both constitutional and outcome variables, beginning with the measurability of constitutional rules. Prima facie, ascertaining currently valid constitutional rules is not a problem, since constitutions (almost) always exist in written form.8 Yet, a formally unchanged section of the constitution can be subject to very different interpretations over time (Voigt 1999: chapter 7). However, let us assume that we are not dealing with constitutional interpretation, but simply with constitutional basics. The more relevant problem now is the potential divergence between a constitution’s formal provisions and their actual implementation. 8

With the United Kingdom, New Zealand, and Israel being the best known exceptions.

22

Conceptual Foundations

Many real-world constitutions promise paradise on earth. Extensive negative rights (protecting citizens from intrusion by the state) are happily married to extensive positive rights (promising citizens everything from the right to a paid nap at work, to the right to housing, employment, and a clean environment). Of course, many of these protections and promises never manifest. Thus, when trying to estimate the consequences of constitutional rules, a formidable problem arises as the economic effects will most likely not be the result of these unenforced constitutional protections and promises, but are far more likely to be due to the way the constitutional rules are actually implemented. Further complicating the issue is that unenforced constitutional provisions are not confined to human rights, but may well involve the organizational structure of the state. For example, the constitution could declare the judiciary independent of the executive but, in fact, this is not always the case.9 A related problem occurs when constitutions contain provisions that can be exercised, but rarely are (e.g., referendums). Some would argue that simply by being possible, these provisions have an effect on political behavior, and this is certainly plausible. However, this assumption might be proven wrong empirically. This is why it is important to not only measure formal provisions, but also investigate their actual use. To date, the most extensive endeavor to code constitutional rules is the Comparative Constitutions Project. The creators of this huge dataset have produced a list of some 650 different variables according to which every available nation state constitution promulgated since 1787 is coded on an annual basis. Overall, 900 constitutions have been coded. This dataset has quickly established itself as a central source for scholars interested in comparing constitutional text. Elkins et al. (2009) is a booklength presentation of the dataset. The Comparative Constitutions Project is interested in constitutional text, i.e., in the de jure side of constitutions. There is no perfect equivalent on the de facto side, but the so-called Varieties of Democracy Project coordinated at the University of Gothenburg is likely to fill in at least some gaps. “V-Dem,” as it is often abbreviated, has collected de facto data for many relevant issues for up to 202 countries since 1900, with plans to extend coverage back to 1850 going back as far as 1789. Many of the measures used in the empirical papers discussed in the next chapters are dummy variables: a country has a presidential or a parliamentary form of government; it has a unicameral or a bicameral legislature; it has a federal or a

9

Of course, one can simply count instances in which de jure rules have not been enforced factually: presidents running for a third term although there is a two term limit, state governments overturned as a consequence of the national government demanding so, and so forth. It is unclear, however, whether this would produce a complete measure of the degree to which de jure rules are factually enforced.

1.7 Measuring Constitutions

23

central structure; and so forth. By necessity, dichotomous variables do not allow taking institutional detail into consideration. But if God (or the devil!) is in the details, perhaps the details are worth a little of our time, and the construction of more precise indicators is a worthy pursuit. Let us now take a look at the outcome variables. Particular emphasis is given to the variables used by Persson, Tabellini, and various co-authors (PT) (2003). PT gave CPE a real boost. It seems appropriate to take a critical look at the outcome variables they employ, because many scholars have followed their lead and have used the same variables. Estimates of labor and total factor productivity must, by necessity, rest on sweeping assumptions. PT use estimates arrived at by Hall and Jones (1999), who rely on an identical Cobb–Douglas production function for all countries. It is assumed that all countries have an identical α (= 1/3), that marginal returns from (school) education are decreasing but identical worldwide, that the amount of investment is known, and that its depreciation is also identical worldwide, and, finally, that GDP is known. On this basis, the Solow residual can be calculated. Obviously, a very crude measure results.10 For government revenue, PT rely on central government. Their chief reason for not relying on total government revenue is “data availability and comparability” (2003: 38). It is conceivable, though, that central government revenue might vary considerably depending on whether the state is federal or unitary. PT control for this with a dummy for federalism, but the control can only be as good as the dummy. Similar arguments can be made with regard to (central) government expenditure and budget surplus, since the last variable is based on the two previously discussed variables. To analyze the effects of constitutional rules on the structure of the budget, PT construct a variable that captures the amount of central government expenditures on social services and welfare as a percentage of GDP. The same criticism, however, applies.11 When it comes to the so-called governance indicators, PT’s true interest lies in the amount of political rents. However, such a measure is not available, so in lieu thereof, they rely on various corruption perceptions indices. Of course, political rents are not identical with perceived corruption, and thus constructing political rent indicators remains a desideratum. There is no getting around that the perceptions indicators are subjective and that subjective perceptions can be heavily influenced by recent events,

10

11

Hall and Jones calculate productivities for 1988. Using their assumptions, we recalculated the measures for the year 2000 (Blume and Voigt 2007). Partial correlations are 0.972 and 0.899 for labor and total factor productivity, respectively. PT choose social and welfare spending as a proxy for broad redistributive programs as opposed to narrow geographical constituencies. They point out (2003: 50) that this interpretation of the variable might not apply to less developed countries. More fine grained indicators are certainly desirable: if one thinks of government spending as either allocation improving or redistributive, on the one hand, and broad (nation wide) or narrow (local, geared at particular districts, etc.), on the other, then a 2  2 matrix results in which social and welfare spending would cover only one cell.

24

Conceptual Foundations

even by recent reports in the press of a long-ago corruption scandal. Objective corruption indicators have been constructed (e.g., road construction in Italy [Golden and Picci 2005], or Indonesia [Olken 2007]), but it hardly seems possible to construct them on a broad basis for many countries, and so this area continues to be a desideratum.12 To estimate the effect of constitutional rules on the efficiency of government, PT rely on the governance indicators provided by the World Bank Institute. Alongside government efficiency, they include a cluster called “graft.” Both indicators have been passionately attacked. The main critique is that they are not based on a thoroughly systematized concept but that the implicit definition of the various concepts is based on the availability of surveys. Since the variables employed change over time, so does the underlying implicit definition of the concept purported to be measured, which makes comparison over various years impossible. The indicators are not supplied as absolute values but as rankings. Once the number of included countries changes, the ranking of a country can change without any substantial change having taken place within that country.13 In sum, neither the operationalization of constitutional rules nor that of the potential outcome variables is without problems, a situation that must be kept in mind when interpreting the estimation results. However, it is also a situation that should encourage the construction of better variables. This chapter is entitled Conceptual Foundations. Its purpose is to make the reader familiar with the various concepts of the constitution that have been used by economists, and to discuss a number of challenges that scholars face when analyzing the effects and determinants of constitutions. The next chapter now asks whether democratic constitutions can be easily told apart from autocratic ones and, if so, in what aspects they systematically differ from each other.

12 13

Gutmann et al. (2019) compare perception based corruption indicators with experience based ones. Critics include Arndt and Oman (2006), Knack (2006), Thomas (2006), and Kurtz and Schrank (2007a, 2007b). Kaufmann et al. (2007) is a reply to critics.

2

Democracy vs. Autocracy

Before we start inquiring into the economic effects of single constitutional traits in considerable detail in the next chapter, it seems to make sense to take a look at the broader picture first. By the broader picture, we mean whether democratically constituted states experience economic outcomes that are systematically different from those experienced by autocratically constituted ones. Theoretical conjectures, as well as empirical insights on this question, occupy the first half of this chapter. The second half is devoted to the question of why some countries are democratic and others autocratic in the first place. Recently, searching for factors determining a change in regime – in either direction – has received considerable attention, which is why we spend some space on reporting results regarding that question here. During the history of mankind, democracy has been the exception rather than the rule. And even today, autocratic regimes clearly outnumber democratic ones. Dealing with the economic effects of both democracy and autocracy is, hence, warranted. In a first take, much institutional detail is explicitly discarded and the focus is on this simple dichotomy.1 According to a dataset by Geddes et al. (2014: 313), of all regime changes that have occurred since the end of World War II, more than half of them were transitions from one kind of autocracy to another kind of autocracy. A change in government is called a regime change here if there is a change in the group of decision-makers. It could be a change regarding the ethnic belonging of the decision-makers, their geographic origins, their religious beliefs, and so on. These issues are highly relevant empirically as most regime changes are not from autocracy to democracy, but from one autocratic regime to another, like Shah Reza Pahlewi being replaced by Ayatollah Khomeini in Iran in 1979.

1

In principle, an even more basic question also deserves attention, namely the one inquiring into the processes that transform anarchy into statehood. Social contract theory is likely to predict the transformation from anarchy to democracy, but this does not seem to be the most frequently observed transition. Inquiring into the impediments to leaving anarchy is relevant to better under stand the situation of failed states. The Fragile States Index, which has been issued by Fund for Peace on an annual basis ever since 2005, lists no fewer than 60 countries belonging to one of the categories between “high warning” and “very high alert” in its 2019 edition.

26

Democracy vs. Autocracy

2.1 How Best to Separate Democracy from Autocracy? If we are interested in empirically ascertaining how the effects of democracy differ from the effects of autocracy, we need to rely on some sort of definition of both political realities. Gallie (1956) coined democracy as “an essentially contested concept.” Not surprisingly, scholars continue to quarrel over an adequate definition. The creators of one of the most frequently used measures of democracy propose to follow Przeworski (1991) who classifies a regime as democratic if it meets all four of these criteria: (1) The chief executive must be chosen by popular election or by a body that was itself popularly elected. (2) The legislature must be popularly elected. (3) There must be more than one party competing in the elections. (4) An alternation in power under electoral rules identical to the ones that brought the incumbent to office must have taken place. By implication, autocracies are defined by default: in case one or more of those criteria is not met, we are dealing with an autocracy. In line with Przeworski (1991), Cheibub et al. (2010) propose a categorical distinction: a country is either democratic or not. They distinguish three types of democracies though: if the head of government is directly accountable to the legislature, a democracy is parliamentary; if the head of government is accountable to the legislature and there is a (different) head of state, a country is semi-presidential; and if the head of government is not accountable to the legislature, the country is presidential. Cheibub et al. (2010) also propose three types of autocrats, namely monarchies (where succession as head of state is hereditary), military (where the head of state is – or was – a member of the military), and civil (no membership in the military). Looking at the implications of such a delineation reveals a number of surprises. For the year 2008, for instance, South Africa was coded as a (civil) autocracy, supposedly because Condition 4 had not been met. The dataset by Cheibub et al. (2010) has been updated recently by Bjørnskov and Rode (Forthcoming). Another frequently used indicator is the one developed by Marshall et al. (2018). This indicator relies on a continuous measure rather than the categorical distinction drawn by Przeworski and Cheibub. The dataset developed by Marshall et al. (2018) contains one variable that codes all independent states on a 21-point scale from 10 (“perfect autocracy”) to 10 (“perfect democracy”). If one is interested in long-term analyses – for example, the spread of democracy over time – this dataset is very helpful because it dates back to 1800. On the other hand, this dataset has been criticized because no one is sure what it means if a country is attributed a score of 0 or 2. Has the reliance on a continuous measure created a third category other than autocracy and democracy? Some scholars have proposed to call these countries “anocracies,” but it is unclear what exactly that means.2 2

Appendix 1 reports how the countries of the world have been coded according to these two indicators.

2.1 How Best to Separate Democracy from Autocracy?

27

For a long time, discussions centered on how best to define democracy and how to meaningfully delineate different types of democratic regimes. In the meantime, it has become apparent that there might also be systematic differences among autocracies and more scholars have tried to come up with proposals how dictatorships could be meaningfully categorized.3 Geddes (1999) is interested in different types of autocrats and distinguishes (1) personalist from (2) military and (3) single-party regimes. In personalist regimes, material rewards to a select group of insiders are exchanged in return for mobilizing political support. Sometimes, researchers add a fourth type, namely (4) monarchies. While Cheibub et al. (2010) and Geddes (1999) focus on who holds power, Wintrobe (1998) proposes to focus on the motivations – or preferences – of the autocrats. Assuming that all autocrats strive to stay in office as long as possible, he argues that they have two means of securing their position, namely by inducing loyalty from the population, or by repressing those who appear to be dangerous to their goal of staying in office. Wintrobe goes on to divide autocrats into four categories: among them, “totalitarians” and “tinpots” are dealt with at most length. Totalitarians are autocrats who enjoy maximizing power over citizens. In a sense, repression not only serves to prolong their stay in office, but they also “consume” repression because they enjoy it as such. Tinpots are autocrats who are more interested in “palaces, Mercedes-Benzes, [and] Swiss bank accounts” (Wintrobe 1990: 849).4 Wintrobe then turns to the differential effects of economic growth on the behavior of tinpots and totalitarians. By assumption, tinpots are interested only in the degree of power that allows them to remain in office. Power is “produced” by the “inputs” of repression and loyalty. If economic growth increases, subjects will be more loyal to the autocrat, and less repression is needed to secure the minimum amount of power necessary. Inversely, less growth leads to more repression in tinpot regimes because less loyalty is substituted with more repression. Totalitarians enjoy power as such and, thus, seek to maximize it. If economic growth increases, subjects will “supply” more loyalty, which enables the totalitarian to be even more repressive. Again, the opposite also holds: if economic growth declines, the totalitarian will reduce the level of repression. Although Wintrobe’s model is an important step toward a more theoretical foundation of the economic theory of autocracy, it is not without problems. The affiliation of a particular autocrat to one of the four types might not be stable over time. A dictator might first enjoy the consumption of luxury goods and then learn how much pleasure he receives from repressing major parts of the population. But this is hardly consistent with the assumption of stable preferences. Wintrobe uses his classification to make predictions regarding the behavior of the four types. If that is interpreted as theory, then it should be formulated such that it is falsifiable; 3 4

The terms “autocracy” and “dictatorship” are being used interchangeably here. The other two types are tyrants and timocrats (or benevolent autocrats).

28

Democracy vs. Autocracy

otherwise, it would lack the status of being scientific. But this implies that the classification of dictators must be independent from their revealed behavior because if it is inferred by their behavior, then the “theory” would turn out to be an unfalsifiable truism. Further, classifications are particularly useful if each of the categories contains at least a sizeable number of real-world cases. This does not seem to hold with regard to both tyrants and timocrats. In fact, the overwhelming number of dictators would appear to be tinpots. Finally, the appropriateness of categories can also be ascertained by looking at those cases that do not easily fit into one of the categories. In a more recent contribution, Wintrobe (2019) speaks of China, North Korea, and Russia as “unclassifiable” – shedding some doubt on the fruitfulness of his taxonomy himself.

2.2 What Role Do Constitutions Play in Autocracy? This primer is on constitutional economics. When speaking of constitutions, many people almost automatically assume that one is discussing democracies. In democracies, constitutions also serve to constrain unwanted effects of majority decisionmaking such as the so-called tyranny of the majority. Because it is well known that democratic decision-making is prone to a number of problems, such as timeinconsistency, some policy areas are explicitly exempt from the purview of generally elected politicians. Today, monetary policy is the responsibility of independent central banks whose governors are not accountable to the general electorate. So what role is there is for constitutions in autocracy? After all, a central function of constitutions is to put constraints on government, whereas autocrats seem to aim at being as unconstrained as possible. But precisely because autocrats are suspected of being unconstrained by any rules, they have a particularly hard time making their own promises credible. If this is a disadvantage – e.g., because people refrain from investing in that country – then a rational autocrat could wish to be constrained by constitutional rules that enhance his capacity to make credible commitments. In addition, no autocrat can truly rule as a single person but needs the support of a loyal group. As described in Chapter 1, a constitution can be thought of as a coordinating device for this group. Every autocrat also needs administrators implementing his ordinances on the ground. Another function of constitutions could, hence, be to serve as a monitoring instrument for the lower charges. To the degree that the constitution collects and explicitly names a number of aspirations, it might serve to increase the autocrat’s domestic legitimacy. Further, to improve his reputation internationally, an autocrat might have incentives to include a number of basic rights in the document. Formulated somewhat differently, constitutions in autocracies can have a commitment function, a coordination function, but can also serve as window dressing. The more important the window dressing is and the less important – conversely – the

2.3 Do Autocracies and Democracies Perform Differently?

29

other functions are, the more we should expect autocratic constitutions to read just like democratic ones (Elkins et al. 2014: 142). Assuming autocrats are power-maximizers, it seems plausible to hypothesize that autocratic constitutions allocate more power to the executive, and correspondingly less to the legislature and the judiciary. This could also imply that the judiciary is granted less independence. Assuming that constitutions are not mere window dressing exercises but are meant to be enforceable documents, one could also expect that they make fewer promises regarding all kinds of basic human rights. So can you tell an autocratic constitution from a democratic one? Elkins et al. (2014) asked exactly that question. They define authoritarian constitutions as those drafted by dictators, implying that the vast majority of their sample is coded as authoritarian, namely 695 out of 846. Controlling for the contents of previous constitutions, the location of the country and the period during which the constitution was drafted, they find that authoritarian constitutions do not delegate more powers to the executive, do not contain fewer rights, but do contain fewer provisions for the independence of the judiciary. The hypotheses just described are, hence, only partially confirmed.

2.3 Do Autocracies and Democracies Perform Differently? 2.3.1 Theory For a long time, Gordon Tullock (1974, 1987) was the only economist seriously concerned with an economic theory of autocracy. One of his central assumptions was that the ousting of autocrats by way of popular revolutions was extremely unlikely, but that the more serious threat for any autocrat would be palace revolutions. Kings would be, in other words, most threatened by their sons, who would all like to become king. But the list of people who would like to replace the current king and take over his position is a lot longer. Tullock, therefore, predicted that autocrats would take particular care to raise entry costs in the sense of making the entry of any competitor more difficult and, hence, less likely. As possible means used to raise the entry barriers of competitors, he mentioned the use of torture, death penalty, press censorship, the regulation of religion, and the amount of resources spent on maintaining an army. In a sense, the paper by McGuire and Olson (1996) takes us even further back than the contributions by Tullock. When we think of autocrats as “bandits,” we would in all likelihood think of them as reigning over a particular territory. Not so McGuire and Olson, who distinguish between “roving” and “stationary bandits” and compare the expected outcomes by evaluating both regimes against those under democracy. Roving bandits plunder area after area. If the likelihood of being plagued by roving bandits is high, incentives to invest in the production of goods that could be stolen

30

Democracy vs. Autocracy

are low. Stationary bandits, by contrast, settle in one area and live off that area’s population. Their incentives are thus entirely different from those of roving bandits as they have an interest in protecting “their” area against other bandits in order to increase the population’s incentive to be productive. Citizens’ utility levels are higher under stationary than under roving bandits. Yet, McGuire and Olson (1996) show that expected utility is higher still under democracy: any additional unit of public goods provided is costly for the stationary bandit and it can be shown that democracy, using a unanimity rule, would lead to a higher level of public goods provision as well as a higher level of income. Although McGuire and Olson argue that democracies should perform better, their theoretical results are heavily dependent on specific assumptions concerning the rules used to make decisions regarding the provision of public goods. If one is faced with the choice between a roving and a stationary autocrat, it would seem that one should wish for a stable one. Wintrobe (2019: 298) is, however, rather skeptical concerning such a conclusion: “Wouldn‘t the victims of Stalin, Hitler, Pol Pot, Sadam Hussein, or Bashir al-Assad have preferred that their bandits were a little less stationary?” One could also wonder if the Mafia is not a group of fairly stationary bandits with rather negative consequences for the utility levels of many people “protected” by it. Other arguments for a democracy having a systematic advantage over an autocracy in terms of growth prospects include its insurance function. In case citizens are unhappy with their ruler, democratic elections allow them to kick her out of office by regular means. In a sense, then, democracy can be interpreted as an insurance against bad rulers. The ability to peacefully remove a bad ruler would lead to more stability (both political and economic), which would, in turn, increase aggregate investment. Barro (1997, 2000) argues that the rule of law and secure property rights are conducive to economic growth, but that the effects of higher levels of democracy are uncertain: pressure to redistribute wealth could compromise property rights and thus reduce the incentive to work hard and invest. He does acknowledge that redistribution can increase stability by making it less attractive to engage in criminal activity, riots, or revolutions, but he argues that even autocrats would be motivated to spend enough on redistribution to prevent such activities. The difference between democracies and autocracies would be that democracies spend too much on redistribution.

2.3.2 Empirical Evidence In the early 1990s, Przeworski and Limongi (1993) collected 18 studies asking whether democracies have growth rates that are significantly different from those realized in autocracies. In these 18 studies, they found 21 different results (some

2.3 Do Autocracies and Democracies Perform Differently?

31

studies thus contained more than one result). Eight results indicate that democratic regimes grow faster, eight of them find the exact opposite, and the remaining five cannot find a significant difference in growth rates. So the answer to the question was as unclear as possible. That survey was published more than 25 years ago and since then quite a few additional studies have been published, but the empirical literature remains inconclusive. For example, in their meta-analysis, Doucouliagos and Ulubasoglu (2008: 62) analyze 483 regression estimates from 84 studies: 15% of the estimates find a negative and significant correlation between democracy and growth, another 21% find a negative but insignificant association, 37% a positive and insignificant one, and 27% both a positive and statistically significant correlation between democracy and growth. So, what to do? A number of options are available. One possible way out is to rely on adequate or (and) more refined estimation techniques. Another possible way out is to ask more precise questions. It might well be that overall growth rates do not differ significantly from each other but that, e.g., the composition of public goods provided by government does. Finally, it might well be that the distinction between democracy and autocracy is simply too coarse. There are not only very different kinds of democrats, but also very different kinds of autocrats, so another option could be to try to take these differences explicitly into account.

2.3.2.1 Use Adequate Estimation Techniques Among those who propose to use adequate estimations techniques is De Haan (2007). He does not present any new estimates but has some suggestions for how they should be run. Frequently, there is more than one measure for a certain concept (in this case, democracy) and the bivariate correlation between the various measures is often low. In such cases, de Haan opts in favor of factor analysis to identify the variation common to all measures. We now discuss a number of studies employing unconventional methods to reach more conclusive results. Roll and Talbott (2001) find that democracy (or, more precisely, the political rights indicator as published by Freedom House, which focuses on the procedural aspects of democracy), in concert with civil rights and freedom of the press, is conducive to higher income. But in all studies interested in estimating the effects of institutions, the direction of causality constitutes a serious problem; is income high because the country is democratic or is the country democratic because income is high? Usually, instrumental variables are used to control for endogeneity. Roll and Talbott (2001) choose a different path; they draw on singular events that made countries either substantially more or less democratic (“democratic event” vs. “anti-democratic event”). They find that democratic events are, on average, followed by a dramatic improvement in country income, whereas anti-democratic events typically lead to substantial reductions in it.

32

Democracy vs. Autocracy

Tavares and Wacziarg (2001) estimate a system of simultaneous equations that enables them to identify the specific transmission channels through which democracy affects growth. They find that democracy is conducive to growth by improving the accumulation of human capital and by lowering income inequality. However, democracy hinders growth by reducing the rate of physical capital accumulation and, again less robustly, by increasing government consumption. The difference-in-differences approach is used in a number of studies (e.g., Persson and Tabellini 2006). Persson and Tabellini (2007) combine the differencein-differences approach with matching estimates based on propensity scores. In very simple language, they compare the growth paths of two (or more) countries that are similar in many respects. One of these countries is then treated by some external shock (economic reform) and the two ensuing growth paths are compared with each other. They conclude that previous estimates might have underestimated the growth effects of democracy. They find that transitions from autocracy to democracy are associated with an average growth acceleration of about one percentage point, whereas the transition in the opposite direction slows growth down by almost two percentage points.5 Finally, the meta-analysis by Doucouliagos and Ulubasoglu (2008) concludes that democracy does not have a direct effect on growth but does have significant and robust indirect effects through higher human capital, lower inflation, higher political stability, and higher levels of economic freedom.

2.3.2.2 Ask More Precise Questions There are also a number of papers that have pursued the second option, namely to ask more precise questions by, for example, using dependent variables that are more narrow than economic growth. Mulligan et al. (2004) do this by asking whether democracies differ from nondemocracies in public policies. Looking at the decades from 1960 to 1990, they find that economic and social policies do not significantly differ between these two regimes. They look at various taxes and find that democracies tend to have less redistributive personal income taxes. This is a surprising result as conventional wisdom has it that democracies are prone to redistribution because a majority of the poor can agree to exploit a wealthy minority. However, the more significant differences between the two regimes involve barriers to entering the political system. Explicitly referring to Tullock (1987), they find that autocracies

5

They classify countries as democratic if they have Polity2 values larger than zero. This means that transitions from autocracy to democracy (and vice versa) could take place without any change in a country’s formal constitution.

2.4 Different Autocracies Perform Differently

33

are more likely to torture, execute, regulate religion, censor the press, and have large military budgets.6 The third possible option to extend our knowledge regarding the economic effects of different types of regimes (namely, to ask whether different kinds of autocracies need to be separated to get clear-cut results) will be discussed in the next section.

2.4 Do We Have Good Reasons to Assume That Different Kinds of Autocracies Perform Differently? 2.4.1 Theory The various taxonomies trying to separate different kinds of autocracies distinguish according to who holds power (Geddes 1999; Cheibub et al. 2010), or what motives drive the behavior of the autocrat (Wintrobe 1998). In her take to answer the question asked in this section, Gandhi (2008) proceeds somewhat differently by asking what institutional setup (might) constrain the behavior of an autocrat. More specifically, she asks whether autocracies with legislatures grow faster than autocracies without legislatures and whether autocracies with multi-party systems grow faster than autocracies with single-party systems. She ventures that legislatures enable autocrats to formally co-opt important groups into their regime. By granting them parliamentary representation, these groups are more likely to accept the regime and less likely to arouse opposition against it. Parliament can be considered as a formalized forum for negotiations between the autocrat and the most important groups in a country. In a sense, it is a device that saves transaction costs when negotiating with groups whose cooperation is needed and (or) beneficial for the autocrat. Legislatures are also useful in providing information regarding the preferences of the groups represented in parliament. This is important for dictators because it might be crucial for them to mitigate what Wintrobe (1998) has termed the “dictator’s dilemma.” According to that dilemma, the dictator would like to know how popular/ unpopular he is, but the more repressive he becomes, the less likely people are to

6

Tullock (1987) also argued that the rules of succession used in autocracies have an impact on whether succession will be orderly or not. Different constitutional succession rules imply different incentives for would be successors to depose the current autocrat. More specifically, Tullock argued that non hereditary autocracies will be less stable than hereditary ones. Regarding the evolution of consti tutional succession rules over time, he predicted that they would move towards hereditary ones. These predictions were tested drawing on the Danish monarchy as a case study by Kurrild Klitgaard (2000). He divided the period from 935 until 1849 into nine shorter periods, in which different succession rules applied. The experiences of the Danes with autocratic succession are supportive of Tullock’s theory.

34

Democracy vs. Autocracy

reveal their true attitudes. In addition, debates in parliament are less costly than protests on the street, and concessions made in parliament can be sold as political compromise – and do not signal capitulation. Because closing a parliament entails some political cost, its creation is expected to have some value in improving the commitment capacity of the autocrat. Summarizing her argument, Gandhi (2008: 12) writes: “Legislatures provide information about the preferences of these groups, a safe forum in which the regime can negotiate with them, and an efficient mechanism by which to target spoils and policy concessions.” The enhanced commitment capacity is assumed to induce more investment, and receiving better information about preferences might lead to improved allocative efficiency. She, hence, expects autocrats with formally institutionalized parliaments (and a multi-party system) to grow faster than autocracies without such an institutional setup. In brief, one could say that she expects autocracies that resemble democracies more to grow faster. Wright (2008) is also interested in knowing whether autocratic institutions make a difference, albeit in a slightly different sense. He draws on the distinction between personalist, military, single-party regime, and monarchic regime, and takes the endogeneity of institutions explicitly into account. His main idea is that autocrats are less likely to create legislatures if they believe that they can live without them. Now, the commitment-enhancing function of parliaments is particularly important in countries that depend on (foreign direct) investment for their development. Inversely, autocrats of resource-rich countries depend less on (foreign direct) investment because they can buy the loyalty of the groups whose support they need by letting them participate in the proceeds acquired from exporting natural resources. In addition, he expects autocrats with a longer time horizon to have a higher propensity to rely on parliaments. His prediction is that oil-rich regimes are likely to be personalist ones. If personalist regimes do rely on parliament, it does not serve as a device to make commitments credible, but rather as a device to reward or punish other members of the elite.

2.4.2 Empirical Evidence Both Gandhi (2008) and Wright (2008) try to test their theories. The results that Gandhi finds are largely supportive of her hypotheses. More precisely, she finds that non-institutionalized dictatorships (those without parliaments) grow slower than those with a single-party parliamentary system. Those with a parliament and a multi-party system do best. However, the coefficients of determination are remarkably small. In a sense, then, autocracies grow faster the more they resemble established democracies that rely on a significant number of veto players. Gandhi interprets her findings as evidence against all those who argue that countries in the early stages of development might profit from autocratic regimes. In the argument

2.4 Different Autocracies Perform Differently

35

underlying her empirical assessment, Gandhi assumes that closing a parliament entails some political cost.7 Wright (2008) proceeds in a number of steps to answer the question whether different kinds of autocrats rely on parliament for quite different functions. He first establishes that regime choice is not random, but is significantly associated with a number of traits, such as whether a country has abundant resource reserves. Per capita oil reserves in personalist regimes are twice as high as in single-party regimes and five times as high as in military ones. To take the time horizon of autocrats explicitly into account, Wright (2008) employs the predicted failure probability assuming that the higher that probability, the shorter the relevant time horizon. His most important findings are as follows: (1) Unstable personalist regimes and stable single-party and military regimes are more likely to have legislatures; (2) when income and regime type are interacted, he finds that richer military and singleparty regimes are more likely to have a legislature, whereas richer monarchies are less likely to have one; (3) when oil revenues and regime type are interacted, very similar results are produced. Specifically, interaction effects for military and single-party regimes are significantly negative, and oil revenues are used as a substitute for legislatures. The next steps in his analysis are to ask whether the existence of legislatures is significantly associated with investment levels. He finds that legislatures are associated with lower investment levels in personalist regimes, whereas they are associated with higher investment in military regimes. They have little impact in single-party regimes and monarchies. The final step in his analysis is to inquire into the growth consequences of all this. The findings are in line with what one would expect after having seen the results regarding investment. In developing his argument, Wright (2008) frequently distinguishes between “binding” and “non-binding” legislatures. In the empirical test, however, the distinction seems to be more between regimes having a legislature from those not having one. Further, countries that are important exporters of natural resources are, of course, not constrained to countries exporting oil. But in the empirical test, only oil export is taken into account. Olson (1993) argued that hereditary rulers develop a longer time-horizon because they derive extra utility from seeing their heirs in office, which would make them choose policies that are more likely to increase general social welfare. Largely relying on that argument, Besley and Reynal-Querol (2017) ask whether hereditary rulers achieve higher growth rates than non-hereditary ones. Based on a dataset covering

7

It would be interesting to deal with this assumption a little more explicitly: Does it make a difference whether the autocrat closing a parliament is also the one who created it, or if he inherited it from one of his predecessors? Analyzing the consequences of closing down parliament, or of attempts to do so, could shed more light on the corresponding costs and would, hence, be very welcome.

36

Democracy vs. Autocracy

the period from 1848 until 2004 and up to 197 countries, they find that hereditary rulers are associated with roughly one additional percentage point in growth every year. This effect is, hence, large and robust to a number of checks. Interestingly, this effect only materializes if the ruler is not subject to strong executive constraints. In a sense, then, hereditary rule can be seen as a substitute for strong executive constraints.

2.5 What Are the (Economic) Effects of Extending the Franchise? A very particular way to measure the degree to which a country is democratic is simply to measure what percentage of the adult (male) population has the right to vote. In the nineteenth century, the right to vote was generally restricted to the male part of the population. Only in the twentieth century was the franchise extended to females. By some (e.g., Aidt et al. 2006), the extension of the franchise during the nineteenth century is also referred to as “economic franchise” because a criterion used for many extensions was the economic position of the prospective voters. During the nineteenth century, the franchise was (often gradually) extended to the poorer echelons of society. The basic conjecture regarding the economic effects of the economic extension of the franchise is straightforward: if more less wealthy people are granted the franchise, this implies that the median voter will be poorer than the mean voter, which is accompanied by a higher demand for redistributive policies (Meltzer and Richard 1981). The subsequent extension of the franchise to women could cause more demand for public goods such as health and education, but also social security. Most of the studies inquiring into the economic effects of franchise extensions are based on either Europe or the USA. Aidt et al. (2006) focused on 12 European countries and the time between 1830 and 1938. They find that the economic franchise contributed to growth in government spending by increasing spending on roads, transportation and communication, as well as on domestic security. The administration and the extension of the female franchise had an impact on government expenditure through an increase in spending on health, education, housing, redistribution, and social insurance “but the effects are statistically weak.”

2.6 What Factors Make a Country Switch from One Regime to Another? 2.6.1 Theory One core issue of Constitutional Political Economy (CPE) is to explain constitutional choices. In our context, the relevant question, thus, is under what circumstances will democracy be chosen and when autocracy is more likely. Therefore, it is relevant to

2.6 What Factors Make a Country Switch from One Regime to Another?

37

explain transitions from one kind of regime to the other. Over the last couple of years, the theory by Acemoglu and Robinson (2001 and 2006) has managed to establish itself as the leading theory. But, of course, these two authors were not the first to inquire systematically into the reasons underlying the choice for different regime types. Before we get to a brief summary of their theory, we describe some answers given before they published their theory. Sutter (1995) assumes that a transition from authoritarian to non-authoritarian government has the potential to improve the general welfare. Because revolutions imply a waste of resources, other transition paths, such as negotiated pacts, are, hence, superior. Sutter then turns to the problems of implementing negotiated pacts. What he calls the “punishment dilemma” is caused by the inability of some actors to credibly commit themselves. Specifically, new rulers have problems to credibly promise not to sanction the exiting rulers in the future. An interesting insight of Sutter’s work is that the optimal policy toward dictators is time-inconsistent. To prevent would-be dictators from becoming dictators, one wants to threaten them with the worst possible punishment. As soon as they are dictators, however, and one is interested in getting rid of them, one wants to punish them only mildly in order to give them an incentive to step down and make room for a non-authoritarian government.8 Barzel (1997, 2000) discusses the role of parliament as a device autocrats can use to credibly bind themselves. He argues that secure kings deliberately give up some powers enabling them to credibly commit themselves to promises not to confiscate their subjects’ property. These credible commitments, in turn, make it possible to realize additional gains from cooperation. His approach, thus, flatly contradicts the more conventional one that conceptualizes the emergence of parliament as the consequence of weak autocrats. This approach ties in very nicely with the commitment function attributed to autocratic constitutions discussed in Section 2.2. Factors contributing to regime change have been analyzed by political scientists for decades. This is not the place to summarize these contributions. The paper by Haggard and Kaufman (1997) is interesting in this context because it is a survey of the most important contributions by political scientists to the topic since 1970. In their own take on the topic, they propose to distinguish between crisis and non-crisis transitions and show that transitions that occur under crises are often more complete. Their argument for this finding is that governments in crisis situations enjoy less support. However, theirs is a “small n”-study, with 10 observations only.9 In his contribution to the topic, Voigt (1999) proposes to interpret constitutional change as the outcome of a bargaining game. Different from the social contract view, 8

9

This insight has immediate implications for the International Criminal Court, which Sutter deals with in a later paper (2006). Haggard and Kaufman (1995) is a book length treatment of the same topic.

38

Democracy vs. Autocracy

where (imagined) deliberation takes place between all members of society, here constitutional change is interpreted as the outcome of a bargaining game that is played by a finite number of actors who represent all groups of society who managed to overcome the dilemma of collective action and who have some positive veto power at their disposal. Veto power is thought of as the power of a group to have a negative impact on the size of the overall number of goods produced by society. It does, hence, not depend on capital ownership, as labor unions could command veto power by calling a general strike. The basic assumption is that constitutions have distributive consequences. It follows that different groups prefer different rule sets, namely those sets that serve their own interests best. The relative power of groups might change over time, e.g. due to technological change. If the constitution does not reflect the relative power of the various groups anymore, those who have gained in power will demand constitutional change in their favor. Daron Acemoglu and James Robinson (2006) proffer a similar explanation of regime change. In their workhorse model, they distinguish between two groups: the powerful elites and the less powerful non-elites. The non-elites demand redistribution of wealth favoring them, e.g., in the form of social and welfare services. If their threat to stage a revolution appears credible, the elites have an incentive to start negotiations and make at least some political concessions to the non-elites. The credibility of threats can change over time, for example, because the non-elites overcome the dilemma of collective action, because of technological progress that makes their input more valuable, and the like. Acemoglu and Robinson assume that negotiations take place and the two groups agree on a modified constitution that entails a more favorable distribution of public goods for the nonelites. At this point, the elites face the problem of being unable to credibly commit to their own promises. If the non-elites are demonstrating in front of the palace, the elites have an incentive to promise improvement. But why would they keep their promise once the heat of the moment has passed and demonstrators are back at home? According to Acemoglu and Robinson (2006), democracy can serve as a tool that helps the elites turn their promises into credible commitments. If most voters are poor, they are likely to vote in favor of (re-)distribution, for example, in the form of social and welfare policies. Extending the right to vote to the poor can thus serve as a commitment device of the elites. Extending the right to vote to ever more groups of society is equivalent to an extension of democracy. The authors, thus, view the transition from autocracy to democracy as the elites’ response to threats from the non-elites. In an extended version of their model, a third group, the middle class, is introduced. If the middle class is sufficiently large, more gradual constitutional change becomes possible and more likely. A competing model explaining the extension of the franchise has been introduced by Lizzeri and Persico (2004). According to

2.6 What Factors Make a Country Switch from One Regime to Another?

39

them, it is not primarily driven by the threat of the non-elites, but rather by conflicts within the elites; more specifically, if a majority of the elites are convinced that too much money is spent on pork barrel goods and not enough on genuine public goods. John Wallis, Barry Weingast, and Douglass North (hereafter, NWW) (2009) have published no less than a “conceptual framework for interpreting recorded human history” – the subtitle of their book. They start from the observation that violence has been a major problem for most societies for the entire history of mankind. In other words, more often than not, there has not been a state, at least not one that meets Max Weber’s definition, i.e., an organization commanding a monopoly on the legitimate use of violence. NWW propose that human history has known only three types of social order: (1) foraging, (2) limited access, and (3) open access. Under this framework, explaining constitutional change implies explaining the transition between these three orders. I here focus only on the transition from (2) to (3). Limited-access orders are based on personal relationships, and the ability to form organizations is limited. In limited-access orders, violence is controlled by relying on a “dominant coalition.” Members of the dominant coalition share the rents that emerge from the absence of violence. In open-access orders, cooperation between individuals is based on impersonal laws that treat all individuals alike. Positions in both politics and the economy are open to everyone and are allocated according to merit. All individuals have the right to found political or business (e.g., corporations) organizations. In open-access orders, all violence is controlled by the state, and the military is under civilian control. The authors group the overwhelming majority of modern-day countries into their second category. Roughly speaking, only the (more established) OECD countries are believed to be open-access orders. NWW’s take on the transition from limited- to open-access orders emphasizes that doing so must be in the interest of the overwhelming majority of the players who make up the dominant coalition. When the elites in a limited-access order society are able to deal with each other impersonally, i.e. based on general rules, then that society will be on the threshold of becoming an open-access order. NWW identify three threshold conditions: (1) the rule of law for elites, (2) perpetually existent organizations in both public and private spheres, and (3) consolidated political control of the military. If the vast majority of the members of the dominant elite coalition realize that general rules, rather than personal relationships, will improve their interactions and make themselves better off, they will be in favor of what NWW call “rule of law for elites.” This sounds like a contradictio in adiecto, but what they really mean by the term is that no member of the elite enjoys particular privileges, and all of them are treated according to the same rules. However, these rules for the elite are not necessarily, indeed not probably, the same rules by which everyone else has to live, hence the term “rule of law for elites.” The main difference between NWW and Acemoglu and Robinson is that the latter basically rely on two actors (the elite vs. the non-elite), whereas the former model the

40

Democracy vs. Autocracy

Figure 2.1 Autocratic transitions and democratic transitions Source: Geddes et al. 2014

elite as a coalition of a multitude of actors. In their explanation of transitions, NWW focus on the bargaining that takes place within the dominant coalition, hence explicitly giving up the monolithic actor assumption relied on by Acemoglu and Robinson. The theories presented above emphasize different aspects of regime transformation. We have not presented a comprehensive theory of regime transformation because there is none. This evaluation is supported by the fact that hardly anybody predicted the demise of the socialist economies in Central and Eastern Europe (Voigt 2019 surveys the literature belonging to the New Institutional Economics that has dealt with the topic). The uprisings in the Middle East and Northern Africa that occurred in the early 2010s were also predicted by almost no one. In the following section, we discuss a particular approach to explaining institutional change that focuses on the diffusion of institutional change across space.

2.6.2 Empirical Evidence Before describing the available empirical evidence, here are some stylized facts regarding regime transitions. Figure 2.1 shows that beginning in the 1950s and throughout the 1980s, transitions from one autocratic regime to another autocratic regime always outnumbered transitions to democracy. The 1990s was the decade during which this trend changed dramatically. Many of the former autocracies in Central and Eastern Europe were transformed into democracies. How did the different autocratic regimes do over time? Figure 2.2 shows that autocracies based on single parties experienced a peak early in the 1980s but have

2.6 What Factors Make a Country Switch from One Regime to Another?

41

Figure 2.2 Autocratic regimes across time Source: Geddes et al. 2014

been in decline ever since. Nevertheless, in 2010, they were still the most frequently encountered type of autocracy (both figures are taken from Geddes et al. 2014 and are, hence, based on their delineation of different types of autocracies). How well do the various theories briefly presented in the preceding section perform in explaining these transitions? As described in the preceding section, Acemoglu and Robinson argue that democratization can be interpreted as the outcome of a credible threat of revolution. In a number of papers, Toke Aidt and his various co-authors have tried to explain the extension of the franchise as a rational response of autocrats to revolutionary threats that were perceived as credible by them. A study by Aidt and Jensen (2014), for example, is based on the idea that revolutionary events taking place anywhere in Europe can reveal information to the rulers all over Europe regarding the degree to which they are threatened by a revolution themselves. They find that one extra revolutionary event is associated with a two percentage point expansion of the (male, adult) franchise. One extra revolutionary event increases the chances of suffrage reform by 75%. It seems, hence, that revolutionary threats perceived as real do cause autocrats to extend the franchise. NWW think of their approach as a “conceptual framework for interpreting history.” It is, hence, not a falsifiable theory in the conventional sense of the term. Asking whether the framework can usefully be applied to various countries is, hence, more useful than asking whether the theory has been put to rigorous empirical tests. And indeed, there is an edited volume (North et al. 2013) in which the framework is applied to nine different countries. The journal Constitutional Political Economy published a special issue applying the framework to the German example.

42

Democracy vs. Autocracy

The most comprehensive empirical test yet is provided by Gassebner et al. (2013). They are actually providing answers to two questions, namely: (1) what determines the emergence of democracy, and (2) what determines its survival? Gassebner et al. run what is called an Extreme Bounds Analysis (EBA). Many times, individual explanatory variables seem to be significant as long as they are not combined with other variables that would also be significant. EBA focuses on one independent variable at a time, but combines them with many other variables in literally thousands of permutations. Scholars implementing EBA frequently run millions of regressions. In this paper, Gassebner et al. (2013) test the significance of 59 factors proposed in the literature. They range from “Africa” and “arable land” to “urban population” and “world democracy” (depicting the global share of democracies) in alphabetical order. It turns out that most of them do not survive. According to the authors, the most robust determinants for transitions to democracy are income growth (with a negative effect) and past transitions (with a positive one). High per capita income is conducive to the survival of democracy, whereas a high number of past transitions display a negative effect.

2.7 Summary and Outlook This chapter served to highlight the possible differences between democracies and autocracies. Among other things, constitutions are often thought of as manifestos that constrain governments. Since it is often assumed that autocrats strive not to be constrained, it is straightforward to ask what role there is for constitutions in autocracies in the first place. As we saw in this chapter, studying the “economics of autocracy” has long been neglected. This has, however, changed over the last 15 years or so. Indeed, the economics of autocracy has become a hot topic. But we also saw in this chapter that many questions in the area still wait to be answered. This not only refers to the degree to which autocrats can be bound by constitutional rules, but also to the effects of constitutions in terms of economic performance. There is some evidence that democracies perform better than autocracies, but little is known about the robustness of these results, the exact transmission channels through which the effects, if any, are mediated, and so on. In the next chapter, we proceed to inquire into the (economic) effects of single constitutional institutions in more detail, such as electoral rules or form of government. We, thus, assume that we are dealing with democracies.

3

Positive Constitutional Economics

3.1 Introductory Remarks In this chapter, the focus will be on the analysis of single constitutional traits, such as the form of government, the structure of the state (unitary or federal), whether direct democracy institutions can be used, and so on. In summarizing research on these traits, I will first deal with their effects, economic and otherwise, and only then ask why different societies have chosen different traits. Why are the USA and Canada federations, whereas the United Kingdom and France are unitary states? In academic lingo, we first deal with constitutional traits as explanans and only then turn to endogenize them, i.e., consider the traits as explanandum. Within this structure, theoretical conjectures are always presented prior to empirical results. As we discuss various constitutional traits, we will follow the structure or design of many recent constitutions. Along those lines, Section 3.2 deals with the rights that the state promises to secure vis-à-vis its citizens. These are not confined to basic human rights (such as the absence of torture), but also comprise economic rights (private property, for instance), civil and political rights (the right to vote, to demonstrate, to form associations, and so on), and the so-called social or emancipatory rights (non-discrimination of women, for example). Many constitutions then move on to deal with the structure of the state organs, their relationship with each other, and so on (the German word Staatsorganisationsrecht is very telling here). We follow this practice here and deal with the horizontal separation of powers in Section 3.3. The horizontal separation of powers deals with the relationship of the three branches of government, i.e., the relationship between the legislature, the executive, and the judiciary. The traits explicitly dealt with here are the form of government (presidential vs. parliamentary), bicameralism, and the judiciary. The vertical separation of powers, in turn, deals with the relationship of the various levels (top or federal level, states or governorates, local communities) of the state with each other. We deal with the vertical separation under the heading of federalism in Section 3.4. In recent years, traditional government functions have ever more frequently been delegated to independent and specialized domestic or international agencies. For instance, monetary policy has often been delegated to independent central banks aiming at price stability, to name only the most prominent example. Additionally, in order to better monitor government, traditional organizations, such as supreme audit

44

Positive Constitutional Economics

organizations, have often been strengthened. These developments may be called “the new separation of powers.” A brief overview of research analyzing these developments is presented in Section 3.5. Section 3.6 focuses on direct democracy. The last trait dealt with in this chapter is electoral rules. Although these were put center stage by Persson and Tabellini, we move them somewhat into the background because most state constitutions do not contain explicit provisions regarding the electoral system to be used. Section 3.8 contains a summary. Appendix 2 is a non-exhaustive list of the empirical papers cited in this chapter.

3.2 Basic Rights 3.2.1 Defining Rights Many constitutions contain an elaborate catalogue of basic rights that are often separated into positive and negative rights. Negative rights can further be delineated into rights establishing freedom from state or third-party interference (such as torture and imprisonment without trial) and freedom to do something (assemble with others, criticize the government in public, etc.). Examples of positive rights are rights to food, housing, paid jobs, and so forth. Negative rights create a protected domain that not even the state is allowed to trespass. For democratically organized states, this implies that negative rights limit the possible scope of majority decision-making. If rights are really basic, even large majorities are bound by them and cannot simply ignore them. Negative rights can thus protect minorities against current majorities and can be thought of as “veto rights,” as they give their holders the right to behave in a certain way, even if a huge majority would like the right holders to behave differently. Blume and Voigt (2007) propose to distinguish four groups of basic rights: (1) basic human rights, which is sometimes also referred to as “integrity rights” and includes the absence of torture, the absence of political killings, the absence of people who disappear, etc.; (2) economic rights, which include primarily private property rights broadly defined; (3) civil and political rights, which include the absence of restrictions on free movement, the possibility of participating in political life, freedom from government censorship, and so forth, although it is not uncommon for political rights to be explicitly separated from civil rights; and (4) social or emancipatory rights, which endow the individual with positive rights vis-à-vis the state.1 1

Often, these rights are also called “economic, social, and cultural rights” as in the International Covenant. We prefer to refer to them as social or emancipatory rights here, as economic rights will be used in a different context and will indicate primarily the security of private property rights. Blume

3.2 Basic Rights

45

3.2.2 Effects of Basic Rights 3.2.2.1 Theory When it comes to basic rights, the congruence between formally granted rights and their actual enforcement is especially crucial. It is plausible to assume that the degree of factual enforcement is a function of other constitutional characteristics. For example, a greater separation of powers (be it horizontal, vertical, new, or a combination thereof ) would appear conducive to a high degree of factual enforcement. Economists have traditionally argued that secure property rights (i.e., economic rights) lead to higher income levels and that social and emancipatory rights have just the opposite effect. It has been pointed out that the two types of rights are in competition as any – factually enforced – promise of the state as to positive social rights must be paid for by someone (see, e.g., Hayek 1976). The potential of civil rights to positively constrain government is often discussed. A free press, for example, provides a safe way of criticizing government behavior (e.g., Sen 1999). On the other hand, many economists are skeptical that democracy (the most important aspect of political rights) is conducive to income and growth (e.g., Barro 2000). Further complicating the issue is that the economic effects of basic human rights have rarely been systematically analyzed. Blume and Voigt (2007) argue that talk of various kinds of human rights makes little sense in the absence of basic rights. For example, discussing the effects of secure economic rights is inane if persons can be imprisoned without a fair trial, tortured into pleading guilty, and further so abused that the existence of “rights” is a hypocrisy of the first level. Farber (2002) argues that a government’s respect for basic human rights is a good indicator of how seriously that same government will take its promises as to other rights, such as property rights. If a government does not refrain from discrimination and physical harm vis-à-vis its citizens, why should it respect private property rights of foreign investors? If government respect for basic human rights is indeed a signal of its credibility in general, the abuse of basic human rights should have consequences for: (1) the country’s creditworthiness, as well as (2) the amount of foreign direct investment it will attract, leading to (3) a generally lower propensity to invest in an abusive country than in one where the government respects basic human rights.

3.2.2.2 Empirics There is a sizable literature inquiring into whether economic rights have a significant effect on economic growth, Knack and Keefer (1995) being a pathbreaker on this and Voigt find their conceptual distinction well supported by data. Running a principal component analysis with a large number of variables shows that these groups are not only conceptually distinct, but also empirically.

46

Positive Constitutional Economics

topic. Their paper, for the first time, uses data from the International Country Risk Guide and Business Environment Risk Intelligence as proxies for the security of property rights and finds that the security of these rights is significant for explaining both investment and growth. A paper by Voigt and Gutmann (2013) inquires more directly into constitutionally protected property rights. Coding these rights and plugging them into a standard growth regression, the authors find only modest effects. They argue that the modest effects could be due to the less-than-complete implementation of these rights. To proxy for their actual implementation, they use a measure of actual judicial independence (JI) based on the conjecture that the highest courts that are independent from government are more likely to ensure the actual implementation of property rights. As soon as property rights are conditioned upon the de facto score of JI, results become much stronger. Knack and Keefer (1995) published their paper around the same time the Economic Freedom Index (Gwartney et al. 1996) and the Index of Economic Freedom first appeared; the latter is produced jointly by the Heritage Foundation and the Wall Street Journal. These data sets gave rise to an entire cottage industry focusing on the interrelations between various aspects of economic freedom and prosperity. This literature is extremely interesting but only marginally relevant for this primer because the explanatory variables found therein often have only a very indirect basis in the constitution.2 Sometimes, the variables are not even proxying for institutions, but simply for policies. Moreover, they are often subjective indicators. Thus, the impact on economic growth of rights that are not only guaranteed by the

2

It would, however, be very interesting to see whether specific constitutional rules are more (or less) conducive to economic freedom. The work by Spindler and de Vanssay (2002) is a first attempt in this direction. The authors find that a higher number of veto players are correlated with higher levels of economic freedom. The homepage of the Fraser Institute, which administers the Economic Freedom Index (EFI), contains a section that lists all the academic papers that rely on the data of the Index (see www .freetheworld.com/papers.html). Scully (2001) is an early survey that discusses some econometric issues, but relies primarily on indicators antedating the EFI. Berggren (2003) is a very accessible survey. De Haan and Sturm (2000) find that changes toward greater economic freedom do indeed foster economic growth, but that the level of freedom is not related to growth. A number of empirical studies dealing with the relationship between economic freedom and growth have analyzed the impact of the single components that make up the EFI based on the assumption that different variables might have different impacts on growth rates. Carlsson and Lundström (2002), e.g., test the sensitivity of the components. They find that four components are positively related to growth: (1) the use of markets, (2) the freedom to use alternative currencies, (3) the legal structure and security of private ownership, and (4) the freedom to exchange in capital markets. Of these, only (3) unequivocally survives the robustness test, but (2) almost does. Monetary policy and price stability are insignificant, whereas more freedom to trade with foreigners actually decreases growth rates. This

3.2 Basic Rights

47

constitution, but also actually enforced, is yet to be conclusively determined.3 In an early paper, de Vanssay and Spindler (1994) conclude that the entrenchment of any single right in the constitution seldom has any significant economic effect on growth. Two follow-up questions suggest themselves. (1) What about the effects of constitutional entrenchment on other variables, such as government spending or corruption? After all, a long list of positive rights means that resources will be needed for their implementation. This should be reflected in the structure of the budget and, most likely, also in the overall level of government spending. (2) What factors determine whether constitutionally guaranteed rights will be factually enforced?

3.2.3 Determinants of Basic Rights 3.2.3.1 Theory Common sense has it that the number of constitutionally guaranteed rights has substantially increased over time. Assuming that this is correct, and that governments would prefer to be less, rather than more constrained, one wonders about possible reasons behind this trend. Different conjectures regarding possible reasons have been promoted. A very basic hypothesis, in line with social contract theory, is that basic rights can be thought of as “guarantees” or “insurance” that people who expect to be in the

3

finding is not only significant but also robust. Further, lower levels of government consumption and transfers decrease growth rates below a certain threshold. In other words, increasing government consumption can increase growth up to a certain level. Past that point, however, growth rates will decrease. The result that increased freedom of trade is detrimental to growth is counterintuitive and has provoked further analysis. Berggren and Jordahl (2005) decompose this subindicator into its single components and find that the result is driven by a single variable, namely, “taxes on international trade.” Analyzing the sensitivity of this result with regard to both model specification and sample composition, they find that it is not robust. In fact, they find that when outliers are ignored, the entire subindicator exerts a positive influence on growth. More recently, Justesen (2008) used Granger causality analysis to discuss whether economic freedom causes growth. Apart from the composite EFI, only two of its components government size and regulatory policies have a robust effect on both economic growth and investment. There is no consensus as to the effects of political and civil rights on economic outcomes. Barro (1997, 2000) finds that the degree of democracy has little significance for explaining the variation in growth rates. Blume and Voigt (2007) find that social or emancipatory rights are not conducive to investment in physical capital but are correlated with higher levels of productivity and basic human rights, which, in turn, have a positive effect on investment, but do not seem to contribute to productivity. None of the four groups of rights ever has a significant negative effect on any of the economic variables included here. Ben Bassat and Dahan (2008) note that they did not find a robust effect of the degree of social rights as postulated in a country’s constitution on public policy.

48

Positive Constitutional Economics

minority in the future demand to be included into the constitution, lest they not consent to a constitutional draft. Inclusion of rights into constitutions serves as insurance because constitutions are difficult to change and they thus insure minorities against majorities. This argument has been put forward, for example, by Ginsburg (2003). In Chapter 2, we saw already that Acemoglu and Robinson (2006) interpret the extension of the franchise, and democratization more generally speaking, as a way for the elite to appease the non-elite. Their general theory can, of course, also be applied to rights. Promising any kind of basic right can serve to appease the nonelite. The constitutionalization of more and more rights can also be interpreted as the consequence of certain ideologies. Governments whose members put strong emphasis on socialist values are also more likely to promise broad social rights, such as a right to work. More recently, the spatial diffusion of rights has come into focus. Goderis and Versteeg (2014) propose to distinguish four different channels through which constitutional rights could spread across countries: (1) coercion, (2) competition, (3) learning, and (4) acculturation. Coercion could be used by powerful states, but also by aid donors. Competition suggests that foreign constitutions are imitated in an attempt to attract foreign direct investment. Learning is thought of as the consequence of learning processes expected to take place between constitution-makers in countries that share many similarities (such as having similar legal systems), whereas acculturation is thought of as getting used to how things are done in other countries without the constitution-makers necessarily being convinced that this is really the best way. Channels (1) and (4) in particular might lead to a “de jure/de facto gap”: If constitution-makers believe additional funding by potential donors can best be secured by adding a couple of formal rights to their constitution, the likelihood of these rights being implemented does not appear to be particularly high. All approaches trying to identify the drivers behind the constitutionalization-ofrights trend mentioned so far are driven by “human” factors. But weather events might also play an important role. In countless instances during the history of humankind, people have looked for someone to blame for a bad harvest. The witch hunts that took place in many parts of Europe during the sixteenth century can be understood thus. Many pogroms against Jews are another example.

3.3 Horizontal Separation of Powers 3.3.1 Introductory Remarks The term “separation of powers” is inseparably linked to Montesquieu’s De l’esprit des loix (1748), in which he describes a functional separation between the legislature,

3.3 Horizontal Separation of Powers

49

the executive, and the judiciary.4 Although Montesquieu’s account is based on a misinterpretation of the British system, it formed the basis for constitutional thinking in what was to become the United States of America (see Hamilton et al. 1788/1961). The term “separation of powers” has been used with different connotations. Governments structured such that different branches can decide without the consent of the other branches in specific areas (a rigid separation of powers) are very different from governments structured such that the various branches monitor each other (a model of checks and balances).5 This latter type has been interpreted as a constraint on politicians for the benefit of the citizens. The horizontal separation of powers is a very broad area, which is why we deal with it in small portions. We start with the most general consequences of having two or three branches of government, instead of just one (Section 3.3.2). We then inquire into different forms of government, more specifically into the question of whether there are systematic differences between presidential and parliamentary systems. This is an aspect of the separation of powers because, at least de jure, presidential systems have a greater degree of separation of powers than do parliamentary ones (Section 3.3.3). The following subsection takes a closer look at the various ways of organizing legislatures. There, we inquire into consequences of bicameral systems (Section 3.3.4). The economic consequences of the judiciary – and the reasons behind different ways of organizing the judiciary – were almost entirely neglected by economists for a long time, but this has changed; and Section 3.3.5 provides an overview of recent insights concerning this topic.

3.3.2 Branches of Government Persson et al. (1997), e.g., argue that politicians have two possibilities for enriching themselves to the detriment of citizens, the first being based on the misuse of power, and the second on exploiting information advantages. The authors show that both possibilities can be reduced by implementing checks and balances between the legislature and the executive. This was one of the first models dealing with the (horizontal) separation of powers. However, it falls short of the traditional notion going back to Montesquieu, as it is based on two government branches only and does not take a third branch (the judiciary) into consideration.6

4

5

6

However, the idea of separation of government power finds its roots in the writings of Aristotle (1932: 1297b 1298a). In the modern context, it was in seventeenth century England that the doctrine emerged for the first time as a coherent theory of the state. Brennan and Hamlin propose calling this second version “division of power,” meaning that one particular power can be exercised only if various agents cooperate (2000: Chapters 11 and 12). Padovano et al. (2003) add the judiciary to the Persson et al. model. Their paper is discussed in Section 3.3.5.1.

50

Positive Constitutional Economics

3.3.3 Form of Government: Presidential vs. Parliamentary Regimes It is often argued that the degree of separation of powers is greater in presidential than in parliamentary systems, because the head of the executive (the president) does not depend on the confidence of the legislature (parliament) to survive. Persson et al. (1997, 2000) argue that it is easier for legislatures to collude with the executive in parliamentary systems, which is why they expect more corruption and higher taxes in those systems than in presidential ones. They further argue that the majority (of both voters and legislators) in parliamentary systems can pass spending programs whose benefits are clearly targeted at themselves, implying that they are able to make themselves better off to the detriment of the minority. This is why Persson et al. (2000) predict that both taxes and government expenditures will be higher in parliamentary than in presidential systems. This view has been challenged on both theoretical and empirical grounds. Robinson and Torvik (2016) point out that Persson et al. (1997) were inspired by the US system and that presidential systems in both Latin America and Africa are different from the US system. Presidents in other countries often enjoy unchecked powers vastly exceeding those of the US president. Lutz (2006: chapter 4) argues that the distinction between parliamentary and presidential systems is a poor proxy for the degree of separation of powers. He proposes an indicator that takes explicitly into account a number of additional aspects, such as federalism and bicameralism. Fish (2006) constructs an index of legislative powers and finds that more legislative power is highly correlated with higher levels of democracy.

3.3.3.1 Effects of Forms of Government To test their hypotheses, Persson and Tabellini (2003; PT for short) needed to code presidential as opposed to parliamentary systems. If a system did not have a vote of no confidence, they coded the country as presidential.7 They derive the following results. (1) Government spending is 6% of GDP lower in presidential compared with parliamentary systems.8 (2) The size of the welfare state is 2–3% lower in presidential systems.

7

8

This implies that Switzerland is coded as “presidential” although it does not have a directly elected president, whereas France is coded as “nonpresidential” because for his or her survival, the French prime minister depends on a confidence vote of parliament. If a country with a proportional and parliamentary system is compared to a country with a plurality and presidential system, government expenditures of the latter are predicted to be 10% of GDP lower than the former, ceteris paribus!

3.3 Horizontal Separation of Powers

51

(3) The influence of the form of government on the budget deficit is rather marginal; the binary variable explains only a small proportion of the variation in budget deficits. (4) Presidential systems seem to have lower levels of corruption. (5) There are no significant differences in the level of government efficiency between the two forms of government. (6) Presidential systems appear to be a hindrance to increased productivity, but this result is significant at the 10% level only. These results are impressive and intriguing. Although government spending is less in presidential systems and they suffer less from corruption than parliamentary systems, parliamentary systems have an advantage over presidential systems in terms of productivity, if only at a low level of significance. This result is reinforced by Persson (2005), who asks whether changes in the form of government matter for the likelihood that governments will implement structural policies conducive to economic growth. He finds that reforms of parliamentary constitutional arrangements have an important growth-promoting effect. Introducing parliamentary democracy in a nondemocracy or in a presidential democracy improves structural policy such that the long-run productivity growth increases a stunning 50%.9 In a study replicating and extending the PT estimates, Blume et al. (2009b) pour some water into PT’s wine. It turns out that PT’s results are not robust, even to minor modifications. Increasing the number of observations from 80 to 92 makes the presidential dummy insignificant in explaining variation in central government expenditure. This is also the case as soon as a slightly different delineation of presidential systems is used. If the dependent variable is changed to total (instead of central) government expenditure, the dummy also becomes insignificant. If “graft” is measured using a longer and more recent period (or the number of observations is increased, or a slightly different definition of presidentialism used, or a combination thereof ), the dummy becomes insignificant. For total factor productivity, PT did not find the presidentialism dummy significant. Interestingly, if a slightly different delineation of presidentialism or estimates of total factor productivity for 2000 (instead of 1988) are used, presidentialism becomes significantly negatively correlated with total factor productivity. PT are, of course, well aware that the form of government is endogenous. Forms of government are not randomly spread across countries, but are the consequence of 9

In an offshoot, PT (2002) ask whether the form of government has any consequences on electoral cycles, i.e., one possible aspect of fiscal policy over time. They find that pre election tax cuts are a universal phenomenon, whereas postelection adjustments, such as spending cuts and tax hikes, occur only in presidential systems. In the long run, parliamentary systems should, hence, suffer from higher levels of debt.

52

Positive Constitutional Economics

deliberate choices made by constitutional assemblies. To mitigate the problem that explanatory variables are not truly exogenous, economists try to find variables that “explain” the explanatory variables. These are called “instruments” or “instrumental variables” and they work well if: (1) they cannot be easily manipulated by humans (such as geographic location of a country, its climate and so forth), and (2) their effect materializes exclusively through the instrumented variable (economists call this the “exclusion restriction”). Finding convincing instruments is extremely difficult. PT use a number of instruments to deal with the problems. Choice of government could be subject to fads and fashion, which is why they instrument it with the age of democracy. They note that the form of government is highly correlated with geographical variables. For example, presidential systems are more likely to be found in Latin America and closer to the equator, which is why they also use the distance of a country from the equator. Acemoglu et al. (2001) argue that settlers brought (good) European institutions to colonies where their mortality was low. Settler mortality rates could thus serve as an instrument predicting the quality of institutions today.10 This is one way to justify the use of the proportion of the current population speaking English or another European language as yet another instrument. The instruments used by PT have been criticized by various scholars, such as Acemoglu (2005) and others. Rockey (2012) also takes issue with these instruments, and in particular the age of democracy variable, arguing that in many cases, it is seriously misspecified. That is why he proposes a new way to measure democracy and its age taking into account four criteria, namely, the date of the first election, the promulgation of the first democratic constitution, the presence of universal male suffrage as a minimum requirement, and that there was no subsequent return to autocracy. In addition to this modified age of democracy variable, Rockey (2012) also uses the square of the regime’s age to control for possible nonlinearities, whether the country ever was a colony and, if so, if it was a British or a Spanish colony. Finally, the percentage of Catholics among the population is used, based on the assumption that Spain (and Portugal) were more interested in converting indigenous people to Christianity than the British (and other) colonizers. Since conversions usually took place before the countries became democratic, Rockey believes the percentage of Catholics to be a good instrument. It turns out that the results are now stronger than the original PT results. 10

Their measure has been heavily criticized. Albouy (2008) argues that 36 out of 64 countries are assigned mortality rates from other countries; these would frequently be based on mistaken or conflicting evidence. Also, Albouy accuses the authors of choosing data that would favor their hypothesis. Acemoglu et al. (2006) have repeatedly replied to Albouy’s reproaches. They state that their results are robust when African data, those most heavily criticized by Albouy, are not included in the sample or when a maximum cap is introduced for African mortality rates. Acemoglu et al. (2005) confirm most of their previous mortality rates drawing on different sources.

3.3 Horizontal Separation of Powers

53

Moving on to the association between form of government and the quality of governance, it has been found that parliamentary systems suffer from significantly less corruption than do presidential ones (Gerring and Thacker 2004). They argue that “effective accountability arises from a highly structured relationship between voters and political parties and from the relatively clear lines of authority instituted by a centralized political apparatus” (Gerring and Thacker 2004: 314). Lederman et al. (2005) also find that parliamentary systems suffer less from corruption than do presidential ones, and also draw on the concept of accountability to explain why. Their argument is that parliamentary systems “allow for a stronger and more immediate monitoring of the executive by the legislature . . ..” They conclude that after “political institutions are accounted for, variables usually found to be important determinants of corruption . . . lose virtually all their relevance.” Treisman (2007) replicates these results but finds that presidentialism becomes insignificant as soon as one controls for Catholicism, or when a dummy for South America is included. The differences in these results clarify an important point: to date, many of the effects supposedly induced by constitutional rules are not very robust but crucially hinge upon the exact specification of the variables, the time period considered, the control variables included, and so on. What this suggests is that further research needs to be as specific as possible in trying to identify possible transmission channels, and needs to take into consideration the possibility that small differences in institutional details can have far-reaching effects. Possible extensions include the use of more fine-grained variables representing the constitutional institutions. The distinction between parliamentary and presidential systems may simply be too coarse. Very different systems are coded as being presidential, and it appears desirable to explicitly recognize more details, such as whether the president is elected directly by popular vote or via an electoral college, whether there are term limits, the power enjoyed by the president, and so forth. These considerations lead to a fundamental inquiry into the theoretical stance regarding the distinction between presidential and parliamentary systems. Cheibub et al. find a large degree of heterogeneity across the characteristics usually attributed to the forms of government and conclude (2013: 3): “Indeed, knowing whether a constitution is parliamentary, presidential or semi-presidential is less helpful in predicting a constitution’s executive-legislative structure . . . than is knowing the geographic region in which the constitution was produced or when it was written.” One of the first papers that takes the heterogeneity of both presidential and parliamentary systems explicitly into account is contributed by Ardanaz and Scartascini (2014). They ask whether the result that presidential governments are strongly associated with smaller government size is robust to the inclusion of the rules according to which a country’s budget is generated. It is frequently argued that more hierarchical budget institutions reduce the common pool problem because no single

54

Positive Constitutional Economics

actor internalizes the full cost of government spending. This might, in turn, lead to smaller government size. Now, Ardanaz and Scartascini (2014) point out that hierarchical budget institutions might also reduce the separation of powers, which might have the exact opposite effect. And indeed, they find that the presidential form of government is only associated with smaller government size if executive discretion regarding the budget is low. This is another example how important institutional details can be. The economic literature on the horizontal separation of powers is still in its infancy and many potentially relevant aspects have yet to be studied. If one is interested in the ability of constitutional rules to constrain politicians, one could, e.g., ask whether governments display systematic differences in complying with, or reneging upon, constitutional rules depending on the form of government. In other words, whether the congruence between formal constitutional rules and their actual implementation is influenced by particular traits of the constitution itself. It could be that presidential systems suffer from a greater likelihood of politicians ignoring the rules of the game, even though the formal degree of separation is higher in these systems.11 Presidents often claim that they are the only ones who represent the people as a whole.12 This might make them more audacious than, e.g., prime ministers, in reneging on constitutional constraints. Political parties are regularly weaker in presidential than in parliamentary systems,13 which might further increase the incentive for a president not to take constitutional rules too seriously: if parties are weak, the possibility of opposition to a president who fails to honor the constitution might be less than in systems with strong political parties. A reduced likelihood of opposition does, of course, make reneging upon constitutional rules more beneficial. There might be yet another transmission mechanism concerned with political parties. Brennan and Kliemt (1994) show that organizations such as political parties often have longer time horizons than do individual politicians. While presidents will be out of office after one or two terms (as in Mexico or the United States), political parties might stay in power indefinitely (like in Japan, for a long time after World War II).

11 12

13

These questions are more extensively dealt with in Section 4.4. As the president of France, Charles de Gaulle declared in 1964 “that the indivisible authority of the State is entrusted completely to the president by the people who elected him, that there existed no other authority, either ministerial, civil, military or judiciary which has not been conferred and was not being maintained by him, and finally that it was his duty to adapt the supreme domain, which is his alone, to fit in with those, the control of which he delegates to others” (quoted in Duverger 1980). Although this quote is certainly illustrative, it may not be representative of all presidents. “Strong” and “weak” refer to the organizational structure of parties; they are “strong” if they have many paying members who are active in political office and also follow political events closely. Due to the organizational structure, strong political parties have the capacity to mobilize many people within a short period of time, which might allow them to produce focal points different from those the executive would prefer. Executives in an environment with strong parties are expected to be more likely to play by the constitutional rules than executives in an environment with weak parties.

3.3 Horizontal Separation of Powers

55

If the discount rate of presidents is indeed higher than that of, say, prime ministers or party leaders, offenses against formal constitutional rules may appear more beneficial to presidents than to prime ministers.

3.3.3.2 Determinants of Forms of Government 3.3.3.2.1 Theory Robinson and Torvik (2016) explicitly endogenize presidentialism. They are particularly intrigued by the observation that most African countries chose a parliamentary system as part of their first postcolonial constitution. By now, however, most African states have switched to presidential systems, regardless of the identity of the former colonial power. Based on a model with two groups (each consisting of citizens, politicians, and political leaders), they show that presidentialism is more attractive when the preferences for public goods of the two groups are more polarized, when ideological differences are more extreme, and when the government budget is low, which Robinson and Torvik equate with poor societies. For Aghion et al. (2004), the choice between the presidential or parliamentary form of government plays a prominent role. They ask how much “unchecked power” a society should optimally delegate to its leaders, then proceed to ask under what conditions societies can be expected to do so, and, finally, turn to some cross-country analysis. They equate “insulation” with unchecked power. Drawing on the dichotomy between autocracy and democracy, autocrats are more insulated than democratically elected governments. Within democracy, presidential systems are more insulated than parliamentary ones. What is the central driving force explaining the variation in insulation as defined here? Aghion et al. (2004) find that insulation is positively and significantly correlated with both ethnic and linguistic fractionalization, meaning that highly fragmented societies are less democratic. However, if they are democratic, they can be expected to be presidential rather than parliamentary. These findings are one small first step toward endogenizing forms of government.

3.3.3.2.2 Empirics In two rather exploratory studies, Hayo and Voigt (2010, 2013) endogenize the form of government and ask when a switch in the form of government is likely to occur and, given that a switch has occurred, why it occurred. Hayo and Voigt (2013) find that changes in the form of government can be explained by characteristics of the political system, a country’s involvement in both domestic and international political conflicts, and a number of biographical traits of political leaders, whereas economic and socio-demographic variables do not seem to matter. They ask whether the constitution-making process has an effect on the form of government and find that

56

Positive Constitutional Economics

countries in which constitutional change is initiated by referendums tend to move toward more presidential systems, and those in which constitutional conventions are employed tend to select more parliamentarian forms of government. When the degree of democratization increases, the probability of a change toward presidential regimes in the subsequent period also increases. Conversely, the aftermath of revolutions is characterized by a higher probability of changing the form of government in the direction of a more parliamentary regime. The loss of dependent territory as an outcome of an external war increases the likelihood of changing the form of government toward a parliamentary one. Finally, if a constitution was revised during the 1990s, it is very likely to incorporate more characteristics of a parliamentary system. This finding can be attributed to the rewriting of most constitutions in Central and Eastern Europe during that period. The search for the reasons leading countries to choose a specific form of government has only begun recently. The empirical results just summarized are not based on a comprehensive theoretical model, but on a number of commonsense assumptions. Comprehensive theoretical models are, however, dearly needed. Nevertheless, the empirical findings contain some interesting observations. If personal traits of government leaders are significant determinants, this implies that future studies should take this into account. As of today, data availability is restricted to a limited number of variables (such as how power was acquired, and whether office was left according to the constitution or otherwise). It is definitely a desideratum to add more dimensions, such as the highest academic degree reached, the area of specialization, and whether a politician used to be a member of the army or another important organization. A second noteworthy finding is that political conflicts matter, whereas economic problems such as high unemployment, high inflation, or low growth only matter when they are transferred into the political realm via general strikes, mass demonstrations, and so on. These events appear as domestic political conflicts, and they do play a role.

3.3.4 The Structure of the Legislature: Bicameralism Presidentialism increases the separation of powers by separating the executive from the legislature; bicameralism does the same by dividing the legislature into separate houses. Although bicameralism is a frequently found constitutional setting (according to Tsebelis and Money 1997, it can be found in approximately one-third of all legislatures), its economic effects remain under-researched. In their survey paper on bicameralism, Cutrone and McCarty (2006) note that it “has not received the scholarly attention that other legislative institutions have.” Similarly, Bradbury and Crain (2002) find that, in contrast to the nineteenth century when it was subject to heated discussion, bicameralism “receives scant attention in modern political economy.”

3.3 Horizontal Separation of Powers

57

3.3.4.1 Effects of Bicameralism 3.3.4.1.1 Theory The differential effects of unicameral and bicameral legislatures were first analyzed from an economic point of view by Buchanan and Tullock (1962: chapter 16). In their analytical frame, that decision-rule is optimal that leads to a minimum of interdependence costs.14 They conjecture that in comparison with unicameral systems, bicameral systems have higher decision costs. They go on to state that “[o]n the other hand, if the basis of representation can be made significantly different in the two houses, the institutions of the bicameral legislature may prove to be an effective means of securing a substantial reduction in the expected external costs of collective action without incurring as much added decision-making costs as a more inclusive rule would involve in a single house” (Buchanan and Tullock 1962: 235f.). The larger the majority required to reach a certain decision, the lower the external costs connected with that decision because the number of opponents is negatively correlated with the required majority. On the other hand, it will become increasingly difficult to reach a decision at all because the decision costs are positively correlated with the required majority. One possibility for keeping external costs down is to require a supermajority (say of 3/4 or 5/6) in the single-house system. Supermajorities in a single-house system and simple majorities in a two-house system can thus be considered as alternatives. Buchanan and Tullock conjecture that, given identical external costs, decision costs will be lower in a bicameral than in a unicameral system. Miller and Hammond (1989) inquire into the effects of bicameralism and the executive veto, sometimes considered the third chamber, on stability in the sense of reducing the probability of cycling majorities à la Condorcet or Arrow (1951). They conclude that bicameralism and the executive veto increase stability. The stabilityenhancing effect of bicameralism depends on preference heterogeneity between the two chambers. Levmore (1992) changes the focus of the analysis somewhat when he conjectures that a bicameral system might be better than a corresponding qualified majority in a unicameral system at reducing the power of the agenda-setter. Bicameral systems are often interpreted as a check against overly active legislatures.

14

Buchanan and Tullock introduce two cost categories. External costs are those costs the individual expects to bear as a result of the actions of others over which he or she has no direct control. Second, there are decision making costs, which the individual expects to incur as a result of the individual’s own participation in an organized activity. These costs include only the estimated costs of partici pating in decisions when the agreement of two or more individuals is required. The sum of these two cost categories is the third category, costs of social interdependency, or simply interdependence costs. For purely private activity, these will be zero. According to Buchanan and Tullock, a rational individual confronted with questions of constitutional choice will opt for minimization of interdependence costs.

58

Positive Constitutional Economics

In sum, then, it is argued that bicameralism protects minorities, reduces cycling, and curbs the power of the agenda-setter. There are conflicting hypotheses regarding the potential effects of bicameralism on fiscal policies. Heller (1997) argues that the larger number of veto players in bicameral systems leads to higher budget deficits. To secure the consent of an additional veto player, additional expenditures need to be agreed upon, which will lead both to higher spending and deficits. But frequently – and following the early advances by Buchanan and Tullock just discussed – the exact opposite is argued. Bradbury and Crain (2002) focus on the composition of government spending and theorize that bicameralism influences the structure of the budget in two ways. On the one hand, spending on redistribution should be lower (minorities being protected against expropriation) and, on the other, spending on potentially productivity-enhancing public goods should be higher than in unicameral legislatures. Levmore (1992) argues that bicameralism should curb corruption. Diermeier and Myerson (1999) show that the cost of lobbying increases when it is harder to pass fresh legislation, which is systematically more difficult under bicameralism than under unicameralism. This implies that there should be less lobbying, as well as less rents and corruption, under bicameralism. After reiterating that bicameralism can be connected to gridlock and status quo bias, Facchini and Testa (2016) draw attention to a third possible disadvantage of having two chambers, namely the time needed to pass fresh legislation. While many of the early supporters of bicameralism, such as the authors of the Federalist Papers, considered delay a virtue because it protected legislators from popular passions and, hence, the tyranny of the majority, Facchini and Testa stress that being subject to harder time constraints reduces the bargaining power of legislators vis-à-vis lobbyists and, therefore, lobbyists might be more successful under bicameralism than under unicameralism. The authors somehow oversell the results of their model though, as they equate lobbyism with corruption.

3.3.4.1.2 Empirics Empirical evidence on the effects of bicameralism is sparse. Congleton (2003, 2006) deserves special mention because he tries to isolate the specific effects of bicameralism by running simulation models. He conjectures that bicameral legislatures can also affect legislation if representation is unbiased, i.e., no group has special representation. Based on his simulation model, he finds that bicameral legislatures are more faithful to the median voter, at least in the long run. The model is complemented by an analysis of the effects observed in two countries that switched from bicameralism to a single chamber (Sweden and Denmark). In these two cases, policy predictability indeed decreased as a consequence of constitutional change. It is certainly a desideratum to analyze the reverse case and investigate whether

3.3 Horizontal Separation of Powers

59

policy predictability increases after a switch from a single-chamber to a twochamber system. Bradbury and Crain (2002) construct a continuous variable (based on the degree of constituent heterogeneity) for bicameralism across US states and find that higher levels of bicameralism reduce government expenditures. In their cross-country study, Bradbury and Crain (2001) find that government spending increases with the number of electoral districts, but that this effect can be curbed by a second chamber. Plümper and Martin (2003) also find that bicameralism is highly significantly and robustly correlated with lower government spending. Regarding the effects of bicameralism on governance indicators, Testa (2010) finds that bicameralism as such is not significant in explaining differences in corruption levels, but that if bicameralism is interacted with both party alignment across the two chambers and the polarization between the incumbent and the challenging party, it turns out that aligned and polarized chambers reduce corruption. Note, however, that this study is confined to 34 democracies and comprises only five years. Some of the theoretically derived hypotheses regarding the effects of bicameralism are, hence, supported by the data: policy predictability is enhanced, and both government expenditure and corruption levels are lower than in unicameral legislatures. Empirical tests of other hypotheses still need to be conducted. For example, do bicameral systems lead to a composition of government spending systematically different from that observed under unicameral legislatures? An even larger step in future research could consist in interacting the effects of bicameralism with other constitutional traits. Here, bicameralism is analyzed as one aspect of the separation of powers. An analysis of potential interactions between bicameralism and the form of government, on the one hand, and federalism, on the other, is an obvious starting point for future research.

3.3.4.2 Determinants of Bicameralism To date, there is very little evidence on the factors that cause countries to choose a bicameral legislature. A very simple, yet broad model regarding the choice of bicameralism is contained in Muthoo and Shepsle (2008). They put special emphasis on whether legislators in the respective chamber are all chosen simultaneously or in staggered elections. Examples for the latter include both the US Senate and the German Bundesrat. Relying on comparative institutional analysis, they compare five different institutional settings: a unicameral legislature whose members are either elected simultaneously (1) or staggeredly (2); a bicameral legislature, where either all legislators are elected simultaneously (3) or staggeredly (4), or where all the members of one house are elected simultaneously and the other house in staggered elections (5). They then have the legislators play a “divide-the-cake” game. Among their many findings, the one that at the constitutional stage, the choice of a bicameral system

60

Positive Constitutional Economics

Pareto dominates the choice of a unicameral system is the most relevant here. The authors point out that their paper can only be a first step into endogenizing bicameralism. Patterson and Mughan (1999: 10) report that smaller countries are more likely to choose a unicameral system, whereas larger countries are more likely to choose a bicameral one.

3.3.5 A Forgotten Branch: The Judiciary 3.3.5.1 The (Economic) Effects of the Judiciary The judiciary has frequently been called “the least dangerous branch” (e.g., Hamilton 1788/1961, Federalist Paper #78), which may be one reason why scholars in constitutional economics turned their attention to its analysis only recently. As one example, in his treatise on Comparative Constitutional Engineering, Sartori (1994) deals extensively with electoral systems and forms of government, but the term “judiciary” does not even appear in the index.

3.3.5.1.1 Theory When considering the separation of powers, independence of the judiciary vis-à-vis the other two government branches is crucial. However, early models of the separation of powers discussed above do not contain the judiciary as an actor. Padovano et al. (2003) extend the Persson et al. (1997) model by incorporating the judiciary as a third actor. The judiciary is modeled as another agent that functions as a constraint on the executive or the legislature. They find that an independent judiciary improves accountability, but that this is not true of an accommodating judiciary. They conjecture that the likelihood of the three branches colluding (with the result that accountability suffers) depends on term lengths. Specifically for the judiciary, they expect longer term lengths to lead to lower accountability and less welfare. That the effects of an independent judiciary are different from those of an accommodating judiciary is intuitively plausible, and yet, in order to be interesting, the conditions under which one can expect a judiciary to be independent or accommodating need to be specified. Unfortunately, the authors do not deal with this issue. Likewise, their conjecture on term lengths is debatable: in fact, shorter term lengths and the option to be reappointed might increase accountability, but this would be accountability to those who have the power to reappoint, not necessarily accountability to the letter of the law. An independent judiciary can be one means of solving the dilemma of the strong state. On the one hand, a state strong enough to protect private property rights is needed; on the other, a state that is sufficiently powerful to protect private property rights is also sufficiently powerful to attenuate or outright ignore them. This situation

3.3 Horizontal Separation of Powers

61

is detrimental for all parties. Citizens who anticipate that their property rights might not be completely respected have fewer incentives to create wealth. The state, in turn, will receive a lower tax income and will have to pay higher interest rates as a debtor. An independent judiciary can reduce this dilemma if its decisions are based on the relevant legislation and are regularly enforced by the other government branches despite not being in their own (short-term) interest. Given these conditions, aggregate investment will rise and the economy will grow more quickly.

3.3.5.1.2 Empirics Is this theoretical conjecture supported by the data? One huge problem, of course, is to make the degree of independence that the judiciary enjoys in various countries measurable and, thus, comparable. Feld and Voigt (2003) introduce two indicators, one dealing with de jure independence (i.e., the independence of the courts as it can be deduced from legal documents) and the other with de facto independence (i.e., the degree of independence that the courts actually enjoy). Drawing on a sample of 73 countries to estimate the impact of JI on economic growth, Feld and Voigt (2006) find that while de jure JI does not have an impact on economic growth, de facto JI positively influences real GDP per capita growth. More recently, Voigt et al. (2015) replicate the earlier studies focusing on a more recent time period and using a partially different sample of countries. The earlier results are not only confirmed but also extended. In their most recent study, the authors show that countries with greater improvements in de facto JI are also likely to grow faster, and thus, getting closer to a causal relationship. Quite a few questions are left to be answered. What are the transmission channels through which de facto JI impacts economic growth? What role does judicial accountability play? Is the specific organization of court systems relevant? To name just one of many possible aspects, over the last couple of decades, many countries have created judicial councils to manage the entire judicial branch. The general idea is to make the judiciary more independent from the executive. In many countries, the administration of the court system is under the auspices of the ministry of justice and, hence, not independent from the executive. But whether judicial councils really help to increase the independence of the judiciary is completely unclear. In fact, there is some evidence (Voigt and El-Bialy 2016) that the exact opposite might be the case. Given that JI has significant economic effects, we would like to know its determinants.

3.3.5.2 Determinants of Judicial Independence 3.3.5.2.1 Theory Conventional wisdom has it that an independent judiciary constrains the other two branches of government. Prima facie, one would therefore expect the other two

62

Positive Constitutional Economics

branches to be keen on a judiciary without teeth. Landes and Posner (1975) were the first to question this conventional wisdom from an economic point of view. According to them, the legislature is not controlled by the judiciary, but legislators have an interest in an “independent” judiciary because its existence makes it more valuable to be a legislator as it can prolong the lifespan of legislative deals that legislators strike with representatives of interest groups. This prolonged expected lifespan of deals increases their value to the interest groups, simultaneously increasing the groups’ willingness to pay the legislators. Landes and Posner define a judiciary as independent if it enforces “existing statutes in accordance with the intent of the enacting legislature” (1975: 883) and, thus, produces stable expectations. The legislators have an interest in being able to make credible commitments vis-à-vis interest groups. An independent judiciary enables them to do this by reducing the possibilities for postcontractual opportunism, either by themselves or by their successors. According to Landes and Posner, the political branches have various means of imposing costs upon the judiciary (“budgetary harassment, tinkering with the courts’ jurisdiction and altering the composition of the judiciary by the creation of many new judgeships” [1975: 885]), which, in turn, can maintain its independence best by enforcing the “contracts” earlier legislatures struck with interest groups. Landes and Posner’s paper can also be interpreted as a theory on the choice of JI. Later papers have both challenged and agreed with that view. The interest group theory of JI can be criticized on various grounds. First, the concept of “independence” seems shaky. Landes and Posner equate JI with “enforcing the original intent of legislative deals.” It thus sees judges as loyal agents of the enacting legislature, a rather unconventional view of JI. In addition, Landes and Posner would need to clarify what incentives judges have to enforce deals made by previous legislatures. After all, their salaries depend on current legislators and not past ones.15 In the meantime, other approaches aiming to explain the determinants of JI have been developed. For Ramseyer (1994), court independence depends on the expectations of politicians: if they expect their own party to remain in power, they will have less incentive to create an independent judiciary than if they expect to lose power to a competing party soon. Ginsburg (2002) can be interpreted as a broadening of Ramseyer’s approach. He believes that politicians are likely to choose a higher degree of judicial review (as one important aspect of JI) the greater the degree of political uncertainty at the time of constitutional design. Hayo and Voigt (2016) inquire into the factors driving changes in constitutionally entrenched de jure JI over time. Inspired by Ramseyer (1994) as well as Ginsburg 15

A number of scholars, mostly from the Virginia tradition of public choice, try to test the Landes and Posner hypothesis empirically (e.g., Crain and Tollison 1979; Anderson et al. 1989, 1990); their results are summarized in Voigt (1997) as well as Shugart and Tollison (1998).

3.3 Horizontal Separation of Powers

63

(2002), the authors develop two theoretical perspectives to test which one better conforms to the data. The first perspective is called the democratic competition view. It argues that institutions fostering JI are more likely to arise under conditions of strong political competition and substantial differences in the positions taken by political parties. The second perspective is called the commitment view according to which an improvement of formally guaranteed JI makes it easier for the ruling elite to make credible constitutional bargains with the rest of society. It argues that rational politicians who are aware of the dilemma of the strong state described in the previous section have incentives to try to mitigate the dilemma by entrenching high levels of de jure JI in the constitution.

3.3.5.2.2 Empirics The scant empirical evidence on determinants of JI is structured by focusing on two dimensions. The first is the distinction between de jure and de facto. Within these two dimensions, a further distinction can be made by asking whether the first time choice of JI is analyzed, or its change over time instead. Hayo and Voigt (2016) identify variations in JI based on changes in the constitutions of as many as 100 countries that occurred between 1950 and 2005. They find strong evidence that changes in political participation and competition, as well as the characteristics of individual leaders (such as how they acquired or lost power), play an important role in explaining changes in constitutionally safeguarded JI. However, no clear-cut evidence supporting one theoretical perspective over the other is found. Ginsburg and Versteeg (2014) study the determinants of constitutional review. Judges who have the power to review the constitutionality of acts passed by the other two government branches need to enjoy actual JI. Hence, JI is a necessary condition for constitutional review. Their results are closely related to the study by Hayo and Voigt just discussed. In the language just introduced, Ginsburg and Versteeg (2014) find support for the democratic competition view. In contrast, they do not find evidence that the adoption of constitutional review is driven by attempts to strengthen the rule of law (“ideational factors” in their terminology), by federalism (because there is greater potential for conflict in federally structured states) or by the international diffusion of innovative constitutional rules. Hanssen (2004) tests two predictions first generated by Ramseyer (1994), namely, that JI will be higher (1) if politicians fear losing power, and (2) the farther apart the ideal points of the rival parties. Using judicial retention procedures as a proxy for JI, he finds empirical support for these hypotheses with regard to states in the United States. Besley and Payne (2003) used a similar approach to explain differences in judicial behavior. They find that judges make decisions that favor important sectors of the electorate, as this can increase their chances of re-election. Both of these

64

Positive Constitutional Economics

empirical studies deal only with the United States. In the meantime, the first crosscountry studies have been published. The studies just presented all focus on the first-time choice of constitutional rules referring to the judiciary or their change over time. They are hence exclusively concerned with constitutional text and neglect constitutional reality. Hayo and Voigt (2007) is a first attempt at explaining variation in de facto JI across countries. They find that high levels of de jure JI, the extent of democratization, and higher degrees of press freedom are good predictors for high levels of de facto JI. Melton and Ginsburg (2014) take issue with the result that de jure JI could be the single most important driving force of de facto JI, arguing that those rules that are more likely to be self-enforcing should be given more importance. For example, if the appointment process of judges is the joint responsibility of different players who have competing interests, this might be an important factor for observing high levels of de facto JI. Moving from a cross-sectional to panel data analysis, results change considerably. Overall, Hayo and Voigt (2019) find a positive trend for de facto JI. However, there is important regional variation. In Latin America, the positive trend is particularly pronounced, whereas in Africa, they only find a weak positive trend. Regarding the question whether de jure JI is a good predictor for de facto JI, they come up with a number of unexpected results. In Eastern Europe, increases in de jure JI are, indeed, a good predictor for improvements in de facto JI. However, in Western Europe and North America, the opposite holds; improvements in de jure JI are followed by a decline in de facto JI. More work is needed that inquires into the utility functions of and the (perceived) constraints on judges. It is interesting that Mueller (1996: chapter 19) does not think that formal institutions could be sufficient to ensure a judiciary that is not only independent but also accountable. For incentives to make judges decide impartially (1996: 284), “one has to rely on ‘the culture of the judiciary’ and the great status (and possibly financial rewards) that surround it.” This statement gives rise to at least two implications. First, an independent and accountable judiciary might simply be impossible, if the respective culture does not support it. Second, informal institutions might be an important factor in the actually realized level of JI. Based on data from the EU Justice Scoreboard, with an attempt to make the central traits of the justice systems of EU member states measurable and comparable, Gutmann and Voigt (2018) identify a puzzle that is largely compatible with Mueller’s observations. National levels of JI (as perceived by the citizens of EU member states) are negatively associated with the presence of formal legislation usually considered as conducive to JI. After attempts to resolve this puzzle based on political economy explanations and specificities of legal systems are unsuccessful, the authors turn to cultural traits and find that, indeed, countries with high levels of generalized trust (and to a lesser extent individualistic countries) exhibit increased levels of de facto JI and, at the same time, reduced levels of de jure JI. The combination of these two

3.4 Vertical Separation of Powers: Federalism

65

effects can explain why judicial reforms that should be conducive to an independent judiciary may seem to have adverse consequences. Gutmann and Voigt (2018) conclude that cultural traits are of fundamental importance for the quality of formal institutions, even in societies as highly developed as the EU member states.

3.4 Vertical Separation of Powers: Federalism 3.4.1 Definitions We now turn to federalism, i.e., the vertical separation of powers. Federalism is closely related to bicameralism as discussed in Section 3.3.4, because all states with a federal constitution have a second chamber which represents the constitutive units. The opposite, however, is not true. Countries can have a bicameral system without having a federal constitution. The United Kingdom is a prominent example. In economics, the branch interested in analyzing the optimal allocation of competences to the various layers of government is usually referred to as “fiscal federalism.” But beyond incorporating a second (or even possibly a third) layer of government into their analyses, scholars working within this approach largely remain within the traditional model that assumes government is efficiency maximizing. They then ask on what governmental-level public goods will be (optimally) provided, taking externalities explicitly into account. In addition, early research in this tradition has also been concerned with economies of scale in the provision of public goods and heterogeneous preferences in their consumption. In the words of Gerring et al. (2006: 7): “Fiscal federalism is a theory of public administration, not of constitutional structure.” In the language of constitutional economics, it deals with the optimal allocation of specific government tasks on the various government levels, but all this at the post-constitutional stage (Inman/Rubinfeld 2016 and Oates 1999, 2005 are excellent surveys of this approach). As we are interested in the economic effects of constitutions, we propose to distinguish between federalism on the one hand, and (fiscal) decentralization on the other. Federalism is a constitutional-level trait; decentralization describes policy choices made at the post-constitutional level. This implies that decentralization can also occur in nonfederally structured states – just think of the process called “devolution” in the United Kingdom.16 Riker (1975: 101) defines federalism as follows: “Federalism is a political organization in which the activities of government are divided between regional governments and a central government in such a way that each kind of government has some 16

Blume and Voigt (2011) have collected more than two dozen indicators of both federalism and decentralization and use factor analysis to test whether the latent variables neatly separate between the two concepts. They find that more than two factors emerge, i.e., there are more than two relevant dimensions that can be distinguished from each other.

66

Positive Constitutional Economics

activities on which it makes final decisions.” In other words, federations consist of constituent governments (the regional governments) and one central government and both levels of government are endowed with final decision-making power in some areas. As we are interested in constitutional provisions, we add that at least one area of final decision-making for each type of government (i.e., regional and central) should be explicitly mentioned in the constitution. Further, this allocation of decision-making competence will be stable over time only if it cannot be changed by either the central government or the regional governments unilaterally. In other words, constitutional change requires the consent of both levels of government. Riker himself (1975: 102) points out that his definition leaves ample room for variation within the group of federally constituted states. At a minimum, the central government of the federation can make decisions in only one narrowly restricted area. At maximum, it can make decisions in all but one area. Ideally, indicators of federalism ought to reflect these differences. One attempt to deal with the large heterogeneity within the group of federally constituted countries is to ask whether they follow a cooperative or a competitive approach. Cooperative federalism assumes that most areas are shared, or that there are some mechanisms that allow the central and regional governments to co-decide in these areas. In more competitively constituted federations, more policy areas are exclusively allocated to either the federal or state level. Although this distinction has been around for a long time, it has not found its way into empirical studies because no established indicator has been available.17 At the end of the day, empirical studies of federalism have also had to struggle with the fact that only around two dozen states the world over have a federal constitution.

3.4.2 Effects of Federalism 3.4.2.1 Theory Very simply speaking, a possible advantage of federally constituted states over unitary ones is that lower government levels are better able to cater to the specific preferences of the citizens in the respective area. People living close to the sea are likely to demand different public goods (e.g., dikes) than people living in the mountains (e.g., protection against avalanches) and local governments are hypothesized to be better able to cater to these preferences than a unitary central government. If local governments compete for taxpaying citizens (or investments), they have incentives to take into account the preferences of those whom they want to attract (Tiebout 1956). A related argument was already made by Hayek back in 1939. 17

It is too early to tell whether the indicator “shared rule,” which has recently been proposed by Hooghe et al. (2016), will be successful in that sense.

3.4 Vertical Separation of Powers: Federalism

67

He posits that competition between governments will reveal information on efficient ways to provide public goods. Assuming that governments have incentives to make use of that information, government efficiency should be higher in federations, ceteris paribus. One possible cost of federal constitutions is that, if the number of states is large, economies of scale in the provision of public goods could remain unrealized. For example, Tanzi (2000) suspects that those providing public goods will be insufficiently specialized. Also, federal states need to deal with a moral hazard problem that is not an issue in unitary states.18 The federal government will regularly issue “nobail-out clauses,” i.e., announcements declaring that in case single member states cannot service their debts anymore, the central government will not come to their rescue. But no-bail-out clauses will not always be credible.19 With regard to the issue of over-borrowing, Wildasin (1997) argues that large states can become “too big to fail.” On the other hand, it has been argued (Rodden and Wibbels 2002) that large member states can internalize more of the benefits generated by responsible fiscal policies. The conceptual notion behind this argument is that of a free rider. Members of a federal union could try to benefit from the services provided by the federal government without paying their fair share. Likewise, they could also pursue policies (like borrowing too much money) that put costs on either the other states of the federal government, or both. A number of factors might mitigate this free-rider problem. If strong, disciplined parties are active throughout most of the federation and one party is in charge of the federal as well as most of the constituent governments, then party leaders may be able to prevent state officials from externalizing the negative effects of overborrowing (Rodden and Wibbels 2002). Notice that this mitigating effect is composed of institutional and noninstitutional aspects. The structure of the party system is a consequence of both the heterogeneity of the country and of its electoral institutions, whereas dissimilitude between federal and constituent governments is the result of voter choice. There is one standard answer to the question of whether corruption is more prevalent under federal or unitary constitutions: Constituent governments are closer to the people, play infinitely repeated games with local constituents, and hence are subject to local capture, i.e., the capture of local government by local elites (see, e.g., Tanzi 2000). Therefore, corruption levels will be higher under federal than under 18

19

The relationship between the central government and the lower units in unitary states might be more aptly described drawing on principal agent theory with its familiar monitoring problems. For such a view, see Seabright (1996). Rodden (2002: 672) points out that the creditworthiness of the federal level might be jeopardized if it does not bail out the constituent governments. On the other hand, there is some evidence that the clause will be perceived as credible after it has been effectively used at least once, i.e., at least one lower level government has gone bankrupt. See, for instance, Bordo et al. (2013).

68

Positive Constitutional Economics

unitary constitutions. The standard argument against the local capture hypothesis is that the behavior of constituent governments is more easily monitored by local populations in federations, and politicians are, hence, more accountable for their actions. This would imply that corruption is lower under federal constitutions. Additionally, corruption can signal an inadequacy in the relevant rule system; under dysfunctional rules, even welfare-enhancing activities will often require corrupt behavior. This assumption leads to the argument that since the constituent units of federal states are closer to the people, it is likely that their rules will be more adequate than those in unitary states. Regarding the stability of government action, policy swings will be less pronounced in federal, as opposed to unitary, countries, when there are changes in the national government. This better predictability of government policy allows private actors to form expectations over a longer period of time, which might, in turn, increase overall productivity. On the other hand, the existence of a number of power centers always entails the possibility of power struggles and instability. This might mean that, ceteris paribus, federal constitutions will result in lower levels of productivity than unitary states. Empirical tests are necessary to resolve these contradictory theories.

3.4.2.2 Empirics Estimating the effects of federalism presupposes the possibility that such effects can be ascertained. As argued in the previous section, federalism is a consequence of a constitutional choice, whereas decentralization is the consequence of policy choices. We are here primarily interested in the former. Various dummy variables indicating whether a country has a federal structure or not have been widely used. The congruence between them is fairly high with only two contentious cases, namely South Africa and Spain. A second type of indicator looks into the allocation of competences between the federal and the constituent governments in more detail; for example, asking what level is granted the residual power in case the constitution remains mute on a specific policy area. A subgroup within this type of indicator enquires into the allocation of fiscal competences as this is of high interest to economists. Blume and Voigt (2011) briefly discuss and evaluate more than two dozen of such indicators. What can we learn from existing empirical studies? For a long time, the evidence concerning the effects of federalism was mixed. It was, for instance, unclear whether federally constituted states had higher overall government spending or not. This appears to have changed now. Rodden (2003), in a cross-country study covering the period 1980–93, shows that in countries in which local and state governments have

3.4 Vertical Separation of Powers: Federalism

69

the competence to set the tax base, total government expenditure is lower. Feld et al. (2003) find that more intense tax competition leads to lower public revenue. Drawing on a dummy variable for federalism first introduced by Daniel Elazar (1995), Treisman (2000) found that federal states have, ceteris paribus, higher corruption levels than unitary states. However, employing data that only became available after this early paper, he later (2007: 235) found that the correlation between perceived corruption and federal structure is not at all robust. Persson and Tabellini (2003: 61) find that federalism is not a significant variable in explaining rent extraction (which they use as a proxy for corruption). The evidence is thus inconclusive, warranting more fine-grained research. With regard to productivity, Persson and Tabellini (2003: 71) find that the federalism variable is highly significant for explaining differences in both labor and total factor productivity, with federations having higher levels of each. The available evidence concerning the effects of federalism on economic growth is just as ambiguous as the theoretical conjectures. There are only half a dozen studies with cross-country evidence and these are often limited to the OECD member states. Based on the various factors that Blume and Voigt (2011) generate in the principal component analysis briefly alluded to above, Voigt and Blume (2012) conclude that institutional detail clearly matters. Their analysis broadens our knowledge about the effects of federalism in two ways: • In those cases in which the federalism dummy is significantly correlated with a dependent variable, it allows more precise insights into the underlying transmission channel. Here is an example: with regard to total government expenditure, the federalism dummy is positive, indicating that federally structured states have higher overall spending than unitary ones. The factors produced by Blume and Voigt now allow us to pin down the factors driving that relationship: it is (1) the unconditional right of lower government levels to a portion of overall government revenues, and (2) the fractionalization of parliament. Interestingly, one variable is institutional, and the other a consequence of institutions. • The second way in which factor analysis broadens our knowledge is cases in which frequently used federalism dummies (such as Elazar 1995) turn out to be insignificant in explaining variation, whereas specific aspects of federalism are significant. For example, federalism dummies are not significantly correlated with budget surplus, i.e., government deficits. However, drawing on more fine-grained indicators, one learns that the local election of municipal governments, as well as the fractionalization of parliament, is correlated with higher government deficits, implying that the presence of these features leads to larger deficits. This study shows that it can be important to take institutional details into account and that simple dummy variables are often insufficient.

70

Positive Constitutional Economics

3.4.3 Determinants of Federalism 3.4.3.1 Theory Our knowledge concerning the conditions under which constitutional assemblies choose federal, rather than unitary, structures leaves much to be desired. Figueiredo and Weingast (2005) introduce a two-stage game to deal with the issue of how the institutions of federalism can be sustained. In the first stage, the institutions of federalism are determined (“institutional game”). In the second stage, a repeated game is played in which participants can act either in accordance with the institutions established in the first stage or renege upon them. Figueiredo and Weingast (2005) stress the need for federal institutions to be self-enforcing. They show that constitutional rules can serve as a coordination device that allows the members of a federation to act collectively in case the center tries to extract rents from one or more members. Figueiredo and Weingast demonstrate that sustained federalism can be an equilibrium of their game. Although they look at a few countries to demonstrate some of the implications of their model, their focus is chiefly theoretical. Still lacking, then, are empirical insights identifying the conditions under which federal constitutions are established in the first place. Riker (1975) argued that being subject to an external threat is a necessary condition for states to unite in a federal structure. Arzaghi and Henderson (2005) propose to think of fiscal decentralization as “effective federalism,” and conjecture that increases in population size make a larger number of autonomous units more effective as fixed costs for certain goods could be borne by more people. This argument is, however, a traditional efficiency argument that does not take political economy aspects into account. Gerring et al. (2011) are interested in a slightly broader question, namely the determinants of “indirect rule” – of which federalism is, however, one occurrence. They argue that if effective regional governments were in place before the formation of the currently existing nation states, this would have made the passing of a federal constitution more likely. Lessmann (2014) is interested in the determinants of spatial inequality, i.e., the distribution of income across the regions of a country. Gutmann and Voigt (2016) argue that a high degree of spatial inequality could be one factor making a federal constitution more likely.

3.4.3.2 Empirics In one of the few attempts to identify the determinants of federalism empirically, Gutmann and Voigt (2016) try to put some of the hypotheses just described to the test. According to them, the land area of a country is the strongest predictor for federal structures: the larger a country in terms of area, the more likely it is to have a federal constitution. Among the more detailed predictions just described, they show that high levels of spatial inequality – i.e., differences in income and wealth across

3.5 The New Separation of Powers

71

the various regions of a country – are correlated with a higher likelihood of finding a federal constitution. One limitation confronting this research strategy is the small number of observations, as there are only a few federal states (some 20 countries worldwide). One possible way of circumventing this problem is to draw on case studies instead of econometric estimates, which is exactly what Blankart (2000) does. Blankart compares the development of two federally organized states over time, namely, Switzerland and Germany. It is often presumed that even in federal states, there are inherent centralizing tendencies. Blankart asks whether the so-called law of the attracting power of the highest budget (also called Popitz’s law) is a law in the sense of natural sciences, or whether it is a consequence of constitutional choices. He conjectures that centralization is a function of the cartelizing tendencies between federal and state government levels, which, in turn, would be a function of constitutional rules. He then shows that the competence of the federal level to appropriate tax competence from the state level is crucial for explaining differences in the centralizing tendencies of different federations. Vaubel (1996) analyzes the process of centralization in federal states empirically. His first result is that federal states are indeed less prone to centralization than states with a nonfederal constitution. Centralization is measured as the share of central government expenditure in total government spending. If one replaces the binary dummy variable for federalism by quantitative constitutional variables, the most powerful single explanatory variable is the age of the constitutional court (for the entire sample), or the independence of the constitutional court from the organs of central government (for the industrialized states). The explanatory power of the model can be raised further if one takes into account: (1) the degree of control that the lower level governments have over changing the federal constitution, and (2) whether tax increases require a popular referendum. It thus seems that some constitutional provisions can make a difference in constraining centralization. Searching for the determinants of federalism implicitly keeps another variable constant, namely the size of the nation. Yet, Alesina and Spolaore (2003) have shown that their size can also be endogenized. They describe the choice as a tradeoff between the benefits of size (i.e., the economies of scale in the provision of public goods) vs. the costs of heterogeneity. They show that optimal size is also a function of whether a country is democratic or not. They predict that transitions from autocracy to democracy will be accompanied, if not by the entire breakup of nation states, at least by a higher level of decentralization.

3.5 The New Separation of Powers: Increasing the Number of Independent Decision-Makers The separation of powers is thought of as involving three branches of government: legislature, executive, and judiciary. When distinguishing between horizontal and

72

Positive Constitutional Economics

vertical separation of powers, vertical separation refers to the different layers of government, in particular, federalism. A modern view of the separation of powers, however, would not stop there. Over the last couple of decades, a large variety of unelected bodies has been created. They have been given various names, such as non-majoritarian institutions (Majone 1996) or independent regulatory agencies (e.g., Gilardi 2002 or Thatcher 2002). They cover many policy areas such as health, water, energy, telecommunications, the environment, the financial sector, competition, and many more. Here, we refer to them as independent decision-makers. The best-known policy area that has been delegated from the executive to independent decision-makers is probably monetary policy. By now, many countries pride themselves with having a central bank that is independent from direct government interference. Governments that are subject to a re-election constraint are tempted to beef up monetary supply shortly before elections because this is likely to improve their re-election chances. The downside of expanding monetary supply is, however, that inflation will rise and citizens will not really be better off. So delegating monetary policy to an actor who is not subject to a re-election constraint seems straightforward. Usually, this sort of delegation is thought of as domestic delegation, but in recent years, more and more competence has been delegated to international organizations. Monetary policy is a good example of both domestic and international delegation. The effects of independent central banks on inflation rates have been analyzed extensively (de Haan and Eijffinger 2019 is an up-to-date survey). But monetary policy is also a good example of delegation at a level beyond the nation-state. The European Central Bank is the most striking example. Others include currency boards and the recognition of foreign currencies as legal tender, as in some Latin American countries. It seems justified to analyze the economic effects of delegated powers within constitutional economics, as at least some of the delegated powers are dealt with on the constitutional level. Voigt and Salzberger (2002) deal with the question of what circumstances lead legislatures to prefer international over domestic legislation if both are feasible. This primer is concerned with the effects as well as the determinants of constitutional issues, not with their normative justification. Before delving into positive issues, it is, however, worth mentioning that the normative foundations of these decision-makers are subject to heated debate: Can they and their agencies be incorporated into established theories of constitutionalism or do we need new theory? (Vibert 2007 is in favor of the latter). More concretely, how much power can be delegated to unelected bodies without undermining the core of democracy, namely, majoritarian decision-making? Beyond these normative issues, at least three questions need to be dealt with when investigating these unelected bodies: (1) What are their functions? (2) How can their status be measured and compared across bodies? (3) What are their effects? The two

3.6 Representative vs. Direct Democracy

73

functions of these independent decision-makers most frequently discussed are as follows: mitigation of the principal–agent problem between the citizens and their government, and to commit against time-inconsistent preferences. The name of the game in principal–agent problems is the underlying asymmetrical information: the agents (the government) know more than their principals (the citizens). Monitoring the behavior of the agents by their principals is therefore costly and, hence, unlikely to take place. Unelected bodies that are independent from the government can help to mitigate the information asymmetry by providing information to the citizens that, in turn, incentivizes government to choose policies closer to the preferences of the median voter. Beetsma and Debrun (2016), for instance, develop a model of a fiscal council that enlightens the understanding of citizens about fiscal policy issues with the intention of encouraging the intended effect of the policy. The other frequently discussed function of unelected bodies is their assumed ability to deal with a time-inconsistency problem that government is subject to. Government might promise non-inflationary monetary policies, but succumb to the temptation of loosening up if that increases re-election chances. This is, of course, the standard justification for independent central banks. But the same logic applies to many other policy areas as, for example, competition policy or the regulation of network industries. If these agencies are to fulfill the functions just described, they need to be independent from government. Most attempts to measure the status of unelected bodies have, indeed, centered on their independence. Hanretty and Koop (2012) are interested in measuring the formal independence of regulatory agencies. They justify their choice to study formal, rather than factual, independence by pointing out that if one is interested in the reasons why competence is delegated, then this is a natural starting point. Other ways to measure the independence of these agencies have been proposed; Gilardi and Maggetti (2011) is one such attempt. Identifying the effectiveness of these agencies is challenging. Many of them have not been around for a long time. Their goals vary considerably between countries. Often, the goals are not formulated very precisely. We are not aware of any crosscountry studies assessing their effectiveness.

3.6 Representative vs. Direct Democracy 3.6.1 Introductory Remarks A broad understanding of the separation of powers also includes direct democracy institutions that allow citizens to act as an additional veto player. It has been hypothesized that such institutions make politicians more accountable and result in policy choices that more closely match citizen preferences. Similar to federalism, it is unclear ex ante whether that should mean lower taxes, budgets, and deficits,

74

Positive Constitutional Economics

although most of the literature appears to assume that this should be the case. As of 2006, more than 100 countries had anchored the possibility of a referendum into their constitution (Gutmann et al. 2017).

3.6.2 Effects of Direct Democracy 3.6.2.1 Theory In real-world societies of a size too large to efficiently vote directly on all issues, representative and direct democracy are complementary institutions. In these societies, a certain degree of direct democratic institutions is combined with representative institutions. With regard to direct democratic institutions, referendums are usually distinguished from initiatives.20 The constitution can prescribe the use of referendums for passing certain types of legislation, in which case agenda-setting power remains with parliament, but citizen consent is required. Initiatives, in contrast, allow citizens to become agenda-setters. The citizens propose a piece of legislation that will then be decided upon, given that they manage to secure a certain quorum of votes in favor of the initiative. Initiatives can be aimed at different levels of legislation (constitutional vs. ordinary legislation), and their scope can vary immensely (e.g., some constitutions prohibit initiatives on budget-relevant issues). In their paper on the effects of direct democratic institutions on total factor productivity in Switzerland, Feld and Savioz (1997: 515) point out the lack of simple theories connecting direct democracy with overall productivity. But if direct democracy makes the provision of public goods more efficient, this should, in turn, also increase total factor productivity. In a principal–agent framework, citizens are the principals and can only very imperfectly control their agent, the government. In this situation, direct democratic institutions can have two effects. The direct effect enables the principals to override the decisions of an unfaithful agent. The indirect effect is where the simple threat of override is sufficient to compel the agent to behave according to the principal’s preferences. Potentially, reducing the principal–agent problem by way of direct democratic institutions could affect all of the economic variables discussed in this paper. If citizens prefer an expenditure level that is higher/lower than that preferred by the government, they should be able to achieve it via direct democratic institutions. It is often assumed that governments prefer higher expenditure levels than do citizens; in this case, we would expect expenditure levels to decrease with the increasing importance of direct democratic institutions in the country.

20

We follow the convention of the literature here and talk of referendums instead of referenda.

3.6 Representative vs. Direct Democracy

75

3.6.2.2 Empirics Empirical research into the effects of direct democracy can be divided into three phases. During the first phase, there was a prevalence of within-country studies, most of them dealing with the United States or Switzerland. Often, they relied on simple cross sections referring to US states or Swiss cantons. Endogeneity was often not a central concern. If there is one stylized fact emerging out of these studies, it is that direct democracy is associated with smaller government. One big question was, of course, whether the results obtained for the United States or Switzerland are valid beyond these two countries. In a second phase, a number of cross-country studies tried to deal with this shortcoming. Endogeneity was still an issue here, but due to lack of data, it was very difficult to deal with in these studies. The stylized fact emerging out of them was that the effects of direct democracy crucially hinge upon the specific kind being used (i.e., referendums or initiatives). While the use of mandatory budget referendums is associated with smaller government size, government size (both spending and revenues) was found to increase with the number of initiatives used. In the third phase of empirical research, scholars of direct democracy have returned to within-country studies. But this time, identification problems are taken a lot more seriously, resulting in studies relying on methods ranging from instrumental variables to differences-in-discontinuities. In these studies, some of the findings from the cross-country phase are confirmed. It is found that the use of initiatives does, indeed, cause government to become larger. Most of the available empirical studies focus on the effects of direct democracy on fiscal policy.

3.6.2.2.1 First-Generation Research: Focusing on the United States and Switzerland Matsusaka (1995, 2004) estimates the effects of the right to an initiative on fiscal policy among all US states except Alaska. States with the right to an initiative have lower expenditures and lower revenues than states without that institution. With regard to Switzerland, Feld and Kirchgässner (2001) deal with the effects of a mandatory fiscal referendum on the same variables. They find that in cantons with the mandatory referendum, both expenditure and revenue are lower by about 7–11% compared to cantons without mandatory referendums. Matsusaka (2004: chapter 4) also deals with the question of whether initiatives have any effect on the distribution of government spending between the state and the local level, and finds that initiative states spend 13% less per capita at the state level than noninitiative states, but spend 4% more on the local level. Proponents of direct democracy would interpret this finding as evidence in favor of the hypothesis that under direct democracy, government spending is more in line with citizen preferences. Feld et al. (2008) ask whether government spending really is

76

Positive Constitutional Economics

more in line with citizen preferences in a direct democracy and answer it by analyzing the Swiss case. Drawing on panel data for Swiss cantons from the years of 1980–98, they find that fiscal referendums at the cantonal level lead to less centralization of both cantonal spending and revenue. However, citizen fiscal preferences are not necessarily always more conservative than those of their representatives: Matsusaka (2000) finds that during the first half of the twentieth century, voters in US initiative states were frequently less conservative fiscally than their elected representatives. Funk and Gathmann (2013) identify preference heterogeneity among voters in different Swiss cantons as a significant variable explaining differences in cantonal spending. The next question we are interested in is whether direct democratic institutions have any effect on rent extraction, i.e., the perceived level of government corruption as well as the efficiency with which public goods are provided. With regard to US states, Alt and Lassen (2003) find that initiative states have significantly lower levels of perceived corruption than noninitiative states. Pommerehne (1983, 1990) deals with the effects of direct democracy on the efficiency with which government services are provided. More specifically, he finds that waste collection in Swiss towns having both a private contractor for the service and direct democratic elements is provided at the lowest cost. Some of the cost effectiveness is lost when waste collection is provided by the town itself, and additional efficiency losses materialize if waste collection is provided in towns without direct democratic elements. Blomberg et al. (2004) ask whether there is any significant difference in the effective provision of public capital between initiative and noninitiative states among the 48 continental US states during the period 1969–86. They find that noninitiative states are 20% less effective in providing public capital than are initiative states. Dalton (2008) looks into the possible effects of direct democracy on the quality of governance and asks whether the availability and the actual use of initiatives across US states between 1960 and 1998 is significantly correlated with a number of proxies for good governance, such as tax policies (reflecting an input into government performance), government management (reflecting the process), and education policies (as a government output). Controlling for a number of standard variables, he finds that the frequent use of initiatives seems to impede state government performance, rather than improve it. Finally, do direct democratic institutions have any discernible effect on productivity and thus on per capita income? Feld and Savioz (1997) find that per capita GDP in cantons with extended democracy rights is 5% higher than in cantons without such rights. Frey and co-authors argue that one should not look only at the outcomes that direct democratic institutions produce, but also at the political processes they induce (e.g., Frey and Stutzer 2006). Kirchgässner and Frey (1990: 63) speculate that the readiness of voters to incur information costs would, ceteris paribus, be higher in

3.6 Representative vs. Direct Democracy

77

democracies with direct democratic institutions because the voters participate more directly in the decisions. Smith and Tolbert (2004) investigate the effects of initiatives on voter turnout rates, civic engagement, and confidence in government. They compare initiative with noninitiative US states and find that the frequent use of initiatives has positive effects on all three aspects. Benz and Stutzer (2004) provide evidence in favor of the conjecture that citizens in states with direct democratic institutions are better informed than citizens in purely representative states. Some European states used referendums to pass the Maastricht Treaty, whereas others did not. Relying on Eurobarometer data, Benz and Stutzer find that citizens in countries with a referendum were indeed better informed both objectively (i.e., knowledge about the EU) and subjectively (i.e., feelings about how well they were informed). The paper is also interesting because it is one of the very few that deals with the effects of direct democratic institutions in a cross-country setting. Based on micro-data controlling for a host of relevant variables such as occupational status, marital status, and health conditions, Frey and Stutzer (2000) find that Swiss citizens who live in cantons with a high degree of direct democracy are happier. Lascher and Wassmer (2007) argue both with the logic underlying the conjecture that broader direct democratic rights could be conducive to happiness and with the empirical evidence. Using data from the General Social Survey for the US states and controlling for variables significant in previous studies, they conclude that the direct democracy proxies do not approach statistical significance in any of their models.

3.6.2.2.2 Second-Generation Research: Cross-Country Studies Blume et al. (2009a) is the first cross-country study to analyze the economic effects of direct democracy. It thus belongs to the second phase of empirical research into the effects of direct democracy. The findings only partially confirm previous results. They do find a significant influence of direct democratic institutions on fiscal policy variables and government efficiency, but no significant correlation between direct democratic institutions and productivity or happiness. Institutional detail matters a great deal; while mandatory referendums appear to constrain government spending, initiatives seem to increase it. The actual use of direct democratic institutions often has more significant effects than their potential use, implying that, contrary to what economists would expect, the direct effect of direct democratic institutions is more relevant than its indirect effect. It is also noteworthy that the effects are usually stronger in countries with weaker democracies (after recoding the Gastil Index such that lower numbers mean lower levels of democracies, Blume et al. create a dummy called “weak democracies” in which all countries with a Gastil-score of five and less are included).

78

Positive Constitutional Economics

In a more recent study, Voigt and Blume (2015) pose a follow-up question. If direct democracy has important effects on substantive outcome variables, does it also impact the political process? More specifically, are voters better informed, is turnout for representative elections affected, is civic engagement higher, and does direct democracy have any effects on the voters’ confidence in government? The answers they provide are rather sobering: voter turnout is not higher when direct democracy is available or used, citizens do not express greater interest in politics, and citizens express lower trust in both government and political parties. Now, these results need to be read with a grain of salt as the comparability of the procedural variables across countries can be questioned. On the other hand, all of these variables have already been used in other cross-country studies.

3.6.2.2.3 Third-Generation Research: Taking Endogeneity Seriously Many of the more recent papers dealing with the economic effects of democracy have returned to within-country analyses. But now, endogeneity concerns are taken a lot more seriously than during the first phase. For example, Funk and Gathmann (2011) focus on the effect of direct democracy on government size in Switzerland. They rely on data covering more than a hundred years (1890–2000) and investigate all Swiss cantons. This allows them to include canton-fixed effects to take care of the unobserved heterogeneity between cantons. To take care of the possibility that high spending levels cause the introduction of more direct democracy, i.e., reversed causality, they instrument their explanatory variables. Their first instrument is based on the degree of signature requirements for initiating change on the cantonal level. They find that the higher the degree, the less likely change is to occur. Further, they reason that the higher the number of neighboring countries already relying on mandatory budget referendums, the greater the likelihood that the canton under scrutiny will also introduce it. Hence, a kind of spatial diffusion argument. They find that mandatory budget referendums reduce expenditures on the cantonal level by 12%. Lowering the signature requirements for the voter initiative by 1% reduces spending by 0.6%. As so often, Swiss results do not necessarily carry over to other countries. Also relying on instrumental variables and covering 13,000 German municipalities, Asatryan (2016) finds that each initiative causes an average expansion of local government by around 8% on both the expenditure and the revenue side, thus confirming the results from the cross-country study summarized above. As with quite a few of the more recent studies, Asatryan (2016) relies on population thresholds for identification: The number of signatures needed to kick off an initiative is a function of the size of the municipality. Asatryan et al. (2016) make use of the introduction of the possibility of initiatives on the local level introduced in Bavaria in the mid-1990s. They use population thresholds to compare government size pre- and

3.6 Representative vs. Direct Democracy

79

post-reform (differences-in-discontinuities). They find that a decrease in signature requirements of one percentage point causes an increase in municipal expenditures of about 12–17%. Also using the expansion of direct democracy rights in Bavaria and relying on population thresholds, Arnold and Freier (2015) find that decreasing the signature requirement by one point increases the probability of an initiative by around 8–10 points. Also drawing on the Bavarian experience, Asatryan and de Witte (2015) find that a higher number of initiatives is associated with higher government efficiency. A sort of cottage industry could be emerging out of this with scholars analyzing the effects of the introduction of direct democracy in their country. Papers using particular experiences of Sweden (Hinnerich and Pettersson-Lidbom 2014) and Spain (Sanz 2017) are the first examples of this possible trend.

3.6.3 Determinants of Direct Democracy Why is Switzerland so strong in direct democracy institutions whereas neighboring Germany has barred any kind of direct democracy entirely from the national level? Although, or maybe because, this question appears puzzling, there have been few attempts to identify the determinants of direct democratic institutions. Decades ago, Lijphart (1984: 206) already admitted “defeat in the search for general propositions and theories” with regard to the presence or absence of the referendum institution. This is definitely not the right place to make an attempt at an encompassing theory, but it might be the right place to give some structure to the basic question. It was just shown that institutional detail is crucial. While referendums are associated with smaller government, initiatives are associated with larger government. It appears plausible that the determinants leading to the installation of these two institutions also differ. With (optional) referendums, government remains the agendasetter, whereas with initiatives, agenda-setting power is moved to the citizens. So it might be that the decisive driver behind referendums is the government, whereas the introduction of initiatives is more driven by the demand of citizens or citizen-groups. It was further shown that for many of the effects of direct democracy to materialize, the mere possibility of using either initiative or referendum is insufficient; they need to be actually used. It would, hence, be desirable to know more about the determinants that lead to the actual use of direct democracy instruments. Again, the reasons of government to call a referendum might be entirely different from the reasons of a group of citizens trying to kick off an initiative. Now, some of the more recent papers relying on instrumental variables in order to deal with endogeneity contain a number of suggestions. This, of course, does not satisfy the demand for a fully fledged theory, but contains information on some interesting associations. The most straightforward seems to be that the higher the cost of kicking off an initiative, the fewer initiatives are initiated. Blume et al. (2014) have looked at some of the covariates of direct democracy institutions. Again, this

80

Positive Constitutional Economics

does not make a theory less desirable but lets us learn something about the context in which we are more likely to encounter these institutions. They find, e.g., that democracy levels are positively associated with direct democracy.21 If countries in Sub-Saharan Africa, the Middle East and Southeast Asia, i.e., countries in regions characterized by rather low democracy scores, introduce direct democracy institutions, these are likely to be referendums rather than initiatives.

3.7 Electoral Systems Literally speaking, electoral systems are rarely included in constitutions. According to Elkins et al. (2009: 40), only every fifth constitution specifies the details according to which parliament is to be elected. But if we rely on the definition of the constitutions proposed in Chapter 1, then electoral systems do belong in such a primer, as they are conjectured to have important effects on how societies make decisions regarding the provision of public goods.

3.7.1 Effects of Electoral Systems 3.7.1.1 Theory A distinction is sometimes made between electoral rules and electoral systems. Electoral rules refer to the way votes translate into parliamentary seats. Under majority rule (also called plurality or first-past-the-post), only the candidate who secures the most votes in a district is elected, and all other votes are “lost” or “wasted.” Under proportional representation, parties are allocated seats according to the proportion of votes they obtain. Electoral systems include more dimensions than electoral rules, e.g., district size and ballot structure. Before issues of how to translate votes into seats become relevant, it is crucial to know who has the right to vote. In a sense, then, the extent of the franchise (discussed in the last chapter) is a central element of electoral systems. District size refers to the number of parliamentarians sent from one district. The ballot structure determines whether citizens can vote only for a party, only for an individual, or some combination thereof. Although theoretically distinct, these dimensions are highly correlated empirically. Countries using majority rule (MR) often have minimum district size (single-member districts) and allow voting for individual candidates. Countries relying on proportional representation (PR) often have large districts and restrict the possibility of deviating from party lists. 21

They rely on Polity IV as a measure for democracy. In generating their measure, Marshall et al. (2017) do not take direct democracy into account. The correlation between direct democracy and Polity IV is, therefore, not high by design.

3.7 Electoral Systems

81

It has been known for many years that electoral rules can have a crucial effect on the number of parties. Duverger’s (1954) observations that MR is conducive to twoparty systems, whereas more parties are apt to arise under PR, has even been called “Duverger’s law” or “Duverger’s hypothesis,” which is some measure of its general validity. We begin this section by rehearsing a number of hypotheses dealing with electoral rules, and then move on to deal with some hypotheses that refer to both district size and ballot structure. The analysis of the economic consequences of electoral systems is of a more recent vintage. Coalition governments are more likely under PR than under MR. Parties participating in the coalition will want to please different constituents, which explains why both government spending and tax rates are, on average, higher under PR than under MR (Austen-Smith 2000). Persson et al. (2007) refer to this as the common pool problem of fiscal policy, because no party in a coalition government internalizes the full costs of public spending. The composition of government spending under alternative electoral rules has been analyzed by Lizzeri and Persico (2001). They distinguish between the provision of a genuine public good, on the one hand, and the so-called pork-barrel projects that serve to redistribute wealth, on the other. They then ask whether incentives to provide these goods differ systematically between MR systems (which they term “winner-take-all systems”) and PR systems. In MR systems, politicians have an incentive to cater to those who can help them obtain a plurality of the votes, and they will do so by promising pork-barrel projects. In PR systems, on the other hand, targeting makes less sense because every vote counts, inducing politicians to provide more general public goods. The hypotheses presented so far all refer to the electoral rules. In their survey of the economic effects of electoral systems, PT (2003) also investigate district size and ballot structure. Suppose single-member districts are combined with plurality rule, which is often the case empirically. Then, a party needs only 25% of the national vote to win the elections (50% of half of the districts; Buchanan and Tullock 1962). Contrast this with a single national district that is combined with PR. Here, a party needs 50% of the national vote to win. PT (2000: chapter 9) argue that this gives parties under a PR system a strong incentive to offer general public goods, whereas parties under plurality rule have an incentive to focus on the swing states and promise policies that are specifically targeted at the constituents’ preferences.22

22

If there is population growth in some areas of a country (e.g., cities), whereas others stagnate or even shrink (e.g., rural areas), it is necessary to redraw district lines every now and then to ensure that each district is inhabited by approximately the same number of voters. This is generally known as redistricting. In the United States, it is referred to as gerrymandering. There, it is often used to manipulate a candidate’s probability of winning. Many countries have delegated the redistricting task to independent agencies, whereas in others (such as the United States), this is done by legislators themselves.

82

Positive Constitutional Economics

Also referring to district size, Milesi-Ferretti et al. (2002) obtain a similar result. In large districts, legislators will primarily represent socioeconomic groups, whereas in small districts, they will primarily represent geographical interests. Transfers are a suitable instrument for paying allegiance to social constituencies, whereas (local) public goods are better suited for paying for geographical allegiance. The authors assume that in majoritarian systems, just one representative is elected in each district, whereas in proportional systems, more than one representative is elected. Given this assumption, governments under proportional systems will spend more on transfers, whereas governments under majoritarian systems will spend more on (local) public goods. The effects of differences in ballot structure are the last aspect of electoral systems to be considered. Often, MR systems rely on individual candidates, whereas proportional systems rely on party lists. Party lists can be interpreted as a common pool, which means that individual candidates can be expected to invest less in their campaigns under PR than under MR. PT (2000: chapter 9) argue that corruption and political rents should be higher, the lower the ratio between individually elected legislators and legislators delegated by their parties. Bordignon et al. (2016) took a further step toward the explicit recognition of institutional detail with regard to electoral systems. In some MR systems, citizens are permitted to vote a second time, namely, to determine the winner among the two candidates who received the most votes in the first round (the most prominent example being French presidential elections). The authors ask whether the dual ballot impacts upon the number of parties (candidates) and their positions. They show that the number of parties is, indeed, larger under the dual ballot, but that the influence of more extreme positions is reduced under this institution.

3.7.1.2 Empirics PT (2003) were the first to put the conjectures just described to an empirical test for a large number of countries (up to 85) over a period of almost four decades (1960–98). They find a significant association between the electoral system and a number of economic variables. (1) In majoritarian systems, central government expenditures are 3% of GDP lower than in PR systems. (2) Expenditures for social services (“the welfare state”) are 2–3% lower in majoritarian systems. (3) The budget deficit in majoritarian systems is 1–2% below that of systems with PR. (4) A higher proportion of individually elected candidates is indeed associated with lower levels of (perceived) corruption.

3.7 Electoral Systems

83

(5) Countries with smaller electoral districts tend to have more corruption. (6) A larger proportion of individually elected candidates is correlated with higher output per worker. (7) Countries with smaller electoral districts tend to have lower output per worker. In a follow-up paper, Persson et al. (2007) take the results on government spending as a starting point to inquire into the specific channels through which the results are brought about. The authors propose to distinguish a direct from an indirect channel. The direct channel works via the incentives that electoral rules have on politicians, whereas the indirect channel is due to differences in party structures and the ensuing types of government (i.e., coalition vs. non-coalition). They find the indirect channel to be more important. PT use instrumental variables to explicitly account for the fact that institutions are never truly exogenous, but their instruments (time period during which the constitution was written, distance of the country from the equator, percentage of population speaking English or another European language as their mother tongue, ethno-linguistic fractionalization, and population size) display a number of weaknesses, as discussed in detail by Acemoglu (2005). Blume et al. (2009b) replicate and extend PT’s analysis, and find that district magnitude and the proportion of individually elected candidates, both substantially and statistically, have a more significant impact on various outcome variables than the electoral rule itself. More attention, hence, should be paid to institutional details. As discussed in some detail in Chapter 1, cross-country studies trying to identify some constitutional rule as causing specific outcomes face important methodological challenges. These include, but are not limited to, the fact that countries vary among many dimensions (and differences in outcomes might well be caused by factors other than the variable of interest; in addition, these factors might even be unobservable), that constitutional rules tend to be stable over time (such that exogenously caused changes cannot be used to identify changes in the outcome variable), and that the transmission channels through which a constitutional rule could affect an outcome variable often remain obscure. In their paper on how changes in electoral rules affect fiscal policy, Funk and Gathmann (2013) take up all three challenges. By focusing on a single country (Switzerland), unobserved differences between countries are excluded as a potential problem. This has a price, though, since it is unclear whether findings that hold in Switzerland also hold elsewhere. Still, the case of Switzerland is very interesting because of the numerous changes made to the electoral rule during the period covered in their analysis. For the first year of their study (1890), all cantons relied on MR. In the last year of their study (2000), all but two cantons had switched to PR. Finally, they also look into possible transmission channels.23

23

Alesina and Glaeser (2004: chapter five) show how proportional representation was introduced into virtually all European countries beginning in the late nineteenth century but was vehemently

84

Positive Constitutional Economics

Funk and Gathmann (2013) are not only interested in overall changes in both expenditures and revenues on the cantonal level, but in particular in the issue whether a switch from MR to PR leads to changes in the composition of government spending. According to theory, we would expect such a switch to cause more spending on genuine public goods and less on targeted public goods that profit only small groups. Funk and Gathmann draw on expenditures for education and welfare as proxies for genuine public goods, whereas expenditures for roads and subsidies for agriculture serve as proxies for geographically targeted public goods. They find that after adopting PR, cantons spend 21% more on education and 42% more on welfare. This is, hence, very strong evidence in favor of the hypothesis that a switch to PR increases spending on genuine public goods. The other effects are less clear-cut. In particular, the transition to PR in Switzerland did not lead to an increase in overall government expenditure or revenue. Probably their most innovative contribution is to inquire into the channels through which PR affects spending for genuine public goods. They ask whether this was driven by more left-wing parties being present in cantonal parliaments, by a higher number of parties being represented within the legislature, or by higher electoral turnout (which could imply a shift in the position of the median voter). They produce their answer in two steps. First, they ask whether the three potential channels just mentioned were actually observed (they did occur) and then, to what degree these channels are associated with higher spending on genuine public goods. Together, these transmission channels explain up to 40% of the changes. A paper by Gagliarducci et al. (2011) can almost be read as a companion paper to the one just summarized. The authors are not interested in some aggregate outcome variable, but in the behavior of individual legislators. This study is possible due to the existence of a peculiarity of the Italian electoral system between 1994 and 2006. During this period, three quarters of Italian MPs were elected by MR in singlemember districts and the remaining quarter was elected under PR. Gagliarducci and his co-authors are interested in two questions. First, are parliamentarians elected under MR significantly more likely than parliamentarians elected under PR to propose bills that are targeted at the region in which they are elected? Second, are they less likely to cash in on rents (because they are more directly accountable to their voters and cannot hide in some common pool as their PR colleagues)? As a proxy for rents, the authors use the absenteeism rate based on the notion that people who are absent will be able to earn higher outside incomes. Both conjectures are confirmed by the data. The model by Milesi-Ferretti et al. (2002) (introduced above) predicts that in large districts, legislators will primarily represent socioeconomic groups, whereas in small opposed by the white majority in the United States that was afraid of African Americans ensuring more seats under it than under first past the post.

3.7 Electoral Systems

85

districts, they will primarily represent geographical interests. They test their model with 20 OECD and 20 Latin American countries and find that, as predicted, transfers are higher under PR. Going beyond the simple dichotomy between majoritarian and proportional systems, they find that higher degrees of proportionality are correlated with more transfer spending (as opposed to public goods spending). The conjecture that larger districts cause representatives to vote not in line with the preferences of the citizens they are supposed to represent has been challenged by Stadelmann et al. (2014). Their argument draws on Condorcet’s jury theorem (1785), which shows that larger juries are more likely to make the correct decision given that the likelihood of an individual jury member making the correct decision is larger than one-half. They agree that an individual representative may be less likely to vote along the lines of the majority the larger the district, but they argue that the majority of representatives are more likely to follow the party line the larger the district. They test their argument by looking at 126 referendums that took place in Switzerland between 1996 and 2010. The referendum results are used as a proxy for the preferences of the citizen voters, and the available roll call votes are used to analyze the voting behavior of the representatives. It turns out that the second effect (the votes of the majority of representatives) far outweighs the first effect (the votes of an individual representative). As of now, it is unclear to what degree this result can be generalized. The existence of a referendum could, for instance, signal to the representatives that their vote will be scrutinized more heavily than everyday parliamentary decisions. Iversen and Soskice (2006) deal with the effects of electoral systems on redistribution. They question the Meltzer–Richard (1981) model, according to which nonegalitarian societies redistribute more than egalitarian ones. They notice that three out of four governments under majoritarian systems were center-right between 1945 and 1998, whereas three out of four governments were center-left under PR. However, if that is the case, a closer look at the transmission mechanism between electoral systems and government expenditures is needed because it is unclear if the difference is due to the constitutional rule, or due to the different government ideologies. Picking up on Acemoglu and Robinson (2000; see below), they argue that PR can be interpreted as a commitment device in favor of redistribution (which they interpret as a kind of insurance against loss of income in case of unemployment, which would, in turn, increase incentives to make specific investments in human capital).

3.7.2 Determinants of Electoral Systems If the choice of electoral systems has important economic consequences, it seems only natural to ask what factors determine this choice. And indeed, there have been

86

Positive Constitutional Economics

heated discussions (mostly among political scientists) on the determinants of electoral systems.24 Colomer (2005) argues that it is the number of parties that determines the electoral system and not vice versa. In other words, he turns Duverger’s law upside down. He assumes that “the presumed line of causality is double.” His empirical test covers 87 countries and 219 elections. He observes 37 changes from MR to PR. He states (Colomer 2005: 12) that “multi-party systems occur before and not only subsequent to the adoption of PR.” He also finds support for a long-term trend toward PR, a finding previously made by Lijphart (1994: 52–6). Boix (1999) asks under what circumstances incumbent parties will change electoral systems. He starts from the assumption that incumbents are interested in maximizing their parliamentary representation (as well as government positions). The emergence of new parties is modeled as an exogenous shock that can, e.g., be caused by a move toward universal suffrage. If the newly emerging parties (usually leftist) are perceived as weak, the electoral system remains unchanged. If, on the other hand, the new parties are perceived as strong, electoral systems change from majoritarian to proportional rule. Boix tests this hypothesis for the electoral systems of 22 countries between 1875 and 1990 and finds it broadly confirmed. Further, he shows that the consequences of a country’s internal fragmentation (measured by ethnic as well as religious fragmentation) depend on the size of the country; in small countries, high levels of fragmentation are one factor leading to PR. Large countries can be highly fragmented at the national level, yet quite homogenous at the local or regional level. If this is the case, federalism serves as a substitute for PR.25 To model the emergence of new parties as an exogenous shock, even though it is the consequence of a move toward universal suffrage and, hence, an endogenous choice, is somewhat unconvincing. Cusack et al. (2007: 373) attack this argument as “wrong in terms of the logic, the interpretation of the historical record and the cross-national evidence.” They endeavor to show that “PR cannot be the equilibrium choice of right parties under the assumptions of the model.” Instead, they divide Europe and North America into two types of countries, namely “proto-corporatist” countries in which negotiations between a higher number of interest groups and parties occur before any policy choices are made, on the one hand, and “liberal” countries, on the other. For the first type of country, majority rule was not suitable any more for a number of reasons, and that is why they changed to PR around the turn of the last century. The liberal 24

25

Benoit (2007) is a very accessible survey of the various arguments made about endogenous electoral systems, up to around 2005. Aghion et al. (2005) endogenize the choice of electoral rules in US cities in a very similar fashion. Here, the external shock is the general extension of the franchise to minorities. The authors show that small minorities will induce the majority to keep first past the post systems, whereas large minorities will induce the majority to switch to proportional representation.

3.7 Electoral Systems

87

countries stayed with MR and, hence, did not experience any change in the electoral system. In a sense, then, this whole discussion revolves around a chicken–egg problem: is it electoral systems that determine party structure or is it the party structure determining electoral systems? The Cusack et al. (2007) paper was challenged by Kreuzer (2010), but Cusack et al. (2010) contributed a heated reply offering additional evidence supporting their position. Benoit’s (2004) intention is to develop a rational choice model of electoral system change, which will, according to the model, occur if a coalition of parties not only expects to be better off under alternative rules but also has the power to modify the currently valid rules. However, his model comprises so many potentially relevant determinants that an empirical test of it is extremely difficult. Ticchi and Vindigni (2010) are interested in an even broader question, namely, the factors determining the choice between “majoritarian” and “consensual” constitutions (the dichotomy having been introduced by Lijphart 1999). “Majoritarian” constitutions are characterized not only by having a plurality rule, but by a number of other characteristics, inter alia, that government is dominated by the executive and that governments are usually one-party governments. “Consensual” constitutions, on the other hand, are characterized not only by having PR but also by the fact that the legislative and executive powers are more balanced and that governments are usually coalition, i.e., multi-party governments. Ticchi and Vindigni hypothesize that the factor driving the choice between majoritarian and consensual systems is the ex ante degree of income inequality: if it is relatively high, a majoritarian constitution is more likely; if it is relatively low, a consensual constitution is more likely. The German state of Hesse changed the electoral law applicable to municipal elections in 1999. The 5% threshold that lets only those parties get a seat on the municipal council that have received at least 5% of all votes was abolished, and the possibility to cumulate votes for a single candidate, as well as the possibility to vote in favor of candidates running on different party lists, was introduced. The declared aim of these measures was to increase the competitiveness of local elections. The law allows different council sizes depending on the population size of the community. Representatives of established larger parties might not only be happy by the prospects of enhanced competition. One instrument at their disposal to make life hard for small and (or) new parties is to reduce council size: The smaller the council size, the lower the likelihood for parties to secure a seat thereon. The larger the council size, the higher the additional competition due to small parties. Baskaran and da Fonseca (2016) run an innovative difference-in-discontinuities study by using the fact that differently sized communities are differently treated by the reform. They find that the higher the intensity by which communities have been treated in this way, the higher the likelihood that they reduced council size. This is yet another study that nicely documents that politicians react in a predictable way to the incentives entailed in the electoral system they are subject to.

88

Positive Constitutional Economics

In sum, the estimated size of the effects of differences in electoral systems on a number of economic variables is often very significant. It is noteworthy that institutional detail matters a great deal. When electoral systems are discussed, the electoral rule is often given pride of place, but the empirical results show that district magnitude must not be overlooked. As soon as electoral institutions are not assumed to be exogenously given, but are analyzed as the result of political power games, it becomes less clear whether the connection between electoral rules and economic outcomes is causal or primarily a correlation. First results regarding transmission channels are interesting and encouraging for future research, which is definitely needed.

3.8 Summary Quite a few constitutional rules appear to have important economic consequences, including electoral rules (proportional vs. majority rule), form of government (presidential vs. parliamentary), and whether a country has a federal structure or direct democratic institutions. These results clearly indicate that endogenizing these institutions is desirable. Yet, it has been argued (Acemoglu 2005) that these results might merely be strong correlations rather than causal effects. It might be (omitted) third variables that determine both institutions and policies. However, even if this argument is correct, it is still necessary to identify the conditions that give rise to different institutions as doing so will also clarify whether these initial conditions have a direct impact on policies.

4

Conclusions and Possible Future Issues

4.1 Introductory Remarks In the Introduction and the first three chapters of this primer, we offered an extended survey of constitutional political economy. We summarized the major research questions, discussed the methods used in a number of studies, and presented the results regarding the effects of constitutional rules as well as their determinants. This chapter is different. After a brief conclusion, I move on to deal with questions that are underresearched and deserve to be analyzed in depth. This is, of course, a subjective endeavor and other scholars in this area might come up with a completely different list. Section 4.3 discusses the relevance of procedural rules, both for producing entirely new constitutions, as well as for amending existing ones. With regard to these issues, we have a number of case studies and comparative work, but precious little in terms of general insights. In Section 4.4, I deal with what I call the de jure de facto gap, i.e., the observation that many constitutions promise many things but fail to deliver time and again. There is evidence that the effects of constitutions are mainly driven by the way the constitutional rules are actually implemented. If we are interested in certain effects of constitutional rules, it is, hence, crucial to understand under what conditions constitutional rules are likely to be implemented – and under what conditions constitutional rules will simply be a nice prose. The final section of this chapter flags the importance of emergency constitutions, i.e., those constitutional rules dealing with emergency provisions. It looks into possible determinants of emergency provisions but is mainly concerned with their effects. Although used frequently, we know extremely little about emergency constitutions. These three topics are closely related, as they all deal with constitutions in difficult times. More precisely speaking, they deal with situations when entirely new constitutions are written (Section 4.3), when some actors prefer to ignore them (Section 4.4), or when their survival is at stake (Section 4.5). Coined differently, all sections in this chapter deal with moments crucial for the creation and (or) the maintenance of the rule of law.

4.2 Conclusions Constitutional political economy has come a long way since the Calculus of Consent. Over the last couple of years, the positive branch has made remarkable progress,

90

Conclusions and Possible Future Issues

thanks to the availability of sizable data sets and the development of a host of indicators built thereon. This has made it possible to subject constitutional issues to econometric analyses that would have been impossible just a couple of decades ago. Some scholars report very strong economic effects of constitutional rules; however, upon closer examination, many of these effects are not as robust as initially assumed. It has become evident that institutional detail matters a great deal when investigating any number of constitutional rules. Therefore, in the years to come, explicit recognition of institutional details should be high on the agenda. Many consider describing and explaining the choice of constitutional rules as the core task of constitutional political economy. If this course can be agreed upon, there is much work to be done as only a few studies endogenizing single constitutional rules or even entire constitutions have been conducted. As we have seen, endogenizing constitutions is truly challenging. Constitutions are written under very special circumstances that do not lend themselves to generalization.1 Furthermore, although constitutions are rewritten far more frequently than often assumed, the overall number of observations is still small. The recent progress in positive constitutional economics has been achieved primarily by drawing on cross-country studies. In Chapter 1, we saw that a discussion about the most adequate tools for cross-country analysis is taking place. In empirical economics, a so-called credibility revolution (Angrist and Pischke 2010) has taken place. To be credible when making causal claims today, convincing identification strategies are key, ideally based on natural experiments. But these are very difficult to find in constitutional economics. Moreover, cross-country studies as such have come under attack. The main methodological concerns seem to be simultaneity, reverse causality, and omitted variable bias. But behind these technical concerns hides the implicit – and problematic – assumption that one model is sufficient to “explain” various developments in vastly different environments and development stages (see Rodrik’s 2008 plea for second-best institutions). Ever more frequently, scholars seem to recommend case studies and yet there are no commonly agreed upon criteria for judging their quality. Whether case studies on constitution-making can lead to general insights remains to be seen. Academic papers often conclude with the policy implications of their results, but in the case of this primer, it is questionable whether policy advice, sound or otherwise, can even be derived from the many and various results discussed here. I believe that we should be extremely careful in jumping to conclusions. The effects of many

1

See Hume’s essay “Of the Original Contract” (1777, 1987: 474): “and were one to choose a period of time, when the people’s consent was the least regarded in public transactions, it would be precisely on the establishment of a new government. In a settled constitution, their inclinations are often consulted; but during the fury of revolutions, conquests, and public convulsions, military force or political craft usually decides the controversy.”

4.2 Conclusions

91

constitutional rules were estimated in isolation, but in the real world, effects will be the result of many interactions. Remember the results of Persson and Tabellini reported in Chapter 3. They find that governments in countries with a majority rule electoral system and a presidential form of government spend some 10% of GDP less than governments in countries with proportional representation and a parliamentary form of government. But it is, of course, questionable whether government spending would decrease by 10% of GDP if parliamentary PR systems switched to a combination of MR cum presidentialism. It should make us cautious that it is so difficult to show that democracies have a systematic advantage over autocracies in terms of growth rates as discussed in Chapter 2. It should also make us wonder how it is possible that a number of constitutional rules have significant effects on fiscal policies and a number of governance indicators, but fail to have a significant impact on either labor or total factor productivity. So, at the end of the day (and this primer), we come smack up against the question that is at the very core of constitutional economics: Do constitutions really matter? Can societies really improve their lot by agreeing on a specific set of rules? In answering this question, it is not sufficient to point to some empirical evidence that seems to indicate that the answer is yes. We need to be able to explain the difference between de jure rules and their de facto enforcement: What are the factors that change a few printed pages into a document around which collective decisionmaking is organized? What are the factors that prevent the factual enforcement of constitutional rules? It might well be that the enforcement of some constitutional rules depends on the existence of “preconstitutional” rules. These could be the informal or internal institutions of a society. If such are fairly stable over time and of the sort unconducive to the implementation of (desirable) constitutional rules, then the factual effect of constitutions could be quite limited. Much of Constitutional Political Economy (CPE) research implicitly assumes the nation-state to be the natural form of government, even though supranational and international agreements are becoming ever more important. There are various possibilities for incorporating these agreements into CPE. One example is to view them as a new level in a multilevel governance structure, as has been done by some social scientists as well as legal scholars. One can then analyze under what conditions politicians of the nation-state level are ready to delegate competence to a supranational or international level. (Voigt and Salzberger (2002) made the first attempt at analyzing delegation decisions from a CPE point of view, conceptualizing delegation to national bureaus and to international organizations as an alternative.) This positive approach needs to be complemented by some normative analysis, however, and to date, in the discussions over Europe’s constitutional future, contributions by economists have been rather scarce. The European Union is only the most obvious supranational development. Analyzing international organizations such as

92

Conclusions and Possible Future Issues

the World Financial Organizations or the United Nations from a CPE perspective should also be highly relevant.

4.3 Rules for Choosing and Amending Constitutions 4.3.1 Introductory Remarks The third chapter of this primer contains a number of examples showing that constitutional content does matter in the sense of causing different outcomes regarding fiscal policies, the quality of governance, and also overall productivity. Because content does matter, studies endogenizing single constitutional rules by trying to identify possible drivers of their implementation into the constitution were reported. This section also deals with the endogeneity of constitutions, but from a different angle. Assuming preferences of those who have the task of preparing a new constitution as given, I ask whether different procedures used to generate a new constitution have a significant impact on its content. More than 30 years ago, Elster (1993: 174) deplored the fact that there was not one single paper that described the process of constitution-making from a general perspective. There are, however, quite a few edited volumes collecting a large number of case studies. Within a framework of questions developed by the editors, participants in the constitution-making process of nine countries share their experiences in a study by Goldwin and Kaufman (1988). Ludwikowski (1996) made a heroic attempt to describe the constitution-making processes in Central and Eastern Europe between 1990 and 1995. The volume edited by Miller (2010) is probably the most comprehensive collection of case studies to date, containing studies about the constitutionmaking of 19 countries all over the world. The rules used to aggregate individual preferences into draft constitutions seem to be of paramount importance in this regard, but there are many other aspects. The following list is inspired by Elster (2012) who deals with all aspects from a normative point of view, trying to establish something like an optimal design. In this chapter, these aspects are taken up by asking how different ways of dealing with them might impact the content of constitutions. Here are a number of questions: (1) Does it matter how the members of the constitutional assembly are chosen? Do they have the exclusive task of preparing a draft constitution or do they also act as legislators? (2) Does the size of the constitutional assembly impact the content of the constitution? (3) Does it matter whether the constitutional assembly is subject to time limits? (4) Does it matter whether deliberations of the assembly are public or secret? (5) What are the consequences of different voting rules?

4.3 Rules for Choosing and Amending Constitutions

93

(6) Does involvement of the public in the process of constitution-making matter for its content? (7) Does it matter how the constitution is ratified? Elster (1995: 373) suggests that constitution-makers are subject to both upstream and downstream constraints. While upstream constraints are imposed on the assembly before it starts meeting, hurdles that need to be jumped to have the draft ratified are called downstream constraints. According to this delineation, Questions 1–3 are definitely upstream constraints. Questions 4–6 might be upstream constraints if the assembly is not given the competence to answer these questions on its own. Question 7 qualifies as a downstream constraint. This section is divided into two major sub sections. The first (Section 4.3.2) deals with various relevant aspects for the choice of entirely new constitutions, whereas the second (Section 4.3.3) deals with procedural rules underlying constitutional change, i.e., constitutional amendments. Due to the lack of either elaborate theories or systematic empirical insights, we discuss theoretical conjectures and empirical cases together here. Empirical evidence is particularly difficult to come by, since every instance of constitution-making is, in a sense, unique. In attempts to answer the questions, the main focus will be on the content of the constitution. There are other potentially interesting outcome variables such as the endurance of constitutions, their success, legitimacy, and so on. These are, however, only dealt with in passing. In this chapter, the answers to the seven questions just posed are assumed to be already given; the primary interest here is in inquiring into their effects on constitutional content. Of course, one could go one step further and inquire into the factors determining the answer to these questions. This would amount to something like an “endogeneity chain.” The first element of that chain endogenizes constitutional rules. The second element endogenizes the rules used to create constitutional rules. It is this second element that we are concerned with here.

4.3.2 Rules for Choosing Constitutions 4.3.2.1 Choice of Members At least three different options for choosing the members of a constitutional convention can be thought of (Mueller 1996: chapter 21). They could be selected at random, which would likely lead to impartial constitution-makers. Mueller also conjectures that this way of selecting them is likely to increase the loyalty of the entire population to the document produced. As a downside, most members are unlikely to be constitutional experts. Second, members to the constitutional assembly could be chosen in an election separate from the election to the legislature. According to Mueller, this would increase the likelihood that individuals elected as

94

Conclusions and Possible Future Issues

members of the assembly would enjoy a high reputation, but not have held political office. If it is clear from the outset that they will not move on to be parliamentarians after the constitution has been created, this might induce them to insist on rules that will make the provision of genuine public goods more likely – and, in turn, the provision of pork barrel goods less so. Third, parliament can simultaneously serve as the constitutional assembly. As pros of this solution, Mueller mentions that it would save time and that the constituent legislators can be held accountable in the next parliamentary election. In other words, a one-shot game is transformed into a repeated one. Further, such election is likely to lead to shorter constitutions, as the members of the assembly might leave more topics to the legislature.2 But constitution-makers-cum-legislators are subject to a time constraint. They have to decide whether to engage in constitution writing or in passing new legislation. Mueller is afraid that the short-term perspective (passing new legislation) might win over the long-term one (preparing a draft constitution). Also, delegates are more likely to vote for a combination of their own personal interests and those of their constituents. A fourth option not mentioned by Mueller comes to mind; in many real-world constitution-making processes, the members of the constitutional assembly are hand-picked by the dictator who wishes to have a novel constitution written. Empirically, the first two options are relatively rare, with some form of the third option being the most frequently chosen (Ginsburg et al. 2009). If there is a constitutional convention that does not simultaneously serve as legislature, it often consists of members who are appointed by the executive or the legislative branch, or a combination of the two. According to public choice theory, one would expect constitutional conventions to be dominated by members who are appointed by the executive to produce constitutional rules allocating relatively more power to the executive. Analogously, one would expect conventions dominated by members who are appointed by the legislative branch to allocate relatively more power to the legislature. In their empirical study, Ginsburg et al. (2009) find that assemblies dominated by the legislature do not necessarily implement strong legislative powers, but the assemblies dominated by the executive do tend to assign more powers to the executive. They, hence, find a curious asymmetry and the public choice prediction only partially backed by the data.

4.3.2.2 Size of Assembly In his classic study on constitution-making, Elster (2000) compared the eighteenthcentury constitution-making of the federal US Constitution with that of 2

The factors determining the length of constitutions (simply measured by the number of words) have been analyzed in Voigt (2009). The two most significant determinants are having a British colonial history (associated with significantly longer constitutions), and being a Muslim dominated country (associated with significantly shorter constitutions).

4.3 Rules for Choosing and Amending Constitutions

95

revolutionary France and noted that the Federal Convention had 55 members, whereas the French Assemblé Constituante had 1,200 members. What are the expected consequences of differences in size for the resulting documents? The smaller the constitutional assembly, the easier it is to vote strategically. Also, logrolling, i.e., mutual vote trading, is easier with small assemblies. Thus, small assemblies are likely to privilege a number of groups to the detriment of those groups who are not part of any logrolling deal.3 On the other hand, the incentive to become informed and be well-prepared for the deliberations are likely to decrease in the size of the assembly, as individual representatives are less likely to be the pivotal members of the assembly. This issue can also be framed as a common pool problem. Finally, the internal consistency of the draft constitution might suffer if it is produced by a large number of delegates. I am not aware of any systematic analysis of the effect of assembly size on constitutional content. However, there are evaluations referring to specific assemblies. Bannon (2007) believes that the internal consistency of the Kenyan constitution suffered as a consequence of the large size (600 members and 2 chambers) of its constitutional assembly.

4.3.2.3 Time Limits Constitutions are often written in times of “revolutions, conquests, and public convulsions” as David Hume (1777/1987: 474) observed. In other words, speed is often of the essence when constitutions are drafted and time limits given to the constitutional assembly might appear intuitively appealing. But do they make sense? Elster (2012) is skeptical that time limits have much of an effect, except if the aim is to keep the constitutional deliberations and intermediate results secret. In that case, a time limit might be helpful. Basic insights from bargaining theory imply that time limits improve the bargaining position of the least patient actors (because they put a cap on patience and reduce the advantages that patience might otherwise convey), and are likely to have an impact on the content of the constitution. But why should they be complied with in the first place? Who would enforce the time limit and how? And what would be the consequence if a constitutional convention did not come up with a draft constitution by the date set? The constitutional assembly of Nepal is a recent example of a constitutional assembly that was subjected to time limits. After a massacre within the royal family of Nepal, the monarchy was terminated in 2008 and an explicitly elected constitutional assembly tried to produce a republican constitution. It failed to meet any of

3

The likelihood of being included in logrolling deals also depends on the voting rule used by the constitutional assembly, which is the topic of Section 4.3.2.5.

96

Conclusions and Possible Future Issues

various deadlines, and the Supreme Court of Nepal finally decided that further extensions were unconstitutional. A second constitutional assembly was elected and began its work in 2013. It also failed to deliver within the time limit originally declared, but a new constitution was finally promulgated in September 2015. Ginsburg et al. (2009) provide a number of observations regarding the time needed to create a new constitution, i.e. not necessarily under time constraints. While it took US representatives only around a week to draft Japan’s 1946 constitution, it took Myanmar 17 years to create a new constitution.

4.3.2.4 Public or Secret Meetings Given that the purpose of a constitution is to facilitate the coordination of all citizens of a country, it seems almost self-evident that the process by which such a constitution is generated should be open to the public. This presumption is also in line with many arguments regarding high-quality governance, which would be easier to achieve if there was a high degree of transparency. But here, we are not interested in normative argument, but the question whether public (or secret) deliberations can impact the content of the constitution. Powerful arguments that this can indeed be the case have been raised. If all positions ever taken by all members of the constitutional assembly become public knowledge, this makes it more difficult for assembly members ever to change their mind. In other words, achieving compromise and consensus on a draft could become more difficult if deliberations are public. I am not aware of any cross-country study analyzing the effects of public – as opposed to secret – deliberations. Elster (2000) claims that publicity explains some of the failures of the post-1789 French constitutional assembly.

4.3.2.5 Voting Rules Public choice scholars are well aware of the difficulties of aggregating individual preferences into collective outcomes (Arrow 1951). Introducing their interdependence cost calculus, Buchanan and Tullock (1962) show that more inclusive decisionmaking rules are connected with higher decision-making costs, but have the advantage of lower external costs. Taking social contract theory as a starting point, unanimous consent on a limited number of basic questions seems normatively desirable. But many more questions need to be dealt with: Who has the competence to put proposals on the agenda of the assembly? Often, constitutional assemblies rely on committees that are responsible for specific topics. Should they be allocated a monopoly in the power to make proposals? If competing proposals regarding specific

4.3 Rules for Choosing and Amending Constitutions

97

topics are made, in what order should they be voted upon? These are only a few of many questions that could be raised.4 The issue of whether deliberations should be secret or public was just discussed. This leads directly to another aspect regarding voting: Should the constitutional assembly delegates’ voting behavior be made public? Keeping it secret makes logrolling significantly more difficult. In principle, secret deliberations are not systematically related to secret voting behavior. If debates are public, it might be useful to make voting secret to mitigate the unwanted effects of public deliberations just discussed above.

4.3.2.6 Public Participation There are a number of possibilities to involve the public in the process of constitution-making. The two lateral options are the most obvious. The public is encouraged to send statements regarding the desired content of the constitution before the constitutional assembly first meets. At best, these statements then serve as inputs for the drafting process. Such involvement has been observed various times, e.g. by the cahiers de doléances in revolutionary France, but also in the 1936 constitution of the Soviet Union or the 1996 constitution of South Africa. Elster (2012) claims that there is no evidence showing that these expressions had any impact on the document finally passed. The other lateral option is to consult the public via referendum after the assembly has finished its work. If the members of the assembly are aware that their draft needs to be welcomed by a majority of voters, they have an incentive to take the presumed preferences of their citizens into account while preparing their draft. Michel and Cofone (2017) have recently analyzed constitutional referendums and focused on the applicable majority rule. They find that the current default of a simple majority rule in the referendum is without theoretical justification. They go on to highlight the circumstances under which a qualified majority rule in the referendum is superior. In a nutshell, qualified majority rules are particularly useful when the legitimacy of the new constitution is of utmost importance, and in situations when additional checks on the drafters are needed. Eisenstadt et al. (2015) have inquired into the effects of participation on the likelihood of a country becoming more democratic. They find that increased participation in constitution-writing positively impacts levels of democracy after the 4

Riker (1984) proposed a new label for analyzing dynamic decision making processes, namely heresthetics or the art of political strategy. By using heresthetics, actors try to structure the situation in such a way that others will readily accept it. Riker names several categories of heresthetics: strategic voting (the avoidance of “wasted” votes, the creation of a voting cycle, vote trading) and agenda manipulation (arranging the sequence of decisions in a certain manner, introducing new voting alternatives) (1983: 63f.).

98

Conclusions and Possible Future Issues

constitution has been promulgated. Simultaneously, their findings shed some doubts on the lasting effects of using referendums to ratify constitutional drafts (on levels of democracy). Their study is highly welcome but also subject to critique. Reading a causal relationship into their findings seems far-fetched as countries allowing a high (low) degree of citizen participation in constitution-making are not randomly chosen. Of course, citizen participation is chosen because some members of the elite preferred it over the alternatives. One should, hence, be careful not to read any causality into these results. Further, their measure of democratic content is debatable, to say the least. It takes into account whether the head of state has decree power, whether the constitution places restrictions on the right to vote, and whether the constitution provides for a human rights commission (Eisenstadt et al. 2015: 604). These dimensions are, however, not the first that come to mind when trying to ascertain the degree of democracy.

4.3.3 Procedural Rules for Changing Rules Constitutional economics is not only interested in identifying the factors that determine the content of the constitution at the time it is first promulgated, but also those factors that lead to constitutional change later on. In Chapters 2 and 3 of this primer, some potential fundamental causes for constitutional change were dealt with. In Chapter 2, the most relevant theories dealing with these causes are summarized. But beyond those basic factors, procedural rules might impact the kind, as well as the frequency and extent of constitutional change. To be as clear as possible regarding the term “constitutional change,” I propose to explicitly recognize two dimensions, namely, the legality and formality of constitutional change. Regarding the first dimension, constitutional change can occur within the amendment rules (and thus be legal), or it can ignore the rules (and thus be illegal). Regarding the second dimension, explicit change occurs when the text of the constitution is modified; implicit change occurs over time as the interpretation of the constitution evolves. The taxonomy can thus be put into a simple matrix (Table 4.1).Voigt (1997) surveys the literature connected with each of the four cells. Rasch and Congleton (2006) pick up that taxonomy and deal exclusively with the first cell. They note that almost all constitutions specify procedures for their amendment. A straightforward hypothesis is that the more stringent the amending procedures, the fewer (explicit) constitutional changes will occur. Lutz (1994) appears to confirm this hypothesis. Ferejohn (1997) disaggregates Lutz’s “difficulty-of-ratification index” and finds that the requirement of special majorities or separate majorities in different organs of government are key to explaining amendment rates. Rasch and Congleton (2006) use the number of veto players relevant for bringing about constitutional change as an alternative indicator and find that this number has systematic

99

4.4 Mind the Gap

Table 4.1 Conceptualizing constitutional change Legality

Legal (constitutional)

Illegal (unconstitutional)

Formality Explicit Implicit

1 2

3 4

effects on the amendment rates in 19 OECD countries. The authors themselves point out their amendment counts might be counting “both apples and oranges”: minor amendments and major changes are all counted in an identical fashion. More recently, Ginsburg and Melton (2015) have argued that all of the previous measures have flaws, and that their validity seems dubious because correlation among them is very low. They propose to draw on a society’s amendment culture as an important determinant of constitutional change. Unfortunately, this is equivalent to abandoning the search for determinants of constitutional amendment as long as a society’s amendment culture cannot be determined independently of the observed number of constitutional changes.5 A desideratum concerning future research is to deal with the question of whether more stringent amendment rules lead not only to less legal and explicit constitutional change, but also to more implicit and/or illegal constitutional change (i.e., change depicted by Cells 2 through 4 in the matrix). If it costs little to change the constitution, many such changes are to be expected, which will impact the uncertainty-reducing function of the constitution. On the other hand, if legal and explicit change is expensive, it seems more likely that constitutional rules will be ignored altogether, also leading to a low degree of predictability. Thus, it seems that there is a real tradeoff between rigidity and flexibility. Elkins et al. (2009) summarize their plea in favor of flexibility in a single sentence: “The overall thrust of the argument can be stated simply: constitutions endure when they are most like ordinary statutes.”

4.4 Mind the Gap: Analyzing the Divergence between Constitutional Text and Constitutional Reality 4.4.1 Preliminary Remarks Many constitutions all over the world promise paradise on earth, but very few seem to deliver. Formulated as a question and with less fanciful rhetoric: How much can one learn about the reality of a country by studying its constitution? And, what are 5

Bucur and Rasch (2019) is an up to date survey of the literature analyzing constitutional amendment from a political economy point of view.

100

Conclusions and Possible Future Issues

the factors leading to a high (low) convergence between constitutional text and constitutional reality? It is important to know the answers to these questions because the effects of constitutional rules crucially depend on the degree to which they are implemented, not simply on how well constitutional text reads. Chapter 3 of this primer contains one very explicit example of this finding: de jure judicial independence does not have much of an effect, it is the de facto independence of the judiciary that is conducive to investment and, at the end of the day, to income and growth. Asking these questions with regard to basic human rights seems to suggest itself as the gap between text and reality seems to be particularly wide when addressing basic human rights. But these questions are definitely not confined to human rights. One might, for example, ask why presidential term limits are only complied with some of the time, under what conditions elections are likely to be postponed or entirely cancelled, under what conditions the relationship between the federal level and the regions is not in line with constitutional text, and under what conditions governments are more likely to ignore court rulings, to pose only a few possible queries. Amazingly few scholars have attempted to answer these questions. Diverging somewhat from the other parts of this primer, we go beyond summarizing the very little empirical evidence on the question, and deal with some of the conceptual difficulties that need to be tackled before the questions can be answered. The two most important ones are: How to define the gap? And, how to measure the gap?

4.4.2 How to Define and Measure the Gap? Before a de jure/de facto gap can be measured, a precise definition of the de jure constitution is needed, just as a precise measurement of its de facto element. Here, we do not set out to provide such definition, but simply offer a sample of questions that need to be answered before a definition can be offered. For example, do we only rely on the text of the document for a de jure definition, or are court decisions interpreting the constitution included? A decision in favor of the latter would cause a serious headache for the empiricist as knowing “the constitution” of a country would be truly mind-boggling. Questions dealing with the measurement of de facto implementation might include: How should aspirational articles be dealt with? Should state capacity explicitly be taken into account? After all, the non-implementation of some social rights might be due to missing resources, and not missing willingness on the side of politicians. How to deal with the very slow implementation of constitutionally guaranteed services? If the judiciary is independent but needs years to come up with decisions, this can cause severe problems as the old saying “justice delayed is justice denied” implies. How to deal with situations in which the articles of the constitution are not explicitly broken, but circumvented by making important decisions in forums not specified in the constitution (Caruso et al. 2015)?

4.4 Mind the Gap

101

When trying to come up with a measure, does it make sense to come up with a single aggregate measure? But many debatable decisions regarding the aggregation of many different aspects need to be made. Should the dimensions simply be added up? Should different weights be attached to different dimensions? If yes, how to decide upon the relative weights? Prima facie, more fine-grained measures promise to be more useful. It might make sense to separate the rights part of a constitution from its organizational part. Within the rights part, one might want to distinguish between negative and positive rights and so on.

4.4.3 Possible Determinants of the Gap: Theory In Chapter 1, various concepts of the constitution were presented. One of them promises to be particularly helpful in identifying possible causes for a de jure/de facto gap, namely the constitution as a coordination device. At the end of the day, all constitutional traits need to be self-enforcing as they are the most basic layer of formal law. Constitutional rules have been interpreted as means for citizens to coordinate their behavior against a government that tries to renege against specific traits of the constitution (e.g., Weingast 1997). If there is a gap, this could indicate an insufficient coordination quality of the rules: the limits that government is supposed to comply with could, e.g., be formulated in an imprecise manner. One could, hence, ask if the coordination qualities of the constitution are sufficiently salient and precise. This logic can be taken a step further. Precise constitutional constraints can be thought of as a necessary condition for the absence of a gap, but they are not sufficient. After all, citizens still need to act collectively to oppose government attempts to transgress constitutional boundaries. It has been argued (Voigt 1999) that constitutions need to be backed by informal institutions. An informal institution telling people that they should oppose government if it transgresses against their neighbors could be immensely helpful in strengthening opposition against government attempts to renege upon the constitution. This implies that one should investigate what informal institutions might decrease the likelihood of a gap between de jure and de facto.

4.4.4 Challenges for Establishing Causality It might very well be the case that particular constitutional traits are important factors for explaining the (non-)convergence between de jure and de facto. For example, both a free press and an independent judiciary might make it more costly for government to deviate from constitutional text. But both the freedom of the press and the independence of the judiciary are also constitutional provisions. And as pointed out in previous chapters, they are not exogenous.

102

Conclusions and Possible Future Issues

4.4.5 Attempts to Analyze the Gap: Empirics Although understanding the underlying reasons causing a gap between de jure and de facto promises to be hugely important, very few explicit attempts have been made to analyze it. In their study on “sham constitutions,” Law and Versteeg (2013) look into three categories of constitutional rights, namely personal integrity rights, civil and political freedoms, and socioeconomic and group rights. For the de jure part, they code whether a particular right is mentioned in the constitution, and for the de facto part, they rely on a number of publicly available datasets. Potential determinants of underperformance are separated into country characteristics on the one hand and constitutional characteristics on the other. The empirical results are only offered in a summary fashion, and neither coefficients nor standard errors are reported. The strength of their results thus remains unknown. They present results for an overall indicator, but also for the three different groups just mentioned. It is noteworthy that none of the potential determinants turns out to be significantly correlated with all four indicators. This finding thus supports our plea for fine-grained indicators and against aggregate ones. If we look at country characteristics, both the level of income and democracy are conducive to constitutional reality and constitutional text converging. The experience of civil war, a large population, and a high degree of ethnic fractionalization all make constitutional reality diverge from constitutional text. Law and Versteeg differentiate between a constitution that “contains only generic rights” and one that “also encompasses more esoteric provisions.” They refer to this difference as “comprehensiveness,” and find that more comprehensive constitutions are less likely to reach convergence. The paper by Law and Versteeg can only be a very first step in analyzing government compliance with constitutional rights. Chilton and Versteeg (2015) is a follow-up paper based on a simple and straightforward conjecture: Rights that are addressed to groups and allow for the formation of associations like unions or parties (“organizational rights”) are more likely to be actually enforced than rights addressed to individuals only. Once parties or unions are formed, the cost of denying them their right to existence is supposed to be higher than denying an individual its right to free movement or freedom of expression. The empirics are in line with their conjecture. In sum, our state of knowledge regarding the reasons for a gap between constitutional text and constitutional reality leaves much to be desired. Empirical research into the issue is confronted with a number of serious challenges. But such challenges should make it all the more exciting to delve into this topic!

4.5 Emergency Constitutions 4.5.1 Introductory Remarks At any given time, a large part of humankind lives under a state of emergency. Between 1985 and 2014, we know of 137 countries that declared a state of

4.5 Emergency Constitutions

103

emergency at least once. This is almost exactly two-thirds of all states. In some countries, the state of emergency has even become the rule: Syria has been under an emergency since 1920, Israel has never left it since its founding of 1948, and Egypt has lived under it almost continuously since 1981. But governments in other world regions also rely on declaring states of emergency. For example, in Europe, the French government declared an emergency after the terrorist attacks on the Stade de France and the Bataclan music hall in November 2015 which was prolonged a number of times. Turkey has been under a state of emergency ever since the failed coup in the summer of 2016. Calling a state of emergency has far-reaching consequences. It is usually accompanied by a shift in the balance of powers where the executive expands its power at the expense of the legislature, and sometimes even the judiciary. A state of emergency is also often accompanied by curtailing certain civil rights. At present, nine out of ten states have provisions for the state of emergency in their constitution. For the sake of simplicity, we refer to these provisions as a country’s “emergency constitution.” Considering the frequency that governments declare a state of emergency and the fact that there are far-reaching consequences for both the rule of law and the level of democracy in a country, and if one takes into account its potential relevance for both revocation of individual rights and the division of powers, it is astonishing how little we know about emergency constitutions.

4.5.2 Measurement Issues Emergency constitutions have a function only if the powers of the government are constitutionally limited. Governments, whose competencies are unlimited, do not need clauses that extend their competences. Absolutist governments, therefore, do not need emergency constitutions. In fact, the rise of emergency constitutions in the modern era begins with the French Revolution. By the middle of the nineteenth century, most countries whose legal systems had been directly or indirectly affected by France had already passed emergency constitutions. This is true not only of the constitutions of Spain and Portugal, but also of almost all of the constitutions of Latin America. Are there general determinants that lead countries to introduce an emergency constitution? And are there specific factors that allow us to predict what kind of emergency constitution a country is likely to have? These questions have been dealt with extensively by Bjørnskov and Voigt (2018a). They argue that the composition of the constitutional convention is an important predictor of who gets what powers under a state of emergency. If the constitutional convention is dominated by legislators, the legislature is expected to keep more powers than if the convention is dominated by members who represent the executive. This hypothesis is confirmed by the data. Constitutional conventions dominated by the legislature make it more

104

Conclusions and Possible Future Issues

difficult for the executive to declare a state of emergency, and they also ensure that the executive is not given too many extra competencies if the emergency is actually called. Besides these restrictions placed on the executive branch, it is more difficult to dissolve the parliament, and basic rights are afforded more protection. Beyond this central result, Bjørnskov and Voigt (2018a) also find that countries having experienced more coups delegate more powers to the executive. Compared to the German/Scandinavian legal family, both the common law and the French legal family delegate more competencies to the executive. Likewise – and still in comparison with the German/Scandinavian legal family, the common law and the French legal family have a longer list of events that justify the proclamation of an emergency. Some non-expected results found by Bjørnskov and Voigt (2018a) are no less interesting: countries prone to natural disasters are less likely to allow the government to dissolve parliament. Also, the abolition of civil rights is more difficult there than in other countries. Finally, autocratic countries do not assign significantly more rights to their executive during an emergency than parliamentary democracies.

4.5.3 Who Uses Emergency Constitutions? Having provisions for an emergency constitution does not mean it will actually be used. Therefore, in this section, I ask whether it is possible to describe the situations in which a state of emergency is actually called. But, of course, it is plausible to assume that the provisions of an emergency constitution and the likelihood of an emergency being actually called are closely correlated. The easier it is to call a state of emergency, the more often politicians will do just that. Further, the greater the benefit politicians draw from the emergency, the more often we have to expect an emergency to be actually called. Here is what Bjørnskov and Voigt (2018b) discovered. First of all, it is extremely important to distinguish between natural catastrophes and political events such as domestic general strikes or mass demonstrations. This seems intuitively plausible because these are very different events. As was to be expected, a greater number of domestic conflicts is associated with a higher likelihood of an emergency being called. Governments of countries with parliamentary systems are less likely to declare a state of emergency than countries with a presidential system. The authors expected the opposite because parliamentary governments are likely to gain more additional powers than presidential governments under a state of emergency. Therefore, parliamentary governments should be more likely to declare a state of emergency – but do not. There are no significant differences in the likelihood of calling a state of emergency between parliamentary democracies and autocracies. Finally, and probably most interestingly, the smaller the cost of proclaiming a state of emergency, the more likely its proclamation.

4.5 Emergency Constitutions

105

4.5.4 Are Emergency Constitutions Effective? Emergency constitutions are conservative documents in the sense that their ultimate goal is to return to the status quo ante. If emergency constitutions are actually used to achieve these goals, the question “who benefits from their use?” seems superfluous. However, it cannot be ruled out that politicians call a state of emergency in order to intimidate the population, to silence the opposition, or simply to remain in power. The question of the effectiveness of emergency constitutions can be asked in relation to political events as well as to natural disasters. Political events are more complex than natural disasters. Natural disasters themselves, not their consequences, are highly exogenous. This is different with regard to political events. In the case of political events, a government could instigate a demonstration towards violence, thereby giving them the sought-after opportunity to call a state of emergency. Assuming that the main goal of using emergency constitutions in the aftermath of a natural disaster is to save as many lives as possible and controlling for the severity of a natural catastrophe, Bjørnskov and Voigt (2017) find that autocracies do not perform worse than democracies, and presidential systems suffer fewer fatalities than parliamentary systems. Their most astonishing result is that the more the executive profits from calling a state of emergency, the higher the fatalities. Regarding a specific type of political event, namely terrorist attacks, Bjørnskov and Voigt (2020) analyze the behavior of governments in 79 countries all having a Western-style constitution. Interestingly, right-of-center governments are less likely to suffer from terrorist attacks but once they do, they are more likely to declare a state of emergency. Once a state of emergency has been declared, it generally leads to higher levels of government repression (measured in the degree to which governments respect basic human rights). Bjørnskov and Voigt (2020) also find that states of emergency are less likely to be declared in election years. Governments in countries with a Westernstyle constitution seem to believe that declaring a state of emergency is unpopular among the population. In Bjørnskov and Voigt (2019), the behavior of governments in the Muslim world subsequent to terrorist incidents is analyzed. This is of particular relevance because the region stretching from Morocco in the West to Afghanistan and Pakistan in the East is the most terror-prone region in the world. Interestingly, countries explicitly close to Islam (not only in terms of the proportion of the population professing to be Muslim but also having constitutionalized Sharia in some way) are less likely to declare a state of emergency, whereas revenues from exporting oil do not seem to play a significant role in making that decision.

4.5.5 Summary and Possible Implications for Future Research In sum, it is very clear that the content of emergency constitutions does channel the behavior of governments, albeit not always in the intended direction. A number of

106

Conclusions and Possible Future Issues

possible implications for future research almost seem to suggest themselves. First, the effectiveness of emergency constitutions with regard to domestic disputes ought to be analyzed. Then the results presented here should be used to identify the probable transmission channels. Why do emergency regulations in parliamentary systems differ from those in presidential systems? Why is the number of victims affected by natural catastrophes positively correlated with the benefits a government can draw from calling a state of emergency for itself? Even though emergency constitutions can clearly channel the behavior of governments, one must ask whether this is always the case. It can be assumed that governments often ignore the emergency clauses. If this assumption can be confirmed empirically, one must ask which determinants are responsible for the divergence between de jure and de facto.

Appendix 1 Coding Countries according to Two Governance Scores

Country Name

Regime Type

Afghanistan Albania

Civilian autocracy Parliamentary democracy Civilian autocracy Civilian autocracy Presidential democracy Mixed democracy Parliamentary democracy Mixed democracy Civilian autocracy Royal dictatorship Civilian autocracy Civilian autocracy Parliamentary democracy Presidential democracy Parliamentary democracy Presidential democracy Civilian autocracy Mixed democracy Presidential democracy Mixed democracy Military dictatorship Civilian autocracy Civilian autocracy

Algeria Angola Argentina Armenia Australia Austria Azerbaijan Bahrain Bangladesh Belarus Belgium Benin Bhutan Bolivia Bosnia and Herzegovina Botswana Brazil Bulgaria Burkina Faso Burundi Cambodia

Democracy Dummy

Polity2 Score

0 1

1 9

0 0 1

2 2 9

1 1

5 10

1 0 0 0 0 1

10 7 10 1 7 8

1

7

1

5

1

7

0 1 1

0 8 8

1 0 0 0

9 6 1 4

108

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Cameroon Canada

Civilian autocracy Parliamentary democracy Mixed democracy Military dictatorship Military dictatorship Presidential democracy Civilian autocracy Presidential democracy Presidential democracy Military dictatorship Civilian autocracy

Cape Verde Central African Republic Chad Chile China Colombia Comoros Congo Congo, Democratic Republic Costa Rica Cote d'Ivoire Croatia Cuba Cyprus Czech Republic Denmark Djibouti Dominican Republic Ecuador Egypt El Salvador Equatorial Guinea Eritrea

Presidential democracy Civilian autocracy Mixed democracy Military dictatorship Presidential democracy Parliamentary democracy Parliamentary democracy Civilian autocracy Presidential democracy Presidential democracy Military dictatorship Presidential democracy Military dictatorship Civilian autocracy

Democracy Dummy

Polity2 Score

0 1

4 10

1 0 0 1

10 6 2 10

0 1

7 7

1

9

0 0

4 3

1

10

0 1 0 1

4 9 7 10

1

9

1

10

0 1

3 7

1

5

0 1

4 8

0 0

6 7

109

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Estonia

Parliamentary democracy Royal dictatorship Civilian autocracy Military dictatorship Mixed democracy Mixed democracy Civilian autocracy Presidential democracy Mixed democracy Parliamentary democracy Presidential democracy Parliamentary democracy Presidential democracy Presidential democracy Presidential democracy Presidential democracy Civilian autocracy Presidential democracy Parliamentary democracy Parliamentary democracy Presidential democracy Civilian autocracy Military dictatorship Mixed democracy

Eswatini Ethiopia Fiji Finland France Gabon Gambia Georgia Germany Ghana Greece Guatemala Guinea Guinea-Bissau Guyana Haiti Honduras Hungary India Indonesia Iran Iraq Ireland

Democracy Dummy

Polity2 Score

1

9

0 0 0 1 1 0 0

9 3 2 10 9 3 4

1 1

7 10

1

8

1

10

1

8

1

4

1

6

1

7

0 1

5 5

1

10

1

9

1

9

0 0 1

7 6 10

110

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Israel

Parliamentary democracy Parliamentary democracy Parliamentary democracy Parliamentary democracy Royal dictatorship Civilian autocracy Presidential democracy Military dictatorship Presidential democracy Royal dictatorship Mixed democracy Civilian autocracy Parliamentary democracy Civilian autocracy Parliamentary democracy Civilian autocracy Civilian autocracy Mixed democracy Parliamentary democracy Mixed democracy Civilian autocracy Presidential democracy Mixed democracy Mixed democracy Military dictatorship Parliamentary democracy

Italy Jamaica Japan Jordan Kazakhstan Kenya Korea, North Korea, South Kuwait Kyrgyzstan Laos Latvia Lebanon Lesotho Liberia Libya Lithuania Luxembourg Macedonia Madagascar Malawi Malaysia Mali Mauritania Mauritius

Democracy Dummy

Polity2 Score

1

6

1

10

1

9

1

10

0 0 1

3 6 9

0 1

10 8

0 1 0 1

7 8 7 8

0 1

6 8

0 0 1 1

6 0 10 10

1 0 1

9 6 6

1 1 0 1

5 5 2 10

111

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Mexico

Presidential democracy Parliamentary democracy Mixed democracy Civilian autocracy Royal dictatorship Civilian autocracy Civilian autocracy Civilian autocracy Parliamentary democracy Parliamentary democracy Parliamentary democracy Civilian autocracy Mixed democracy Presidential democracy Parliamentary democracy Royal dictatorship Parliamentary democracy Presidential democracy Parliamentary democracy Presidential democracy Presidential democracy Presidential democracy Mixed democracy Mixed democracy

Moldova Mongolia Montenegro Morocco Mozambique Myanmar Namibia Nepal Netherlands New Zealand Nicaragua Niger Nigeria Norway Oman Pakistan Panama Papua New Guinea Paraguay Peru Philippines Poland Portugal

Democracy Dummy

Polity2 Score

1

8

1

9

1 0 0 0 0 0 1

10 9 4 5 8 6 6

1

10

1

10

0 1 1

6 5 7

1

10

0 1

8 7

1

9

1

5

1

9

1

9

1

8

1 1

10 10

112

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Qatar Romania Russia Rwanda Saudi Arabia Senegal Serbia Sierra Leone

Royal dictatorship Mixed democracy Civilian autocracy Military dictatorship Royal dictatorship Mixed democracy Mixed democracy Presidential democracy Civilian autocracy Mixed democracy Parliamentary democracy Parliamentary democracy Civilian autocracy Civilian autocracy Military dictatorship Parliamentary democracy Presidential democracy Military dictatorship Presidential democracy Parliamentary democracy Presidential democracy Military dictatorship Mixed democracy Civilian autocracy Military dictatorship Military dictatorship Mixed democracy Civilian autocracy

Singapore Slovakia Slovenia Solomon Islands Somalia South Africa South Sudan Spain Sri Lanka Sudan Suriname Sweden Switzerland Syria Taiwan Tajikistan Tanzania Thailand Timor-Leste Togo

Democracy Dummy

Polity2 Score

0 1 0 0 0 1 1 1

10 9 4 3 10 7 8 7

0 1 1

2 10 10

1

8

0 0 0 1

5 9 0 10

1

6

0 1

4 5

1

10

1

10

0 1 0 0 0 1 0

9 10 3 3 3 8 2

113

Appendix 1 Coding Countries

(cont.) Country Name

Regime Type

Trinidad and Tobago

Parliamentary democracy Presidential democracy Civilian autocracy Civilian autocracy Civilian autocracy Mixed democracy Royal dictatorship Parliamentary democracy Presidential democracy Presidential democracy Civilian autocracy Civilian autocracy Civilian autocracy Military dictatorship Presidential democracy Civilian autocracy

Tunisia Turkey Turkmenistan Uganda Ukraine United Arab Emirates United Kingdom United States Uruguay Uzbekistan Venezuela Vietnam Yemen Zambia Zimbabwe

Democracy Dummy

Polity2 Score

1

10

1

7

0 0 0 1 0 1

4 8 1 4 8 10

1

8

1

10

0 0 0 0 1

9 3 7 0 6

0

4

Sources: Bjørnskov and Rode (2019) and Marshall et al. (2018); the democracy dummy is coded 1 if the respective country is a democracy and 0 otherwise; the polity2 variable can take on values between 10 (perfect autocracy) and 10 (perfect democracy). All values for 2017 except for those for Bosnia and Herzegovina which are for 2014.

Appendix 2 Empirical Results at a Glance: Constitutional Rules As Explanatory Variables and Cross-Country Results Unless Otherwise Noted

Dependent Variables I 1

Constitutional Rule Source

Result

Security of Property Rights Property Rights  de facto independence of judiciary (Factual) Judicial Independence Decentralization (of revenue)

Knack and Keefer (1995) Voigt and Gutmann (2013)

+, * +; *

OLS, N = 97, 1974 89 GLS, N = 77

Feld and Voigt (2003) Enikolopov and Zhuravskaya (2007) Thießen (2003b)

+, *

OLS, N = 73

,

OLS, N = 73

,*

GLS, N = 26

Thießen (2003b)

Hump shape , (*)

Panel, N = 26, 1975 95 N = 46

+, *

OLS, N = 25

+, *

OLS, N = 82, 1965 90

Macro Outcome Variables Economic Growth

Decentralization (of revenue) Decentralization (of revenue) Decentralization (of spending) Decentralization (of spending) Number of Veto Players 2

Estimator, # of Countries, Period, . . .

Per Capita Income

Davoodi and Zou (1998) Thießen (2003a) Henisz (2000)

115

Appendix 2 Empirical Results at a Glance

(cont.) Dependent Variables 3

4

II 1

Constitutional Rule Source

Result

Estimator, # of Countries, Period, . . .

Proportional Representation Proportion of Individually Elected Candidates Size of Electoral District Presidential Form of Government Presidential Form of Government

PT (2003)

+, (*)

OLS, N = 74

PT (2003)

+, *

OLS, N = 73

PT (2003)

,*

OLS, N = 73

PT (2003)

, (*)

OLS, N = 74

Blume et al. (2009a)

,

OLS, N = 79

Proportional Representation Proportion of Individually Elected Candidates Size of Electoral District Presidential Form of Government Presidential Form of Government Direct Democracy (fiscal referendums)

PT (2003)

+,

OLS, N = 73

PT (2003)

+,

OLS, N = 72

Proportional Representation Proportional Representation

Labor Productivity

Total Factor Productivity

PT (2003)

,*

OLS, N = 72

PT (2003)

,

OLS, N = 73

Blume et al. (2009a) Feld and Savioz (1997)#

,*

OLS, N = 79

+, *

GLS, N = 26

PT (2003)

+,

OLS, N = 76

Blume et al. (2009)

+, (*)

OLS, N = 78

Fiscal Policy Variables Government Revenue

116

Appendix 2 Empirical Results at a Glance

(cont.) Dependent Variables

Constitutional Rule Source

Result

Estimator, # of Countries, Period, . . .

Presidential Form of Government Presidential Form of Government Presidential Form of Government Presidential Form of Government  executive discretion re the budget Decentralization

PT (2003)

,*

OLS, N = 76

Blume et al. (2009)

,

OLS, N = 78

Rockey (2012)

,*

Fuller, CUE, N = 80

Ardanaz and Scartascini (2014)

,*

OLS, N = 76

Feld et al. (2003)#

,*

Panel, 2SLS, N = 26, 1980 98

Direct Democracy (initiative) Direct Democracy (mandatory fiscal referendums)

Matsusaka (1995)#

,*

Feld and Kirchgässner (2001)#

,*

Panel, OLS, 49, 1960 Panel, OLS, 26, 1986

State and Local Tax Revenue

2

N= 90 N= 97

A. Central Government Expenditure Proportional Representation Presidential Form of Government Bicameral Legislature Bicameral Legislature

PT (2003)

+, *

OLS, N = 80

PT (2003)

,*

OLS, N = 80

Bradbury and Crain (2001) Plümper and Martin (2003)

,*

WLS, N = 37, 1971 89 OLS, N = 83, 1975 97

,*

B. Total Government Expenditure Proportional Blume et al. Representation (2009a) Presidential Form Blume et al. of Government (2009a) Federalism (“own Rodden (2003) source” revenue)

+,

OLS, N = 77

,

OLS, N = 77

,*

Panel, N = 44, 1980 93

117

Appendix 2 Empirical Results at a Glance

(cont.) Dependent Variables

Constitutional Rule Source

Result

Estimator, # of Countries, Period, . . .

Direct Democracy (mandatory referendums) Direct Democracy (possibility initiative)

Blume et al. (2009a)

,*

OLS, N = 62

Blume et al. (2009a)

+, *

OLS, N = 62

Direct Democracy (initiative) Direct Democracy (mandatory fiscal referendums) Direct Democracy (mandatory fiscal referendums) Direct Democracy (initiative)

Matsusaka (1995)#

,*

Feld and Kirchgässner (2001)#

,*

Funk and Gathmann (2011)#

,*

C. State and Local Government Expenditure

3

N= 90 N= 97

Panel, N = 25, 1890 2000

Asatryan (2016)#

+;*

2SLS, N  13,000

PT (2003)

+, *

OLS, N = 60

,*

OLS, N = 45

Milesi Ferretti et al. (2002)

+, *

OLS, N = 40

PT (2003)

?,

OLS, N = 56

Budget Deficit Proportional Representation Direct Democracy (frequency of use)

4

Panel, OLS, 49, 1960 Panel, OLS, 26, 1986

Composition of Government Spending A. Transfers Proportional (Social Representation Security plus (district size) Subsidies to Firms) B. Social Proportional Security and Representation Welfare Presidential Form of Government

Blume et al. (2009a)

PT (2003)

, (*)

OLS, N = 56/69

118

Appendix 2 Empirical Results at a Glance

(cont.) Dependent Variables III 1

2

3

Constitutional Rule Source

Result

Estimator, # of Countries, Period, . . .

Governance Variables Corruption Proportion of Individually Elected Candidates Size of Electoral District Presidential Form of Government Presidential Form of Government Presidential Form of Government Bicameral Legislature Federalism (fiscal decentralization)

PT (2003)

,*

WLS, N = 68

PT (2003)

+, *

WLS, N = 68

PT (2003)

,

Direct Democracy (initiative)

Alt and Lassen (2003)#

,*

OLS, N = 45

Proportion of Individually Elected Candidates Size of Electoral District Presidential Form of Government Presidential Form of Government Number of Veto Players

PT (2003)

,*

WLS, N = 78

PT (2003)

+, *

WLS, N = 78

PT (2003)

,*

WLS, N = 78

Blume et al. (2009a) Panizza (2001)

,

WLS, N = 79

,*

2SLS, N = 111

Direct Democracy (initiatives)

Asatryan and de Witte (2015)#

+, *

FDH, N = 705

Gerring and Thacker (2004) Lederman et al. (2005) Testa (2010)

+, *

WLS, N = 125

+, *

Treisman (2007)

No effects,

Probit/OLS, 1987 97 Panel, N = 34, 1996 2000 OLS, N = 54, 2005

,*

Graft

Government Efficiency

119

Appendix 2 Empirical Results at a Glance

(cont.) Dependent Variables IV 1

2

Constitutional Rule Source

Result

Estimator, # of Countries, Period, . . .

Others Propensity to Cheat on Taxes Direct Democracy (mandatory referendums)

Blume et al. (2009a)

Direct Democracy

Frey and Stutzer (2006)# Lascher and Wassmer (2007)# Blume et al. (2009a)

,*

OLS, N = 49

Reported Life Satisfaction (“Happiness”)

Direct Democracy (various measures) Direct Democracy (various measures)

+, *

WLS, N = 6,134

No effects,

Various, N = up to 2,324

No effects,

OLS, N = 54

Persson and Tabellini are abbreviated with “PT”; +/ = sign of coefficient in corresponding regression; * = claimed to be significant; (*) = marginally significant; claimed to be insignificant; “#” refers to within country studies.

=

References Acemoglu, D. (2005). Constitutions, politics and economics: A review essay on Persson and Tabellini’s The Economic Effects of Constitutions. Journal of Economic Literature, 43(4), 1025 48. Acemoglu, D., Johnson, S., & Robinson, J. (2001). The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5), 1369 401. Acemoglu, D., Johnson, S., & Robinson, J. (2005). A response to Albouy’s “A Reexamination Based on Improved Settler Mortality Data.” Mimeo: MIT. Acemoglu, D., Johnson, S., & Robinson, J. (2006). Reply to the revised (May 2006) version of David Albouy’s “The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data.” Mimeo. Acemoglu, D., & Robinson, J. (2000). Why did the West extend the franchise? Quarterly Journal of Economics, 115, 1167 99. Acemoglu, D., & Robinson, J. (2001). A theory of political transitions. American Economic Review, 91(4), 938 63. Acemoglu, D., & Robinson, J. A. (2006). Economic Origins of Democracy and Dictatorship. Cambridge, MA: Cambridge University Press. Aghion, Ph., Alesina, A., & Trebbi, F. (2004). Endogenous political institutions. Quarterly Journal of Economics, 119(2), 565 611. Aghion, Ph., Alesina, A., & Trebbi, F. (2005). Choosing Electoral Rules: Theory and Evidence from US Cities. Harvard Institute of Economic Research, Discussion Paper 2065. Aidt, T. S., Dutta, J., & Loukoianova, E. (2006). Democracy comes to Europe: Franchise extension and fiscal outcomes 1830 1938. European Economic Review, 50(2), 249 83. Aidt, T. S., & Jensen, P. S. (2014). Workers of the world, unite! Franchise extensions and the threat of revolution in Europe, 1820 1938. European Economic Review, 72, 52 75. Albouy, D. (2008). The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data (June 2008). NBER Working Paper, W14130. Alesina, A., & Glaeser, E. L. (2004). Fighting Poverty in the US and Europe: A World of Difference. Oxford: Oxford University Press. Alesina, A., & Spolaore, E. (2003). The Size of Nations. Cambridge, MA and London: MIT Press. Alt, J., & Lassen, D. (2003). The political economy of institutions and corruption in American States. Journal of Theoretical Politics, 5(3), 341 65. Anderson, G., Martin, D., Shugart II, W., & Tollison, R. (1990). Behind the veil: The political economy of constitutional change. In: W. Crain & R. Tollison (eds.), Predicting Politics: Essays in Empirical Public Choice. Ann Arbor: University of Michigan Press, 89 100. Anderson, G., Shugart II, W., & Tollison, R. (1989). On the incentives of judges to enforce legislative wealth transfers. The Journal of Law & Economics, 32, 215 28.

References

121

Angrist, J. D., & Pischke, J. S. (2010). The credibility revolution in empirical economics: How better research design is taking the con out of econometrics. The Journal of Economic Perspectives, 24(2), 3 30. Ardanaz, M., & Scartascini, C. (2014). The economic effects of constitutions: Do budget institutions make forms of government more alike? Constitutional Political Economy, 25(3), 301 29. Aristotle (1932). The Politics, with an English translation by H. Rackham, London: W. Heinemann. Arndt, Chr., & Oman, Ch. (2006). Uses and Abuses of Governance Indicators. Paris: OECD. Arnold, F., & Freier, R. (2015). Signature requirements and citizen initiatives: Quasi experimental evidence from Germany. Public Choice, 162(1), 43 56. Arrow, K. (1951). Social Choice and Individual Values. New York: John Wiley & Sons. Arzaghi, M., & Henderson, J. V. (2005). Why countries are fiscally decentralizing. Journal of Public Economics, 89(7), 1157 89. Asatryan, Z. (2016). The indirect effects of direct democracy: Local government size and non budgetary voter initiatives in Germany. International Tax and Public Finance, 23(3), 580 601. Asatryan, Z., Baskaran, T., Grigoriadis, T., & Heinemann, F. (2017). Direct democracy and local public finances under cooperative federalism. The Scandinavian Journal of Economics, 119(3), 801 20. Asatryan, Z., & de Witte, K. (2015). Direct democracy and local government efficiency. European Journal of Political Economy, 39, 58 66. Asian Development Bank (2005). Country Governance Assessment: Philippines. www.adb.org/ Documents/Reports/CGA. Austen Smith, D. (2000). Redistributing income under proportional representation. Journal of Political Economy, 108(6), 1235 69. Bannon, A. L. (2007). Designing a constitution drafting process: Lessons from Kenya. The Yale Law Journal, 116, 1824 72. Barro, R. J. (1997). Determinants of Economic Growth: A Cross Country Empirical Study. Cambridge: MIT Press. Barro, R. J. (2000). Rule of law, democracy, and economic performance. In: G. O’Driscoll, Jr., K. Holmes, & M. Kirkpatrick (eds.), 2000 Index of Economic Freedom. Washington, DC: The Heritage Foundation, 31 49. Barzel, Y. (1997). Parliament as a wealth maximizing institution: The right to the residual and the right to vote. International Review of Law and Economics, 17, 455 74. Barzel, Y. (2000). Property rights and the evolution of the state. Economics of Governance, 1(1), 25 51. Baskaran, T., & da Fonseca, M. L. (2016). Electoral competition and endogenous political institutions: Quasi experimental evidence from Germany. Journal of Economic Behavior & Organization, 122, 43 61. Beard, C. (1913/1986). An Economic Interpretation of the Constitution of the United States; with a New Introduction by Forrest Mc Donald. New York: Free Press. Beath, A., Christia, F., & Enikolopov, R. (2017). Direct democracy and resource allocation: Experimental evidence from Afghanistan. Afghanistan Journal of Development Economics, 124, 199 213.

122

References

Beetsma, R. M., & Debrun, X. (2016). Fiscal Councils: Rationale and Effectiveness. Washington, DC: International Monetary Fund. Ben Bassat, A. & Dahan, M. (2008). Social rights in the constitution and in practice. Journal of Comparative Economics, 36, 103 19. Benoit, K. (2004). Models of electoral system change. Electoral Studies, 23, 363 89. Benoit, K. (2007). Electoral laws as political consequences: Explaining the origins and change of electoral institutions. Political Science, 10(1), 363. Benz, M., & Stutzer, A. (2004). Are voters better informed when they have a larger say in politics? Evidence for the European Union and Switzerland. Public Choice, 119, 31 59. Berggren, N. (2003). The benefits of economic freedom: A survey. The Independent Review, 8(2), 193 211. Berggren, N., & Jordahl, H. (2005). Does free trade really reduce growth? Public Choice, 122, 99 114. Besley, T., & Payne, A. (2003). Judicial accountability and economic policy outcomes: Evidence from employment discrimination charges. Institute for Fiscal Studies (IFS), London, IFS Working Papers No. 03/11, http://dx.doi.org/10.1920/wp.ifs.2003.0311, Mimeo, June. Besley, T. & Reynal Querol, M. (2017). The Logic of Hereditary Rule: Theory and Evidence. CEPR Discussion Paper 11777. Bjørnskov, Ch., & Rode, M. (Forthcoming). Regime types and regime change: A new dataset. Review of International Organizations. Bjørnskov, Ch. and Voigt, S. (2017). Profiting from Natural Disaster? Inquiring into the Effectiveness of Emergency Constitutions. Typescript. Universities of Aarhus and Hamburg. Bjørnskov, Ch. and Voigt, S. (2018a). The architecture of emergency constitutions. International Journal of Constitutional Law, 16(1), 101 27. Bjørnskov, Ch. and Voigt, S. (2018b). Why do governments call a state of emergency? European Journal of Political Economy, 54, 110 23. Bjørnskov, Ch. and Voigt, S. (2019). Terrorism and Emergency Constitutions in the Muslim World. Typescript. Universities of Aarhus and Hamburg. Bjørnskov, Ch. and Voigt, S. (2020). When does terror induce a states of emergency? And what are the effects? Journal of Conflict Resolution, 64(4) 579 613. Blankart, Ch. (2000). The process of government centralization: A constitutional view. Constitutional Political Economy, 11(1), 27 39. Blomberg, S., Hess, G., & Weerapana, A. (2004). The impact of voter initiatives on economic activity. European Journal of Political Economy, 20(1), 207 26. Blume, L., Hayo, B., & Voigt, S. (2014). Correlates and Determinants of Direct Democracy. Available at SSRN 2542156. Blume, L., Müller, J., & Voigt, S. (2009a). The economic effects of direct democracy A first global assessment. Public Choice, 140, 431 61. Blume, L., Müller, J., Voigt, S., & Wolf, C. (2009b). The economic effects of constitutions: Replicating and Extending Persson and Tabellini. Public Choice, 139, 197 225. Blume, L., & Voigt, S. (2007). The economic effects of human rights. Kyklos, 60(4), 509 38. Blume, L., & Voigt, S. (2011). Federalism and decentralization A critical survey of frequently used indicators. Constitutional Political Economy, 22(3), 238 64.

References

123

Boix, C. (1999). Setting the rules of the game: The choice of electoral systems in advanced democracies. American Political Science Review, 93(3), 609 24. Bordignon, M., Nannicini, T., & Tabellini, G. (2016). Moderating political extremism: Single round vs. runoff elections under plurality rule. American Economic Review, 106, 2349 70. Bordo, M. D., Jonung, L., & Markiewicz, A. (2013). A fiscal union for the euro: Some lessons from history. CESifo Economic Studies, 59(3), 449 88. Bradbury, J. C., & Crain, W. M. (2001). Legislative organization and government spending: cross country evidence. Journal of Public Economics, 82, 309 25. Bradbury, J.C., & Crain, W. M. (2002). Bicameral legislatures and fiscal policy. Southern Economic Journal, 68(3), 646 59. Brennan, G., & Buchanan, J. (1980). The Power to Tax. Cambridge: Cambridge University Press. Brennan, G. & Buchanan, J. (1985). The Reason of Rules: Constitutional Political Economy. Cambridge: Cambridge University Press. Brennan, G., & Hamlin, A. (2000). Democratic Devices and Desires. Cambridge: Cambridge University Press. Brennan, G., & Kliemt, H. (1994). Finite lives and social institutions. Kyklos, 47(4), 551 71. Brennan, G., & Kliemt, H. (2019). Fiscal powers revisited: The Leviathan model after 40 years. In: R. Congleton, B. Grofman, & S. Voigt (eds.), The Oxford Handbook of Public Choice. Oxford et al.: Oxford University Press, vol. 2, 263 85. Brennan, G., & Lomasky, L. (1993). Democracy and Decision The Pure Theory of Electoral Preference. Cambridge: Cambridge University Press. Buchanan, J. (1975). The Limits of Liberty Between Anarchy and Leviathan. Chicago: University of Chicago Press. Buchanan, J. (1987). The constitution of economic policy. American Economic Review, 77, 243 50. Buchanan, J., & Tullock, G. (1962). The Calculus of Consent Logical Foundations of Constitutional Democracy. Ann Arbor: University of Michigan Press. Bucur, C. & Rasch, B. E. (2019). Institutions for amending constitutions. In: R. Congleton, B. Grofman, & S. Voigt (eds.), The Oxford Handbook of Public Choice. Oxford et al.: Oxford University Press, vol. 2, 156 76. Carlsson, F., & Lundström, S. (2002). Economic freedom and growth: Decomposing the effects. Public Choice, 112, 335 44. Caruso, G., Scartascini, C., & Tommasi, M. (2015). Are we all playing the same game? The economic effects of constitutions depend on the degree of institutionalization. European Journal of Political Economy, 38, 212 28. Cheibub, J. A., Elkins, Z., & Ginsburg, T. (2013). Beyond presidentialism and parliamentarism. British Journal of Political Science, 44(3), 515 44. Cheibub, J. A., Gandhi, J., & Vreeland, J. R. (2010). Democracy and dictatorship revisited. Public Choice, 143(1 2), 67 101. Chilton, A. S., & Versteeg, M. (2015). Do constitutional rights make a difference? American Journal of Political Science. Colomer, J. (2005). It’s parties that choose electoral systems (or, Duverger’s law upside down). Political Studies, 53, 1 21.

124

References

Congleton, R. D. (2003). On the merits of bicameral legislatures: Policy stability within partisan polities. Jahrbuch für Neue Politische Ökonomie, 22, 29 49. Congleton, R. D. (2006). On the merits of bicameral legislatures: Intragovernmental bargaining and policy stability. In R. D. Congleton (ed.), Democratic Constitutional Design and Public Policy: Analysis and Evidence. Cambridge: MIT Press, 163 88. Congleton, R. D. (2011). Perfecting Parliament: Constitutional Reform, Liberalism, and the Rise of Western Democracy. Cambridge: Cambridge University Press. Crain, W. M., & Tollison, R. (1979). Constitutional change in an interest group perspective. Journal of Legal Studies, 8, 165 75. Cusack, T., Iversen, T., & Soskice, D. (2007). Economic interests and the origins of electoral systems. American Political Science Review, 101(3), 373 91. Cusack, T., Iversen, T., & Soskice, D. (2010). Coevolution of capitalism and political representation: The choice of electoral systems. American Political Science Review, 104 (02), 393 403. Cutrone, M., & McCarty, N. (2006). Does bicameralism matter? In: B. Weingast & D. Wittman (eds.), Oxford Handbook of Political Economy. Oxford: Oxford University Press, 180 95. Dalton, R. (2008). Direct democracy and good governance: Does it matter? In: Sh. Bowler & A. Glazer (eds.), Direct Democracy’s Impact on American Political Institutions. New York: Palgrave. Davoodi, H., & Zou, H. F. (1998). Fiscal decentralization and economic growth: A cross country study. Journal of Urban Economics, 43(2), 244 57. Dicey, A. V. (1985). Introduction to the Study of the Law of the Constitution. Indianapolis, IN: Liberty Fund. Diermeier, D., & Myerson, R. (1999). Bicameralism and its consequences for the internal organization of legislatures. American Economic Review, 89(5), 1182 96. Djankov, S., Glaeser, E., La Porta, R., Lopez de Silanes, F., & Shleifer, A. (2003). The new comparative economics. Journal of Comparative Economics, 31(4), 595 619. Doucouliagos, H., & Ulubasoglu, M. (2008). Democracy and economic growth: A meta analysis. American Journal of Political Science, 52(1), 61 83. Duverger, M. (1954). Political Parties: Their Organization and Activity in the Modern State. New York: Wiley. Duverger, M. (1980). A new political system model: Semi presidential government. European Journal of Political Research, 8(2), 165 87. Eggers, A., Freier, R., Grembi, V., & Nannicini, T. (2018). Regression discontinuity designs based on population thresholds: Pitfalls and solutions. American Journal of Political Science, 62(1), 210 29. Eisenstadt, T. A., LeVan, A. C., & Maboudi, T. (2015). When talk trumps text: The democratizing effects of deliberation during constitution making, 1974 2011. American Political Science Review, 109(03), 592 612. Elazar, D. (1995). From statism to federalism: A paradigm shift. Publius, 25(2), 5 18. Elkins, Z., Ginsburg, T., & Melton, J. (2009).The Endurance of National Constitutions. Cambridge: Cambridge University Press. Elkins, Z., Ginsburg, T., & Melton, J. (2014). The content of authoritarian constitutions. In: T. Ginsburg & A. Simpser (eds.), Constitutions in Authoritarian Regimes. Cambridge: Cambridge University Press, 141 64.

References

125

Elster, J. (1984). Ulysses and the Sirens: Studies in Rationality and Irrationality. Cambridge: Cambridge University Press. Elster, J. (1993). Constitution making in Eastern Europe: Rebuilding the boat in the open sea. Public Administration, 71(1/2), 169 217. Elster, J. (1995). Forces and mechanisms in the constitution making process. Duke Law Journal, 45(2), 364 96. Elster, J. (2000a). Ulysses Unbound: Studies in Rationality, Precommitment, and Constraints. Cambridge: Cambridge University Press. Elster, J. (2000b). Arguing and bargaining in two constituent assemblies. University of Pennsylvania Journal of Constitutional Law, 2, 345 421. Elster, J. (2012). The optimal design of a constituent assembly. In: H. Landemore & J. Elster (eds.), Collective Wisdom: Principles and Mechanisms. Cambridge: Cambridge University Press, 148 72. Enikolopov, R., & Zhuravskaya, E. (2007). Decentralization and political institutions. Journal of Public Economics, 91(11 12), 2261 90. Facchini, G., & Testa, C. (2016). Corruption and bicameral reforms. Social Choice and Welfare, 47(2), 387 411. Farber, D. (2002). Rights as signals. Journal of Legal Studies, 31, 83 98. Feld, L. P., & Kirchgässner, G. (2001). The political economy of direct legislation: Direct democracy and local decision making. Economic Policy, 33, 329 63. Feld, L. P., Kirchgässner, G., & Schaltegger, Chr. (2003). Decentralized Taxation and the Size of Government: Evidence from Swiss State and Local Governments. CESifo Working Paper 1087, December. Feld, L. P., & Savioz, M. (1997). Direct democracy matters for economic performance: An empirical investigation. Kyklos, 50(4), 507 38. Feld, L. P., Schaltegger, Ch., & Schnellenbach, J. (2008). On government centralization and fiscal referendums. European Economic Review, 52, 611 45. Feld, L. P., & Voigt, S. (2003). Economic growth and judicial independence: Cross country evidence using a new set of indicators. European Journal of Political Economy, 19(3), 497 527. Feld, L. P., & Voigt, S. (2006). Making judges independent Some proposals regarding the judiciary. In: R. Congleton and B. Swedenborg (eds.), Democratic Constitutional Design and Public Policy: Analysis and Evidence. Cambridge: MIT Press, 251 88. Ferejohn, J. (1997). The politics of imperfection: The amendment of constitutions. Law and Social Inquiry, 22, 501. Figueiredo, R. & Weingast, B. (2005). Self enforcing federalism. Journal of Law, Economics, and Organization, 21(1), 103 35. Fish, M. S. (2006). Stronger legislatures, stronger democracies. Journal of Democracy, 17(1), 5 20. Frey, B., & Stutzer, A. (2000). Happiness, economy and institutions. The Economic Journal, 110, 918 38. Frey, B., & Stutzer, A. (2006). Direct democracy: Designing a living constitution. In: R. Congleton (ed.), Democratic Constitutional Design and Public Policy: Analysis and Evidence. Cambridge: MIT Press, 39 80.

126

References

Frey, B., & Stutzer, A. (2018). Public choice and happiness. In: R. Congleton, B. Grofman, & S. Voigt (eds.), The Oxford Handbook of Public Choice. Oxford et al.: Oxford University Press. Frohlich, N., Oppenheimer, J., & Eavey, Ch. (1987). Laboratory results on Rawls’s distributive justice. British Journal of Political Science, 17, 1 21. Fuller, L. (1964). The Morality of Law. New Haven, CT: Yale University Press. Funk, P., & Gathmann, C. (2011). Does direct democracy reduce the size of government? New evidence from historical data, 1890 2000. The Economic Journal, 121(557), 1252 80. Funk, P., & Gathmann, C. (2013). How do electoral systems affect fiscal policy? Evidence from cantonal parliaments, 1890 2000. Journal of the European Economic Association, 11(5), 1178 203. Gagliarducci, S., Nannicini, T., & Naticchionia, P. (2011). Electoral rules and politicians’ behavior: A micro test. American Economic Journal: Economic Policy, 3(3), 144 74. Gallie, W. (1956). Essentially contested concepts. Proceedings of the Aristotelian Society, 56, 167 98. Gandhi, J. (2008). Dictatorial institutions and their impact on economic growth. European Journal of Sociology, 49(01), 3 30. Gassebner, M., Lamla, M. J., & Vreeland, J. R. (2013). Extreme bounds of democracy. Journal of Conflict Resolution, 57(2), 171 97. Geddes, B. (1999). What do we know about democratization after twenty years? Annual Review of Political Science, 2(1), 115 44. Geddes, B., Wright, J., & Frantz, E. (2014). Autocratic breakdown and regime transitions: A new data set. Perspectives on Politics, 12(02), 313 31. Gerring, J. (2006). Case Study Research: Principles and Practices. Cambridge University Press. Gerring, J., & Thacker, Str. (2004). Political institutions and corruption: The role of unitarism and parliamentarism. British Journal of Political Science, 34, 295 330. Gerring, J., Thacker, Str., & Moreno, C. (2006). Are Federal Systems Better Than Unitary Systems? Mimeo: Boston University. Gerring, J., Ziblatt, D., Van Gorp, J., & Arevalo, J. (2011). An institutional theory of direct and indirect rule. World Politics, 63(03), 377 433. Gilardi, F. (2002). Policy credibility and delegation to independent regulatory agencies: A comparative empirical analysis. Journal of European Public Policy, 9(6), 873 93. Gilardi, F., & Maggetti, M. (2011). The independence of regulatory authorities. In: D. Levi Faur (ed.), Handbook on the Politics of Regulation, Cheltenham: Edward Elgar, Chapter 14, 201 15. Ginsburg, T. (2002). Economic analysis and the design of constitutional courts. Theoretical Inquiries in Law, 3(1), 49 85. Ginsburg, T. (2003). Judicial Review in New Democracies: Constitutional Courts in Asian Cases. Cambridge: Cambridge University Press. Ginsburg, T., Elkins, Z., & Blount, J. (2009). Does the process of constitution making matter? Annual Review of Law and Social Sciences, 5(5), 84 105. Ginsburg, T., & Melton, J. (2015). Does the constitutional amendment rule matter at all? Amendment cultures and the challenges of measuring amendment difficulty. International Journal of Constitutional Law, 13(3), 686 713.

References

127

Ginsburg, T., & Versteeg, M. (2014). Why do countries adopt constitutional review? Journal of Law, Economics, and Organization, 30, 587 622. Goderis, B., & Versteeg, M. (2014). The diffusion of constitutional rights. International Review of Law and Economics, 39, 1 19. Golden, M., & Picci, L. (2005). Proposal for a new measure of corruption, illustrated with Italian data. Economics & Politics, 17(1), 37 75. Goldwin, R., & Kaufman, A. (eds.). (1988). Constitution Makers on Constitution Making. Washington, DC: American Enterprise Institute. Gutmann, J., Hayo, B. & Voigt, S. (2017). Correlates and Determinants of Direct Democracy. Typescript. Gutmann, J., Padovano, F., & Voigt, S. (2019). Perception vs. Experience: Explaining Differences in Corruption Measures Using Microdata. Munich: CESifo Working Paper No. 8027, December 2019. Gutmann, J., & Voigt S. (2016). Why Adopt a Federal Constitution? And Why Decentralize? Searching for Proximate and Deep Causes. Typescript. Institute of Law & Economics, University of Hamburg. Gutmann, J., & Voigt, S. (2018). Judicial independence in the EU: A puzzle. European Journal of Law and Economics, 1 18. Gwartney, J., Lawson, R., & Block, W. (1996). Economic Freedom of the World: 1975 1995. Vancouver et al.: The Fraser Institute et al. de Haan, J. (2007). Political institutions and economic growth reconsidered. Public Choice, 131, 281 92. de Haan, J. & Eijffinger, S. (2019). The politics of central bank independence. In: R. Congleton, B. Grofman, & Voigt, S. (eds.), The Oxford Handbook of Public Choice. Oxford et al.: Oxford University Press, vol. 2, 499 519. de Haan, J., & Sturm, J. E. (2000). On the relationship between economic freedom and economic growth. European Journal of Political Economy, 16, 215 41. Hadfield, G. K., & Weingast, B. R. (2014). Constitutions as coordinating devices. In: S. Galiani & I. Sened (eds.), Institutions, Property Rights, and Economic Growth: The Legacy of Douglass North. Cambridge: Cambridge University Press, 121 50. Haggard, S., & Kaufman, R. R. (1995). The Political Economy of Democratic Transitions. Princeton: Princeton University Press. Haggard, S., & Kaufman, R. R. (1997). The political economy of democratic transitions. Comparative Politics, 29(3), 263 83. Hall, R. & Jones, Ch. (1999). Why do some countries produce so much more output per worker than others? Quarterly Journal of Economics, 114(1), 83 116. Hamilton, A., Madison, J., & Jay, J. (1788/1961). The Federalist Papers With an Introduction, Table of Contents, and Index of Ideas by C. Rossiter. New York: Mentor. Hanretty, C., & Koop, C. (2012). Measuring the formal independence of regulatory agencies. Journal of European Public Policy, 19(2), 198 216. Hanssen, A. (2004). Is there a politically optimal level of judicial independence? American Economic Review, 94(3), 712 29. Hardin, R. (1989). Why a constitution? In: B. Grofman & D. Wittman (eds.), The Federalist Papers and the New Institutionalism. New York: Agathon Press, 100 20.

128

References

Hayek, F. (1939). Economic conditions of inter state federalism. The New Commonwealth Quarterly, 5(2), 131 49. Hayek, F. (1945). The use of knowledge in society. The American Economic Review, 35(4), 519 30. Hayek, F. (1973). Law, Legislation and Liberty, Vol. 1: Rules and Order. Chicago: Chicago University Press. Hayek, F. (1976). Law, Legislation and Liberty, Vol. 2: The Mirage of Social Justice. Chicago: Chicago University Press. Hayek, F. (1979). Law, Legislation and Liberty, Vol. 3: The Political Order of a Free People. Chicago: Chicago University Press. Hayek, F. (1988). The Fatal Conceit The Errors of Socialism, ed. by W. W. Bartley III. London: Routledge. Hayo, B., & Voigt, S. (2007). Explaining de facto judicial independence. International Review of Law and Economics, 27(3), 269 90. Hayo, B., & Voigt, S. (2010). Determinants of constitutional change: Why do countries change their form of government? Journal of Comparative Economics, 38(3), 283 305. Hayo, B., & Voigt, S. (2013). Endogenous constitutions: Politics and politicians matter, economic outcomes don’t. Journal of Economic Behavior & Organization, 88, 47 61. Hayo, B., & Voigt, S. (2016). Explaining constitutional change: The case of judicial independence. International Review of Law and Economics, 48, 1 13. Hayo, B., & Voigt, S. (2019). The long term relationship between de jure and de facto judicial independence. Economics Letters, 183, 108603. Heller, W. (1997). Bicameralism and budget deficits: The effect of parliamentary structure on government spending. Legislative Studies Quarterly, 22(4), 485 516. Henisz, W. (2000). The institutional environment for economic growth. Economics and Politics, 12(1), 1 31. Hinnerich, B. T., & Pettersson Lidbom, P. (2014). Democracy, redistribution, and political participation: Evidence from Sweden 1919 1938. Econometrica, 82(3), 961 93. Holmes, S. (1988). Precommitment and the paradox of democracy. In: J. Elster & R. Slagstad (eds.), Constitutionalism and Democracy. Cambridge: Cambridge University Press, 195 240. Hooghe, L., Marks, G., Schakel, A., Niedzwiecki, S., Chapman Osterkatz, S., & Shair Rosenfield, S. (2016). Measuring Regional Authority: A Postfunctionalist Theory of Governance, vol. I. Oxford: Oxford University Press. Hume, D. (1777/1987). Essays Moral, Political, and Literary, ed. and with a Foreword, Notes, and Glossary by Eugene F. Miller, Indianapolis, IN: Liberty Classics. Inman, R., & Rubinfeld, D. (2016). Economics of federalism. In: F. Parisi (ed.), Oxford Handbook of Law and Economics, Oxford: Oxford University Press. Iversen, T., & Soskice, D. (2006). Electoral institutions and the politics of coalitions: Why some democracies redistribute more than others. American Political Science Review, 100, 165 181. Justesen, M. (2008). The effect of economic freedom on growth revisited: New evidence on causality from a panel of countries 1970 1999. European Journal of Political Economy, 24, 642 60. Kaufmann, D., Kraay, A., & Mastruzzi, M. (2007). Worldwide Governance Indicators Project: Answering the Critics. World Bank Policy Research Working Paper No. 4149, Available at SSRN: http://ssrn.com/abstract=965077.

References

129

Kirchgässner, G., & Frey, B. (1990). Volksabstimmung und direkte Demokratie: Ein Beitrag zur Verfassungsdiskussion. In: H. D. Klingmann & M. Kaase (eds.), Wahlen und Wähler Analysen aus Anlaß der Bundestagswahl. Westdeutscher Verlag: Opladen, 42 69. Knack, St. (2006). Measuring Corruption in Eastern Europe and Central Asia: A Critique of the Cross Country Indicators. World Bank Policy Research Working Paper 3968. Knack, St., & Keefer, Ph. (1995). Institutions and economic performance: Cross country tests using alternative institutional measures. Economics and Politics, 7(3), 207 27. Kocher, M., Strauß, S., & Sutter, M. (2006). Individual or team decision making Causes and consequences of self selection. Games and Economic Behavior, 56, 259 70. Kreuzer, M. (2010). Historical knowledge and quantitative analysis: The case of the origins of proportional representation. American Political Science Review, 104(02), 369 92. Kurrild Klitgaard, P. (2000). The constitutional economics of autocratic succession. Public Choice, 103(1 2), 63 84. Kurtz, M., & Schrank, A. (2007a). Growth and governance: Models, measures, and mechanisms. Journal of Politics, 69(2), 538 54. Kurtz, M., & Schrank, A. (2007b). Conceptualizing and measuring institutions: A view from political science. Mimeo: Prepared for Second Workshop on Measuring Law & Institutions held in Paris December 14 15, 2007. Landes, W., & Posner, R. (1975). The independent judiciary in an interest group perspective. The Journal of Law & Economics, 18(3), 875 911. Lascher, E., & Wassmer, R. (2007). Reconsidering Claims about the Secondary Benefits of Direct Democracy. Mimeo: Department of Public Policy and Administration, California State University at Sacramento. Law, D., & Versteeg, M. (2013). Sham constitutions. California Law Review, 101(4), 863ff. Lederman, D., Loayza, N., & Soares, R. (2005). Accountability and corruption. Economics & Politics, 17(1), 1 35. Lessmann, C. (2014). Spatial inequality and development: Is there an inverted U relationship? Journal of Development Economics, 106, 35 51. Levmore, S. (1992). Bicameralism: When are two decisions better than one? International Review of Law and Economics, 12, 145 62. Lijphart, A. (1984). Democracies. New Haven, CT: Yale University Press. Lijphart, A. (1994). A Study of Twenty Seven Democracies 1945 1990. New York: Oxford University Press. Lijphart, A. (1999). Patterns of Democracy Government Forms and Performance in Thirty Six Countries. New Haven, CT and London: Yale University Press. List, J. A. (2011). Why economists should conduct field experiments and 14 tips for pulling one off. The Journal of Economic Perspectives, 25(3), 3 15. Lizzeri, A., & Persico, N. (2001). The provision of public goods under alternative electoral incentives. American Economic Review, 91(1), 225 39. Lizzeri, A., & Persico, N. (2004). Why did the elites extend the suffrage? Democracy and the scope of government, with an application to Britain’s “Age of Reform”. Quarterly Journal of Economics, 119(2), 705 63. Ludwikowski, R. (1996). Constitution Making in the Region of Former Soviet Dominance. Durham, NC: Duke University Press.

130

References

Lutz, D. (1994). Toward a theory of constitutional amendment. American Political Science Review, 88(2), 355 70. Lutz, D. (2006). Principles of Constitutional Design. Cambridge: Cambridge University Press. Majone, G. (1996). Temporal Consistency and Policy Credibility: Why Democracies Need Non majoritarian Institutions. San Domenico di Fiesole: Robert Schuman Centre Working Paper RSC No. 95/57. Marshall, G., & Trotter, D. (1984). Constitutional Conventions: The Rules and Forms of Political Accountability. Oxford: Oxford University Press. Marshall, M. G., Gurr, T. R., & Jaggers, K. (2018). Polity IV Project: Political Regime Characteristics and Transitions, 1800 2016. Vienna, VA: Center for Systemic Peace. Matsusaka, J. (1995). Fiscal effects of the voter initiative: Evidence from the last 30 years. Journal of Political Economy, 102(2), 587 623. Matsusaka, J. (2000). Fiscal effects of the voter initiative in the first half of the twentieth century. The Journal of Law & Economics, 43, 619 50. Matsusaka, J. (2004). For the Many or the Few. The Initiative, Public Policy, and American Democracy. Chicago: The University of Chicago Press. McCormick, R., & Tollison, R. (1981). Politicians, Legislation, and the Economy An Inquiry into the Interest Group Theory of Government. Boston: Nijhoff. McGuire, M., & Olson, M. (1996). The economics of autocracy and majority rule: The invisible hand. Journal of Economic Literature, 34, 72 96. Melton, J., & Ginsburg, T. (2014). Does de jure judicial independence really matter? A reevaluation of explanations for judicial independence. Journal of Law and Courts, 2(2), 187 217. Meltzer, A., & Richard, S. (1981). A rational theory of the size of government. Journal of Political Economy, 89, 914 27. Metelska Szaniawska, Katarzyna (2016). Economic Effects of Post Socialist Constitutions 25 Years from the Outset of Transition. Polish Studies in Economics, vol. 7. Frankfurt a. M.: Peter Lang. Michel, S., & Cofone, I. N. (2017). Majority rules in constitutional referendums. Kyklos, 70(3), 402 24. Milesi Ferretti, G. M., Perotti, R., & Ristagno, M. (2002). Electoral systems and public spendine. Quarterly Journal of Economics, 117(2), 609 57. Miller, G., & Hammond, T. (1989). Stability and efficiency in a separation of powers constitutional system. In: B. Grofman & D. Wittman (eds.), The Federalist Papers and the New Institutionalism. New York: Agathon Press, 85 99. Miller, L. (2010). Framing the State in Times of Transition Case Studies in Constitution Making. Washington, DC: United States Institute of Peace. Muthoo, A., & Shepsle, K. (2008). The constitutional choice of bicameralism. In: E. Helpman (ed.), Institutions and Economic Performance. Cambridge, MA: Harvard University Press, 249 92. Mueller, D. (1996). Constitutional Democracy. Oxford: Oxford University Press. Mulligan, C. B., Gil, R., & Sala i Martin, X. (2004). Do democracies have different public policies than nondemocracies? Journal of Economic Perspectives, 18(1), 51 74. North, D. (1981). Structure and Change in Economic History. New York: Norton. North, D. (1990). Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press.

References

131

North, D., Wallis, J., & Weingast, B. (2009). Violence and Social Orders A Conceptual Framework for Interpreting Recorded Human History. Cambridge: Cambridge University Press. North, D. C., Wallis, J. J., Webb, S. B., & Weingast, B. R. (eds.). (2013). In the Shadow of Violence: Politics, Economics, and the Problems of Development. Cambridge: Cambridge University Press. Oates, W. (1999). An essay on fiscal federalism. Journal of Economic Literature, 37, 1120 49. Oates, W. (2005). Toward a second generation theory of fiscal federalism. International Tax and Public Finance, 12, 349 73. Olken, B. (2007). Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy, 115(2), 200 49. Olson, M. (1993). Dictatorship, democracy, and development. American Political Science Review, 87(03), 567 76. Ordeshook, P. (1992). Constitutional stability. Constitutional Political Economy, 3(2), 137 75. Ordeshook, P. (1993). Some rules of constitutional design. In: P. E. Frankel, F. Miller, & J. Paul (eds.), Liberalism and the Economic Order. Cambridge: Cambridge University Press, 198 232. Padovano, F., Sgarra, G., & Fiorino, N. (2003). Judicial branch, checks and balances and political accountability. Constitutional Political Economy, 14(1), 47 70. Panizza, U. (2001). Electoral rules, political systems, and institutional quality. Economics & Politics, 13(3), 311 42. Patterson, S. C., & Mughan, A. (1999). Senates: Bicameralism in the Contemporary World. Columbus, OH: Ohio State University Press. Persson, T. (2005). Forms of Democracy, Policy and Economic Development. Mimeo, January. Persson, T., Roland, G., & Tabellini, G. (1997). Separation of powers and political accountability. Quarterly Journal of Economics, 112, 310 27. Persson, T., Roland, G., & Tabellini, G. (2000). Comparative politics and public finance. Journal of Political Economy, 108(6), 1121 61. Persson, T., Roland, G., & Tabellini, G. (2007). Electoral rules and government spending in parliamentary democracies. Quarterly Journal of Political Science, 2(2), 155 88. Persson, T., & Tabellini, G. (2000). Political Economics Explaining Economic Policy. Cambridge et al.: The MIT Press. Persson, T., & Tabellini, G. (2002). Do Electoral Cycles Differ across Political Systems? Mimeo: Universities of Stockholm and Bocconi. Persson, T., & Tabellini, G. (2003). The Economic Effects of Constitutions. Cambridge, MA: The MIT Press. Persson, T., & Tabellini, G. (2006). Democracy and development. American Economic Review, 96(2), 319 24. Persson, T., & Tabellini, G. (2007). The Growth Effect of Democracy: Is It Heterogeneous and How Can It Be Estimated? CESifo Working Paper, 2016. Pommerehne, W. (1983). Private vs. öffentliche Müllabfuhr: Nochmals betrachtet. Finanzarchiv, 41, 466 75. Pommerehne, W. (1990). The empirical relevance of comparative institutional analysis. European Economic Review, 34, 458 68.

132

References

Plümper, Th., & Martin, Chr. (2003). Democracy, government spending, and economic growth: A political economic explanation of the Barro effect. Public Choice, 117, 27 50. Przeworski, A. (1991). Democracy and the Market Political and Economic Reforms in Eastern Europe and Latin America. Cambridge: Cambridge University Press. Przeworski, A., & Limongi, F. (1993). Political regimes and economic growth. Journal of Economic Perspectives, 7(3), 51 69. Ramseyer, M. (1994). The puzzling (in)dependence of courts: A comparative approach. Journal of Legal Studies, 23(2), 721 47. Rasch, B. E., & Congleton, R. (2006). Constitutional amendment procedures. In: R. Congleton (ed.), Democratic Constitutional Design and Public Policy: Analysis and Evidence. Cambridge: MIT Press, 319 42. Raudla, Ringa (2010). Explaining Constitution Makers’ Preferences: The Cases of Estonia and the United States. Constitutional Political Economy, 21(3):249 69. Rawls, J. (1971). A Theory of Justice. Cambridge: Belknap. Reynolds, A. (2002). The Architecture of Democracy Constitutional Design, Conflict Management, and Democracy. Oxford: Oxford University Press. Riker, W. (1975). Federalism. In: Fred I. Greenstein & Nelson Polsby (eds.), The Handbook of Political Science, Vol. V: Government Institutions and Processes. Reading, MA: Addison Wesley. Riker, W. (1983). Political theory and the art of heresthetics. In: A. Finifter (ed.), Political Science: The State of the Discipline. Washington, DC: American Political Science Association, 47 67. Riker, W. (1984). The heresthetics of constitution making: The presidency in 1787, with comments on determinism and rational choice. American Political Science Review, 78, 1 16. Robinson, J., & Torvik, R. (2016). Endogenous presidentialism. Journal of the European Economic Association, 14(4), 907 42. Rockey, J. (2012). Reconsidering the fiscal effects of constitutions. European Journal of Political Economy, 28(3), 313 23. Rodden, J. (2002). The dilemma of fiscal federalism: Grants and fiscal performance around the world. American Journal of Political Science, 46(3), 670 87. Rodden, J. (2003). Reviving Leviathan: Fiscal federalism and the growth of government. International Organization, 57, 695 729. Rodden, J., & Wibbels, E. (2002). Beyond the fiction of federalism Macroeconomic management in multitiered systems. World Politics, 54, 494 531. Rodrik, D. (2008). Second best Institutions. Mimeo. Downloadable at: http://ksghome.harvard .edu/~drodrik/Second best%20institutions%20paper.pdf. Roll, R., & Talbott, J. (2001). Why Many Developing Countries Just Aren’t. Mimeo: University of California, Los Angeles. Sanz, C. (2017). Direct Democracy and Government Size: Evidence from Spain. Available at ssrnc.com. Sartori, G. (1994). Comparative Constitutional Engineering. New York: New York University Press. Scully, G. (2001). Institutions, policy, and economic growth. In: W. Shughart & L. Razzoli (eds.), The Elgar Companion to Public Choice, Cheltenham et al.: Edward Elgar, 611 27.

References

133

Seabright, P. (1996). Accountability and decentralization in government: An incomplete contracts model. European Economic Review, 40, 61 89. Sen, A. (1999). Development as Freedom. Oxford: Oxford University Press. Shugart, W. & Tollison, R. (1998). Interest groups and the courts. George Mason Law Review, 6, 953 69. Smith, D., & Tolbert, C. (2004). Educated by Initiative. Ann Arbor: University of Michigan Press. Spamann, H. (2015). Empirical comparative law. Annual Review of Law and Social Science, 11, 131 53. Spindler, Z., & de Vanssay, X. (2002). Constitutions and economic freedom: An international comparison. South African Journal of Economics, 70(6), 1135 46. Stadelmann, D., Portmann, M., & Eichenberger, R. (2014). The law of large districts: How district magnitude affects the quality of political representation. European Journal of Political Economy, 35, 128 40. Sutter, D. (1995). Potholes along the transition from authoritarian rule. Journal of Conflict Resolution, 39(1), 110 28. Sutter, D. (2006). The deterrent effects of the International Criminal Court. In: S. Voigt, M. Albert, & D. Schmidtchen (eds.), International Conflict Resolution Conferences on New Political Economy, vol. 23. Sutter, M., Haigner, S., & Kocher, M. (2010). Choosing the carrot or the stick? Endogenous institutional choice in social dilemma situations. Review of Economic Studies online first. Tanzi, V. (2000). Some politically incorrect remarks on decentralization and public finance. In: J. J. Dethier (ed.), Governance, Decentralization and Reform in China, India and Russia. Boston et al.: Kluwer, 47 63. Tavares, J., & Wacziarg, R. (2001). How democracy affects growth. European Economic Review, 45, 1341 78. Tergiman, C. (2015). Institution design and public good provision: An experimental study of the vote of confidence procedure. Experimental Economics, 18(4), 697 717. Testa, C. (2010). Bicameralism and corruption. European Economic Review, 54(2), 181 98. Thatcher, M. (2002). Delegation to independent regulatory agencies: Pressures, functions and contextual mediation. West European Politics, 25(1), 125 47. Thießen, U. (2003a). Fiscal decentralization and economic growth in high income OECD countries. Fiscal Studies, 24(3), 237 74. Thießen, U. (2003b). Fiscal Federalism in Western European and Selected Other Countries: Centralization or Decentralization? What Is Better for Economic Growth. Berlin: DIW. Thomas, M. (2006). What Do the Worldwide Governance Indicators Measure? Mimeo: Johns Hopkins University. Ticchi, D., & Vindigni, A. (2010). Endogenous constitutions. Economic Journal, 120(543), 1 39. Tiebout, Ch. (1956). A pure theory of local expenditures. Journal of Political Economy, 64, 416 24. Treisman, D. (2000). The causes of corruption: A cross national study. Journal of Public Economics, 76, 399 457. Treisman, D. (2007). What have we learned about the causes of corruption from ten years of cross national empirical research? Annual Review of Political Science, 10, 211 44. Tsebelis, G., & Money, J. (1997). Bicameralism. Cambridge: Cambridge University Press.

134

References

Tullock, G. (1974). The Social Dilemma: The Economics of War and Revolution. Blacksburg, VA: University Publications. Tullock, G. (1987). Autocracy. Dordrecht: Kluwer. de Vanssay, X., & Spindler, Z. (1994). Freedom and growth: Do constitutions matter? Public Choice, 78, 359 72. Vaubel, R. (1996). Constitutional safeguards against centralization in federal states: An independent cross section analysis. Constitutional Political Economy, 7(2), 79 102. Vibert, F. (2007). The Rise of the Unelected: Democracy and the New Separation of Powers. Cambridge: Cambridge University Press. Voigt, S. (1997). Positive constitutional economics A survey. Public Choice, 90, 11 53. Voigt, S. (1999). Explaining Constitutional Change A Positive Economics Approach. Cheltenham et al.: Elgar. Voigt, S. (2009). Explaining constitutional garrulity. International Review of Law and Economics, 29, 290 303. Voigt, S. (2015). Veilonomics: On the use and utility of veils in constitutional political economy. In: Louis M. Imbeau & Steve Jacob (eds.), Behind a Veil of Ignorance? Power and Uncertainty in Constitutional Design. Heidelberg: Springer, 9 33. Voigt, S. (2019). Institutional Economics An Introduction. Cambridge: Cambridge University Press. Voigt, S., & Blume, L. (2012). The economic effects of federalism and decentralization A cross country assessment. Public Choice, 151(1 2), 229 54. Voigt, S., & Blume, L. (2015). Does direct democracy make for better citizens? A cautionary warning based on cross country evidence. Constitutional Political Economy, 26(4), 391 420. Voigt, S., & El Bialy, N. (2016). Identifying the determinants of aggregate judicial performance: Taxpayers’ money well spent? European Journal of Law and Economics, 41(2), 283 319. Voigt, S., & Gutmann, J. (2013). Turning cheap talk into economic growth: On the relationship between property rights and judicial independence. Journal of Comparative Economics, 41(1), 66 73. Voigt, S., Gutmann, J., & Feld, L. P. (2015). Economic growth and judicial independence, a dozen years on: Cross country evidence using an updated set of indicators. European Journal of Political Economy, 38, 197 211. Voigt, S., & Kiwit, D. (1998). The role and evolution of beliefs, habits, moral norms, and institutions. In: H. Giersch (Hrsg.), Merits and Limits of Markets. Berlin et al.: Springer, 83 108. Voigt, S., & Salzberger, E. (2002). Choosing not to choose: When politicians choose to delegate powers. Kyklos, 55(2), 247 68. Wagemann, C., & Schneider, C. (2010). Qualitative Comparative Analysis (QCA) and fuzzy sets: Agenda for a research approach and a data analysis technique. Comparative Sociology, 9, 376 96. Weingast, B. R. (1997). The political foundations of democracy and the rule of the law. American Political Science Review, 91(02), 245 63. Wildasin, D. (1997). Externalities and Bailouts: Hard and Soft Budget Constraints in Intergovernmental Fiscal Relations. Nashville, TN: Mimeo.

References

135

Wintrobe, R. (1990). The tinpot and the totalitarian: An economic theory of dictatorship. American Political Science Review, 84(3), 849 72. Wintrobe, R. (1998). The Political Economy of Dictatorship. Cambridge: Cambridge University Press. Wintrobe, R. (2019). Are there types of dictatorship? In: R. Congleton, B. Grofman, & S. Voigt (eds.), The Oxford Handbook of Public Choice. Oxford et al.: Oxford University Press, vol. 2, 286 310. Wright, J. (2008). Do authoritarian institutions constrain? How legislatures affect economic growth and investment. American Journal of Political Science, 52(2), 322 43.

Index access orders limited, 39 open, 39 autocracy, 25 7, 29 32, 36, 38, 41 2, 52, 55, 71 definition, 26 economic theory of, 29 bandits roving, 29 stationary, 29 30 bargaining game, 37 bargaining theory, 95 behavioral neutrality, 13 beliefs, 10, 21, 25 bicameralism, 43, 50, 56, 58 9, 65 Bicameralism, 59 Brennan, 8, 12 13, 49, 54 Buchanan, 1, 4 5, 7 8, 10 11, 13, 57 8, 81, 96 budget deficits, 58 budget institutions, 53 case studies, 3, 15, 17, 71, 89 90, 92 causation, 13 central bank, 72 centralization, 71, 76 checks and balances, 49 commitment capacity, 34 commitment view, 63 common pool, 53, 81 2, 84, 95 common pool problem, 53, 81, 95 competition institutional, 11 systems, 11 competitiveness of local elections, 87 Condorcet’s jury theorem, 85 constitution, 3 6, 9 13, 15, 18 19, 21 2, 24, 28, 32, 38, 46 8, 52 5, 63, 65 6, 68, 70 1, 74, 80, 83, 87, 90, 92 5, 97 103 definition, 5 Constitution as bundle of conventions, 9 coordination device, 101 precommitment device, 8

principal agent relationship, 7 result of an evolutionary process, 10 social contract, 7 constitution making, 5, 18, 55, 90, 92 4, 97 8 constitutional assembly, 94 constitutional change, 37 9, 56, 58, 66, 93, 98 9 constitutional convention, 93 5, 103 constitutional referendums, 97 constitutional review, 63 constitutionalism, 6 constitutions consensual, 87 majoritarian, 87 corruption, 19, 23, 47, 50 1, 53, 58 9, 67, 69, 76, 82 costs decision, 57 external, 57, 96 credible commitments, 28, 38, 62 credibly commitments, 37 8 democracy definition, 26 direct, 15, 17, 43 4, 73 9 as an insurance, 30 democratic competition view, 63 Dicey, 6 difference principle, 16, Siehe Rawls difference in differences, 14, 32 dilemma dictator‘s dilemma, 33 punishment dilemma, 37 dilemma of strong state, 60 dominant coalition, 39 40 Duverger’s law, 81, 86 EBA 54, Siehe extreme bounds analysis, 42 electoral rules, 80 ballot structure, 80 2 district size, 80 1 majority rule, 80 2, 86, 88, 91, 97

137

Index

proportional representation, 18, 80 3, 86, 91 switch from MR to PR, 84 electoral systems Siehe electoral rules, 80 endogenizing constitutional rules, 2 endogenizing constitutions, 2, 90 enforcement, 5, 9, 45, 91 expenditure, 19, 51, 74 5, 78 experiments natural, 14, 90 randomized, 13 extreme bounds analysis, 42 federalism, 23, 43, 50, 59, 63, 65 6, 68 73, 86 cooperative, 66 fiscal, 65 fiscal decentralization, 70 fragmentation, 86 France, 18, 43, 50, 54, 95, 97, 103 franchise, 36, 38, 41, 48, 80, 86 gap de facto, 48, 89 de jure, 48, 100 1 government deficits, 69 government efficiency, 19, 24, 51, 67, 77, 79 government expenditure, 23, 36, 50 1, 59, 69, 71, 82, 84 5 happiness, 20, 77 hereditary rulers, 35 Homo economicus, 12 identification strategies, 90 income inequality, 32, 87 independent decision makers, 72 3 independent regulatory agencies, 72 Index of economic freedom, 46 indirect rule, 70 initiative, 74 9 institutions external, 4, 9, 32, 56, 70, 86, 96 informal, 4, 64, 91, 101 internal, 4 5, 91, 95 insulation, 55 interdependence cost calculus, 96 interdependence costs, 57 judiciary independence, 46, 60 4 de facto, 61, 64, 100 de jure, 61, 64

legitimacy, 19, 28, 93, 97 logrolling, 95 matching, 14, 32 methodological individualism, 5, 10, 12 mixed methods, 17 Montesquieu, 48 Nepal, 95 new institutional economics, 2, 4 non majoritarian institutions, 72 North, 4, 28, 39, 41, 64, 86 oil reserves, 35 omitted variable bias, 14, 20, 90 party system, 34, 67 policy advice, 2, 90 P OPITZ ’s law, 71 pork barrel projects, 81 precommitment, 5, 8, 11 principal agent framework, 74 principal agent problem, 8, 17, 73 4 productivity, 18 19, 23, 47, 51, 58, 68 9, 74, 76 7, 91 2 property rights, 30, 44 5, 60 Rawls, 16 redistribution, 30, 32, 36, 38, 58, 85 referendum, 17, 71, 74 5, 77, 79, 85, 97 regime change, 25, 37 8 regression discontinuity, 14 reversed causality, 14, 78 right to vote, 17, 36, 38, 43, 80, 98 rights basic human, 29, 43 5, 47, 100 civil/political, 43 4 definition, 44 5, 47, 102 economic, 43 5 negative, 44 positive, 22, 44, 47, 101 property right, 30 social/emancipatory, 43 4, 47 selection bias, 14 separation of powers horizontal, 72 vertical, 43, 65, 72 size of the nation, 71 South Africa, 26, 68, 97

138

Index

time inconsistency, 28, 73 time inconsistent preferences, 8, 73 tinpots, 27 8 totalitarians, 27 tradeoff between rigidity and flexibility, 99 transition, 2, 25, 32, 37 42, 71, 84 transmission channel, 18 19, 32, 42, 53, 61, 69, 83, 88, 106 trial and error, 11 tyranny of the majority, 28, 58

Soviet Union, 97 spatial inequality, 70 Spending government spending, 19, 23, 36, 47, 51, 54, 58 9, 68, 71, 75, 77, 81, 83 4, 91 social and welfare spending, 19 20, 23 strong, disciplined parties, 67 supermajority, 57 synthetic control approach, 14 systems parliamentary, 34, 49 51, 53 6, 104 5 presidential, 22, 26, 43, 49 51, 53 6, 82, 88, 91, 100, 104 5

US constitution, 17, 94 universal suffrage, 86

tax revenue, 19 taxes, 8, 19, 32, 47, 50, 73

veil of ignorance, 16 violence, 39, 105