73825 Understanding Policy Change Understanding Policy Change How to Apply Political Economy Concepts in Practice Cristina Corduneanu-Huci Alexander Hamilton Issel Masses Ferrer THE WORLD BANK Washington, D.C. © 2013 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 16 15 14 13 This work is a product of the staff of The World Bank with external contributions. Note that The World Bank does not necessarily own each component of the content included in the work. The World Bank therefore does not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http://creativecommons.org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Cristina Corduneanu-Huci, Alexander Hamilton, and Issel Masses Ferrer. 2013. Understanding Policy Change: How to Apply Political Economy Concepts in Practice. Washington, DC: World Bank. DOI: 10.1596/978-0-8213-9538-7. License: Creative Commons Attribution CC BY 3.0 Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ISBN (paper): 978-0-8213-9538-7 ISBN (electronic): 978-0-8213-9539-4 DOI: 10.1596/978-0-8213-9538-7 Cover design: Debra Naylor, Naylor Design, Inc. Library of Congress Cataloging-in-Publication Data Corduneanu-Huci, Cristina, 1976–  Understanding policy change : how to apply political economy concepts in practice / by Cristina Corduneanu-Huci, Alexander Hamilton, Issel Masses Ferrer.   p. cm.  Includes bibliographical references.  ISBN 978-0-8213-9538-7 — ISBN 978-0-8213-9539-4 (electronic)  1. Policy sciences. 2. Economics. I. Hamilton, Alexander, 1984– II. Ferrer, Issel Masses, 1987– III. Title.  H97.C668 2012  320.6—dc23 2012011964 CONTENTS Foreword xi Acknowledgments xiii Abbreviations xv Introduction 1 What Is This Handbook About? 2 A Guide for Reformers, Journalists, and Civil Society Activists 3 The Main Theoretical Narrative and the Guiding Principle of the Handbook 5 Collective Action: The How To Change Solution 6 The Rest of This Book 14 Summary 15 1.  Political Economy: What It Is and What It Is Not 17 Objectives of Chapter 1 17 Political Economy Analysis, Diagnostics, and Tools 18 Reform Stories: Seeing the World through the Lens of Political Economy 22 Deconstructing and Understanding Poverty Reduction Reform: An Empirical Puzzle 28 The Limits, Perils, and Promises of Political-Economy Analysis 40 Summary 44 PART I 49 2.  Accountability and Corruption: The What Question 51 Objectives of Chapter 2 52 What Is Accountability? 52 Common Symptoms of Malfunctioning Institutions: Corruption and Its Relatives 62 Summary 73 v The Collective Action Problem in Development: 3.  The Why Question 79 Objectives of Chapter 3 81 Collective Action Problems: At the Heart of Development 81 What Are Public Goods? 82 What Factors Inhibit Collective Action? 85 Social Dilemmas of Delivering Public Goods 88 Identifying and Evaluating Collective Action Problems 91 Game Theory and Collective Action: Modeling Social Dilemmas with Nash Equilibria 95 Summary 98 Exercise 3.1: The Unscrupulous Diner’s Dilemma 99 Theories and Mechanisms of Political Economy: 4.  Institutions and Equilibria 111 Objectives of Chapter 4 113 Institutions, Incentives, and Collective Action 113 Institutional Origins, Stability, and Change 131 Symptoms of Institutional Failure: Lack of Enforcement 140 Modeling the Role of Institutions with Game Theory 144 Summary 146 Exercise 4.1: The Prisoner’s Dilemma Game 147 5.  Collective Choice and Agenda Setting 157 Objectives of Chapter 5 159 What Is Agenda Setting? 159 Types of Agenda Setting 163 Agenda Setting as a Strategic Tool for Policy Making: A Note on Positivist and Normative Angles 171 Summary 173 Exercise 5.1: The Agenda-Setting Game 174 Delegation Problems and the Principal-Agent 6.  Relationship 181 Objectives of Chapter 6 183 Delegation 183 The Principal, the Agent, and Their Incentives 186 What Are the Main Technical Features of a Principal-Agent Model? 187 Who or What Will Make Sure That the Agent Complies with the Mandate? Solving Delegation Problems 189 Analyzing Principal-Agent Dilemmas Using Game Theory 193 vi Understanding Policy Change Summary 195 Exercise 6.1: The Principal-Agent Game 196 Causes and Effects of Information Asymmetries in   7.  the Policy-Making Process 207 Objectives of Chapter 7 209 The Political Economy of Information 210 Broader Implications of Information Asymmetries 217 Understanding the Nonlinear Relationship between Information Asymmetries and Pro-Development Collective Action 224 Summary 229  8. Credible Commitment 233 Objectives of Chapter 8 234 The Problem: Time Inconsistency 234 What Is “Credible Commitmentâ€?? 235 Credible Commitment, Collective Action, and Development 241 The Importance of Collective Action–Induced Credible Commitment to Development: What the Data Say 242 Modeling Credible Commitment as a Challenger Game 243 Summary 244   9.  Solutions to Collective Action Problems 247 Objectives of Chapter 9 248 When Do Mobilization and Production of Public Goods Occur? 249 Tools for Solving Collective Action Problems 254 Conceptualizing Leadership, Trust, and Repeated Interaction through Games 264 Summary 266 Exercise 9.1: The Prisoner’s Dilemma with Different Time Horizons/Institutional Incentives Game 267 PART II 279 How to Conduct Political-Economy Analyses: 10.  First Steps 281 Objectives of Chapter 10 282 Who Are the Stakeholders, and How Can Their Preferences Be Mapped? 283 The Role of Institutions 286 The Role of Constraints 287 Contents vii Methods and Instruments for Diagnostics and Analysis 288 From Theory to Action 291 Summary 293 Exercise 10.1: The Leadership Game 294 11.  Recognizing Opportunity for Policy Change 303 Objectives of Chapter 11 304 Time Frames for Reform: Timing, Sequencing, Pacing, and Maintaining Momentum 304 Complementary versus Trade-offs of Different Reform Areas 310 Summary 315 Appendix A: Fundamental Concepts in Game Theory 317 Appendix B: Technical Appendix 321 Appendix C: Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines 323 Appendix D: Political-Economy Homework Exercises 333 Index 345 Box 11.1 Political-Economy Analysis: A Practical Checklist 311 Concepts in Practice 2.1 Democratic Institutions, Famines, and Food Shortages 54 2.2 The Case of Budgetary Decisions in Iran 62 2.3 Prebendalism and Artisanal Gold Mining in the Democratic Republic of Congo 65 2.4 Clientelism, Ethnicity, and Gender in Benin 67 2.5 The “New Public Managementâ€?—from New Zealand to the World 73 3.1 Prices of Agricultural Products in Sub-Saharan Africa 92 3.2 Political Parties as Collective Action Problem Solvers 95 4.1 Constitutions and Individual Rights to Public Goods— Water Access and Health Care 119 4.2 Water Resource Management Reforms in Zimbabwe 120 4.3 Effects of Electoral Laws on Public Goods in Brazil and República Bolivariana de Venezuela 127 4.4 Informal Institutions in China 130 4.5 Labor Regulations in India 133 viii Understanding Policy Change 4.6 The Power of Ideas and Tax Reforms 138 4.7 Neo-Patrimonialism in Africa and Historical Legacies 140 4.8 Lifespan of National Constitutions around the Globe 141 5.1 Optimum Outcomes and Constraints 162 5.2 The Power of Formal Agenda Setting: Using the Presidential Veto to Make Policy in Uruguay 170 5.3 The Power of Informal Agenda Setting: Using the Issue of Slavery to Get Elected 171 5.4 Agenda Manipulation in Competitive Authoritarian Elections 172 6.1 Solving Delegation Problems and Data Collection in the Iraqi Pension System 186 6.2 Can Politicians Directly Control Bureaucrats? Evidence from Tax Agency Reform in Argentina 191 6.3 Principal-Agent and Collective Action 194 7.1 Individual Characteristics and the Working Environment in Bureaucracies 213 7.2 Taming the Bureaucracy through Broadcasting: The Case of Ghana’s Public Account Committee 217 7.3 Feeding the Starving: The Role of the Media in Generating Incentives for Politicians to Respond to Famines and Flood Damage 218 7.4 Using Technology to Monitor the Provision of Public Goods 221 7.5 The Consequences of Information Asymmetries on Development Outcomes 228 8.1 The Rise of Independent Central Banks 237 8.2 The Importance of Veto Players in Policy Making 239 8.3 Collective Action and Credible Commitment in the Dominican Republic 242 8.4 Political Responses to the AIDS Epidemic and Ethnic  Affiliations 243 9.1 Case Study: Social Capital and Development 259 9.2 Leadership, Trust, and Participation in São Tomé and Príncipe 261 9.3 Leaders in the Democratic Republic of Congo 262 10.1 Using a Survey Instrument to Map the Political Economy of Health Care Litigation 291 Figures O.1 Institutional Constraints That Affect Collective Action 12 1.1 Political Economy: The Intersection of Politics and Economics 20 1.2 Implementation Gaps—Spending versus Outcomes in Selected Countries, 1980s and 1990s 29 Contents ix 2.1 Conceptual Map of Accountability, Corruption, and Related Governance Problems 52 2.2 The Path between Citizen Preferences and Policy Outcomes 56 2.3 Varieties of Accountability 58 2.4 The Link between per Capita Income and Corruption 70 3.1 Conceptual Map of Collective Action Problems 80 3.2 Impact of Individual and Collective Incentives on a Public Good 82 4.1 Conceptual Map: Understanding Institutions and Their Influence in Development 112 4.2 Influence of Collective Action on Institutional Change 114 4.3 The Intersection of Formal and Informal Institutions 121 5.1 Agenda-Setting Conceptual Map 158 6.1 Delegation and Principal-Agent Conceptual Map 182 6.2 Decision Making in a Sequential Form Game 195 7.1 Conceptual Map of Information Problems and Asymmetries 209 7.2 Influence of Transparency, through the Media, on Government 223 8.1 Conceptual Map of Credible Commitment 234 9.1 Conceptual Map: Solutions to Collective Action Problems 249 10.1 Conceptual Diagram of Political-Economy Diagnostics 282 C.1 Five-Year Timeline of the Philippines Procurement Reform 325 C.2 Round 1 Net Map: Who Influenced the Passage of the Procurement Reform Law? 328 C.3 Round 2 Net Map: Who Influenced the Passage of the Procurement Reform Law? 331 Tables 2.1 Modes of Accountability and Time Horizon for Performance Evaluation 56 3.1 Goods Quadrant—Public Goods versus Club (Toll) Goods, Common-Pool Resources, and Private Goods 83 4.1 Political Incentives for Development and Collective Action Opportunities Induced by Formal Institutions 123 4.2 Incentives of Informal Institutions and Their Impact on Collective Action for Development 132 B11.1 Comparison of Stakeholder Preferences and Resources 312 B11.2 Information Imperfections: Costs and Benefits of Correction 313 B11.3 Mechanisms, Priorities, and Solutions 313 B11.4 Institutional Stability and Agenda-Setting Opportunities 314 A.1 An Illustration of Payoffs 319 C.1 Stakeholder Map for the Philippines Procurement Reform 329 x Understanding Policy Change FOREWORD This book is the outcome of a rather unique conjunction of circumstances that prevailed at the World Bank Institute in Washington, DC, in the years from 2008 to 2012. There and then, a talented group of young staff in the Governance and Leadership Unit, working in the atmosphere of creativity and the supportive authorizing environment provided by the renewal the World Bank Institute was undergoing, began putting their minds together to explore a number of questions few others at the World Bank had addressed. What are the types of constraints that conspire to derail otherwise sen­ sible policies? When can these constraints prevent a reform process or a project from achieving its declared objectives? How can these constraints to change be unlocked? Under what conditions do individuals and organiza- tions engage in collective action, and when does this lead to the sort of last- ing institutional change that is conducive to sustainable development? These were the sorts of questions the team probed tirelessly over three years. In line with the World Bank Institute’s global learning and training man- date, this book was initially conceived as an instructor’s guide, with games and exercises, and addressed to those charged with equiping practitioners with the analytical tools and practical exercises to navigate the treacherous and often murky waters of policy reform. Upon closer scrutiny, it dawned upon the team that few books explain these tools to lay readers, and even fewer had ventured to jump the fence and reach out to practitioners. It is to attempt to address these gaps that the idea of this book in its present form first emerged. Policy making does not occur in a vacuum; rather, it takes place amidst complex political and social environments in which different stakeholders, operating within a changing institutional context, interact strategically as they pursue conflicting agendas. Here, political economy analysis (PEA) xi ­ rovides a set of powerful tools to diagnose and analyze these factors and p evaluate the feasible space for policy change when designing projects for an optimal chance of success. Therefore, by providing a systematic way of in- corporating the impact of such factors into a project’s benefit–cost analysis, PEA enables a more realistic appraisal of what reformers can hope to achieve in practice. This volume is intended to provide policy makers, civil society activists, students, and bureaucrats alike with a comprehensive introduction to the core PEA concepts and building blocks, including collective action, in- stitutional analysis, agenda setting, information asymmetries, principal- agent dynamics, credible commitment, leadership, and coalition building. By using real world examples and case studies, as well as a basic game theoretic exposition of key concepts, it builds for readers a step-by-step understanding of how political and social factors may work to shape incen- tives and facilitate or impede collective action. In addition, several chap- ters provide a practical guide on how to use the portable tools of PEA in specific policy contexts. This book aims to develop a basis for more suc- cessful project identification, feasibility assessment, and realization. We hope it has achieved its objectives. J. Edgardo Campos Edouard Al-Dahdah Practice Manager Senior Economist World Bank Institute World Bank Institute xii Understanding Policy Change ACKNOWLEDGMENTS This book is a product of the World Bank Institute’s Governance and Lead- ership Unit, headed by J. Edgardo Campos. The project was housed with the Political Economy Team under the supervision of Edouard Al-Dahdah. The main authors of the book are Cristina Corduneanu-Huci, Alexander Hamilton, and Issel Masses Ferrer.1 The manuscript benefited greatly from the immensely helpful input provided by its formal reviewers, Philip Keefer, Margaret Levi, and Nimah Mazaheri. Skillful editorial support was provided by Hilary Hendricks and Dana Lane. Yianna Vovides provided critical guidance for the pedagogical ele- ments of this book. Stephen McGroarty, Rick Ludwick, and Andres Meneses ushered the project to completion. Furthermore, the book benefited from the work of Audrey Sacks who drafted an early outline of this project, as well as from the helpful feedback of Antonio Lambino. Participants at a training event organized by World Bank Institute and the Ghana Institute of Management and Public Adminis- tration (GIMPA), held in Accra in January 2012, offered useful comments on the theoretical and pedagogical approaches of this project. We are also grate- ful for the support provided by Brett Beasley, Monali Chowdhurie-Aziz, Craig Hammer, Hirut McLeod, Jeesun Lee, Ceren Ozer, Ajay Tejasvi, and Kay Winning. Finally, we would like to thank Ray Duch, Keith Poole, Cambridge Uni- versity Press, The University of Chicago Press, Oxford University Press, and Prince­ton University Press for their kind permission to reproduce citations and/or diagrams. The authors can be contacted at: cc97@duke.edu, a.hamilton@oxon.org, im3050a@american. 1 edu. xiii ABBREVIATIONS ANC African National Congress CDF Constituency Development Fund DRC Democratic Republic of Congo GDP gross domestic product LCA logic of collective action MP member of parliament NGO nongovernmental organization NPM new public management PEA political-economic analysis PAC public account committee PD the prisoner’s dilemma PR proportional representation PRI the Institutional Revolutionary Party of Mexico RTC regional transport committee TC the tragedy of the commons xv Introduction The introductory chapter sets the stage and outlines the logic of the rest of the handbook. First, we present the main learning objectives; second, we in- troduce the pedagogical approach, methodology, and structure of the book. This handbook is intended to introduce the concepts of political economy to a wide audience of development practitioners, including civil society ac- tivists, journalists, students, and bureaucrats. Since the target readers vary widely in their previous exposure to the subject matter, the book summa- rizes a vast academic field and presents a comprehensive repertoire of concepts, theories, and empirical examples. Rather than offering a “do-it-­ ­ yourselfâ€? framework, we opted for developing a step-by-step analytical puzzle. First, we introduce the core mechanisms of political economy and their inner logic, and, subsequently, we help our readers learn how to recog- nize these mechanisms in their daily development-related work. By the end of the book, we hope that readers will be able to:  ecognize core development problems stemming from the political- • R economic environment Link theoretical concepts to real-life situations •  • Diagnose the symptoms and the root causes of malfunctions • Understand the short-term and long-term consequences of poor gover- nance and low institutional equilibria 1 • Grasp the intersections of collective action, types of political institutions, the incentives they generate, and development outcomes • Recommend possible solutions for pro-development collective action and reform. In addition, readers will explore why political-economy analysis is a valuable tool for assessing the underlying causes of poor governance in a more nu- anced way and learn what it actually entails, as well as its many virtues and pitfalls. This handbook is also designed to provide trainers with some of the peda- gogical materials they need to develop an introductory course on political- economy analysis for policy practitioners. The content focuses on the what, the why, and the how to of policy change. The readers or trainees will en- counter key theories and concepts and learn how to apply the analysis to an understanding of their own policy-making environment. Pedagogically, the handbook uses interactive classroom exercises and the case study method to reinforce learning objectives and to capture the concepts, methods, experiences, and challenges relevant for practitioners. Structured learning activities at the end of most chapters and a comprehen- sive group exercise in appendix D will also give readers and trainers the op- portunity to apply the knowledge and tools of political economy to simulated or specific development puzzles. What Is This Handbook About? This book explains and illustrates how incentives shape the decisions of pol- icy makers. These decisions determine (intentionally or unintentionally) who gets what, when, and how. Therefore, being able to identify the sources and intensity of such incentives can help explain development policy out- comes. Here are some questions that political economy can help us answer:  hy do some states promptly initiate and implement relatively healthy • W economic reforms, while others opt for policies that are blatantly ineffi- cient from the point of view of development? What factors explain policy and institutional change and continuity? •  How do institutions shape the content, feasibility, timing, and pace of nec- •  essary economic reforms? Why is economic development stalled in many contexts? •  How can large shares of non-tax revenue derived from natural resources •  or foreign aid be harnessed to promote development instead of hamper- ing it? 2 Understanding Policy Change  hy is tax revenue in many developing countries so low? • W What explains the divergent paths of states at similar stages of develop- •  ment with respect to their ability to reduce poverty? What role do collective action and leadership play in development? •  How do incentives at work in different political systems shape govern- •  ment responsiveness to natural calamities such as famines, floods, or earthquakes? Why do bad or wasteful policies persist even though it is technically fea- •  sible to adopt a better modus operandi? Do bad policies endure because of voter ignorance or because of the •  power of entrenched interest groups? What are the opportunities for effective pro-development collective •  action? Will providing a development loan to country x facilitate development •  because policy makers have strong incentives to better the lives of mar- ginalized groups, or will the loan be diverted to the bank accounts of the politically connected? Identifying how the incentives of politicians and bureaucrats emerge, evolve, and differ across policy-making contexts and time requires an under- standing of some basic tools and concepts. These heuristic devices can help practitioners break down a complex environment into manageable analyti- cal components and think with greater nuance about the development issues they encounter in their daily work. Beyond standard diagnostics, conventional Western-centric primers and textbooks on political economy often fail to capture many context-specific innovations. We attempted to remedy or bypass this problem by incorporat- ing numerous examples of problems and solutions from around the world. Readers or trainees can also contribute to enriching the analytical repertoire of solutions to basic political-economic problems and dilemmas encoun- tered in their specific contexts. A Guide for Reformers, Journalists, and Civil Society Activists Despite their inherent limitations and pitfalls, political-economy diagnostics are recognized as useful for answering the what and the why of development (that is, What are the central development problems? Why do they persist?) and for guiding the how to change. In a landscape of challenges and obstacles, a map of all possible routes is the first piece of navigational equipment needed. Introduction 3 Our approach to political economy attempts to be pragmatic and aims to allow enough flexibility for readers to fill in the blanks of the conceptual building blocks with their own development work experience or to use the analytical tools for concrete problems and contexts. To achieve this goal, we survey many theories and empirical findings in the fields of political science and development economics. In doing so, we hope to avoid the really big questions and big answers as much as we can and yet still capture the logic of quite abstract concepts such as collective action and institutions. The examples are quite heterogeneous, gathered from a variety of studies dealing with development and reform. ­ The approach is more didactic than meta-theoretical and seeks to accom- plish four things:  irst, we introduce a vast theoretical and empirical literature for read- • F ers who have had little or no previous exposure to political economy. Therefore, the text errs on the side of comprehensiveness and provides a “one-stop shopâ€? for relevant concepts and tools. Needless to say, like any survey book, it cannot cover individual components in depth. In an attempt to compensate for this shortcoming, the reference list at the end of each chapter offers in-depth readings on the specific topics covered. Second, we present analytical devices that practitioners can use in ad- •  dressing their own questions related to development issues. These con- cepts are also accompanied by blank do-it-yourself templates designed to help readers think analytically and systematically about the actors, insti- tutions, and constraints with which they deal.   Games and class activities at the end of most chapters are meant to il- lustrate the concepts, to entertain, and to make analytical thinking in gen- eral more fun and interactive. Third, the case study we use in appendix C gets into the specifics of the •  “how-toâ€? of reform, once the analysis is done and the problem clear. Most handbooks on political economy address big questions, and their answers focus on the what and the why. Why does the problem persist? What works? We try to go beyond the why and the what into the everyday de- tails of the how to navigate the tricky terrain of policy change. Readers will be guided through the story of the 2002 public procurement reform in the Philippines. Finally, appendix D provides the materials necessary for a group activity •  designed to give readers or trainees the feel of a real diagnostic produced through team work. 4 Understanding Policy Change The Main Theoretical Narrative and the Guiding Principle of the Handbook To what extent do political incentives interact with, undermine, or bolster the economic and administrative targets of policy reforms? The main cul- prits—corruption, poor governance, and pervasive patron-client networks— hamper and derail development in many parts of the world. Sometimes, these pathologies are simply a “second best,â€? either because they are deeply embedded in complex systems of social relations or because attempts to dis- mantle them might cause more harm and disruption than correction. In con- trast, good governance entails political accountability, the capacity of the government to design and implement development policies demanded by the citizens, and the commitment of key political actors to the rules of the game.1 The first two chapters will ponder the various manifestations of de- velopment failure, or the what, in greater detail. From the what, we will then move to the why and the how of change. Why do development failures persist? Our anchoring stories will suggest a list of potential explanations. Without mechanisms of accountability that align good policies with good politics, bureaucrats and politicians have incentives to satisfy only narrow segments of the electorate, to build patronage ma- chines, and to divert public spending to further their own careers instead of boosting literacy and saving lives. Why do reforms often fail? How can change occur? Is there a repertoire of solutions for diagnosed problems? Last but not least, we address a perennial question, What kinds of incentives can align good policies with good politics? We will isolate theoretical mechanisms and walk through their logic with the help of game theory, a methodological tool- kit developed at the intersection of economics and decision sciences. From diagnostics, the handbook will break down real-world examples into analytical pieces, generalize and abstract from their specific context, and try to guide readers to the bare-bones theoretical mechanisms of causa- tion. This process entails distilling the core political-economic logic from the thick description of numerous details surrounding specific cases (Teune and Przeworski 1970). Accordingly, from concrete examples of electoral rules that give incentives to Honduran or Brazilian politicians to obey party discipline, we will move up a level of abstraction to institutions, or the rules of the game, and explore what they are, what they do to behavior, why they persist, and when they change. The independent supreme audit institutions and central banks from our vignettes are just outward manifestations of and labels for broader credible- commitment mechanisms through which politicians signal to voters that Introduction 5 they mean business and take their mandate seriously. Politics, by definition, implies delegation of mandates. Voters delegate decision making to their po- litical representatives, and these, in turn, delegate policy implementation tasks to bureaucracies. In decentralization reforms, central states give away power and resources to lower-tier governments. We will take this whole pro- cess apart, call it a principal-agent relationship, and deconstruct it into ana- lytical pieces. A major obstacle to pro-development change is the large discrepancy in information among voters, politicians, and bureaucracies. This information gap is an intrinsic part of a principal-agent mechanism. Voters are often uninformed about elections, policies, institutions, or the existence of other voters, just like them, who share exactly the same policy concerns. Legisla- tures are often left in the dark when the annual budget proposal coming from the powerful offices of the executive has only 10 pages or excludes half the actual spending. We will call these discrepancies information asymme- tries. If they are large enough, they give rise to moral hazard problems: the executive who conceals information has incentives to go against voters’ preferences. Finally, we will look at how acting together and mobilizing effectively for better development outcomes can make change happen. Participating in city budgeting in Porto Allegre, Brazil; increasing electricity access in Uttar Pradesh, India; reforming corrupt public procurement in the Philippines; or building successful irrigation systems at the grass roots in Nepal are all cases of costly efforts to change the rules of the game. We will call this process collective action and explore the theoretical obstacles against it and some of ­ the conditions under which it can happen against all odds of success. Collective Action: The How To Change Solution Collective action is the organizing principle of this narrative and our answer to the how question. To paraphrase Nobel Prize winner Elinor Ostrom, col- lective action is the very heart of development. Pro-development reforms entail moving from a “badâ€? to a “goodâ€? equilib- rium, or changing the rules of the game if they generate suboptimal out- comes. This path is not really controversial, as the importance of institutions has long been recognized. Changing institutions that produce the wrong kind of incentives requires the concerted action and cooperation of dedi- cated individuals and organizations. In this sense, understanding when and under what conditions collective action occurs—despite seemingly insur- mountable challenges—gives us the magic key to the how of change. 6 Understanding Policy Change There are many modes of collective action or ways in which citizens come together to influence public policies:  olitical parties aggregate preferences of voters and aim to participate in • P decision making. Interest groups represent industries and causes, raise funds, and lobby •  politicians to obtain favorable policy outcomes for their constituencies. Social movements amass large numbers of citizens to protest policies per- •  ceived as detrimental to their welfare. Reform teams like the one featured in the Philippines case study build •  networks of organizations and influential politicians to push institutional change. Collective arrangements (cooperatives, irrigation systems, and the like) at •  the grass roots fill the role of a missing or malfunctioning state and deliver public services. Some of these groups join forces in networks, platforms, and coalitions that increase their power and influence vis-à-vis the opposed vested inter- ests. Many political parties have to participate in coalition governments or form broad electoral coalitions. Citizens and civil society organizations often build platforms for social accountability. Others either work alone or are un- able to sustain fragile coalitions within which many divergent policy direc- tions cannot be reconciled. As the sequence of events in the Philippines case study will show (see ap- pendix C), reform also goes through many stages over its life cycle. Between initiation and consolidation, the main protagonists and the ties among them might even change considerably. More certain is the fact that consolidation of policy gains requires higher standards of commitment from stakeholders than initiation (Haggard and Kaufman 1995, 9–11). If, under certain condi- tions, a single organization or individual can spark change, reform consoli- dation requires coalition building, as well as the cooptation or—more chal- lenging—the re-creation of various societal interests in the process. To reach this point, successful reformers must solve several problems simultaneously: sustaining collective action, mediating distributional conflicts, and building political credibility (Haggard 1990; Waterbury 1989; Haggard and Kaufman 1995; Crisp 2000). Juggling so many challenges is an art that requires rare leadership skills. When we think of collective action, probably because of the emphasis on action, we tend to associate it with change, reform, and institutional shifts. Several chapters will argue that these are only one part of the story. Routine collective action capabilities also ensure that the “goodâ€? institutions that align good politics with good policies work as they are supposed to. For Introduction 7 ­ xample, in democracies, elections do less to discipline political leaders if e voters cannot act collectively and effectively to punish them because of lack of information, skewed electoral laws, malapportionment, gerrymandering, or ethnic fractionalization. Politicians do not have incentives to make credi- ble commitments not to overspend before elections if groups of citizens can- not monitor and collectively sanction such behavior. Normatively, for pro-development policy, not all forms of successful col- lective action are benign. In fact, more often than not, necessary reforms are blocked by groups that have the resources and power to coalesce, whereas the interests and preferences of the majority (that is, the poor) remain latent and are never articulated. In public procurement processes, as we will see, influential private sector companies can either lobby government agencies effectively or bribe bureaucrats. As a result, under tight budgets and a poor regulatory environment, the favorable contracts they obtain come at the ex- pense of basic public services. Political and administrative networks of cor- ruption and patronage are also forms of very well solved collective action problems that often block change and severely stall development. They are hard or impossible to dismantle, even when domestic reformers and interna- tional donors have considerable organizational capacity and are able to mount concerted action. Needless to say, political parties and social move- ments are not always benign either. Sometimes, their leaders rally support on radical political platforms, fuel ethnic tensions, or initiate conflicts to di- vert attention away from their own inability to boost development and alle- viate poverty. With these caveats in mind, the handbook will deal primarily with those types or modes of collective action that relate to economic devel- opment policies. Collective Action and Institutions: Coming Full Circle The rules of the political game generate incentives that often undermine the technical targets of reforms. Unfortunately from a theoretical perspective, the same rules or institutions that lock in the wrong incentives for develop- ment policies are also often responsible for obstructing pro-reform collec- tive action, that is, the ability of groups and individuals with shared agendas to work together. Political institutions are about collective choice and preference aggrega- tion. They add up, combine, and filter the ideal policy goals of voters or stakeholders to produce concrete outcomes. They come into existence to solve collective action problems in the first place. By definition, then, they 8 Understanding Policy Change arbitrate power relations among different groups in society. However, by solving the motivational and informational problems that undermine coop- eration between individuals and groups, they also create the incentives that facilitate certain types of collective action but not others. These might or might not always be favorable for development. In some autocratic states, the poor are completely eliminated from policy making because they lack all channels of collective action and recourse for their complaints. Political parties and trade unions are banned, and protests suppressed. Martial law often legally forbids “publicâ€? gatherings of more than five people. To give an extreme example, in 1979, in Bangui in the ­ Central African Empire, food riots were severely repressed, and the head of state participated directly in the beating deaths of some protesters. Several months later, around 100 schoolchildren protested against a mandate to buy expensive school uniforms from the government, featuring the picture of no other than Jean Badel Bokassa, emperor of the country. Many of the students were tortured, killed, and, allegedly, eaten.2 During his public trial later that year, Bokassa faced allegations of cannibalism. In contrast, democracies allow political parties, labor unions, and popular protests against unpopular government policies. However, these rights do not guarantee that the poor can collectively and effectively mobilize and pursue policies that alleviate poverty. None of the political parties might put forward pro-poor platforms, the electoral rules might be unfavorable to the geographical representation of the poor, the parties in government might need to make policy concessions to their coalition partners, or the legislature might not even be politically relevant in the presence of a powerful execu- tive. By the same token, sometimes, in very unfavorable environments, grass- roots communities find internal solutions to their collective action problems and organize successful governance arrangements for dealing with scarce resources. Information, monitoring and sanctioning, commitment, leader- ship, and clear rules of engagement all interact with the overall institution (say, democracy or federalism) to lead to collective action—or not. Thus, institutions or rules of interaction, on the one hand, arise to solve collective action problems and induce people to cooperate. On the other hand, sustained joint action of citizens and groups can sometimes lead to major pro-development changes in such rules and dismantle forms of coop- eration that blocked them in the past from joining forces. This circular rela- tionship between the rules of the game and collective action is also the source of a major theoretical challenge. If institutions structure a certain type of actor interaction and political incentives to pursue development or Introduction 9 not but are also simultaneously responsible for collective action opportuni- ties, the amount of information, the resource endowments of actors, or the technologies of political commitment that politicians have, how can change happen at all? There are no easy answers. Sometimes, as we will see in chapter 4, institutional change and major equilibrium shifts occur in the total absence of collective action (that is, in the aftermath of major economic or political shocks such as war, natural calamities, or dramatic changes in world prices of goods and services). At ­ other times, internally sparked collective action drives change and overhauls long-established institutions whose stability has been taken for granted. The 1989 post-Communist transitions and the Arab Spring are powerful and inspirational examples of this type of change. Uppercase INSTITUTIONS and lowercase institutions, a distinction re- cently made by Abhijit Banerjee and Esther Duflo (2011, 243), might have different susceptibilities to change through collective action. The former re- fer to the fundamental rules of society (democracy, federalism, types of checks and balances), whereas the latter are more granular manifestations of institutions (local norms of interaction, specific electoral rules, and so on). Even if the former rarely change except in unusual conditions, the latter are sometimes much easier to dismantle or replace. Development reforms, or positive policy steps, can happen in surprising circumstances. For example, despite the fact that the Saudi Arabian Majlis (legislature) lacks both author- ity over the expenditure side of the annual budget and institutional power, it can veto tax bills. When the government wanted to impose an income tax on foreign workers, the members of parliament (MPs) and the businesses reli- ant on foreign labor mobilized and successfully blocked the reform. In some cases, the mere availability of information, or the correction of information asymmetries, has sparked popular collective action and changed overnight the INSTITUTIONS that skewed development outcomes. In Peru, it took only one “bribe-freeâ€? television station to show a video of ­ Vladimir Montesinos—the powerful head of the Peruvian secret police and right hand of President Fujimori—bribing an opposition MP to switch to F ­ ujimori’s party, to bring people to the streets and take down a very corrupt regime. For pro-development reform, collective action is often the engine as well as the source of hope. Even if INSTITUTIONS and the power relations that underpin them rarely change, if the institutions at least can be transformed to align good politics with some good policies, reformers will have achieved their goals. Chapter 4 will continue the conversation about the relationship between collective action and institutions, the two major theoretical pillars of this handbook. 10 Understanding Policy Change Besides this general debate, individual chapters will go more deeply into the specific constraints of the institutional environment that shape the pos- sibilities, as well as the aspirations, of participants in collective action. Without information, reformers cannot come together (the problem of in- formation asymmetries). When communities cannot monitor their mem- bers to prevent lack of cooperation or overconsumption of a public good or when politicians go against the wishes of the voters (principal-agent prob- lems), collective action is not effective. If political leaders lack technologies of commitment through which they can signal to voters and citizens that they will respect their preelectoral promises or if those leaders are weak, uncertainty about the policy environment persists, and underinvestment is the result, unless collective action redresses the problem. On a different scale, leaders of organizations such as political parties and nongovernmen- tal organizations facilitate joint efforts when they manage to signal to their rank and file that they will not break their promises. Often, pro-­development collective action is either hampered or facilitated by institutional manipula- tion or agenda setting. In the Philippines procurement case, for example, the reform team greatly benefited from the actions of an influential agenda- setting MP, who managed to pass an important bill in the legislature on the day when many opposition MPs were absent. Agenda setting, which by defi- nition solves problems of preference aggregation, also raises concerns. In certain circumstances, even if the poor are able to act collectively (through, say, an effective pro-poor party represented in the legislature), institutional manipulation can make their collective action irrelevant and instead push legislation that avoids sensitive development issues altogether. The chap- ters of the handbook will follow the conceptual map illustrated by figure 0.1, and explore how these four institutional constraints facilitate or inhibit joint efforts for policy change. Most of the theoretical mechanisms will be illustrated with the help of game theory, a methodological field developed at the intersection of eco- nomics and decision sciences that captures best the strategic interaction of actors or stakeholders. Game Theory and the Dynamics of Political Economy By its very nature, decision making is usually not just about what policy mak- ers think and want but also about how they expect other stakeholders (or actors) to react to their choices. Because of the inherently strategic nature of policy making, it is useful to include some of the basic tools of game theory to make sense of the strategic environment. Without being overly formal, most chapters of the book conclude with a discussion of how the key tools Introduction 11 FIGURE O.1  Institutional Constraints That Affect Collective Action Information asymmetries Principal-agent Collective Credible relationships commitment action Institutional manipulation/ agenda setting Source: Authors. and concepts of political economy work in situations of strategic interaction. Therefore, by illustrating them intuitively, we hope that the reader will iden- tify real-world situations in which a similar logic applies. In order to understand social and political dynamics in terms of a game, we need to learn to use a few simple tools and concepts that will be further expanded and refined as the chapters progress. Appendix A is an essential starting kit, and readers should get acquainted with its basic concepts in or- der to follow the games and interactive exercises presented in subsequent chapters. Appendix B introduces the more technical definitions and solution concepts of game theory. Besides interactive exercises, our case study, detailed in appendix C but referred to throughout the handbook, will give the reader the feel of a real reform that demonstrates the power of collective action for reaching better development outcomes. 12 Understanding Policy Change Procurement Reforms in the Philippines: A Case Study in the Political Economy of Change The processes through which government agencies purchase goods and ser- vices from the private sector are plagued by severe corruption in many coun- tries. It is estimated that globally some US$400 billion is lost in bribery every year, at the expense of about 20–25 percent of all costs incurred by govern- ments (Ware et al. 2007, 295–96). Because public procurement lies at the interface of politics, bureaucracies, and private companies and involves high financial stakes, attempts to reform the system are daunting, if not often fu- tile. Government contracts (especially for large infrastructure projects) are a lucrative source of commissions, bribes, illegal transactions, and collusion among various stakeholders. Often, these funds supplement the regular in- come of bureaucrats, are used for political campaigns, and fuel patronage machines. Despite being imperative for better development outcomes, changing the rules of the game through new regulations and better imple- mentation is rarely successful. Powerful vested interests opposing reform are likely to be responsible for a vast cemetery of aborted attempts. The Philippines case gives us a unique chronological log, or a reform di- ary, of many events, institutions, incentives, and individuals involved in the attempt to reform procurement. The aim of this convoluted journey was to overhaul a system that lost about US$48 billion to corruption over a 20-year period. The case is particularly fascinating for two reasons. First, it is a mini- laboratory of political-economy analysis, narrating the stories of two reform attempts, one of which was prematurely halted and the other successful. Comparing the configurations of actors, institutions, time sequencing, and pacing of the two episodes clarifies the factors that made or broke these ef- forts. Second, the case study was codeveloped by one of the reformers, mak- ing it a primary source of firsthand information from a participant-observer. Setting Up the Case Study. The text discusses the analytical components that facilitate or hinder collective action for change. Appendix C summar­ izes the main events and provides a detailed chronology. We hope that the broad outlines of the story will set the stage for analysis. What institutions were relevant for reformers? Who were the key actors? Did agenda setting push the reform bill through the legislature? As with all case studies, reality is complex and multidimensional, never quite fitting nicely into our constricted analytical models. The danger of breaking the story piecemeal into components is that our understanding of the final outcome is overdetermined. Simply put, there are too many factors Introduction 13 at too many points in time responsible for the initial failure and later success of the procurement reform. How do we know which one was crucial? Were the politicians’ incentives to stop the bill before an upcoming election more important than agenda setting in legislative committees? Did the electoral rule according to which representatives were elected skew incentives more than interest group politics? To clarify, we are not presenting a theoretical explanation but are rather taking the reader along on a journey through a complex world, with adaptive processes in which every potential analytical mechanism contributed some- thing, no matter how marginally, to the final reform outcomes. In this sense, our take on the case study approaches the “analytical narrativeâ€? method used in social science research to combine the factual richness of the case with a rigorous understanding of the analytics underpinning it and the use of game theory (Bates et al. 1998). The Rest of This Book The structure of this book is designed so that readers are walked through the different tools and concepts of political economy that can be used to explain policy outcomes. Chapter 2 examines how accountability (and the lack thereof ) and general governance issues may help explain variations in the degree of corruption, political clientelism, and waste in policy making. This is the what question of development, as it reveals pathologies and symptoms of failure and malfunctions. Chapter 3 introduces readers to collective action problems. As we have argued, the existence of coordination and joint efforts or their absence is absolutely central to understanding and enacting change. The ability of stakeholders to turn their aspirations into actual policy means that collective action problems have to be solved. Of course, to identify which stakeholders are more likely to be effective in incentivizing policy makers in a specific time or place, we must be able to map the institutional context; that is the objective of chapter 4. Because in- stitutions determine the rules of the game, they are critical to knowing which stakeholders enjoy more or less power than they would in a different institu- tional setting. Thus, chapters 3 and 4 explore the why. Then we move into the how, which is the conceptual territory of change. After the consequences of collective action problems have been identified and how to map the institutional context has been explained, chapter 5 brings the two together by demonstrating how the (institutional) power of agenda setting can enable some stakeholders to exercise disproportionate 14 Understanding Policy Change policy-making power. This discussion leads naturally into chapter 6, which introduces the principal-agent framework as a useful tool for analyzing how information asymmetries, monitoring, sanctioning, and other delegation di- lemmas may work to amplify or check power and hence create the effective incentives that different policy makers face. Chapter 7 is more generally concerned with exploring the existence of political-economic market imperfections and how these can enable some stakeholders or policy makers to exercise more influence than others. The role of information in the policy process is central. Chapter 8 then introduces credible commitment and how this concept can be used to understand the conditions under which it might be possible for principals (voters, elected officials, and so forth) to incentivize agents (elected officials, bureaucrats) to comply with their agenda rather than engage in blatant corruption or rent extraction. The conceptual section of ­ the book ends with chapter 9, which comes full circle and provides a ­ comprehensive introduction to solutions to collective action problems and how these can be realized more or less effectively in different institutional contexts. The second major section of the book, chapters 10 and 11, is concerned with providing readers with the practical ability to use these problem- solving tools in their own projects. Through a generic political-economy checklist, the chapters show how stakeholder preferences and power (determined by their potential for collective action) and institutional ­ ­ constraints (agenda-setting power, delegation issues, information discrep- ancies, and credible commitment) can enable reformers to map the policy- making process. To help make the material in the book more accessible, each chapter be- gins with an introductory section that includes a discussion of how and why the analytical concept highlighted in the chapter affects development out- comes. A game theory section at the end of each chapter provides readers with the tools and exercises for rigorous analysis. Summary The diagnostics and tools of political economy help focus analyses on the actors, their potential for collective action, the costs and benefits of reform, and the relevant institutions and incentives. They also provide a navigational compass for reformers. Political-economy analysis helps explain why subop- timal development outcomes occur. Introduction 15 Note 1. See Kaufman, Kraay, and Zoido-Lobatón (1999). 2. According to Amnesty International’s investigations of this episode, it seems that the public rumors and allegations were more extreme than what actually happened. It is beyond doubt, though, that most of the students died by either beating or suffocation (Borgenicht and Regan 2008). References Banerjee, Abhijit, and Esther Duflo. 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. New York: Public Affairs. Bates, R., A. Greif, M. Levi, J. L. Rosenthal, and B. Weingast. 1998. Analytic Narra- tives. Princeton, NJ: Princeton University Press. Borgenicht, D., and T. Regan. 2008. Worst-Case Scenario Almanac Politics. San Francisco: Chronicle Books Llc. Crisp, Brian F. 2000. Democratic Institutional Design: The Powers and Incentives of Venezuelan Politicians and Interest Groups. Palo Alto, CA: Stanford University Press. Haggard, Stephen. 1990. Pathways from the Periphery: The Politics of Growth in the Newly Industrializing Countries. Ithaca, NY: Cornell University Press. Haggard, Stephen, and Robert Kaufman. 1995. The Political Economy of Democratic Transitions. Princeton, NJ: Princeton University Press. Kaufmann, D., A. Kraay, and P. Zoido-Lobatón. 1999. “Governance Matters.â€? Policy Research Working Paper 2196. Washington, DC: World Bank. Teune, Henry, and Adam Przeworski. 1970. The Logic of Comparative Social Inquiry. New York: John Wiley & Sons. Ware, Glenn, Shaun Moss, Ed Campos, and Gregory Noone. 2007. “Corruption in Public Procurement: A Perennial Challenge.â€? In The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level, ed. J. Edgar Campos and Sanjay Pradhan, 295–334. Washington, DC: World Bank. Waterbury, John. 1989. “The Political Management of Economic Adjustment and Reform.â€? In Fragile Coalitions: The Politics of Economic Adjustment, ed. Joan Nelson, 39–56. New Brunswick, NJ: Transaction Books. 16 Understanding Policy Change CHAPTER 1 Political Economy: What It Is and What It Is Not In this chapter, we begin to define political economy and look at how it seeks to explain development outcomes by identifying the incentives of the actors and the context in which they make decisions and interact strategically. Objectives of Chapter 1 By the end of this chapter, readers should have a clear map or guide to the main concepts, mechanisms, and theories related to political economy that will be covered in this handbook. They should be able to do the following:  efine political economy and recognize its products • D • Understand the value that political economy can add to the design and implementation of reforms • Clarify what political-economy analysis entails and what it does not • Be aware of the main virtues and shortcomings of the approach • Critically evaluate analytical and practical alternatives. 17 Political-Economy Analysis, Diagnostics, and Tools The powerful role that governance and political factors play in a country’s development path is increasingly recognized. Policies in sectors as diverse as natural resource management, transportation, procurement, and urban planning, to name just a few, are continuously shaped by the interaction among governments, civil society actors, the private sector, and organized citizens. An analytical approach to governance and the political-economic environment that accounts for these complex interactions is essential to un- derstanding root problems, why they persist, and how they can be changed. Uncovering development traps requires knowing the policy landscape well and identifying the obstacles to reform before navigating it. Political Economy: The Meeting of Good Politics and Good Policies Political economy is an established academic field that studies the intersec- tion (or the relationship) between politics and economics. Its first parent, politics, is in no need of introduction. We all recognize it in our everyday lives. It affects us, we love it, we hate it, and we participate directly or indi- rectly in its course. To quote a famous definition, politics is about who gets what, when, and how (Lasswell 1936). Collective decision-making processes generate the an- swers to these three fundamental questions. The way in which societies se- lect leaders enshrines power relations; the way they deal with conflicts over the allocation of scarce resources determines policies and ultimately out- comes. The nature of political regimes, the electoral laws, the transparency of elections, the access of interest groups to the decision-making arena, the quality of the judiciary, and the degree of political stability determine the laws of the land and their implementation. The second parent, economics, deals with the production, consumption, and allocation of goods and services. Supply, demand, and market equilibria are more or less household concepts. Governments regulate markets and the behavior of economic agents, collect taxes, and reallocate resources. In relation to specific development issues and reforms, political economy brings the two fields of inquiry together to identify problems and look for solutions. Why are health services or education underprovided to the poor or available in certain regions only? Why are bridges and highway projects started but never finished in many locations? Why does widespread poverty 18 Understanding Policy Change persist in some areas despite foreign aid, institutional reforms, or entire batteries of policies meant to address it? ­ Economics would answer market failure. Significant information discrep- ancies among parties in a transaction, monopoly powers, and negative exter- nalities lead to underinvestment and inefficiencies. Therefore, either mar- kets or governments fail to provide crucial goods and services to the poor. Politics focus less on efficiency and more on the distributional impact of choices. Its answer to the question of high poverty rates would be collective action problems or the difficulty of some groups—say, the poor—in joining forces and demanding that governments enact policies favorable to them. Political economy combines the two approaches and asks bidirectional questions: • How do political factors and institutions affect development outcomes? For example, are democracies providing more public goods than autocracies? Do federal systems lead to higher fiscal deficits than unitary modes of government? Does political clientelism undermine the ability of govern- ments to provide education and health services to the majority of the population? Do certain political parties provide more electricity, water, or better services to the poor? Does the lack of political mobilization of cer- tain groups in society hamper policy change? • How do economic factors and institutions related to development affect po- litical outcomes? Does socioeconomic inequality make transitions from autocracy to democracy more likely? Does democracy survive longer in countries that have already reached a certain level of development? Do regional patterns of economic production and resource endowments lead to the choice of decentralized as opposed to centralized systems of gov- ernment? What tax and spending policies extend the political tenure of leaders (for example, presidents or prime ministers)? The tools and diagnostics of political economy comprise a range of the- matically diverse streams of research: the politics of regulation, campaign finance, trade policies, municipal service delivery, legislative institutions, budgetary processes, poverty reduction, and decentralization, among many others. Methodologically, political economy uses economic tools to under- stand development problems and guide reform (Gerber 2003). For readers exposed in any way to economics, some of the chapters of this textbook will sound familiar. Major analytical blocks—principal-agent relationships, in- formation asymmetries, and credible commitment—are the bread and butter of economic analysis. To this mix, political economy adds political parties, bureaucracies, interest groups, and civil society organizations as key actors Political Economy: What It Is and What It Is Not 19 FIGURE 1.1  Political Economy: The Intersection of Politics and Economics Politics PEA Economics Source: Authors. and policy ingredients and looks at how states and markets interact. All the exercises and interactive activities in this text build on game theory, another family of key tools closely related to economics and decision sciences. Simply put, political economy is the politics of economics (see figure 1.1). The field is built around a simple but powerful intuition: good policy is not always good politics (Bueno De Mesquita et al. 2003; Guriev 2012). If politi- cians face a clear trade-off between promoting policies beneficial to society and staying in power, they are likely to choose the latter. This choice, caused by incentives generated by the rules of the political game, often derails devel- opment and stalls reforms. We also know that institutions are not all the same: they generate different incentives that lead to degrees of match or mis- match between “good policyâ€? and “good politics.â€? If political leaders are elected, monitored, and kept accountable for their campaign promises, their reelection incentives often align with policies demanded by voters. If, how- ever, the political survival of politicians depends in reality only on a handful of powerful local elites or on vote buying, or if their performance in office is not monitored, they lack incentives to allocate resources to health and edu- cation. For them, it pays off politically to keep their small circle of cronies happy through favorable government contracts, subsidies, and tax loopholes and to build patronage machines or invest in a repressive apparatus to deal with public discontent in case it arises. This (mis)alignment between politi- cians’ incentives to pursue good policies and good politics is central to politi- cal economy. 20 Understanding Policy Change What Is—and What Is Not—Political-Economy Analysis? Some rules of thumb can help explain what political-economy analysis (PEA) is and is not (Gerber 2003, 4). Not all analyses of political economy pertain to the study of the economy per se: many PEA products model politi- cal actors as strategic utility maximizers1 operating in different institutional contexts and use important insights from this assumption. The study of leg- islatures is an example of political-economy work that has only indirect im- plications for economic development policy or reform. At the other end of the spectrum, while some research agendas use the tools, methods, and language of modern economics to study politics, they do not capture the intrinsically political mechanisms at work in their studies of a particular problem. For example, classic public finance scholars simply assumed that the public sector would have an incentive to increase trans- parency, provide information, and prevent fiscal deficits without consider- ing whether the political incentive structure would be conducive to that objective. This being said, political economy covers a vast territory of concepts, problems, methods of research, and standards of evidence. Rather than dwelling on meta-theoretical or methodological issues, we focus simply on raising awareness about the conditions under which various configurations of political actors and institutions can skew incentives and, as a result, hin- der or advance a reform agenda. To be fair, politics is not always the culprit. Sometimes it corrects the incentives of markets, it regulates, it redistributes, or it internalizes externalities generated by markets. We operate under the assumption that there is no silver bullet. Otherwise, the world would be a better place by now. That point aside, a critical way of thinking about policy reform is necessary in order to gain at least a glimpse of what is behind organizational mantras, fads in the international commu- nity, or whatever constitutes the latest spin, if nothing else. This handbook tries to survey a vast field and to organize information that development practitioners can then transfer to their own specific questions and contexts. It is also important to note what political-economy analysis is not about. Even if the discipline of study and the set of diagnostics and tools it provides contain the word political, it is not prescriptive or normative, and it does not intend to provide endorsements or recommendations to readers of any spe- cific political institution or arrangement. Its only aim is to point out core concepts and mechanisms of causation that shed light on the allocation of scarce resources and apply to a wide variety of cultural contexts, organiza- tions, and institutions. In fact, the core logic of political economy has been Political Economy: What It Is and What It Is Not 21 used to explain dynamics and outcomes generated by entities as different as family units, religious organizations, business associations, social networks, prison gangs or other forms of organized crime, and state bureaucracies. In this sense, this handbook showcases the analytical tools that can help prac- titioners and researchers diagnose problems, navigate around obstacles, generate alternative scenarios of action, and come up with feasible and cre- ative solutions. Moreover, the content discussed throughout this book evolves constantly, and our hope is to encourage readers to develop the analytical thinking that can create less conventional models of problem solving in development. Var- ious cultural contexts and communities around the world generate innova- tive ways to solve collective problems and take full advantage of technology, preexisting social networks, and windows of opportunity for action. Despite the fact that good policies and good politics are rarely in sync, leadership and local solutions to such misalignments identified throughout the handbook are emerging every day and generate good policies even under conditions of bad politics. In many cases, self-help groups and cooperatives organize around the mistrust of corrupt state institutions. Hometown associations in Nigeria, for example, fulfill the de facto role of a missing (or malfunction- ing) local government. Lineages, elders, faith-based communities, women’s credit groups, and self-help associations have played a similar role in Tanza- nia and Uganda. Communities of farmers in Nepal, despite employing less sophisticated irrigation techniques than the ones provided by the govern- ment, have managed to grow more crops than ever before because of clear community rules about contributions and sanctions. Reform Stories: Seeing the World through the Lens of Political Economy The following examples briefly sketch the stories of several reforms that seem, at a first glance, dry and technical. The problems appear to be clearly identifiable and the solutions seem to reside in either standard economics textbooks or mainstream public administration manuals. Yet, despite this first impression, without accounting for political (or “nontechnicalâ€?) expla- nations, we cannot fully understand their course. Public Procurement Reform Perhaps very few policy areas are as prone to corruption, bribery, graft, and bad governance as public procurement, a sector that lies at the dangerous 22 Understanding Policy Change intersection between powerful private and public interests (see Ware et al. 2007; Frøystad, Heggstad, and Fjeldstad 2010). It is estimated that at least US$400 billion annually is lost to corruption in procurement worldwide. Be- cause corruption raises the costs incurred by governments and distorts bud- get allocations for universal public services (such as education and health), dubious procurement practices lead to inferior development outcomes. Pro- curement brings together politicians, bureaucrats, and firms that want to win lucrative contracts and that have incentives to collude. Members of the legislature often provide “pork-barrelâ€? spending to their constituency to be reelected. Civil servants, sometimes closely connected to the politician, have enough discretion to award the contract to private companies. The chain of political influence and the large sums of money involved render reforms in this area extremely difficult even in developed countries, not to mention in contexts with weaker regulatory frameworks and rule of law. The package of technical solutions contains a variety of demand- and supply-side interventions that have worked in some contexts: the use of technology for e-procurement systems, mechanisms of forensic auditing, voluntary disclosure programs, engagement of civil society organizations for better monitoring, and information-sharing platforms, among others. Nevertheless, the technical correctives of incentives are by far overshad- owed by the larger political-economic environment in which reform occurs. Even in cases in which some level of technical reform has been achieved, implementation still remains an unresolved issue. Despite the fact that many local companies active in the road construction sector in Uganda, for exam- ple, prefer and support a proper tender-based system for contracting, fears of political retaliation or blacklisting in case they challenge the winners still persist (Frøystad, Heggstad, and Fjeldstad 2010, 23). This reality is recurrent in many countries where even the anticipation of weak incentives for imple- mentation and the fear of political retaliation push potential reform support- ers to refrain from actively seeking justice and promoting change. The Setting of Interest Rates The cost of borrowing (that is, interest rates) affects the willingness of busi- nesses and consumers to spend money: lower interest rates stimulate de- mand, and higher interest rates increase savings (Fischer 1995). Interest rates that are too low can precipitate too much short-term borrowing and result in a boom that ends when unsustainable debt levels force consumers and businesses to cut back on spending. Therefore, the interest rate has to be set so that the level of spending is not unsustainable. This is a classic example of a technical element in financial policy making. Political Economy: What It Is and What It Is Not 23 However, elected politicians are a crucial part of the story. Governments may wish to overspend just before elections to make certain categories of voters feel better off. If government can control the setting of interest rates, the result may be what is usually called “a political business cycleâ€? in which rates are driven by the career concerns of the incumbent policy makers rather than by the fundamentals of economics. An approach based entirely on the standard economic prescriptions would not explain what really hap- pens in most of the world. This political-economy dynamic has been one of the reasons why after World War II many countries have sought, with varying degrees of success, to make the central banks that set interest rates independent of political pressure. This institutional reform is the product of an explicit recognition that “nontechnicalâ€? or political factors can often skew incentives and derail the outcomes predicted by textbook best-practice types of policy. Civil Service Reforms Making bureaucracies function well has been at the forefront of develop- ment work for quite a long time. The recognition that governance matters and that administrative corruption thwarts development outcomes has led donors and governments to make concerted efforts to improve the quality and the probity of the civil service. In an initial wave of reforms, the problem was approached from a fiscal-administrative angle and focused on employ- ment reduction in the public sector. The subsequent shift to governance concerns brought in a combination of managerial solutions to boosting per- formance but also the realization that politics is a crucial part of the story. Even if the right financial and professional incentives are in place, even if the salaries of public servants are high enough, and even if there are estab- lished rules of meritocratic recruitment on paper, bureaucracies are, more often than not, colonized by the political supporters of the leaders. In the absence of institutions that stand in the way, the incumbent political parties incentives to politicize bureaucratic agencies (Grzymala-Busse 2007; have ­ O’Dwyer 2006; Shefter 1994). Just as in the previous example, civil service reform is a typical policy domain that was initially thought to have purely fiscal and administrative characteristics, only to end up explicitly incorporating prescriptions aimed at depoliticization. Nevertheless, that realization by itself, without a context- specific understanding of political incentives, has led, in the best-case sce- narios, to only marginal improvements. An evaluation of 124 World Bank lending projects designed to address civil service reform between 1980 and 24 Understanding Policy Change 1997 found a poor level of outcomes (Shepherd 2003). Civil service commis- sions established overnight—or tenure and pension provisions supposed to detach bureaucrats from political influence—could not do much in the ab- sence of robust party competition that could effectively keep in check the incumbent’s incentives to staff the bureaucracy with supporters. In some cases, the timing and sequencing of democratization, state building, and party building have shaped the degree of bureaucratic autonomy, as well as the nature of the political coalitions protecting it (Shefter 1994). Decentralization Economic theories of fiscal federalism, seconded by a search for pragmatic ways to bring decision making closer to citizens, have generated a world- wide wave of reforms aiming to devolve power and resources to lower tiers of government. Decentralization was expected to correct democratic defi- cits, improve public service delivery, make local governments more likely to foster a better investment climate for attracting businesses, and reduce in- centives for corruption. In many cases, these expectations never material- ized or went in the opposite direction. In some contexts, like Argentina and Brazil, the economic performance deteriorated as a result of the reforms. In others, like Nigeria, it exacerbated ethnic tensions (Wibbels 2006). What went wrong? It turns out that the normative assumptions that un- derpinned initial models were highly problematic (see Wibbels 2006, 168; Remmer and Wibbels 2000; Rodden 2002; Gelineau and Remmer 2006; Ea- ton, Kaiser, and Smoke 2010). Often, local governments either were more corrupt than the center or, for political or bureaucratic reasons, were de- prived of capacity and resources. On the demand side, the aspirations to correct democratic deficits also often fell short of expectations. Voters lacked information on who should provide services and therefore became more confused and were unable to keep politicians accountable. Businesses and citizens were not as mobile across subnational units as initially thought. The very origins of decentralization often concealed political incentives that distorted development outcomes. The pace and scope of reforms, in many cases, were dependent on electoral calculations of the incumbent po- litical party and on the regional and local power of the opposition. These frictions resulted in lack of coordination among tiers of government, over- taxation of citizens, overspending by local units, and widespread misunder- standing over which level of government was responsible for providing public goods. Overall, these side effects greatly diminished the intended benefits of the reforms. Political Economy: What It Is and What It Is Not 25 Reforms of Public Financial Management Finally, budgets are the true battleground for distributive decisions (see Dorotinsky and Pradhan 2007). In many ways, annual budgets are the offi- ­ cial script of the social contract between rulers and citizens. Budget propos- als go to legislatures, are debated, are voted on, and decide the amount and structure of revenue to be collected and public expenditure to be allocated in the coming fiscal year. Afterward, they are usually submitted to an audit in- stitution that makes sure that commitments were adequately fulfilled. Dur- ing the full budget cycle, the draft involves almost every political institution, from the presidency to the line ministries all the way to parliaments and specialized public accounts committees. Therefore, it is widely recognized that reforms entail building the mechanisms of accountability within and outside the government as much as they entail dealing with the technical issues related to resource management. Corruption and political clientelism (the usual suspects by now) go hand in hand with budget systems that lack capacity, internal controls, external accountability, and transparency. Beyond informal networks of patronage, the formal rules of the political systems create a whole world of incentives that shape the fiscal deficit levels, revenue shares, and expenditure alloca- tions for certain constituencies. Presidential and parliamentary systems as- sign different powers to the legislature during the budget formulation and approval process. If a powerful executive is unchecked because the parlia- ment receives the budget late for review—and is therefore deprived of key information on tax and spending—or because the legislature lacks the power to amend the budget draft, the accountability of the system is weak. Even the number of pages of budget proposals varies dramatically around the world, from one or two pages to thousands. The percentage of off-budget accounts in total spending also ranges from zero to over 60 percent. In many contexts, these accounts do not go through the legislative process, remaining entirely at the discretion of the executive. Without supreme audit institutions to trace budget implementation and ask for explanations in case of discrepan- cies, the government can engage in corrupt or clientelistic practices at will. Even when legislatures have a significant say in the budget process, elec- toral rules, the strength of political parties, and coalitional requirements cre- ate different incentives for members of parliament (MPs). For example, par- liamentary regimes seem to lead to a larger government than presidential regimes do, as politicians have incentives to claim higher individual shares from a common-pool resource.2 Other political institutions such as federal- ism have been found to interact with fiscal policy outcomes in general, and expenditures in particular, in both developed and developing countries (see 26 Understanding Policy Change Beramendi 2007; Remmer and Wibbels 2000; Rodden 2002; Wibbels 2005; Haggard and Kaufman 1995; Niskanen 2003). In federal arrangements, for example, subnational units do not bear the costs of tax collection but can overspend in anticipation of bailouts from the center. These incentives cre- ate the so-called “tragedy of the commonsâ€? discussed at length. Even the electoral rules of candidate selection affect budgets. In some Latin American countries, for example, regional governors have more weight in the selection of political candidates than party leaders. This arrangement often leads to fragmented party organizations, and to incentives of individ- ual politicians to secure funds for their own geographical districts as­ opposed to more universalistic expenditure programs (Mainwaring and Shugart 1997, 83). This list of political incentives at play in public financial management is not exhaustive, but it aims to show that any reform has to take into account first the political-economic dynamics and mechanisms of accountability. These examples of reforms showcased some of the political or “non- technicalâ€? factors that skewed incentives, intersected with the initial policy targets, and changed final outcomes. How can development practitioners identify the specific sets of incentives that are likely to travel this long path of causation to affect their own outcome of interest? What kinds of con- straints and opportunities inherent in the political-economic environment are they likely to face when funding or managing concrete development projects as diverse as bridge building, financial management information systems, or public health care policies? Why, despite a continuous stream of financial resources allocated to these projects, are the results lagging? Why do other operational contexts achieve surprising results with much lower budgets? To help readers think systematically through the list of factors that might help or hamper their own project path, we designed a hypothetical exercise. Let us imagine that the reformer or practitioner sees only the two ends of the policy process: the initial investment and the tangible change in outcomes following the intervention. How can he or she unpack the black box of the policy process in between? We illustrate the task with a specific develop- ment puzzle: why do some countries manage to achieve better primary school enrollment and infant mortality indicators despite lower health and education spending? How do we explain the development failure in some contexts and success of others? First, we will introduce the puzzle. Then, we will build an inventory of economic, administrative, and political constraints that could be responsible for outcome divergence. Let’s try to organize and analyze the individual pieces. Political Economy: What It Is and What It Is Not 27 Deconstructing and Understanding Poverty Reduction Reform: An Empirical Puzzle One of the major debates on what really matters for poverty reduction fea- tures at least four key positions:  he first argues that underdevelopment persists because the objective • T conditions that some countries face are abysmal: difficult weather, unfa- vorable geographical locations (landlocked or arid areas), exposure to pandemics, and sheer lack of resources. According to this argument, it is hard for people to escape the so-called poverty trap without major up- front investments that lead to higher incomes. Foreign aid, in this version of the story, is crucial. If rich countries committed more funds to poverty reduction, not only could ambitious goals be achieved but also global poverty could be eliminated as early as 2025 (Sachs 2005). The second version of the story pays more attention to the institutions, •  corruption, and bad governance that filter aid funds in the recipient coun- tries and transform them into development outcomes (World Bank 1997). The third version is less keen on aid interventions from donors and more •  focused on building the right incentives for domestic actors to promote growth through their own participation in markets and self-reliance (Easterly 2002). Finally, the fourth approach is the most pragmatic (Banerjee and Duflo •  2011). Instead of asking big and sweeping questions about the relation- ship between more spending and outcomes, it concentrates on very spe- cific questions: What kinds of aid programs work well and in what kinds of institutional settings in country x or y? Why? Without getting into the theoretical details of this debate, we will now explore how political economy can weave into these stories (see Banerjee and Duflo 2011, 9–11 for a concrete application of the development polemic). Figure 1.2, which appeared in the 2004 World Development Report Making Services Work for the Poor, captures variations in implementation gaps be- tween public spending on health and education and the tangible outcomes (reducing infant mortality and achieving a high primary school completion rate). What explains divergent results achieved with similar spending or similar results generated by dramatically different expenditure allocations? If the problem were more or less technical (that is, lack of development funds as one of the positions in the development debate suggests), then we would expect that increased allocations for health and education would lead to better results for infant mortality and primary school enrollment. This is 28 Understanding Policy Change FIGURE 1.2 Implementation Gaps—Spending versus Outcomes in Selected Countries, 1980s and 1990s Public spending Under-five mortality Public spending Primary school on health on education completion 60 200 25 60 Malawi Malawi deaths per thousand 180 Côte d’Ivoire 20 U.S. dollars 40 40 U.S. dollars 160 15 Percent Ethiopia Ethiopia 140 10 20 Côte d’Ivoire 20 120 Haiti 5 Haiti 0 100 0 0 100 100 1,000 100 Peru deaths per thousand Thailand 90 80 800 Mexico Thailand U.S. dollars U.S. dollars 80 60 600 80 Percent Jordan Peru 40 400 70 60 Mexico Jordan 20 200 60 40 0 0 50 1980s 1990s 1990 2000 1980s 1990s 1990 1999 Source: World Bank 2004, 37. Note: Spending is measured as total annual public spending for education on each primary-school-age child, in 1995 U.S. dollars, aver- aged for two decades: the 1980s and the 1990s. For the health graphs, the data show total annual per capita public spending on health, in 1995 U.S. dollars, averaged for two decades: the 1980s and the 1990s. not the case. What could account for this wide gap between spending on education and health, on the one hand, and outcomes, on the other hand? We know by now that development failure can have many manifestations: severe corruption, rent seeking, political patronage, poor service quality, low capacity, and lack of financial resources. What, then, are the incentives that lead to better outcomes? Often, this question simply lacks a satisfactory answer. Too many factors filter the amount of spending and decouple it from real results on the ground. For the purpose of this handbook, however, it is still worth thinking system- atically about these intermediary steps, because they introduce the core ana- lytical blocks that the chapters later explore in greater depth. A discussion of the economic, administrative, and political factors or con- straints that could, at least in theory, account for this stunning divergence between spending and outcome follows. Readers should feel free to add po- tential explanations based on their own research and development work experience. Political Economy: What It Is and What It Is Not 29 Examples of Economic Constraints Governments offer different policy incentives to increase demand for educa- tion or health care services. Some use direct cash transfers to households to stimulate school enrollment and regular visits to clinics, while others, for an equivalent budget allocation, focus more on the supply side (delegation of service delivery to localities, for example). The choice to contract out or to provide services is another difference. In health care, some governments choose to delegate selected services to non- governmental organizations (NGOs) or private sector providers. In some cases, studies suggest that similar levels of spending can result in significant differences in the efficiency of service delivery. Economic crises (debt, currency, financial, and so forth) can trigger sud- den country- or region-specific effects on infant mortality and educational enrollment. In the wake of crises, increased unemployment, poor nutrition, low caloric intake, and program cuts in prenatal care have been found to lead to increases in neonatal mortality. A study looking at the effects of the 1997 Asian currency crisis on child health in Indonesia, for example, found a dras- tic hike in neonatal mortality, despite foreign aid flows meant to supplement declines in spending. In 1998, the food prices skyrocketed by 250 percent. In a country in which food accounted for about 50 percent of household expen- ditures, and in which public hospitals had scarce supplies at the peak of the crisis, the visits to health clinics declined, and the infant mortality rate in- creased 3.2 percent in both rural and urban areas (Rukumnuaykit 2003).3 Countries also adopt policies at different development baselines. Reducing infant mortality or improving the primary school completion rate depends on a complex web of institutions and services, not on budgetary allocations to line ministries alone. Without electricity and clean water to improve sani- tation and prevent infant death and without proper infrastructure and trans- portation to allow children to go to school, positive development outcomes are harder to achieve. It is thus important to take into account the level of development, since it has an obvious impact on the complex ecology of the complementary institutions needed for achieving real infant mortality and literacy targets. The gap between spending and outcomes also depends on preexisting so- cial and economic inequalities. Access to basic services in health and educa- tion disproportionately affects the poorest segments of the population. In Cambodia, for example, the poorest fifth have up to three times higher infant mortality rates than the richest fifth (World Bank 2004, 3). High differences between the rich and the poor increase the magnitude of the problem and make targets harder to achieve. 30 Understanding Policy Change Macroeconomic factors, like currency fluctuation and inflation, have an impact. Under conditions of high inflation, the real budget allocations to health and education can plummet dramatically very quickly and become insignificant. Thus, they never translate into the outcomes planned in the beginning of the budget year. The structure of public spending matters: higher allocations for infrastruc- ture investments or wages and salaries of health and education staff, for ex- ample, as opposed to spending on actual services in health or education, might not translate into outcomes at all. Moreover, schools without teachers or health clinics without qualified health care professionals but with a high number of “ghostâ€? workers on the payroll, do not do much for lowering in- fant mortality and increasing school enrollment. Another spending typology relevant for achieving outcomes breaks down budgets into primary, second- ary, and tertiary allocations, depending on the level of health and education services provided. Just as in the case of socioeconomic inequalities, this ­ classification has profound distributive implications. The middle class and the richest quintile of the population are more likely to benefit from uni­ versity education than the poor. In Nepal, for example, 46 percent of total education spending accrues to the richest fifth, while the poorest quintile receives only 11 percent (World Bank 2004, 3). Examples of Administrative or Bureaucratic Constraints Low salaries for public health professionals and schoolteachers might lead to corruption. Having to pay bribes for health care and education services acts as a regressive tax and affects the poor disproportionately. In many post- Communist countries, between 70 and 90 percent of patients were found to have paid bribes to medical personnel. In contexts where the wages are very low for health and education workers, or where they are paid with large de- lays, these workers have no incentive to show up to work or to enter the profession in the first place. Bureaucracies often have low implementation capacity. In some countries, the qualifications and number of civil servants involved in the health and education sectors are greater than in others, despite similar budget alloca- tions. Following decolonization in Africa, for example, many new states lacked qualified civil servants in general. The Belgian Congo had only 6 col- lege graduates, and Zambia had only 76 university graduates in spite of being one of the relatively modernized societies. In the 1970s, fewer than 7,000 Africans in Sub-Saharan Africa were pursuing college degrees. Paradoxi- cally, this shortage of qualified bureaucrats was remedied fast, but many countries witnessed “dramatic improvements of individual capacity accom- Political Economy: What It Is and What It Is Not 31 panied by an equally dramatic decline in institutional capacityâ€? (Van de Walle 2001, 133). Another example illustrates the challenge of staffing in health and edu- cation. In Papua New Guinea in the 1990s, many newly built schools had to close because of the insufficient number of teachers with credentials to fill the positions (World Bank 2004, 22). In contrast, many Eastern European and Central Asian countries have a reasonable ratio of staff to patients or teachers to students. In addition to the low number of teachers and doctors, inadequate qualifications have also been identified as a key determinant of soaring inefficiencies in many countries. A cross-national study conducted in the mid-1990s found that a very low number of cases of diarrhea in chil- dren under the age of five had been properly diagnosed or treated (for ex- ample, in Zambia 30 percent of all cases were accurately diagnosed and only 19 percent adequately treated) (World Bank 2004, 25). The low quality of health care services induced by lack of trained and properly equipped medical personnel led to poor diagnostics and treatment of patients. Data on outcomes (infant mortality and primary school completion in- cluded) are sometimes not reliable. The data collection process itself depends on the quality of the bureaucracy and on political incentives to trace and disclose outcomes accurately. For a number of countries in Eastern Europe and Central Asia, studies that compared data on infant mortality reported by national bureaucracies to international databases with survey data collected in the field have found important discrepancies (Aleshina and Redmond 2003). Similarly, for electricity—a service essential to better health and edu- cational outcomes—satellite-reported data on nighttime energy consump- tion do not always match the data reported by state agencies in charge of delivering electricity (Min 2010). Many governments simply lack the capacity, the tools, and the mechanisms to monitor policy implementation. An often-quoted study in the development community attempted to trace capitation grants awarded to schools for non- wage primary education expenses in Uganda between 1991 and 1995. The findings were puzzling: only 13 percent reached the beneficiaries. Moreover, richer communities managed to claim these funds, whereas in poor areas, most schools received nothing (Reinikka and Svensson 2004). In many con- texts, qualified medical personnel do not want to work in remote rural areas. In some cases, even when they are assigned to clinics in these locations, the lack of monitoring mechanisms and of professional punishment leads to high absenteeism and underprovision of services. A survey in Bangladesh found that the absentee rate among doctors was 74 percent (World Bank 2004, 4). Lack of bureaucratic coordination between governmental agencies is an- other issue. The long and twisted path linking budget allocations with actual 32 Understanding Policy Change outcomes brings together many corners of the bureaucracy: the ministry of finance, different tiers of government (central, local, municipal, and so forth), line ministries (education, health), and the like. In some cases, task duplica- tion creates unnecessary inefficiencies. In other cases, the lack of communi- cation between various agencies blocks important information transfers. Now, after exploring some of the potential economic and bureaucratic constraints of the development puzzle, we will see what politics adds to the mix. Examples of Political Constraints Just like corruption, clientelism, the exchange of concrete benefits for a citi- zen’s vote, is often a conspicuous manifestation of malfunctioning develop- ment institutions because it concentrates public services in certain constitu- encies at the expense of universal provision of health and education services that would benefit all citizens. Patronage networks connecting political par- ties, bureaucrats, governments, and voters often provide disincentives for development reforms because they lock in complex exchanges of access to public services and political support between citizens and politicians. In 1989, for example, the Mexican government introduced PRONASOL (the National Solidarity Program), a large-scale poverty alleviation campaign providing school construction, water, electricity, and nutrition to the poorest segments of the population. Despite a large budget allocation of approxi- mately 1.2 percent of GDP, the results were quite insignificant (a reported 3 percent rate of poverty reduction compared to an initial target of close to 64 percent). Why this large gap? According to several studies that analyzed spending in states and municipalities, the political affiliation of the benefi- ciaries was a major factor in program targeting. Communities that supported the PRI (the Institutional Revolutionary Party), the political incumbent, ­ received significantly higher PRONASOL funds per capita than localities that voted for other parties (Diaz-Cayeros, Estévez, and Magaloni 2002; Magaloni 2006). The system of political representation can create obstacles or opportunities for the access of certain groups to policy making. In autocracies, opposition parties, trade unions, and many other channels of political participation are officially blocked. In democracies, the poor often make up the large majority of the population. At least theoretically, then, their votes should be impor- tant to politicians, and policies such as those touching on health and educa- tion should closely reflect their preferences. Paradoxically, this electoral power is the exception rather than the rule. Why? Political institutions inter- mediate both the way in which the poor are represented in the legislature Political Economy: What It Is and What It Is Not 33 and the incentives of legislators to enact and implement policies favorable to their constituencies. Therefore, even if the majority of voters prefer policies that would increase literacy rates in the rural areas, this preference might not always translate into actual outcomes. Strong executives and irrelevant legislatures skew the political process toward the policy preferences of the executive. In some circumstances, these might closely match the policy pre- ferred by voters. In others, they could not be further apart. Keeping other factors constant, formal electoral rules governing the democratic process can provide their own incentives for legislators to be ­ ­ accountable or not to voters. The design of electoral districts—that is, gerrymandering—might prevent higher representation of the poor in the ­ parliament. The processes of candidate nominations by political parties and the rules by which they are elected to parliaments shape different patterns of interaction and accountability between voters and politicians. In systems of proportional representation, where citizens vote party lists, for example, the control that the party leaders have over candidate selection determines the type of bond that the candidate wants to cultivate. Closed lists in which party leaders exert most control are less likely than open-list systems (in which voters have a say in the candidate order on the ballot) to build a symbiotic link between individual politicians and voters. District size also matters, be- cause in small districts the competition for a limited number of positions gives incentives to politicians to cultivate personalistic votes. A study of elec- toral politics in Honduras, for example, shows how clientelism interacts with formal electoral rules to produce outcomes. On the one hand, political caudillos, all the way from the president to local politicians, are expected ­ to award personal favors, jobs, and tangible benefits to supporters and not to make or implement nationwide pro-development policy. On the other hand, a closed-list proportional representation rule gives no incentives to politi- cians to cater to local constituencies but, instead, to acquiesce to the party leaders. As a result, the majority of MPs never initiate any bills in the legisla- ture. The ones who do, however, are more concerned with locally targeted bills that would keep their rural constituencies satisfied, rather than national legislation that would improve education and health (Taylor-Robinson 2006, 2010). Delegation of authority can also generate political incentives and affect development. The gap between policies and outcomes is sometimes related to the capacity of the particular tier of government responsible for imple- mentation. For instance, many countries have decentralized education and health policies. Strong political incentives to decentralize or not in the first place are often responsible for the lack of coordination and capacity mis- matches between different tiers of governments. In other cases, decentral- 34 Understanding Policy Change ization occurs on paper at the insistence of donors, without the actual re- sources that would enable local governments to execute policies. The timing of elections and the assessment of the probability of winning or losing them create strong incentives for politicians to deliver visible results or push reforms consequential for health and education outcomes. Reforms such as decentralization that affect the gap between expenditures and out- comes are sometimes initiated and implemented when the incumbent party estimates that it will lose national elections, in an attempt to preserve at least some power in the regions. An approaching election makes political representatives want to avoid sensitive or difficult-to-solve development issues. After all, difficult reforms are commonly initiated in the “honeymoonâ€? period or at the beginning of the term in office because the potential political unpopularity would have enough time to dissipate by the time of the next election. In other cases, election time also provides good opportunities for credit taking. The Constituency Development Fund (CDF) in Kenya, for example, allocates US$1 million to each member of the Parliament for projects in his or her own district. In the absence of a formal accountability mechanism, suspicions of corruption often plague the management of these funds. In 2007, a civil society organization wanted to conduct social audits of these funds and involved community members, but the data on CDF allocations were very hard to obtain. The group approached the representative of the Changamwe district and persuaded him that if he disclosed the accounts, he would be the first MP ever to cooperate with civil society on information release. Understanding that cooperation would boost his public image in an election year, he agreed to share detailed data on 14 projects sponsored by his own district CDF. The strategy paid off for both civil society and the politi- cian. The former discovered and corrected many irregularities at the imple- mentation stage, whereas the MP was reelected in a year in which most sit- ting MPs lost elections. The representative acknowledged that opening CDF accounts played a big role in his electoral success.4 We have already alluded to the resilience of clientelistic politics in many countries around the world. One of the most common reasons for it is the politicians’ lack of credibility in delivering universalistic policies, such as re- ducing infant mortality rates or increasing school enrollment. Often, voters do not trust political party platforms since they have not seen genuine re- sults in the past. Therefore, they would rather accept concrete material ben- efits in exchange for their turnout and vote. To quote an eloquent evaluation of this type of voters’ cynicism: “Honduran voters do not understand that a deputy’s job is to legislate, and they do not value legislation because you can- not eat a lawâ€? (Taylor-Robinson 2006, 111). Political Economy: What It Is and What It Is Not 35 In addition to this type of commitment, at the policy implementation stage politicians often lack incentives to build credible mechanisms that would guarantee that budget allocations translate into genuine outcomes. Positions in public bureaucracies can be used to reward supporters, and un- controlled or unaudited budgets can provide campaign funds for the incum- bent party. Depoliticizing the ministry of education or creating a politically independent supreme audit institution that would audit health expenditures is a typical example of “good policiesâ€? that are not “good politics.â€? Lack of information about policies and outcomes can also limit the extent to which citizens can award or sanction political behavior. Unless citizens and voters are informed about the gap between planned and achieved out- comes or, more generally, about government’s performance in providing ef- fective health and education services, they cannot sanction politicians at the ballot box. Media and civil society organizations play a key role in translating the process of policy making and implementation for citizens and in raising awareness about the noncompliance of politicians or bureaucrats with their mandate. The government of Uganda, for example, launched a universal primary education program with the goal of increasing enrollments and literacy rates. In 10 years, this ambitious effort gave tangible results: it increased pri- mary school enrollment from 2.9 to 6.3 million children. The success itself generated its own challenges, however: a severe shortage of school buildings and classrooms. A school facilities’ grant incorporated in the national budget disbursed US$600,000 per district to build new school buildings. The com- munity officials’ lack of managerial experience coupled with incentives for graft made monitoring imperative. In 2002, a civil society organization mo- bilized citizens to obtain information on expenditures from local politicians and bureaucrats and match it with the actual construction sites. Finding large mismatches between spending and the existence or quality of school buildings, the citizens petitioned the Office of the Prime Minister and the Ministry of Education and Sports. As a result, problematic subcontractors were fired. Moreover, the community started to post financial information on community boards, reducing teachers’ absenteeism and the corruption incentives of local officials.5 The process through which development priorities are selected significantly affects development outcomes. Often, the geographical maps designed to al- low technical targeting of resources to the areas that are most in need do not overlap with the political maps of development. For many political represen- tatives, rewarding key constituencies pays off more than objectively allocat- ing aid to the poor. Visibility of outcome matters for politicians, since they 36 Understanding Policy Change can take credit for good policy implementation. Lack of visibility also mat- ters: if political representatives or bureaucrats can slip poor performance under the rug or pass the blame to somebody else, they have incentives to do so to avoid electoral defeat. Occasionally, the search for results that voters can easily identify and attribute to the officials or representatives leads to unexpected results. In some cases, national governments invest in and moni- tor the implementation of innovative education or health policies, mostly in areas with the capacity already in place to implement them, since they are more likely than localities or regions with more questionable capacity to achieve tangible and visible results quickly. In such cases, policy visibility comes at the expense of growing inequalities. Regions that already have re- sources and capacity are empowered, while areas that lack them in the first place suffer from neglect. In two countries that spend the same share or amount on education and health, such selection effects can nonetheless lead to discrepant outcomes. Variations in the capacity of groups to mobilize will fundamentally shape the result of their efforts to demand services. Low-income groups often lack the resources and organizational capacity to campaign effectively. Further- more, the lack of information, skewed electoral rules, the ways in which electoral districts are drawn, and the existence and power of pro-poor politi- cal parties can all constitute severe obstacles or opportunities for the poor to mobilize and demand bureaucratic or political accountability. In some cases, the poor are well mobilized and regimented by patronage networks, have a large vote turnout, but fail to select and hold accountable politicians who would credibly commit to provide and implement health and education. The appeal of tangible clientelistic benefits handed out in front of the ballot box is simply higher in some contexts than the noncredible electoral promises of other parties. On the other side of the spectrum, in many cases the capacity for mobilization and collective action has led to sig- nificant improvements in policy implementation. The by-now famous case of participatory budgeting in the Brazilian town of Porto Alegre initiated in 1989 by a left-wing mayor committed to large-scale community mobilization and citizens’ involvement in city finances led to tangible outcomes: over a 15-year period, the access to water and sewage reached almost 100 percent, and the number of schools quadrupled. In the Indian state of Uttar Pradesh, a low-caste political party managed to increase the electricity coverage of villages (Min 2010). Despite such inspiring stories, more often than not, the poor fail to mobi- lize as effectively as other interest groups. This failure skews resource allo- cation in favor of the most politically vocal and influential. The health sector Political Economy: What It Is and What It Is Not 37 provides an example. If the public procurement process for medications is plagued by corruption, then a large share of public spending on health goes toward purchasing overpriced drugs from powerful private sector compa- nies rather than toward basic services that would, for example, significantly lower infant mortality rates among the poor. In the late 1990s, Albania spent a large share of its health care budget on the problematic procurement of drugs through direct purchase and leakage of confidential information to pharmaceutical companies, as opposed to competitive bidding and a trans- parent process (Cohen, Mrazek, and Hawkins 2007, 32). Mechanisms of accountability are key to providing politicians with incen- tives for implementing good policies. Some cross-national studies have found that poor democracies have higher primary school enrollment than authoritarian equivalents, since electoral pressures make politicians more responsive and accountable to the educational needs of voters (Brown 1999). In the same vein of argument, a study of education spending in Brazil found that the periods of democratic rule, as opposed to autocratic episodes, sig- nificantly affected the allocation of primary education funds between state and local governments, as well as the share of primary versus tertiary educa- tion spending (Brown 2002). But democracy and elections might not be the only institutions that can generate incentives for public service delivery. As other findings indicate, autocratic forms of government with alternative mechanisms of accountability might outperform democracies in delivering selected basic health and education outcomes. Cuba has better health and literacy indicators than many countries with much higher income, and China has reduced infant mortality to a great extent and achieved almost universal primary school enrollment (World Bank 2004). An effective and independent judiciary can protect citizens against politi- cal attempts to promote policies that may hinder their basic rights. Even in the absence of a responsive government, a robust system of checks and bal- ances and separation of powers creates paths other than the ballot box through which citizens can influence outcomes and hold the executive ac- countable. Especially in the case of health and education policies, many na- tional constitutions inscribe basic services as human rights that states should provide to all citizens. If the courts are free of political interference from the government or the legislature and if citizens have the informa- tional and financial capacity and the ability to mobilize and access the­judi- cial system, then tangible outcomes are more likely to be achieved. The Treatment Action Campaign Case, for example, brought in front of the South African Constitutional Court in 2002 by a civil society organization, claimed that an antiretroviral medication crucial to the prevention of HIV 38 Understanding Policy Change transmission from mother to child at birth should become available in pub- lic hospitals. Despite the government’s resistance to the dissemination of the drug on financial grounds, the court ultimately awarded the case to the claimants. The case is now considered a major milestone on the long and winding path of HIV/AIDS policy reform in South Africa (Gauri and Brinks 2008). Last, but certainly not least, conflict and political instability—such as wars, civil unrest, or other fragile situations—can significantly affect development outcomes. Several cross-national studies have found that fragile states have, on average, 10 percent higher infant mortality rates than comparable coun- tries, irrespective of the line ministry spending (Gates et al. 2010). By the same token, according to estimates, 38 million out of approximately 230 mil- lion children in countries affected by conflict do not attend primary school- ing. In fact, 30 percent of all the children in the world who do not complete primary education live in fragile states (Gates et al. 2010, 19). Economic, Administrative, and Political Factors: Blurry Areas of Conceptual Overlap The factors behind development failure and success just discussed are by no means a complete list. Moreover, only problem- and context-specific empiri- cal testing can confirm whether each of these constraints is relevant. The list is hypothetical. It is rather supposed to start a conversation about a broader range of potential causes of development failure, beyond a narrow “techni- calâ€? (or nonpolitical) core of explanations and solutions. A good understanding of political economy helps us sort through these constraints and incentives responsible for such a divergence in final out- comes. Sometimes the intersection of distinct political and economic causes is the best explanation. Other times, what may appear to be a purely eco- nomic or political factor might hide incentives characteristic of the other two categories. Here are some examples:  conomic inefficiencies in the implementing agency might stem from sev- • E eral causes: lack of basic capacity, lack of adequate coordination among bureaucratic units, and problematic public procurement practices for pri- vate contractors, among others. These constraints pertain to the bureau- cratic and political domain as well, as they skew incentives across all levels. To an equal extent, seemingly political factors that decouple health or edu- •  cation outcomes from the spending levels (such as lack of political will to Political Economy: What It Is and What It Is Not 39 implement policies effectively, weak rule of law, weak voice and account- ability, and the like) are often problems closely associated with or exacer- bated by resource scarcity, the degree of development, a low level of urbanization and modernization, or sheer poverty. ­ Bureaucracies malfunction for many reasons. On the one hand, malfunc- •  tions might be the product of insufficient economic development or the lack of basic state capacity to establish a critical mass of competent bu- reaucrats. On the other hand, in the absence of mechanisms that prevent politicization, bureaucracies are often at the mercy of the political incum- bent. The winning party can staff the state with party supporters and use regulatory agencies to retaliate or persecute opponents. Even data on budgets and outcomes—which we usually take for granted— •  are not “politicsâ€? free. Sometimes key budgetary indicators are not re- leased at all or are distorted precisely because politicians try to prevent being sanctioned electorally if such data reveal underperformance. The rest of this handbook will deal mainly with the third category, politi- cal constraints, and explore how they derail or promote development out- comes and reform attempts. The Limits, Perils, and Promises of Political-Economy Analysis Before we move forward, we present some of the main arguments for and against political economy that the development community has raised. The arguments in favor of this approach claim that it provides the ana- lytical tools, methods, and diagnostics that allow us to understand and map the policy context. Without explicitly incorporating political institutions, actors, and their interaction into our analyses of the policy-making environ- ment, we leave out at least half the story and fail to properly account for context specificity. The arguments opposed to it say that political economy focuses on large-scale institutions and processes that rarely change but ­ neglects or ignores the marginal improvements feasible even in bleak envi- ronments. In addition, political economy is less practical in addressing how reform should proceed and is often overly pessimistic, given that the con- straints in many developing countries are severe and the feasibility of policy change is low. According to its critics, political economy is also not sensitive enough to context because of its rational choice underpinnings that over- simplify incentives, payoffs, and the nature of information flows between actors. 40 Understanding Policy Change In the following section, we briefly introduce some arguments and alter- natives that aim to respond to these issues. Then we will clarify how this handbook plans to address (or bypass) at least some of the criticisms. Insurmountable Problems, Policy Resignation One of the critiques of applying the approach of political economy to devel- opment comes from its fatalistic flavor. The analytical focus generally falls on institutions, the incentives they generate, and the effects these incentives have on development outcomes. The implication is that if a reformer wants to shift the development equilibrium and reverse the path—which means changing outcomes, which means changing incentives, and which means al- tering or creating institutions—the set of potential solutions is usually overly challenging to tackle. Voter-politician accountability relations, decentral- ized government, or anticorruption agencies cannot be transformed over- night to suddenly become functional without the necessary buy-in from key actors. Especially if the fundamental institutions (that is, democracy, federalism, and the like) are a reflection of underlying conditions and power configura- tions in the society, institutional change requires a fundamental overhaul of social relations and background conditions and might not be feasible or even desirable. This view is often quite cynical with respect to the idea of “trueâ€? institutional change often showcased by international donors. Provocatively, one observer argued that even the editorial template chosen for the presen- tation of reform outcomes in the policy reports of international organiza- tions often signals the uniqueness of such success stories. In-text boxes usu- ally highlight the rarity of institutionally induced change, possible only when the underlying conditions are ripe for success. Otherwise, if institutions were the true drivers of results on the ground, irrespective of deeper struc- tural factors, the readers would expect to see tables or graphs that would record empirical patterns more systematically (Przeworski 2004, 530). In practice, this is the “go home and cryâ€? scenario. If, say, we know that democratic accountability through elections gives the “rightâ€? incentives to politicians to perform and that as a result, it leads to the prevention of large- scale famines, we might understand the problem analytically while being further than ever from a feasible solution. Promoting clean elections and de- veloping accountability in a poor country over a short period of time in order to apply an institutional solution to food crises are as challenging as reducing inequality or promoting economic growth overnight. We argue, however, that practitioners and reformers should be aware of the limits and perils of Political Economy: What It Is and What It Is Not 41 change, even if in the end of their analysis they discover that there is nothing to be done. Otherwise, they run the risk of institutional engineering with limited or missing domestic ownership and possibly detrimental conse- quences for development. “The Quiet Revolution of Small Successesâ€? versus “Big Questions and Big Answersâ€? In a recent influential book in development economics, the authors Abhijit Banerjee and Esther Duflo make a useful theoretical distinction between uppercase INSTITUTIONS and lowercase institutions. Whereas the former are the broad fundamental rules in a society (democracy, federal arrange- ments, or parliamentary or presidential regimes), the latter include more granular and localized forms of interaction between actors. A second re- lated critique argues that political economy is overly focused on political INSTITUTIONS: democracy and autocracy, federal versus unitary systems, or any other broad set of core rules that keep the machinery running ­ (Banerjee and Duflo 2011, 238–42). Reform in this view, again, is hard or impossible because the approach is simply too ambitious and too pessi- mistic. While key rules might explain why things do not work well, they are often hard or impossible to change precisely because they lie at the core of the beast. An alternative approach has furthered our understand-ing of how development works on the ground and has infused some optimism in the debate between academia and practice. Its proponents adopt instead a more pragmatic plan of action, leaving aside the big questions such as why de- mocracies foster growth and development and focusing instead on manage- able marginal successes based on the concrete institutional arrangements on the ground rather than on the big INSTITUTIONS themselves. Here are examples of questions that this approach answers: Are the spe- cifics of democratic institutions—characteristics of elections, checks and balances, patterns of popular participation, the nature of the legislative pro- cess—in country x conducive to less corruption, better public procurement systems, more budget transparency? How are various leadership structures and forms of authority affecting the production of public goods and collec- tive action at the village level? Why are there fewer malaria cases in region x than in region y, despite similar environmental conditions? From this per- spective, the view from below, based on the realities on the ground and not on meta-theoretical questions, is more likely to be better equipped to inform policy choices. Even changes in small rules can adjust incentives and ulti- mately lead to better outcomes. 42 Understanding Policy Change This approach is complementary to our view of political economy in practice as it restores hope in the feasibility of change. We argue for empiri- cally grounded analysis built around a concrete problem that uses a variety of methodological tools and conceptual building blocks. However, big ques- tions and big answers related to INSTITUTIONS often bring theoretical and empirical insights that inductive, narrow, or localized findings do not always provide. A mixture of both forms of inquiry into why development problems persist can paint a complete feasibility map for policy navigation. It can also build an active repository of cumulative analytical findings and policy solu- tions to overcome external validity issues stemming from exclusive reliance on context-specific findings. Actionable Political Economy Analysis as Opposed to Static Research Outputs Another prevalent criticism of the political-economy approach persistent in the development community comes from the “so what?â€? problem. Once we have conducted an analysis and understand the policy dynamics taking place, what do we do next? Most studies or reports are valuable in drawing the landscape and the obstacles surrounding reform but are less useful in guiding us through that landscape. One of the responses to this shortcoming is the “actionableâ€? component (Pradhan 2010). Accordingly, political econ- omy has to go beyond analysis and directly fuel a process of knowledge- driven change by translating analysis into actual reform. The type of dy- namic knowledge that has local legitimacy and ownership would not only identify the actual actors and their obstacles but also actively catalyze and nurture their collective capacity for action. This view aspires to organically connect the knowledge production pro- cess with action. Successful and sustainable operationalization, therefore, requires a paradigm shift along three dimensions: analysis for whom, by whom, and how? The answers require the redefinition of roles, as follows: (1) the client—the final beneficiary of the analysis—should be the change agent on the ground and not the donor agencies; (2) the producer—the local teams—should undertake political-economy analysis to develop under- standing and trigger ownership of reform; and (3) the process—instead of generating a passive research output—should result in analyses commis- sioned and produced locally for the change agents and organically integrated into the process that they are about to initiate. As a one-stop source of basic information, interactive activities, and case studies, this handbook is de- signed to encourage and serve this precise purpose. Political Economy: What It Is and What It Is Not 43 Incorporation of Political-Economy Analysis in Donor Operations Last but not least, the World Bank, the United Nations Development Pro- gramme, and the U.K. Department for International Development, as well as many other development agencies and think tanks, have generated a pleth- ora of studies, frameworks, guidelines, and manuals to help their staff better understand the operational environments and think in a more nuanced way about policies. Pockets of reluctance still persist, however, because factoring in the political process occasionally seems to be at odds with the techno- cratic mandate of these organizations. Nonetheless, we argue that main- stream recognition of the fact that politics matters is long overdue. Donor organizations have always been an integral part of the political-economic environment, given their own incentives and interactions with local actors. A deep understanding of the underlying politics of reform, as well as one’s own role in it, is crucial for successful policy design and implementation. Summary As the opening vignettes suggested, the rules of the political game generate incentives that often undermine the technical targets of reforms. Despite all their limitations, political-economy diagnostics are recognized as useful for answering the what and the why of development and for guiding the how to change. In a landscape of nontechnical challenges and obstacles, a map of all possible routes is the first piece of navigational equipment needed. Notes 1. Rational utility maximization is a key assumption of economics; simply put, it means that, when making a decision, the individual (as a consumer, voter, or the like) settles for the greatest “valueâ€? possible derived from the choice, for the lowest cost (or investment). 2. According to some studies, public spending is, on average, 10 percent of gross domestic product (GDP) lower in presidential than in parliamentary regimes (Persson and Tabellini 2003). 3. See http://pdf.usaid.gov/pdf_docs/PNACT315.pdf. 4. International Budget Partnership (IBP), “Social Audits in Kenya: Budget Transparency and Accountability,â€? http://internationalbudget.org/wp-content/ uploads/Impact-Story-Kenya-English.pdf. 5. IBP, “School Building Fund Provides Lessons on Community Mobilization in Uganda,â€? http://internationalbudget.org/wp-content/uploads/UgandaStory English.pdf. 44 Understanding Policy Change References Aleshina, Nadejda and Gerry Redmond. 2003. “How High is Infant Mortality In Central and Eastern Europe and the CIS?â€? UNICEF Innocenti Research Center, Florence: Innocenti Working Papers No. 95. Banerjee, Abhijit V., and Esther Duflo. 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. New York, NY: Public Affairs, a Member of the Perseus Books Group. Beramendi, Pablo. 2007. “Inequality and the Territorial Fragmentation of Solidar- ity.â€? International Organization 61 (4): 783–820. Brown, David S. 1999. “Reading, Writing, and Regime Type: Democracy’s Impact on Primary School Enrollment.â€? Political Research Quarterly 52 (4): 681–707. ———. 2002. “Democracy, Authoritarianism and Education Finance in Brazil.â€? Journal of Latin American Studies 34 (1): 115–41. Bueno De Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2003. The Logic of Political Survival. Cambridge, MA: MIT Press. Cohen, Jillian C., Monique F. Mrazek, and Loraine Hawkins. 2007. “Corruption and Pharmaceuticals: Strengthening Good Governance to Improve Access.â€? In The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level, ed. J. Edgar Campos and Sanjay Pradhan, 29–62. Washington, DC: World Bank. Diaz-Cayeros, Alberto, Federico Estévez, and Beatriz Magaloni. 2002. “The Erosion of One-Party Rule: Clientelism, Portfolio Diversification and Electoral Strategy.â€? Paper presented at the American Political Science Association Annual Meeting, Boston, MA, August 29–September 1. Dorotinsky, William, and Shilpa Pradhan. 2007. “Exploring Corruption in Public Financial Management.â€? In The Many Faces of Corruption: Tracking Vulnerabili- ties at the Sector Level, ed. Jose Edgar Campos and Sanjay Pradhan, 267–94. Washington, DC: World Bank. Easterly, William. 2002. The Elusive Quest for Growth: Economists’ Adventures and Misadventures in the Tropics. Cambridge, MA: MIT Press. Eaton, Kent, Kai Kaiser, and Paul Smoke. 2010. “The Political Economy of Decen- tralization Reforms: Implications for Aid Effectiveness.â€? Washington, DC: World Bank. Egel, Daniel. 2009. “Tribal Diversity, Political Patronage and the Yemeni Decentral- ization Experiment.â€? Unpublished manuscript, University of California, Berkeley. Fischer, Stanley. 1995. “Central-Bank Independence Revisited.â€? American Economic Review 85 (2): 201–06. Frøystad, Mona, Kari Heggstad, and Odd-Helge Fjeldstad. 2010. “Linking Procure- ment and Political Economy.â€? Chr. Michelsen Institute Report commissioned by U.K. Department for International Development and World Bank Institute, http://www.cmi.no/publications/file/3955-linking-procurement-and-political- economy-a-gui.pdf. Gates, Scott, HÃ¥vard Hegre, HÃ¥vard M. NygÃ¥rd, and HÃ¥vard Strand. 2010. “Conse- quences of Civil Conflict.â€? Background paper for the World Development Report Political Economy: What It Is and What It Is Not 45 2011, World Bank, Washington, DC, http://wdr2011.worldbank.org/sites/default/ files/pdfs/WDR%20Background%20Paper_PRIO.pdf. Gauri, Varun, and Daniel M. Brinks, eds. 2008. Courting Social Justice: Judicial Enforcement of Social and Economic Rights in the Developing World. Cambridge: Cambridge University Press. Gelineau, François, and Karen L. Remmer. 2006. “Political Decentralization and Electoral Accountability: The Argentine Experience, 1983–2001.â€? British Journal of Political Science 36 (1): 133–57. Gerber, Elisabeth. 2003. “What Is Political Economy?â€? Political Economist 11 (2):1–3. Grzymala-Busse, Anna. 2007. Rebuilding Leviathan: Party Competition and State Exploitation in Post-Communist Democracies. New York: Cambridge University Press. Guriev, Sergei. 2012. “Why Good Policy Is Not Good Politics.â€? In Policymic. Next Generation News and Politics. Office Hours, http://www.policymic.com/group/ showCompetition/id/1973. Haggard, Stephan, and Robert R. Kaufman. 1995. The Political Economy of Demo- cratic Transitions. Princeton, NJ: Princeton University Press. Lasswell, Harold. 1936. Politics: Who Gets What, When, and How. New York: McGraw-Hill. Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico. Cambridge Studies in Comparative Politics. New York: Cambridge University Press. Mainwaring, Scott, and Matthew Soberg Shugart. 1997. Presidentialism and Democracy in Latin America. Cambridge Studies in Comparative Politics. New York: Cambridge University Press. Min, Brian. 2010. “Distributing Power: Public Service Provision to the Poor in India.â€? Unpublished manuscript. Ann Arbor, MI: University of Michigan. Niskanen, William A. 2003. Autocratic, Democratic, and Optimal Government: Fiscal Choices and Economic Outcomes. Northhampton, MA: Edward Elgar Publishing, Inc. O’Dwyer, Conor. 2006. Runaway State-Building: Patronage Politics and Democratic Development. Baltimore, MD: Johns Hopkins University Press. Persson, Torsten, and Guido Tabellini. 2003. The Economic Effects of Constitutions. Cambridge, MA: MIT Press. Pradhan, Sanjay. 2010. Keynote address at the World Bank Institute and CommGap, “Political Economy Analysis to Action, a Global Learning Event.â€? June 21–22, Washington, DC., http://siteresources.worldbank.org/EXTGOVACC/Resources/ PoliticalEconomyAnalysistoActionFinalReport.pdf. Przeworski, Adam. 2004. “Institutions Matter?â€? Government and Opposition 39 (4): 527–40. Reinikka, Ritva, and Jakob Svensson. 2004. “Local Capture: Evidence from a Central Government Transfer Program in Uganda.â€? Quarterly Journal of Economics 119 (2): 679–705. 46 Understanding Policy Change Remmer, Karen L., and Erik Wibbels. 2000. “The Subnational Politics of Economic Adjustment.â€? Comparative Political Studies 33 (4): 419–51. Rodden, Jonathan. 2002. “The Dilemma of Fiscal Federalism: Grants and Fiscal Performance around the World.â€? American Journal of Political Science 46 (3): 670–87. Rukumnuaykit, Pungpond. 2003. “Crises and Child Health Outcomes: The Impacts of Economic and Drought/Smoke Crises on Infant Mortality and Birthweight in Indonesia.â€? Unpublished manuscript, Department of Economics, Michigan State University, http://pdf.usaid.gov/pdf_docs/Pnact315.pdf. Sachs, Jeffrey. 2005. The End of Poverty: Economic Possibilities for Our Time. New York: Penguin Press. Shefter, Martin. 1994. Political Parties and the State: The American Historical Experience. New York: Cambridge University Press. Shepherd, Geoffrey. 2003. “Civil Service Reform in Developing Countries: Why Is It Going Badly?â€? Paper presented at the 11th International Anti-Corruption Conference, May 25–28, Seoul, Republic of Korea. Taylor-Robinson, Michelle. 2006. “The Difficult Road from Caudilliismo to Democracy.â€? In Informal Institutions and Democracy: Lessons from Latin America, ed. Gretchen Helmke and Steven Levitsky, 106–24. Baltimore, MD: Johns Hopkins University Press. ———. 2010. Do the Poor Count? Democratic Institutions and Accountability in a Context of Poverty. University Park, PA: Pennsylvania State University Press. van de Walle, Nicolas. 2001. African Economies and the Politics of Permanent Crisis, 1979–1999. Cambridge: Cambridge University Press. Ware, Glenn, Shaun Moss, Ed Campos, and Gregory Noone. 2007. “Corruption in Public Procurement: A Perennial Challenge.â€? In The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level, ed. J. Edgar Campos and Sanjay Pradhan, 295–334. Washington, DC: World Bank. Wibbels, Erik. 2005. Federalism and the Market: Intergovernmental Conflict and Economic Reform in the Developing World. New York: Cambridge University Press. ———. 2006. “Madison in Baghdad? Decentralization and Federalism in Compara- tive Politics.â€? Annual Review of Political Science 9: 165–88. World Bank. 1997. The State in a Changing World. Washington, DC: World Bank. World Bank. 2004. World Development Report: Making Services Work for Poor People. Washington, DC: World Bank. Political Economy: What It Is and What It Is Not 47 PART I The first several chapters of this book describe the major mecha- nisms and diagnostic tools of political-economy analysis. Public pol- icy making is a complex process, often characterized by the strategic interaction of stakeholders in continuously evolving contexts. To navigate reform successfully, practitioners need to identify the theo- retical logic behind development problems, as well as the political- economy tools that will allow the analysis of their causes and the identification of potential solutions. Thinking in a nuanced fashion about the way in which political incentives can promote or hinder development can lead to feasible and desirable agendas for change. CHAPTER 2 Accountability and Corruption: The What Question Let us start with the first essential questions any analysis should ask: What is the development problem to be addressed? What are its main manifesta- tions? We begin the chapter with some generic answers to these “whatâ€? questions and introduce the reader to a preliminary diagnosis of the symp- toms. Corruption, pervasive political clientelism, and endemic rent-seeking behavior are signs of development failure. They are highly visible marks of “badâ€? equilibria, and it is likely that they affect the poor disproportionately. What are their immediate or proximate causes? Corruption, patronage, and other outcome-distorting phenomena imply that somewhere, somehow, in the institutional chain of task delegation and political representation, some- thing is not working well. These malfunctions are generated by a broken or weak relationship of responsibility or accountability. In some cases, politicians are not doing their job, and the electorate lacks the capacity to sanction them through collective action. In other cases, bureaucrats shirk their duties and fail to implement accurately the mandates conferred on them by politicians. Even if the locus and depth of the problem differ across policies, sectors, and countries, the generic diagnostic of problematic accountability mechanisms applies widely. 51 FIGURE 2.1  Conceptual Map of Accountability, Corruption, and Related Governance Problems Types of Addressing Accountability Corruption accountability accountability • What does it • Accountability • What is • Mechanisms mean? in democratic corruption? to address • How is it relevant regimes • How is it accountability to development? • Accountability relevant to and governance • Consequences in single-party governance and • Unintended of accountability regimes accountability? consequences • Accountability • Types of of such in authoritarian corruption mechanisms regimes Source: Authors. Therefore, this chapter focuses on accountability and the varieties of gov- ernance problems as key areas of interest for any development practitioner. Among the entire universe of governance problems, we focus mostly on cor- ruption and political clientelism (or patronage) as prime suspects for derail- ing development in many parts of the world. Figure 2.1 introduces the con- ceptual map of the chapter. Objectives of Chapter 2 By the end of this chapter, readers should be able to do the following:  nderstand the varieties of accountability. • U Recognize the different types and manifestations of poor governance, •  corruption, and patronage. Go beyond normative connotations of accountability, clientelism, and •  corruption. Link general concepts of accountability, corruption, and clientelism to •  real-life situations. What Is Accountability? The concept of accountability is closely related to other notions such as ­ political representation, responsiveness, reliability, responsibility, mandate ­responsiveness, and governance (based on Przeworski, Stokes, and Manin 1999, 8–10).1 In a general sense, accountability refers to the normative as- 52 Understanding Policy Change sumption that governments act in the best interest of the people or their citizens.2 Accountability implies a relationship between citizens (as “prin- cipalsâ€?) and politicians (as “agentsâ€?), through which the agents are rou- tinely monitored and sanctioned to ensure effective delivery of benefits to their constituencies (Kitschelt et al. 2009, 742). However, given the diversity of individual preferences in society, how can citizens make sure that the aggregated opinions of the majority lead to policy outcomes that reflect the will of the people? More precisely, in what forms of political representation do government actions and outcomes most closely align with citizen mandates for politicians? Nobel Prize winner Amartya Sen endorsed multiparty democracy as the best vehicle for accountability by positing that no famine has ever occurred under that type of rule: It is not surprising that no famine has ever taken place in the history of the world in a functioning democracy—be it economically rich (as in contempo- rary Western Europe or North America) or relatively poor (as in postindepen- dence India, or Botswana, or Zimbabwe). Famines have tended to occur in colonial territories governed by rulers from elsewhere (as in British India or as in an Ireland administered by alienated English rulers), or in one-party states (as in the Ukraine in the 1930s, or as in China during 1958–1961, or Cambodia in the 1970s. (Page 16 in Sen, Amartya ©. 1999. Development as Freedom. By permission of Oxford University Press, www.oup.com) Elected officials, according to the argument, do not want to be voted out of office for failing to prevent a famine. The explanation implies that at the opposite end of the spectrum, authoritarian governments do not have incen- tives to respond promptly to food crises since the public cannot hold them accountable (see Concepts in Practice 2.1). Complicating Sen’s observation is the fact that both representation and accountability come in many shapes and forms. In fact, there is no single path linking political representation to any uniquely robust type of ac- countability. Democracies rely on regular elections, through which voters can sanction nonperforming politicians. Despite the fact that this possibil- ity of penalizing representatives is a crucial component of mandate re- sponsiveness, it does not automatically translate into accountability and good governance. If elections are not free and fair, or if the poor lack the capacity for collective action, voters fail to punish corrupt representatives. In contrast, autocracies either dilute or completely lack direct electoral ac- countability. In some cases, however, they feature sophisticated systems of institutional checks and balances that ensure effective sanctioning of poli- ticians and bureaucrats who shirked their mandates. The following sec- Accountability and Corruption: The What Question 53 CONCEPTS IN PRACTICE 2.1 Democratic Institutions, Famines, and Food Shortages What is the logic that brings together famine After 1947, India has experienced many epi- occurrence and democratic elections in devel- sodes of severe hunger, but the death toll has oping countries? If politicians fail to prevent rarely reached the threshold that qualifies food catastrophic deaths as a result of widespread scarcity as famine (a possible exception is the famines, they are penalized by the voting public Bihar famine of 1966–67). In contrast, China and lose elections. Therefore, the political in- lost tens of millions of lives in the aftermath of centives generated by elections in democracies the severe famine of 1958–61. Not only did the drive food policy to a better outcome than could government fail to prevent this devastation, but be achieved otherwise, in the absence of the the real damage was carefully hidden from the mechanisms of accountability. public eye. The data on the number of starva- Before Amartya Sen brought to the fore the tion deaths were publicized only after Mao’s role of democratic institutions and accountabil- death. In the absence of an alert media or any ity, the dominant theory on the cause of fam- form of political opposition and accountability, ines focused on severe food shortages brought the autocratic government failed to respond to about by natural causes (drought, floods, unfa- the crisis, with dramatic consequences. vorable weather, and the like). The author ana- Despite the fact that the argument has been lyzed the great famine of Bengal, which oc- criticized for not explaining the occurrence of curred in 1943 and caused around 3 million famines in some democratic countries (like deaths, and concluded that food decline was Ethiopia) or numerous starvation deaths in not the major cause of the catastrophe. The many Indian states since independence, it still problem, instead, resided in the declining real sets the standard for the debate on “nontech­ wages of the rural workers and farmers, com- nicalâ€? solutions to development problems. In- bined with increasing prices for food. Since deed, bringing in political institutions, incen- these factors pertain to the political-economic tives, collective action, and broad questions of domain, the argument suggested that the solu- the accountability of politicians to voters has tion might lie in the institutional capacity of the enhanced our understanding of the variation in public to sanction politicians and bureaucrats policy responses to food crises around the who fail to enact preventive policies. world. Source:  Sen 1999, http://www.nytimes.com/2003/03/01/arts/does-democracy-avert-famine.html?pagewanted=all&src=pm. See also Brass 1986. tions examine accountability mechanisms in democratic and nondemo- cratic settings. This chapter next discusses different forms of political representation and the variety of accountability relations they trigger. For now, let us start with democracies. 54 Understanding Policy Change Government Accountability in Democratic Settings To illustrate accountability pathways in democracies, let us imagine that in democratic Country X, where the infant mortality rate is among the highest in the world, the public demands a reform of the health care system that ­ addresses the issue promptly and effectively. Subsequently, the majority preference for health policy is transmitted to politicians through a number of channels: voting for the party or candidate that runs on a health care plat- form in elections, writing letters to the elected district officials, and conduct- ing national opinion polls confirming the popular will. Once politicians who ran on an electoral platform to address health care reform are in office, vot- ers can check government responsiveness by comparing campaign promises with actual policies. Has the government promptly initiated a health care bill? Did Party A, which won a majority in the legislature, maintain the politi- cal position on health care reform advertised during the campaign? If the answers to these questions are positive, based on the best abilities of voters to evaluate the match between campaign promises and actual policies, then the government is responsive to the preferences of the electorate. Achieving robust accountability, however, requires two extra steps:  n evaluation of the implementation of the health care reform voted on in • A the parliament, approved by the president, and actually carried out on the ground must be conducted. Has the infant mortality rate, for example, improved at a reasonable pace following the new law or reform? Contingent on the answer to that question, the voters must be able to pun- •  ish or reward the government for the concrete results achieved. For ex- ample, if the outcome satisfied or did not satisfy the majority of citizens, then the political party in office could win or lose the next election.3 Figure 2.2 shows the causal path between the heterogeneous preferences of citizens and final policy outcomes. The institutionalization of such a sanc- tioning mechanism through which citizens can punish or reward politicians based on their policy performance is key to the concept of government ac- countability. Democracy is the only such institutional arrangement, or “rule of the game,â€? that formalizes a direct relationship of responsibility between citizens (as principals) and politicians (as agents) through three mechanisms:  egular free and fair elections • R Guaranteed political rights and civil liberties •  Institutional checks and balances. •  A necessary, if not sufficient, component of democracy and the main ingredi- ent for an accountable democratic government, periodic free and fair elec- Accountability and Corruption: The What Question 55 FIGURE 2.2  The Path between Citizen Preferences and Policy Outcomes Individual preferences → Public signals to politicians → Political mandates following elections → Policies → Outcome Source: Adapted from Przeworski, Stokes, and Manin (1999, 9). Copyright © 1999 Adam Przeworski, Susan C. Stokes, and Bernard Manin. Eds. Democracy, Accountability, and Representation. Vol. 2. Reprinted with the permission of Cambridge University Press tions ensure that voters have unconstrained access to the political process and that each vote has equal weight in determining the electoral outcome. What Are the Inner Mechanics and Modes of Democratic Accountability Relations? Politicians use campaigns to identify themselves with the poli- cies preferred by the majority of voters. Ideally, regularly held, free and fair elections guarantee a certain consistency between campaign promises and policies, if the voters have the capacity to act collectively; that is, voters must be informed, be able to adequately monitor politicians and bureaucracies, and be able to elicit credible promises from politicians. However, accountability can manifest itself in many ways and depends on the timing of events (when can voters monitor the delivery of goods—before or after election day?) as well as on the nature of the promises (tangible pri- vate goods such as cash or consumer appliances, or public goods such as health or education?). Not all relationships of accountability are ideal or desirable. Some, such as the clientelistic exchanges in table 2.1, might be the ­ second-best alternatives, when politicians cannot promise anything else, or might even be detrimental to development. Types of Political Accountability in Democratic Politics. Table 2.1 sum- marizes various types of links between citizens and politicians, with implica- tions for the relationship between the two sets of actors.4 Of the main modes TABLE 2.1  Modes of Accountability and Time Horizon for Performance Evaluation Mode of accountability  Time horizon for evaluation  of agent’s performance Indirect (programmatic) Direct (clientelistic) Past performance Retrospective evaluation of public Retrospective evaluation of private good delivery good delivery Future performance Prospective evaluation of public Prospective evaluation of private good delivery good delivery Source: Authors, building on Kitschelt (2000) and on Kitschelt et al. (2009). 56 Understanding Policy Change of political accountability, the “indirectâ€? category—which corresponds to candidate selection based on electoral platforms and the delivery of pub- lic goods—often decouples both timing and b ­enefits of the delegation relationship: • First, there is a temporal lag between the date of election and the timing of policy evaluation. Usually, pre-electoral policy promises, as opposed to tangible handouts or public sector jobs, take longer to yield fruits. Even more importantly, parties lump many issues together in order to offer policy packages before elections (Kitschelt 2000, 850–851). Improving literacy or reducing infant mortality are only a subset of a much broader political agenda advertised to voters. At the ballot box, citizens vote for labels that often combine their preferences on education and health pol- icy with areas as diverse as gay rights or foreign affairs. This process fil- ters majority will and policy stances over multiple dimensions through concrete forms of political representation. Therefore, the temporal ties of accountability as well as voters’ individual policy gains are decoupled and only indirect. • Second, the benefits and costs of post-election policies, by definition, ac- crue to all voters, not only to the supporters of the electoral winners (Kitschelt 2000, 845). The type of relationship that rests on parties offering complex issue pack- ages and on citizens voting the match between their preferences and party platforms is called programmatic linkage between voters and politicians (Kitschelt 2000, 850). At the other end of the spectrum, political clientelism refers to an ex- change of tangible benefits for a citizen’s vote. Clientelism (often associated with corruption) reduces the time lag between vote and benefits but has as its main consequence the delivery of private goods targeted to small con- stituencies rather than to society at large.5 Monitoring and sanctioning con- stitute institutional dilemmas for both modes of accountability. In the case of indirect or programmatic ties, they are key devices for disciplining politi- cians and bureaucrats. For clientelistic ties, they are also tools in the hands of politicians who want make sure that voters comply with their part of the deal (Kitschelt et al. 2009, 744). Depending on the types of actors involved in a relationship of account- ability, we can also distinguish among three accountability types: political, managerial (internal), and social accountability. This primer will explore all forms of accountability. Some chapters apply more to political or to social accountability. Others, such as chapter 6 on principal-agent theory, are more useful for understanding managerial accountability and dwell less on Accountability and Corruption: The What Question 57 FIGURE 2.3  Varieties of Accountability Politicians (policy makers) Political Managerial (internal) accountability accountability Voters/citizens Bureaucracy/Service providers Social accountability Source: Adapted from World Bank (2004), 6–10. direct relationships between voters and politicians or between citizens and bureaucracies. Figure 2.3 illustrates the three-way relationship among citizens, politi- cians, and the bureaucrats charged with implementing policies, along with the types of accountability that are shared by those groups. Governments function best when politicians are accountable to voters, bureaucrats are ­ accountable to both voters and politicians, and individuals and groups are able to act collectively to bring about needed reforms. In reality, sadly, this condition is rare. In diagnostic political-economy analysis, it is worth asking repeatedly, On what leg of the triangle is the problem really located? Government Accountability in Nondemocratic Settings Despite the fact that free and fair elections provide a direct channel for mon- itoring and sanctioning politicians who did not deliver on their promises, elections might not be the only mechanism of accountability that leads to provision of public services and development. Chapter 1 pointed out that even if, relative to other regimes, competitive elections create incentives for good governance, the theoretical and empirical literature suggests that dem- ocratic institutions do not automatically eliminate the potential conflict of interest between voters and (elected) public policy makers. As later chapters will demonstrate, problems of information asymmetries or agenda manipu- lation mean that elections do not necessarily correct this divergence (see 58 Understanding Policy Change Barro 1973; Besley 2006; Ferejohn 1990). As long as voters’ potential for col- lective action is severely constrained by informational or institutional prob- lems, democracies cannot do much to improve development outcomes. Elections by themselves are not sufficient for ensuring robust accountability. In fact, studies show an ambiguous empirical relationship between democ- racies and the redistribution of wealth and income, as well as between de- mocracies and the provision of better public services for the majority of the population. This area of research deserves further attention. We will now discuss forms of accountability under various types of nondemocratic political regimes, as well as the importance of institutional mechanisms in ensuring government responsiveness. First, a clarification is in order. Both democracy and authoritarianism are umbrella categories for many subtypes of institutions that generate their own incentives and obsta- cles to collective action. One-party regimes tend to rely heavily on coercion and ban other parties or political challengers. Authoritarianism includes a plethora of institutional arrangements, such as military regimes (which may or may not allow party competition), monarchies, partyless polities, theocra- cies, and hybrid regimes that combine elements from both democracies and autocracies and, in many cases, allow elections and limited multiparty com- petition.6 We explore varieties of accountability in some of these institutions in the section that follows. Single-Party Regimes and Accountability. As a prominent extreme ex- ample of a one-party political arrangement, the classic Communist system adopted in the former Soviet Union and Eastern Europe was characterized by the fusion of the unique party apparatus and the state.7 The window dressing of formal elections combined with repression or totalitarianism allowed only weak or nonexistent accountability mechanisms. Several fac- ­ tors underlie the lack (or weakness) of government accountability within this type of regime:  itizens can make no direct institutionalized appeal to government; the • C members of the legislature are often selected either directly by the bu- reaucracy or by citizens participating in elections that are usually neither free nor fair. State power is self-legitimized; because the officials are self-declared •  “permanent repositories of public good,â€? accountability, by definition, is not an issue.8 The majority of policies are designed by a small nucleus of party leaders, •  are often passed by decree, and relatively few even reach the compliant legislature. Accountability and Corruption: The What Question 59  he principle of duplication of functions applies: a small number of • T party members execute the same de facto governing functions as the bureaucracy. Therefore, there are no robust institutional checks state ­ and balances. The system, in general, lacks processes of routine policy contestation: the •  factions or party “wingsâ€? are usually purged, and the repressive mecha- nisms keep the civil society under tight control. These obstacles do not automatically imply that there are no private or pub- lic spaces of dissent. In some contexts, local councils or municipal govern- ments become vibrant forums of policy contestation; in others, selected poli- cies are submitted for real debates in the national parliaments. Overall, however, single-party regimes have significantly fewer and weaker pockets of dissent and accountability than their democratic counterparts. Ordinary citizens have little or no recourse to public policies. Authoritarianism and Accountability. In authoritarian states, in gener- al, because of the lack (or poor quality) of elections that would penalize rulers for reneging on policy promises or provision of public goods, there is little or no direct and institutionalized link of accountability between citi- zens and the government. Nevertheless, as we have already seen, with re- spect to certain development indicators, some countries that do not hold free and fair elections perform better than their democratic counterparts. Cuba, as noted, has better health outcomes than Chile or Costa Rica, and China has reached high literacy rates. What explains this pattern? Accord- ing to political economists seeking to understand the mechanics of auto- cracies, some alternative modes of political accountability might give incentives to the executive to deliver public services (Ames 1987; Tullock 1987; Wintrobe 1998; Haber, Razo, and Maurer 2003; Bueno de Mesquita et al. 2003; Crystal 1989). These incentives derive from the following ­features:  nstitutional checks and balances • I Certain rules of selection and succession for the top echelon of political •  leaders Time horizon and survival considerations for the political leader •  Specific mechanisms through which autocratic politicians were able to •  commit to bureaucratic cadres that they will refrain from expropriation of investments (Gehlbach and Keefer 2011) Ruling-party institutionalization (Gehlbach and Keefer 2011) •  60 Understanding Policy Change  olitical career incentives (that is, climbing up the hierarchy of the hege- • P monic political party might require demonstrating competency at the local or regional level) High capacity of monitoring and sanctioning bureaucratic behavior (for •  example, prohibitively high penalties for corruption, noncompliance, and the like). However, these varieties of alternative accountability differ significantly from their democratic counterparts in both domain (accountability along a given dimension) and scope (accountability to whom). The spaces of public policy contestation are narrower in nondemocracies, and state repression is a disciplinary tool often used to block potential dissenters from effectively articulating sensitive political claims. Moreover, autocratic leaders face threats and must therefore be “more accountableâ€? to a small subset of soci- ety that constitutes their coalition of support, not to the citizenry at large (Bueno de Mesquita et al. 2003; Haber 2006). Testing this theoretical logic, some empirical studies have found a strong correlation between regime type and the delivery of public goods: accord- ingly, because political accountability is monitored and sanctioned through elections, democracies perform better overall than autocracies in delivering the policies and goods demanded by the vast majority of society (Bueno de Mesquita et al. 2003; Deacon and Saha 2006; Lake and Baum 2001; Acemo- glu and Robinson 2006; Boix 2003; Meltzer and Richards 1981). On the other hand, other strands of research contradict these findings, arguing that dicta- torships redistribute more than democracies (Wintrobe 1998), or that there simply is no systematic difference in fiscal or development policies across regime types (Cheibub 1998; Mulligan, Gil, and Sala-I-Martin 2004). Be- tween 1990 and 2004, for example, there was no statistically significant dif- ference between democracies and autocracies in total inflows of foreign di- rect investment (Gehlbach and Keefer 2011). Beyond the generic debate, increasing evidence indicates that in the nondemocratic regimes that per- form well in economic growth and development outcomes, some form of institutional checks and balances has evolved (see Concepts in Practice 2.2 for an example of checks and balances in Iran). Over time, the credibility of commitment to some good policies on the side of the government has re- duced graft opportunities and corruption. In sum, accountability is an essential component of governance, whether a regime is autocratic or democratic. When mechanisms for accountability are missing from political institutions, corruption flourishes and growth and development stall. In the next section, we address the symptoms often pres- ent when accountability problems arise and persist. Accountability and Corruption: The What Question 61 CONCEPTS IN PRACTICE 2.2 The Case of Budgetary Decisions in Iran Iran’s post-1979 constitution conferred political not have full authority over it. The speaker of the power on two different institutions: a directly Majlis pleaded the case in front of the supreme elected president and the supreme leader who leader, arguing that, in the past, parliamentary is not directly elected. The president formulates commissions had conducted public audits of the the yearly budget, and the Majlis-e Shora-ye broadcasting organization. He also asked for a ­Islami (the Islamic Consultative Council, or leg- revision of Article 198 of its Internal Affairs Code islature) approves it. The supreme leader has so that the Majlis could exercise its full audit au- large constitutional powers, ranging from con- thority. In the end, the Expediency Council re- trol of bonyads (foundations that own many en- vised the code and allowed the Majlis to exer- terprises and that constitute a significant share cise full audit authority on all state organizations, of the economy) all the way to the right to guide including those under the direct supervision of general economic policies and supervise their the supreme leader. implementation. In addition, two other bodies Another example relates to annual budget have significant powers in the legislative pro- institutions in Iran. Since 1979, no budget had cess. The Expediency Council mediates dis- ever been contested. However, the Guardian putes between the Guardian Council and the Council rejected the draft of the 2002–03 bud- Majlis, whereas the Guardian Council ensures get draft ratified by the Majlis. As a result of the compatibility of proposed legislation with the disagreement, the Expediency Council con- Islamic law and the constitution. vened a meeting with the representatives of On many occasions, the interaction between the Majlis, members of the Guardian Council, these branches has highlighted effective mech- and members of the Budget Committee in the anisms of alternative accountability and the ex- legislature to mediate the dispute in a timely istence of institutional checks and balances. In manner. The Expediency Council ruled in favor June 2002, for example, the Majlis decided to of the Majlis with respect to some objections audit the Iranian broadcasting entity, based on of the Guardian Council. Interestingly, out of an inquiry into its public procurement practices. the five areas of budget dispute, the one most The debate was blocked by the Guardian Coun- closely related to economic development and cil, which had invoked the fact that the broad- growth (the provision of financial guarantees casting agency was under the direct jurisdiction for foreign investors) was awarded to the of the supreme leader and that the Majlis did legislature. Source: World Bank (2005). Common Symptoms of Malfunctioning Institutions: Corruption and Its Relatives One of the most common and detrimental signs of poor governance and lack of accountability in developing countries is corruption, or “the mis- 62 Understanding Policy Change use of public office for private gain.â€? The classic illicit transaction usually involves “(. . .) a bribe by a private citizen to a public official in return for some service that the official should either provide free of charge (. . .) or not provide at all.â€? (Treisman 2007, 211–212). First, it is important to understand what actions constitute corruption. In the context of a principal-agent relationship, corruption occurs when the agent (implementer) shirks the policy implementation tasks in ex- change for a private gain, and the principal (boss) lacks either the will or the capacity to monitor or sanction deviant behavior. Some argue that corrup- tion might also have different meanings in different cultural settings. For example, the density of tight solidarity networks in Africa compared to Europe has been put forward as a potential determinant of the social ac- ­ ceptability of corruption in Sub-Saharan Africa (de Sardan 1999). Beyond different potential meanings, corruption also occurs in various forms. One of its close relatives, political clientelism or patronage, is a prevalent phe- nomenon in many countries and has often been identified as an essential symptom of market failure and suboptimal development. Types of Corruption and Clientelism Some political economists distinguish between different forms of corruption based on two criteria: (1) the scale on which it occurs; and (2) the content and systemic implications. “Low-levelâ€? corruption implies deliberate policy misimplementation for the extraction of private gains when the existing laws should normally sanction such behavior (based on Rose-Ackerman 1978, 1999, 53–54; 2006, xviii–xx). Three examples of low-level corruption follow:  public official solicits a bribe to tolerate a certain behavior that would • A otherwise be at odds with extant laws (for example, no traffic ticket in exchange for a small amount of money). The bureaucrat in charge of allocating a scarce public benefit selects the •  beneficiaries based on their capacity to pay an illegal fee (for example, an official in the ministry of education extracts bribes to award coveted scholarships). An official can make a certain public benefit appear scarce and take ad- •  vantage of this distortion (for example, issuing driver’s licenses or busi- ness registrations in exchange for a bribe). “Grandâ€? corruption implies a more systemic machinery of rent extrac- tion involving both low- and top-level bureaucrats and politicians (Rose- Accountability and Corruption: The What Question 63 Ackerman 2006, xix–xx). For example, an entire branch of the state bureau- cracy (such as the police or the judiciary) cooperates with leaders of organized crime organizations in exchange for generous bribes. Another in- stance of grand corruption is also reflected in the efforts of high-level politi- cians to extract rents from firms that want to use their influence to obtain government contracts. Other typologies of corruption based on the systemic implications of the pathology distinguish among several forms of behavior: administrative cor- ruption, nepotism, state capture, prebendal predation, and kleptocracy. Each is described below. Administrative corruption consists of illegal payments and gifts to bureau- crats for a laxer policy implementation (Hellman, Jones, and Kaufmann 2000, 2). Commonly employed, this type of corruption is the “looking the other wayâ€? behavior of many regulatory agencies around the world in ex- change for the favorable interpretations of laws and regulations. It occurs everywhere, but the scope of the problem and its probability of detection vary dramatically. For example, the U.S. Department of the Interior was sued for blocking auditors from pursuing US$30 million worth of oil royalties not paid by large petroleum companies operating in the Gulf of Mexico. Thou- sands of miles away, in the squatter settlement of Kibera, located in the out- skirts of Nairobi, Kenya, small water providers pay significant bribes to the officials of the water utility for timely delivery and reasonable pricing. These bribes translate into higher prices for the customers of the water kiosks. The courts cannot do much to remedy administrative corruption in a squatter settlement that is excluded from the city water utility networks altogether (McPherson and MacSearraigh 2007). Bureaucratic nepotism, another type of common governance problem, refers to favoritism, such as favors granted, contracts awarded, or appointments to office made by those in power, shown to narrowly targeted interests. This preferential treatment might or might not be related to clientelism or politi- cal patronage. Another systemic type of corruption is state capture, defined as the influ- ence of powerful economic interests “in the formation of laws, regulations, and policiesâ€? through the illegal provision of “private benefits for public of- ficialsâ€? (Campos and Pradhan 2007, 3; Hellman, Jones, and Kaufmann 2000, 2). State capture can be conceptualized as “the privatization of public policyâ€? in which large companies manipulate the rules of the game to their own ad- vantage by providing illicit payments to politicians and bureaucrats (Hell- man, Jones, and Kaufmann 2000, 2). Post-Communist Russia and Ukraine 64 Understanding Policy Change are classic examples of close ties between business oligarchs and the top echelons of the state. Prebendal predation refers to a special form of state capture generated by state bureaucrats who, empowered by their official position and in the ab- sence of state resources directly allocated to them, abusively extract their own discretionary rents directly from citizens (Englebert 2009; Joseph 1998). The term prebendal originated from the system of benefits granted to clergy by the Anglican Church. In a development context, it refers to bureau- cratic agencies of failed states and individuals associated with them that ex- tract as much as they can from the population, while never providing any public or private services in return. In many African countries throughout the 1980s and 1990s, despite poststructural adjustment fiscal austerity that significantly reduced the salaries of public officials, public employment steadily increased. Bureaucrats, organized in pyramidal prebends, were able to extract resources from officials occupying a lower hierarchical position. Going down the ladder, the lowest-placed officials abusively extracted bribes and illegal fees directly from citizens. In the 1990s, it is estimated that in Malawi, for example, up to one-third of the total collected revenue was “stolenâ€? by civil servants. Similarly, artisanal gold miners in the Democratic ­ Republic of Congo pay exorbitant random taxes and fees to government workers (see Concepts in Practice 2.3). CONCEPTS IN PRACTICE 2.3 Prebendalism and Artisanal Gold Mining in the Democratic Republic of Congo Taxation or rent extraction through the exercise of sovereign authority, legal or not, is a typical prebendal behavior in many African countries. According to World Bank estimations, the “effectiveâ€? tax rate that poor artisanal miners have to pay to many tiers of bureaucrats, army, and police approximates 60 percent in the Democratic Republic of Congo. In some of the gold mines of the Orientale Province, studies have identified 30 separate taxes and fees levied on gold min- ers, traders, and operators. For artisanal miners, these include buying mining permits, paying 20–30 percent of their production to mine operators, being taxed several times by multiple layers of officials, and putting aside a certain amount of money each week for payoffs to the army or police. Failing to pay subjects miners to administrative and police harassment. Similar practices have been documented in other DRC provinces such as Katanga and the Kivus. Sources: Fahey (2008); Pole Institute (2007). Accountability and Corruption: The What Question 65 Finally, another type of grand corruption, kleptocracy, is known as the most extreme form of state capture, usually associated with highly personal- istic rule. It implies that the control of the state is in the hands of a highly discretionary individual ruler or his or her small coalition of supporters, who use their political power to pocket a large part of society’s resources (Acemoglu, Robinson, and Verdier 2004). There are many examples of heads of states and their close associates amassing large fortunes while in office. In response to the discovery of Swiss bank accounts in his name, Mobutu Sese Seko, the former president of Zaire, replied: “I would estimate my accounts total less than fifty million dollars. What is that for twenty-two years as head of state in such a big country?â€? (Borgenicht and Regan 2008, 116). Evidence shows that during the 1970s, around 15–20 percent of the operating state budget went directly to the Mobutu family’s numerous bank accounts (Les- lie 1987; Acemoglu, Robinson, and Verdier 2004). In content, the broad concept of corruption is closely associated with re- lated theoretical constructs such as political clientelism, or the exchange of material benefits for votes. However, despite the fact that corruption and clientelism overlap significantly, their relationship might be more compli- cated than initially thought. The conceptual territory, for example, of “non- corruptâ€? clientelism or “nonclientelisticâ€? corruption is currently under the- oretical scrutiny (Piattoni 2001; Keefer 2007; Kitschelt and Wilkinson 2007). Political patronage or clientelism is a “transaction, or a direct exchange of a citizen’s vote in return for concrete paymentsâ€? (such as cash, consumption goods, food, clothes, or household appliances), “or for continuing access to employment, goods, and servicesâ€? (such as public sector jobs, access to food or housing subsidies, or educational scholarships)9 (Kitschelt and Wilkinson 2007, 2). This is one of the most pervasive intersections between skewed po- litical incentives and underprovision of public services around the world. A growing body of literature looks at the causes, manifestations, and ef- fects of political patronage. What is its effect on voting patterns? Are some groups more susceptible than others to handouts at the expense of services? Is there always a trade-off between programmatic and clientelistic politics? Is clientelism independent of coercion, or do the two travel together? How does clientelism affect the provision of public goods and economic develop- ment? How do political parties or individual candidates monitor voters’ be- havior on election day? Do voters or citizens always correctly attribute the tangible goods received in exchange for their votes to the right party or tier of government? The answers to such questions are consequential for our un- derstanding of the relationship among political patronage, accountability, and, ultimately, development. As previously suggested, in some circum- stances, clientelism might just constitute a “second-bestâ€? alternative form of 66 Understanding Policy Change CONCEPTS IN PRACTICE 2.4 Clientelism, Ethnicity, and Gender in Benin In a unique case set in Benin, the author of a study was able to randomize the messages of electoral candidates to voters during the first round of the March 2001 presidential election. In some villages, candidates running for office com- municated to voters a broad public policy platform. They emphasized political issues related to national unity, eradication of corruption, poverty alleviation, and education reform, among many others. In other local communities, politi- cians used electoral appeals focused more narrowly on patronage jobs or finan- cial support for local fishermen and cotton producers. This field experiment enabled the researcher to compare the effect of the two types of messages— programmatic (public good) messages and clientelistic (narrowly targeted goods) messages—on the electoral behavior of voters. The findings suggested that (1) keeping ethnic affiliation constant, clientelistic appeals were particularly ef- fective when undertaken by incumbents or local politicians and (2) men were more likely than women to respond to clientelistic messages communicated by local candidates for office. The study found strong support for the hypothesis that clientelistic campaign messages affect individual voting behavior. Source: Wantchekon (2003). accountability and substitute for the void of general policy responsiveness of politicians to voters. In other cases, patronage might act as a de facto form of corruption and undermine robust electoral accountability. Concepts in Prac- tice 2.4 examines clientelism and voting patterns in Benin. In all their forms, severe clientelism and corruption can have devastating consequences for development and economic well-being. Occasionally, however, positive side effects emerge. Both negative and positive features are discussed below. Negative Consequences of Clientelism and Corruption The most immediate consequences for economic development, social wel- fare, and government legitimacy are severe misallocation of public resources, budgetary distortions, expanding underground economies, and low state ca- pacity for revenue extraction. Clientelism or patronage implies conditional relations between citizens and politicians. By definition, then, the poor who do not vote or support their patrons do not receive cash, water, electricity, access to public jobs, and the like. Accountability and Corruption: The What Question 67 Consider the impact of clientelism on traffic control in Dhaka, Bangla- desh, for example. The traffic chaos generated by inefficient transportation scheduling and the poor condition of the 6,700 buses operating in the city have many negative effects: the high number of pedestrian accidents, deaths, general lack of safety, air pollution, delays, and so on. One traffic corridor alone has 50 percent of total buses operating on less than 30 percent of the total number of routes. Despite an extensive regulatory framework and stakeholder involvement, problems have persisted for a long time. The chal- lenge seems technical at first glance: with better scheduling and newer buses, traffic could improve significantly. It turns out, though, that political patronage is a key part of the story. The Regional Transport Committee (RTC) is the public agency that grants route permits to bus owners. Political party leaders dominate bus owner associations, and they are represented in the RTC. Therefore, the RTC is the real hub of patronage distribution and has to allocate many more routes than technically feasible to satisfy multiple clients with political connections. Moreover, bus control is strategically im- portant for political parties whenever they want to provide transportation for supporters to the sites of protests and demonstrations. As a consequence, all bus companies need political contacts even to obtain route allocations and scheduling favors. The result is severe overcrowding, increased pollu- tion, and heightened risk of accidents (World Bank 2009). In some cases, specific economic sectors or geographic locations are more prone to patron-client relations than others. The following example illus- trates the intersection between reconstruction and political patronage in Malawi. In 2008, months before the parliamentary and presidential election, the Ndirande market in Malawi burned to the ground. Many small vendors lost all the booths and merchandise on which their livelihood depended. When similar incidents happened in other markets in the past, the recon- struction was relatively fast and efficient, with joint efforts of vendors and local officials. In contrast, in Ndirande, one year later, the first phase of re- construction was not yet completed. Why? The story starts many decades ago, during the nondemocratic period, when the single political party dominated the committee that regulated the market, coordinated vendors, and mediated disputes. At the time, there were no factions in the market-management institution. The advent of multiparty competition politicized the committee and splintered it into three sections, each one backed up by a party. What makes Ndirande special compared to other markets in Malawi is the overlap of ethnicity and political support. Be- cause most vendors are ethnic Yao, Ndirande is the political fiefdom of one of the parties, the UDF, whose leader—a former president—is also Yao. The party leader himself, as well as other politicians, donated money to the mar- 68 Understanding Policy Change ket committee for reconstruction efforts. It so happens that during the recon- struction attempts, a competing party, the DPP, won the national elections and replaced the committee leadership. When vendors went to the office in charge of collecting and disbursing money, they were told that the donated amount was much lower than what was needed and that some of it was “bor- rowedâ€? by an official. Feeling a sense of injustice, vendors went to the DPP office and set it on fire. In retaliation, the incumbent party significantly cut funds for reconstruction and delayed the whole process (Cammack 2011). Obviously, corruption takes a heavy toll on development. Bribes, extor- tion, and unpredictability of costs act as a deadweight tax on businesses, NGOs, and other organizations that may generate wealth and the provision of desirable public goods. To the extent that corrupt officials do not use such bribes to pursue progrowth investment, corruption may reduce the rate of economic growth and modernization of a country. Therefore, it can have the effect of making a society worse off. Effects of Corruption on Resource Allocation. If public policy goals can be easily altered by the payment of bribes to policy makers, some individuals and organizations may devote resources to furthering their own agenda. Be- cause small organized groups (for example, trade unions or business associa- tions) are able to lobby more effectively than larger groups (the general pub- lic), corruption and rent seeking opportunities may facilitate the growth of special interest legislation and regulation. Such regulations are likely to im- pose costs on large groups but generate benefits for only a narrow subset of the population (those who can lobby) (Tullock 1987). For example, protec- tionist tariffs against foreign competitors may increase the profits of an in- dustrial sector even if, in some circumstances, they condemn the general public and other business sectors to consuming more expensive or inferior domestic goods and services. In addition, to the extent that corruption is a form of arbitrary decision making by government and bureaucrats, the increased uncertainty of doing business may result in the outflow of money or in less foreign direct invest- ment in a country (Wei 2000). Even if some countries have simultaneously experienced growth and ­ corruption, they face higher economic risks than contexts in which corrup- tion is not prevalent. Systemic corruption also undermines government legitimacy, efficiency, and equity. Sometimes it can lead to the reversal of the ­ political regime, or it can consolidate forms of predatory nondemocratic rule (Rose-Ackerman 1997). Finally, as political power is correlated with economic wealth, it may be the case that the existence of corruption further entrenches inequality, as Accountability and Corruption: The What Question 69 FIGURE 2.4  The Link between per Capita Income and Corruption 10 Honduras Paraguay 8 Pakistan Uganda Ukraine Russia Ecuador Bangladesh Bolivia Venezuela, RB Colombia Nicaragua India Philippines Thailand Mexico Latvia Bulgaria Zambia Senegal Romania Guatemala Brazil Jamaica Argentina 6 Turkey Slovak Republic corruption Malawi Zimbabwe Belarus El Salvador Uruguay Korea, Rep. Italy Peru Poland Mauritius Czech Republic Greece South Africa Hungary Malaysia Taiwan, China Namibia Estonia Costa Rica Spain 4 Botswana Belgium Portugal Japan France Chile Israel Austria USA 2 Ireland Australia Germany UK Switzerland Netherlands Luxembourg Norway Sweden Singapore New Zealand Iceland Canada Finland 0 Denmark 6 7 8 9 10 (log) per capita income country fitted values Source: Data from Persson and Tabellini (2003). Note: Lower score = less corruption. The perceptions of corruption are an average score of Transpar- ency International’s Corruption Perceptions Index between 1995 and 2000. The per capita data is from the World Development Indicators (1993–1998). the rich and politically powerful are able to protect themselves from the ­ government (through bribes or political action), while more marginalized groups are less able to prevent the abuse of public office for private gain ­(Olson 1982). As seen in figure 2.4, corruption affects—negatively and signifi- cantly—the general level of economic development.10 Can Corruption and Clientelism Have Some Positive Consequences? The earlier literature on the causes and consequences of corruption argued that both low-level and grand corruption may play several political- economic roles in the functioning of developing states. The two types of cor- ruption affect individuals and states in different ways. Small-scale corruption might be a symptom of a state experiencing the inherent tensions of rapid modernization. In some contexts, corruption and clientelism are just the natural expression of social capital; they divert politi- 70 Understanding Policy Change cal claims from the class-based Western-centric left-right axis to ethnic and tribal affiliations (Chabal and Daloz 1999; de Sardan 1999). As for large-scale corruption, under some circumstances, its side effects were thought to lead to more resources for private investment, open up ac- cess to marginalized segments of the population, and compensate for the low level of tax revenues characteristic of many developing states (Leff 1964, Huntington 1968, Bayley 1966, for a brief review see Szeftel 1998, 227). In some developing countries, certain forms of corruption might be a necessary compromise between the state and vested interests, guaranteeing ­ opportunities for corruption in some sectors while gradually creating “pockets of efficiencyâ€? in others (see Nye 1967 for Asia; see Geddes 1996 for Latin America). In other contexts characterized by complicated ethnic, clan, regional, and religious loyalties and by factionalism, political clientelism has been hypoth- esized to help the democratic process. In other words, rather than being the problem or cause of malfunctions, it could allow a second-best equilibrium through which patronage relations and concrete material benefits to selected groups prevent or impede polarization and open conflict (Szeftel 1998). A more recent and related theoretical contribution (North, Wallis, and Weingast 2009) seeks to explain the relationship between rent seeking and political violence. According to this narrative, in some contexts there is a clear trade-off between a certain level of patronage (or corruption) and con- flict, with the implication that suddenly dismantling these rents would have unintended consequences for political stability. In other words, the feasibil- ity of reform is constrained by the probability that violence would erupt if key groups in society stopped receiving the benefits they are used to. In a vicious circle, however, these rents generate a low developmental equilib- rium. In contrast, societies that have solved this dilemma allow economic and political competition and are more developed in general. Nevertheless, the more recent research on the topic overwhelmingly converges toward the negative consequences of corruption: perverse politi- ­ cal and economic incentives, lower levels of accountability, and derailed development. Both analytically and practically, the conceptual angles from which policy makers view corruption and political clientelism have profound repercus- sions for reform trajectories. With few exceptions, the governance reform agendas focus heavily on the civil service and public financial management as classic areas for dismantling political rents. According to several research- ers, however, understanding the root causes of the phenomena might lead to different context-tailored recommendations. For example, efforts to create Accountability and Corruption: The What Question 71 more job opportunities in the private sector could be an effective alternative to combating corruption and clientelism in job-scarce environments where public sector employment is heavily politicized (Grindle 2007; Khan 2006). In some sense, political-economy analysis gives the development practi- tioner both the feasible and the normative answer to such reform questions. “Second-bestâ€? solutions, or “good enoughâ€? governance, might be more sen- sible than textbook “first-best solutionsâ€? (Grindle 2007; Khan 2006; Rodrik 2008; Fritz, Kaiser, and Levy 2009). Why Is Corruption More Prevalent in Some Countries Than in Others? To date, we have a number of cross-national indicators of corruption that allow researchers to distinguish among possible causes. Although subject to continuous debates in academic and policy circles, several explanations have been proposed for variation in corruption among countries (Shleifer and Vishny 1993; Lambsdorff 2006; Treisman 2000, 2007): •  arge size of government L •  Low economic development •  Low openness to trade •  Heavy reliance on fuel exports • Short or nonexistent history of democracy  •  The absence of a vibrant media • Heavy government regulations  •  Federalism • The type of colonial legacy.  What Policy Interventions Have Governments Pursued to Reduce Corruption? Although the menu of policy interventions is diverse and often the effect of specific policies on outcomes is far from clear, a few examples of reforms that governments have pursued to reduce corruption follow: •  eform of civil servants’ wages R •  Investments in a free media •  Establishment of an independent judiciary •  Promotion of some degree of market competition through privatization • Participation in international anticorruption conventions  72 Understanding Policy Change CONCEPTS IN PRACTICE 2.5 The “New Public Managementâ€?—from New Zealand to the World Although initially developed in high-income democracies (especially New Zea- land, the United Kingdom, and the United States), the new public management approach raised hopes that it could be adapted to developing countries to help reduce incentives for corruption and government waste. The main focus of NPM was to disaggregate bureaucracies in smaller units, incentivize competi- tion with the private sector, foster budgetary accountability, and focus on effi- ciently providing citizens with the services they need. Whatever its impact on developed economies, NPM’s effect on developing countries was very limited for several reasons: Many public sector providers did not face real competition from private sec- •   tor companies, a fact further exacerbated by weak judicial institutions. A public sector ethos (that is, service to the community) was not fostered •   by aspects of NPM; therefore, its ability to thrive in a context in which there was little or no social expectation of serving the community limited its effectiveness. Source: Manning (2001).  stablishment of independent anticorruption committees • E Adoption of effective oversight institutions. •  As discussed in Concepts in Practice 2.5, some of these reforms were initi- ated as part of an approach called “new public managementâ€? (NPM) ad- opted with varying success by many governments around the world (Dun- leavy et al. 2011). Summary Accountability and governance are essential ingredients for durable devel- opment. This chapter disentangled the analytical mechanisms at work and suggested that different policy contexts may feature alternative forms of ac- countability. The second part of the chapter went into the details of corrup- tion and political clientelism and patronage and their complex relationship Accountability and Corruption: The What Question 73 to accountability and development. Finally, the chapter briefly summarized common policy solutions for tackling corruption. Notes  1. Reliability is used by Downs (1957); responsibility by Riker (1993); representation by Pitkin (1967); responsiveness by Dahl (1971); mandate responsiveness by Stokes (2001); accountability by Przeworski, Stokes, and Manin (1999); and governance, by Kaufmann, Kraay, and Zoido-Lobaton (1999).  2. This classic definition of political representation comes from Hannah Pitkin (1967). These concepts have a long and complicated intellectual history, but for the purposes of this exposition, it suffices to maintain some level of generality.  3. Needless to say, this example assumes a unidimensional (single issue) policy space and oversimplifies the electoral mechanism. It serves here as just an illustration.  4. This classification builds on Kitschelt (2000, 845–846).  5. The definition of public good versus private and club goods will be elaborated in chapter 3.  6. Typology based on Hadenius and Teorell (2007).  7. This section is based on Kornai (1992, 37–39).  8. Ibid, 56.  9. Adapted from the definition by Kitschelt and Wilkinson (2007, 2). 10. However, the relationship between corruption and economic development is complicated by the direction of causality: which causes which? Some empirical findings contradict the idea that corruption breeds low development and point at potential positive effects on the economy (Rose-Ackerman 2006). References Acemoglu, D., and J. A. Robinson. 2006. Economic Origins of Dictatorship and Democracy. Cambridge, MA: Cambridge University Press. Acemoglu, Daron, James Robinson, and Thierry Verdier. 2004. “Kleptocracy and Divide-and-Rule: A Model of Personal Rule.â€? Journal of the European Economic Association 2 (2/3): 162–92. Ames, Barry. 1987. Political Survival: Politicians and Public Policy in Latin America. Berkeley: University of California Press. Barro, R. J. 1973. “The Control of Politicians: An Economic Model.â€? Public Choice 14 (1): 19–42. Bayley, David H. 1966. “The Effects of Corruption in a Developing Nation.â€? Western Political Quarterly 19 (4): 719-32. Besley, T. 2006. Principled Agents? The Political Economy of Good Government. New York: Oxford University Press. Boix, Carles. 2003. Democracy and Redistribution. Cambridge: Cambridge Univer- sity Press. 74 Understanding Policy Change Borgenicht, D., and T. Regan. 2008. Worst-Case Scenario Almanac Politics. San Francisco: Chronicle Books Llc. Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2003. The Logic of Political Survival. Cambridge: MIT Press. Cammack, Diana. 2011. “Local Governance and Public Goods in Malawi.â€? IDS Bulletin 42 (2): 43–52. Campos, Jose E., and Sanjay Pradhan. 2007. The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level. Washington, DC: World Bank. Chabal, Patrick, and Jean-Pascal Daloz. 1999. Africa Works: The Political Instrumen- talization of Disorder. Oxford: James Currey; Bloomington: Indiana University Press. Cheibub, J. A. 1998. “Political Regimes and the Extractive Capacity of Governments: Taxation in Democracies and Dictatorships.â€? World Politics 50 (3): 349–76. Crystal, Jill. 1989. “Coalitions in Oil Monarchies: Kuwait and Qatar.â€? Comparative Politics 21 (4): 427–43. Dahl, Robert A. 1971. Polyarchy: Participation and Opposition, vol. 54. New Haven: Yale University Press. Deacon, R. T., and S. Saha. 2006. “Public Goods Provision under Dictatorship and Democracy: A Survey.â€? In The Companion in Public Economics: Empirical Public Economics, ed. A. F. Ott and R. J. Cebula, chap. 6. Northampton, MA: Edward Elgar Publishing. Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper. Dunleavy, P., H. Margetts, S. Bastow, and J. Tinker, 2006. “New Public Management is Dead—Long Live Digital Era Governance.â€? Journal of Public Administration Research, and Theory 16 (3): 467–494 Englebert, Pierre. 2009. Africa: Unity, Sovereignty, and Sorrow. Boulder, CO: Lynne Rienner Publishers. Fahey, Dan. 2008. “Le Fleuve D’Or: The Production and Trade of Gold from Mongbwalu, DRC.â€? In L’Afrique des Grands Lacs, Annuaire 2007–2008, ed. S. Marysse, F. Reyntens, and S. Vandeginste. Paris: L’Harmattan. Ferejohn, John A., and James H. Kuklinski. 1990. Information and Democratic Processes. Champaign: University of Illinois Press. Fritz, V., K. Kaiser, and B. Levy. 2009. Problem-Driven Governance and Political Economy Analysis. Washington, DC: World Bank. Geddes, Barbara. 1996. Politician’s Dilemma: Building State Capacity in Latin America. Berkeley, CA: University of California Press. Gehlbach, Scott, and Philip Keefer. 2011. “Investment without Democracy: Ruling- Party Institutionalization and Credible Commitment in Autocracies.â€? Journal of Comparative Economics 39 (2): 123–39. Grindle, M. S. 2007. Going Local: Decentralization, Democratization, and the Promise of Good Governance. Princeton, NJ: Princeton University Press. Haber, S. H., A. Razo, and N. Maurer. 2003. The Politics of Property Rights: Political Instability, Credible Commitments, and Economic Growth in Mexico, 1876–1929. New York: Cambridge University Press. Accountability and Corruption: The What Question 75 Hadenius, A., and J. Teorell. 2007. “Pathways from Authoritarianism.â€? Journal of Democracy 18 (1): 143. Hellman, J. S., G. Jones, and D. Kaufmann. 2000. “Seize the State, Seize the Day: State Capture, Corruption, and Influence in Transition.â€? Policy Research Working Paper 2444, World Bank, Washington, DC. Huntington, Samuel P. 1968. Political Order in Changing Societies. New Haven: Yale University Press. Joseph, R. 1998. “Africa, 1990–1997: From Abertura to Closure.â€? Journal of Democracy 9: 3–17. Kaufmann, D., A. Kraay, and P. Zoido-Lobatón. 1999. “Aggregating Governance Indicators.â€? Working Paper 2195, World Bank, Washington, DC. Keefer, Philip. 2007. “Clientelism, Credibility, and the Policy Choices of Young Democracies.â€? American Journal of Political Science 51 (4): 804–21. Khan, Mushtaq. 2006. “Determinants of Corruption in Developing Countries: The Limits of Conventional Economic Analysis.â€? International Handbook on the Economics of Corruption, ed. Susan Rose-Ackerman, 216–44. Northampton, MA: Edward Elgar Publisher. Kitschelt, H. 2000. “Linkages between Citizens and Politicians in Democratic Polities.â€? Comparative Political Studies 33 (6/7): 845–79. Kitschelt, Herbert, Kent Freeze, Kiril Kolev, and Yi-Ting Wang. 2009. “Measuring Democratic Accountability: An Initial Report on an Emerging Data Set.â€? Revista de Ciencia Política 29 (3): 741–73. Kitschelt, H., and S. Wilkinson. 2007. Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition. New York: Cambridge University Press. Knack, S. 2007. “Measuring Corruption: A Critique of Indicators in Eastern Europe and Central Asia.â€? Journal of Public Policy 27: 255–291 Kornai, Janos. 1992. The Socialist System: The Political Economy of Communism. New York: Oxford University Press. Lake, D. A., and M. A. Baum. 2001. “The Invisible Hand of Democracy.â€? Comparative Political Studies 34 (6): 587. Lambsdorff, J. G. 2006. “Causes and Consequences of Corruption: What Do We Know from a Cross-Section of Countries?â€? In International Handbook on the Economics of Corruption, ed. Susan Rose-Ackerman, 3–52. Northampton, MA: Edward Elgar Publisher. Leff, N. H. 1964. “Economic Development through Bureaucratic Corruption.â€? American Behavioral Scientist 8 (3): 8. Leslie, Winsome J. 1987. The World Bank and Structural Transformation in Develop- ing Countries: The Case of Zaire. Boulder, CO: L. Rienner Publishers. Manning, Nick. 2001. “The Legacy of New Public Management in Developing Countries.â€? International Review of Administrative Sciences 67 (2): 297–312. McPherson, Charles, and Stephen MacSearraigh. 2007. “Corruption in the Petro- leum Sector.â€? In The Many Faces of Corruption. Tracking Vulnerabilities at the Sector Level, ed. Jose E. Campos and Sanjay Pradhan, 191–221. Wash­ ington, DC: World Bank. 76 Understanding Policy Change Meltzer, Allan H., and Scott F. Richards. 1981. “A Rational Theory of the Size of Government.â€? Journal of Political Economy 89 (5): 914–27. Mulligan, C. B., R. Gil, and Xavier Sala-I-Martin. 2004. “Do Democracies Have Different Public Policies Than Nondemocracies?â€? Journal of Economic Perspec- tives 18 (1): 51–74. North, Douglass, John Wallis, and Barry Weingast. 2009. Violence and Social Orders. New York: Cambridge University Press. Nye, Joseph. 1967. “Corruption and Political Development: A Cost-Benefit Analysis.â€? American Political Science Review 61 (2): 417–27. Olson, M. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven: Yale University Press. Persson, T., G. E. Tabellini, and I. Brocas. 2000. Political Economics. Cambridge, MA: MIT Press. Piattoni, S. 2001. Clientelism, Interests, and Democratic Representation: The Euro- pean Experience in Historical and Comparative Perspective. New York: Cam- bridge University Press. Pitkin, Hanna F. 1967. The Concept of Representation. Berkeley: University of California Press. Pole Institute. 2007. Rules for Sale: Formal and Informal Cross-Border Trade in Eastern DRC. Goma: Pole Institute. Przeworski, Adam, Susan C. Stokes, and Bernard Manin, eds. 1999. Democracy, Accountability, and Representation, vol. 2. New York: Cambridge University Press. Riker, William H. 1993. Agenda Formation. Ann Arbor: University of Michigan Press. Rodrik, Dani. 2008. “Second-Best Institutions.â€? NBER Working Paper 14050, National Bureau of Economic Research, Cambridge, MA. Rose-Ackerman, Susan. 1978. Corruption: A Study in Political Economy. New York: Academic Press. ———. 1997. “The Political Economy of Corruption.â€? In Corruption and the Global Economy, ed. Kimberly Ann Elliott, chap. 2. Washington, DC: Institute for International Economics. ———. 1999. Corruption and Government: Causes, Consequences and Reform. Cambridge: Cambridge University Press. ———. 2006. International Handbook on the Economics of Corruption. Northampton, MA: Edward Elgar Publishing. Sen, Amartya K. 1999. Development as Freedom. New York: Oxford University Press. Shleifer, Andrei, and Robert Vishny. 1993. “Corruption.â€? Quarterly Journal of Economics 108 (3): 599–617. Stokes, Susan. 2001. Mandates and Democracy. Neoliberalism by Surprise in Latin America. Cambridge: Cambridge University Press. ———. 2007. “Political Clientelism.â€? In The Oxford Handbook of Comparative Politics, ed. Carles Boix and Susan Stokes, 604–27. New York: Oxford University Press. Szeftel, Morris. 1998. “Misunderstanding African Politics: Corruption and the Governance Agenda.â€? Review of African Political Economy 76: 221–40. Accountability and Corruption: The What Question 77 Transparency International. 2012. Corruption Perceptions Index. http://www. transparency.org/. Treisman, D. 2000. “The Causes of Corruption: A Cross-National Study.â€? Journal of Public Economics 76 (3): 399–458. ———. 2007. “What Have We Learned about the Causes of Corruption from Ten Years of Cross-National Empirical Research?â€? Annual Review of Political Science 10 (1): 211–44. Tullock, Gordon 1987. “Public Choice.â€? In The New Palgrave: A Dictionary of Economics 1040–44, 3 vols. New York: Palgrave MacMillan. Wantchekon, L. 2003. “Clientelism and Voting Behavior: Evidence from a Field Experiment in Benin.â€? World Politics 55 (3): 399–422. Wei, Shang-Jin. 2000. “How Taxing Is Corruption on International Investors?â€? Review of Economics and Statistics 82 (1): 1–11. Wintrobe, Ronald. 1998. The Political Economy of Dictatorship. Cambridge: Cam- bridge University Press. World Bank. 2004. World Development Report 2004: Making Services Work for Poor People. Washington, DC: World Bank. ———. 2005. “The Islamic Republic of Iran. Report on Public Financial Manage- ment, Procurement, and Expenditure Systems in Iran.â€? Washington, DC: World Bank. ———. 2009. “Operationalizing Political Economy: Urban Bus Operations in Dhaka.â€? South Asia Political Economy and Governance Notes 1: 1–8. ———. 2012. World Development Indicators. www.worldbank.org. 78 Understanding Policy Change CHAPTER 3 The Collective Action Problem in Development: The Why Question Chapter 2 explored the most common manifestations of development fail- ure: corruption, patronage, state capture, and the like. Missing or weak ac- countability relations among citizens, politicians, and bureaucrats skew in- centives, allow abuses of power, and lock in suboptimal outcomes. In this chapter, we start asking why accountability fails and poor governance thrives. Why do citizens not act together to challenge their nonperforming represen- tatives? Why are some interest groups effectively shaping policy, while the poor can rarely act collectively to claim public goods? Why do some groups mobilize despite all odds of success, challenge corrupt governments, and take up ambitious pro-development reforms? Once the practitioner or reformer has diagnosed the specific develop- ment problem or the answer to the what question discussed in the previous chapter, joint action of various actors is necessary in many cases to correct it. In legislatures, party coalitions are often needed to form governments or to pass reform bills. Broad platforms of civil society organizations are some- 79 times the solution for reaching a well-functioning social accountability rela- tion between citizens and bureaucrats. Popular protests against corrupt poli- ticians or governments come to life because of individuals and organizations that invest significant resources and efforts in shaping social movements. In some circumstances, the poor obtain favorable outcomes when they vote for political parties committed to bringing development to their con- stituencies. Referred to as collective action, this act of coming together can introduce a host of challenges. If the expected costs of political action are high or if they fall exclusively on a small group of entrepreneurs—whereas the benefits of policy change are equally distributed among actors—reform might never be initiated. Mobilizing relevant stakeholders, building coali- tions, and making claims for policy change require time and dedication and can entail considerable political, financial, and fiduciary risks. The case study in appendix C will take the reader through all the trials and tribula- tions along the path of a real reform in the Philippines. This chapter first introduces the major social dilemmas that citizens and groups face whenever they have to act together for a better development outcome. Going up a level of abstraction, we explore the concept of public goods. In a very broad sense, public goods are the final development goals for which actors and stakeholders have to come together in the first place and from which all of them benefit. The chapter then explores how different groups in society might be more or less able to overcome pro-development collective action problems and discusses the possible implications for policy making (see figure 3.1). FIGURE 3.1  Conceptual Map of Collective Action Problems Collective action problems Motivation Informational problems problems How do How do motivational informational problems inhibit problems hamper collective action? collective action? Source: Authors. 80 Understanding Policy Change Objectives of Chapter 3 By the end of this chapter, readers should be able to do the following:  rasp the concept and typology of private, public, and club goods and • G how they relate to development Understand why social dilemmas occur in development with respect •  to the production and maintenance of public goods or common-pool resources Explore the implications of collective action problems •  Think about some concrete factors that may facilitate or inhibit collective •  action Be familiar with major modes of collective action •  Link collective action obstacles to real-life situations. •  Collective Action Problems: At the Heart of Development The title of this section paraphrases Elinor Ostrom, the recent Nobel Prize winner in economics, who argues that collective action is the answer to the why question (Ostrom 2010). Development deals primarily with the produc- tion of public goods, ranging from provision of basic water and electricity to the release of information about the way governments raise taxes and allocate spending to their constituents. Individuals and organizations often face sev- eral social ­dilemmas whenever the time to join forces and act together comes. These dilemmas refer to the mismatch between individual and group incen- tives to participate in the production of a public good. The logic is quite straightforward: if someone else incurs the costs of pro-development action while I derive full benefits, why would I contribute in the first place? There- fore, the individual’s incentives not to pay the costs but to reap the benefits turn out to undermine the general interest of the group. Consequently, devel- opment efforts are often derailed by short-term self-interest that prevents reform from reaching ultimate outcomes. For instance, if citizens cannot act collectively to sanction corruption and abuses of political power, relation- ships of accountability fail, and distribution of public goods such as basic edu- cation and health suffers. In contrast, at the local or national level, effective collective action, through which actors articulate claims for public goods and monitor implementation, often translates into better development outcomes. Motivation is not the only source of social dilemmas. Large discrepancies in the information available to citizens, voters, or stakeholders make coordina- tion and alignment of individual and group incentives hard to achieve. The Collective Action Problem in Development: The Why Question 81 FIGURE 3.2  Impact of Individual and Collective Incentives on a Public Good Collective incentives Public goods Individual incentives Source: Authors. Figure 3.2 illustrates the impact of the two sets of incentives (individual and collective) on the production of a public good or the maintenance of a common-pool resource. Individual incentives resulting from short-term self-interest tend to undermine development, whereas collective (or group) incentives to act together lead to a superior outcome. This mismatch (or misalignment) between individual and collective incentives is the central so- cial dilemma that hampers efforts to act jointly for the common good. To allow for a better understanding of such a mismatch and its conse- quences on development, this chapter will introduce in greater detail the key concepts of public goods and social dilemmas and will present some of the factors that exacerbate or inhibit collective action problems. What Are Public Goods? The term public goods refers to goods and services that simultaneously meet the following criteria:  hey are nonrivalrous, meaning that one individual’s consumption of the • T good does not reduce the quantity or quality of that good or service for other participants in its consumption. They are nonexcludable, meaning that individuals cannot be stopped from •  consuming the good, even if they pay nothing toward the cost of provid- ing the good. 82 Understanding Policy Change TABLE 3.1  Goods Quadrant—Public Goods versus Club (toll) Goods, Common-Pool Resources, and Private Goods Rivalry of use  Difficulty of excluding  potential beneficiaries High Low     High Common-pool resources or common Public goods: for example, peace goods: for example, lakes, irrigation and security of a community, public systems, fisheries, forests, annual safety, national defense, knowledge, fire budgets, intergovernmental fiscal protection, air, government transparency, transfers in federal systems corruption reduction or eradication     Low Private goods: for example, food, Club (toll) goods: for example, theaters, clothing, general consumption goods private clubs, day-care centers Source: Adapted from Ostrom (2005, 24). Ostrom Elinor. Understanding Institutional Diversity. 2005 © by Princeton University Press. Reprinted by permission of Princeton University Press.  he pure public goods are also nonexhaustible, meaning that the good • T cannot be depleted by consumption (Samuelson 1954). Examples of pure public goods are air, natural landscapes, and national defense. Goods or services that do not meet at least the first two characteristics listed above are called private goods, meaning either that their consumption is competitive or that only those who pay can consume them. Table 3.1 intro- duces a typology of goods based on two dimensions: excludability of use (can members prevent noncontributors from access to the good or service?) and rivalry or subtractability of use (if some users participate in the unrestrained consumption of the good, will this diminish other users’ consumption?) (Ostrom 2005). Most goods and services, however, are never purely public or private but fall somewhere in between. If group members who contributed to the pro- duction of the goods have the power to exclude noncontributors from using the goods or services, these are often labeled club goods. Sometimes, club goods are provided on a large social scale and resemble public goods. Other times, the club size is so small that they are de facto private goods. Diagnosing the development problem implies thinking first about the good to be produced or maintained and its nature. It is well worth keeping in mind that the good or service may constitute a variety of tangible or intan- gible ­outcomes (national defense, water, electricity, information provided by governments to citizens, knowledge, popular protests demanding greater ac- countability, and so on). In chapter 1, we briefly discussed the difference be­­­­ tween INSTITUTIONS (with capital letters) and institutions (lower-cased) (Banerjee and Duflo 2011). To paraphrase the authors, we could also think of PUBLIC GOODS that fulfill the conditions above. They are the broadest The Collective Action Problem in Development: The Why Question 83 and most general examples of nonexcludability, nonrivalry, and nonexhaust- ibility of use that all of us recognize. This type includes the following: •  lean air C • Universal primary education  •  Universal health care •  National defense •  Knowledge • Transparency of government  • Corruption-free bureaucracy  •  Economic growth • Development.  In the broadest possible sense, economic growth and development are pub- lic goods. Therefore, pro-development policy reforms require concerted ac- tion and the coordination of many different actors. Their effects are benefi- cial over the medium or long term for all participants, even if they entail short-term costs. In some cases, policy goals and outcomes that intuitively resemble public goods still generate competitive or excludable behavior among participants. For example, in its pure state, universal primary education is often designed as a public good. However, school location, staffing policies, or delays in teachers’ salaries make it de facto a club good, since certain constituencies who lack access to schools are excluded from its benefits. If school building and staffing occur only in areas of the country that provide political support for the incumbent party, other regions or localities will be deprived of access to education. Similarly, national defense is a public good since everyone ben- efits from security, but some actors—for example, government contractors— derive substantial private benefits from it (Aranson and Ordeshook 1985; Bueno de Mesquita et al. 2002). Often, if not always, public policies are mix- tures of private and public goods. As an analytical starter, the following questions might be useful: •  hat is the good to be produced or maintained? W •  Who are the producers of the good? •  Who benefits from the consumption of the good? • How does it score on the three dimensions: nonrivalrous, nonexcludable,  and nonexhaustible? Is the good or service private, public, or club? •  What are the costs and benefits of public good production or common- •  pool resource maintenance? 84 Understanding Policy Change  re they concentrated (incurred by a narrow subgroup only) or diffused • A (spread out widely in the group or population at large)? Diagnosing development problems implies clarifying first the what of re- form. The answers to this question can take many forms and shapes, includ- ing underproduction of the public good or inadequate maintenance of a common-pool resource. Once we grasp what exactly is being underproduced or undermaintained, we can then analytically move to understanding the individual and group incentives surrounding the good or service. ­ What Factors Inhibit Collective Action? Two general categories of factors lead to the collective action problem or to the difficulty groups have in mobilizing effectively for policy change. First are motivational problems: if somebody else can provide the public good without contributions from those who will benefit, there is always the temp- tation to free ride. Free riding means participating in the consumption but not in the production of the good. Second are information problems, meaning either that groups lack crucial information or that one group of stakeholders derives power from having more information than others (Ostrom 2010; Gibson et al. 2005). Motivational Problems and the Temptation to Free Ride The incentive to free ride—or, simply put, to reap benefits without incurring costs—occurs for a variety of reasons. In some circumstances, as rational choice theory would predict, it is a simple matter of individual morality. In such situations, the free rider is a typical case of noncompliance with the rules of the group or other community norms. In other cases, however, even if individuals (or groups) are willing to participate in the production of pub- lic goods, they might not have the necessary resources to do so. For example, the poor can rarely afford to make the initial individual investment that a public good requires, irrespective of their rational calculations, incentives, or sense of morality (McKean 2000). Not paying water fees, for example, could mean either classic free riding or the inability of individual water users to afford the fee. Collective action, or the act of mobilizing stakeholder groups to produce or maintain a good, is also challenging because it entails additional types of investments: pooling of costs and pooling of preferences. The former suffers from the inherent free-rider incentives just discussed. The latter dimension implies that, in addition to the individual motivation to contribute or not to The Collective Action Problem in Development: The Why Question 85 the collective efforts, the way in which all preferences of actors get aggre- gated to produce policy outcomes can obstruct or promote certain modes of collective action. Preference aggregation depends not only on resources but also on the similarity of views and goals, as well as on information available to groups and individuals in society. Often, the poor fail to mobi- lize and ­effectively claim public goods precisely because they lack infor­ mation about other groups that might have common interests. In addition, poverty—understood as the number of people whose income is lower than a given threshold—is in continuous flux. Some poor households, neighbor- hoods, or generations manage to escape it; others fall into it despite not being born poor (Krishna 2010). These differences in social mobility across groups and individuals sometimes induce significant heterogeneity of policy interests and goals, lack of cohesion, and overall disincentives to act jointly for specific policy outcomes. In other contexts, even if collec- tive action happens and groups of actors decide to pursue reform together, divergent views with respect to either final policy goals or the very process of reaching them can easily dismantle coalitions. As we will see in the next chapter, institutions arise to solve collective action problems (albeit not always ­pro-development collective action) and to facilitate cost and pref- erence pooling. Let us briefly unpack the two categories of collective action problems. Motivational problems usually stem from a high individual ratio of costs to benefits induced by the size and heterogeneity of the group trying to mobi- lize (Olson 1965, 48 for the first three categories):  he larger the group, the smaller the fraction is of the total benefit an in- • T dividual receives from cooperating. The larger the group, the less likely it is that some subgroups in society •  will mobilize to obtain the public good, given the ratio of costs to benefits they would incur. The larger the group, the greater are the organizational costs (in the ab- •  sence of coercion or other incentives from third parties). The more heterogeneous the assets, preferences, information, stakes, or •  identity of the participants, the less likely it is that collective action will be successful. For example, the poor and the rich, or members of different ethnic groups, more often than not fail to cooperate in reaching policy goals that might be beneficial to all. As a result, large numbers of the poor or diffuse citizens are less likely to solve their collective action problems than, say, corporate interest groups. The discrepancy is indeed at the very heart of development, as it leads to skewed outcomes. 86 Understanding Policy Change Information Problems Information problems, in contrast, plague groups of all sizes. Information asymmetries occur when an actor has information that others do not have. In this context, stakeholders who know more tend to have significant bargain- ing advantages over other groups and, hence, have high stakes in conserving such discrepancies. Voters often lack basic information about the electoral process, rules, procedures, political candidates, and policy options. Low turnout generated by missing information prevents groups from electing the candidate or party that would best represent them in legislatures and is thus a manifestation of a collective action problem. Opaqueness of budgetary pro- cesses and general lack of transparency about government activities impede the coordination of civil society or opposition groups around policy issues that could trigger greater accountability. A later chapter is dedicated entirely to informational constraints in policy change. In addition to missing information, moral hazard and adverse selection, two special cases of information asymmetries that will be elaborated in chapter 7, can be serious obstacles to collective action. Moral hazard refers to a situation in which the better-informed individuals or actors engage in risky actions for an organization or collective precisely because they do not bear the full costs of such actions. Leaders of political parties, social movements, or reform teams do not always know enough about the real incentives and stakes of the rank and file within their group. This lack of information about the true actions of the members of an organization often does not allow lead- ers to adequately monitor commitment to the movement or party. For ex- ample, in many countries legislators may win an election because of their association with a party with a reputation for competency, prudence, and the like. However, if after the election such legislators can support risky bills and easily change political parties for opportunistic reasons, without penalty or negative career consequences, the collective action capacity of parties to pursue policies in parliaments is greatly reduced, and volatility increases. Adverse selection refers to circumstances in which, for example, the lead- ers of organizations or coalitions have no way of properly assessing their new recruits’ or members’ true degree of individual commitment to the or- ganizational goals. In cases of adverse selection, the actors who lack ideo- logical or organizational commitment are also paradoxically more likely to join. For example, if a civil society organization offers concrete material ben- efits to all the new recruits to a citizen group advocating social accountabil- ity, as opposed to only the ones who prove commitment, it might attract members who are not interested in collective action per se and greater accountability but rather in the short-term tangible handouts. In this case, ­ The Collective Action Problem in Development: The Why Question 87 adverse selection during recruitment may attract high numbers of partici- pants with low true commitment to collective action. Acquiring the necessary information to act collectively is costly. Political parties, for example, reduce significantly the voters’ cost of collecting in­ formation about policy platforms and electoral candidates, as well as the politicians’ costs of disseminating information to voters during campaigns (Aldrich 1995; Kitschelt 2007, 525). Imagine a world without parties: indi- vidual politicians would have to bear the entire informational costs of cam- paigns during the process of running for office, and individual voters would have to gather information on all candidates’ policy positions, with very lim- ited or nonexistent prior knowledge of them. In this sense, parties solve in- formational and coordination problems and thus facilitate collective action. Papua New Guinea is one of the very few countries that has democratic com- petition without political parties. Social Dilemmas of Delivering Public Goods Political economists commonly use three closely related models to demon- strate the social dilemmas of cooperation between individuals and groups in society for achieving the delivery of public goods: the prisoner’s dilemma, the tragedy of the commons, and the so-called logic of collective action. Think of these models as metaphors for human interaction. They are short parables that illustrate the strategic reasons underpinning an important puzzle: why cooperation for public goods and development does not occur even when it is clear that it would make everybody better off. In fact, we play varieties of the prisoner’s dilemma game, and we face tragedy-of-the-commons chal- lenges in our daily lives: in the family, at work, and in social settings. In large communities, these problems only get worse—especially in the absence of institutions that can correct for misaligned incentives. Prisoner’s Dilemma The prisoner’s dilemma refers to a situation in which actors achieve subop- timal outcomes because they distrust each other in a strategic environment. The exercise of the next chapter will explain in greater detail why prisoners are the main characters of the story. A basic setup for a two-player prisoner’s dilemma game demonstrates that in the absence of institutions that can shift the equilibrium, two individual actors decide not to cooperate to produce the good that would have benefited them both collectively. The prisoner’s dilemma is an often-used analytical model that helps us understand why in- 88 Understanding Policy Change dividuals or groups do not cooperate in the realization of optimal develop- ment or reform outcomes (producing or maintaining the good). The Game Theory 3.1 exercise following this chapter demonstrates how this concept applies to a community of farmers deciding to cooperate on joint irrigation work. Despite the fact that contributing to the building of an irrigation sys- tem would have benefited both farmers, neither of them trusts that the other will want to incur costs while she could just reap benefits, and the system is never built. This behavioral deadlock is especially prevalent in cases in which there have been no previous interactions or trust-building mecha- nisms developed between the two farmers. Additionally, large perceived dif- ferences among potential cooperators may induce a focus on dissimilarities rather than a search for common grounds. For instance, to give a real world example, in Tamil Nadu, India, and Guanajuato, Mexico, farmers who be- longed to irrigation organizations were significantly more likely to cooperate for water conservation if they perceived each other as equals in landholding and social status (Bardhan and Dayton-Johnson 2007, 123). Tragedy of the Commons The tragedy of the commons is a multiplayer prisoner’s dilemma and illus- trates a scenario in which many actors attempt to maximize their individual consumption of the good. However, by doing so collectively, they deplete or exhaust the good, with negative results for everyone. The tragedy of the com- mons usually plagues the use and maintenance of common-pool resources as diverse as the environment; forests, lakes, and open land; fish and wildlife; budgetary processes in parliaments; and intergovernmental fiscal arrange- ments in federal or decentralized systems. As an explanatory tool and diag- nostic, the tragedy of the commons is particularly useful in cases in which the long-term maintenance rather than the production of the good is at stake. We use the management of fisheries to illustrate this collective action problem. Many local communities around the globe depend on fishing for their livelihood. Individual fishermen have incentives to overuse resources. If they fail to cooperate by setting and enforcing consumption ceilings, they reduce the overall population of fish. For example, in Bodrum, Turkey, and in the Bay of Izmir several obstacles impeded such cooperation in the 1980s: overcrowding, quick gain from fishing, high demand associated with the tourism industry, and, above all, the lack of a viable institution that would mediate conflict between competing groups. In Bodrum, for example, the rules that prevented overfishing were not enforced by the Ministry of Agri- culture. In addition, the time horizons and stakes of the fishermen groups were too large and too heterogeneous. Because they could not mobilize to The Collective Action Problem in Development: The Why Question 89 enforce compliance with fishing rules, the result was a depletion of fish and an overall decline in the industry. By contrast, the Alanya community found innovative ways to solve its collective action problem. How? In the 1970s, the fish population reached low levels, and violent conflicts among local fisher- men erupted. Therefore, a local co-op experimented with a trial-and-error system of allocating fishing licenses and monitoring compliance. Every year, the lists of fishing sites were endorsed by all individual fishermen and stored with the local mayor and police. Second, a lottery decided on the assignment of fishermen to a specific site. Third, to provide equal access to the most pro- ductive fishing sites, between September and January each fisherman had to move east each day and rotate the lot. This local arrangement managed to mediate disputes, enforce norms of fairness, prevent overcrowding, and optimize the capabilities of each site (Ostrom 1990, 19, 144–146). The Logic of Collective Action in Small and Large Groups The logic of collective action1 alludes to a similar mismatch between indi- vidual and collective incentives and explains theoretically why large groups in society lack effectiveness and power when they attempt to mobilize politi- cally. According to this logic, group effectiveness relies on members’ assess- ment that sufficient individual gain will come to them through participation or that free riders (those who do not share the cost of producing the good) will not receive the same level of benefits: that is, for collective action to suc- ceed, the gains that result from collectively advocating for a policy outcome must significantly exceed the costs of mobilization. This misalignment of individual and group incentives means that only certain groups that derive high individual benefits with lower administrative costs of mobilization (business lobbies, for example) are likely to mobilize successfully and pursue favorable outcomes. In contrast, diffuse groups of citizens affected by development policies or reform are less likely to mobi- lize. Hence, the power, resources, and intensity of the actors’ preferences will be severely unbalanced. As a consequence, the final outcome may be skewed toward groups that can effectively solve collective action problems. The electric power sector in India provides an example (Lal 2006). It is estimated that up to three-fourths of poor Indian households, especially in the rural areas, lack access to electric power. What are the observable causes? The public sector owns 90 percent of electricity production, is politically captured, and is plagued by severe rent seeking and corruption. Regressive subsidies benefiting large farmers contribute to a large sector deficit, esti- mated to have reached up to 1.5 percent of gross domestic product. Reform attempts began in the early 1990s and have been at the top of electoral party 90 Understanding Policy Change platforms ever since. Despite these efforts, severe underdelivery of electric- ity still persists. In the federal system of India, most of the political obstacles related to electricity provision occur at the state level, where rich farmers benefit dis- proportionately from subsidies at the expense of all other users. The reliance on electric pumps for irrigation varies dramatically among states (12 percent in Rajasthan; 67 percent in Punjab), but across the board, medium and large farmers who own more than two hectares of land and benefit from subsi- dized electric pumps have tended to block reforms that would lead to the reduction of subsidies. Surveys have consistently indicated that users (in- cluding small farmers) would be willing to pay realistic electricity prices if the quality of delivery were adequate. Despite high political salience and users’ willingness to contribute to the production and maintenance of the public good, vested interests (large farm- ers) have traditionally blocked effective change that would entail subsidy reduction. What explains their success? First, big farmers are more politi- cally consequential than the poor. As landowners and notables in their com- munities, they have the capacity to deliver many votes for the incumbent parties in exchange for benefits such as subsidies. Second, large farmers are better organized than the rural users who have a lower socioeconomic pro- file; therefore, they manage to lobby and secure subsidies at the expense of widely scattered rural communities that lack the capacity and resources to act collectively. Another similar collective action problem involving agricul- tural interests is discussed in Concepts in Practice 3.1. These three theoretical constructs illustrate classic social dilemmas and emphasize the key idea that self-interested, rational individuals acting inde- pendently of each other have strong incentives not to cooperate (or to free ride) on the production or delivery of public goods (the prisoner’s dilemma and the logic of collective action) and to deplete a limited shared resource (tragedy of the commons), even if long-term cooperation would be in their best interest. Accordingly, an individual will only rarely, contingent on the size of the group and the cost-benefit ratio, have incentives to join in the production of public goods. The immediate implication is that only powerful and cohesive interest groups are successful in achieving favorable policy. Identifying and Evaluating Collective Action Problems When identifying a collective action problem in the course of a political- economy analysis, the analyst should do five things: (1) understand the type The Collective Action Problem in Development: The Why Question 91 CONCEPTS IN PRACTICE 3.1 Prices of Agricultural Products in Sub-Saharan Africa One of the seminal political-economy studies of lobbying power managed to obtain generous agricultural policies in Africa found that the dif- subsidies from the state. This means that in ferent potential for collective action of various countries with greater populations of large stakeholder groups best explained the rural and farmers (Côte d’Ivoire and Kenya, for instance) urban differences in development (Bates 1981). the agricultural policies were more favorable to In East African urban centers, oligopolistic in- the rural dwellers than in contexts character- dustries and labor unions wanted high con- ized, paradoxically, by a more equal distribution sumer prices for the manufactured goods they of land among smaller farmers. produced but low food prices, since 60 percent Why did small-scale farmers lack the capacity of the budget of an average urban dweller was for collective action that would have enabled spent on food. As a consequence, the state bu- them to mobilize against the state and effec- reaucrats (representatives of marketing boards) tively demand change? First, the costs of politi- offered farmers prices below the world market cal voice were too high, as small farmers tend to price, using the difference either as rents or for be more geographically dispersed and the state the political co-optation of well-organized urban occasionally relied on coercion to enforce unpop- interests. Thus, both urban labor and capital ular policies. Second, in some circumstances, were allied against farmers and managed to these farmers have discovered alternative exit lock in policies that favored them at the ex- routes that allowed them to avoid the policy trap: pense of rural areas. Among farmer groups, be- the use of cross-crops that combined different cause of their weak capacity for collective ac- crop prices, hedging financial risks, abandoning tion, only small farmers suffered the policy joint ventures that were not advantageous, and consequences. In contrast, larger farmers with labor market migration, among others. Source: Bates 1981. of public good to be produced or maintained; (2) think analytically about the type of mismatch between individual and collective incentives to cooperate or join forces; (3) carefully examine the factors that led to the dilemma (motivational or informational) in the first place; (4) sift through the poten- ­ tial political-economic roots of these factors (does the motivational problem occur because of rational calculations, community or cultural norms, or discrepancy in resource endowments among individuals or groups?); and ­ (5) consider whether correcting the collective action problem might lead to another similar dilemma (for example, will punishing noncompliant mem- bers of the group require a few compliant members to pay the full costs of enforcing the rules, whereas the benefits of enforcement extend to all?). Often, adequate answers to these five questions have major implications for policy solutions. For example, if free riding occurs because of strictly ra- 92 Understanding Policy Change tional considerations, monitoring and sanctioning might work well (Ostrom, Gardner, and Walker 1994). If, however, the root cause of free-riding behav- ior resides in poverty or, more broadly, in the incapacity (as opposed to lack of will) to contribute to the public goods, positive selective incentives for participation could do a better job of solving the problem. On a more granular empirical level, research has identified a series of con- crete factors that hamper mobilizational effectiveness and generate the logic of the collective action problem:  he number of interest groups. Theoretically, the larger the number of in- • T terest groups, the more likely it is that they would offset each other’s influ- ence in shaping government policy. However, empirical research suggests that the number of such groups might not necessarily be a crucial factor. Many interest groups can receive policy concessions from the government if they are able to mobilize effectively (Bates 1981; Keefer 2004, 251). Cohesion of the (economic) sector. The less fractionalized the mobilizing •  group representing a certain sector, industry, or constituency is, the more likely it is to have an impact on policy (Frieden 1991, Keefer 2004, 252). Magnitude of gains or losses caused by a potential policy change. Higher •  stakes increase interest group effectiveness (Frieden 1991). For example, firms that are capital intensive, operate in natural resource extraction and management, and would incur high costs in redeploying their assets, have a stronger interest in acting collectively and deriving policy gains from governments (Bates 1981; Keefer 2004, 252). The number of participants and the risks entailed by collective action. As •  already argued, large groups do not often organize effectively to act in their shared interests. In many contexts, the poor or the consumers fail to influence public policy, despite their large numbers. Also, not all forms of collective action pose the same risks of participation. It is certainly less risky for individuals to take part in participatory budgeting exercises or- ganized by civil society organizations in cooperation with municipal gov- ernments than to protest against corrupt political leaders who do not shy away from using force against contenders. Some rough estimations of the number of individuals who support a cause and choose to actively partici- pate in contentious forms of collective action to achieve it (protests, dem- onstrations, and the like) suggest the so-called 5 percent rule: only 5 per- cent of all members of large social groups that would benefit over the long term from the outcome get involved, with the number of activists being overwhelmingly exceeded by bystanders (Lichbach 1998, 12, 17). In con- clusion, smaller groups with high stakes have a greater likelihood of orga- nizing effectively, potentially to the detriment of larger groups. The Collective Action Problem in Development: The Why Question 93 Are individuals and groups thus doomed to be trapped in collective action dilemmas? Despite both motivational and informational problems that lead to social dilemmas and plague collective action, some groups do manage to solve their incentive problems, organize, and attempt to lobby the govern- ment on behalf of their policy interests. Then the question becomes, When and under what conditions are these groups successful in obtaining their preferred policy result? The good news is that despite grim predictions, nu- merous studies confirm that several collective action problems are overcome every day (Ostrom 2000, Banerjee and Duflo 2011). The examples of mobilizational incentives from India and Africa intro- duced above focused on powerful interest groups that manage to act jointly. These are not the only modes of collective action in town. Political parties, for example, routinely aggregate policy and ideological preferences, information flows between voters and politicians, discipline members who do not comply with organizational rules, and mediate interac- tion between citizens and decision makers. Social movement organizations often rely on their core activists who manage to mobilize many participants in protests, demonstrations, or other forms of popular mobilization against bad governance. Reform teams sometimes develop as part of preexisting pol- icy networks and bypass social dilemmas. For pro-development collective action, all these modes often build coalitions. How do these modes of collec- tive action relate to accountability and a better provision of public goods (see Concepts in Practice 3.2)? The subsequent chapters explore in depth how delegation or informa- tion constraints in organizations impede effective collective action for pol- icy change. In addition, even when it occurs, agenda manipulation and lack of credible commitment might still prevent collective action from fully translating into the final development outcomes. Chapter 9 will review the broad repertoire of possible solutions to collective action problems in detail. Game Theory and Collective Action: Modeling Social Dilemmas with Nash Equilibria Collective action problems can be understood and analyzed using the basic tools of game theory.2 In essence, a collective action problem is a situation in which actors have no incentive to organize and obtain a favorable change in the status quo. That is, they are “stuckâ€? in a suboptimal equilibrium. Game theory has a specific name for all situations, whether optimal or sub- 94 Understanding Policy Change CONCEPTS IN PRACTICE 3.2 Political Parties as Collective Action Problem Solvers One of the greatest controversies in the political economy of development is whether democratization fosters development by ensuring that large groups of voters can hold policy makers to account. The fact that many East Asian coun- tries appeared to pursue successful development outcomes despite having nondemocratic regimes that could, in theory, engage in arbitrary rule seems to partially refute this claim. The existence of institutionalized incumbent parties with clear rules of selection, succession, monitoring, and sanctioning created an alternative mechanism through which broad coalitions of stakeholders could act jointly and keep the unelected leaders accountable. More broadly, the exis- tence of institutionalized parties has been linked by some studies to better de- velopment outcomes. In contrast, weak accountability of party leaders to their members and constituencies might not automatically foster development even in democratic contexts. Source: Keefer (2011). optimal, in which actors have no incentive to change their behavior: Nash equilibria. A Prisoner’s Dilemma3 Two neighboring farmers grow a crop to feed their families, and if there is some surplus, they sell to the nearby market.4 The farms are equidistant to a nearby river. If the farmers could build an irrigation system and bring water to the farms, then the crops they could grow and harvest would increase dramatically, and they would be able to send their children to a nearby school. If the irrigation system is built, neither farmer can be excluded from its use (free-rider problem). Furthermore, building the irrigation system is costly and would require both farmers to work together to build the ditch, if they were still to have time to grow their crops. If one farmer tries to build the irrigation system alone, she will succeed; but because she will not have time to grow any crops, she would not be able to feed her family. The preferences of Farmer 1 (the farmers’ preferences are symmetrical), based on the payoffs she receives, are as follows:  f Farmer 1 does not help build the irrigation system but Farmer 2 does, • I Farmer 1 receives the benefit of the irrigation system but does not bear any of the costs. From a purely selfish perspective, this is the optimal out- The Collective Action Problem in Development: The Why Question 95 come (highest payoff ) for her. In this scenario, Farmer 2 bears the entire cost of the project but does not live to enjoy the benefits, as she had no time to tend to her fields. If Farmer 1 builds the irrigation system by herself, she bears the entire •  cost of the project but does not get to enjoy the benefits as she has no crops to harvest. Conversely, Farmer 2 enjoys the benefits of the irrigation system but bears none of the costs. If both Farmer 1 and Farmer 2 build the irrigation system, they both bear •  some of the costs (less time to grow their crops), but they both enjoy the benefits of higher crop yields. If neither farmer builds the irrigation system, they will enjoy a normal •  crop yield and incur no costs. The diagram below conveys all this information but in a more concise format. There are two players, just as in the scenario of the two farmers. Each player has two options: cooperate or defect (not cooperate). The choices the players make determine their payoffs (benefits versus costs). To generalize the findings, the diagram uses the ordinal rankings of the players; a higher number denotes more utility—greater crop yield—allowing the farmer to buy more food for her family minus any costs associated with building the irrigation system and minus the initial loss of time tilling the field in order to dig the ditches. Cooperate Defect Cooperate 3 3 1 4 Defect 4 1 2 2 Now let us think about how we can use the information to make predic- tions about what the farmers will do (hint: look back at the definition of the Nash equilibrium above). Let’s start with Farmer 1. Like any rational person, she wants to maximize her payoff (that is, maximize her crop yield and minimize her costs). How- ever, she also knows that this outcome does not depend only on her actions but also on those of Farmer 2. Therefore, to decide what her optimal strategy should be, she needs to think about what Farmer 2 might do and condition her strategy on that expectation. She can do this by running the following two-step thought experiment: if she knows that Farmer 2 would definitely build the irrigation system whether she herself cooperates or not, then her best action would be to defect (as she would get a payoff of four instead of three, the benefits of the ­ irrigation system but none of the costs): 96 Understanding Policy Change Cooperate Defect Cooperate 3 3 1 4 Defect 4♦ 1 2 2 If she knows that Farmer 2 will definitely choose to defect, then her best ac- tion would also be to defect (a payoff of two versus a payoff of one): Cooperate Defect Cooperate 3 3 1 4 Defect 4♦ 1 2♦ 2 Now let’s think about Farmer 2. Like Farmer 1, she knows that her payoffs depend on the strategy of the other farmer. Through the same “thought experiment,â€? it is possible to identify the strategy of Farmer 2. If she knows that Farmer 1 will definitely choose to build the irrigation system, then her best action would be to defect (as she would get a payoff of four instead of three): Cooperate Defect Cooperate 3 3 1 4♦ Defect 4♦ 1 2♦ 2 If she knows that Farmer 1 will definitely choose to defect, then her best ac- tion would also be to defect (a payoff of two versus a payoff of one): Cooperate Defect Cooperate 3 3 1 4♦ Defect 4♦ 1 2♦ 2♦ Given this scenario, both farmers have an incentive to defect, given how they expect the other farmer to react. As neither farmer can be made better off by unilaterally cooperating, this outcome constitutes a Nash equilibrium. The prisoner’s dilemma game is a simple but powerful tool for conveying how socially suboptimal outcomes can emerge from the rational behavior of individuals. For example, citizens in many countries would benefit if they could cooperate to fight corruption or perceived injustices. However, be- cause every citizen would benefit from such an outcome, every citizen has an The Collective Action Problem in Development: The Why Question 97 incentive to try and free ride—let others fight for change—and, therefore, nobody cooperates. Summary This chapter has introduced collective action problems and explained how they can affect the provision of public goods. Given that some groups or modes of preference and cost aggregation may be more effective at overcom- ing collective action problems than others, unequal collective action capa- bilities can result in the persistence of skewed development outcomes. In addition to being symptomatic of underprovided public goods, collective ac- tion problems often plague many attempts at reform. This chapter helps policy entrepreneurs analytically identify and assess the costs, benefits, and feasibility of concerted action before they engage in the change process. 98 Understanding Policy Change Exercises for Chapter 3: Collective Action Dilemmas Exercise 3.1: The Unscrupulous Diner’s Dilemma5 The facilitator explains the steps involved in the Unscrupulous Diner’s Dilemma. Steps in the exercise (time frame: ~40–45 minutes): 1. Read out instructions (1 minute) 2. Allow participants to read and answer (5–10 minutes) 3. Collect responses, tally responses, facilitate group discussion (10 minutes) 4. Class discussion (10 minutes) 5. Technical explanation (5–10 minutes) Preparation and materials: • Envelopes to be placed in the middle of the round table (one at each table) • Pens should be available to all participants • Copies of the handout of instructions • Copies of the answer sheet • Handout of key questions and group discussion • Facilitator instructions NOTE: There are three different scenarios to be used. Each table works through one of the scenarios. Procedure: 1. The facilitator reads aloud the following instructions: I am going to distribute instructions and an answer sheet. Read the in- structions and answer the questions on the answer sheet. Once you have read the instructions and completed your answers, place your answer sheet in the envelope in the middle of the table. DO NOT WRITE YOUR NAME ON EITHER DOCUMENT and MAKE SURE NO ONE SEES YOUR ANSWERS. Keep the instructions handy for the follow-up discus- sion. I will give you about 5–10 minutes to complete this. 2. The facilitator distributes the instructions handout and the answer sheet to all the participants. 3. The facilitator gauges the rate of completion of the assignment and gives participants a two-minute warning so that the activity can wrap up within its allotted 10-minute time frame. As part of the two-minute warn- ing, the facilitator reminds participants that they should not write their The Collective Action Problem in Development: The Why Question 99 names on the answer sheet; when they have finished, they should put their answer sheet in the envelope in the middle of the table. 4. The facilitator explains: I will come around to collect the envelopes with the answer sheets and tally your responses. At the same time, I will give you a handout that in- cludes two key questions. While I tally the responses on the answer sheets, please go ahead and discuss at your tables the two questions on the handout. Spend approximately 10 minutes in this discussion. 5. To tally the votes, the facilitator uses a blank answer sheet and notes for each option the total number of votes. 6. The facilitator walks around the room among the tables in a nonintrusive manner and pays attention to what is discussed. The facilitator notes a few comments from the tables that he or she can highlight as part of the transition to the lecture portion. 7. The facilitator gives the participants a two-minute warning to wrap up their discussion. A simple way to illustrate the tragedy of the commons is to remind par- ticipants of the coordination game (chapter 4). The suboptimal equilib- rium (3,3) reflects a situation in which players do not cooperate (over- graze the commons), while the optimal equilibrium (4,4) represents a situation in which collective action problems are somehow resolved. Coordination Game Cooperate Unilateral Cooperate 4 4 1 1 Unilateral 1 1 3 3 8. The facilitator explains: Now that you have played the game and have had a chance to discuss with others at your tables some of the key underlying issues, let’s find out the results of your individual decisions. 9. The facilitator announces the results and poses the following questions to the entire group:  1. How many of you were surprised by the entire group’s final tally?  2. If yes, why?  3. If no, why? 10. The facilitator then explains the case study activity. 11. The facilitator concludes with an explanation of the postexercise activity: 100 Understanding Policy Change Exercise 3.1: Instructions Handout 1 You and some acquaintances are at a restaurant and have decided to split the bill evenly among yourselves, regardless of what anyone ordered (splitting the bill = dividing the total bill by the number of people at the table). There are two items on the menu: (1) an expensive meal and (2) an inexpensive meal. If you had to pay the full cost of the meal yourself, you would prefer the inexpensive meal to the expensive meal. However, if you think that one or two others around the table will choose the inexpensive meal and the rest might choose the expensive meal, you would prefer to buy the expensive meal because you will be paying the average cost of the entire meal. The worst-case scenario, from your perspective, is that most other people or everyone else orders the expensive meal and you order the inexpensive meal, resulting in your paying a higher price for the inexpensive meal. You know a few things about the other dinner guests: One dinner guest is on a diet and the inexpensive meal has a lower caloric •  content. Another dinner guest has health complications and does not like to eat too •  much, and the inexpensive meal is a smaller portion. The Collective Action Problem in Development: The Why Question 101 Exercise 3.1: Instructions Handout 2 You and some acquaintances are at a restaurant and have decided to split the bill evenly among yourselves, regardless of what anyone orders (splitting the bill = dividing the total bill by the number of people at the table). There are two items on the menu: (1) an expensive meal and (2) an inexpensive meal. If you had to pay the full cost of the meal yourself, you would prefer the inexpensive meal to the expensive meal. However, if you think that one or two others around the table will choose the inexpensive meal and the rest might choose the expensive meal, you would prefer to buy the expensive meal because you will be paying the average cost of the entire meal. The worst-case scenario, from your perspective, is that most other people or everyone else orders the expensive meal and you order the inexpensive meal, resulting in your paying a higher price for the inexpensive meal. You know a few things about the other dinner guests:  wo diner guests are known to “enjoy their foodâ€? and will definitely be • T ordering the expensive meal. 102 Understanding Policy Change Exercise 3.1: Instructions Handout 3 You and some acquaintances are at a restaurant and have decided to split the bill evenly among yourselves, regardless of what anyone orders (splitting the bill = dividing the total bill by the number of people at the table). There are two items on the menu: (1) an expensive meal and (2) an inexpensive meal. If you had to pay the full cost of the meal yourself, you would prefer the inexpensive meal to the expensive meal. However, if you think that one or two others around the table will choose the inexpensive meal and the rest might choose the expensive meal, you would prefer to buy the expensive meal because you will be paying the average cost of the entire meal. The worst-case scenario, from your perspective, is that most other people or everyone else orders the expensive meal and you order the inexpensive meal, resulting in your paying a higher price for the inexpensive meal. You know a few things about the other dinner guests:  wo of the other dinner guests, who are close friends, are in financial dif- • T ficulty, and having to split an expensive bill would certainly not help their finances. The Collective Action Problem in Development: The Why Question 103 Exercise 3.1: Answer Sheet 1 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Buy the expensive meal Buy the inexpensive meal Briefly explain why you decided to select one meal rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. 104 Understanding Policy Change Exercise 3.1: Answer Sheet 2 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Buy the expensive meal Buy the inexpensive meal Briefly explain why you decided to select one meal rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. The Collective Action Problem in Development: The Why Question 105 Exercise 3.1: Answer Sheet 3 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Buy the expensive meal Buy the inexpensive meal Briefly explain why you decided to select one meal rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. 106 Understanding Policy Change Exercise 3.1: Key Questions Handout – Group Discussion In your group, discuss the following questions: 1. How did the fact that you were splitting the bill affect your decision to select one meal instead of the other? 2. How did the information you had about the other participants affect your decision? The Collective Action Problem in Development: The Why Question 107 Notes 1. The logic of collective action was first introduced by Mancur Olson in 1965. 2. Readers interested in a more formal definition can consult the technical appendix (appendix B). 3. Source: Inspired from (Bardhan 1993). 4. Instructor note: If in a classroom setting, do exercise 3.1 (found at the end of the chapter) before having participants read this section. 5. Source: Inspired from (Gneezy et al. 2004). References Aldrich, John H. 1995. Why Parties? The Origin and Transformation of Party Politics in America. New York: Cambridge University Press. Aranson, Peter H., and Peter C. Ordeshook. 1985. “Public Interest, Private Interest, and the Democratic Polity.â€? In The Democratic State, ed. Roger Benjamin and Stephen Elkin, eds., 87–177. Lawrence: University Press of Kansas. Banerjee, Abhijit, and Esther Duflo. 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. New York. Public Affairs. Bardhan, P. 1993. “Analytics of the Institutions of Informal Cooperation in Rural Development.â€? World Development 21(4): 633–639. Bardhan, Pranab, and Jeff Dayton-Johnson. 2007. “Inequality and the Governance of Water Resources in Mexico and South India.â€? In Inequality, Cooperation, and Environmental Sustainability, ed. Jean-Maries Baland, Pranab Bardhan, and Samuel Bowles, 97–130. New York: Russell Sage. Bates, Robert. 1981. States and Markets in Tropical Africa: The Political Basis of Agricultural Policy. Series on Social Choice and Political Economy. Berkeley: University of California Press. Bueno de Mesquita, Bruce, James D. Morrow, Randolph M. Siverson, and Alastair Smith. 2002, “Political Institutions, Policy Choice and the Survival of Leaders.â€? British Journal of Political Science 32 (4): 559–90. Frieden, J. A. 1991. “Invested Interests: The Politics of National Economic Policies in a World of Global Finance.â€? International Organization 45 (4): 425–51. Gibson, Clark C., Krister Andersson, Elinor Ostrom, and Sujay Shivakumar. 2005. The Samaritan’s Dilemma: The Political Economy of Development Aid. New York: Oxford University Press. Gneezy, U., E. Haruvy, and H. Yafe, 2004. “The Inefficiency of Splitting the Bill.â€? The Economic Journal 114 (495): 265–280. Keefer, Phil. 2004. “What Does Political Economy Tell Us about Economic Development—and Vice Versa?â€? Policy Research Working Paper 3250, World Bank, Washington, DC. ———. 2011. “Collective Action, Political Parties, and Pro-Development Public Policy.â€? Asian Development Review 28 (1): 94–118. 108 Understanding Policy Change Kitschelt, Herbert. 2007. “Party Systems.â€? In The Oxford Handbook of Comparative Politics, ed. Carles Boix and Susan C. Stokes, 522–54. New York: Oxford Univer- sity Press. Krishna, Anirudh. 2010. One Illness Away: Why People Become Poor and How They Escape Poverty. New York: Oxford University Press. Lal, Sumir. 2006. “Can Good Economics Ever Be Good Politics? Case Study of the Power Sector in India.â€? Working Paper 83, World Bank, Washington, DC. Lichbach, Mark I. 1998. The Rebel’s Dilemma. Ann Arbor: University of Michigan Press. McKean, Margaret A. 2000. “Rational Choice Analysis and Area Studies.â€? In Beyond the Area Studies Wars: Toward a New International Studies, ed. Neil L. Waters, 29–63. Hanover, NH: University of New England. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Harvard Economic Studies. Cambridge: Harvard University Press. Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press. Ostrom, Elinor. 2000. “Collective Action and the Evolution of Social Norms.â€? The Journal of Economic Perspectives 14(3): 137–158. ———. 2005. Understanding Institutional Diversity. Princeton, NJ: Princeton University Press. ———. 2010. “Overcoming the Samaritan’s Dilemma in Development Aid.â€? Presenta- tion at the Workshop in Political Theory and Policy Analysis, Indiana University Center for Study of Institutional Diversity, May 31, Bloomington, IN. Ostrom, Elinor, Roy Gardner, and James Walker. 1994. Rules, Games, and Common- Pool Resources. Ann Arbor: University of Michigan Press. Ostrom, Vincent. 1997. The Meaning of Democracy and the Vulnerability of Democracies: A Response to Tocqueville’s Challenge. New York: Cambridge University Press. Samuelson, Paul. 1954. “The Pure Theory of Public Expenditure.â€? Review of Economics and Statistics 36 (4): 387–89. The Collective Action Problem in Development: The Why Question 109 CHAPTER 4 Theories and Mechanisms of Political Economy: Institutions and Equilibria Pro-development reform entails some form of institutional change. This chapter explores the inner mechanics of institutions and the incentives that they generate. Why do institutional malfunctions persist? Why are they so durable if they generate obviously suboptimal outcomes? When do institu- tions change? What is the relationship between collective action and institu- tional change? The decisions that policy makers (politicians, bureaucrats) make, as well as the reactions of stakeholders (citizens, interest groups, and businesses, for example) to these decisions, are shaped by the rules or the institutional context in which this interaction occurs. The “rules of the gameâ€? shape policy makers’ choices (that is, Should I provide public goods, steal the money, or pander to key constituents?). They also constrain or boost the collective action potential of stakeholders. The policy landscape usually consists of formal and informal institutions, and both types of rules influence incentives for action. Therefore, by developing the ability to map the institutional context, reformers can identify the feasi- 111 FIGURE 4.1  Conceptual Map: Understanding Institutions and Their Influence in Development What are institutions? How do institutions differ from organizations? Formal institutions Informal institutions • How do institutional incentives shape pro-development collective action? • When do institutions change? • What are the symptoms of institutional malfunctions? Source: Authors. bility of different projects in light of the context-specific opportunities and constraints they face. This chapter first defines institutions, links them to the collective action problem, distinguishes between different types of institutions (formal and informal), and offers theoretical explanations for their origins and change. Second, it introduces the basic tools of game theory, using the prisoner’s dilemma to illustrate how the institutional context can determine which ­ outcomes (equilibria) are viable. This overview of institutional analysis forms the basis of subsequent chapters that focus on how specific institu- tional factors—such as agenda setting, agency problems, information asym- metries, and credible commitment—affect collective action and policy outcomes. 112 Understanding Policy Change Objectives of Chapter 4 By the end of this chapter, readers should be able to do the following: • R  ecognize institutions, organizations, formal institutions, informal in­ stitutions, institutional continuity and change, and the concept of equilibrium •  Understand the relationship among types of institutions, collective ac- tion, and development outcomes •  Link concepts of political economy to real-life situations. Institutions, Incentives, and Collective Action In chapter 3, we explored why cooperation between individuals and groups in society does not occur, even when the potential gains from joint effort would make everyone better off. Social dilemmas stemming from the pro- duction and maintenance of public goods impede effective mobilization for economic development. Collective action problems occur, conceptually, in a world that lacks the right rules and incentives to make everyone participate. Institutions, or rules of interaction, arise for the very purpose of solving such collective action problems. They stabilize expectations and help actors real- ize the gains from cooperation. Simply put, institutions give actors the incen- tives to cooperate rather than opt out. Nevertheless, as we will emphatically suggest again and again, just like collective action, not all forms of institutions have a positive impact on devel- opment. Powerful interests with plenty of resources and networks usually cooperate very well and build well-functioning institutions or rules that work to their advantage and block challengers. In development terms, the collective action potential of some groups and the institutions channeling it do not always lead to beneficial outcomes for all and often work against the poor. This fundamental problem explains skewed policies and outcomes. Thus, for reformers, the search for the Holy Grail becomes a search for a specific kind of joint effort: the pro-development type of collective action that can lead to a better institutional equilibrium. When does it occur? Are there institutions that can help mobilization succeed? Can pro-development col- lective action occur within unfavorable institutional boundaries? Can a group navigate key institutions to help move an agenda forward? Institutions are the larger conceptual umbrella here, because they can ex- ist and thrive in the absence of pro-development collective action, whereas pro-development collective action cannot exist without institutions. Politi- Theories and Mechanisms of Political Economy: Institutions and Equilibria 113 FIGURE 4.2  Influence of Collective Action on Institutional Change Institutions Collective action Institutional incentives and constraints Source: Authors. cal parties, civil society organizations, and protest movements need sets of rules to ensure cohesion and joint action. As the big conceptual context sur- rounding collective action, institutions can change dramatically, sometimes even in the absence of groups that can overcome collective action problems. Nevertheless, this is not a one-way street. Even if the rules of the game significantly shape and confine the likelihood of pro-development mobiliza- tion, this handbook argues that successful collective action sometimes in- duces endogenous institutional change as well. Figure 4.2 illustrates this scenario. To personalize this illustration, remember the 100 students who pro- tested against the government led by emperor Bokassa in the Central African Empire (see the introduction to this volume). Their tragic fate set an exam- ple for many other potential protesters against abuses of power. In this case, brutal state repression dramatically increased the costs of direct action for similar groups or individuals who disagreed with the mandate to buy expen- sive school uniforms featuring a picture of the emperor. Therefore, in the vocabulary of political economy, the institution of autocracy diminished the potential for collective action by severely increasing the costs of protest. North of the Central African Republic, many years later, a story with a similar beginning had a very different end. In the small rural town of Sidi Bouzid, on December 16, 2010, Mohamed Bouazizi, a street vendor support- ing his family by selling fresh produce, was yet again (after many previous episodes) harassed by the local police who confiscated his wares because he 114 Understanding Policy Change could not pay a bribe. When he attempted to approach the governor to make a formal complaint, he was refused an appointment. In an economy that de- nied him any other alternative for survival, Bouazizi set himself ablaze in protest against the state. The country where this happened is Tunisia, and this episode catalyzed waves of protest that finally led to the toppling of Ben Ali’s regime and ushered in the Arab Spring. In this case, despite autocratic rules banning dissent, effective and inspiring collective action led to pro- found institutional change. Understanding the incentives of actors or stakeholders to support the provision of public goods or to promote change requires taking a step back to look at the rules that generate such incentives and constraints in the first place. Critical thinking about the origins and internal mechanics of rules or institutions can shed light on the conditions under which they are likely to change, as well as on the specific roles that economic and political actors can play during the change process. Later in the chapter, tables 4.1 and 4.2 summarize a series of classic formal and informal institutions and link them to politicians’ incentives and pro- development collective action. This list is not complete by any means, nor are the conceptual links as clear-cut. Its purpose is just to summarize some findings of the literature and engage the reader in systematic thinking about the effects of rules on outcomes. Since every policy reform or attempt at change is complex, in reality many rules must interact to produce successful collective action and positive development outcomes. Institutions generate incentives for politicians and bureaucrats to pur- sue—or not—development and the provision of public goods. Understanding how institutions function and the incentives they generate matters. After all, policy reform in development entails fundamental changes of core rules. To identify those institutions that can help groups act collectively and achieve positive change, the following section will describe in detail what institu- tions are and how they generate incentives. What Are Institutions? In the simplest possible conceptualization, institutions are the rules of the game. More formally, they can be defined as “rules, compliance procedures, and moral or ethical norms designed to constrain the behavior of individualsâ€? (North 1981, 201–02.) All the components of this definition have implications for the origins, stability, and possibility of change. Please note, however, that the understanding of institutions as rules differs slightly from institutions as equilibria or norms (Crawford and Ostrom 1995). The focus on the rules of the game emphasizes an explicit sanction for noncompliance. By contrast, Theories and Mechanisms of Political Economy: Institutions and Equilibria 115 conceiving of institutions as equilibria relies on the evolution of shared expec- tations that certain actions automatically trigger unfavorable change in other actors’ responses. The concept of institutions as norms rests on the internal sense of proper and improper behavior that constrains the actions of the ac- tors (Crawford and Ostrom 1995, 583). This normative approach refers to the obligations and duties of the participants defined in a particular cultural con- text and reinforced through symbols, rituals, ceremonies, and the like. Often, in everyday vocabulary, institutions and organizations are used in- terchangeably. Political economy, however, commonly imposes a conceptual distinction between them (North 1990). Before we go any further, the follow- ing clarification is in order. How Are Organizations Different from Institutions? Organizations bring together individuals or social groups with a common policy goal;1 func- tion according to their own rules, procedures, and culture; and seek to influ- ence outcomes. They include political parties, bureaucratic units, firms, civil society groups, and international donor organizations, among many others. They are the players (actors or stakeholders) that interact strategically with- in the confines of the institutional context (North 1990). Let us briefly explore the intersection of institutions and organizations. A trade union is an organization representing workers that tries to influence policy on such issues as wages or working conditions in a specific institu- tional context. In turn, the rules of the game resulting from the interaction among union, government, and employers shape the union’s actual power and influence (cost of striking, the existence of closed shops, a formal policy- making role, and the like). As suggested in chapter 3, organizations also facilitate collective action among individuals. Leaders of organizations monitor their boundaries, sanc- tion free riding, and reward compliance. The internal institutional con- straints on leadership matter very much for the organizational effectiveness of collective action. For example, if the rules governing the selection of lead- ers block challengers and make it harder for members to remove the incum- bent, then accountability might be weak, and the “iron law of oligarchyâ€? iso- lates the executive team from the rank and file (Michels 1911). At the other extreme, if the leadership is weak or rapidly changing, leaders might not be able to impose antishirking requirements on free riders. Political party disci- pline, strength, and cohesiveness have a lot to do with the organizational structure, leadership, and mechanisms of the parties as well as the internal accountability to members. A similar logic applies to civil society. For a political-economy analysis, it is also important to assess whether or- ganizations are monolithic or heterogeneous in their incentives, stakes, and 116 Understanding Policy Change preferences. Often, different factions, divisions, units, and bureaus within the same organization have divergent goals and fail to act collectively. After this clarification, we begin by mapping the institutional landscape. First, we will explore the inner logic of institutions consequential for devel- opment. Second, we look at how to recognize the main types of political- economic rules of the game and identify the incentives they produce. Institutional Mechanics: Inside the Machinery. Institutions as equilibria. The concept of institutions as equilibria implies that the development out- come they produce is self-enforcing. Once the outcome is achieved, no actor or stakeholder has an incentive to try and change it, even if it is far from his or her ideal policy preference (Weingast 2002, 682).2 The conceptual focus is on the self-enforcing logic of institutions; that is, the parties to the pact must believe three things:  hey are better off accepting the outcome rather than rejecting it. • T Each actor will change his or her behavior when the other actors do so. •  • All the individual actors will defend the gains of cooperation of the other participants (that is, no participant will be excluded from the benefits ac- crued in equilibrium). Here is an example of a low institutional equilibrium from which actors did not have incentives to deviate. After the end of communism, many coun- tries in Eastern Europe and Central Asia fell into a self-reinforcing “reform trap,â€? whereby a predatory state raised taxes and diverted revenue away from the provision of public goods to satisfy politically powerful interest groups. As a result of the perceived misallocation of resources, many entrepreneurs and households began operating in the informal economy to avoid taxation. In turn, the expansion of the informal economy led to the further spiraling de- cline in tax revenue and provision of public goods. In an attempt to tackle the problem, some countries transitioned to a lower tax rate meant to stimulate the growth of the formal economy and increase tax revenue. Despite these policies, governments could not alter the institutional context by adopting the rule of law and effectively and credibly guaranteeing that the collected taxes would not benefit a narrow circle of elites. Firms operating in the informal economy were thus reluctant to participate in the formal economy. An equi- librium of low tax collection, heavy tax burden, and high corruption persisted because of this institutional configuration (Aslund, Boone, and Johnson 2002). Institutions as rules. The concept of institutions as rules presupposes a third- party enforcer that can monitor and sanction noncompliant behavior of one or more participants. Traffic rules emerged as the result of coordination Theories and Mechanisms of Political Economy: Institutions and Equilibria 117 among participants and are enforced by the traffic police. This is a classic ex- ample of a rule-based institution. The absence or lack of capacity of the moni- tor or enforcer (that is, corrupt or inefficient police officers) alters the func- tioning of the institution. As the primary laws of the land, modern constitutions are the rules of the game par excellence. They arise from the strategic interac- tion of actors endowed with a certain distribution of power and enshrine ex- pectations about how these arrangements will affect their power in the fu- ture. In most political systems, supreme or constitutional courts, as third-party enforcers, watch over constitutional rights and obligations. However, if the judiciary is not independent and strong, executives can interfere with its rul- ings, and the rules of the game are not enforced. This institutional malfunc- tion can have detrimental consequences for development. Concepts in Prac- tice 4.1 shows, in the case of access to basic water supplies in South Africa, how altering the rules of the game, and entrusting their enforcement to an independent third party, can provide disenfranchised groups with the institu- tional means for securing access to previously denied goods and services. Institutions as norms. Institutions that have normative meaning within a specific cultural setting entail gains and costs of cooperation that are context specific. This mechanism is consistent with the previous understandings of institutions as equilibria or rules whenever the actors’ stakes are clearly identifiable within a given cultural context. Norms themselves have often evolved from the strategic interaction of groups and individuals over long periods of time. Historically, for example, tax rebellions in medieval Japan developed a leaderless pattern because the revolt leaders, if caught by authorities, were harshly punished (White 1995). Contributing to the evolution of community norms, in this case, the technology of law enforcement, created unique and enduring organizational structures and shaped the stakes of popular mobili- zation. Concepts in Practice 4.2 shows that cultural norms often define or constrain the sequence and protocol of actors’ interactions, as well as the gains and losses derived from policy change. As Concepts in Practice 4.2 illustrates, it is important for development analysts and practitioners to understand the specific normative meanings, dimensions, and gains or losses attached to policy reform by various stake- holder groups. Formal and Informal Institutions A useful typology differentiates between formal institutions (those that exist in law or statute) and informal institutions (those without written rules or 118 Understanding Policy Change CONCEPTS IN PRACTICE 4.1 Constitutions and Individual Rights to Public Goods—Water Access and Health Care Some constitutions around the world, attempt- ability. In 2000, for example, residents of Wal- ing to redress severe and long-standing inequal- lacedene—one of the poorest settlements of ities of access, have sought to grant all citizens Capetown—won their class action suit in the basic rights to public services. South Africa, for Constitutional Court on grounds of the right to instance, has a very high demand for water but proper housing facilities. Similarly, in 2009, the fifth-lowest water availability among the 41 5,000 poor inhabitants of Phiri, Soweto, made Sub-Saharan African nations. Under these con- legal history as the first test case of the consti- ditions of extreme resource scarcity, the apart- tutional right to free water. The claimants won heid regime concentrated the right of access in the case in the Supreme Court of Appeal but the hands of the white minority, which repre- lost in the Constitutional Court (Danchin 2010). sented 13 percent of the population and used Similarly, over the past 20 years, India, South around 95 percent of the available water for ag- Africa, and many countries in Latin America ricultural irrigation on large farms. According to have begun guaranteeing the constitutional estimates, in the early 1990s, about 15.2 million right to health care. This expansion of rights has South Africans did not have access to a basic translated into a wave of individual and class ac- water supply, and 20.5 million lacked access to tion litigation, as citizens have been able to de- basic sanitation (Mehta and Ntshona 2004, 4). mand that the public sector provide and cover The postapartheid era made access to water the costs of specific medications and services. and sanitation a policy priority, and the govern- Despite the fact that the jury is still out on the ment instituted a policy that committed to pro- fiscal and equity implications of health care liti- viding universal access to a basic supply of safe gation, citizens have often managed to demand water for all households. Thus, South Africa the compliance of the executive with the rules became the only country in the world that con- of the game. In some cases, these formal rules stitutionally guaranteed the right to basic drink- enforced by the judiciary as an independent ing water and water for sanitation. Resetting third party have led to dramatic and undeniable the rules of the game and inscribing them in the changes: universal lifetime access to basic HIV constitution meant that poor citizens could take medication in India, increased health care cov- the local or national government to the Consti- erage and benefits for all workers operating in tutional Court as the third-party enforcer if it re- the informal sector in Colombia, and preventing neged on this obligation. Indeed, the mere exis- mother-to-child HIV transmission for the entire tence of legal grounds led many inhabitants of population of South Africa. shantytowns to demand government account- Sources: Mehta and Ntshona (2004); Danchin (2010); Yamin and Gloppen (2011). formal structures that still determine outcomes). On the one hand, formal institutions such as constitutions and international treaties are drafted, de- bated, and voted upon during a long process through which all key actors set and clarify contractual expectations and ensure that the agreement they Theories and Mechanisms of Political Economy: Institutions and Equilibria 119 CONCEPTS IN PRACTICE 4.2 Water Resource Management Reforms in Zimbabwe In the late 1990s, several countries in southern the implementation of the reform, water is a Africa created decentralized water manage- God-given good for the Ndau community in ment institutions to increase the efficiency of Zimbabwe and is at the center of religious wor- water use and stimulate broader popular partici- ship and ancestral beliefs. The owners of the pation in resource management. Zimbabwe water are the spirits of the ancestors, whereas passed a water act in 1998 and began imple- the custodians are the tribal chiefs. Ndau be- mentation two years later. This law promoted a liefs about water access, ownership, and user new institutional configuration under the guise fees were at odds with the new law and the of integrated water management at the catch- central state’s definition of resource ownership. ment level. Following technical trends in the Similarly, in several communal areas of Zimba- international water management community, bwe, traditional definitions of community (as this institutional shift profoundly altered the settlements extending along entire rivers) di- way in which resources were managed and rectly contradicted the prescriptions imposed paid for in the localities. The former Water Re- by the new law. As a result, significant numbers source Department became the Zimbabwe of users did not pay the fees the state expected National Water Authority. Following the policy ­ to collect for water consumption. recommendations of the IMF Structural Adjust- In both instances, the preexisting rules of ment Program, this new bureaucratic unit was actors’ interaction and the informal authorities supposed to operate along commercial lines that governed water consumption conflicted di- and generate its own resources through user- rectly with state policy and undermined formal fee collection for access to water. institutional change. By not engaging with local Imposing such fees on consumption in rural actors to understand the real stakes and rules areas often conflicted directly with local cus- of the community, reformers failed to anticipate toms. For example, according to an analysis of the negative consequences of their policy. Source: Nicol and Mtisi (2003). reach will be binding. On the other hand, informal institutions—such as gen- der roles, family obligations, manners, and rules of appropriate behavior within a particular ethnic or faith-based group—may or may not be codified or written. Nonetheless, they produce strong compliance, sometimes even more so than written rules. Why should we care about the complete institutional landscape of policy reform? Political economy brings politicians and political processes explic- itly into development equations. The mechanisms or the rules through which political representatives are selected, monitored, and sanctioned gen- erate individual or organizational incentives to provide public goods, engage in collective action, and shift (or not) development equilibria. These formal 120 Understanding Policy Change FIGURE 4.3  The Intersection of Formal and Informal Institutions Formal institutions Informal institutions Organizational and individual incentives Strategic interaction Development equilibria Source: Authors. rules of selection often interact with the less visible territory of tacit expec- tations and lock in final outcomes. Understanding this intersection allows decision makers to correctly identify the feasible space for change, and avoid policy mistakes that harm more than help development. In the case of water management policy in Zimbabwe, for example, reformers ignored the infor- mal rules of interaction at the local level, while expecting the formal rules designed by the central government in Harare to be fully enforced and to improve water management. Not taking into account informal institutions led to skewed incentives and outcomes: detrimental redefinition of tradi- tional community boundaries and widespread noncompliance with the new policy. Political-economy analysis should therefore take into account both cate- gories and focus on the intersection of formal and informal institutions (as shown in figure 4.3). Formal Institutions. Most recognizable institutions of modern states— such as the form of government, the electoral system, and the arrangements ensuring political checks and balances—fall into the category of formal institutions. They are written or codified rules of behavior for all social, Theories and Mechanisms of Political Economy: Institutions and Equilibria 121 economic, and political actors. These sets of rules are sometimes called “parchmentâ€? institutions to denote their legalistic and codified character (Carey 2000). But how effective are the rules or laws on paper in shaping economic, political, and social behavior? Especially in the case of the developing world, it has been widely recognized that not all written rules effectively govern behavior and that many effective individual incentives derive from tacit, unwritten rules (Carey 2000, 737). For example, the vast majority of countries have some sort of legislature with elected representatives who have a mandate given by their constituencies to enact legislation. However, some assemblies are merely “window dressingâ€? forums that fulfill various purposes other than democratic representation. They can serve, for in- stance, as platforms through which powerful executive or hegemonic par- ties identify and co-opt political challengers. Similarly, in spite of relatively clear rules for the selection of the executive inscribed in constitutions, leaders often circumvent the legal constraints of their mandate. Instead, the real rules of access to power and allocation of scarce resources might reside in a narrow elite organized around the executive or the military, or they might entail mechanisms inherited from ethnic, kinship, or faith- based cleavages. By the same token, if the formal rules of representative democracy are weakly institutionalized and robust checks and balances on power are missing, executives have the space to abuse their mandate, build clientelistic voting networks, and consolidate a hard-to-defeat in- cumbency advantage. Here are some examples of classic formal rules associated with the politi- cal process that have been found to generate incentives regarding the provi- sion of public goods and collective action for reform: • Political regime types (for instance, democracy versus authoritarianism) • Federalism and decentralization • Presidential versus parliamentary systems • Bicameral versus unicameral legislature • Executive-judiciary relations • Electoral laws such as proportional representation (PR) (voting rules  through which political candidates run for election on party lists and in which the number of votes received translates into the number of seats in the legislature) or single-member district plurality vote for candidate selection (“winner-take-allâ€? systems in which only the candidate who wins the most votes at the constituency level gets elected). Table 4.1 links these institutions to political incentives for pro-development collective action. 122 Understanding Policy Change TABLE 4.1  Political Incentives for Development and Collective Action Opportunities Induced by Formal Institutions Political incentives for provision Pro-development collective action:  Formal institutions of public goods. opportunities and constraints. Regime type Democracy Elections provide accountability Political parties, interest groups, and mechanisms. social movements are allowed. Autocracy Less or no direct accountability; Difficult for groups to pursue repression; potential existence development agendas if they are not of alternative mechanisms of aligned with the priorities of political institutional checks and balances leadership. in some states. Hybrid regimes Partial accountability relations. Some reforms can be pursued; “approvedâ€? parties are allowed to function, as long as they do not threaten the core distributive logic of the regime. Federalism Federations or Regional interests are well Political party cohesion and discipline decentralized systems represented; regional and local might be lower since politicians have governments might overspend incentives to satisfy regional governors and engage in common-pool and not national party leaders; resource dilemmas; in some implementing policies could entail cases, if accountability is cooperation from both the center and less clear, decentralization creates subnational units, making cooperation opportunities for corruption. more costly. Unitary or centralized National-level institutions shape Some parties might be more cohesive; systems political incentives. national-level trade unions and business lobbies can pursue agendas more effectively than in decentralized systems; likely to act against decentralization reforms, for example. Executive-legislative relations Parliamentary system Incentives for common-pool resource Collective action for lasting policy behavior in the budget process to changes might depend on the satisfy individual constituencies; relationship between the cabinet independent parliamentary behavior (minority versus majority) and versus party discipline might depend legislature, cabinet duration, etc.; on the type of coalition necessary for cabinet minority in some cases has cabinet formation. induced lower party discipline. Presidential system Strong presidential powers; ability to Collective action is contingent on the use decrees; control of the budget political relation between the head of process. state and legislature: if from the same party and high party discipline, effective reform; if not, potential policy stalemate. (table continues on next page) Theories and Mechanisms of Political Economy: Institutions and Equilibria 123 TABLE 4.1  continued Political incentives for provision Pro-development collective action:  Formal institutions of public goods. opportunities and constraints. Type of legislature Unicameral legislature Smaller coalitions necessary to pass reforms. Bicameral legislature Sometimes it allows greater It increases the size of the coalitions representation of regional or and time required to pass pro- occupational interests that balance development laws; potential policy out purely political incentives; stalemate if ideologically divergent traditionally designed to balance out chambers. the liberal popular chamber with a conservative counterpart (like the United Kingdom). Executive-judiciary The degree of judiciary Without judicial independence, relations independence might give incentives collective action of groups trying to to the judiciary to behave challenge executive decisions is strategically and preempt conflicts reduced. with the executive. Constitutional review rules Approval by ordinary Collective action for major changes or majority amendments of the constitution becomes easier. Approval by supermajority More difficult to acquire the necessary majority for change. Electoral rules Single-member plurality It creates incentives for politicians to First-past-the-post systems are likely to pursue personalized constituency have a lower number of parties (usually service (pork barrel politics; in some two major parties) but are less cases, clientelistic exchanges); some disciplined. studies argue that it reduces corruption because it strengthens personal accountability.a Proportional representation Since politicians are elected based Coalitions and policy trade-offs are on party lists, it might decrease common; collective action for reform direct accountability to voters; it depends on party discipline, coalition allows greater representation and a size. In some cases, PR coalitions are higher number of parties. fragile because of the high number of parties and policy divergence. Closed-list proportional Politicians have to cater to party Less accountability to voters (some representation leaders rather than local studies argue that they produce higher (party leaders decide on constituencies. corruption levels), but also likely to the order of candidates on increase party discipline for pushing the ballot) reforms through. 124 Understanding Policy Change TABLE 4.1  continued Political incentives for provision Pro-development collective action:  Formal institutions of public goods. opportunities and constraints. Open-list proportional More direct accountability as voters This electoral rule generates a mix of representation decide on the rank of candidate accountability to voters and national (Voters have a say in terms nomination. party leaders. Its effects on collective of deciding the order in action opportunities and party discipline which candidates appear might be contingent on other factors. on the ballot) Districting rules District magnitude The smaller the district, the higher It might personalize constituency the candidate competition for a benefits (pork barrel, clientelistic representative seat. handouts) and reduce party cohesion at the national scale. Overlap between local and Might create incentives for or against If districts overlap, then legislators national electoral districts decentralization. might fear that the mayors or governors of decentralized units would want to undermine their interests;b they might resist or oppose fiscal decentralization. Cabinet formation/ Coalition type Concentration of executive Either a minimal winning coalition or The collective action effect related to power (minority cabinets an undersized coalition forms the policy reform depends on the support or minimum winning cabinet. of the legislature for the cabinet (low coalition) for undersized cabinets, higher for minimal winning); the minimal winning coalition cabinets could entail less policy concessions than oversized coalitions. Sharing of executive More parties than necessary form Potential coalition fragility, factionalism, power (oversized cabinets) the government. policy trade-off. Interest group system Pluralism No major interest group coordination; Many small organizations, no umbrella labor, business, and government do associations (i.e., business not meet regularly to set wages, etc. associations, large labor unions, no tripartite agreement), less influence in policy and reform. Corporatism Involves the incorporation of The collective action potential of these coordinated interest groups in policy interest groups to promote or block formation (labor, business); some policies is very high (usually peak studies found better macroeconomic umbrella associations, large, few in performance, others debate numbers at the national level). findings.c (table continues on next page) Theories and Mechanisms of Political Economy: Institutions and Equilibria 125 TABLE 4.1  continued Political incentives for provision Pro-development collective action:  Formal institutions of public goods. opportunities and constraints. Political party systems Two-party systems Politicians appeal with policies to Produce single-party cabinets that are satisfy the median voter. more stable than coalition governments and, according to some, more effective for policy making.d Multiparty systems Better representation of the Generate coalition governments that preferences of a diverse electorate. are often fragile and nondurable. Party discipline and Organizational incentives to comply Strong party discipline can effectively cohesion or not comply with national party block or promote pro-development leaders and policies. reforms. Intraparty selection rules Some studies argue that intense Collective action for public good and competition intraparty competition might lead to provision through parties might be clientelism.e hampered by intraparty factionalism. Source: Authors; selected examples of formal institutions in democracies come from Lijphart 1999. a. Persson and Tabellini (2005). b. Eaton, Kaiser, and Smoke (2010). c. Lijphart (1999). d. Lijphart (1999). e. Carey and Shugart (1995); Golden and Chang (2001). Sometimes, even small changes within the same family of formal institu- tions (for example, the rules governing the election of political candidates) can lead to different outcomes related to the provision of public goods. Con- cepts in Practice 4.3 illustrates the effects of variations of electoral laws on the provision of public goods and economic reform in Brazil and República Bolivariana de Venezuela. As we will see, there is no easy answer, recipe, or immediate causality, as the effect of such rules is often complex. Therefore, mapping the web of core institutions governing the policy domain of interest allows reformers to identify the real origin or locus of skewed incentives, come up with correct diagnostics, and design feasible strategies of action. Written, codified institutions often interact with unwritten, less formal- ized shared expectations of behavior among actors or stakeholders. These normative templates are called informal institutions. Development out- comes are in many cases the products of such interaction. Ignoring what lies behind the immediately visible landscape, as well as the real power relations underpinning and sustaining formal rules, and attempting to export blue- prints of formal institutions on grounds that they functioned elsewhere can lead to either ineffective or detrimental outcomes. To paraphrase Luiz Car- los Bresser Pereira, a former Brazilian minister, “Institutions can be at most imported, never exportedâ€? (Przeworski 2004, 540). 126 Understanding Policy Change CONCEPTS IN PRACTICE 4.3 Effects of Electoral Laws on Public Goods in Brazil and República Bolivariana de Venezuela Proportional representation is a voting method district level, and reform deadlock (Ames 1999, commonly used for national and local elections 2001). in which the number of votes won by a political In República Bolivariana de Venezuela, in con- party translates into a proportional number of trast, a closed-list system of proportional repre- seats in legislatures. For example, if a party sentation has been singled out as one of the fac- wins 20 percent of votes nationally, it obtains tors leading to reform deadlock (Crisp 2000). In roughly 20 percent of the seats in the parlia- 1984, falling oil prices, large foreign debt, and ment. The procedure through which political declining real wages led to a crisis in a country parties list candidates for elections and decide previously commended for its stability and on their order on the ballot is also consequential growth. The government, led by Jaime Lusinchi for the final result. in collaboration with a small group of bureaucrats The “open-listâ€? proportional representation and a few interest groups, was in charge of navi- rules give voters direct influence over the order gating the necessary package of reforms. In the in which political candidates appear on the bal- process, the executive immediately started to lot, as they can alter the ranking of candidates rule by decree, structured and dominated the leg- on party lists (individual mandate). In contrast, islative agenda, and consolidated unprecedented in “closed-listâ€? PR systems, party leaders, not power. However, despite this wide-ranging politi- voters, decide on the order. cal mandate, no ­ significant reforms were under- How do these rules affect the provision of taken. On the contrary, several policies solidified public goods? In Brazil, for example, open-list the vested interests of powerful groups to the PR led to the creation of pockets of voters detriment of diffuse groups of citizens. supporting individual candidates who offered ­ How can this puzzle be explained in a demo- either “porkâ€? (narrowly targeted private goods cratic polity in which a vibrant legislature and an to a specific constituency) or bribes for votes anxious public could have prevented the reform delivered en masse by local party bosses. deadlock? Part of the explanation lies in the These incentives led to very weak parties, in­ closed-list proportional representation rule (sur- dividualistic deputies, and strong municipal viving until 1993) that made political candidates party bosses. Thus, because of the fragmenta- dependent on the central leadership of the tion, weakness, and lack of cohesion in the party for their selection, not directly on voters. party system, the incumbent president always This institution created individual incentives for needed governing coalitions that were hard high party discipline, no opposition majority, to obtain and stabilize in order to pursue re- and, overall, an ineffective congress unable to form agendas. This volatility of reform coali- challenge a powerful president and interest tions at the center subsequently translated into groups. In the absence of a vocal civil society, increased corruption, pork barrel politics at the reform was stalled. Informal Institutions. Informal institutions are social norms, conven- tions, or context-specific prescriptive templates of behavior. Unlike formal rules, they are usually unwritten and function outside official channels (Helmke and Levitsky 2006). Formal institutions, because of their legalistic Theories and Mechanisms of Political Economy: Institutions and Equilibria 127 and codified nature, are relatively easy to identify. Constitutions, laws, de- crees, organizational charts, and electoral laws stipulate rights, responsi- bilities, future contingencies, and the monitoring technology that ensures compliance. Tacit, informal rules or norms are sometimes trickier to pin down and analyze. Here are several rules of thumbs with respect to how to (or not to) identify them (Helmke and Levitsky 2006, 6–10):  nformal institutions entail shared social expectations of behavior, as well • I as some form of external punishment for noncompliance. Not all unwritten behavioral patterns are informal institutions. For ex- •  ample, if bureaucratic corruption is tolerated or even encouraged by cen- tral state authorities and if there is an underlying expectation or tacit norm in the wider public sector community that bribes should be ex- tracted, corruption becomes the informal rule of the game. If, however, bureaucrats solicit bribes in response to low public sector salaries, cor- ruption is just a behavior triggered by the lack of adequate incentives. • Informal institutions are often conflated with cultural values and infor- mal organizations that have usually been overlooked by Western-centric institutional analyses (for example, tribes and clans). However, caution is needed in identifying and analyzing them. Sometimes, even if the infor- mal organization (say, a particular ethnic tribe, for example) is a stake- holder in a reform process, it does not mean that the shared expectations within the group automatically shape the policy stakes and affect the broader process of policy change. If, nevertheless, the reform stakes as- sessed by this group are derived from common expectations of costs and benefits, the informal institutions are at work. As the example related to water management reforms in Zimbabwe showed, the informal institu- tions of the Ndau ethnic group that normatively defined concepts of own- ership and property rights related to water had a direct impact on policy implementation. The mere incorporation of the same ethnic group in the list of local stakeholders for another policy (say, decentralization reforms) for which the group does not have a priori shared expectations, however, does not qualify as an informal institution. • Some institutions that might be perceived as informal by the analysts are in reality highly formalized. Many African countries have created consti- tutional structures and responsibilities for traditional authorities (cus- tomary councils, chieftaincy committees, and the like) for different pur- poses: the unification of customary laws across regions, compiling and registering lines and succession of authority, coordinating customary policies with the regulatory framework of the central or local govern- 128 Understanding Policy Change ments, institutionalizing patronage networks, and so forth. In Yemen, for example, the political survival of the incumbent party and the president traditionally depended on alliances with key tribal leaders. The Depart- ment of Tribal Affairs, a bureaucratic agency, pays stipends to 5,000 tribal sheikhs all across the country. In addition to this formal recognition, a parallel network awards government jobs to either family members or close allies of tribal leaders (Egel 2010; ARD 2006).  Similarly, in the 1990s, several Latin American countries codified and formalized indigenous laws to achieve better compliance with national policies. Treating them as informal ignores an entire political process of codification, formalization, and official recognition of these rules by the central state. Strong informal rules of behavior can either undermine or dominate formal institutions. The interaction between the two types of rules of the game often explains development outcomes, as well as the success or failure of ­ policy reform. Sometimes, by eroding or circumventing written rules, conflicts between informal rules of behavior and formal institutions can have negative effects and block change. For example, even if the legal framework gives women equal rights (such as the right to vote or participate in the labor market), patriarchal, societal, ethnic, or family norms can de facto provide disincen- tives for acting on such rights (social stigma, exclusion, or physical punish- ment). In some cases, top-down reforms aiming to replace traditional insti- tutions that regulate behavior with formal rules derived from “best practicesâ€? end up being either ineffective or counterproductive. Informal institutions can also play a positive role by providing strong rules and mechanisms of accountability in cases where formal institutions are weak or missing. In China, for example, despite the lack of development funds in the villages and the absence of democratic mechanisms of account- ability, some rural units perform better than others in delivering public goods. The cause of such variation in development outcomes lies in strong unofficial rules and norms that effectively stabilize expectations of actors. Villages with solidary social groups, such as temples or villagewide lineages (clans), typically provide more public goods. Concepts in Practice 4.4 con- siders some of the factors that led to this puzzling variation in development across rural localities. Given that institutional weakness characterizes many developing coun- tries (Bardhan 2002), the relevance of informal institutions and of their ­ interaction with the formal rules of the game is essential for development policies and reforms. Theories and Mechanisms of Political Economy: Institutions and Equilibria 129 CONCEPTS IN PRACTICE 4.4 Informal Institutions in China A study conducted in four rural provinces in informal codes of duties and responsibilities China that surveyed 316 villages showed that can derive from many sources: for example, development funds rarely reach the villages similar religious beliefs (in the case of village since counties and townships prefer to invest temples) and family- or kin-related codified be- their resources in infrastructure projects that havior (in the case of clans). Interestingly, in the connect localities rather than providing public Chinese context, the variation in development goods inside the villages. Village governments outcomes seems to be best explained by the are nevertheless responsible for irrigation, gar- joint effect of the two criteria. While village bage collection, road maintenance, and primary temples cover all the inhabitants of the village, education, with very few financial resources at including the officials responsible for public their disposal. Despite resource scarcity across goods, village churches are encompassing but the board, the production and maintenance of not embedded: they provide social services to a public goods vary significantly in these rural lo- large segment of the village population but calities. What explains this divergence in de­ rarely include officials since they are nonindig- velopment outcomes? The study found that in- enous institutions. In contrast, subvillage lin- formal institutions—specifically, village solidary eages, by definition, do not encompass the groups such as temples or clans—lead to supe- entire village population, although they might ­ rior delivery and maintenance of public goods. include the locality official. According to the For these informal institutions to provide in- study, village temples and villagewide lineages centives to village officials to perform, they that are simultaneously encompassing and em- have to meet two criteria: (1) they need to be bedding score higher in governance indicators encompassing, meaning that they must apply than villages featuring informal institutions that to everyone under the village jurisdiction; and lack one of the two criteria (sublineages that in- (2) they need to include the officials as mem- clude just a segment of the population or village bers. According to this argument, if officials churches). The findings demonstrate that, in and group members share a similar code of certain contexts, informal mechanisms of ac- moral conduct and ethical obligations, their in- countability can lead to better development out- centives are more likely to be aligned, with comes even in the absence of formal electoral positive consequences for development. These alternatives. Source: Tsai (2007). Although understanding the political-economy environment surround- ing reform—especially the configurations of informal and formal institu- tions—is challenging, it is crucial for a successful outcome. Following are some examples of informal institutions relevant in many parts of the world (Helmke and Levitsky 2006, 276–280): • Political clientelism or patronage • Informal rules of bureaucratic meritocracy (Singapore has the same post- colonial bureaucratic legacies as the Philippines and Indonesia but has unspoken norms of strong meritocracy) 130 Understanding Policy Change • Unwritten norms of bureaucratic coordination between agencies • Informal rules of interaction between bureaucrats and private sector companies that participate in public procurement processes (for exam- ple, Amakudari or “descent from heavenâ€? practices in Japan, whereby re- tiring civil servants obtain lucrative jobs in companies for which they se- cured public contracts in the past) • Blackmail and regulatory uncertainty (Kuchma in Ukraine, Fujimori in Peru) • Informal interparty or executive-legislative consultations. For a brief summary of the incentives provided by informal institutions and their impact on collective action for development, see table 4.2. Sometimes, the causal effect of formal or informal institutions on devel- opment outcomes can be more complex than initially thought. Unintended consequences or specific interactions of many institutional layers skew the original goals. Concepts in Practice 4.5 presents an example of such unin- tended consequences in the case of regulations governing labor relations and the development outcomes they generated in Indian states. Institutional Origins, Stability, and Change Why Do Institutions Exist? Generally two major explanatory frameworks seek to account for the origins and evolution of institutions: demand-side and supply-side approaches. Demand-side approaches tend to emphasize that institutions are designed to help (some or all) stakeholders enhance their ability to alleviate collective action problems: “Institutions arise in part to help create the conditions for self-enforcing cooperation in an environment where there are gains from cooperation but also incentive problems that hinder a community’s ability to maintain cooperationâ€? (Weingast 2002, 674).3 Thus, for example, compul- sory trade union memberships (closed-shop agreements) were established so that individual workers could not free ride on the efforts of a union to se- cure better working conditions for all workers in a factory. Supply-side approaches tend to emphasize how the unintended conse- quences of institutions may alter the relative power of stakeholders in un- foreseen ways and independently of the original raison d’être on which the institution was founded. For example, closed-shop agreements, by providing a union with significant power to disrupt production in an industry, could enable it to make political demands that have nothing directly to do with ensuring that workers’ salaries or working conditions meet certain stan- Theories and Mechanisms of Political Economy: Institutions and Equilibria 131 TABLE 4.2  Incentives of Informal Institutions and Their Impact on Collective Action for Development Political incentives to provide Pro-development collective action:  Informal institution public goods opportunities and constraints Political clientelism/ It generates incentives for politicians Clientelism and political machines are patronage and parties to cultivate personalized very effective with respect to vote relations with local notables, turnout and collective action since intermediaries, organizations for they mobilize large numbers of voters; distribution of clientelistic handouts, they also commonly undermine and vote monitoring. programmatic policies and public good provision. Unwritten norms of They give incentives for cooperation Facilitate bureaucratic collective action bureaucratic coordination across agencies, irrespective of at the policy formulation and political affiliation. implementation stage. Regulatory uncertainty In some environments, it creates The fear and anticipation of arbitrary incentives for corruption and implementation of laws and regulations self-refraining of opposition (for example, discretionary tax members from challenging the collection or inspections) reduce the executive. collective action potential of opposition to the executive. Rules of career They give incentives for collusion According to studies, they reduce the advancement in between bureaucrats and private general level of oversight and increase bureaucracy and companies. risk taking from companies; they can engagement with the hamper pro-development collective private sector (for action. example, the Japanese Amakudari or “descent from heavenâ€? system) Informal rules of interparty Informal consultation forums or Facilitate bipartisan coalition formation consultation caucuses across party lines might and consolidation; prevent conflicts; it increase incentives for policy might entail policy trade-offs. cooperation; create incentives for accountability (programmatic or clientelistic) to voters for individual politicians. Clans, lineages, ethnic Give incentives to politicians, voters, These norms are usually very effective faith-based group norms and bureaucrats to stick together. at solving in-group collective action problems; high ethnic fractionalization or in-group/out-group competition for scarce resources reduces the potential for collective action, and, according to some studies, leads to public good underprovision.a Source: Authors. a. Alesina, Baqir, and Easterly (1999). 132 Understanding Policy Change CONCEPTS IN PRACTICE 4.5 Labor Regulations in India The relatively low rate of manufacturing growth In short, while all Indian states began with in India stands in stark contrast to the manufac- the same legal-institutional framework, by track- turing growth of neighboring East Asian states, ing the state-level reforms of the 1947 act over as well as to India’s own service and information a period of 35 years, we can examine how those technology industries. What explains this rela- changes in the formal legal rules of the game tively poor performance? The role of formal in- affected final outcomes (economic growth, for- stitutions and how they affect the business en- mal employment, and poverty rate). By coding vironment could be one plausible, although by whether each reform of the 1947 act, at the no means the only, factor, which may help ex- state level, was pro-employer, neutral, or pro- plain part of this empirical puzzle. Many of the employee, the authors of this study were able to laws and regulations governing industry in India identify whether marginal changes in the formal are set at the center and apply to all the states regulatory environment had a significant effect (the cornerstone legislation being the Regula- on development outcomes. The authors found tion and Development Act of 1951). However, that pro-worker amendments were associated the regulation of industrial relations is a concur- with reduced output, reduced formal employ- rent issue in which both the central and the ment, less investment, lower productivity, in- state governments have the ability to “change creased employment in the informal economy, the rules of the gameâ€? by altering legislation. and higher urban poverty rates. Thus, changing Thus, while in 1947 virtually all Indian states had the rules of the game in a legally “pro-workerâ€? the same regulatory framework for industrial way had the perverse effect of reducing the op- relations (governed by the Industrial Disputes­ ­ portunities for formal employment and the bet- Act of 1947), over time each state was able to terment of living conditions. change this act, in every jurisdiction. Source: Besley and Burgess (2002). dards. Thus, unions may lobby for subsidies to declining industries or for perpetuating jobs and benefits at the expense of the taxpayer (see, for ex- ample, North 1990 or Hayek 1962). Irrespective of their genesis, for institutions to survive they must reach a self-enforcing equilibrium.4 As already suggested, this means that the actors should be better off by obeying the rules, that they should change their be- havior in response to other parties doing so, and that they will effectively counter attempts or political transgressions to change the rules of the game. When Do Institutional Equilibria Change? Self-enforcing institutions embed and lock in context: power discrepancies among social groups, economic development, state capacity, and many other background conditions (North 1981, 18–50; 1990). Therefore, the prospects Theories and Mechanisms of Political Economy: Institutions and Equilibria 133 for profound change are likely to be shaped by significant shifts in the fol- lowing factors: • Power endowments of major actors • Demography • Technology • Property rights • Global economic production, distribution, and exchange • International and domestic political competition • Ideology. Change in Power Endowments of Actors. Institutions formalize and pre- serve preexisting power configurations; they reflect the underpinning bal- ance of power between political groups in society. Historically, institutions (especially INSTITUTIONS, or the core societal rules of interaction) devel- oped around certain power configurations and were created to preserve those power discrepancies.5 Because these rules simultaneously shape col- lective action and development outcomes, changing them is always chal- lenging and often impossible. Here is an eloquent metaphor that captures the difficulty of major institu- tional change. In an influential article pondering the causal link between in- stitutions and development outcomes, the author draws a comparison that all sport fans can relate to, and poses the following conundrum. Let’s say that two teams play basketball (or any other sport for that matter), but one of them has a significant height advantage over the other team. The rules of the game and the referees—as third-party enforcers—are designed to be impar- tial. The conceptual problem here, the analyst argues, is that independently of all institutional guarantees designed to insure objectivity, the taller team is still more likely to win. It also follows, according to this line of argument, that the rules that would bring the game closer to fairness, for example low- ering the basket, will not change in a world in which the taller or more pow- erful players make the rules, precisely because their team derives advantages from the status-quo (Przeworski 2004, 529). Therefore, major institutional change occurs only when the underlying balance of power changes. One example of such institutional shifts comes from the historical evo- lution of succession and inheritance rules and provides an interesting illus­tration. Throughout the Middle Ages, primogeniture (or the right of succession granted exclusively to the first-born son) coexisted with a single-income source available to the household (usually land assets). Pro- gressively, as the income alternatives diversified (the rise of industry, man- ufacturing, and the like), the bargaining power of the other siblings in- 134 Understanding Policy Change creased and ultimately succeeded in establishing more equal rules of succession and property division. Historians have shown that the main explanation behind equality-versus-first born favoritism inheritance rules in the 19th century was the availability of multiple sources of income for household members (Knight 1992, 169). In more contemporary settings, transitions from autocracy to democracy occur when the nondemocratic coalition revolving around the political leader (that is, the military, one-party regime elites, powerful bureaucrats or notables, and broad societal groups that support the nondemocratic regime) fear defeat or replacement, have depleted the resources that bought political support, or have lost an open conflict with contenders. Lower-case institutions also change when the power endowments of ma- jor actors in society shift. South Africa, for example, pursued a comprehen- sive decentralization policy at the beginning of the 1990s. Anticipating elec- toral defeat by the transition incumbent (the African National Congress, or ANC), the two establishment parties (the National Party dominant in the Eastern Cape Province and the Inkatha Party with a strong support base in Kwa Zulu Natal) pushed for enhanced fiscal and decision power in the re- gions to try to preserve some political gains, despite the loss at the national level. The ANC accepted the plan within the context of the transition pact. Years later, when the party consolidated, it withdrew support from decen- tralization to weaken the opposition parties in the provinces and internal party rivals. The government dramatically reduced the number of local gov- ernments, from nearly 900 to 238, in a process of amalgamation officially justified as a bureaucratic measure meant to address the precarious capacity problems of local governments. As some analysts pointed out, the policy marked a reassertion of centralized control underneath an administrative guise (Eaton, Kaiser, and Smoke 2010, 54). Understanding institutional change as such has policy implications for re- formers because it reveals the limits and perils of institutional engineering. If the underlying power relations are essential for equilibria, then just adopt- ing or importing formal institutions that worked elsewhere can either fail to produce any results or do more harm than good. Demographic Change. Increases or decreases in population or changes in the demographic structure alter economic demand and production capabili- ties and create potential for changing the previously prevailing rules of distri- bution. For example, gender-friendly policy reforms in postgenocide Rwanda illustrate how institutions changed to reflect shifting demographics (Powley 2005). In October 2003, Rwanda found itself with the most gender-balanced legislature in the world, with women holding 48.8 percent of the seats. Prior Theories and Mechanisms of Political Economy: Institutions and Equilibria 135 to the civil war and genocide of the mid-1990s, they had never held more than 16 percent of the seats in the legislature. This puzzling reversal was partially caused by the massive demographic change triggered by the ethnic conflict that claimed over 800,000 lives. In the immediate aftermath of the genocide, the entire population was 70 percent female (women and girls). To this day, the legacy of gender imbalance persists. In a postconflict situ- ation in which women survived their husbands and sons, they had to acquire new skills rapidly and assume new gender roles. Into the organizational vacuum, multiethnic women from nongovernmental organizations (NGOs) such as Pro-Femmes stepped in, advocated for ethnic reconciliation, and ad- vised the government on gender-friendly policies. The inclusive, reconcilia- tory, and gender-sensitive constitution adopted in the postconflict period, coupled with the role of women NGOs in fostering national reconciliation, was critical in raising the political profile of prospective female politicians. The greater formal representation of women resulted in the formation of a women’s caucus and the passage of legislation that sought to promote wom- en’s issues, including the right of women to inherit land (1999), which had previously been prohibited. Thus, while Rwanda faced many developmental and political challenges, the change in the demographic composition of the broader population, as well as of the legislature, still had a clear and visible effect on policy-making priorities. Technological Change. The introduction of new technology increases effi- ciency, productivity, and power, thereby altering the relative costs of advocat- ing for a new structure of redistribution. For example, with the increased ability of developing countries to use electronic transfers to send benefits to recipients rather than going through bureaucratic intermediaries, the discre- tionary power and potential for corruption of the public sector apparatus may be reduced. Over the past 10 years, the Internet and social media (such as Facebook and Twitter) have radically transformed social movements, pop- ular protests, and political party appeals and fundraising and have thus facili- tated collective action and empowered marginalized groups of citizens. Change in Property Rights. Property rights are, broadly speaking, rules “defining ownership, use, rights of income and alienability of resources and assets as expressed in laws and regulationsâ€? (North 2005, 57). They emerge as a result of an ongoing negotiation between the rulers of a state and its constituents and specify the duties and obligations of the ruler with respect to the assets (land, labor, capital) of individuals and companies. Changes in the specification or enforcement of the system of property rights can trigger institutional change. 136 Understanding Policy Change Changes in Global Production, Distribution, and Prices. For Nobel Prize–winning economist Douglas North, changes in tastes and in the ratio of the prices of the factors of production are the most important drivers of incremental institutional change. In medieval Europe, for example, the plague caused dramatic changes in the land-labor ratio and triggered an im- portant resetting of political and economic rules. Similarly, changes in the relative prices of work, leisure, and contraception have modified the struc- ture of the family unit and have altered behavior (North 1990, 84). Shifts in Domestic and International Political Competition. The inten- sity and robustness of political competition will limit the monopoly power of the state. Shifts in either international or domestic competition can trigger institutional change (North 1990). According to increasing evidence in many developing countries, greater political competition has led to the creation of independent regulatory agen- cies, oversight institutions, civil service reforms, and ombudsmen (Geddes 1994; Grzymala-Busse 2006; O’Dwyer 2006). If political parties are uncer- tain about their electoral future when taking office, the argument goes, they have incentives to build in oversight institutions and public bureaucracies that will not punish them if they lose elections. Following the transition from communism, some Eastern European countries (Estonia, Hungary, Lithuania, Poland, and Slovenia) adopted rela- tively fast civil service reforms and established independent supreme audit institutions or public procurement agencies. Others (Bulgaria, the Czech Re- public, Latvia, and Slovakia) took almost 10 years to initiate a similar process of institution building. Paradoxically, some of the laggard state reformers like the Czech Republic were well ahead of other countries in market re- forms as well as in prospects for accession to the European Union. What explains this puzzle? A study found that robust political competition among parties limited the discretion of executives who might have wanted to use state coffers to maintain their incumbency. Fearing replacement in future elections and potential retaliation if the opposition were to take power, ­ incumbent politicians created relatively neutral monitoring and oversight institutions that could not be colonized by political rivals and that would guarantee a level playing field (Grzymala-Busse 2006). Major Ideological Shifts. The power of ideas to transform institutions has long been noted and documented (Hayek 1978). The transition from capi- talism to communism and vice versa triggered profound changes across many institutions: property rights were reconfigured through waves of na- tionalization and privatization, the type of political regime was transformed Theories and Mechanisms of Political Economy: Institutions and Equilibria 137 overnight, and individual incentives were considerably altered. To under- stand the powerful exogenous impact of such ideological shifts, just imag- ine the sweeping capacity of some states to modify preexisting social norms that had survived for centuries, as well as the very fabric of property rights and basic social interactions. For example, Mao’s China banned the practice of women’s foot binding overnight. At the other end of the spectrum, post- Communist regimes attempted to reverse the logic of land property rights following decollectivization but in the process interfered with the society’s understanding of kinship and family networks (Verdery 2003). As Concepts in Practice 4.6 will show with respect to Estonia’s transition from commu- CONCEPTS IN PRACTICE 4.6 The Power of Ideas and Tax Reforms Mart Laar, a historian by training, became prime form a very cool reception. Instead, the idea minister of newly independent Estonia in 1992. came from Milton Friedman’s book, Free to During his first term as prime minister, from Choose, which advocated a flat tax and was the 1992 to 1994, he pursued a radical change in only economic text that Mart Laar claimed economic policies that saw Estonia’s economy to have read. Thinking Friedman’s idea logical, transformed from a part of the USSR to one of Laar simply implemented it. the most liberal market economies. A particularly Following transition, many Estonian eco­ contentious component of this reform policy nomists thought that the imposition of such was the implementation of a flat rate on income radical economic policies was unworkable. meant to simplify the tax code, increase revenue Laar, however, assumed that rapid privatization (decrease tax evasion), and help cement Esto- and the flat tax reforms inspired by Friedman’s nia’s reputation as a business-friendly country. work have already been tested, successfully Eventually, Estonia’s flat-tax example was emu- adopted, and well routinized in many Western lated by many other formerly Communist coun- countries. In his own words, his reform team tries, including Bulgaria, Latvia, Lithuania, Roma- ‘walked on water’ not knowing it to be impos- nia, the Russian Federation, Slovakia, Ukraine, sible (CATO 2006). and, most recently, Hungary (Bemis 2011). In this way, a major element of Milton Fried- How did Estonia become Eastern Europe’s man’s fiscal policy prescriptions became the pioneer in the flat-tax revolution? Western insti- cornerstone of tax policy in most of Eastern Eu- tutions such as the European Union did not sup- rope and the former USSR—not because of the port the flat tax, and in fact the International prescriptions of the economics profession, but Monetary Fund initially gave the proposed re- because of the reading priorities of a historian. Source: CATO Institute (2006); Bemis (2011). 138 Understanding Policy Change nism in the early 1990s, shifting ideologies and theories can have a big im- pact on reform opportunities and processes. The Other Side of the Coin: Institutional Resilience, Path Dependency, and Critical Junctures One theoretical perspective6 departs from the idea of institutions as “self- reinforcing equilibriaâ€? and focuses instead on particular historical junctures that have led to lasting arrangements. In this account, the resilience of seem- ingly underperforming institutions sheds light on the long-lasting mecha- nisms of the reproduction of rules, despite shocks that could have disrupted them (Thelen 1999). This concept, often called path dependence, means that outcomes and pat- terns are locked in for long periods of time, making reversal of the processes hard or impossible (Pierson and Skocpol 2002). Even if the rules to which the stakeholders subscribe stall development, they survive because of the reinforcing feedback mechanisms that generated suboptimal incentives for the actors in the first place. An often quoted example is the development of the so-called QWERTY typewriter keyboard at the end of the 19th century (David 1985).7 Despite the later invention of more efficient alternative products,8 QWERTY still be- came the most commonly used keyboard on the market because a number of factors converged to lock it in as the predominant choice. Once a critical mass of consumers started using it, the switching costs (the time one would need to invest in learning a new typing system and transitioning to a new keyboard) were prohibitively high. In addition, the larger the pool of con- sumers using QWERTY, the more profitable for specialized stores to sell QWERTY typewriters or computers and the more useful for new users to learn this style as opposed to direct competitors like the Dvorak keyboard. Economists often call this reinforcing process “network externalitiesâ€? or “external increasing returns.â€? This analytical perspective emphasizes the role of history, temporal sequences, macro political-economic contexts, and complex configura- tions of structural conditions over long periods of time. The story of the QWERTY typewriter might have some lessons for development as well. As shown in Concepts in Practice 4.7 with respect to the rise of neo- patrimonialism in Africa, a legacy of weak institutions and strong tribal and ethnic loyalties can generate powerful incentives that perpetuate ethnicity- based patronage even when it is clearly suboptimal for development. Theories and Mechanisms of Political Economy: Institutions and Equilibria 139 CONCEPTS IN PRACTICE 4.7 Neo-Patrimonialism in Africa and Historical Legacies After decolonization, many African states inherited weak political institutions. This void created the space for elites associated with governments to rely on patronage. Thus, the state became neo-patrimonial, combining modern bureau- cracies with traditional patron-client networks built on the exchange between political support and concrete material benefits for narrow pockets of elites. This pattern of rule significantly constrains the possibilities for policy change to this day. Other long-lasting historical legacies have also led to highly resilient struc- tures of political influence. In former settler colonies like Kenya, Namibia, and Zimbabwe, the early European farmers established powerful interest groups whose political clout is still observable today in agricultural policy making. Source: Van de Walle 2001. Institutional (In)stability Institutions are stable or resilient when they survive major economic and po- litical changes and endure over longer periods of time (Levitsky and Murillo 2006). As much as path dependency and long periods of time might be responsible for locking in some very resilient underperforming institutions and the bad equilibria they generate, if institutions attempt to solve coordina- tion and collective action problems and to stabilize actors’ expectations, sud- den changes and high volatility can also have a negative effect on beliefs, ex- pectations, and preferences. If constitutions, legislatures, supreme courts of justice, or laws change overnight or if the civil service is staffed with incum- bent party supporters following each election, then societal actors will fail to make long-term investments, with immediate consequences for economic development. As Concepts in Practice 4.8 below illustrates, the lack of institu- tional predictability creates a vicious circle that some call the “instability trapâ€? (Helmke and Levitsky 2006; Levitsky and Murillo 2009), perpetuating a self-reinforcing trajectory of weak institutions over a long period of time. Symptoms of Institutional Failure: Lack of Enforcement Institutions generate pro-development incentives when they function prop- erly. In the real world, rules vary on how consequential they are for policy 140 Understanding Policy Change CONCEPTS IN PRACTICE 4.8 Lifespan of National Constitutions around the Globe In the entire universe of national constitutions, on average, institutional ar- rangements last 19 years. The range of variation is huge: the U.S. Constitution, drafted in 1789, has survived for more than 220 years; at the other extreme, the Dominican Republic and Haiti together account for about 7 percent of the global number of constitutions because of their dramatic instability. Let us take a quick look at the island of Hispaniola—a true “constitutional graveyardâ€? (Elkins, Ginsburg, and Melton 2009, 180–185). During the 20th century, both the Dominican Republic and Haiti, the two countries sharing the tiny island, oscillated between democratic and autocratic regimes. The wide polarization between incumbents and oppositions increased the stakes of staying in power. Accordingly, when one of the parties came to office, it changed the constitution to set rules that would favor it politically. Paradoxi- cally, during periods of strong executive rules in both countries (the Trujillo years in the Dominican Republic or the Duvalier regime in Haiti), presidents did not just make amendments—the easier path—but rather rewrote the legal document completely. This self-reinforcing constitutional death pattern cre- ated a spiral of all-or-nothing gains that, in turn, fed a climate of severe politi- cal and economic instability. Source: Elkins, Ginsburg, and Melton (2009). outcomes. Most reforms are attempts to remedy one form or another of in- stitutional failure or malfunctioning. If actors ignore the rules of the game, believe that the rules are unfair, or think they themselves will not be sanc- tioned for noncompliance, the rules fail to produce incentives of any kind. Good enforcement and some degree of stability are symptoms or mani­ festations of well-functioning institutions (Levitsky and Murillo 2009). Enforcement refers to the degree of actors’ compliance with the rules of ­ the game. Weak Institutional Enforcement Many developing countries are characterized by weak formal institutions, created intentionally or unintentionally. In some cases, the actors who wrote the rules genuinely want to enforce them, if they have the capacity to do so. More often, they lack either the capacity or the will to enforce them (Lev- itsky and Murillo 2009, 120). Theories and Mechanisms of Political Economy: Institutions and Equilibria 141 A weak civil service or a limited reach of the central state into remote regions of a country, for example, has been found to severely curtail the state’s ability to monitor and sanction deviation from the rule of law. Some studies suggested that, in the 19th century, the Peruvian state, because of its unwillingness to challenge local hierarchies, had limited reach in the coun- tryside and therefore failed to effectively raise taxes. This failure signifi- cantly hindered the development of the rural regions relative to their urban counterparts. In contrast, Chile was able to exercise more effective control over its rural areas and, therefore, has been able to craft and implement more successful development projects in both rural and urban regions (Soifer 2006). Why is it that Chile and Mexico had a greater state capacity than Colom- bia and Peru, despite their shared cultural and institutional legacies? One historical factor affecting the capacity and effectiveness of the state seems to have been the ability and willingness of key elites in Chile and Mexico to use national bureaucrats to make and implement local policies, thereby freeing decision making from the grip of entrenched local patrons. While such a policy has its limits—bureaucrats’ incentives were not necessarily fully aligned with those of local populations—they were effective in ensuring that national policies (infrastructure projects, tax collection) were implemented more effectively than when local power barons could capture national re- sources with little oversight on how these funds were used.9 In general, lack of political incentives is a key factor in weak institutional enforcement. Many governments in the developing world often adopt “window-dressingâ€? institutions in order to comply with the donor commu- nity’s conditions on borrowing but never plan to enforce them (Levitsky and Murillo 2009, 120). This is known as the Potemkin (or “fakeâ€?) village phe- nomenon. Politicians usually lack ­ incentives to build “realâ€? villages because citizens cannot act collectively and sanction them for deviation, preferring instead to pocket the money ­ allocated for development projects. In other cases, weak institutional enforcement is the product of the bar- gaining strategies of political actors. For example, even if initially Brazil and South Africa could not guarantee enforcement of certain racial- and gender- equality rights inscribed in their constitutions, human rights activists still pushed for their official inclusion in the legislation, hoping that they could become effective at a later point in time (Htun 2003; Levitsky and Murillo 2009, 121). Weak institutional enforcement, in some cases, is also the product of ­ demand-side factors such as societal compliance (or lack thereof ). High socioeconomic inequality can hamper systematic enforcement. In the early ­ 1990s, Brazil and South Africa were two of the most unequal societies on 142 Understanding Policy Change earth. Despite the similar degree of income inequality, the ability of the state to collect taxes and implement fiscal policies varied remarkably. South Africa was able to collect taxes effectively and ensure that even rich citizens complied. Conversely, in Brazil rich elites were able to exempt themselves from most direct taxation, ensuring that the tax burden fell disproportion- ately on the poor. One of the reasons that tax compliance was so different in the two coun- tries stemmed from the social cleavages characterizing these societies. In South Africa, as a result of the solidarity between rich and poor whites, the rich were more willing to comply with the tax demands of the state. Con- versely, in Brazil, where official discrimination was prohibited—although it was pervasive in practice—the rich felt less solidarity with any segment of the rest of the population. Thus, paradoxically, it was easier to establish fiscal policies that sought to redress past injustices against the Black population in South Africa (following the 1994 transition to democracy) than in Brazil, where fiscal compliance by the rich continues to be erratic and the state thus relies on more regressive and distortionary forms of taxation.10 Sometimes, the core logic of certain institutional arrangements stems from weak compliance. In many low- to middle-income countries, as public sector jobs decreased and incumbent politicians failed to credibly commit to generating employment for young college graduates, the informal economy became a necessary exit opportunity for the unemployed. Any attempts to reduce it by formalization met strong political discontent and resistance (Hibou 2006). Similarly, in some contexts, selective enforcement of institutional rules is a powerful strategic tool of political co-optation or punishment. In Ben Ali’s Tunisia, for example, access to bank credits or development subsidies for entrepreneurs was often granted based on personal relationships with key political actors, not on market criteria (Hibou 2006). In general, clientelistic political systems, in which the incumbent party channels benefits only to its own voters, function on a logic of weak institu- tional enforcement of formal rules. Symptoms of weak institutional enforce- ment may include: • Weak capacity (lack of resources) • Weak political will • Eroding informal institutions (clientelism) • Adopted institutions that fail to achieve equilibrium because they do not fit the underlying distribution of power in society • High stakes of the game (socioeconomic inequalities or wide political polarization). Theories and Mechanisms of Political Economy: Institutions and Equilibria 143 The underlying causes of institutional failure are also the obstacles to collec- tive action that we will be exploring in depth in the following chapters, which include information discrepancies, absence of credible commitment on the part of politicians, agency problems, institutional manipulation, and other forms of imperfections that erode the capacity of citizens to act effectively. To sum up, taking into consideration the factors discussed in this chapter, when conducting a political-economy analysis, the researcher should ask the following questions about institutions and organizations: • What are the formal and informal rules governing your policy domain of choice? • Are they enforced or not? Why? • Are they durable? Did they survive any major exogenous shocks in the recent past? • Why does stability vary across different institutions? • What is the analytical link among institutions, incentives, and the poten- tial for pro-development collective action in your policy area or sector? • Can you identify organizations involved in a project you have worked on or a public policy issue you are interested in? • How did the institutional context in which these organizations operate affect their ability? • Did the rules of the game make it easier or more difficult for each indi- vidual organization to achieve its goal? • Are these organizations monolithic in preferences and reform stakes? • Which organizational subunits would gain most from the passing of a cer- tain piece of legislation? Do they benefit or not from the current rules? Institutions affect how actors interact with each other, determining the outcomes of a social situation. Therefore, for a better knowledge of policy consequences, it is essential to understand the logic of strategic interaction between actors. This process will enable us to think about how different insti- tutions, by altering the nature of this interaction, can affect outcomes. Game theory provides the tools by which we can study the interaction of actors in different institutional contexts. The core ingredients of a game follow. Modeling the Role of Institutions with Game Theory11 Let’s recall the game introduced in chapter 3, dealing with the collective ac- tion problem two farmers faced. Because each farmer had an incentive not to cooperate (but to defect), both farmers were collectively worse off because 144 Understanding Policy Change they could not cooperate on building an irrigation system for both farms. Institutions can potentially mitigate such collective action problems by pro- viding mechanisms to reconcile the interests of individuals with those of the group. Assume, as in chapter 3, that the two farmers would benefit from con- structing an irrigation system for their farms. Furthermore, assume that they now live in a country with a very robust and efficient legal system that en- ables them to write contracts with each other. Specifically, if the farmers en- ter into an agreement to construct an irrigation system jointly, but one of them does not cooperate (that is, the farmer defects), the courts will speedily and efficiently fine the noncooperating (defecting) farmer, effectively taking her crop yield and using it to compensate the farmer who did not renege on her agreement. Assuming that the farmers have entered a legal agreement to cooperate and build the irrigation system; the preferences of the first farmer (the farm- ers’ preferences are symmetrical) are based on the payoffs she would receive, as follows: • If the first farmer reneges on her construction commitments but the sec- ond farmer does not renege, the first farmer is fined, losing all her crops. The second farmer bears the costs of the irrigation system, but she also has enough food to sustain her family (due to the fine) and subsequently benefits from the existence of the irrigation system. • If the first farmer builds the irrigation system but the second farmer re- neges on her construction commitment, then the first farmer bears the full cost but also receives compensation (able to feed her family) and the benefit of the irrigation system. The second farmer loses all her crops (due to the fine). • If both the first and the second farmer build the irrigation system, they both bear some of the cost (less time to grow their own crops), but they both enjoy the benefits of higher crop yields. • If both farmers renege on the agreement, the agreement is rendered void, and they both have a normal yield (equivalent to the fine). The diagram below shows these payoffs. Cooperation is now a strictly dominant strategy; that is, both farmers will cooperate regardless of what they think the other farmer will do. Cooperateb Defecta Cooperate b 3 3 2 1 Defect a 1 2 1 1 a. Defect here refers to noncooperation. b. “Cooperate, cooperateâ€? for both farmers is the equilibrium of the game, denoted by the asterisk. Theories and Mechanisms of Political Economy: Institutions and Equilibria 145 Summary The outcomes of the political process depend on how the institutional con- text determines the feasible courses of action that different stakeholders can take. Therefore, being able to map and analyze the institutional context can significantly increase the value added of a political-economy analysis. By identifying the formal and informal institutions and the conditions under which they may change, a researcher can anticipate the equilibria that exist or that may emerge. As the next few chapters will demonstrate, specific in- stitutional settings can engender various opportunities and challenges to stakeholders’ ability to realize their own agenda. The reformer will thus be able to identify how some rules may empower stakeholders either to support or to oppose a reform proposal or set project priorities appropriately. 146 Understanding Policy Change Exercises for Chapter 4: Theories and Mechanisms of Political Economy: Institutions and Equilibria Exercise 4.1: The Prisoner’s Dilemma Game12 The facilitator explains the steps involved in the Prisoner’s Dilemma. Steps in the exerciser (time frame: ~40–45 minutes): 1. Read out instructions (1 minute) 2. Allow participants to read and answer (5–10 minutes) 3. Collect responses, tally responses, facilitate group discussion (10 minutes) 4. Lecture (10–15 minutes) 5. Class discussion (10 minutes) Preparation and Materials: 1. Envelopes (A4 size) to be placed in the middle of the round table (one at each table) 2. Pens should be available to all participants 3. Copies of the instructions 4. Copies of the answer sheet 5. Copies of the key questions handout and group discussion 6. Facilitator instructions Procedure: 1. The facilitator reads aloud the following instructions: I am going to distribute instructions and an answer sheet. Read the in- structions and answer the questions on the answer sheet. Once you have read the instructions and completed your answers, place your answer sheet in the envelope in the middle of the table. DO NOT WRITE YOUR NAME ON EITHER DOCUMENT and MAKE SURE NO ONE SEES YOUR ANSWERS. Keep the instructions handy for the follow-up discus- sion. I will give you about 5–10 minutes to complete this. 2. The facilitator distributes the instructions handout and answer sheet to all the participants. 3. The facilitator gauges the rate of completion of the assignment and gives participants a 2-minute warning so that the activity can wrap up within its allotted 10-minute time frame. As part of the 2-minute warning, the facilitator reminds participants that they should not write their names on the answer sheet; when they have finished, they should put their answer sheet in the envelope in the middle of the table. Theories and Mechanisms of Political Economy: Institutions and Equilibria 147 4. The facilitator explains: I will come around to collect the envelopes with the answer sheets and tally your responses. At the same time, I will give you a handout that in- cludes two key questions. While I tally the responses on the answer sheets, please go ahead and discuss at your tables the two questions on the handout. Spend approximately 10 minutes in this discussion. 5. To tally the votes, the facilitator uses a blank answer sheet and notes for each option the total number of votes. 6. The facilitator walks around the room among the tables in a nonintrusive manner and pays attention to what is discussed. The facilitator notes a few comments from the tables that he or she can highlight as part of the transition to the lecture portion. 7. The facilitator gives the participants a two-minute warning to wrap up their discussion. 8. The facilitator explains: Now, that you have played the game and had a chance to discuss with oth- ers at your table some of the key underlying issues, I will go ahead with a mini-lecture to talk about the technical aspects in relation to the concep- tual aspects that you addressed in your discussions. In fact, I noted the following key words being used at several tables: a. Trust 9. The facilitator presents the lecture beginning with the announcement of the final overall tally by entering the results on the first slide. The facilita- tor ends the lecture by posing the following questions to the entire group: a. How many of you were surprised by the entire group’s final tally that I shared at the beginning of the slide? b. If yes, why? c. If no, why? 148 Understanding Policy Change Handout Exercise Set 4.1: Instructions Handout 1 You and a colleague (whom you do not know well) have been arrested after committing a crime. Upon arrest you were separated (there is no way to communicate with him or her). You are now in a police cell and the police negotiator informs you that they are willing to offer you the following deal, which is simultaneously also being offered to your colleague (you can as- sume this information is true): • If you confess to the crime and your colleague does not confess to the crime, you will receive a reduced sentence of 1 year in prison, while your colleague will receive a full sentence of 12 years in prison. • If you do not confess to the crime but your colleague does confess to the crime, you will receive a full sentence of 12 years in prison while your col- league will receive a reduced sentence of 1 year in prison. • If you both confess to the crime, you will both receive a reduced sentence of 6 years in prison. • If neither one of you confesses, you will both receive a reduced sentence of 3 years. Theories and Mechanisms of Political Economy: Institutions and Equilibria 149 Handout Exercise Set 4.1: Answer Sheet 1 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Confess to the Crime Do Not Confess to the Crime Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE ONLY TICK ONE BOX. He/she will confess to the crime He/she will not confess to the crime Briefly explain why you decided to pursue one action rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. 150 Understanding Policy Change Handout Exercise Set 4.1: Key Questions Handout 1 — Group Discussion In your group, please discuss the following questions: 1. What do you expect the result of the game/tally to be among all partici- pants across all tables? Why do you think this is the case? 2. How does taking into account someone else’s behavior affect your decision? Theories and Mechanisms of Political Economy: Institutions and Equilibria 151 Notes  1. This distinction is specific to the new economic institutionalism.  2. Self-enforcing equilibrium has the following characteristics: (1) best strategy among the feasible strategies for all players; (2) any change in behavior triggers change in the other players with inferior outcomes; and (3) actors have the capacity and willingness to punish deviant behavior.  3. See also North (1984, 1990).  4. See the separate subsection for the concept of self-enforcing equilibrium.  5. Various institutionalism schools operate with different assumptions related to how “endogenousâ€? (embedded in preexisting social relations and power configurations), or “exogenousâ€? (actors have well defined ex ante preferences) institutions are. The Rational Choice School, for example, assumes for the most part that institutions are exogenous (Weingast 2002).  6. Historical institutionalism.  7. For a critique of this interpretation of the QWERTY narrative, see Liebowitz and Margolis (1990).  8. The Dvorak Simplified Keyboard, invented in 1936, was technically superior to QWERTY on many indicators such as finger motion, typing time, and the like.  9. Adapted from Soifer (2006). 10. See Lieberman (2003). 11. Source: Inspired by (Bardhan 1993). 12. Source: Inspired by (Gibbons 1992). References Alesina, Alberto, Reza Baqir, and William Easterly. 1999. “Public Goods and Ethnic Divisions.â€? Policy Research Working Paper 2108. World Bank, Development Research Group, Washington, DC. Ames, Barry. 1999. “Approaches to the Study of Institutions in Latin American Politics.â€? Latin American Research Review 34 (1): 221–36 ———. 2001. The Deadlock of Democracy in Brazil: Interests, Identities, and Institu- tions in Comparative Politics. Ann Arbor: University of Michigan Press. ARD. 2006. “Yemen Corruption Assessment: Technical Report.â€? USAID, Washing- ton, DC. Aslund, Anders, Peter Boone, and Simon Johnson. 2002. “Escaping the Under- Reform Trap.â€? Staff Working Paper 48, International Monetary Fund, Washington, DC. Bardhan, P. 1993. “Analytics of the Institutions of Informal Cooperation in Rural Development.â€? World Development 21 (4): 633–639. Bardhan, Pranab. 2002. “Decentralization of Governance and Development.â€? Journal of Economic Perspectives 16 (4): 185–205. Bemis, T. 2011. ‘Flat-tax countries include Bulgaria, Mongolia, Iraq,’ Market Watch, The Wall Street Journal website, http://blogs.marketwatch.com/thetell/ 152 Understanding Policy Change 2011/10/25/bulgaria-mongolia-guernsey-among-the-23-flat-tax-countries/. Web accessed: 6/18/2012 Besley, Timothy, and Robin Burgess. 2002. “Can Labor Regulation Hinder Eco- nomic Performance? Evidence from India.â€? Quarterly Journal of Economics 119 (1): 91–134. Carey, John. 2000. “Parchment, Equilibria, and Institutions.â€? Comparative Political Studies 33 (6/7): 735–61. Carey, John M., and Matthew Soberg Shugart. 1995. “Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas.â€? Electoral Studies 14 (4): 417–39. Cato Institute. 2006. “Individual Liberty, Free Markets, and Peace: Mart Laar’s Biography.â€? Cato Institute Website, http://www.cato.org/special/friedman/laar/. Crawford, Sues E. S., and Elinor Ostrom. 1995. “A Grammar of Institutions.â€? American Political Science Review 89 (3): 582–600. Crisp, Brian F. 2000. Democratic Institutional Design: The Powers and Incentives of Venezuelan Politicians and Interest Groups. Palo Alto: Stanford University Press. Danchin, Peter. 2010. “A Human Right to Water? The South African Constitutional Court’s Decision in the Mazibuko Case.â€? EJIL: Talk!, http://www .ejiltalk.org/a-human-right-to-water-the-south-african- constitutional-court%E2%80%99s-decision-in-the-mazibuko-case/. David, Paul A. 1985. “Clio and the Economics of QWERTY.â€? American Economic Review 77 (2): 332–37. Eaton, Kent, Kai Kaiser, and Paul Smoke. 2010. The Political Economy of Decentral- ization Reforms: Implications for Aid Effectiveness. Washington, DC: World Bank. Egel, Daniel. 2010. “Tribal Diversity, Political Patronage and the Yemeni Decentral- ization Experiment.â€? Unpublished manuscript, http://www.cgdev.org/doc/ events/Post-Doc%20Seminars/Daniel_Egel.pdf. Elkins, Zachary, Tom Ginsburg, and James Melton. 2009. The Endurance of National Constitutions. Cambridge: Cambridge University Press. Geddes, B. 1994. Politician’s Dilemma: Building State Capacity in Latin America. Berkeley: University of California Press. Gibbons, R. 1992. A Primer in Game Theory. Harlow: Prentice Hall. Golden, Miriam A., and Eric C. C. Chang. 2001. “Competitive Corruption: Factional Conflict and Political Malfeasance in Postwar Italian Christian Democracy.â€? World Politics 52 (July): 588–622. Grzymala-Busse, Anna. 2006. “The Discreet Charm of Formal Institutions: Post-Communist Party Competition and State Oversight.â€? Comparative Political Studies 39 (10): 1–30. Hayek, Friedrich A. 1962. Rules, Perception and Intelligibility. Oxford: Oxford University Press. ———. 1978. Constitution of Liberty. Chicago: University of Chicago Press. Helmke, G., and S. Levitsky, eds. 2006. Informal Institutions and Democracy: Lessons from Latin America. Baltimore: Johns Hopkins University Press. Hibou, BeÌ?atrice. 2006. La Force De L’obeÌ?Issance: Economie Politique De La ReÌ?Pression En Tunisie. Textes aÌ€ L›appui Histoire Contemporaine. Paris: DeÌ?couverte. Theories and Mechanisms of Political Economy: Institutions and Equilibria 153 Htun, Mala. 2003. Sex and the State. New York: Cambridge University Press. Knight, Jack. 1992. Institutions and Social Conflict. Cambridge: Cambridge ­University Press. Levitsky, Steven, and María Victoria Murillo. 2009. “Variation in Institutional Strength: Causes and Implications.â€? Annual Review of Political Science 12: 115–33. Lieberman Evan. 2003. Race and Regionalism in the Politics of Taxation in Brazil and South Africa. Cambridge, UK: Cambridge University Press. Liebowitz, S. J., and Stephen E. Margolis. 1990. “The Fable of the Keys.â€? Journal of Law and Economics 33 (1): 1–25. Lijphart, A. 1999. Patterns of Democracy. New Haven: Yale University Press. Mehta, Lyla, and Zolile Ntshona. 2004. “Dancing to Two Tunes? Rights and Market-Based Approaches in South Africa’s Water Domain.â€? Research Paper 17, Institute of Development Studies, University of Sussex. Michels, Robert. 1911. Political Parties: A Sociological Study of the Oligarchical Tendencies of Modern Democracy. Piscataway, NJ: Transaction Publishers. Nicol, Alan, and Sobona Mtisi. 2003. “Politics and Water Policy: A Southern Africa Example.â€? IDS Bulletin 34 (3): 41–53. North, Douglass C. 1984. “Government and the Cost of Exchange in History.â€? Journal of Economic History 44 (2): 255–64. North, Douglas. 1981. Structure and Change in Economic History. New York: Norton. ———. 1990. Institutions, Institutional Change, and Economic Performance. New York: Cambridge University Press. ———. 2005. Understanding the Process of Economic Change. Princeton: Princeton University Press. O’Dwyer, Conor. 2006. Runaway State-Building: Patronage Politics and Democratic Development. Baltimore: Johns Hopkins University Press. Olson, Mancur. 1981. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven: Yale University Press. Persson, Torsten, and Guido Tabellini. 2005. The Economic Effects of Constitutions. Cambridge: MIT Press. Pierson, Paul, and Theda Skocpol. 2002. “Historical Institutionalism in Contempo- rary Political Science.â€? In Political Science: State of the Discipline, ed. Ira Katznel- son and Helen V. Milner, 693–721. New York: W.W. Norton. Powley, E. 2005. “Women Hold Up Half the Parliament.â€? In Women in Parliament: Beyond the Numbers, ed. F. Ginwala and A. Karam. Stockholm: International Institute for Democracy and Electoral Assistance. Przeworski, Adam. 2004. “Institutions Matter?â€? Government and Opposition 39 (4): 527–40. Soifer, H. D. 2006. Authority over Distance: Explaining Variation in State Infrastruc- tural Power in Latin America. PhD Dissertation. Cambridge, MA: Harvard University Press. Thelen, K. 1999. “Historical Institutionalism in Comparative Politics.â€? Annual Review of Political Science 2 (1): 369–404. 154 Understanding Policy Change Tsai, Lily. 2007. Accountability without Democracy: Solidary Groups and Public Goods Provision in Rural China. Cambridge: Cambridge University Press. Van de Walle, Nicolas. 2001. African Economies and the Politics of Permanent Crisis, 1979–1999. Cambridge: Cambridge University Press. Verdery, K. 2003. The Vanishing Hectare: Property and Value in Postsocialist Transylvania. Ithaca: Cornell University Press. Weingast, Barry. 2002. “Rational-Choice Institutionalism.â€? In Political Science: State of the Discipline, ed. Ira Katznelson and Helen V. Milner, 660–92. New York: Norton. White, James. 1995. Ikki: Social Conflict and Political Protest in Early Modern Japan. Ithaca, NY: Cornell University Press. Yamin, Alicia, and Siri Gloppen, eds. 2011. Litigating Health Rights: Can Courts Bring More Justice to Health? Cambridge: Harvard University Press. Theories and Mechanisms of Political Economy: Institutions and Equilibria 155 CHAPTER 5 Collective Choice and Agenda Setting Collective decision making often entails making a choice between several different viable alternatives. Sometimes, how decision makers present and compare these alternatives—that is, how they set an agenda—can determine which policy alternative is chosen, regardless of merit. In many policy- making contexts (especially in legislatures), the final decision can depend on the agenda-setting power of certain actors rather than on the inherent desir- ability of any of the possible choices faced. Specifically, agenda setting is an institutional mechanism (see chapter 4) that mitigates the collective action problems faced by some or all of the indi- viduals or groups with a stake in policy outcomes (see chapter 3). Agenda setting can facilitate collective action by helping groups economize on infor- mation; by allowing a subset of the group to develop specialist knowledge (for example, committees in legislatures, such as foreign affairs, public fi- nance, defense, etc.); and by promoting group cohesion (for example, cohe- sion of a political party can be strengthened if committee appointments and membership are based on seniority or on a record of disciplined adherence to party policy). 157 FIGURE 5.1  Agenda-Setting Conceptual Map Characteristics of formal agenda setters Formal agenda setting How can formal agenda setting What is agenda affect policy? setting? Is it relevant to development? Characteristics of informal agenda setters Informal agenda setting How can informal agenda setting affect policy? Source: Authors. Therefore, understanding the role of agenda setting can help reformers both identify powerful agenda setters who might help or hinder their reform proposals, and design institutions to take into account future agenda-setting dynamics. Normatively, the mere existence of agenda setting raises fundamental questions about the ways in which majority will translates into concrete policy outcomes. Of course, it is also important to note that the absence of agenda-setting mechanisms can also present challenges. Specifically, the lack of agenda control can generate information problems and a lack of co- hesion among stakeholders. Learning about agenda setting is also relevant because in everyday development work the absence of such institutional mechanisms can lead to a lack of coordination, a weak detection and moni- toring capacity, and, subsequently, suboptimal organizational outcomes. Therefore, it is useful to acquire the analytical tools that help understand the origins, organizational location, and probability of change associated with specific institutional problems (see figure 5.1). This chapter uses the concept of agenda setting to show how a develop- ment practitioner can recognize and use institutional rules to alter policy outcomes. Reformers should ask themselves the following questions: • Among the variety of rules governing the policy domain of interest, which ones are most relevant for change? • Are these rules enforced? 158 Understanding Policy Change • What strategic tools can be used to navigate existing formal and informal rules to bring the policy outcomes closer to the ideal goals? • Who is responsible for scheduling the order in which policy proposals are submitted for debate and voted on according to the organizational procedure? Objectives of Chapter 5 By the end of this chapter, readers should be able to do the following: • Identify the relevant formal and informal agenda-setting institutions that govern the policy domain • Understand how agenda setting can induce changes in the final outcome irrespective of the distribution of stakeholder support for a certain policy alternative • Explore potential positive or negative effects on policy change and reform caused by the absence or existence of agenda-setting power. What Is Agenda Setting? Agenda control, or setting, refers to how the process of decision making af- fects outcomes irrespective of the substantive merit of each policy proposal.1 The sequencing of decision making, the timing of voting procedures, and the skillful use of rules that govern the form in which bills or amendments are brought up for legislative debate are classic examples of agenda setting and will be explored in depth throughout this chapter. An often-quoted example illustrates the point: John Dingell, a former Democratic representative in the U.S. Congress, once said, “I’ll let you write the substance . . . and you let me write the procedure, and I’ll screw you every time.â€?2 The tremendous politi- cal power held by the Committee on Rules in the U.S. Congress since the 1880s is clear evidence that the use of voting rules in policy making matters significantly for the final legislative results. The importance of agenda setting first came to prominence in the 1950s and 1960s as scholars in political science increasingly began to question the received orthodoxy that the raw power (number of voters, seats in parlia- ment, and the like) of a group translated directly into an increased ability to influence outcomes. One of the major criticisms of this “classical pluralistâ€? approach was that it did not take into account the ability of certain groups— not necessarily the largest groups but those that could overcome collective Collective Choice and Agenda Setting 159 action problems most effectively—to use their superior capabilities to set the agenda and manipulate outcomes in their favor (Riker 1981). We can think of agenda setting as having two dimensions (Bachrach and Baratz 1962): the first dimension is the choice between the alternatives on offer, and the second is the ability to determine what the choices are in the first place. Because agenda-setting individuals or groups may well not be the same as those who exercise the most power when deciding among alterna- tives, the importance of institutional rules and procedures governing the se- lection and scheduling of policy proposals can be significant. The following sections briefly explain the process through which the pol- icy preferences of stakeholders are aggregated to generate collective policy, thereby making it easy to see how strategic agenda setters might be able to use their power to influence outcomes. From Individual Preferences to Collective Choice Individuals and organizations have different interests, stakes, and incentives in policy change. For example, let’s assume that a local council has to deliber- ate over and vote on the location of a bridge. At the council, each of the three districts is represented by a councilor elected by the voters from that district, with each district having an equal number of votes on the council. One com- munity might prefer the bridge to be built in its own district, but if that is not possible, it will prefer that the bridge be built in the neighboring district, rather than the district that is farther away. Another community, depending on its geographical location, will order its preferences differently. Collective choice is the process that converts such heterogeneity of policy positions into concrete outcomes. Agenda setting (that is, the use of preference-aggregation rules such as voting) intermediates the way in which individual preferences translate into collective choice. Thus, agenda setting helps solve a fundamental problem characterizing majoritarian decision-making settings,3 namely, the fact that the aggregation process through which individual choices become policy decisions leads to instability or “cyclingâ€? among multiple possible equilibria (Riker 1981; Shepsle 1989). One of the most consequential institutions in many legislatures around the world is the use of committees to filter legislation. Committees develop spe- cialist knowledge of a policy topic and have the power to select bills for debate in the legislature. As a consequence, few bills reach the floor. Committees limit the number of policy proposals to a more manageable number, although they may also have the effect of limiting the ability of the full legislature to amend bills. In this way, the existence of committees plays a critical role in ensuring that the workload of legislatures remains manageable. 160 Understanding Policy Change The agenda-setting powers of legislative committees often translate into concrete development outcomes. Let’s take a look at an example of “handing out the porkâ€? in the German parliament (Stratmann and Baur 2002). Because committees decide which bills are voted on in the full leg- islative chamber or how the content of proposed legislation is altered, many committees have the power to ensure that the final version of a bill is closest to the interests of their (majority) membership. It has long been known that, because of these agenda-setting powers, election-seeking pol- iticians may attempt to obtain committee positions through which they can use this power to further the interests of their own constituents. For example, in the case of the German Parliament, in which the elections of some members are determined by a constituency plurality (first past the post) rule and others under proportional representation (or through party lists), there is significant self-sorting on committees as politicians jockey to provide rents to their voters. Thus, unsurprisingly, committees that allo- cate spending to certain geographical areas are dominated by politicians elected from constituencies, while committees that allocate spending in- dependent of geography but whose expenditure can be targeted to diffuse groups across the country are dominated by legislators elected by party lists. This desire to obtain relevant committee membership highlights the importance of agenda setting as a mechanism to please core voters and se- cure reelection. Cycling To return to our example of the new bridge to be constructed, assuming four policy alternatives (x = bridge construction in District 1, y = bridge construc- tion in District 2, and z = bridge construction in District 3), three groups of voters have the following ranking of preferences (notation P to be read as voter from District 1 prefers x to y, y to z, x to z, etc.): Voter from District 1: x P1 y P1 z Voter from District 2: y P2 z P2 x Voter from District 3: z P3 x P3 y We need a majority (50 percent plus one, in this case, two voters) to agree on their rankings of individual preferences in order to aggregate them opti- mally in the final ordering of the group preference. What alternative ranking would win two votes? Notice the unstable equilibria for x, y, and z. No matter how hard one tries to find a sequence of x, y, and z that would aggregate the will of the majority best, the efforts will fail (for the entire group, x P y P z P x P y P z P x). Therefore, x, y, and z form a cycle (Johnson 1998). Collective Choice and Agenda Setting 161 Sometimes in efforts to aggregate preferences, some policy alternatives will win independently of the sequencing of the voting or the ordering of any other alternative. This concept is often referred to as a Condorcet winner.4 Unlike alternatives that are subject to cycling, a Condorcet winner is an alternative that can defeat each of the other alternatives in any pairwise ma- jority vote. In this case, if our voters preferred a fourth alternative w (say, a bridge built at the joint border of all three districts) that defeats x, y, and z in all three cases described above, then w is a Condorcet winner. As Concepts in Practice 5.1 explains in more detail, this problem of iden- tifying the majority-preferred outcome in the presence of cycling is a major CONCEPTS IN PRACTICE 5.1 Optimum Outcomes and Constraints Given cycling, what is the major dilemma a pol- when individual preferences for other alter- icy maker faces when trying to identify the natives change, say w). ­ majority-preferred outcome? As Nobel Prize– • There is no dictator who can impose his or winning economist Kenneth Arrow formally her preferences on the group. This means demonstrated (Arrow 1951), no voting system that if one person in the group cares in- can aggregate individual preferences in a way tensely about having the bridge built in that satisfies the basic criteria of fairness. There- District 1, she cannot unilaterally impose ­ fore, it is impossible to know whether an out- her will so that the entire group prefers come reflects the will of the majority or is the the same location for the infrastructure product of agenda manipulation by politicians. project. Arrow’s famous (im)possibility theorem has While it is beyond the scope of this work to had a profound effect on our understanding of formally show how Arrow’s theorem is proved, the normative implications of collective deci- its implications for social choice are clear: in the- sion making. It has proved that no voting sys- ory, the fact that there is no aggregation method tem can aggregate individual preferences in a that satisfies all three conditions at the same manner that does not violate at least one of time implies that no voting procedure is ideal for the following three conditions, which are often reflecting the “willâ€? of the majority. Thus, as identified as being inherently desirable (fair) in Riker (1981) showed, the nonexistence of such collective decision making: an “objectiveâ€? social optimum enables political • If every individual in a group prefers alterna- entrepreneurs who can manipulate agendas to tive x to y, then the group will also prefer x impose their private preferences on society. In to y. other words, group decisions (public policy out- • If every individual’s preferences for alterna- comes) may reflect the private preferences of tives x and y remain the same, the group the most adept political manipulators rather than preferences will also remain the same (even the “will of the people.â€? 162 Understanding Policy Change preoccupation of social science, as it makes it difficult to argue that, even if a benevolent policy maker exists, he or she can actually know what is “bestâ€? for society. Why Agenda Setting? Given the inherent instability of aggregation (reflected in the previously in- troduced concept of cycling and the general impossibility theorem), it has been argued that agenda setters (the committee system in some legislative institutions, for example) can (1) eliminate or prioritize policy alternatives and (2) strategically instrumentalize the voting power of political actors in a way that leads to the so-called structure-induced equilibrium (Shepsle and Weingast 1984, 1987; Riker 1986; Ordeshook and Schwartz 1987; Tsebelis 1994, 1997). Simply put, by manipulating voting procedures, strategic agenda setters can structure the political debate in such a way that they are more likely to get what they want (Pollack 1997, 121–122). Types of Agenda Setting Often, in political economy, agenda setting refers to two distinct processes, formal and informal. Most of the official voting rules briefly described below fall into the category of formal agenda setting. The locus of agenda control and the political power of the agenda setter depend on the following factors (Pollack 1997): • Initiative rules. Every legislative body has an idiosyncratic set of actors endowed with the right to initiate legislation. For example, in the U.S. Congress, specialized committees have disproportionate power over leg- islative proposals. In the European Union, the European Commission holds the right as the sole initiator of legislation.  In the case of the Philippines public procurement reform, documented in greater detail in appendix C, the ability to initiate a new bill rested with specific committees in the legislature. In 1998, during the early days of re- form attempts, when President Joseph Estrada was facing corruption al- legations and had incentives to give legislative priority to a “cleanâ€? public procurement bill, his office needed to identify a member of Congress who was willing to sponsor the proposal and who belonged to a committee that would be sympathetic to the reform. The Budget Department’s Legislative Liaison Office, keeping a low profile to avoid attracting open opposition to the reform attempt, found an influential member of the Committee of Public Works who was interested in the reformers’ proposal. Collective Choice and Agenda Setting 163  The congressman who decided to sponsor the bill had previously initi- ated a legislative proposal on sanctions for violations of government con- tracting regulations and agreed to substitute the procurement package for his original initiative. Only two weeks before a legislative recess, he man- aged to use his influence with the Rules Committee to introduce the bill for debate with unprecedented speed, despite not being scheduled, and to bring it to the floor when most opposing legislators were absent and a neutral quorum was met. • Voting rules. Under majority voting, a strong agenda setter can structure the decision process in a way that ensures a stable equilibrium close to its de- sired policy. Many times, in both parliaments and organizations, the choice of the voting rule can have profound implications for the final policy out- come. We have already suggested in previous chapters that the electoral rules through which political representatives are selected (on national party lists or through a “winner-take-allâ€? system at the district level) may give politicians different incentives to offer clientelistic or universalistic public goods to their constituents. Similarly, legislatures, ­ political party congresses, caucuses, organizations, or coalitions often ­employ various vot- ing methods that count and aggregate individual preferences differently.  In one-vote systems, all members of an organization pick only one among all possible options. In our bridge construction example, the vot- ers of District 1 choose option x. In ranked or preferential systems, voters get to rank their options and assign points according to the rank. Accord- ing to one such rule (Borda counting), District 2 voters will give a score of 3 to y, a score of 2 to z, and a score of 1 to x. In the end, the sum of total votes obtained by all alternatives decides the winner.  In rated voting systems, voters can assign any grades or points within a range (say, 0 to 10) for all three options (District 3 voters could assign a score of 10 to z, 7 points to x, and 1 point to y). In our case, a Borda voting system will give the same number of 6 points to all options. In other cases, however, because ranked or rated voting systems are more sensitive to the averages of preferences across districts, they might lead to the selection of broadly acceptable policy options as opposed to those simply preferred by the majority. If agenda setters can select organizational voting rules, they are likely to influence the final outcome and either empower or bypass majorities. The interactive exercise at the end of this chapter will show how the voting sequence, or the order in which these three bridge con- struction site alternatives are submitted for majority decisions, can render a winner desired by the agenda setter. • Amendment rules. Sometimes, amendment rules require just 50 percent plus one vote; other times, it is easier to accept an agenda setter’s full bill 164 Understanding Policy Change proposal than it is to amend it. The European Commission is a case in point. Because of the unanimity rule governing its amendment powers, it is virtually impossible to amend its proposals.  In many legislatures around the world, committee rules with respect to who can propose, how, or in what shape or form amendments can go to the floor of the assembly for debate are numerous and granular. Their level of detail and technicality is often discouragingly complex for anyone not intimately acquainted with the legislative process. The standing U.S. Senate Manual of rules, orders, laws, and resolutions for the 112th Con- gress is a huge tome of no less than 1,429 pages detailing the minutiae of senators’ work, protocols, and procedures. As we will see in the next sec- tion of this chapter, in some legislative assemblies, even small changes in rules affecting the order of amendments to bills, or the form in which the amendment is proposed (for example, if it is to be included in the original text of the bill or kept separate) can lead to policy outcomes that are quite different from what the majority really prefers.  For reformers, this factor can be both bad and good news. On the one hand, the high level of complexity makes the legislative process intimi- dating and difficult to navigate. On the other hand, the devil is in the de- tails. Sometimes, paradoxically, key amendments to preexisting laws and regulations or consolidation of scattered regulations might be more im- portant for providing quick gains to reformers than submitting for debate an entirely new bill that would meet open opposition.  In 2001, in Kenya, during the last days of the Moi administration, the public procurement system was plagued by severe corruption and colo- nized by entrenched political patronage networks. A reform team decided to tackle the challenge. The reformers’ initial instinct was to prepare a new procurement act, make the attempt politically salient, and garner public support for reform. They were instead advised to take advantage of amendment procedures to the preexisting Exchequer and Audit Act rather than making their efforts public and attracting intense opposition from vested interests. As a result, as part of an omnibus act clarifying the implementation of the Exchequer and Audit Act, the Kenyan legislature passed a new set of public procurement regulations that unified all the disparate acts governing procurement and created key institutions; thus, the reform team bypassed the fierce opposition it would have generated by passing a new law.  The Philippines public procurement reform features a similar trajec- tory. In 2001, the reform team took a step-by-step approach and first per- suaded President Arroyo to pass an executive order that consolidated all previous procurement regulations, despite the fact that this act did not Collective Choice and Agenda Setting 165 allow amendments to the preexisting laws. This first step created a sense of policy ownership in the executive’s office. In a second stage, the reform team garnered public support through the activities of a nongovernmen- tal organization, Procurement Watch, and promoted the bill as a broader omnibus package designed to ensure transparency and not just a narrow change in the workings of public procurement. Thus, by building support incrementally, the reformers were finally able to gain sufficient leverage in both chambers of the legislature to ensure passage (see appendix C). • Heterogeneity of actors’ preferences. Often, the policy arena has a complex web of institutions, each having its own voting rules. In some legislative bodies, if the two chambers that need to jointly pass a bill have directly antagonistic interests (say, the majorities are held by two opposing politi- cal parties), the agenda setter (as the initiator of legislation) will refrain from even formulating the proposal in the first place. In the case of the Philippines procurement bill, the supporters’ first attempt at passage in 1998 was obstructed because, while it passed in the lower house, the Sen- ate had different political priorities and did not consider the bill. Whereas the congressman who sponsored the proposal in the lower house was a representative of the opposition and tried to distance himself from a pres- ident facing a corruption scandal, the chairman of the relevant Senate committee was closely associated with the president, did not have incen- tives to promote the bill, and strategically blocked it. Passage was finally assured after an election, which produced a more balanced distribution of party seats and a unified opposition party presence in both chambers. Subsequently, this new configuration allowed the reformers to identify a small number of pivotal supporters, from both the majority and opposi- tion parties, to champion the bill, thereby allowing the media to portray its passage as a bipartisan effort (see appendix C). • The impatience (or time horizon) of the various actors to change the status- quo. In many legislative settings, for example, as long as the agenda set- ter’s proposal is better than the status quo, even if far from other actors’ ideal or preferred policy, the bill could still pass. However, if the latter anticipate that the agenda setter is impatient to push the legislation (or has a short time horizon because of reputational or electoral risks), they can engage in a series of proposal rejections, until the agenda setter brings the policy closer to their preferences.  This type of strategic voting caused by different time horizons reduces the power of the agenda setter (Pollack 1997). In contrast, if the other decision makers are impatient, the influence of the agenda setter in- creases significantly (Shepsle 1989; Pollack 1997, 124). The same logic ap- 166 Understanding Policy Change plies to many legislative agenda-setting processes as well. To go back to the case of the Philippines procurement bill, the need for President Es- trada, who faced a corruption scandal and the possibility of impeach- ment, to pass a law that tried to minimize corruption in public procure- ment became attractive five months before an election. As a result, the president became a key supporter of the bill in order to boost his reelec- tion prospects and thus increased the agenda setter’s bargaining position. Unfortunately, by the time the bill had reached the Senate, many of its members were busy trying to bolster their own reelection prospects and were able to outlast the president (see appendix C). How Do Formal Rules Influence Outcomes? As a sophisticated setting of agenda control, the U.S. House of Representa- tives has a vast range of rules for the order, content, and form in which amend- ments to legislative bills can be debated by the majority of members. This institutional setting allows the powerful Committee on Rules to “set the agendaâ€? either by eliminating certain alternatives right from the beginning or by ordering their vote in a way that produces a result different from what another ordering alternative would have produced, even under conditions of sophisticated voting. Order or sequencing of voting can alter the final outcome dramatically. The results change for alternative amendment voting agendas: for example, given a set of three proposals (x, y, and z), an agenda setter decides on pitting x against y, the winner of which is pitted against z out of all three sequencing possibilities. The two discarded sequences are x being pitted against z, the winner of which is pitted against y; and z being pitted against y, the winner of which is pitted against x. Below are some examples of rules governing the content of policy propos- als that can yield different policy results (Patty and Penn 2008). • “Closedâ€? or restrictive voting rules in some legislatures (including the U.S. Congress) allow only certain policy amendments to reach the floor for debate. • “Openâ€? rules, in contrast, allow any germane legislative amendments to be debated, thus aligning more closely with the majority will. • Ordered open rules, as a subset of the previous category that combines effects of both content and order, prespecify the order in which these ­ amendments are considered (Patty and Penn 2008; Krehbiel 1991; Olezsek 2007). Collective Choice and Agenda Setting 167 However, even if some rules are more open to policy alternatives than oth- ers, they still allow ample room for strategic manipulation by agenda setters. This aspect raises important questions about the design of collective choice institutions, as well as about their outcomes (Patty and Penn 2008, 20; Dum- mett 1984). For example, in the case of open rules, even if all amendments are sub- mitted for debate, the form of submission will still alter the result. In some cases, the committee on rules will consider amendments to a specific piece of legislation that would first be fully incorporated or printed in the original bill and sent out for debate as a substitute bill, instead of simply being submitted as disparate individual amendments. However, this ver- sion of the amended piece of legislation is also pitted against the un- amended original bill when voted on the floor. What is the reason for the inclusion of the original bill as well? Despite the seeming redundancy, studies show that in such cases, the outcomes differ fundamentally, de- pending on whether the original bill is included or not. When the original bill is included, the floor is more likely to choose the substitute bill as is. In the other case, the substitute bill will be chosen only after being amended on the floor by the majority (Patty and Penn 2008; Bach and Smith 1988). By anticipating the voting outcome, the strategic agenda setter (in this case, the Committee on Rules in the U.S. House of Representatives) is thus able to deliberately select specific voting rules that would bring the policy content closer to its ideal point.5 Informal Agenda Setting While formal agenda influence is conferred by clear and codified decision- making rules, for informal agenda control, the agenda setter must be a skillful policy entrepreneur who, under conditions of imperfect information6 and legislative impasse or polarization, can construct “focal pointsâ€? for bargain- ing and bypassing obstacles. Rather than deciding on voting rules as in the case of formal agenda setting, the entrepreneur just offers an idea that can rescue the bargaining process when there is no other equilibrium (Garrett and Weingast 1993; Pollack 1997, 125). Informal agenda-setting power requires policy creativity. The interest- ing thing about it is that many actors (lobbyists, legislative committees, ­ influential individuals, media representatives, and civil society organiza- tions, among others) can be policy entrepreneurs, unlike the formal agenda setter, whose role is clearly inscribed in institutional codes. 168 Understanding Policy Change Regardless of whether agenda setting is formal or informal in nature, the tools of agenda manipulation are often similar (Riker 1981, 1986): • Heresthetics can broadly be defined as the attempt by an individual policy entrepreneur or a cohesive group to try and manipulate the con- text or structure of a decision-making process to ensure a more favor- able outcome. • Rhetoric is the art of using language effectively to inform, persuade, or manipulate an audience. • Agenda control means using formal legislative rules of proposal and amendment to obtain favorable voting outcomes. • Strategic voting means using voting procedures to control outcomes. • Manipulation of dimensions involves redefining a situation to create a stronger coalition. Informal agenda-setting power also closely relates to the concept of lead- ership, as individuals who can persuade supporters of a reform to form a vi- able coalition or can use their influence at strategic moments can be critical to achieving policy goals. Some studies have identified three main factors that, combined, can generate a “policy windowâ€? for an innovative entrepre- neurial idea that could serve as a key bargaining point (Kingdon 1984, 165– 167; Pollack 1997): • The identification of the problem • The proposal of feasible and acceptable policy alternatives • Political changes (alternations of political parties in the legislature, up- coming elections, and the like). The confluence of these three simultaneous processes could possibly lead to a viable alternative to policy deadlock. What Are the Characteristics of a “Goodâ€? Formal and Informal Agenda Setter? The formal agenda setter has to have a thorough understanding of the rules, the anticipated policy outcomes, and the strategic behavior of the major ac- tors involved in the decision-making process. However, institutional power is often already assigned to the agenda-setting bodies and recognized by all participants from the very beginning. Concepts in Practice 5.2 offers an ex- ample of the power of formal agenda setting and the way in which it was used in Latin America to align development policy outcomes with presiden- tial mandates. Collective Choice and Agenda Setting 169 CONCEPTS IN PRACTICE 5.2 The Power of Formal Agenda Setting: Using the Presidential Veto to Make Policy in Uruguay Traditionally, the ability of the president of Uru- amended version of the law, which can then be guay to influence lawmaking has been consid- approved by a simple majority. Therefore, if ered weak. While the president can veto a law Congress is better off accepting the president’s passed by Congress, it is possible for the legis- amendment by a simple majority rather than lature to override such a veto, albeit with a overriding the veto, which requires a three- three-fifths majority. Therefore, the conven- fifths majority, the president can influence tional wisdom was that, when there was con- ­ policy outcomes through her agenda-setting sensus, Congress would be able to get its way power. because the president’s ability to stop legisla- The agenda-setting power of the president tion could be circumscribed. of Uruguay was clearly evident in the passage However, this understanding of the relative of the Five Year Budget Bill (2000–2004), which power of the president and Congress does not saw the president significantly modify the act take into account the agenda-setting power of with respect to tax policy, salaries, the school the former. The president of Uruguay is not only curriculum, benefits, and funding. As Congress able to veto legislation, but also to make ‘amen- faced a 30-day deadline to pass the new bill and datory observations’—that is changes to a bill a supermajority is required to override the presi- after it has been approved by Congress. The dent’s proposals, support to drop the presi- implications of this amendatory power are that, dent’s proposals was sufficient only to elimi- in effect, the president has conditional agenda- nate 6 of a total of 34 amendments. In short, by setting power that enables him or her to alter setting the agenda, the president was able to final legislation. The president effectively sets significantly modify the final outcome of the the agenda by proposing to Congress an policy-making process. Source: Tsebelis and Aléman (2005). Unlike the formal agenda setter, the informal agenda setter requires less clear and more eclectic skills, and his or her influence or power may vary. Studies have attempted to identify a set of basic features: the person usually possesses considerable authority as a policy expert or negotiator, patience and other skills necessary for navigating a complex institutional landscape (Kingdon, 1984, 188; Pollack 1997, 126). Influential policy networks of techni- cal experts, coupled with incomplete information available to the political representatives, have often led to legislation proposed and passed without any amendment from legislators (see the Philippines case study on procure- ment in appendix C for further elaboration). 170 Understanding Policy Change As Concepts in Practice 5.3 will now show with respect to the election of Abraham Lincoln in 1860, the roles of informal agenda setting and issue ma- nipulation have long been effectively used by political entrepreneurs to fur- ther policy goals. CONCEPTS IN PRACTICE 5.3 The Power of Informal Agenda Setting: Using the Issue of Slavery to Get Elected Sometimes being able to identify an issue that to power? Abraham Lincoln was able to use the resonates with voters but is neglected by the issue of slavery (agenda setting) to divide the establishment can help propel a seemingly mar- Democrats’ coalition. Following the Missouri ginal or extreme candidate into office. During Compromise of 1820, the issue of slavery had the 1850s, attempts to dislodge the Democratic become increasingly important to parts of the Party from power in the United States resulted electorate for a variety of reasons (for example, in failure because the party had solid electoral westward expansion, the rise of the abolitionist support from both northern farmers and cotton movement, and so forth). By emphasizing this planters in the South given its support for free issue during his election campaign, Lincoln was trade, making it easier for these groups to ex- able to create a new winning coalition that pit- port goods to Europe. As a result, industrialists ted northern farmers and industrialists against (in the North) were isolated in their support for southern planters. In this way, by setting the tariffs and hence protection from their Euro- electoral agenda, Lincoln was able to win the pean competitors. election of 1860. The U.S. Civil War of 1861–65 How were the Democrats dislodged and the resulted in the abolition of slavery across the Republicans, under Abraham Lincoln, brought United States. Source: Riker (1981). Agenda Setting as a Strategic Tool for Policy Making: A Note on Positivist and Normative Angles So far, we have argued that agenda setting intermediates the aggregation of individual preferences in collective choice, bypassing cycles. In addition, and perhaps more important from the point of view of the general develop- ment practitioner, agenda setting is a strategic tool for pursuing policy or po- litical goals. Normatively, agenda setting as a policy tool can be used for both “goodâ€? and “nefariousâ€? purposes. In some contexts, skillful manipulation of formal rules or the use of informal channels of political influence and communica- Collective Choice and Agenda Setting 171 tion can block any attempt to advance pro-development reform and act against positive change. As will now be seen in the case of election rigging in a post-Communist context, framing and manipulating the issue of the fairness of the electoral process, by preemptively making allegations about the motives of election monitors, can detract from the real question of whether elections were held in a fair manner in the first place (see Concepts in Practice 5.4). In other circumstances and institutional settings, a reform-minded agenda setter (for example, the head of the committee on rules or budgets in the national legislature) or a talented policy entrepreneur who manages to substantively influence the agenda at the right time can be crucial for pro- development policy change. The Philippines case study in appendix C will illustrate in greater detail the key importance of agenda setters for the overhaul of the corrupt public procurement system in the Philippines. Despite unfavorable conditions for the passage of a bill targeting the vested interests that had blocked previous attempts at public procurement reforms, change agents skillfully located and co-opted a potential agenda setter—the chairman of an influential com- CONCEPTS IN PRACTICE 5.4 Agenda Manipulation in Competitive Authoritarian Elections Nondemocratic regimes rely on a combination of coercion and persuasion to remain in power. In particular, competitive authoritarian regimes need to be- come adept at persuading the public that their rule is beneficial and better than any alternative the opposition can offer. For example, in a recent election, opposition groups were allowed to moni- tor voting irregularities and even publish their findings through the electronic media. However, in anticipation of the opposition’s possible allegations of ballot irregularities, rumors were circulated before the election that the opposition was planning to publish a list of alleged irregularities that it had already manu- factured. Thus, when the opposition parties promptly published their list of ir- regularities after the election, some media outlets questioned the speed and the authenticity of the findings. The regime managed to diminish the credibility of the opposition’s initial charges of voting irregularities and shifted the debate (agenda) toward the trustworthiness of the opposition rather than the tactics of the government. Source: Schatz (2008). 172 Understanding Policy Change mittee in the legislature (Committee for Public Works). Taking advantage of the absence on a particular voting day of members of Parliament opposed to reform, the agenda setter pushed the bill to the floor and managed to circum- vent entrenched interests while acting within the prerogatives and legal constraints of his position. Summary Policy outcomes are not always a reflection of the mobilization potential of large groups or majorities. Because of cycling, it is possible for strategically minded agenda setters, with either formal or informal power, to influence the decision-making process to further their agendas. Understanding how agenda setting works in a specific decision-making context is, therefore, crit- ical in knowing how and under what conditions certain reforms may be- come feasible. Considering the dynamics of agenda setting in the design of new institutions is undeniably important. Collective Choice and Agenda Setting 173 Exercises for Chapter 5: Collective Choice and Agenda Setting Exercise 5.1: The Agenda-Setting Game8 The facilitator explains the steps involved in the Agenda-Setting Game. Steps in the exercise (time frame: ~25 minutes): 1. Read out instructions (1 minute) 2. Allow participants to read and play game (10–15 minutes) 3. Class discussion (10 minutes) Preparation and materials: 1. Pens should be available to all participants 2. Copies of the Group Briefing handouts 3. Copies of the Head Official handout Procedure: 1. The facilitator explains: For the following activity, at each table, you are going to be voting officials from three different districts, and you will need to decide where to build a new bridge. One of you will be the head official responsible for facilitat- ing and recording the votes. Before I pass out the handout that includes the detailed briefing for the activity, please decide who the head official will be at each table. Once you do that, go around the table to indicate if you are in District A, B, or C. I will now go ahead and distribute the handouts. 2. The facilitator goes to each table and hands out Exercise 5.1, the Group Briefing handout, to all participants. In addition to this handout, the ­ person selected as head official also receives Exercise 5.1: Head Official Briefing handout. 3. The facilitator asks the participants to read the detailed briefing and in- vites them to ask questions about the task at hand, if they have any. 4. The facilitator gauges the rate of completion of the assignment and gives participants a 2-minute warning so that the activity can wrap up within its allotted 10-minute time frame. 5. The facilitator asks the head official from one of the tables to report out the result of Scenario 1, then asks another table to report out the result of Sce- nario 2, and a third table to report out the result of Scenario 3. Then the facilitator poses the following questions for an overall group discussion: • How did the head official’s agenda affect which outcome prevailed? • What, if anything, does this tell us about the role of institutions? 174 Understanding Policy Change After the class game, the facilitator explains the logic of agenda setting by providing the technical explanation for agenda setting (Duch 2009, 1). “How do committees decide? How do decision-making procedures influence outcomes? Assume a set of alternatives and a set of committee members. It’s sim- ple if there are two options. If there are three alternatives, x,y,z, we might imagine a process such as Vote x a Vote y z b c We describe processes such as these as binary agendas as at each stage of voting the choice is between two options. For example, at the first stage we consider whether we want to adopt x or not. The number of terminal histo- ries is at least the number of alternatives, and each alternative is associated with at least one terminal history.â€? (ibid, 2009, 1–2) Three alternatives: A, B, and C, and three voting blocs of equal size in the city council. The preferences for the bridge construction project are ranked as follows: Bloc 1 (33 percent): A > B > C Bloc 2 (33 percent): B > C > A Bloc 3 (33 percent): C > A > B Here are the diagrams for results depending on the voting sequence: 7 C Wins B Wins A Wins A B A C B C 1 + 3 = 66% 2 = 33% 1 = 33% 2 + 3 = 66% 1 + 2 = 66% 3 = 33% A C B C A B C B B A C A 1 = 33% 2 + 3 = 66% 3 = 33% 1 + 2 = 66% 2 = 33% 1 + 3 = 66% Collective Choice and Agenda Setting 175 Exercise 5.1: Group Briefing Handout You are officials in three different districts and must decide on where to build a new bridge. Decision making is by majority voting. That is, a decision can be made only if more than 50 percent of the officials vote for it. Further- more, decisions are made using pairwise comparisons: that is, any two alter- natives are put to the vote at any one time. There is a head official running this meeting who does not vote but sets the agenda. Each district represented has only one vote even if there are multiple officials from each district. Thus, for example, if District A has two people at your table and District B has three people, both districts still have only a single vote. You represent one of the three districts (Districts A, B, or C) and would like to see the bridge built in your district. Failing that, you would rather see that the bridge is built in the adjoining district to yours. This means that • District A officials prefer A>B>C • District B officials prefer B>C>A • District C officials prefer C>B>A. The head official conducted an initial vote, and based on the preferences of each district, no location obtained a majority as each location has an equal vote (33.3 percent support). For a proposal to be accepted and pass the 50 percent plus threshold, two of the three groups must actually vote in favor of the proposal. The head official at each table will announce a legislative agenda with three different scenarios. Each scenario will have two stages. In the first stage, all districts will have the opportunity to vote between two possible lo- cations. In the second stage, the winning location of the first stage will be up against the remaining location. 176 Understanding Policy Change Exercise 5.1: Head Official Handout 1. After all participants at the table have reviewed the Group Briefing handout, explain that there will be two rounds of voting based on three scenarios. 2. Explain that you will read out the voting scenario so that participants can decide how to vote. 3. Remind them that if there are multiple officials present from one district, that district still has only a single vote. 4. Record the first-round vote results. 5. Conduct the second-round vote. 6. Record which location wins for Scenario 1. Once this has been done for Scenario 1, repeat the process for using Scenarios 2 and 3. Scenario 1 First-Round Vote: Vote between District A and District B. What is the outcome of the first round?_____________________ Second-Round Vote: Vote between the [district that won in the first round vote] and District C. What is the final outcome? _____________________ Scenario 2 First-Round Vote: Vote between District A and District C. What is the outcome of the first round?_____________________ Second-Round Vote: Vote between the [district that won in the first round vote] and District B What is the final outcome?_____________________ Scenario 3 First-Round Vote: Vote between District C and District B. What is the outcome of the first round?_____________________ Second-Round Vote: Vote between the [district that won in the first round vote] and District A. What is the final outcome?____________________ Collective Choice and Agenda Setting 177 Notes 1. In fact, agenda manipulation can ensure that the “best collective outcomeâ€? does not become policy. 2. “I’ll let you write the substance and you let me write the procedure, and I’ll screw you every time.â€? See Regulatory Reform Act: Hearings on H.R. 2327 Before the Subcommittee on Administrative Law and Governmental Relations of the House Committee on the Judiciary, 98th Cong. 312 (1983) (Patty and Penn 2008, 2). 3. For purposes of clarity, a majoritarian system is one in which public decisions require the consent of at least 50 percent plus one of voters. 4. The name comes from Marquis de Condorcet, a French philosopher who, in 1785, wrote a treatise entitled Essay on the Application of Analysis to the Probabil- ity of Majority Decisions, where he defined the concept and explained for the first time the paradox of social choice and preference aggregation. 5. An “ideal pointâ€? refers to the policy desired by a certain actor, and any depar- tures from it decrease his or her utility or overall gain. 6. Defined as a situation in which one actor (the agenda setter) has information that other actors do not possess. 7. The game and diagrams illustrating the “paradox of votingâ€? are adapted from Poole (2011), http://voteview.com/paradox_of_voting.htm. 8. Source: (Duch 2009) and (Poole 2011) . References Arrow, Kenneth. 1951. Social Choice and Individual Values. New Haven: Yale University Press. Bach, S., and S. S. Smith. 1988. Managing Uncertainty in the House of Representatives: Adaptation and Innovation in Special Rules. Washington, DC: Brookings Institu- tion Press. Bachrach, Peter, and Morton Baratz. 1962. “The Two Faces of Power.â€? American Political Science Review 56 (4): 947–52. Duch, Raymond. 2009. “Formal Analysis Lecture Notes.â€? Nuffield College: Oxford, http://www.raymondduch.com/course/hilary2009/formalanalysis/ lecture_2009_06.pdf. Dummett, Michael. 1984. Voting Procedures. New York: Oxford University Press. Garrett, G., and B. Weingast. 1993. “Ideas, Interests, and Institutions.â€? In The Role of Ideas in Foreign Policy, ed. Judy Goldstein and Robert Keohane, 173–206. Ithaca: Cornell University Press. Johnson, Paul. 1998. Social Choice. Theory and Research. Thousand Oaks, CA: A Sage University Paper. Kingdon, J. W. 1984. Agendas, Alternatives, and Public Policies. London: Longman. Krehbiel, Keith. 1991. Information and Legislative Organization. Ann Arbor: University of Michigan Press. 178 Understanding Policy Change Oleszek, Walter. 2007. Congressional Procedures and the Policy Process. Washington, DC: CQ Press. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge: Harvard University Press. Ordeshook, P. C., and T. Schwartz. 1987. “Agendas and the Control of Political Outcomes.â€? American Political Science Review 81 (1): 179–99. Patty, J. W., and E. M. Penn. 2008. “Amendments, Covering, and Agenda Control: The Politics of Open Rules.â€? Manuscript. Washington University, St. Louis. Poole, Keith. 2011. The Paradox of Voting (The Condorcet Paradox). Houston: University of Houston, http://voteview.com/paradox_of_voting.htm. Pollack, Mark. 1997. “Delegation, Agency, and Agenda Setting in the European Community.â€? International Organization 51 (1): 99–134. Riker, William. 1981. Liberalism against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice. Long Grove, IL: Waveland Press. ———. 1986. The Art of Political Manipulation. New Haven: Yale University Press. Schatz, Edward. 2008. “Transnational Image Making and Soft Authoritarian Kazakhstan.â€? Slavic Review 67 (1): 50–62. Shepsle, Kenneth. 1989. “The Changing Textbook Congress.â€? In Can the Government Govern? ed. John Chubb and Paul Peterson, 355–68. Washington, DC: Brookings Institution. Shepsle, Kenneth, and Barry Weingast. 1984. “Uncovered Sets and Sophisticated Voting Outcomes with Implications for Agenda Institutions.â€? American Journal of Political Science 28 (1): 49–74. ———. 1987. “The Institutional Foundations of Committee Power.â€? American Political Science Review 81 (1): 85–194. Stratmann, Thomas, and Martin Baur. 2002. “Plurality Rule, Proportional Represen- tation, and the German Bundestag: How Incentives to Pork-Barrel Differ across Electoral Systems.â€? American Journal of Political Science 46 (3): 506–14. Tsebelis, G. 1994. “The Power of the European Parliament as a Conditional Agenda Setter.â€? American Political Science Review 88 (1): 128–42. ———. 1997. Bicameralism. Cambridge: Cambridge University Press. Tsebelis George, and Eduardo Aléman. 2005. “Presidential Conditional Agenda Setting in Latin America.â€? World Politics 57 (3): 396–420. Collective Choice and Agenda Setting 179 CHAPTER 6 Delegation Problems and the Principal-Agent Relationship Accountability implies a well-functioning relationship of delegation among citizens, politicians, and bureaucrats. Voters elect representatives and ex- pect them to fulfill the policy mandate for which they were elected. Politi- cians enact laws and regulations but leave the implementation to bureau­ cracies. As chapter 2 suggested, mismatched incentives of these actors in fulfilling their tasks on any leg of the accountability triangle generate delega- tion problems. In turn, the noncompliance of delegates with the tasks of their mandate is consequential for public good delivery. Whenever politi- cians or bureaucrats shirk their duties, without adequate monitoring and penalties, corruption and poor governance thrive. More pragmatically, to achieve an outcome—say, the realization of a devel- opment project or a reform target—the actors concerned with reaching this goal, known as principals, must often delegate tasks to those who can help implement the objective, referred to as the agents. Given the need for delega- tion and the fact that the interests of principals and agents do not always fully converge, a host of “principal-agent problemsâ€? arise, particularly the possibil- ity that agents will shirk their duties. For example, bureaucratic and political agents do not always have incentives to allow their principals to scrutinize their use of taxpayers’ money, even though the country may have passed free- 181 dom of information legislation that gives citizens the right to do so. As a re- sult, some agents actively obstruct the effective implementation of these laws. In general, the principal-agent problem, as the key analytical backbone of delegation, is very helpful in understanding a wide variety of relationships, including those between legislators (the principals) and bureaucrats (the agents); voters (as principals) and politicians (as agents); members of civil society or political party organizations and their leaders; employers and em- ployees. Delegation and principal-agent relations are at the heart of political, bureaucratic, and social accountability. The position in the delegation chain and the scope of information about the task to be performed generate the sets of incentives for actors. But as we all know, delegation processes occur in all sorts of organizations. Political parties, civil society groups, or coalition leaders often entrust member recruitment, treasury tasks, or organizational expansion to agents. In these cases, shirking has negative consequences for the collective action potential of these groups and acts as a constraint on joint efforts to achieve policy change. This chapter introduces readers to the challenge of delegation, to issues surrounding the principal-agent relationship, and to the different types of institutionally induced incentives (sanctioning and monitoring) that can sometimes mitigate conflict of interests between the two (see figure 6.1). Therefore, the chapter provides a basis for diagnosing the delegation chal- lenges inherent in many projects and public policy-making contexts. The central dilemma that this chapter addresses is how to get the politi- cian or bureaucrat (the agent) to act in the best interests of voters and citi- zens (the principal), when the former has an informational advantage and interests that diverge from those of the latter. FIGURE 6.1  Delegation and Principal-Agent Conceptual Map Principal-agent Principal-agent Solutions to model problems principal-agent problems • Principal-agent • Why do • About control relationship principal-agent schemes: • Delegation problems ex ante and occur? ex post checks • Who can be • Unintended affected by consequences these of principal- problems? agent solutions Source: Authors. 182 Understanding Policy Change Objectives of Chapter 6 By the end of this chapter, readers should be able to do the following: • Understand the reasons for delegation, along with its potential benefits and pitfalls • Describe various strategies for monitoring and sanctioning agents • Recognize how the competing agendas of principals and agents may make one or the other more receptive to reform goals • Learn to explore analytically the relationship among political institutions, collective action, and principal-agent problems and solutions. Delegation Principal-agent relationships are inextricably related to the concept of dele- gation: that is, the resolution of a boss (the principal) to transfer consulting, decision, or implementation power to a worker (the agent). Given private in- formation that favors agents and gives them incentives to shirk their duties on the mandate given by the principal, political economists have attempted to find answers to several theoretical questions. Why delegate in the first place? Why Do Bosses Willfully Give Up Control and Delegate? The classic answer of delegation focuses on knowledge discrepancies be- tween the principal and the agent. It is argued that actors who are more knowledgeable about the operational context will make better choices. More recent theoretical accounts focus less on the better outcome or better choice and more on how delegation can make “the outcomes less riskyâ€? (Bendor and Meirowitz, 2004, 294). The information and expertise about specific policies that the agent possesses or acquires can hedge against the risk of random shocks that could prevent the busy or less informed principal from achieving her target. Under conditions of uncertainty over how exactly policies will translate into outcomes, for instance, principals are likely to delegate tasks to agents, irrespective of how risk averse or risk prone they are. As the distance between the outcome uncertainty known by the bureaucracy, on the one hand, and by politicians, on the other, increases, delegation is more likely, and the mandate will be broader (Epstein and O’Halloran 1994, 1999; Bawn 1995). The risk rationale is central to the relationship between politicians as princi- pals and bureaucracies as agents (Bendor and Meirowitz 2004, 294). Uncer- tainty over their own electoral fortune also makes incumbent political princi- Delegation Problems and the Principal-Agent Relationship 183 pals more likely to protect their bureaucratic agents from future interference. Thus, political uncertainty is also crucial to the nature and scope of the dele- gation mandate (Huber and Shipan 2006, 269) (Moe 1989). Given that the delegation problems are inherent in a principal-agent re- lationship and occur when agents have an informational advantage over their boss, it is important to ask, Who will the principal trust enough to have as his or her agent? Bureaucrats know more about policy implemen- tation than politicians. From the point of view of the executive and the in- cumbent political party, however, irrespective of their technical expertise and competence, not all bureaucrats are equally trustworthy. In empirical tests, given equally qualified agents, principals will choose the one that is the closest ideologically to them because principals care not only about “objectiveâ€? competency but also about the realization of their own elec- toral agenda (the “ally principleâ€?) (Huber and Shipan 2006). The mandates of agencies (that is, bureaucratic units) often depend on the partisan game. The delegation mandate is likely to be narrower and leave less policy dis- cretion to the agents in cases in which the policy preferences of politicians and bureaucrats do not converge. In situations of high policy conflict be- tween the legislature and the executive and divided government, the del- egation mandate contains less bureaucratic autonomy and less policy dis- cretion, as politicians fear divergence. Now, let us see in more detail how and why delegation happens. How Does the Process of Delegation Occur? Once principals have selected an agent, they also have to choose both the degree of discretion (how broad or narrow and how generic or specific the agency instructions should be) and the estimate of the probability that the agent will shirk and get caught. The precision with which the principal de- fines the tasks to be completed by the agent establishes the limits of action or the scope of the delegation mandate. There are many cross-national studies of the principal-agent problem and the degree of policy discretion given to agents (measured as the level of de- ­ ureaucratic tail in statutes or laws defining the authority and jurisdiction of b agencies). The main findings can be summarized as the following: • Presidential systems seem to be able to control bureaucracies better than parliamentary systems. • Agents will have more limited discretion in an environment character- ized by high policy conflict, high legislative capacity, and no bicameral 184 Understanding Policy Change polarization, as well as when the legislative majority cannot rely on non- statutory factors. • Coalition and minority governments are more likely to write policy details into statutes, thus restricting the agent’s discretion. Anticipating fragility (short time in office), they have incentives to bring the policy preferences of the bureaucrats closer to their political agenda (Huber and Shipan 2002). In addition to looking at how much policy discretion the agent has, some studies also classify types of delegation according to other criteria (Strøm, Muller and Bergman 2003, 64–66): • Direct versus indirect delegation. In parliamentary regimes, for example, the voters elect legislatures, who then elect the prime minister and the cabinet, who further delegate authority to bureaucratic agencies (a form of indirect delegation). In presidential regimes, both the president and the legislatures are directly elected (direct delegation). • Singularity or plurality of agents and principals. In parliamentary institu- tions, each agent (implementation unit) corresponds to one principal (the specific cabinet minister directly accountable to the Prime Minister who is accountable to the legislature), whereas in presidential regimes, ­ ureaucratic agent has at least two principals (the legislature and the the b president). • Competition between different agents’ delegation mandates. To use the same example of the two distinct institutional settings, in parliamentary contexts, there is no overlap between agents’ jurisdictions: the ministry of education and the ministry of health do not directly monitor or compete with each other; in contrast, in presidential contexts, some bureaucratic agencies place bids for policy implementation whose winner is selected by the legislature and/or presidency; the system of checks and balances also allows agents to crosscheck mandates. Sometimes, particularly in settings with strong authoritarian legacies, delegation is rare altogether, because top-level politicians usually make pol- icy decisions and are not always comfortable with bureaucratic discretion or autonomy. As Concepts in Practice 6.1 shows, administrative reform in Iraq, a country with a top-down chain of command, entailed developing a middle tier of bureaucrats who could perform basic tasks such as data collection. Acquiring a grasp on when, how, and why delegation occurs in organiza- tions and institutions of political representation is an essential step in under- standing why problems arise. Now, let’s go up a level of abstraction. The fol- lowing section will introduce in more detail the principal-agent theory and Delegation Problems and the Principal-Agent Relationship 185 CONCEPTS IN PRACTICE 6.1 Solving Delegation Problems and Data Collection in the Iraqi Pension System In 2009, the government of Iraq requested the new pension team (with a mix of members World Bank’s assistance in implementing a uni- from different offices) that had the task of fied pension law. The law aimed to increase ­ collecting as much relevant pension data as the number of Iraqis contributing to and bene- possible in a short time. As a result, the team fiting from the social protection programs and was able to come up with new managerial and creating enrollee contribution systems. How- data-gathering mechanisms that increased effi- ever, the World Bank and the Iraqi National ciency. Besides compiling data on beneficiaries Board of Pensions quickly realized that, to of social protection, they also decided to renew, achieve successful implementation, the team simplify, and standardize the documentation re- and the organization itself would have to ad- quirements of the National Board of Pensions, dress serious delegation challenges. A top- making the rules clear to all. Communication down culture in the bureaucracy was prevalent, was also improved between pension officers top bureaucrats were reluctant to delegate and teams. Moreover, more training helped tasks, and authority was highly concentrated. team members better understand and use new Teams and staff worked in silos, and collabora- technology and innovative methods to increase tion was rare. Procedures in place were old and their confidence when executing tasks. Overall, inefficient, and managers did not receive con- these solutions successfully addressed delega- stant training. tion problems, created a managerial middle To address data collection by first address- level that was previously missing, and boosted ing delegation problems, reformers created a the organization’s performance. Source: World Bank (2011). discuss the actors’ incentives in play. Having them in mind can help readers think of potential mechanisms for solving this type of constraint. The Principal, the Agent, and Their Incentives Technically, the principal-agent relationship is built on two major assump- tions: The boss wants an agent who is capable and willing to perform his tasks well - - the voter prefers a competent to an incompetent legislator; and second, the contract is feasible precisely because both parties prefer the su- perior option of delegating and working to one in which the boss retains full control of all tasks and the delegate shirks his duties (Bendor, Glazer, and Hammond 2001, 237–238). Democracy and elections, for example, function under the assumption that the interests of the citizenry, bureaucracy, and 186 Understanding Policy Change politicians are fully aligned However, two sets of problems—agency losses and slippage—can emerge (Pollack 1997, 108). The first problem—agency loss—is inherent in any delegation relation- ship because principals and agents often have conflicting interests (Kiewiet and McCubbins 1991, 5). For example, voters as principals would like to see their preferred policies enacted by politicians as agents, but the latter also face incentives to shirk and opportunistically maximize rents instead of pub- lic goods in the absence of constraints designed to keep them in line with their mandate. The same logic applies to most organizational dynamics be- tween rank and file members and leaders. Moreover, beyond classic shirking, a second dynamic called slippage is in- duced by the institutional setup and structure of delegation (Pollack 1997, 108). Slippage often occurs when independently of the agent’s propensity to shirk or comply, the black box of the decision making process within the agency, or the informal norms of policy implementation, lead to outcomes that are not aligned with the principal’s mandate. What Are the Main Technical Features of a Principal-Agent Model? A principal-agent model is essentially about how the different tasks, attri- butes, and roles of the principal and his or her agents shape delegation out- comes. Thus, the principal-agent dynamic involves at least two actors and has several standard features (or core assumptions) (this section builds heavily on Miller 2005, 205–06; Holmstrom 1979; Shavell 1979): • Information asymmetry. This refers to any situation in which one actor has more information than another (the concept will be discussed in de- tail in chapter 7). Gathering complete information is expensive for the principal, one of the reasons for which delegation occurred in the first place. Legislators do not have the time and resources to actually monitor the daily routine of bureaucrats to make sure that they comply with their mandate. Even if at the end of the fiscal year, specialized legislative com- mittees will assess the policy outcomes of an implementation agency, they will still not be able to clearly understand the major causes of un- derperformance: is it because of the economic context (for example, a crisis), the lack of coordination among various agencies, or a bureaucrat’s inadequate effort? • Risk asymmetry. When the agent’s preferences and risks are not aligned with those of the principal (for example, the agent may be more risk Delegation Problems and the Principal-Agent Relationship 187 averse than the principal), the agent has a propensity to shirk his or her duties (see the Principal-Agent Game at the end of this chapter). • The “unifiedâ€? principal initiates the principal-agent relationship, by dele- gating tasks to the agent. This means that even in cases of multiple princi- pals (several institutional veto points such as parliament, executive, and bicameral legislature, as opposed to just one legislative chamber, for ex- ample), the “unifiedâ€? principal (that is, all the institutional veto points jointly) acts based on an aggregated set of preferences and decides to delegate tasks to the bureaucracy. Individual principals have a mecha- ­ nism for agreeing on a unified strategy or on the process by which one or all of them will come to a decision. • Backward induction and common knowledge. In many situations, the prin- cipal-agent dynamic has a sequential nature. Usually, the principal selects the agent (stage 1); the agent performs an action (stage 2); the principal observes and reacts to the final outcome (stage 3). It is usually assumed that both the principal and the agent know the structure of the game (stages), the payoffs, and the probability of outcomes; and the principal can see the agent’s best response function down a “game treeâ€?—that is, how a typical agent will react if the incentives are changed. • Ultimatum bargaining. The principal makes a “take it or leave itâ€? offer to the agent in delegating power to him or her, or not (Miller 2005; Sapping- ton 1991). If the principal can correctly calibrate her offer, she may be able to ensure that the agent has an incentive to deviate only minimally from the principal’s optimum preferences. What Are the Results of the Principal-Agent Dynamic? If the probability of detecting the agent’s shirking is low enough through the monitoring process and the sanctions attached to shirking are insignificant, the principal will choose not to delegate at all, and the agent will shirk if del- egation occurs. However, if delegation does take place, the principal and the agent will bear certain consequences and tradeoffs in terms of risk transfer and efficiency—both the principal and the agent may be overwhelmed with their tasks. The Problem of Multiple Chains of Principals and Agents. In an analysis principal-agent relationship in a particular context, it is also useful to of a ­ think of it as embedded in a broader ecosystem of principal-agent transac- tions. For example, the legislature is the principal and the revenue adminis- tration is the agent in one such mechanism, given that politicians specify the delegation mandate of bureaucratic agencies. 188 Understanding Policy Change However, the legislators are elected by voters, who confer a certain elec- toral mandate on them (the platform on which candidates ran) and thus cre- ate another principal-agent relationship. If outcome-based incentives such as performance bonuses for tax collectors (the solution to the first principal- agent relationship) lead to coercive and abusive taxation, the negative out- come is transferred to the second principal-agent transaction: the voters will attempt to stop such abuse by formulating claims to their elected representa- tives, who, in a circular turn, will attempt to shift the incentives of the tax collection agency to limit abuse (again, the first principal-agent transaction). Thinking about the whole chain of principal-agent transactions, rather than analyzing them in isolation, can lead to a more sophisticated design of outcome-based incentives and monitoring techniques, as well as promoting ­ greater accountability of politicians, bureaucrats, and voters. Who or What Will Make Sure That the Agent Complies with the Mandate? Solving Delegation Problems The classic solutions to principal-agent problems are related to incentives given to the agent in exchange for committing to perform the tasks with due diligence, as well as to different types of monitoring or control mechanisms to detect the agent’s deviations from the mandate conferred upon her by the principal. Because these are broad solutions that apply to the whole range of delegation issues, we will also briefly mention them in the next chapter, which explores information asymmetries as a subtype of larger principal- agent problems. Incentives to Achieve the Outcome Desired by the Principal Outcome-based incentives are often used to overcome the problems inher- ent in a principal-agent relationship. For example, performance bonuses for bureaucrats and high sanctions for low effort are common techniques that encourage compliance with the agent’s mandate. In some countries, for ex- ample, tax officers receive individual bonuses depending on the total reve- nue that they manage to collect. However, while outcome-based incentives are central to the principal-agent relationship, they can also generate unin- tended consequences. As in the example of tax or customs officers, the in- centive of individual performance bonuses can lead to individual opportun- ism (abusive extraction) as well as to a lack of coordination among the many agents (tax collectors), with overall negative consequences for revenue col- lection and perceptions of procedural fairness.1 Delegation Problems and the Principal-Agent Relationship 189 Monitoring and Control Schemes The principal can employ several different control schemes to prevent the agent’s shirking. However, one should keep in mind that despite all precau- tions, an agent might at times choose to disobey and avoid delegation prob- lems (Bendor and Meirowitz 2004, 302). Having said that, the control mecha- nisms can help minimize incentives for the agent to disobey the principal and thus are considered potential solutions for delegation problems that typically arise in development. Depending on the sequencing of monitoring and sanctioning, there are two types of checks: ex ante checks, which are forms of controls put in place before the actual principal-agent transaction takes place, and ex post checks, which are put in place after the transaction. Examples of ex ante checks include • Contracts. Detailed contracts that specify the aligned goals of principals and agents. • Recruitment screening and selection. For example, electoral candidates may have to prove themselves first in the political party hierarchy; simi- larly, the head of a bureaucratic agency is often a civil servant with a long career in the bureaucracy (Kiewiet and McCubbins 1991, 27–31; Strøm 2000; Strøm, Muller, and Bergman 2003). Ex post control schemes. Oversight is legislators’ usual instrument for con- trolling bureaucratic information. According to the probability of detection embedded within the principal-agent relationship, oversight mechanisms can take two forms: (1) so-called “police patrols,â€? in which the cost is high since it requires the principal to perform routine or random observations (monitoring) of the principal-agent relationship; and (2) “fire alarms,â€? in which the mechanism of accountability implies that informed second par- ties (that is, oversight bureaucrats) or informed third parties (policy con- sumers or groups of citizens) send signals to the principals about the devia- tion of the agent (McCubbins, Noll, and Weingast 1987; Lupia and McCubbins 1994). As opposed to ex ante checks, ex post monitoring occurs either during the delegation process itself (as with the “fire alarmsâ€? and “police patrolsâ€? dis- cussed above) or at the end of it (through official reports and audits). Some authors argue that such controls are inferior to ex ante oversight mecha- nisms because some principals might gain from the agent’s noncompliance (for example, the political party of the executive, but not the majority party in the legislature in cases of divided government). Therefore, in some cases, 190 Understanding Policy Change the interested principals have incentives to allow shirking (McCubbins, Noll, and Weingast 1987, Bendor, Glazer, and Hammond 2001, 246). In addition to the mechanism of detection, the penalty for shirking is also consequential for the quality of the relationship. Studies have found that dele- gation will be equivalent to abdication (the principal simply gives up) when the penalties for lying and shirking are too weak and where agents (bureaucracies) perceive the ideological or policy compatibility and overlap between fire-alarm institutions and current legislators as weak, given that the cost of monitoring and sanctioning is high (Lupia and McCubbins 1994). An example of how some of these controls may help is presented below in Concepts in Practice 6.2. CONCEPTS IN PRACTICE 6.2 Can Politicians Directly Control Bureaucrats? Evidence from Tax Agency Reform in Argentina Reforming Argentina’s tax collection agency compliance. Therefore, they had to rely on was an important priority for the first Menem costly routine “police patrolâ€? checks on the tax administration (1989–95). Legislators in many agency. developing countries face significant hurdles in This complex context shaped the nature of controlling the bureaucracy. In this context, the policy change targeting the legislative oversight control over the tax agency was particularly lax, of the bureaucracy. Specifically, the tax author- and its losses were financed through loans. The ity was ultimately reformed to take into account attempts to correct agency losses and improve the fact that legislators could not rely on inter- the principal-agent relationship interacted with est groups to alert them to the actions of the political incentives: members of the parliament bureaucracy; instead, because “deck-stackingâ€? faced increased difficulties in maintaining their (or cumbersome evidentiary procedures im- offices because of political violence and the posed on the agency) was very expensive, overall instability of the democratic process. ­ politicians chose to focus on streamlining the The closed electoral lists rendered the politi- agency’s internal procedures, simplifying the cians’ careers dependent on party leaders, cul- organizational hierarchy, and reducing the gen- tivated strong party loyalty, and provided disin- eral organizational costs of oversight such as centives for particularistic benefits. In addition, auditing and hearings. Therefore, even though elected officials in Argentina operated within interest groups were less likely to be able to cohesive parties that were not responsive to incentivize individual legislators to alter the be- interest groups that could have acted as in- havior of bureaucrats, legislators could more formed and interested third parties activat- easily and directly monitor the activities of the ing “fire alarmsâ€? in cases of bureaucratic non- tax agency. Source: Eaton (2003). Delegation Problems and the Principal-Agent Relationship 191 Ex post mechanisms have two major categories of effects on the principal-agent problem that shape the scope of the mandate that the prin- cipal gives the agent (Huber and Shipan 2006): • The substitution effect. Strong ex post monitoring will give incentives to the politicians to write a more discretionary statute of delegation (Huber and Shipan 2002; Bendor and Meirowitz 2004). • Political Uncertainty. Uncertainty about the future preferences of the pol- iticians (due to changes in the legislature after elections) requires the right balance of ex ante and ex post auditing mechanisms so that politi- cians can ‘lock in’ their policy preferences. Generally, if current politi- cians fear that their successors will renege on their preferred policy, they will be more willing to grant bureaucrats less discretion when delegating (Moe 1989). It is also important to keep in mind that the type of monitor- ing tools chosen by the principal induces strategic behavior on the side of the agent (Huber and Shipan 2006, 260–269). In general, monitoring is less onerous when there is close ideological proximity between bureaucrats and governments. According to the ally principle—that is, the degree of policy conflict between politician and bu- reaucrats—if the policy-ideal points of the principal and those of the agent converge, delegation will be more extensive. Some studies also argue that monitoring effectiveness and the type of monitoring devices differ significantly between parliamentary and presi- dential regimes. In parliamentary regimes, prime ministers (acting as agents) are often screened intensely by the legislatures, and they have al- ready gained their reputation by climbing the incumbent party ranks. Therefore, strong ex ante checks are in place. In contrast, presidential re- gimes often bring in an agent (the president) who is an outsider and with whom the principals in the legislature might not be well acquainted. In presidential regimes, ex post checks (such as impeachment procedures or referendums) prevail as mechanisms of ensuring compliance with the mandate (Strøm, Muller, and Bergman 2003). As many of the chapters of this handbook emphasize, the slippage or agency problems among citizens, politicians, and bureaucrats do not de- pend exclusively on the quality of monitoring or the particular incentives adopted. Like everything else in the political-economy universe, principal- agent relations do not exist in an institutional void. The incentives to shirk or not to shirk also depend on a host of other contextual rules of the game that may completely alter the initial intention of policy designers. For in- 192 Understanding Policy Change stance, one of the most common recipes recommended by international donors for reforming corrupt and inefficient tax administrations was the creation of semiautonomous revenue agencies to replace the more conven- tional central tax offices. Not surprisingly, evaluations of the performance of such organizations across Latin America and Africa showed mixed re- sults: they were not a delegation panacea and worked as intended only in contexts characterized by strong political commitments (Minh Le 2007). Again, synergistic, complementary institutions matter for the final devel- opment result. Therefore, the practitioner should develop a holistic under- standing of the institutional web of the policy-making environment before acting. Finally, since collective action is at the heart of development and the guiding principle of our narrative, how do delegation and better bureau- cratic accountability relate to the ability of citizens to mobilize? What con- straints do principals or agents place on the space and capacity for the col- lective action of voters? Concepts in Practice in 6.3 suggests one of the many potential theoretical answers. Analyzing Principal-Agent Dilemmas Using Game Theory Some of the basic intuitions of the logic of delegation can be conceptualized through a simple game. Before a discussion of the details of the game, it is critical to introduce a new solution concept—backward induction. Backward induction is a method for solving games in which there is a limited number of moves and players make decisions sequentially (Shor 2005). First, one has to identify the optimal or best strategy of the last player, and then work ‘backwards’ determining the optimal strategy of the second last player, etc., until the optimal strategies of all players have been worked out (ibid).2 For example, let us look at a game in which an agent (who has already been appointed to a policy-making position by taxpayers) is assumed to be solely concerned with maximizing her private income. She has a choice over whether to use the position either to maximize her income by being corrupt (rent maximizing) or to minimize her income by being honest (rent minimizing). Given these preferences and choices, it might seem as if the agent will always choose to pursue a rent-maximizing course of action.3 Delegation Problems and the Principal-Agent Relationship 193 CONCEPTS IN PRACTICE 6.3 Principal-Agent and Collective Action If politicians determine the rules that govern to deliver goods to core voters (such parties are the civil service—for example, the level of bu- beholden to the demands of patrons for private reaucratic professionalism and the degree of goods). transparency in the budgetary process—then it In short, if voters can overcome collective becomes possible to understand under what action problems directly and hold political par- conditions politicians have an incentive for en- ties to account for the management of the civil suring that the civil service is an efficient orga- service through programmatic party organiza- nization as opposed to being simply a mecha- tions, higher standards and less corruption may nism for distributing jobs to supporters, with be the outcome. If, however, political parties scant regard for the quality of public service are beholden to patrons, then narrow clientelis- provision. tic concerns will dominate resource allocation. Using a large dataset, the authors of one Thus, who the politician’s principal is (voters or study found that the existence of programmatic patrons)—which itself depends on the outcome political parties—that is, those that do not rely of collective action (can voters organize into pro- on patrons to organize voters but can instead grammatic parties and keep them program- be held directly accountable by voters—is much matic?)—determines the nature of the principal- more likely to favor a transparent and efficient agent dynamic, as well as the overall quality of civil service than parties that depend on patrons service delivery. Source: Cruz and Keefer (2010). However, what if the agent can potentially be removed from office by tax- payers who have the ability to detect rent extraction and punish any agent who engages in such an activity? In this case, an agent who is more con- cerned with keeping her position will have an incentive to pursue a rent- minimizing agenda because she faces a trade-off: stay compliant and keep the post (payoff of 3) or try and rent maximize and eventually get caught and removed from office (payoff of 2). In short, by looking down the game tree, agents can anticipate the reac- tion of the taxpayers to their different strategies (rent maximize; rent mini- mize) and can, therefore, select the strategy that will make them better off in the end. Thus, backward induction can be used to find the solution for this type of game (known as the “subgame perfectâ€? Nash equilibrium) (see figure 6.2). 194 Understanding Policy Change FIGURE 6.2  Decision Making in a Sequential Form Game Revolt â…· 2 2 Max Rents â…· Taxpayers Accept â…· 4 1 â…· Tax agent Revolt â…· 1 1 Min Rents â…· Taxpayers Accept â…· 3 4 Source: Authors. Summary The need to delegate tasks to politicians and bureaucrats can have signifi- cant effects on the implementation of a project’s objectives because of the potential conflict of interest between principal and agent. Understanding the logic of principal-agent delegation and the tools that might, under certain conditions, mitigate this conflict is therefore a critical component ­ of a good political-economy analysis. Grasping principal-agent dynamics makes it easier to identify some of the implementation challenges, to think about informational discrepancies between actors and groups in society and the unintended consequences of delegation, and to explore attempts to remedy the consequences. By understanding the components and logic of a principal-agent relationship, the development practitioner will be able to do the following: • Make better-informed decisions about delegating tasks within his or her organization • Introduce the most effective monitoring and incentive mechanism within concrete institutional contexts • Recognize entry points for reforms (that is, who is more likely to have in- centives aligned with the reform goals: the principal or the agent? Why? Delegation Problems and the Principal-Agent Relationship 195 Exercises for Chapter 6: Principal-Agent Theory, Delegation, and Accountability Exercise 6.1: The Principal-Agent Game The facilitator explains the steps involved in the Principal-Agent Game. Steps in the exercise (time frame: ~40–45 minutes) 1. Read out instructions (1 minute) 2. Allow participants to read and answer (5–10 minutes) 3. Collect responses, tally responses, facilitate group discussion (10 minutes) 4. Class discussion (10 minutes) 5. Technical explanation (10–15 minutes) Preparation and materials a. Envelopes to be placed in the middle of the round table (one at each table) b. Pens should be available to all participants c. Copies of the Instructions handouts d. Copies of the Answer Sheet e. Copies of the Key Questions handout f. Facilitator instructions NOTE: There are three different scenarios (one for each separate table). If there are only two tables, then Scenarios 1 and 3 should be used. If there are more than three tables, one or more of the three scenarios can be used again. Procedure: 1. The facilitator reads aloud the following instructions: I am going to distribute instructions and an answer sheet. Read the in- structions and answer the questions on the answer sheet. Once you have read the instructions and completed the answer sheet, place your answer sheet in the envelope in the middle of the table. DO NOT WRITE YOUR NAME ON EITHER DOCUMENT and MAKE SURE NO ONE SEES YOUR ANSWER. Keep the instructions handy for our follow-up discus- sion. I will give you about 5–10 minutes to complete this. 2. The facilitator distributes the Instructions handout and Answer Sheet to all participants. See Note above on the three scenarios. 3. The facilitator gauges the rate of completion of the assignment and gives participants a 2-minute warning so that the activity can wrap up within its allotted 10-minute time frame. As part of the 2-minute warning, the 196 Understanding Policy Change facilitator reminds participants that they should not write their names on the answer sheet; when they have finished, they should put their answer sheet in the envelope in the middle of the table. 4. The facilitator explains: I will come around to collect the envelopes with the answer sheets and tally your responses. At the same time, I will give you a handout that in- cludes two Key Questions. While I tally the responses on the answer sheets, please go ahead and discuss at your tables the two questions on the handout. Spend approximately 10 minutes in this discussion. 5. To tally the votes, the facilitator uses a blank Answer Sheet and notes for each option the total number of votes. 6. The facilitator walks around the room among the tables in a nonintrusive manner and pays attention to what is discussed. The facilitator notes a few comments from the tables that he or she can highlight as part of the transition to the technical explanation portion. 7. The facilitator gives the participants a two-minute warning to wrap up their discussion. 8. The facilitator explains: Now that you have played the game and had a chance to discuss with others at your tables some of the key underlying issues, let’s find out the ­ results of your individual decisions. 9. The facilitator announces the results and poses the following questions to the entire group: 1. How many of you were surprised by the entire group’s final tally? 2. If yes, why? 3. If no, why? Delegation Problems and the Principal-Agent Relationship 197 Exercise 6.1: Instructions Handout 1 You are an extremely busy executive working in a field you are passionate about. You have several important projects whose deadlines are all coming up. You know that if you were not under a time constraint, you could com- plete all the projects yourself. Given the time constraints, you have the fol- lowing choices: 1. Try and complete all the projects yourself, doing a satisfactory job overall. 2. Delegating one project to an assistant, allowing you to complete all the other projects extremely well but forcing you to accept whatever quality outcome your assistant produces. You know that your assistant is an extremely competent employee who wants to impress you, as she will soon be up for promotion. 198 Understanding Policy Change Exercise 6.1: Instructions Handout 2 You are an extremely busy executive working in a field you are passionate about. You have several important projects whose deadlines are all coming up. You know that if you were not under a time constraint, you could com- plete all the projects yourself. Given the time constraints, you have the fol- lowing choices: 1. Try and complete all the projects yourself, doing a satisfactory job overall. 2. Delegating one project to an assistant, allowing you to complete all the other projects extremely well but forcing you to accept whatever quality outcome your assistant produces. Your assistant is new, so you are unsure of how well he or she will be able to complete the project. However, it is relatively easy for you to monitor the work, so that if he or she seems to be falling behind you would know almost immediately. Delegation Problems and the Principal-Agent Relationship 199 Exercise 6.1: Instructions Handout 3 You are an extremely busy executive working in a field you are passionate about. You have several important projects whose deadlines are all coming up. You know that if you were not under a time constraint, you could com- plete all the projects yourself. Given the time constraints, you have the fol- lowing choices: 1. Try and complete all the projects yourself, doing a satisfactory job overall. 2. Delegating one project to an assistant, allowing you to complete all the other projects extremely well but forcing you to accept whatever quality outcome your assistant produces. Your assistant is extremely incompetent. He or she was hired only because he or she is a relative of the big boss upstairs. If the assistant does a bad job, there will be no consequences for him or her. 200 Understanding Policy Change Exercise 6.1: Answer Sheet 1 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Do all the work yourself Delegate some of the work to your assistant Briefly explain why you decided to pursue one action rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. Delegation Problems and the Principal-Agent Relationship 201 Exercise 6.1: Answer Sheet 2 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Do all the work yourself Delegate some of the work to your assistant Briefly explain why you decided to pursue one action rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. 202 Understanding Policy Change Exercise 6.1: Answer Sheet 3 Please indicate below, by ticking (î?µ) the box next to the appropriate action, what you would like to do. PLEASE TICK ONLY ONE BOX. Do all the work yourself Delegate some of the work to your assistant Briefly explain why you decided to pursue one action rather than the other. Once you have completed your answers, put them in the envelope in the middle of the table. Delegation Problems and the Principal-Agent Relationship 203 Exercise 6.1: Key Questions Handout—Group Discussion In your groups, please discuss the following questions: 1. What are the main reasons people are giving for the action they have taken? 2. How does the information you have on the way in which your assistant was selected and your ability to monitor him or her affect your answer? (If applicable) 3. How does the information you have on the way in which you can monitor your assistant’s progress affect your answer? 4. What does the example suggest about the importance of information asymmetries in determining decisions to delegate? 204 Understanding Policy Change Notes 1. U4 Anti-Corruption Resource Center, “Revenue Administration and Corrup- tion,â€? http://www.u4.no/themes/pfm/Revenueissue/revenue5.cfm#7. 2. Shor, M. 2005. “Backward Induction,â€? Dictionary of Game Theory Terms, Game Theory.net, Web accessed: 6/18/2012. 3. Inspired by the congressional dominance literature as summarized in (Muller 2003, 386–405). References Bawn, Kathleen. 1995. “Political Control versus Expertise: Congressional Choices about Administrative Procedures.â€? American Political Science Review 62–73. Bendor, J., A. Glazer, and T. Hammond. 2001. “Theories of Delegation.â€? Annual Review of Political Science 4 (1): 235–69. Bendor, J., and A. Meirowitz. 2004. “Spatial Models of Delegation.â€? American Political Science Review 98 (2): 293–310. Cruz, C., and P. Keefer. 2010. “Programmatic Political Parties and Public Sector Reform.â€? Paper presented at the American Political Science Association annual meeting, July 19, Washington, DC. Eaton, K. 2003. “Can Politicians Control Bureaucrats? Applying Theories of Political Control to Argentina’s Democracy.â€? Latin American Politics and Society 45 (4): 33–62. Epstein, D., and S. O’Halloran. 1994. “Administrative Procedures, Information, and Agency Discretion.â€? American Journal of Political Science 38 (3): 697–722. ———. 1999. Delegated Powers: A Transaction Cost Politics Approach to Policy Making under Separate Powers. Cambridge: Cambridge University Press. Holmstrom, Bengt. 1979. “Moral Hazard and Observability.â€? Bell Journal of Economics 10: 74–91. Horn, M. J. 1995. The Political Economy of Public Administration: Institutional Choice in the Public Sector. New York: Cambridge University Press. Huber, J. D., and C. R. Shipan. 2002. Deliberate Discretion: The Institutional Founda- tions of Bureaucratic Autonomy. New York: Cambridge University Press. Huber, John and Charles Shipan. 2006. “Politics, Delegation, and Bureaucracyâ€? in The Oxford Handbook of Political Economy. Ed. Barry R. Weingast and Donald A. Wittman, 256–271. Oxford: Oxford University Press. Kiewiet, D. R., and M. D. McCubbins. 1991. The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago: University of Chicago Press. Lupia, Arthur, and Mathew McCubbins. 1994. “Who Controls? Information and the Structure of Legislative Decision Making.â€? Legislative Studies Quarterly 19 (3): 361–84. McCubbins, Mathew D., and Terry Sullivan. 1987. Congress: Structure and Policy: Political Economy of Institutions and Decisions. New York: Cambridge University Press. Delegation Problems and the Principal-Agent Relationship 205 McCubbins, Mathew, Roger Noll, and Barry Weingast. 1987. “Administrative procedures as instruments of political control.â€? Journal of Law, Economics, and Organization 3:243–77. Miller, Gary J. 2005. “The Political Evolution of Principal-Agent Models.â€? Annual Review of Political Science 8 (1): 203–25. Minh Le, Tuan. 2007. “Combating Corruption in Revenue Administration: An Overview.â€? In The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level, ed. Jose E. Campos and Sanjay Pradhan, 355–38. Washington, DC: World Bank. Moe, T. M. 1989. “The Politics of Bureaucratic Structure.â€? In Can the Government Govern? ed. J. Chubb and E. Peterson, 267–329. Washington, DC: Brookings Institution Press. Muller, D. 2003. Public Choice III. Cambridge: Cambridge University Press Powell, B., and G. Whitten. 1993. “A Cross-National Analysis of Economy Voting: Taking Political Context into Account.â€? American Journal of Political Science 37: 391–414. Sappington, David E. M. 1991. “Incentives in Principal-Agent Relationships.â€? Journal of Economic Perspectives 5 (2): 45–66. Shavell, Steven. 1979. “Risk Sharing and Incentives in the Principal and Agent Relationship.â€? Bell Journal of Economics 55–73. Shepsle, K. A. 1992. “Bureaucratic Drift, Coalitional Drift, and Time Consistency: A Comment on Macey.â€? Journal of Law, Economics, and Organization 8 (1): 111. Shor, M, .2005. “Backward Induction,â€? Dictionary of Game Theory Terms, Game Theory.net, Web accessed: 6/18/2012 Strøm, K. 2000. “Delegation and Accountability in Parliamentary Democracies.â€? European Journal of Political Research 37 (3): 261–90. Strøm, K., C. Wolfgang Muller, and T. Bergman. 2003. Delegation and Accountability in Parliamentary Democracies. Oxford: Oxford University Press. Tsebelis, George. 1999. “Veto Players and Law Production in Parliamentary Democracies: An Empirical Analysis.â€? American Political Science Review 591–608. Weber, Max. 1918. “Science as a Vocation.â€? Speech at Munich University, http:// www.ne.jp/asahi/moriyuki/abukuma/weber/lecture/science_frame.html. World Bank. 2011. “Accelerating Reforms within Iraq’s National Board of Pensions,â€? http://siteresources.worldbank.org/INTMENA/Resources/QuickNote51.pdf >. 206 Understanding Policy Change CHAPTER 7 Causes and Effects of Information Asymmetries in the Policy-Making Process The role of information in fostering pro-development collective action is crucial. Information asymmetry refers to any situation in which one indi- vidual, group, or actor has some information that another actor does not possess. Significant information asymmetries among voters, politicians, and bureaucrats can lead to ineffective representation and undermine the ability of voters to monitor and sanction noncompliance through elections. The magnitude of information asymmetries varies widely around the world. For instance, the 2005 Latino Barometer opinion survey found a large percent- age of respondents agreeing with the statement that “politics is so compli- cated that people like us often do not know what is happening.â€? The varia- tion across countries was also substantial, ranging from 40 percent in República Bolivariana de Venezuela to 69 percent in El Salvador and 68 per- cent in Paraguay (Taylor-Robinson 2010, 15). Information problems are not confined to developing countries only. Sig- nificant information asymmetries characterize all electorates in all coun- tries. Studies done in the United States, for example, show that the Gini coef- 207 ficient of the distribution of political information across U.S. voters is around 0.60. This indicator takes values between 0 (indicating equally distributed information across all voters) and 1 (indicating that only one person knows what is going on in policy making) (Converse 2000, 333). In the absence of strong technologies of political commitment and monitoring, this informa- tion asymmetry significantly reduces effective collective action and leads to less accountable forms of political representation. Despite persistent information asymmetries, in some contexts well- functioning institutional mechanisms of monitoring and precommitment (politicians’ ability to commit to a certain course of action in the future) can substitute for this gap and make sure that development is not entirely de- railed. As a political columnist has suggested, “Never overestimate the infor- mation of the electorate, but never underestimate its intelligenceâ€? (Mark Shields, cited in Converse 2000, 331). In other contexts, attempting to cor- rect information asymmetries can have unintended consequences and can even hamper pro-development collective action. This chapter will explore in greater detail the complicated and often nonlinear relationship between information asymmetries and collective action. Chapters 3 and 4 showed how collective action capabilities and the institu- tional context can determine the ability of different stakeholders to translate their latent preferences into actual policy outcomes. However, because policy is made by politicians and bureaucrats (whose interests may only partly align with those of voters), it is important to consider how they strategically use in- formation flows and gaps to garner support, get elected, or get promoted. To be sure, some participants in economic or political markets simply have more information than others. For example, bureaucrats may know the true cost of providing a publicly financed good and service, but voters and elected officials may not, thereby providing an opportunity for bureaucrats to push for a larger budget than is justified by their outputs. Even if laws and regulations clearly establish the procedural blueprint for budget formulation and approval, some ministries or bureaucratic agencies may have greater informational resources that give them advantageous bargaining positions in a given institutional envi- ronment. Thus, even when institutions function relatively well, actors (as stakeholders) have different degrees of knowledge about the details of the transaction, the process, or the real policy outcome. Information asymmetries are important to understand because they in- fluence the likelihood of change. Voters or policy consumers need informa- tion, first and foremost, to evaluate the performance, motives, or competence of the policy producers. Reformers need to acquire information about other stakeholders with similar policy positions in order to build lasting coalitions. As one of the major ingredients of accountability and effective collective ac- 208 Understanding Policy Change FIGURE 7.1  Conceptual Map of Information Problems and Asymmetries Information Main sources of problems and information asymmetries asymmetries Consequences of Potential solutions Types of information imperfect to information asymmetries information in asymmetries development Source: Authors. tion, information that is available and interpretable by all participants in eco- nomic transactions—voters, politicians, and bureaucrats—is crucial for de- velopment. The availability, transparency, and interpretability of information in the process of policy change can make substantive differences in the final outcome. In fact, under certain conditions, even small amounts of informa- tional correctives can have large effects on development outcomes. While information is critical, it is not a panacea. Under certain conditions, it may actually increase the divergence of interests between stakeholders and policy makers, if the latter resort to pandering or concealing their true agenda in the short run so that they may remain in power long enough to pursue a private agenda at a later point in time. In other circumstances, as we will see, keeping a low profile, bypassing full contestation, and avoiding a politically polarized environment benefit reform teams. This chapter intro- duces the concept of information asymmetries as part of a broader family of economic and political market imperfections (see figure 7.1). Objectives of Chapter 7 The analytics introduced in this chapter will help development practition- ers to: • Understand information problems and asymmetries • Develop an intuitive understanding of the strategic origins of informa- tional imperfections Causes and Effects of Information Asymmetries in the Policy-Making Process 209 • Diagnose the scope and type of information problems within their own policy change process and linking chapter concepts to relevant life situations • Reflect on the relationship between information asymmetries and collec- tive action in their own work • Identify the most context-appropriate strategy for correcting such imperfections. The Political Economy of Information As noted above, information asymmetries arise when a systematic discrep- ancy exists between the information available to the different market par- ticipants or contractual parties. Simply put, one actor knows more than the others. Producers know more about their own products than consumers, sellers know more than buyers, and employees know more their own abili- ties in the workplace than employers. In economics, the earlier general equilibrium theories were derived from three main assumptions: perfect information among market participants in transactions; a complete set of markets (that is, all buyers and sellers were able to trade goods and services with everybody else without incurring transaction costs); and no enforcement problems (contracts are fully en- forced) (Arrow and Debreu 1954). In this theoretical universe, institutions and collective action did not matter at all. Supply and demand were solely responsible for market equilibrium. However, as the 2001 Nobel Prize win- ning team (George Akerlof, Michael Spence, and Joseph Stiglitz) argued, markets suffer from imperfect information problems, which in turn signifi- cantly affect the nature of transactions. Because some parties have more in- formation about their own ability to fulfill the contractual obligations, their actions and choices convey some sort of signals to the other market partici- pants. In return, these participants react to such informational cues (Stiglitz 2000, 1444; Stiglitz 2002, 460). For example, when a firm offers a three-year guarantee on an electronic appliance, it signals product confidence to con- sumers in addition to its commitment to absorb the risk of paying for main- tenance if the appliance breaks. The price of the product, per se, does not convey all necessary information (Stiglitz 2002, 468). Firm managers also have incentives to increase information asymmetries between them and consumers to gain or retain market power. Corporate governance, finance, and organizational design are key domains where man- agers and firm employees have more information than shareholders and consumers, as well as incentives to perpetuate this lack of transparency. One 210 Understanding Policy Change of the key theoretical contributions of this seminal body of work demon- strates that markets alone do not give incentives to participants to disclose information about their characteristics and activities in the absence of insti- tutions that oblige them to do so. Therefore, explicitly incorporating infor- mation problems into the analysis has changed the paradigm in modern economics. Similar (if not larger) information gaps apply to political markets as well. Governments and politicians, by definition, know more about their own competence, policies, and regulations than do citizens. In extreme cases, they do not share anything with the opposition or voters. In some countries, for example, the annual budget document sent to parliaments for review has no more than a few pages. In the recent past, several severe financial crises with worldwide development consequences were triggered by governments’ lack of fiscal transparency: the 1997 Asian financial crisis or the more recent case of Greece in the Euro Area. Information problems are much more acute in the case of governments than in the case of firms. Lack of transparency or “sunshineâ€? conceals mis- takes, corruption, and abuses of power. Moreover, in economic markets, cus- tomers can switch to another product if firm managers use secrecy to cover mismanagement or abuse. In politics, the exit opportunity, especially when the executive perpetuates information asymmetries, is always difficult and often impossible (Stiglitz 2002, 487–488). Information about political repre- sentatives’ behavior and policies also has distributional consequences. Be- cause in many ways such information is a scarce good, it empowers some groups that have access to it at the expense of others. We have already suggested that lack of voter information (due to voter characteristics or a lack of government transparency) is one of the most im- portant sources of information asymmetry in politics. Even in developed contexts, lack of knowledge about candidates’ characteristics, as well as the manipulation of voters’ decisions through political campaigning and adver- tising, give electoral advantage to special interests that can mobilize more effectively to extract favorable policies through behind-the-scenes lobbying, advertising, and the like (Grossman and Helpman 1996). Given that voters cannot adequately monitor the behavior of politicians in contexts with high information asymmetries, politicians can extract rents with little risk of be- ing sanctioned. Therefore, the informational discrepancies among various stakeholders have profound implications for institutions, accountability re- lations between citizens and governments, the collective action potential of stakeholders, and ultimately policy outcomes. Let us also recall from chapter 3 that public information and transparency provided by governments to citi- zens is a public good and the benefits extend to everyone in society (Stiglitz Causes and Effects of Information Asymmetries in the Policy-Making Process 211 2008). Precisely because it is a public good, it also suffers from some of the inherent collective action problems discussed earlier, such as free riding. As we will see in this chapter, citizens want government transparency, but sometimes the costs of individual participation for obtaining it exceed the benefits. Where Do Information Asymmetries Come from? Types of Information Problems Information problems originate in principal-agent relations, a topic that chapter 6 explored in greater depth. This relationship, which characterizes all processes of delegation (say, from politicians to bureaucrats or from vot- ers to politicians), generates so-called agency problems. Simply put, one of the participants in the transaction (the one who delegates, or the voter) runs the risk of not having the delegatee (say, the legislator representing his or her district) fulfill its obligations. While incentives generated by principal-agent relationships are broader than informational problems, transparency consti- tutes a crucial component of agency issues. Two central issues lie at the core of the concept of information asym- metries (Stiglitz 2000, 1447): • The problem of selection. Identifying the characteristics of political participants. • The problem of behavior. Identifying divergent or skewed incentives em- bedded in the relationship between actors. For a more detailed discussion of monitoring and enforcement, see also chapter 9 Both types of information problems refer to the negative consequences of information asymmetries between two groups of actors (say, voters and pol- iticians). We will now present several examples. Self-Selection. “Self-selection is the process through which individuals re- veal information about themselves through the choices they makeâ€? (Roths- child and Stiglitz 1976; Stiglitz 2000, 1450). The typical insurance example of self-selection is the following: an insurance company does not know the real likelihood of an accident for the individual seeking insurance. However, if the latter chooses to purchase an insurance package with a higher deduct- ible than other alternatives available, the insurer can infer from this choice that the individual assesses himself as less likely to have an accident in the short run. As another example, governments often do not know the real eco- nomic ability of taxpayers, but for tax purposes, it bases its assessment of their capacity to pay on what it can observe (such as wealth, income, or, as 212 Understanding Policy Change in medieval Europe, the number of house windows). This process of self- selection in regard to taxation is crucial in developing countries, where gov- ernments are less likely to observe the real income of market participants ­ because of large informal economies or undeclared wealth and income. Hence, they cannot impose an effective income tax schedule (Stiglitz 2000, 1451). As Concepts in Practice 7.1 below illustrates, the role of self-selection in bureaucracies can be a powerful factor in shaping organizational outcomes. Moral Hazard. Moral hazard occurs when one actor has incentives to take a higher risk because he is not bearing the full costs of his actions. Typical examples of moral hazard also come from the insurance market. Homeown- ers’ insurance, for example, generates incentives either to be careless about home safety or, in more extreme cases, to file insurance claims for a loss that never actually occurred. Because insurers cannot monitor behavior ade- quately, they have to find ways to reduce moral hazard by taking indirect preventive measures (for example, installing sprinklers, in the case of home insurers, or setting up institutions that prevent blatant corruption of politi- cians, in the case of voters) (Arrow 1971; Stiglitz 2000, 1453; 2001). CONCEPTS IN PRACTICE 7.1 Individual Characteristics and the Working Environment in Bureaucracies In large and complex institutions, the ability to realize goals or collective man- dates requires the interaction and coordination of a large number of individuals, not all facing the same incentives. The development of an administrative and management system generates incentives for political control (hierarchies). Such mechanisms, in turn, affect who is most likely to rise to the top and make decisions. As Friedrich Hayek noted with respect to the public sector at large, the management of big organizations tends to favor the advancement of those with a relatively unscrupulous “politicalâ€? personality. Even though these people may not be as technically competent as others, or as cooperative, they are es- sential to the realization of the organization’s goals. Thus, paradoxically, the self-selection process generated by the growth of an organization can under- mine its long-term productivity, as resources are diverted to the private agen- das of administrators less closely aligned with the organization’s mission, than more technically competent and cooperative personality types. Source: Hayek (1944). Causes and Effects of Information Asymmetries in the Policy-Making Process 213 Moral hazard types of information asymmetries characterize all relations between voters, political representatives, and bureaucracies. Voters do not know how the legislators they elect in their district will actually behave in the legislature. Therefore, once elected and “insuredâ€? for at least one term of office, politicians have incentives to behave opportunistically, engage in cor- ruption, and deviate from their mandate, in the absence of effective monitor- ing. While the risky action is undertaken by politicians, often voters are the ones incurring the costs, and the relationship of accountability suffers. Another example of moral hazard comes from the historical relationship between imperial centers and provincial notables. Let’s imagine a feudal world with a center and many vassal kingdoms. If a central government gives authority to local rulers to raise taxes on its behalf, that government incurs the potential risk of making enough resources available to the local or re- gional rulers to mobilize armies against the center and claim political power. Therefore, the center has often imposed a lower tax rate on the provinces than it would have done in the absence of the moral hazard problem. Volun- tarily limiting its own tax-collection capacity has also, paradoxically, limited the military threat posed by its local agents. The Ottoman Empire during its classical age (1453–1600s) is a case in point. Through the “fief system,â€? the sultan had incentives to implement provincial taxes below what he could have feasibly extracted, given the possible armed resistance that provincial elites could have mounted once they had accumulated enough resources. This central dilemma of dual delegation of taxation and coercion also char- acterized administrations in medieval Europe, Japan, and Persia (Karaman 2009, 692). Moral hazard problems also characterize more contemporaneous federal arrangements or decentralized institutions around the world. If local or re- gional governments in decentralized systems anticipate bailouts from the central government in case they overspend, they lack incentives not to en- gage in overspending in the first place. The incentives, in this case, generate a collective action problem, because subnational units can engage in a ver- sion of the tragedy of the commons by “overfishingâ€? the total amount of transfers from the center. Since the likelihood of bailouts leads subnational governments to engage in strategic behavior vis-à-vis the central govern- ment, moral hazard problems have profound consequences for the design of intergovernmental transfers. Correcting moral hazard incentives is not an easy task. Furthermore, at- tempts to correct moral hazard incentives can create even more serious problems. In an effort to prevent such incentives, for example, health insur- ance companies set up complex systems to screen for preexisting condi- 214 Understanding Policy Change tions but in the process exclude many from qualifying for insurance in the first place.1 Adverse Selection. Another prevalent form of information asymmetries, closely related to moral hazard, is adverse selection. Adverse selection, to return to our insurance examples, refers to a situation in which the actors who pose the highest risk are also the ones most likely to enter an insurance contract. For example, smokers are more likely than nonsmokers to buy health or life insurance, but, at the same time, they are generally more likely to actually incur the health risks associated with smoking. Therefore, this problem is the “adverseâ€? self-selection of a higher-risk group into the insur- ance contract. One of the classic examples of adverse selection made famous by Nobel Prize winner George Akerlof refers to the market for “lemonsâ€? or bad used cars (Akerlof 1970). Buyers want a quality product but are not sure if the car they are about to buy is bad or good since they cannot observe and thor- oughly evaluate all product dimensions. Therefore, consumers expect to pay an average price that is lower than the real value for a good used car. The sellers of good cars will thus lose in this transaction and have incentives to withdraw from the market. Because only the bad cars stay up for sale, a downward spiral of lower prices and lower-quality products results in a mar- ket dominated by bad cars. Eventually, this process could result in the disap- pearance of the market altogether as consumers are unwilling to pay any- thing at all for the substandard products on offer. Adverse selection problems in politics have a similar logic. Assume there are two kinds of politicians: one who is interested in keeping voters happy through policies and another who is running for office just for the high sal- ary, opportunities for corruption, and side payments. In the absence of in- formation about performance in office, reputation, or the like, voters will not know what kind of politician a candidate is. Especially under conditions of high monetary incentives for members of legislatures, it is likely that more candidates from the second group will run for office. Uninformed vot- ers who cannot tell the two types apart will be more likely to vote for an opportunistic candidate, given that they are more likely to run for office, and the overall quality of political representation will go down. This dy- namic has important consequences for accountability and the provision of public goods (Mattozzi and Merlo 2007). The process of adverse selection can sometimes make voters accept clientelistic payments or enter patron- age networks, as they are skeptical of the overall competence of their politi- cal representatives. Causes and Effects of Information Asymmetries in the Policy-Making Process 215 The Repertoire of Solutions to Problems of Information Asymmetries Two possible solutions for correcting information asymmetries in the principal-agent relationship relate to monitoring and incentive pay (Stiglitz 2000, 1454). Monitoring the behavior of bureaucrats and politicians to prevent moral hazard and adverse selection problems can be done by the media, civil so- ciety, and oversight institutions. One example is public account commit- tees (PACs). The ability of bureaucrats to implement public policy without direct oversight can result in incentives for corruption and waste. Over- sight bodies that undertake random or targeted audits can mitigate such tendencies. In European countries, the power of parliamentary PACs, whose role is to scrutinize the use of public funds, has had varying degrees of success (Buzaljko et al. 2010). Specifically, countries with more power- ful PACs, such as Denmark, waste less public funding than countries in which the powers of the PACs are severely circumscribed, such as Greece. Concepts in Practice 7.2 illustrates a case of bureaucratic oversight in Ghana, where such institutions corrected information asymmetries and significantly improved public finances. Incentive pay for the agent (that is, incentive pay for bureaucrats or politi- cal representatives) is one of the classic prescriptions of civil service reforms. Paying civil servants high wages while also having a clear mechanism for rewarding or punishing bureaucratic performance has sometimes been sug- gested as a good way to tackle corruption and inefficiency. Singapore pays some of the highest salaries for civil servants and politicians in the world. These salaries, combined with a tough penal system and reward bonuses for delivering long-term economic growth, may be one of the reasons behind the success of Singapore and its relatively low rate of corruption, according to some studies (Behnke et al. 2008). However, a caveat is in order: as the adverse selection example showed in the case of political candidates for office, high material rewards for poli- ticians can attract the least ideal candidates who do not enter politics to provide public goods and boost development but rather to enjoy the spoils of office. Organizations that solve collective action problems such as politi- cal parties or civil society organizations face similar challenges, as their leaders are not often able to gauge the degree of commitment to the group of new adherents to the cause. Selective incentives linked to actual perfor- mance in office or in the organization might under certain conditions do a better job of solving adverse selection problems, as they can discourage nonperformers. 216 Understanding Policy Change CONCEPTS IN PRACTICE 7.2 Taming the Bureaucracy through Broadcasting: The Case of Ghana’s Public Account Committee Ghana’s 1992 constitution gave parliament ness and to ensure that its hearings could be oversight over the executive. To supervise pub- held in public where they would receive media lic finances, a public account committee was attention. The public account committee be- established. The committee was headed by a gan holding public meetings in 2007, and the member of the opposition and was responsible first such hearing became a major media for scrutinizing public finances, as in most Com- event. The presence of the media—especially monwealth countries (former U.K. colonies). the private media, which did not stop broad- Formally, the 25-member committee had the casting when ministers were called to answer power to call for “persons and papersâ€?—that is, potentially embarrassing questions about pub- all government officials and documents it consid- lic financial management—not only ensured ered necessary for fulfilling its obligations and high attendance but also led to the uncover- effectively scrutinizing the use of taxpayers’ ing of irregularities in payments and public money. However, in practice, the public account procurement. committee’s powers were severely diminished Thus, the use of the media to televise and because many members did not feel compelled disseminate information about the actions of to attend meetings and the committee’s recom- the bureaucracy ensured, for the first time, that mendations were not implemented. legislative oversight significantly affected the Starting in 2005, a concerted effort was behavior of the executive. made to increase the committee’s effective- Source: Sallas-Menshah (2011). Broader Implications of Information Asymmetries Information asymmetry problems can have profound implications for the role of institutions. According to traditional paradigms, marginal or incre- mental changes in information were not supposed to dramatically affect ac- tors’ beliefs and equilibria. Nevertheless, because of the major role that in- formation asymmetries play in the strategic interaction of different actors (such as voters, politicians, and bureaucrats), we now know that even small amounts of information can have large consequences for institutional equi- libria (Stiglitz 2000, 2002). Because of the significant difference that correc- tive information can make for markets, firms, and consumers, we recognize that some degree of regulation is necessary for market efficiency (Shapiro and Stiglitz 1985). In political markets, transparency and freedom of infor- mation laws, along with demands for increased accountability, can alter sig- Causes and Effects of Information Asymmetries in the Policy-Making Process 217 nificantly the relationship between the government (as agent) and the citi- zens (as principals). Disclosure or freedom of information laws that oblige public authorities to provide information to the general public are sometimes powerful tools for correcting information asymmetries. Some countries, such as Thailand, incorporated them into the constitution. Others joined the wave of global adopters. Unfortunately, laws on paper are not enough. As we will see later in this chapter, voters and citizens must be able to solve collective action problems and act upon them to keep the executive accountable and solve moral hazard or adverse selection problems. Correctors of Information Asymmetries: The Role of the Media, Civil Society, and Donors Lack of information, as well as lack of commitment to deliver on campaign promises on the part of elected politicians, renders accountability relations problematic. Information asymmetries make monitoring difficult and can induce politicians to focus on very narrow constituencies through targeted exchanges as opposed to providing public goods. Alternatively, as the exam- ple presented in Concepts in Practice 7.3 shows, solving information asym- metries can incentivize politicians to provide previously undersupplied pub- lic goods. Mass media has a crucial role to play in the process. Increased transparency can encourage collective action and public participation built around demands for better performance from the government. In line with the Indian example, several studies have found newspaper circulation and media freedom to be inversely correlated with the level of corruption and positively correlated with the security of property rights (Ad- sera, Boix, and Payne 2003; Brunetti and Weder 1999; Ahrend 2002). Coun- tries with more vibrant media seem to have, on average, more accountability, fewer governance problems, and more robust political turnover ( ­ Besley and Prat 2001). Furthermore, strengthening the capacity of the media can make it easier for voters and elected officials to control the executive because it helps reveal the “trueâ€? character of the politicians with more accuracy. Sometimes information collected and disclosed by donors or other mem- bers of the development community in collaboration with domestic media also had an immediate effect on correcting asymmetries. A famous study conducted in Uganda in 1995 attempted to trace central education spending to the final recipients of funds: individual schools. Shockingly, only around 20 percent of total allocations ever reached them. After the negative public- ity following the release of these findings, the government started to share month-to-month information on disbursements to local districts. In addi- 218 Understanding Policy Change CONCEPTS IN PRACTICE 7.3 Feeding the Starving: The Role of the Media in Generat- ing Incentives for Politicians to Respond to Famines and Flood Damage India is an ideal case study of how elections and the role of the media can inter- act to influence officials’ incentives to take certain actions. Despite the variation in political competition across the country—with some states more dominated by a single party than others—and wide differences in the availability of local- language newspapers, it is possible to observe how a more competitive politi- cal system and a more vibrant media might enhance transparency, affect the actions of elected officials, prevent moral hazard, and make officeholders more responsive or accountable in implementing development policies. Using data from 1958 to 1992, a study identified the extent to which varia- tions in political competition and the circulation of the local newspapers af- fected the public distribution of food and calamity relief to marginalized groups most at risk of hunger in times of famine. The study found that representative democracy and the presence of free and independent regional presses were significant factors in ensuring the protection of vulnerable citizens because these elements increased public visibility of relief action or lack thereof. Source: Besley and Burgess (2002). tion, as a result of an ambitious national newspaper campaign that followed this first Public Expenditure Tracking Survey, when the same study was re- peated five years later, 80 percent of the funds reached their beneficiaries. The school headmasters, acting on this information, successfully claimed their share of development funds whenever they did not fully receive them (Banerjee and Duflo 2011, 236–37; Reinikka and Svensson 2004, 2005). Paradoxically, in this particular instance, the information-correction mechanism has also led to a rethinking of the entire principal-agent relation- ship embedded in the process of decentralization that, in turn, contributed to the reversal of the initial reform. The public revelation of dramatic inefficien- cies in allocation of the education fund to the localities created an opportunity for bureaucratic agencies opposing decentralization to recentralize. As a re- sult of this new information that became available in 2002, the revised fiscal decentralization strategy significantly restricted the fiscal autonomy of local governments. Until that point, Uganda, together with South Africa, was the leading adopter of decentralization reforms in Sub-Saharan Africa (Eaton, Kaiser, and Smoke 2010, 57–58). The indirect role that increased transparency played in the process of reform reversal sheds light on some of the potential unintended consequences of information correctives. It also calls attention to Causes and Effects of Information Asymmetries in the Policy-Making Process 219 the strategies employed by some segments of the bureaucracy that used the newly available evidence to consolidate their position and recentralize. In addition to the availability of news sources, patterns of media owner- ship and their implications for overall transparency also matter significantly. Some cross-national studies have shown that state ownership of the media generally correlates with higher corruption (Djankov et al. 2003). Other re- cent in-depth case studies have analyzed the impact of commercial liberal- ization of Chinese media on regulatory implementation and argued that this process has helped reinforce the state’s legitimacy rather than challenging it (Stockmann and Gallagher 2011). Private media outlets effectively bolstered the state narrative of a proworker bias in judicial cases related to labor con- flicts, and boosted the overall confidence in the legal system. In addition to mass media, civil society organizations have a crucial role to play in promoting transparency. In some cases, the executive has private information that voters do not possess and has moral hazard in- centives. In other cases, the government itself lacks the capacity to collect crucial data from citizens that could help provide public goods more ef- fectively. Sometimes, civil society intermediaries provide key solutions to these bidirectional information problems. The government of Bangla- desh, for example, responded much more promptly to the floods of 1998 than it did to a similar natural calamity that had occurred 10 years earlier, in 1988. During the 1988 floods, the government attempts to distribute emergency relief, food and water were severely curtailed by a poor distri- bution network, weak engagement with civil society organizations, and a general lack of information about the specific needs of victims across the damaged areas (Beck 2005, 11). These factors, coupled with pervasive and institutionalized corruption, resulted in a severe misallocation of re- sources. Shelters were built in sub-optimal locations, and some victims even had to pay rent. The 1998 floods affected approximately 68% of the country and directly impacted around 30 million people (ibid, 3). Despite a disaster magnitude comparable with the previous episode and higher population densities, the death toll was 60 percent lower. Why? The role of civil society organi- zations turned out to be critical in facilitating a more effective response to the crisis. First, the NGOs had effective local distribution networks. Sec- ond, they were better able to collect crucial and timely information about the victims’ needs, especially as the advent of mobile phones and GPS de- vices facilitated this effort (Beck 2005, 11–14; Young et al. 2000). The gov- ernment coordinated with thousands of NGOs, took advantage of their information collection and distribution capacity, and finally ensured a sig- nificantly better response compared to 1988. 220 Understanding Policy Change The Role of Technology in Correcting Information Asymmetries Let us recall the discussion from chapter 4 about the role that major technol- ogy shifts have had on institutional change. Technology is increasingly being used to help stakeholders correct information asymmetries, to monitor the provision of public goods and services in a principal-agent relationship, and to solve collective action problems. Many organizations are using techno- logical innovations to oversee public service delivery, to measure its impact on development around the world, and to create platforms and virtual com- munities of citizens with interests in development issues. Concepts in Prac- tice 7.4 describes how these technologies have been applied to improve out- comes in both developed and developing countries. CONCEPTS IN PRACTICE 7.4 Using Technology to Monitor the Provision of Public Goods Some development programs are using cell For instance, in Kenya, a free technology plat- phones to track whether promised goods have form was created to allow the monitoring of the arrived at their destination. Others are using 2010 constitutional referendum and presented technology on a bigger scale, developing pro- a space for citizens to report acts of violence grams like Mapping for Results, which moni- (by webmail or mobile phones). The platform tors where World Bank aid goes and to whom.a became so popular that it had 45,000 users.c In Moreover, whether it is to express their discon- the case of India, a technology tool was created tent or to share their data and exchange their to track the behavior of political leaders through- stories, citizens are increasingly using technol- out their time in office, showing when and how ogy to learn, communicate, and solve collective citizens solicit information from their officials.d action problems for institutional change. The Another study developed in Indiae shows that role of Facebook or Twitter for mobilizational using cameras to monitor teachers’ attendance efforts is well established by now. The Arab ­ could contribute significantly to a reduction in Spring is a case in point. teacher absenteeism. This study financially re- Technology also allows better monitoring of warded teachers who proved, with a camera, agents by principals and the efficient correction their attendance at school. While many factors of information asymmetries. A recent studyb are involved in increasing accountability, these presents an analysis of cases from Brazil, Chile, and many other studies suggest that technol- India, Kenya, and Slovakia, where technology ogy can certainly become helpful in the process interventions seem to have increased the ac- of correcting information asymmetries between countability of public organizations, expanded citizens (as principals) and bureaucrats and poli- their services, and improved their outcomes. ticians (as agents). a. See http://maps.worldbank.org/. b. Fung, Gilman, and Shkabatur (2011). c. See http://www.ushahidi.com/about-us. d. See http://www.transparency-initiative.org/wp-content/uploads/2011/05/impact_case_studies_final1.pdf. e. See Duflo and Hanna (2006), http://www.hks.harvard.edu/inequality/Seminar/Papers/Duflo06.pdf. Causes and Effects of Information Asymmetries in the Policy-Making Process 221 Can Transparency Have Unintended Consequences? Despite the normative halo surrounding transparency in development cir- cles, in some cases full information availability and policy contestability have created pitfalls. Under certain circumstances, induced transparency (as a corrective mechanism for information asymmetries) may distort outcomes or provide disincentives for collective action and claims making: • More transparency might cause excessive politicization, with potentially negative consequences for legislative outcomes (Heald 2003). • In some cases, increased transparency might lead to decreased efficiency of high-quality service providers in the public sector. In general, teachers and health workers unions have mobilized against transparency. Increased visibility of performance may attract more service users. Paradoxically, un- like in the private sector, under conditions of rationing and budget con- straints, this phenomenon might put high pressure on high performers and decrease overall efficiency of operation (Gavazza and Lizzeri 2007). • More transparency might encourage low-competency incumbents who do not stand a chance of being reelected to give up trying to appease vot- ers and pursue a strategy of aggressive short-term rent extraction (Besley 2006). • Transparency can also induce politicians to keep costly preelectoral promises at the expense of policy flexibility and prevent “goodâ€? candi- dates from running for office in the first place (Geraats 2005). • Sometimes, a certain degree of secrecy or low visibility may lead to better allocation of resources or reform outcomes, because those that would lose from reforms will not be able to coalesce around a clear and highly visible target and block the attempt at change. • As a result of transparency laws aimed at correcting information asym- metries, the process of requesting information may entail high costs or risks for the actors involved. It can sometimes allow autocratic incum- bents to target those who are seeking information as political opponents. • Information symmetries between the executive and private firms com- peting for public procurement contracts may, under certain conditions, cause collusion, with negative consequences for the bidding process (Olken 2007). Given that the effects of any attempt to correct information asymmetries are contingent on specific contexts and preexisting institutions, the analyst should be cautious when selecting the solution, and explore its potential un- intended consequences. 222 Understanding Policy Change FIGURE 7.2  Influence of Transparency, through the Media, on Government More informed voters mobilized effectively and demanded accountability Mechanism 1 Transparency Faster government response (newspaper circulation in to crop flood damage local languages) Mechanism 2 Broader media appeals allowed politicians to take credit for targeting policies to narrow electoral constituencies Source: Authors. In addition to theoretical or conceptual mechanisms, empirical evidence may also require a careful selection of indicators and interpretation of find- ings. Figure 7.2 presents an example of increased transparency that might have unintended consequences. Using the example of connecting higher newspaper circulation in India with faster government response to crop flood damage and famines, let us think systematically about the different ways in which we could interpret the results (this point was developed by Keefer 2004, 266, based on Besley and Burgess 2002). Whereas mechanism 1 in figure 7.2 is aligned with the general hypothesis that increased transparency leads to better development outcomes through collective action, mechanism 2 demonstrates an alternative explanatory route for the empirical finding. This second explanation suggests that, inde- pendent of collective action and popular claims for greater accountability, local politicians might want to target the affected political constituencies because of the high policy visibility that allows them to take immediate credit for reelection purposes. At the same time, they can reduce the deliv- ery of public goods or programmatic spending in other “less visibleâ€? do- mains (education, health, or districts not affected by crop damage, for ex- ample). Therefore, the causal relationship among indicators of transparency, collective action, and observed policy outcomes should always be inter- preted with caution. Causes and Effects of Information Asymmetries in the Policy-Making Process 223 Understanding the Nonlinear Relationship between Information Asymmetries and Pro-Development Collective Action Does transparency or the correction of information asymmetries always fa- cilitate pro-development collective action? Under what conditions do they fulfill that role? Does newly acquired information always bolster the poten- tial of reform groups to mobilize? Analysts and reformers interested in un- derstanding the political economy of information should carefully think about all possible causal configurations and contingencies. It needs to be emphasized that robust accountability of politicians to citi- zens or voters entails two components: transparency and collective action. Government transparency by itself does not always trigger public participa- tion that automatically translates into socially beneficial policies. Freedom of information laws on paper, without accompanying mechanisms that en- sure that citizens know their rights and can collectively act upon them to solicit information, do not mean much. In fact, sometimes they could even be detrimental in the short run when claiming these rights attracts political retaliation from powerful incumbents. This outcome is most obvious in non- democratic regimes where the cost of acting on existing laws (often adopted at the pressure of international donors) and requesting sensitive data on, say, public expenditure is prohibitively high for individuals and groups. Even in the case of democratic polities, it may be difficult for transparency alone to increase accountability, assign individual responsibility, or effec- tively sanction officials in contexts in which the performance of individual legislators in office is difficult to gauge. If political parties are weak and fail to solve their own collective action problems, it is not easy for elected offi- cials to provide anything other than redistribution to a narrow set of partisan supporters to be reelected, whatever their individual competence. For ex- ample, evidence from Benin suggests that increased access to information, through broader radio reception, has not heightened demand for better pub- lic education, possibly because of the weakness of the national party system and inadequate complementary mechanisms of accountability (Keefer and Khemani 2011). Increased access to information did, however, induce some behavioral changes. In general, the relationship between information asymmetries and collec- tive action is under theoretical scrutiny, but it seems to be more nuanced than a simple linear dynamic through which transparency always positively and unconditionally affects pro-development collective action. Synergistic institutions of accountability might be necessary to complement increased transparency for positive outcomes. 224 Understanding Policy Change The following vignettes tell four different stories about this relationship. The first case shows how citizens mobilized successfully, collected data, and increased transparency. Despite these accomplishments, in the absence of complementary accountability mechanisms, they could not align politicians’ incentives with their policy goals. As a result, basic water services remained underprovided in rural South Africa (collective action and transparency, but no results). The second story is equally pessimistic and shows that in Peru’s legislature a higher degree of transparency and increased accountability of individual legislators to constituencies paradoxically reduced the collective action potential of opposition to the corrupt Fujimori government (transpar- ency, no collective action, no results). The third vignette is more optimistic. Despite death threats, a high-rank- ing bureaucrat in the Nigerian government decided to disclose key public finance data in national newspapers and thereby increased accountability (no collective action, transparency, and concrete results). The fourth instance, also optimistic, shows how a team of Kenyan reformers kept a low profile to avoid open political confrontation with powerful vested interests and was able to change key regulations governing a very corrupt and clientelistic public procurement system (collective action, no transparency, and results). The puzzling variation in the capacity of key driving factors such as trans- parency and collective action to produce tangible development outcomes shows that, in these four stories, there was no necessary or sufficient condi- tion for reform success. Both transparency and collective action might need to interact with other specific sets of incentives and institutions to generate robust and lasting accountability relations between voters and politicians. Information, Collective Action, and Water Access in Rural South Africa A project initiated by a civil society organization in a small South African municipality involved local governments, tribal elders, and citizens in an exercise meant to facilitate community assessment of water services and demand policy action from the state agencies responsible for implementa- tion (Hemson and Buccus 2009). The community first participated in ­ training workshops to learn how to use scorecards and then collected ac- tual information on water sources, conditions, and access throughout the municipality. Empowered by the newly acquired information, the community appealed to the local officials who were initially interested in the project, recognizing that the local council would have never obtained these valuable data without this ambitious effort of community mobilization and collective action. De- Causes and Effects of Information Asymmetries in the Policy-Making Process 225 spite this first enthusiastic feedback, however, the election time was ap- proaching, and internal divisions within the African National Congress (the incumbent party) created competition between two tiers of local govern- ment. Therefore, to avoid signaling incompetency to the upper tier, an elected official blocked the dissemination of the data showing problematic water management, and the problems of water services were never ad- dressed. In this case, although effective collective action led to information correction, the outcome was suboptimal, because of electoral incentives and lack of robust political accountability. Transparency and Collective Action—Trade-Offs in Peru’s Legislature The Peruvian Congress under Fujimori provides a counterintuitive story in which the correction of information asymmetries and increased capacity of voters to monitor the performance in office of their political representatives led, paradoxically, to weaker collective action among the opposition parties and, as a result, to the political dominance of a corrupt party and head of state (Carey 2003). This vignette reveals the benefits of nonlinear thinking that would a priori suggest that transparency and information asymmetries automatically solve pro-development collective action problems. In 1997, at the advice of the Inter-American Development Bank, the Peruvian Congress bought an electronic vote-counting machine to replace hand counts prone to mistakes and irregularities. Initially, legislators opposed the technological upgrade, fearing that their votes would be on record. Ultimately, with buy-in from the incumbent party, the machine was purchased, voting began to be recorded, and the media took full advantage. Newspapers and TV stations began to report on the performance in office of individual legislators (num- ber and type of bills, for example). Paradoxically, close to the year 2000, when Fujimori’s party had a nar- row majority in parliament, this technological upgrade that increased transparency also gave individual politicians incentives to cater to their lo- cal constituencies rather than to act collectively as a party. All of a sudden, it became possible for voters to identify the extent to which their represen- tative had “brought home the baconâ€?—that is, provided her supporters with goods and services. As a consequence, the opposition parties became frag- mented and lacked discipline and cohesiveness, whereas the incumbent party—under the strong leadership of Fujimori—behaved as a compact vot- ing bloc and locked in opportunities for vetoing all reforms that attempted to dismantle its patronage and corruption. Information asymmetries be- 226 Understanding Policy Change tween politicians and voters were indeed corrected, but in this case greater transparency came at the expense of opposition legislators and undermined their effective collective action. Risky Disclosure of Key Budget Information in Nigeria In Nigeria, in 2004, under conditions of severe budgetary opaqueness and high political stakes associated with the oil revenues, finance minister Okonjo-Iweala started to publish the state-by-state central transfers to the 36 state governments in national newspapers each month. This public dis- closure made some of the state governors particularly uncomfortable (Val- lely 2006). As a result of her commitment to enhance fiscal transparency, the minister made many political enemies and even received a number of death threats. Similar consequences of information disclosure affect many journalists around the world every day. In cases like this, because of the high individual cost of action and claims making, generating transparency is a public good and suffers from collective action problems. As already ar- gued, if the benefits of information on revenue and spending extend to all citizens but the prohibitively high costs fall on only a handful of individuals (for example, death threats or actual political persecution), the fight for in- formation is not likely to take place. It is precisely in these settings that leadership, activism, and entrepreneurship become necessary to undertake risks and open up the path to the collective action needed to correct infor- mation asymmetries. The Nigerian junior finance minister is a good exam- ple of these requirements. Strategic Obfuscation, Collective Action, and Reform in Kenya The case of Kenya we mentioned in chapter 5 demonstrates that, quite coun- terintuitively, reform sometimes has better chances under conditions of rela- tive opaqueness.2 In 2001, during the last days of the Moi administration, the incumbent party was desperately trying to use government funds to finance its patronage networks to maintain power. Public procurement turned out to be one of the main venues of funding the party through the discretionary award of contracts to private firms and individuals. A small circle of Kenyan reformers decided to make the issue salient, increase its visibility, and garner public support. According to the conventional wisdom corroborated by the advice of international donors, making the issue salient and initiating a na- tional debate that would help coordinate unorganized reform supporters from different segments of society would weaken resistance and opposition Causes and Effects of Information Asymmetries in the Policy-Making Process 227 CONCEPTS IN PRACTICE 7.5 The Consequences of Information Asymmetries on Development Outcomes How can we formally conceptualize the persis- (cooperate, cooperate as equilibrium). If only tence of suboptimal equilibria due to informa- one group tries to promote the issue, both tion costs? One way is to think of political and groups will lose out, because the issue will be market imperfections as leading to suboptimal publicly perceived as lacking significant and equilibrium outcomes (that is, “steady statesâ€?; joint support. If both groups do not cooperate, see appendix A and chapter 3). A good equilib- the issue will be considered a low priority but rium is one in which the outcome is the most will not be neglected as if it were an issue that favorable possible. Conversely, a bad equilib- had been poorly supported by only one group rium is one in which, for some reason (such as (don’t cooperate, don’t cooperate equilibrium). information asymmetries or costs) resources If the ability of these two groups to coordi- are being inefficiently allocated. nate their strategies depends on the cost of in- At the most abstract level, it is possible to formation they face, then changes in informa- conceptualize the existence of “goodâ€? and tion costs can have a profound effect on “badâ€? equilibria using the tools of game theory. whether the optimal (cooperate, cooperate) or As the table below shows, two players can the suboptimal (don’t cooperate, don’t cooper- choose to either cooperate or not cooperate. ate) equilibrium emerges. For example, each There are two Nash equilibria in this game (coop- group might need to gather information on the erate, cooperate) and (don’t cooperate, don’t co- size, mobilization capacity, commonality of inter- operate). It is not possible to deduce, a priori, ests, competency of the leadership, and influ- which equilibrium emerges. However, if informa- ence of the other group before it decides tion costs between players affect their ability to whether to try and coordinate in pressing a com- coordinate, then it becomes possible to estab- mon cause. However, obtaining such informa- lish the conditions under which the optimal or tion can be costly. Therefore, as information suboptimal equilibrium is most likely to emerge. costs fall (or rise), the ability of the two groups For example, two civil society organizations to communicate effectively and jointly coordi- might benefit from jointly promoting a certain nate increases (or decreases); and, in turn, the issue, thereby, raising public awareness and result increases (or decreases) the possibility mobilizing resources in favor of their agendas that the higher equilibrium outcome will emerge. Cooperate Don’t Cooperate Cooperate 4 4 1 1 Don’t Cooperate 1 1 3 3 Source: Inspired from (Cartwright et al 2009). to policy change. Animated by this impulse, the reformers approached the permanent secretary in the Ministry of Finance. They were in for surprising advice. The strategy suggested to them was exactly the opposite. Instead of becoming easy targets of political vengeance 228 Understanding Policy Change under conditions of making open claims, they were advised to proceed as “quietlyâ€? as possible through procedural channels. Taking the advice, the group of reformers identified a package of implementation rules (an omni- bus bill) that would automatically follow the legislative vote on the annual budget, ensuring the precision and scope of regulatory implementation. Cir- cumventing debate altogether with help from within the bureaucracy, the team—to its surprise—was able to insert the desired amendments as if they were yet another strictly technical regulation. As a result, crucial reforms that would never have succeeded if the issue had been clear and open to contestation to all political actors passed because of strategic obfuscation. Our Philippines case study paints a similar portrait (see appendix C). As illustrated by the above examples, solving information asymmetries and enhancing transparency can sometimes help or jeopardize reform coali- tions and their opportunities to generate change. Information sharing allows coordination, joint action, and trust-building among members of reform teams. Sometimes, as we have seen, disclosure and transparency may have unintended consequences. By understanding this nuanced relationship be- tween correctives to information asymmetry and the potential for collective action, development practitioners can identify the best feasible strategy for achieving results (see Concepts in Practice 7.5). Summary Information asymmetries and the incentives they generate are crucial con- straints to pro-development collective action and policy change. Problems of missing information, moral hazard, or adverse selection pose serious prob- lems for the accountability of political representatives to voters and often lead to suboptimal delivery of public goods. However, the relationship be- tween transparency and collective action, on the one hand, and development outcomes, on the other hand, is nonlinear and might be contingent on paral- lel accountability mechanisms, monitoring devices, or institutional features. Accounting for nonlinearity allows practitioners to think in a context-­ specific way about the theoretical and empirical path leading from transpar- ency to tangible development reforms. Notes 1. We are grateful to Margaret Levi for this point. 2. WBI and CommGap, “Political Economy Analysis to Action: A Global Learning Event,â€? June 2010, Washington, DC. Causes and Effects of Information Asymmetries in the Policy-Making Process 229 References Adsera, Alicia, Carles Boix, and Mark Payne. 2003. “Are You Being Served? Political Accountability and Quality of Government.â€? Journal of Law, Economics, and Organization 19 (2): 445–90. Ahrend, Rudiger. 2002. “Press Freedom, Human Capital, and Corruption.â€? Working Paper 2002-11, DELTA, Paris. Akerlof, George A. 1970. “The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism.â€? Quarterly Journal of Economics 84 (3): 488–500. Arrow, Kenneth J. 1971. “A Utilitarian Approach to the Concept of Equality in Public Expenditure.â€? Quarterly Journal of Economics 85 (3): 409–15. Arrow, K. J., and G. Debreu. 1954. “Existence of an Equilibrium for a Competitive Economy.â€? Econometrica 22: 265–90. Banerjee, A., and E. Duflo. 2011. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. New York: Public Affairs. Beck, T. 2005. “Learning Lessons from Disaster Recovery: The Case of Bangladesh. Disaster Risk Management.â€? Working Paper Series 11. Washington, DC: The World Bank. Behnke, Kathrin, Alexander Hamilton, Leo Pagnac, and Paulina Terrazas. 2008. The Dynamics of Legislative Rewards. London: London School of Economics and World Bank. Besley, T. 2006. Principled Agents? The Political Economy of Good Government. New York: Oxford University Press. Besley, Timothy, and Robin Burgess. 2002. “The Political Economy of Government Responsiveness: Theory and Evidence from India.â€? Quarterly Journal of Economics 117 (4): 1415–51. Besley, Timothy, and Andrea Prat. 2001. “Handcuffs for the Grabbing Hand? Media Capture and Government Accountability.â€? CEPR Discussion Paper, Center for Economic Policy Research, Washington, DC. Brunetti, Aymo, and Beatrice Weder. 1999. “More Open Economies Have Better Governments.â€? Economic Series 905, University of Saarland, Saarbrücken, Germany. Buzaljko, Karolina, Anna Marlene Kanis, Alexandra Tamasan, and Frans Verkaat. 2010. “Public Financial Oversight: A Comparative Analysis of Parliamentary Committees across Europe.â€? Maastricht Graduate School of Governance, Maastricht, Netherlands, and World Bank Institute, Washington, DC. Carey, John M. 2003. “Transparency versus Collective Action.â€? Comparative Political Studies 36 (9): 983–1006. Cartwright, E., J. Gillet, and M. Van Vugt. 2009. “Endogenous Leadership in a Coordination Game with Conflict of Interest and Asymmetric Information.â€? School of Economics Discussion Papers, No. 0913, University of Kent, http://hdl. handle.net/10419/50571’ Web accessed: 6/18/2012. Converse, Philip. 2000. “Assessing the Capacity of Mass Electorates.â€? Annual Review of Political Science 3 (1): 331–53. 230 Understanding Policy Change Djankov, D., C. McLiesh, T. Nenova, and A. Shleifer. 2003. “Who Owns the Media?â€? Journal of Law and Economics 46 (2): 341–82. Duflo, Esther, and Rema Hanna. 2006. “Monitoring Works: Getting Teachers to Come to School.â€? NBER Working Paper 11880, National Bureau of Economic Research, Cambridge, MA, http://www.hks.harvard.edu/inequality/Seminar/ Papers/Duflo06.pdf. Eaton, Kent, Kai Kaiser, and Paul Smoke. 2010. The Political Economy of Decentral- ization Reforms: Implications for Aid Effectiveness. Washington, DC: World Bank. Fung, Archon, Hollie Gilman and Jennifer Shkabatur. 2011. Impact Case Studies from Middle Income and Developing Countries, http://www.transparency-initiative. org/wp-content/uploads/2011/05/impact_case_studies_final1.pdf. Gavazza, A., and A. Lizzeri. 2007. “The Perils of Transparency in Bureaucracies.â€? American Economic Review 97 (2): 300–05. Geraats, Petra. 2005. “Transparency of Monetary Policy: Theory and Practice.â€? CESIfo Economic Studies, Vol. 52 (March): 111–52. Grossman, Gene, and Elhanan Helpman. 1996. “Electoral Competition and Special Interest Politics.â€? Review of Economic Studies 63 (2): 265–86. Hayek, Friedrich A. 1944. The Road to Serfdom. Chicago: University of Chicago Press. Heald, D. 2003. “Fiscal Transparency: Concepts, Measurement and UK Practice.â€? Public Administration 81 (4): 723–59. Hemson, David, and Imraan Buccus. 2009. “The Citizen Voice Project: An Interven- tion in Water Services in Rural South Africa.â€? IDS Bulletin 40 (6): 60–69. Karaman, K. K. I. 2009. “Decentralized Coercion and Self-Restraint in Provincial Taxation: The Ottoman Empire, 15th–16th Centuries.â€? Journal of Economic Behavior and Organization 71 (3): 690–703. Keefer, Philip. 2004. “What Does Political Economy Tell Us about Economic Development and Vice Versa?â€? World Bank Annual Reviews on Political Science 7: 247–72. Keefer, Philip, and Stuti Khemani. 2011. “Mass Media and Public Services: The Effects of Radio Access on Public Education in Benin.â€? Policy Research Working Paper 5559, World Bank, Washington, DC. Mattozzi, Andrea, and Antonio Merlo. 2007. “The Transparency of Politics and the Quality of Politicians.â€? American Economic Review 97 (2): 311–15. Olken, B. 2007. “Monitoring Corruption: Evidence from a Field Experiment in Indonesia.â€? Journal of Political Economy 115 (2): 200–49. Reinikka, Ritva, and Jakob Svensson. 2004. The Power of Information Evidence from a Newspaper Campaign to Reduce Capture. Washington, DC: World Bank, http:// www.econ.worldbank.org/view.php? type=5&id=34028. ———. 2005. “Fighting Corruption to Improve Schooling: Evidence from a Newspa- per Campaign in Uganda.â€? Journal of the European Economic Association 3 (2/3): 259–67. Rothschild, Michael, and Joseph E. Stiglitz. 1976. “Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information.â€? Quarterly Journal of Economics 90 (4): 630–49. Causes and Effects of Information Asymmetries in the Policy-Making Process 231 Sallas-Menshah, Samuel. 2011. “Engaging Civil Society: Ghana’s First Public PAC Hearing.â€? In African Parliamentary Reform, ed. Frederick Stapenhurst, Rasheed Draman, Andrew Imlach, Alexander Hamilton, and Cindy Kroon, 5–14. New York: Routledge. Shapiro, Carl, and Joseph E. Stiglitz. 1985. “Equilibrium Unemployment as a Worker Discipline Device.â€? American Economic Review 75 (4): 892–93. Stiglitz, J. E. 2000. “The Contributions of the Economics of Information to Twenti- eth Century Economics.â€? Quarterly Journal of Economics 115 (4): 1441–78. ———. 2001. “Challenges in the Analysis of the Role of Institutions in Economic Development.â€? In The Institutional Foundations of a Market Economy, ed. Gudrun Kochendorfer-Lucius and Boris Pleskovic, 15–28. Bonn: German Foundation for International Development. ———. 2002. “Information and the Change in the Paradigm in Economics.â€? American Economic Review 92 (3): 460–501. ­ ———. 2008. “Fostering an Independent Media with a Diversity of Views.â€? In Information and Public Choice: From Media Markets to Policymaking, ed. R. Islam, 139–52. Washington, DC: World Bank. Stockmann D., and M. Gallagher. 2011. “Remote Control: How the Media Sustains Authoritarian Rule in China.â€? Comparative Political Studies 44 (4): 436–67. Taylor-Robinson, M. 2010. Do the Poor Count? Democratic Institutions and Account- ­ niversity ability in a Context of Poverty. University Park: Pennsylvania State U Press. Vallely, Paul. (2006, May 16). “Transcript of Interview with Mrs. Ngozi Okonjo- Iweala, Nigerian Finance Minister.â€? The Independent (UK). Retrieved April, 2012. Young, R. and Associates. 2000. “Bangladesh: 1998 Flood Appeal: An Independent Evaluation.â€? Active Learning Network for Accountability and Performance in Humanitarian Action. London: Overseas Development Institute. Available at: http://www.alnap.org/pool/files/erd-2859-full.pdf 232 Understanding Policy Change CHAPTER 8 Credible Commitment Credible commitment mechanisms that give different actors the incentives to follow through on the actions to which they are formally committed con- stitute an important ingredient for successful reform (see figure 8.1). With- out “tying their own hands,â€? leaders can violate or renege on their own agen- das, and voters may be unable to oust incumbents who are pursuing bad policies. Leaders can signal their credible commitment to specific policies by es- tablishing some form of institutional limits on their power or by following through on their campaign promises. The following examples introduce some of the conventional credible commitment mechanisms adopted in many countries: • Control of budgets by parliaments can make it difficult, if not impossible, for leaders to violate key expenditure and macroeconomic commitments. • An independent judiciary and politically insulated bureaucracies can help monitor and enforce policy commitments and deter politicians from colonizing institutions before or after elections. 233 FIGURE 8.1 Conceptual Map of Credible Commitment Time inconsistency and the Credible commitment as policy problems it can create a solution Mechanisms of credible commitment Conditions that enable credible commitment mechanisms Source: Authors. Objectives of Chapter 8 By the end of this chapter, readers should be able to do the following: • Explain the logic of credible commitment as a solution to time inconsis- tencies problems • Think about the connection between collective action and credible commitment • Explore how institutions may induce opportunities to make credible commitments • Understand the role of independent third-party monitoring and enforcement • Link the logic of credible commitment to real-life situations. The Problem: Time Inconsistency The concept of “time inconsistencyâ€? or “dynamic inconsistencyâ€? refers to a change in the preferences of a decision maker that occurs between an initial policy promise and a policy decision that takes place later. Because politicians want to be reelected, their preferences might change over time. Without insti- tutional guarantees, however, it is not possible for politicians today to ensure that they will take the same action in the future. Time inconsistency is often one of the key determinants of whether credible commitment problems exist. 234 Understanding Policy Change As an example, politicians often campaign on reducing inflation in order to win elections. However, once they are in office, implementing inflation- reducing policies might have politically unpopular consequences, such as increased unemployment. Therefore, even if the initial preferences of the government policy makers were consistent with the goals of reducing infla- tion, once the time has come to implement them, they may lack incentives to follow through. One of the classic examples of time inconsistency and credible commit- ment is quite familiar, drawing on the dilemma of Ulysses in Homer’s Odys- sey. On the one hand, Ulysses and his sailors want to hear the mesmerizing Sirens’ song. On the other hand, Ulysses knows that once he hears it, it could be lethal. He would be tempted to swim toward them at the cost of his life. Therefore, Ulysses asks his crew to tie him to the mast of the ship. The hero instructs his sailors not to release him during the song even if, enchanted by the Sirens, he begs to be released. In other words, he designed a mechanism to protect himself from his own incentives to undertake a risky action at a later point in time (Elster 2000). The concept of time inconsistency can also be applied to the issue of property rights, as it has significant consequences for investment and eco- nomic development. “Economic activities often involve time-inconsistent exchanges between the state and private agentsâ€? (Frye 2004, 455). For ex- ample, a government promises favorable investment conditions for a given company (such as preferential tax treatment or regulatory measures). However, once the company decides to invest, the government may renege ex post on the ex ante promises, despite written laws or agreements that stipulate differently. If the company could foresee the time-inconsistent behavior of the government, it would not invest in the first place (Frye 2004; Diermeier et al. 1997). What Is “Credible Commitmentâ€?? The concept of credible commitment, closely related to “precommitment,â€? refers to a situation in which politicians (and, more specifically, the gov- ernment) “tie their own handsâ€? against future discretionary use of state institutions for political or private gains. Given the time inconsistency problems illustrated above, commitment mechanisms block politicians from taking advantage of institutions for political purposes when their in- centives may change (say, right before an election). For example, by mak- ing the judiciary independent, politicians may limit their own ability to engage in corruption even when they have strong career incentives to do Credible Commitment 235 so. Refraining from such politically gainful but socially detrimental behav- ior implies the design of strong rules of the game recognized by all actors concerned. Understood as such, credible commitment is a solution to the problem of time inconsistencies. The creation of an independent central bank is another standard exam- ple of a commitment device because it stabilizes inflation rates and infla- tionary expectations by guaranteeing that populist governments will not engage in overspending right before elections. Similarly, civil service reform in a democratic system that would uproot corruption and clientelism pre- supposes that legislative parties are institutionally committed not to popu- late the bureaucracy with their party members when they have the chance (if they win the next elections). The logic of credible commitment can be applied to a whole variety of institutional choices through which politicians tie their own hands in different settings and policy domains: democracy, bu- reaucratic reforms, fiscal rules, and natural resource governance are just a few examples. What Are the Mechanisms of Credible Commitment? External independent enforcement and mutual monitoring are the most ob- vious mechanisms for ensuring credible commitments (Schelling 1960; Bates 1988), especially when paired with clear rules of election, appointment, and decision making. When all parties recognize such institutional arrangements as “the only game in townâ€? and when violations of the rules trigger sanctions, the effect is to stabilize expectations, reduce the uncertainties associated with the political process, and thus contribute to economic development. Chapter 1 has already introduced the technical and political-economic dimensions of independent central banks. As Concepts in Practice 8.1 shows, the temptation of elected officials to manipulate the economy before and after elections can help explain why central banks were made independent ­ in the first place. Other examples of institutions created as commitment devices include the following: • Independent courts of public accounts or supreme audit institutions • Independent electoral commissions • Civil service commissions. Independence on paper, however, does not automatically translate into de facto insulation from political pressure. Although the presence of such institutions does signal a certain degree of commitment of governments and politicians to not interfering with elections, bureaucracies, or public budget- 236 Understanding Policy Change CONCEPTS IN PRACTICE 8.1 The Rise of Independent Central Banks Since the end of World War II and especially political variables are good determinants of since the 1970s, a growing number of countries how fast a country adopted central bank inde­ have granted operational independence to their pendence. Countries with federal systems, in central banks, meaning that while politicians which political accountability and the attribution set inflation targets, the technocrats in the cen- of responsibility for economic policy making tral banks implement them. As already dis- were more fragmented, were among the earlier cussed in previous chapters, the case for cen- adopters. Unitary states, in contrast, in which tral bank independence is that elected officials attribution of inflation policy was clearer, were are prone to abuse monetary policy to make late comers. voters feel better off just before an election by In short, according to this study, federal in- overinflating the economy. Thus, setting the stitutions tend to dilute the clarity of macroeco- central bank free from day-to-day political inter- nomic tasks and potentially increase political ference is a way of avoiding this outcome. temptations to manipulate monetary policy. However, incentives generated by political Therefore, in such contexts, calls for credible institutions, and not by purely technical con­ commitment mechanisms were likely to be siderations, often explain the decision to com- more imperative and finally led to the earlier mit or not. As some empirical evidence shows, adoption of independent central banks. Source: Farvaque (2002). ing for political purposes, that commitment might not be especially credible in the absence of robust political opposition in legislatures or adequate checks and balances. There may be nothing to impede a powerful executive from interfering in the work of these institutional devices without incentives to refrain. This dynamic is problematic because it raises the question of how many institutional safeguards are needed to ensure real credibility. If independence on paper is not enough, what guarantees that the hands of the politicians are genuinely tied? What stops them from using the same institutions to tie their competitors’ hands? To quote the eloquent words of Norwegian philosopher Jens Arup Seip, “In politics, people never try to bind themselves, only to bind othersâ€? (Elster 2000, ix). The political-economy literature identifies several core mechanisms that ensure credible commitments. Discussion of some of these mechanisms follows. Constitutional Limits on Government Power. In 1688, following the Glorious Revolution in England, the establishment of Parliament as an insti- ­ tutional constraint on the power of the monarchy allowed the king to “com- Credible Commitment 237 mitâ€? to the repayment of debts and the security of property rights (North and Weingast 1989). Guardian Judiciary. A guardian judiciary is an independent body that can prevent the government from violating the rights of investors and asset holders (Landes and Posner 1975). A strong and independent judiciary may enable the government to credibly commit to repaying its debts, as the con- stitutional power of the judges might be greater than the political incentives the executive might otherwise have. “Hostageâ€? Mechanisms. A “hostage mechanismâ€? as a means of inducing credible commitments occurs when the government agrees to allow one of its assets to be seized by a third party if it fails to adhere to its promise to other actors (Williamson 1983). Some studies outline what is needed to cre- ate an effective hostage mechanism: (1) the value of the asset has to be great enough so that the government has significant financial incentives not to al- low its seizure; and (2) the asset needs to be held abroad so that the govern- ment has no control over the seizure (Haber, Razo, and Maurer 2003, 25). An example of a successful hostage mechanism comes from República Bolivariana de Venezuela and has to do with the management of the foreign Venezuela, the country’s state-owned petroleum com- assets of Petróleos de ­ pany.1 The foreign assets function as a hostage mechanism because they can be seized if the government overextracts revenue from the company and fails to ensure the rights of foreign investors (Haber, Razo, and Maurer 2003). As the loss of such assets would hurt politicians and exceed any popu- larity or rents they would have otherwise obtained from the seizure, these assets act as a credible commitment against expropriation. In addition, governments can also use a third party to enforce contracts and guarantee the rights of economic actors (Haber, Razo, and Maurer 2003, 26–32). However, the problem with implementing these rules or commit- ments in some countries (those that are resource dependent) is that they are often under authoritarian rule where a third-party mechanism would be in- effective and a constitution is unlikely to be upheld in the first place. Multiple “Veto Playersâ€? Systems. Another example of credible commit- ment mechanisms is illustrated in Concepts in Practice 8.2. This example shows how the number and configuration of veto players—different actors who have the power to block policy—can prevent sudden policy switches or reversals induced by shifting political incentives. The number of veto play- ers, however, comes at the expense of a potential shortcoming: the same institutional device that prevents policy shifts and reversals that could be 238 Understanding Policy Change CONCEPTS IN PRACTICE 8.2 The Importance of Veto Players in Policy Making “Veto players can be defined as individual or collective actors whose agree- ment is required for a change in the status quoâ€? (Tsebelis 1995, 289). Two types of veto players are very common in political systems: “institutional veto playersâ€? (presidents, chambers) and “partisan veto playersâ€? (such as political parties) (Tsebelis 2002). Generally, as the number of veto players (with noniden- tical interests) increases, the scope for policy change decreases: that is, the status quo—whether better or worse than a proposed alternative—is more likely to persist. For example, in the United Kingdom there are generally few veto players. The majority in Parliament can pass any legislation it likes without agreement from a strong executive or from any other major actor. Conversely, in the United States the president can veto legislation passed by Congress, and the Supreme Court can nullify (veto) legislation passed by Congress and ap- proved by the president. Source: Tsebelis (1995). detrimental to development might also be responsible for blocking pro- development change. In multiple “veto playersâ€? systems, achieving agreement over policy ­ reversals requires the approval of several political actors rather than only one (executive, legislature, regional representatives, the general population through a referendum, and so on). Institutionalizing checks and balances that place limits on policy makers can signal some policy stability in the long run to members of the opposition, to voters, or to private investors. Policy stability can lead to greater investments and potentially to better develop- ment outcomes. Reputational Mechanisms. Reputation has also been shown to fill several roles with implications for establishing the credibility of commitment (Kreps and Wilson 1982; Mazaheri 2008; Milgrom and Roberts 1982): • They often preserve contracts as well as other relationships of exchange between actors and firms, political parties and voters (North 1990; North and Weingast 1989). • Reputation encourages or discourages transactions or political candida- cies even before the actual occurrence (Greif 1994; Kollock 1994). • Reputational signals serve as a blueprint for day-to-day interaction among firms, agents, bureaucrats, and other actors. They lock in expectations Credible Commitment 239 that parties or governments will not renege on the initial policy promise under conditions of time-inconsistent behavior. Reputation is most useful as a mechanism of credible commitment in political-economic contexts with high volatility and risk, high information asymmetries, little third-party enforcement (Kreps and Wilson 1982), scarce capital, and in highly personalized transactions between exchange agents (informal economy) (Geertz 1978). But How “Credibleâ€? Is a Credible Commitment? Despite the conceptual appeal of “credible commitmentâ€? as a solution to the time inconsistency problems that actors face, the degree and “credibilityâ€? of a commitment that can be achieved through constitutions or third-party en- forcers have been heavily debated. Some scholars have argued that creating multiple institutional veto points is neither a necessary nor a sufficient con- dition for credibility (Stasavage 2002a, 2002b). Specifically, it has been sug- gested that for credibility of commitment to be established, control of veto points by different political parties and the possibility of cross-issue coali- tions are prerequisites. Accordingly, with the example of 1688 England in mind, some studies have traced the evolution of interest rates on government debt over time and found that a country’s credibility for not defaulting on its sovereign debt was not automatically triggered by the mere institutional existence of the Parliament, as an institutional veto point and a significant check on the power of the monarch. What seems to matter most for achieved credibility, in this particular account, was a cohesive cross-issue Whig coalition of government debt holders, characterized by a disciplined party organiza- tion and covering issues as diverse as religion, foreign policy, and finance. In contrast, the majority of land-owning interests represented in the Par- liament would have not cared (and were even often against) sovereign debt repayment. By the same token, even if the Bank of England had been cre- ated to lead to commitment, as a nonpolitical institution, by itself, it would not have led to credibility without the debt holders’ cohesive efforts to block any other actors’ interference with its apolitical nature (Stasavage 2002a, 2002b). In 18th century France, the credible commitment to repay debts was never achieved because government debt holders were never represented in the parliament. These two stories bring us back to the recurrent question of this hand- book: how does credible commitment help or hinder pro-development col- 240 Understanding Policy Change lective action? What is the relationship between the two concepts? In the case of Great Britain, according to this account, the Whigs used well- documented devices to enforce voting cohesion and party discipline be- tween 1688 and 1715. The collective action of this multi-issue coalition man- aged to enforce the credible commitment of debt repayment. In France, in contrast, the political coalitions that developed in the Assembly after the Revolution were highly unstable, and parties could not discipline their rank and file in pursuit of cohesive policies on government debt (Stasavage 2002b). In this instance, the collective action problem hindered the credibil- ity of government commitment. Credible Commitment, Collective Action, and Development The ability of groups to overcome collective action problems can be a criti- cal factor in determining the extent to which credible commitments can exist. Collective action–induced credible commitment can facilitate good governance and hence development in several important ways (Keefer 2011, 3–5): • Mitigating rent-seeking. The ability of citizens to remove low performance office holders depends on the existence of a credible alternative, accom- panied by effective electoral sanctions. Collective action is essential in ensuring that citizens can replace incumbents with challengers who can credibly commit to providing public goods, and therefore is the corner- stone for incentivizing rent minimization amongst policy-makers. • Altering political incentives. Politicians lack incentives to enact pro- development policies whenever citizens cannot act collectively to hold them accountable for reneging on their promises and diverting resources towards organized lobbies. • Affecting private investment. Private investment will decrease if individ- ual investors fear that government expropriation of their assets will not result in a collective response from other investors. • Restraining patronage. The absence of a collective action–induced credi- ble commitment can have implications for the provision of public goods. If political parties and individual politicians cannot commit to providing universalistic goods, such as health or education, then one possible out- come is the emergence of “patrons,â€? who act as intermediaries between politicians and citizens (Keefer and Vlaicu 2008). Credible Commitment 241 The Importance of Collective Action–Induced Credible Commitment to Development: What the Data Say One of the key puzzles in development is why the introduction of competi- tive elections in new democracies does not necessarily lead to less clien- telism, corruption, and rent seeking activities (Keefer 2011). Although stable, well-established democracies do exhibit significantly less corruption than other regime types, this is not the case with young democracies. One explanation for this puzzling phenomenon is that competitive elec- tions do not, in and of themselves, create incentives for political elites to minimize rents. As noted above, if candidates (especially those associated with new and less established party labels) cannot credibly commit to a set of programmatic or universalistic policies that benefit a wide range of citi- zens, then voters will not believe any politicians who pledge that they will pursue a good governance agenda. As Concepts in Practice 8.3 shows, because of a programmatic void, pa- tronage networks filled the postauthoritarian political space in the Domini- can Republic and led to inferior development outcomes (Keefer and Vlaicu 2008). Concepts in Practice 8.4 offers another example of how political parties’ lack of credibility in delivering public goods to all citizens affects a large number of developing countries. Many societies are deeply divided along CONCEPTS IN PRACTICE 8.3 Collective Action and Credible Commitment in the Dominican Republic The Dominican Republic emerged as a democracy after the elections of 1966, following the assassination of Rafael Trujillo in 1961 and the ousting of a demo- cratically elected leader by a military coup in 1962. The winner of these elec- tions, Joaquin Balaguer, managed to remain in office after two more elections in 1970 and 1974, not because he established a reputation for providing good value for taxpayers’ money but by building and exploiting the clientelistic net- work developed by Trujillo. This network included everything from the distribu- tion of bicycles to village children, to the doling out of public sector jobs. This clientelistic modus operandi did not end with the election of the opposition to power (1978 and 1982). Given that the credible commitment mechanism was still lacking, the opposition similarly relied on patronage networks to remain popular. 242 Understanding Policy Change CONCEPTS IN PRACTICE 8.4 Political Responses to the AIDS Epidemic and Ethnic Affiliations AIDS epidemics affect a large percentage of the population in many African countries. Yet some governments are more prompt than others in presenting the contagion as a public health issue and in implementing policies to contain it. Why? Recent studies have found that in countries with clearly defined ethnic cleavages, the HIV epidemic was not perceived as a shared threat to the nation but instead was framed as a threat to a specific group. In such circumstances, the response of the government to the epidemic was slow and indecisive. Irrespective of the ethnic group in power, institutions based on clear ethnic boundaries tended to shift the blame for the spread of the disease, creating social stigma and ultimately putting in place less effective policies for address- ing national pandemics. In contrast, in contexts with more blurred lines of eth- nicity, policy advances were more likely. South Africa and Brazil, two countries where the evolution of the pandemic was similar, illustrate this large difference in policy outcomes. Whereas the Brazilian government built bureaucratic capac- ity and a relatively autonomous agency to address HIV early on in the 1980s, the South African National AIDS Council, a comparable organization, saw the light only in the year 2000, and its authority is still not widely recognized (Gauri and Lieberman 2006; Lieberman 2009). ethnic, linguistic, or religious lines. Because of their weak commitment to delivering universalistic policies across such cleavages, politicians often send identity-based signals to voters and mobilize them along ethnic or religious issues. This tendency becomes a vicious circle, as deeply fragmented societ- ies fail to cooperate to provide public goods to the detriment of economic development (Varshney 2001; Chandra 2004; Miguel and Gugerty 2005). Modeling Credible Commitment as a Challenger Game2 To further help readers understand the implications of credible commit- ment to reform and development, this section will present two games that illustrate how credible commitment, or the lack of it, can affect outcomes. First, consider a case in which two players cannot credibly commit to a specific course of action. Say, for example, that an incumbent politician has two strategies he or she can pursue, given the existence of a challenger can- didate. The incumbent can either fight the election campaign (by raising and Credible Commitment 243 expending resources) or give in. Conversely, the challenger can decide either to enter the race or to stay out. In the normal-form simultaneous-move game below, neither actor can commit, a priori, to a specific course of action. However, given the assump- tion of complete information, the challenger candidate can anticipate that “give inâ€? strictly dominates (see definitions in chapter 2) “fightâ€? for the in- cumbent. Therefore, the challenger will enter the contest (the Nash equilib- rium is give in; enter). Enter Stay Out Fight 1 1 3 2 Give In 2 4 4 3 However, what if the incumbent can credibly commit to fighting before the challenger has a choice of whether to enter or not? Say, for example, the in- cumbent can spend his or her time in office cultivating links with supporters and engaging in preemptive campaign financing. This situation can be mod- eled as an extensive-form game (below). Here, the outcome associated with the subgame perfect Nash equilibrium3 is fight; stay out, because the incum- bent can anticipate that if he or she does not credibly commit to fight (in the first stage), the challenger will have an incentive to enter (in the second stage), obtaining a payoff of 4 instead of 3. However, if the incumbent fights in the first stage, the challenger will elect to stay out (challenger payoff of 2 versus 1). In such a setting, the incumbent is better off than before, because by credibly committing to fight, he or she gets a payoff of 3 instead of 2 (in the first game). Enter 1 1 â—? Fight â—? Challenger Stay Out â—? 3 2 â—? Incumbent Enter â—? 2 4 Give In â—? Challenger Stay Out â—? 4 3 Summary Many policy change initiatives require those involved in decision making to be able to credibly commit to certain actions. Why should citizens pay taxes 244 Understanding Policy Change if their governments cannot commit to not expropriating them as private rents? By critically evaluating and understanding the problem of credible commitment, as well as its relationship with the capacity for collective ac- tion of various groups, reformers can identify any commitment-related problems they may face and evaluate whether, in light of these problems, the project remains feasible. Notes 1. http://www.stanford.edu/class/polisci313/papers/MonaldiFeb04.pdf, 2–4. 2. Source: Inspired by (McCarty and Meirowitz 2006). 3. For simplicity and exposition purposes, the sub game perfect equilibria do not include off-path analysis. References Bates, R. H. 1988. Toward a Political Economy of Development: A Rational Choice Perspective. Berkley: University of California Press. Chandra, Kanchan. 2004. Why Ethnic Parties Succeed. Cambridge: Cambridge University Press. Diermeier, D., J. M. Ericson, T. Frye, and S. Lewis. 1997. “Credible Commitment and Property Rights: The Role of Strategic Interaction between Political and Economic Actors.â€? In The Political Economy of Property Rights: Institutional Change and Credibility in the Reform of Centrally Planned Economies, ed. David Weimer, 20–41. Cambridge: Cambridge University Press. Elster, Jon. 2000. Ulysses Unbound: Studies in Rationality, Precommitment, and Constraints. Cambridge: Cambridge University Press. Farvaque, Etienne. 2002. “Political Determinants of Central Bank Independence.â€? Economics Letters 77 (1): 131–35. Frye, T. 2004. “Credible Commitment and Property Rights: Evidence from Russia.â€? American Political Science Review 98: 453–66. Gauri, V., and E. S. Lieberman. 2006. “Boundary Institutions and HIV Policy in Brazil and South Africa.â€? Studies in Comparative International Development 41 (3): 47–73. Geertz, C. 1978. “The Bazaar Economy: Information Search in Peasant Bargaining.â€? American Economic Review 68 (2): 28–32. Greif, A. 1994. “Cultural Beliefs and the Organization of Society: A Historical and Theoretical Reflection on Collectivist and Individualist Societies.â€? Journal of Political Economy 102 (5): 912–50. Haber, S. H., A. Razo, and N. Maurer. 2003. The Politics of Property Rights: Political Instability, Credible Commitments, and Economic Growth in Mexico, 1876–1929. Cambridge: Cambridge University Press. Kasara, K. 2007. “Tax Me If You Can: Ethnic Geography, Democracy, and the Taxation of Agriculture in Africa.â€? American Political Science Review 101 (1): 159–72. Credible Commitment 245 Keefer, Philip. 2011. “Collective Action, Political Parties, and Pro-Development Public Policy.â€? Asian Development Review 28 (1): 94–118. Keefer, Philip, and Razvan Vlaicu. 2008. “Democracy, Credibility, and Clientelism.â€? Journal of Law, Economics, and Organization 24 (2): 371–406. Kollock, P. 1994. “The Emergence of Exchange Structures: An Experimental Study of Uncertainty, Commitment, and Trust.â€? American Journal of Sociology 100 (2): 313–45. Kreps, D. M., and R. Wilson. 1982. “Reputation and Imperfect Information.â€? Journal of Economic Theory 27 (2): 253–79. Landes, William M., and Richard A. Posner. 1975. “The Independent Judiciary in an Interest-Group Perspective.â€? Journal of Law and Economics 18 (3): 875–901. Lieberman, E. 2009. Boundaries of Contagion: How Ethnic Politics Have Shaped Government Responses to AIDS. Princeton: Princeton University Press. Mazaheri, Nimah. 2008. “An ‘Informal’ Revolution: State-Business Conflict and Institutional Change in Iran.â€? Middle Eastern Studies 44 (4): 585–602. McCarty, N., and A. Meirowitz. 2006. Political Game Theory. Cambridge, MA: Cambridge University Press. Miguel, Edward, and Mary Kay Gugerty. 2005. “Ethnic Diversity, Social Sanctions, and Public Goods in Kenya.â€? Journal of Public Economics 89 (11/12): 2325–68. Milgrom, P., and J. Roberts. 1982. “Predation, Reputation, and Entry Deterrence.â€? Journal of Economic Theory 27 (2): 280–312. Monaldi, F. 2001. “Sunk-Costs, Institutions, and Commitment: Foreign Investment in the Venezuelan Oil Industry.â€? Unpublished manuscript, Stanford University, Department of Political Science. North, D. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. North, D. C., and B. Weingast. 1989. “Constitutions and Credible Commitments: The Evolution of the Institutions of Public Choice in 17th Century England.â€? Journal of Economic History 49 (4): 803–32. Schelling, T. C. 1960. The Strategy of Conflict. Cambridge: Harvard University Press. Stasavage, David. 2002a. “Private Investment and Political Institutions.â€? Economics and Politics 14 (1): 41–63. Stasavage, David. 2002b. “Credible Commitment in Early Modern Europe: North and Weingast Revisited,â€? London, LSE Research online, http://eprints.lse.ac. uk/227/1/credible_commitment.pdf. Tsebelis, George. 1995. “Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism, and Multipartyism.â€? British Journal of Political Science 25 (3): 289–326. ———. 2002. Veto Players: How Political Institutions Work. Princeton, NJ: Princeton University Press. Varshney, Ashutosh. 2001. “Ethnic Conflict and Civil Society: India and Beyond.â€? World Politics 53 (3): 362–98. Williamson, O. E. 1983. “Credible Commitments: Using Hostages to Support Exchange.â€? American Economic Review 73 (4): 519–40. 246 Understanding Policy Change CHAPTER 9 Solutions to Collective Action Problems Collective action is the engine of change. Reform occurs when individuals and groups act together, share information, and keep politicians and bureau- crats accountable. Pro-development mobilization is not easy, and institu- tional constraints often act against reform. Lack of information, motivational factors, missing mechanisms for credible commitment, delegation problems in organizations with unaccountable leaders, and institutional manipulation against policy reform all stifle attempts to effectively mobilize and change the status quo. Fortunately, despite grim predictions, pro-development collective action does happen. Groups manage to mobilize in spite of formidable obstacles. Sometimes they make history. Mohamed Bouazizi, a street vendor in a rural town in Tunisia, became the catalyst of the previously unthinkable Arab Spring; the city of Porto Alegre in Brazil has a vibrant civil society that ac- tively monitors municipal spending and has built more schools in 10 years than were ever thought possible; in the Indian state of Uttar Pradesh, a low- caste political party significantly increased the electrification rate of vil- lages; and a coalition of bureaucrats, nongovernmental organizations (NGOs), youth groups, and leading politicians took advantage of institu- 247 tional opportunities and reformed a corrupt system of public procurement in the Philippines. How do good results come about? The honest answer is that nobody re- ally knows why good things happen despite unfavorable odds, but we can do our best to look at a repertoire of classic solutions to collective action prob- lems that have been identified in a very rich and diverse body of knowledge that cuts across all social sciences. Let us see what it has to offer future change agents. Chapter 3 introduced the logic of the collective action problem and its major implications: the undersupply of public goods, the overuse of common resources, and the general inability of large groups in society to formulate claims for a better development outcome. We have also explored many modes of collective action that we encounter every day: social movements, political parties, interest groups, coalitions, and reform teams. This chapter will wrap up the theoretical section of the handbook by ex- amining in more detail specific ways in which collective action problems can be solved. The first section will explore some of the organizational and coali- tional characteristics, as well as effective institutional arrangements that build and maintain collaborative efforts for public goods. The second section will deal with various monitoring and sanctioning mechanisms, trust, dis- trust, and selective incentives, and their effect on governance and develop- ment outcomes. Finally, the chapter will introduce the concept of leadership as a potential solution to collective action problems. The chapter establishes a basis for helping change agents analytically sort through some of the ways out of key dilemmas that would otherwise hamper the realization of their agenda (see figure 9.1). By improving information flows, cultivating trust and communication among the members of a group or organization, and using different institutional mechanisms of monitoring and sanctioning, reformers might be able to overcome the problems that pre- vent the provision of public goods and effective pro-development action. Objectives of Chapter 9 By the end of chapter 9, readers should be able to do the following: • Understand how different mechanisms can help resolve collective action problems. • Explore organizational and coalitional features that lead to successful pro-development action. • Think about potential institutional arrangements that govern effective joint efforts in development. 248 Understanding Policy Change FIGURE 9.1  Conceptual Map: Solutions to Collective Action Problems Monitoring and Conditions for sanctioning effective monitoring Selective Solutions to incentives collective action Sources of trust problems Trust-building Consequences of trust Leadership Source: Authors. • Appreciate the importance of different monitoring techniques, trust building, selective incentives, and leadership, as well as their potential role in resolving collective action problems. • Be able to link concepts of collective action solutions to real-life situations. When Do Mobilization and Production of Public Goods Occur? We know that many successful cases of collective action have taken place despite the chronic misalignment between individual and group incentives postulated by theoretical work. Why and under what conditions does align- ment happen? How do communities come together to regulate the use of common resources? How do individuals succeed in effectively mobilizing for policy change in high-risk political-economic environments? What roles do mutual trust, face-to-face interaction, and reciprocity play in facilitating successful collective action? How can organizations monitor and sanction noncompliant behavior of members, correct informational discrepancies, and delegate crucial coordination tasks to leaders? The three central dilemmas of cooperation described in chapter 3—the prisoner’s dilemma, the tragedy of the commons, and the logic of collective action—have been challenged on both theoretical and empirical grounds. Let us briefly review some of the reasons for optimism. Solutions to Collective Action Problems 249 Group Size Matters . . . but Not Always Chapter 3 argues that the size of the group matters since it is much more dif- ficult to pool risks, preferences, and organizational costs among many citi- zens as opposed to just a few. Large groups of diffuse citizens are less likely to solve collective action problems than small, more concentrated business lobbies. Theoretically, however, some researchers have shown that the large number does not always hamper collective action, as long as a subgroup of individuals cares intensely about providing a public good, even if all other members are free riders (Hardin 1971, 1995). Smaller privileged subgroups (sometimes referred to as k-groups to indi- cate that they include only k members out of a total of n group members, with kb and b>c), this is defined as having complete preferences; and (2) that these c ­ omplete preferences must be transitive (if a>b and b>c, then a>c). For example, citizens and politicians, incumbents and challengers, competing firms, etc. 317 • Actions available to each player. For example, confess/don’t confess, cooperate/don’t cooperate, or work hard/shirk. • Strategies. This is the set (potentially a combination) of actions a player can pursue given how she expects or knows other players will react to her behavior. For example, a bureaucrat’s strategies may include the follow- ing: (1) engage in increased rent-seeking activities if she anticipates that politicians will not be able to observe her behavior; or (2) minimize her rent-seeking activities if she anticipates that politicians will detect such activities and punish her for her actions. • Rules of Interaction. Players do not usually operate in an institutional vac- uum. Context matters as, for example, it determines what information players have, whether they make decisions simultaneously, whether they can communicate with each other, etc. • Outcomes or results. For example, one possible outcome of a bureaucrat’s pursuing a rent-seeking agenda is that her actions are detected and she loses her job. • Payoffs. These are the utility accrued by each player as a result of each possible outcome. In other words, payoffs are the value or utility that players gain from different outcomes. The utility of an outcome depends on the costs and benefits it generates. Specifically, an outcome will be more favorable the larger the gap between the benefits minus the costs. Benefits and costs may be a variety of things—money, votes, fame, etc. Payoffs can be represented in different ways. When faced with a set of (discrete) alternative actions, as in table A.1, one simple way of represent- ing the desirability of payoffs is to rank them in order, say from 1 to 4 (if there are four discrete choices). It is important to note that such ordinal rankings do not convey the magnitude of the difference between the al- ternatives but only their ordering. Thus, in the table below 4,000 votes is preferred to 3,000 votes, which is preferred to 2,500 votes, although the gap between these alternatives differs (3,000 is 1,000 less than 4,000 while 2,500 is 500 less than 3,000).  For example, a bureaucrat may receive a higher payoff if she pursues a rent-seeking strategy and is not detected (keeping her job) than if she pursues a rent-seeking strategy and is detected, resulting in her losing her job and facing criminal proceedings.  In the case of the Philippines procurement reform example, the payoffs to the general public of the passage of the bill were positive (less of their tax money was stolen). For corrupt officials, however, the payoffs are neg- ative, as the bill would lead to fewer possibilities to generate illicit income from corrupt procurement dealings. 318 Understanding Policy Change TABLE A.1 An Illustration of Payoffs  Payoff (ordinal ranking) Benefit-cost Example (votes of a candidate) 4 Greatest 4,000 votes 3 3,000 votes 2 2,500 votes 1 Smallest 1 vote Game Theory and Party Election Platforms. Let’s illustrate the power of this solution concept by using a very simple example. The diagram below represents a normal form game in which two players can act simultaneously. In this case, let us assume the players are two political parties competing for office. Each party can decide on its own electoral platform. In this case, a party can run on either a promise to focus on domestic issues or a promise to engage in a pointless war with a friendly neighboring state. Given that the electorate would never vote for a party that promised to embark on a ruinous war, both parties will avoid this issue. To see this, let’s examine the payoffs for the first party (red). Knowledge of electoral prefer- ences allows the parties to rank their electoral strategies based on how they will appeal to voters, given the other party’s behavior. The optimal outcome (vote maximization) of a party is that it pursues the most popular policy and the other party pursues the least popular campaign strategy (highest pay- off ); the second most favorable outcome is that both parties pursue a popu- lar campaign strategy; the third best alternative is that both parties pursue an unpopular campaign strategy; and the least favorable outcome, for any given party, is that it pursues an unpopular campaign strategy while the other party pursues a popular strategy. Let’s say the first party (light grey) believes that the second party (dark grey) will pursue a domestic electoral campaign. In this case, the first party would obtain a payoff of 3 if it also pursued a domestic campaign and a payoff of 1 if it pursued a war-focused campaign. If, alternatively, the first party be- lieved the second party would pursue a war-focused campaign, it would ob- tain a payoff of 4 if it pursued a domestic campaign and a payoff of 2 if the party also pursued a war-focused campaign. In short, regardless of what strategy the second party chooses, the first party is always better off pursu- ing a domestically f ­ ocused agenda (3> 1 and 4>2). Therefore, pursuing a do- mestic-focused campaign strictly dominated pursuing a war-focused cam- paign for the first party independently of the platform choice of the second party. Given that the preferences of the second party are identical to those of Appendix A: Fundamental Concepts in Game Theory 319 the first (a domestic-focused campaign strictly dominated a war-focused campaign), it is therefore possible to use the iterated elimination of strictly dominated strategies to find the unique solution to the game: namely, that both parties will pursue a domestically focused agenda. Domestic War Domestic 3 3 4 1 War 1 4 2 2 Reference Game Theory Net. 2012. A resource for educators and students of game theory. Owen Graduate School of Management, Vanderbilt University: Nashville, Tennessee, http://www.gametheory.net/. Rapoport, A. (ed.). 1974. Game Theory as a Theory of Conflict Resolution. Boston: D. Reidel Publishing Company. Shor, M. 2012. Dictionary of Game Theory Terms, Game Theory.net, Smith S. 2003. “Game Theory.â€? Beyond Intractability. Eds. Guy Burgess and Heidi Burgess. Conflict Information Consortium, University of Colorado, Boulder: Posted August (2003) < http://www.beyondintractability.org/bi-essay/prisoners- dilemma > 320 Understanding Policy Change APPENDIX B Technical Appendix Nash Equilibrium Refers to a situation in which an action undertaken by a player is optimal given the actions of every other player. No player has an incentive to deviate from her course of action without the others doing so (Osborne 2003, 21). Weakly Dominated Strategy Refers to any strategy that yields payoffs at least as great as any alternative strategy for a specific player. Such strategy results in a higher payoff for at least a subset of the strategy profiles of other players (Shor 2006a). Strictly Dominated Strategy Refers to a strategy which yields a higher payoff for a player independently of the actions of any other players in the game (Shor 2006b). 321 References Osborne, M. 2003. Introduction to Game Theory. Oxford: Oxford University Press Shor, M. 2006a. ‘Weakly Dominated Strategies,’ Dictionary of Game Theory Terms, Game Theory.net, Shor, M. 2006b. ‘Strictly Dominated Strategies,’ Dictionary of Game Theory Terms, Game Theory.net, Shor, M. 2012. “Dictionary of Game Theory Terms.â€? Game Theory.net, http://www. gametheory.net/dictionary/url_of_entry.html 322 Understanding Policy Change APPENDIX C Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines Corruption has been identified as one of the major factors that can hinder development. Given a weak rule of law, the environment becomes more hos- pitable to corruption, thus significantly affecting economic growth and dis- tributive outcomes. The following case study summary describes a public procurement reform in the Philippines that began in 1998 and was approved in 2003 (Campos and and Syquia 2005). This case illustrates how, in a highly complex political-economy environment, a group of reformers was able to form a coalition to advocate for policy change in the crucial area of public procurement transparency. Note: This is a summary of the World Bank Policy Study by Jose Edgardo Campos and Jose Luis Syquia, Managing the Politics of Reform: Overhauling the Legal Infrastructure of Public Procurement in the Philippines (Washington, DC: World Bank, 2005). 323 The story begins when a group of policy entrepreneurs recognized that the weak rule of law and poor governance in the Philippines prevented growth and development. To address this challenge, the group identified a concrete reform goal: they would overhaul the regulations and operational procedures related to public procurement. Reaching this goal entailed over- coming many political-economy obstacles. The team first had to acquire a comprehensive understanding of the operational context. Then it had to create feasible and effective strategies for bypassing opposition. Indeed, their understanding of context enabled their successful strategy design; in turn, the strategies minimized the bargaining leverage of vested interests, allowing a comprehensive bill to pass in the legislature. This reform presents a vivid example of how political-economy concepts, key identified constraints, and strategic solutions apply in practice. Given the richness of this case, the following summary will illustrate how the theo- retical building blocks discussed throughout the handbook affect real-life scenarios and how using some of the existing political-economy tools can help change agents’ understanding of their context and develop feasible and effective strategies to address constraints and achieve positive results. Identifying the Problem According to Transparency International, in the early 1990s the Philippines was among the most corrupt countries of East Asia. Corruption took many forms, flourishing in many sectors of the country; procurement was one of the venues where corruption thrived (Social Weather Station, 2002). With con- tradictory procurement regulations that allowed opportunities to award cor- rupt contracts, companies winning public contracts were rarely the most ef- ficient, qualified, or cost effective. By paying bribes to government officials, a company could earn a contract and deliver costly but poor services to the people. As a result, the general public received suboptimal services and goods. Timeline of the Reform Passage Reform of government procurement in the Philippines involved multiple actors and occurred over a period of at least five years (see figure C.1) (Cabañero-Verzosa and Garcia 2009). The following sections integrate political-economy concepts to describe how passage of reform legislation was ultimately achieved. 324 Understanding Policy Change FIGURE C.1  Five-Year Timeline of the Philippines Procurement Reform July: August: Boncodin August: BRTF October: revives the October: Analysis is organizes President GPPB and September: Diokno conducted workshops to Estrada TWG; Procurement requests a and draft develop a new sends bill to February: procurement Watch procurement bill is proposal and Congress as PWI is work begins strategy is analysis proposed get buy-in urgent created once more launched 1998 1999 2000 2001 2002 2003 August: October: May: January: December: January: 11th Congress, Corruption President 12th Congress Lower President Arroyo Lower House scandals Estrada steps starts with House and signs bill, making Committee of Public about down and new Senate it a law Works endorses bill; President Arroyo configurations approve bill Lower House approves Estrada becomes bill and sends to Senate; interim President Estrada’s impeachment dominates agenda and the bill dies Graph adapted from Cabañero-Verzosa and Garcia 2009, 85. Round One In 1998, Benjamin Diokno saw a window of opportunity for a reform path. As head of the Department of Budget and Management (DBM), Diokno under- stood how the existing rules and laws governing the procurement processes created opportunities for clientelistic bids. Determined to work toward a so- lution, he began by requesting the help of the U.S. Agency for International Development (USAID) in conducting an analysis to better understand the procurement system. Based on this study, the analysts would draft a bill to streamline the public procurement system. With much enthusiasm, USAID hired two consultants to develop a rigor- ous analysis and draft a bill proposal. Soon a complication arose. Both con- sultants had previously worked in countries where top-down leadership was the most effective method of change. In the Philippines, by contrast, in- dividualism is very strong. Because different agencies and bodies have differ- ing incentives, it is difficult for them to come together with one voice. Per- haps based on prior experiences, the consultants failed to engage the DBM staff during the development phase of the analysis, which generated discon- Appendix C: Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines 325 tent and lack of support. In great need of buy-in, Diokno urged his Budget Reform Task Force (BRTF), and later the Technical Working Group (TWG), to organize a series of workshops where staff could critique the analysis and create their own draft of a procurement law. As BRTF expected, the work- shops provided opportunities for staff and bureaucrats to participate and created a sense of ownership and support for the bill. This approach, used frequently throughout the entire reform effort, coincides with the chapter 7 strategies for solving information asymmetries and generating trust among coalition members (chapter 9). The new collaborative draft was sent to President Joseph Estrada, who had a variety of incentives to support this bill. First, his campaign empha- sized budget reforms and transparency. Second, he was facing corruption accusations that were increasingly threatening his incumbency. Presumably hoping to demonstrate a credible commitment (see chapter 8) to campaign promises in advance of the election, scheduled only five months away, the president gave legislative priority to the bill. After gauging the president’s interest in supporting the bill, the Budget Department’s Legislative Liaison Office identified the congressional commit- tee that would have the capacity and will to work towards reform. The team found the perfect match for the job and sent the draft to the Committee on Public Works. This decision stemmed from the fact that Congressman Nep- tali Gonzalez II, a member of the Committee, had previously initiated a re- lated bill on government contracting regulations and sanctions for noncom- pliance. Acknowledging the similarities of both versions, Congressman Gonzalez II agreed to sponsor the legislation and substitute it for his original bill. Because this original draft was already scheduled for debate, the reform team was not only able to find a strong supporter, but was also able to insert the bill for prompt debate. It is also important to note, in line with our obser- vations in Chapter 5, that during the initialization phase, the reform team kept a relatively low profile in order to prevent vested interests supporting the status quo from blocking their efforts. Time dynamics were an important part of this decision, with reformers using windows of opportunity (see chapters 4 and 11) to incentivize politi- cians to commit. Specifically, the draft discussion coincided with highly pub- licized calls for the president’s impeachment on allegations of corruption. Essentially, the team hoped that the bill’s anticorruption flavor would moti- vate legislators to support the measure to increase their popularity and cred- ibility with the general public. In addition, political competition generated by the upcoming election made politicians eager to distance themselves from the unpopular president by calling for reform. (Chapter 4 discusses the influ- ence of political competition on the likelihood of institutional change.) Con- 326 Understanding Policy Change sequently, the proposed measure had immediate acceptance in the commit- tee. Now that the influential Committee on Public Works had accepted to sponsor the bill, the next step entailed finding the right time to bring it to the floor of the Congress and secure approval. Congressman Gonzalez II, a sea- soned agenda setter, was able to significantly influence the scheduling of the bills for floor debate, and waited until most opposing legislators were absent to introduce the Procurement Reform in Congress. This strategic move pres- ents a powerful illustration of the agenda-setting mechanism discussed in chapter 5: reformers waited for the right time and took advantage of institu- tional rules governing voting in order to move forward. In this concrete case, agenda setting allowed for the bill to receive unanimous approval. Once Congress had approved the bill, the draft went to the Senate, where majority floor leader Francisco Tatad worked to identify another committee willing to sponsor it. This time, in the second chamber, the senators were less eager to work on the bill, as the presidential impeachment trial as well as their individual incentives to survive electorally by credibly dissociating themselves from corruption allegations loomed high on the horizon. As a re- sult of lack of interest in the legislation, the bill died in the Senate. This meant that during the next congressional session, the reform project would have to start from scratch to gather approval from both Lower House and the Senate. Figure C.2 details the involvement of various actors in the first round of legislation. Round Two Significant changes occurred from the time of lost momentum until the be- ginning of the new Congressional session. In May 2001, President Estrada resigned. Vice president, Gloria Macapagal-Arroyo succeeded him as the in- terim president of the Philippines. The Lower House and the Senate reshuf- fled most influential positions. With these tectonic shifts in the background, Emilia Boncodin was now asked to serve as the DBM’s secretary. Boncodin saw her appointment as a renewed opportunity to further the procurement bill, and thus asked undersecretary Laura Pascua, to manage the reform pro- cess. As a result, Boncodin and Pascua worked jointly to reactivate the Gov- ernment Procurement Policy Board (GPPB) and to prepare an executive or- der that would consolidate the desired changes with a set of new rules governing consulting services. This strategy aimed at increasing support for the bill, especially from those who were promoted by the new administra- tion. The team also decided to revive the Technical Working Group (TWG). After two months of intense weekly meetings held by the TWG, the GPPB approved the new draft. On October 8, 2001, and following GPPB’s approval, Appendix C: Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines 327 FIGURE C.2  Round 1 Net Map: Who Influenced the Passage of the Procurement Reform Law? Legisla- ture + Office of the – Govern- President (Joseph ment Estrada) Congressman linked to Civil Contractors Society + + + + Donors BRTF Head of DBM Chair, Chair, and (Diokno) Lower Senate TWG + House Private Sector USAID DBM’s Lower House Senate + Legislative liaison office + – + – + Supporters + – Neutrals – Opponents + – + – + – + – Financial Link General TV Radio Media Formal Authority Public Informal Authority Source for the graphic tool: NetMap, Schiffer 2012. interim president Arroyo, who signaled commitment to improving gover- nance in the Philippines, signed Executive Order 40. The effective delegation of tasks that is apparent in this round—particu- larly with regard to the various bureaucratic units, the DBM and GPPB—can be partially explained by a greater alignment in policy preferences between President Arroyo and the Lower House, than under the leadership of Presi- dent Estrada. The work of the DBM helped create this preference alignment, which in turn translated into expanded delegation to the agency (see chapter 6 on the principal-agent theory). In other words, civil servants at the DBM, as reform actors, correctly identified the changing political space and were able to react effectively. Several chapters of the handbook suggested that monitoring and sanction- ing are essential solutions to collective action problems, correctors of infor- mation asymmetries in principal-agent relationships, and effective mecha- nisms of credible commitment. Neutral third parties can often fulfill effective 328 Understanding Policy Change roles of ‘fire alarms’ or ‘police patrols’ that make sure that ex-ante policy commitments are not reneged upon. Let’s see what happens in the procure- ment reform story. While passing the bill was a major accomplishment, the BRTF and the Technical Assistance Team were aware of the important role that a nongovernmental organization would play in monitoring implemen- tation and compliance with the law. To this end, Procurement Watch, Inc., established in 2001, had the main mission to “fight corruption in public pro- curementâ€? (Campos and Syquia 2005, 6). To develop credibility and build a platform among key stakeholders, the organization invited respected civil society representatives to become part of the Procurement Watch Board. Its staff became highly involved in the de- liberations of the new bill. Eventually, Procurement Watch, in addition to playing the role of a third party monitor, also became an important vehicle for solving collective action problems among civil society representatives and bureaucrats interested in reform, stemming from missing information and lack of credibility (see chapters 3 and 9). Collective action was realized through Procurement Watch’s interaction with stakeholder groups and the media (see table C.1). Getting the omnibus procurement law passed required sufficient public support for the bill. Pro- curement Watch brought together the support of key civil society organiza- tions and increased the salience of the issue related to corruption in public procurement through various media platforms: print, radio, and television. TABLE C.1  Stakeholder Map for the Philippines Procurement Reform Benefits from Intensity of Incentive for mobilization  Stakeholder reform? Resources preferences (resources × intensity) Government   reform teams Yes High High High Donors Yes Medium High Medium Investigative media Yes Medium Medium Medium Big contractors Yes High High High NGOs and CS Yes Low Medium Medium General public Yes Medium-low Medium-low Low General media Yes Medium Low Low Church Yes Medium Medium Medium Legislators linked   to contractors No High High High Source: Authors. Appendix C: Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines 329 Why this strategy? Chapter 7 discussed the example of increased govern- ment responsiveness to flood damage triggered by media exposure. It also suggested that media impacts political incentives on a two-way street. On one hand, an informed public is likely to formulate demands for greater ac- countability; on the other hand, increased issue salience and visibility at- tracts politicians because of electoral credit taking opportunities. In this case, Procurement Watch skillfully appealed simultaneously to both targets: the general public and individual politicians. The organization worked to educate civil society about the importance of the new legal changes and helped gain public support from major stakeholders such as youth (students and associations), the Catholic Church, private sector groups represented by chambers of commerce, media outlets, and civil society organizations (in- cluding the Transparency and Accountability Network). As a result, attempt- ing to take advantage of increased visibility that media coverage could bring, individual politicians hurried to associate themselves with the procurement measure These partnerships brought enough public support (mobilization) to pressure lawmakers into considering the bill. The reform team had to think strategically about its next steps, under conditions of a shifting political landscape in both Chambers of Congress. During the new administration, the two chambers of the legislature had a more balanced power distribution, but opposition was much more unified (see chapter 4). Reformers approached legislators from both major compet- ing parties, the National People’s Coalition Party and the Philippine Demo- cratic Party, and convinced an influential senator to sponsor the bill jointly. This strategic arrangement allowed the media team to portray the bill as a bipartisan effort and successfully minimized the risk of interparty conflicts that would hinder the passage of the reform (Campos and Syquia 2005, 23). During the opening debates taking place in the Lower House, Speaker Jose de Venecia firmly endorsed this version of the bill. Then, he entrusted the project to his colleague Congressman Rolando Andaya Jr., chairman of the Appropriations Committee. Congressman Andaya immediately decided to assign senior bureaucratic staff to the secretariat of the Congressional Technical Working Group, and requested that the technical assistance (TA) team be incorporated in the secretariat as well. After numerous work ses- sions, the Congressional Technical Working Group developed a draft sub- mitted to the Lower House on May 16, 2002. Once again, Congressman Gonzalez II was now waiting for the right time to introduce the bill to Congress (see chapters 5 and 11). Because several key legislators strongly opposed the reform, Gonzalez decided to wait five months to schedule the debate for a day when a majority of the quorum of the legislators in attendance favored the bill. His agenda setting strategy was again successful and the bill passed with a large majority. 330 Understanding Policy Change Three different senators sponsored several versions of the bill in the Sen- ate. To aid the passage, the TA team and the BRTF sought Senator Edgardo Angara’s endorsement, given his status as a powerful member of the opposi- tion and the leader of the legislative caucus. With the help of Senator Anga- ra’s chief of staff and chief policy adviser, the TA team arranged a meeting between Senator Angara and Secretary Boncodin. The meeting was highly successful, resulting in Senator Angara’s endorsement. Senator Angara as- signed his staff to work with the three entities—the TWG, the TA team, and Procurement Watch—to craft a Senate version of the new legislative project. He then submitted a comprehensive report which covered the new Senate Bill and attracted considerable media attention. Two months later, the Sen- ate floor debated the bill, voted and approved it. Figure C.3 details the in- volvement of various actors in the second round of legislation. FIGURE C.3  Round 2 Net Map: Who Influenced the Passage of the Procurement Reform Law? Legisla- ture + – Govern- + Congressman Office of the ment linked to President (GMA) BRTF Contractors Civil Society + + + + Head of DBM Chair, Donors Chair, (Emilia USAID Lower Boncodin) House Senate Private + + Sector TA and TWG Usec. DBM Lower Team (Laura Pascua) House Senate + – + – + Procurement Watch + Supporters + – Neutrals – Opponents + + + – + + Financial Link Youth TAN Radio Media Students Formal Authority Informal Authority + + – + + Investigative General Church TV Public Media Source for the graphic tool: NetMap, Schiffer 2012. Appendix C: Political-Economy Concepts in Practice: Public Procurement Reform in the Philippines 331 Furthermore, given that two versions of the bill adopted in the two cham- bers required reconciliation, the Bicameral Conference Committee con- vened to consolidate them and negotiate differences. The resulting version was approved by both houses, with an added provision that emphasized implementation oversight for the law for an initial period of five years. This final version of the bill was signed on January 10, 2003 by President Arroyo. Thus, after a long saga, the Consolidated Procurement Reform Bill finally became law. References Cabañero-Verzosa, C., and H. Garcia. 2009. Building Commitment to Reform through Strategic Communication: The Five Key Decisions. Washington, DC: World Bank. Campos, J., and J. Syquia 2005. Managing the Politics of Reform: Overhauling the Legal Infrastructure of Public Procurement in the Philippines. Washington, DC, World Bank. Schiffer, Eva. “Net-Map Toolbox. Influence Mapping of Social Networks.â€? Available at: http://netmap.wordpress.com/. Last modified on May 9, 2012. Social Weather Station. 2002. “Survey of Enterprises on Public Sector Corruption.â€? Presentation at the Asian Institute of Management in Manila, Philippines, Intercontinental Hotel, March 21. 332 Understanding Policy Change APPENDIX D Political-Economy Homework Exercises This homework is designed to reinforce the learning outcomes from the classes (especially with respect to coalition building), provide a benchmark on how to write a case study, and facilitate teamwork. You will be randomly assigned into groups of five to complete the home- work. It is important that you work on these exercises together. You will each be given a “character card.â€? During some of the homework exercises, you will use the information on these cards to role play to illustrate the com- plexity of the coalition-building process. When the homework does not use the role-playing cards, you can assume that you are not trying to act in char- acter. The materials needed include group instructions and character cards. Please note: Homework should be distributed after each relevant exercise. Do not distribute all homework together as this may affect the learning outcomes. 333 Exercise 1: The Commonwealth of Rationaltopia Group Briefing: Read this information as a group. Then each one of you should read your individual card for 2 minutes. After doing so, go around the table and individually discuss your character’s policy preferences. Once this process is complete, fill in the answer sheet as a group. The Commonwealth of Rationaltopia is a developing country, which has been independent for the last 60 years. Since independence, the common- wealth has been characterized by the following: 1. Structural changes in government. Within only a few decades, the com- monwealth has shifted from a parliamentary form of government, to a presidential form of government, to a semipresidential form of govern- ment, and now back to a full presidential system. The current president sees himself (or herself ) as the leader of the nation and the region and is very keen to pursue a proactive defense and foreign policy. The president does not want to be entangled with the “pettyâ€? considerations of domes- tic politics. 2. Clientelism in elections. The parliament is hampered by infighting and deadlock. Each legislator seeks to provide supporters with benefits in the form of tangible material handouts throughout the electoral cycle. All seem to care about being reelected, but none seems interested in making long-term policy to benefit anyone beyond their supporters. 3. An insulated bureaucracy. In contrast to the legislature, the bureaucracy is a source of stability. While retention of bureaucrats has ensured that someone is responsible for enacting national policy, this insularity has also resulted in allegations of gross inefficiency and corruption. 4. Divided civil society organizations. While there are some very dynamic or- ganizations in civil society, they tend to be split geographically (and there- fore linguistically and religiously). Many organizations operate in either the south or the north of the country but not in both regions. 5. Politicized private sector. Although the developing status of the Rational- topia means that there are some booming businesses (e.g., tourism, min- ing, and agriculture), other businesses, especially small businesses with- out political connections, are struggling to establish themselves. In particular, some members of the business community argue that political connection and not business acumen determines everything from how long it takes to get a business license to who gets selected for government procurement services. 334 Understanding Policy Change Participant 1: Character Card A You are president of Rationaltopia. You have always known that you were destined for greatness. You feel that, no matter what domestic sacrifices have to be made, it is your duty to the people of Rationaltopia to maintain a strong military and a proactive foreign policy. When it is your turn to speak, introduce yourself as the president of Rationaltopia. Try to convince everyone that your vision is sound and that ­ the long-term benefits of pursuing an active foreign policy—such as security and good relations with neighboring states—are in everyone’s interest. Participant 2: Character Card A You are a member of parliament (MP). You are fed up with everyone blaming the inability of elected politicians to get along and develop coherent policy. You work very hard to make sure your supporters receive any benefits you can deliver for them. After all, isn’t that your role—to look after the people who elected you? When it is your turn to speak, introduce yourself as a member of parlia- ment. Try to convince everyone that MPs, such as yourself, are doing a good job in looking after the people who voted for you. It is unfair to be branded irresponsible just because you don’t want to divert resources away from your supporters. Everyone in Rationaltopia has the vote, so if other people want something, why don’t they ask their MP instead of blaming MPs and parlia- ment as a whole? Participant 3: Character Card A You are a senior civil servant. During your career you have been responsi- ble for ensuring that Rationaltopia has the expertise to develop and imple- ment public policy. Everyone considers the bureaucracy to be insular, inef- ficient, and somewhat corrupt. No one seems to appreciate the fact that without it, all those short-termist members of parliament and presidents with delusions of grandeur would have run the country into the ground long ago. When it is your turn to speak, introduce yourself as a civil servant. Re- mind everyone just how critical your work is. The civil service is not just there to create red tape; in fact, the civil service has kept Rationaltopia func- tioning every time the political system has gone into a tailspin. Appendix D: Political-Economy Homework Exercises 335 Participant 4: Character Card A You are the head of a small nongovernmental organization (NGO) whose aim is to promote awareness about the exclusion of the poor across the country. However, apart from the fact that most politicians and policy makers are too busy to listen to you, you also face a challenge that people from the north and the south do not trust each other. When it is your turn to speak, introduce yourself as a head of a poverty- focused NGO. Express your concern that, apart from the fact that politicians don’t seem to care about the very poor (who generally do not vote), south- erners and northerners are unwilling to come together to resolve common issues like poverty. Participant 5: Character Card A You are the head of the Small Business National Chamber of Commerce. You are very worried that a growing number of existing businesses seem to be using their links to MPs or civil servants to obtain advantages over other firms. This skewed form of competition is not just unfair to new businesses but also to the public, who end up paying higher prices for poorer-quality goods and services. When it is your turn to speak, introduce yourself as the head of the Cham- ber of Commerce. Let the other participants know that you see new busi- nesses as an important engine of growth in Rationaltopia but that you are concerned about the increasingly unhealthy relationships among business executives, politicians, and bureaucrats. Group Work Once all group members have introduced their characters, think about and discuss the following questions together. Record your answers below. 1. Which characters represent formal institutions? 2. Which character is policy seeking (wants to implement her policy no matter what)? 3. What factor seems to be driving the profits of some of the firms in the country? 4. Which character is election seeking (wants to keep her position no matter what)? 5. What role do cultural groups (even informal institutions, including bonds of trust) appear to have in the development of civil society? 6. Which character thinks a little that corruption is not such a bad thing? 336 Understanding Policy Change Exercise 2: The Budgetary Process in Rationaltopia Group Briefing: Read this information as a group. Then, individually, rank your budget priorities (on the Individual Instructions page) according to the character card assigned you in exercise 1. When you are finished, record the preferences of each character (on the Group Work grid) and discuss your findings. According to Rationaltopia’s constitution and laws, the budget is devel- oped by parliament (which must pass it by a simple majority). Once approved by the president, the budget is implemented by the civil service. The imple- mentation is supposed to be overseen by parliament’s Public Accounts Committee. In practice, the budget is proposed by the president. Parliament usually does not add significant amounts of additional expenditure but, depending on who is in charge, does seem to divert a significant amount of money to projects favored by key supporters. In particular, the south of the country, which is more developed and has more people (and thus more votes), ap- pears to do disproportionately well from allocations. With respect to implementation, the Public Accounts Committee does not seem to be working very well. The committee convened only once last year (so that its members could receive their committee allowances). This means that the bureaucracy is not supervised in implementing the budget. The perception that some bureaucrats are developing “cozyâ€? relationships with businesses is now considered a fact. There are five main items in the budget and thus five possible alterations: 1. Defense and foreign affairs expenditure 2. Discretionary spending for miniprojects allocated by parliament 3. Money for the procurement and delivery of large domestic projects (administered by the relevant departments) 4. Money for welfare spending (for schools and health clinics), currently allocated so that, per capita, southerners obtain more funds) ­ 5. Tax cuts and deductions for businesses. Individual Instructions As your character (from the previous exercise), rank the budget options ac- cording to your perceived priorities. Give the item you would like to achieve the most a 1 and the item you care least about a 5. Appendix D: Political-Economy Homework Exercises 337 Budget item Rank (1 = most preferred, 5 = least preferred ) Defense and foreign policy Parliamentary discretionary spending Large project procurement Money for welfare spending Tax cuts and deductions Group Work Decide on a volunteer recorder. As a group, go around the table and discuss the preferences of the different characters. Record the preferences on the grid. Head of Chamber of Budget item President MP Bureaucrat NGO Head Commerce Defense and foreign policy Parliamentary discretionary spending Large project procurement Money for welfare spending Tax cuts and deductions After this process is complete, discuss to what extent the characters agree over the budget priorities Rationaltopia should adopt. Then address the fol- lowing questions: 1. Do you think that members of parliament who decide to sit on a commit- tee have typical characteristics of an MP, or are they likely to be a self- selected type? 2. Does the behavior of Public Account Committee members in this case sound like a case of moral hazard or adverse selection? 3. Why would a parliament want to establish a committee to specialize in public financial oversight? 338 Understanding Policy Change Exercise 3: The Emergence of a Better Budget? Group Briefing: Read this information as a group. Then discuss the question that follows. Given the huge gulf between how the Rationaltopia budget should be implemented and how it is being implemented in practice, a group of con- cerned citizens and donors would like to organize a committee to generate ideas about how to improve budget formulation and implementation. The group has two options for the composition of the committee’s membership: • A random but representative sample of citizens • A group of civil society, public sector, and private sector leaders who are considered experts in the nature of public policy making. Regardless of how it is composed, the committee will have access to a group of international experts who will help identify international best practice and the feasibility of different options of budget response. As a group, discuss and make notes on the following question: • What are the benefits and pitfalls of letting a representative sample of citizens make decisions rather than delegating to a group of experts? Appendix D: Political-Economy Homework Exercises 339 Exercise 4: The Committee Comes Up with a Proposal Group Briefing: Read this information as a group. Then read your new indi­ vidual card for 2 minutes. (You should keep the same character as in the pre­ vious exercises.) Discuss the proposed reform as a group and take a final vote. After recording your vote results, talk about the final question and summarize your conclusions in the space provided. Having decided to adopt a hybrid model (including both a random set of citizens and a hand-selected group of policy-making leaders), the committee convened and has now come up with a set of recommendations for the bud- get process. These include a recommendation to introduce a participatory budgetary procedure; that is, to give ordinary citizens the right to be con- sulted and make authoritative decisions on the budget. This proposal is be- ing widely reported in the media and is gaining traction as a viable solution for the budget situation. After reading your character cards, discuss the proposal as needed, vote, and record your votes below. Yes No Conditional The president The MP The bureaucrat The head of the NGO The head of the Chamber of Commerce Once you have voted, discuss whether the opposition of the different charac- ters will keep the reform from passing. Summarize your conclusions below. 340 Understanding Policy Change Participant 1: Character Card B As the president, you are indifferent to whether this proposal becomes law. After all, you have more important defense and foreign policy issues to deal with. If parliament ratifies the new law, you will not veto it—why agitate any- one?—but if parliament does not, this is of no concern to you anyway. You have a region to stabilize. Participant 2: Character Card B As an MP, you are aghast at this proposal. How will you be able to target re- sources toward your loyal supporters if people who may not help your re- election can alter the budget? You must stop this silly proposal from ever becoming law. Participant 3: Character Card B As a career civil servant, you know that having to deal with the “bright ideasâ€? of even more uninformed people is the last thing you need. Participant 4: Character Card B As the head of an NGO working with the marginalized, this reform is poten- tially great news. Finally, the politically marginalized, low-income citizens will have an opportunity to make their voices heard. Of course, it will be important to ensure that the north is provided with the same opportunities as the rest of the country. But despite these reservations, you are an avid supporter. Participant 5: Character Card B As head of the Small Business Chamber of Commerce, you are concerned, on the one hand, that the uncertainty generated by a participatory budget pro- cedure may adversely affect the business environment. On the other hand, if this makes it more difficult for politically privileged businesses and bureau- crats to collude and unfairly deprive new businesses of opportunities to grow, then maybe it’s not such a bad thing. You are probably going to come out in favor. Appendix D: Political-Economy Homework Exercises 341 Exercise 5: A Tentative Coalition Emerges Group Briefing: Read this information as a group, discuss the questions, and record your answers. It has been a few months since the participatory budget proposal was first launched, and a few actors are trying to come together to support its realiza- tion. Many of the supporters of this process are pro-poor NGOs and small businesses. The most vigorous opposition is emanating from politically ac- tive citizens from across the political spectrum who have traditionally en- joyed privileged access to parliament and the bureaucracy. As a group, discuss and record your answers to the following questions: • How easy is it to build a coalition between actors with very different ex- periences (small profit-making businesses and NGOs)? • What kind of coordination or collective action problems might such a co- alition face? • How might opponents of the reform try to split the coalition, especially given their additional financial resources? (Hint: Think about the defini- tion of corruption.) 342 Understanding Policy Change Exercise 6: Building the Coalition Group Briefing: Again, read this information as a group, discuss the questions, and record your answers. Despite attempts by opponents to bribe some of its members, the coali- tion in favor of the participatory budget proposal is gaining traction. How- ever, to secure passage in parliament, the coalition needs to communicate that it enjoys the support of a broad majority of citizens—unlike the oppo- nents of the reform, who are a (very) vocal minority. As a group, talk about how the following factors might help the nascent reform coalition achieve its objective or prevent it from reaching its goal. Record your discussion below. • Currently, the coalition conducts its business through a council in which different stakeholders have the same power. How might developing a clear leadership structure affect the group’s efficacy? • Would the coalition benefit from using opinion polls and strategic com- munication techniques to dispel the appearance that the measure does not have broad appeal? Appendix D: Political-Economy Homework Exercises 343 Exercise 7: Coalition Case Study Group Briefing: As a group, compile the information from the previous exer­ cises into a case study presentation. If your instructor indicates that Power­ Point is available, consider creating slides; if not, think about other visualiza­ tions that might enhance your presentation. Use the following questions as a guide: • Who are all the relevant actors? • How does the institutional context affect the policy-making environment? • What stakes do different actors have in the reform process? • How can reform actors coordinate to develop a reform coalition? • What kind of leadership structure or other activities can the coalition use to enhance its effectiveness? The timing of reforms is also critical to success. Think about how the follow- ing events might affect the activities of the coalition. Then incorporate into your presentation how the coalition might deal with these events: • Upcoming elections (in the next few months) • Antireform actions (protests, communication campaigns, and the like) organized by powerful antireform groups • The growing perception in the international epistemic community that participatory budgeting constitutes a sound reform that can enhance accountability • Civil war in a neighboring country; the president commits Rationaltopia’s army for peacekeeping, thus necessitating higher defense expenditures • Economic crisis with global markets losing tolerance for opaque budget- ing processes (demanding higher interest rates from countries that have poor budget practices). Present your case study as a group during class, as scheduled by your instructor. 344 Understanding Policy Change INDEX Boxes, figures, notes, and tables are indicated by b, f, n, and t following page numbers. accountability, 51–78. See also corruption agenda control, 169 causal path between citizen agenda manipulation, 162b, 169, 171–72, preferences and policy outcomes, 172b, 178n1 55, 56f agenda ownership, 292 conceptual map of, 52f agenda setting, 157–79 defined, 52–54 conceptual map of, 158f in democratic settings, 53, 55–58 cycling and, 161–63, 162b modes of relations, 56 defined, 159–60 overview, 55–56 elections and, 171b, 172b types of, 52f, 56–58, 56t, 58f for individuals and organizations, media influence on, 218–20 160–61 monitoring and sanctioning and, 55, leadership in, 169, 261b 57, 61 manipulation in, 162b, 169, 171–72, in nondemocratic settings, 58–61 172b, 178n1 authoritarianism, 60–61 objectives of, 163 overview, 58–59 optimum outcomes and constraints single-party regimes, 59–60 in, 162b overview, 51–52 overview, 157–59 principal-agent relationships and, 53 in procurement reform, 11, 163–64, administrative corruption, 64 165–66, 167, 327, 330 adverse selection, 87–88, 215 as strategic tool, 171–73 Africa. See also specific regions and types of, 163–71 countries factors affecting, 163–67 corruption in, 65b formal, 163, 167–68, 169, 170b HIV/AIDS in, 243b informal, 168–69, 170–71, 171b neo-patrimonialism in, 139, 140b agents. See principal-agent relationships tax agency reform in, 193 agricultural policies, 92b African National Congress (ANC, South AIDS. See HIV/AIDS Africa), 135, 226 Akerlof, George, 215 agency loss, 187 Albania, health care in, 38 agency problems, 212 ally principle, 184, 192 345 amendatory observations, 170b Bokassa, Jean Badel, 9, 114 amendment rules, 164–66 Bolivia, civil service reform in, 310 ANC (African National Congress, South Boncodin, Emilia, 327, 331 Africa), 135, 226 Borda voting system, 164 Andaya, Rolando, Jr., 330 Bouazizi, Mohamed, 114–15, 247 Angara, Edgardo, 331 Brazil Anglican Church, 65 civil society organizations in, 306 Arab Spring, 10, 114–15, 221b, 247 collective action in, 247, 251 Argentina decentralization in, 25 decentralization in, 25 education in, 37, 38 tax agency reform in, 191b electoral laws in, 126, 127b unions in, 251 HIV/AIDS in, 243b Arrow, Kenneth, 162b institutional enforcement in, 142–43 Arroyo, Gloria Macapagal, 165, 327, 328, technology interventions in, 221b 332 unions in, 251 articulated interests, 284 water access in, 37 Asia. See also specific countries Bresser Pereira, Luiz Carlos, 126 development outcomes in, 95b bribery. See corruption education and health care in, 32 broadcasting. See media influence financial crisis (1997), 211 budget systems informal institutions in, 117 corruption in, 26 manufacturing growth in, 133b cultural differences in, 26, 27, 62b authoritarian regimes cycles of, 307–8 accountability in, 60–61 data on, 40 agenda setting in, 172b information asymmetries and, 87 corruption in, 61 bureaucracies defined, 59 controlling, 191b delegation in, 185 data collection process and, 32 education in, 38 implementation capacity of, 31 elections in, 172b informal rules of, 130–31, 132t incentives for development and lack of coordination in, 32–33 collective action, 123t malfunctions of, 40 transition to democracy, 135 media influence on, 217b personality types in, 213b backward induction, 188, 193–94, 195f self-selection in, 213b Balaguer, Joaquin, 242b tax authorities in, 191b ballot systems, 34, 127b working environment of, 213b Banerjee, Abhijit, 10, 42 Bangladesh cabinet formation, incentives for clientelism in, 68 development and collective action, 125t flooding in (1988, 1998), 220 California water system studies, 254 health care in, 32 Cambodia, infant mortality rates in, 30 bank independence, 237b, 308 CDF (Constituency Development Fund, Belgian Congo, education in, 31 Kenya), 35 Benin Central African Republic, political clientelism in, 67, 67b violence in, 9, 16n1, 114 collective action in, 224 Central Asia. See Eastern Europe and “best practiceâ€? recommendations, 308 Central Asia 346 Understanding Policy Change central bank independence, 237b, 308 credible commitment and, 240 checks and balances delegation to agents, 182, 185 credible commitment and, 237, 239 exogenous factors and, 309 delegation mandates and, 185 incentives for development and in democratic settings, 55 collective action, 123–26t ex ante checks, 190, 192 institutional incentives and ex post checks, 190, 192 constraints, 313b in formal institutions, 121 leadership of. See leadership in nondemocratic settings, 53, 60, 61 loyalty, importance of, 257 Chile monitoring devices of, 255 institutional enforcement in, 142 Philippines public procurement technology interventions in, 221b reform. See Philippines China procurement reform education in, 38 political-economy analysis and, 315b famine in (1958–61), 54b solving information asymmetries ideological shifts in, 138 and, 229 infant mortality rates in, 38 of stakeholders, 283 informal institutions in, 129, 130b timing and, 305 literacy rates in, 60 trust issues. See trust media influence in, 220 voting rules and, 164 civil service reforms, 24–25, 310 collective action, 6–11, 79–109, 247–78 civil society organizations, 220, 251, 254, conceptual map of, 80f, 249f 306 credible commitment and, 241–43, 242b clientelism decision making. See agenda setting collective action-induced credible defined, 80 commitment and, 241 development and, 81–82, 113, 251–53 corruption and, 66–69. See also five percent rule of, 93, 250 corruption game theory and, 94–98, 264–65 defined, 33, 57, 66 identifying and evaluating, 91–94 education and, 33 incentives for, 122, 123–26t, 131, 132t elections and, 34 information asymmetries and, 85, ethnicity and, 67b, 68 87–88, 224–29 gender and, 67b inhibiting factors, 85–88 health care and, 33 institutional change and, 8–11, 12f, informal institutions and, 130, 132t 113–15, 114f, 253–54 institutional enforcement and, 143 modes of, 7, 94, 95b, 251–53 political violence and, 71 motivational problems and, 85–86, 255 public sector employment and, 72 overview, 79–81, 247–49 closed-list system for ballots, 34, 127b in policy change and reform, 227–29 closed voting rules, 167 political violence and, 9, 10, 114–15 club goods, 83, 83t, 84 principal-agent relationships and, coalition building, 7, 80, 94, 208, 285, 292, 194b 333 problem-solving tools, 254–65 coalitions, 7, 25, 61, 79 leadership, 261–65 agenda ownership and, 292 monitoring and sanctioning, agenda setting and, 169, 315b 254–58 collective action and, 241, 248, trust and communication, 251–53, 309 258–61, 265 Index 347 pro-development, 113 Constituency Development Fund (CDF, public goods and, 82–85, 88–91, Kenya), 35 249–54 constitutions and constitutional rights club goods vs., 83, 83t to health care, 119b defined, 82–85 lifespan of, 141b electoral laws affecting, 126, limits on power, 237–38 127b to water, 119b incentives impact on, 82, 82f contracts in principal-agent mobilization and production of, relationships, 190 90–91, 249–54 control schemes, 190. See also monitoring monitoring of, 221, 221b and sanctioning private goods vs., 83, 83t corruption, 51–78. See also accountability social dilemmas of delivering, administrative, 64 88–91 in authoritarian regimes, 61 stakeholder dynamics and, 314–15b blackmail, 131 transparency and, 226–27 in budget systems, 26 collective arrangements, 7 clientelism and, 66–69 collective choice, 160 collection action-induced credible collective incentives, 82, 82f commitment and, 242 Colombia, health care in, 119b conceptual map of, 52f Committee on Rules (U.S. Congress), 159, cultural differences and, 63 167 defined, 62–63 common knowledge, 188 economic development and, 70, 70f, common-pool resources, 82, 83, 83t, 251, 74n10, 307, 323 253–54, 255 in education, 31 communication. See also information grand, 63–64, 66 asymmetries in health care, 31 in collective action, 258–61, 265 incentive pay and, 216 face-to-face, 259, 261 interventions to reduce, 72–73, 73b institutional arrangements kleptocracy, 66 promoting, 253 large-scale, 71 need to improve, 248 low-level, 63 political-economy analysis and, 315b negative consequences of, 67–70 problems in, 33 nepotism, 64 stakeholder’s resources, 311b overview, 51–52, 62–63 strategic communication to garner political-economy analyses and, 72 support, 306 positive consequences of, 70–72 strategic miscommunication, 290 prebendal predation, 65, 65b communism, 31, 59, 117, 304 prevalence of, 72 comparative statics, 289 in principal-agent relationships, 63 complementarity of reform areas, 310 public procurement and, 13, 22–23, Condorcet winner, 162, 178n4 324 Congo, Democratic Republic of resource allocation and, 69–70, 194b civil society organizations in, 254 small-scale, 70–71 corruption in, 65, 65b state capture, 64–65, 66 leadership and trust in, 262b systemic, 69 monitoring and sanctioning in, 256 types of, 63–67 348 Understanding Policy Change credible commitment, 233–46 incentives for development and central bank independence and, collective action, 123t 237b, 308 mobilization of public goods in, 61 collective action and, 241–43, 242b modes of relations in, 56 conceptual map of, 234f overview, 55–56 defined, 235–36 poverty in, 33 to development, 241–43 transition to, 135 effectiveness of, 240–41 types of, 52f, 56–58, 56t, 58f game theory and, 243–44 demographic changes, 135–36 mechanisms for ensuring, 236–40 Denmark, public account committees constitutional limits on power, in, 216 237–38 Department of. See specific name of guardian judiciaries, 238 department hostage, 238 descriptive political-economy analyses, independent enforcement, 289, 290 236–37 development. See also public goods monitoring, 236 baselines for, 30 reputational, 239–40 collective action and, 81–82, 113, veto players, 238–39, 239b 251–53 overview, 233 credible commitment to, 241–43 time inconsistency and, 234–35 economic, 70, 70f, 74n10, 307, 323 critical junctures, 304–5 information asymmetries and, 228b Cuba social capital and, 259, 259b education in, 38 dictatorships. See authoritarian regimes health outcomes in, 60 Dingell, John, 159 cultural differences Diokno, Benjamin, 325, 326 in budget systems, 26, 27, 62b direct delegation, 185 corruption and, 63 discrimination, 143 cycling and agenda setting, 161–63, 162b distrust. See trust Czech Republic, institutional change in, Dominican Republic 137 constitutional lifespan in, 141b credible commitment in, 242b data collection process, 32 DPP (Malawi), 69 decentralization, 25, 34–35, 135, 219 Duflo, Esther, 10, 42 “deck-stackingâ€? oversight mechanism, dynamic inconsistency, 234–35 191b deductive approach, 284 Eastern Europe and Central Asia. See Defense Department (U.S.), 308 also specific countries delegation, 182, 183–86, 186b communism in, 59, 304 demand-side approaches, 23, 30, 131, 142 education and health care in, 32 Democratic Republic of Congo. See institutional change in, 137 Congo, Democratic Republic of post-communism economy in, 117 democratic settings tax policies in, 138b accountability in, 53, 55–58 economics checks and balances in, 55 corruption’s effects on development, delegation in, 185 70, 70f, 74n10, 307, 323 education in, 38 crises, 30, 307 Index 349 defined, 18, 19 equilibria inefficiencies, 39 changes in, 133–39 information asymmetries in, 210–11 demographic, 135–36 rational utility maximization, 44n1 global production and education distribution, 137 administrative constraints on, 31–33 ideological shifts, 137–39 in authoritarian regimes, 38 political competition, 137 clientelism and, 33 power endowments, 134–35 corruption in, 31 property rights, 136 in democratic settings, 38 technological, 136 economic constraints on, 30–31 information asymmetries effect on, inequalities in, 30–31 217, 228b political constraints on, 33, 34–35, institutions as, 116, 117 36–39 Nash, 95, 97, 194, 228b, 244, 321 political violence and, 39 self-enforcing, 117, 152n2 public spending on, 28, 29f, 30–31 structure-induced, 163 elections Estonia, tax policies in, 138b accountability and, 53, 55–56, 59 Estrada, Joseph, 163, 167, 326, 327 agenda setting and, 171b, 172b ethnicity in authoritarian regimes, 172b clientelism and, 67b, 68 ballot systems, 34, 127b HIV/AIDS and, 243b campaigns, 56 policy implementation and, 128, 132t clientelism and, 34 European Commission, 163, 165 cycles of, 305 Europe/European Union. See also specific game theory and, 319–20 regions and countries incentives for development and agenda setting in, 163, 165 collective action, 124–25t corruption in, 63 information asymmetries in, 87, 88 institutional change in, 134–35, 137 political constraints on, 34, 35, 37 public account committees in, 216 programmatic linkage between public finances in, 308 voters and politicians, 57 tax policies in, 138b, 214 proportional representation and, time horizons in, 166 122, 127b ex ante checks, 190, 192 time horizon for performance Exchequer and Audit Act (Kenya), 165 evaluation, 56, 56t excludability of use, 83, 83t timing of reform and, 305 exogenous institutions, 152n5 voting rules, 164, 167–68 ex post checks, 190, 192 electricity access, 32, 37, 90–91 “external increasing returns,â€? 139 elites, 20, 140b, 258, 309, 310 El Salvador, information asymmetries famines, 53, 54b, 219b in, 207 federalism Elster, Jon, 256 central bank independence and, 237b employment in public sector, 72, 213b economic theories of, 25 endogenous institutions, 152n5 fiscal policies and, 26–27 enforcement. See monitoring and incentives for development and sanctioning collective action, 123t England. See United Kingdom fief systems, 214 350 Understanding Policy Change financial management reform, 26–27 Greece “fire alarmâ€? oversight mechanisms, 190, financial crisis in, 211 191b public account committees in, 216 fisheries and fishing, 89–90, 256 structural adjustment policies in, five percent rule of collective action, 93, 308 250 grim trigger strategy, 265 flooding, 219b, 220 guardian judiciaries, 238 formal agenda setting, 163, 167–68, 169, 170b Haiti, constitution in, 141b formal institutions, 118–26, 121f, 123–26t Hayek, Friedrich, 213b France, credible commitment in, 240–41 health care free riding, 85, 90, 92–93 administrative constraints on, 31–33 Friedman, Milton, 138b clientelism and, 33 constitutional rights to, 119b game theory corruption in, 31 backward induction and, 188, 193– economic constraints on, 30–31 94, 195f HIV/AIDS policies, 38–39, 119b, 243b collective action and, 94–98, 264–65 inequalities in, 30–31 concepts involved in, 317–20 litigation, 291b credible commitment and, 243–44 political constraints on, 33, 34–35, election platforms and, 319–20 36–39 institutions and, 144–45 public procurement of, 38 leadership and, 264–65 public spending on, 28, 29f, 30–31 Nash equilibria and, 95, 97, 194, 228b, stakeholder mapping and, 291b 244, 321 heresthetics, 169 overview, 11–12 HIV/AIDS, 38–39, 119b, 243b payoffs and, 318, 319t hometown associations, 22, 254 principal-agent relationships and, Honduras, elections in, 34 188, 193–94 hostage mechanisms, 238 prisoner’s dilemma and, 88–89, hybrid regimes, incentives for 95–98 development and collective action, strategies, 265, 318, 320, 321 123t trust and, 265 gender differences ideal points, 178n6 clientelism and, 67b ideological shifts, 137–39 demographic change and, 136 implementation capacity, 31 Germany, agenda setting in, 161 incentives gerrymandering, 34 for collective action, 122, 123–26t, Ghana, public account committees in, 131, 132t, 241 216, 217b individual vs. collective, 82, 82f Gini coefficient, 207–8 institutions and, 113–15, 122, 123–26t, global production and distribution, 137 131, 132t, 313–14b Gonzalez, Neptali, II, 326, 327, 330 outcome-based, 189 governance in principal-agent relationships, 189 accountability and, 52 public goods and, 82, 82f systems of, 253–54 salary and, 216, 258 grand corruption, 63–64, 66 selective, 216, 257–58 Index 351 independent enforcement, 236–37 types of, 212–15 India adverse selection, 87–88, 215 collective action in, 247, 251 moral hazards, 87, 213–15 development in, 259b self-selection, 212–13, 213b electricity access in, 37, 90–91 inheritance rules, 134 famine and flood relief in, 54b, 219b initiative rules, 163 health care in, 119b instability trap, 140 labor relations in, 131, 133b institutional enforcement, 141–44 manufacturing growth in, 133b Institutional Revolutionary Party (PRI, media influence in, 218, 219b Mexico), 33 political competition in, 219b institutional veto players, 239b technology interventions in, 221b institutions, 111–55 unions in, 251 changes in, 133–39 water conservation in, 89 demographic, 135–36 indirect delegation, 185 global production and individual incentives, 82, 82f distribution, 137 Indonesia, economic crisis in, 30 ideological shifts, 137–39 inductive approach, 284 political competition, 137 inequalities in education and health care, power endowments, 134–35 30–31. See also poverty property rights, 136 infant mortality rates, 30, 38, 39 technological, 136 inferential political-economy analyses, collective action and, 8–11, 12f, 113– 289, 290 15, 114f, 253–54 informal agenda setting, 168–69, 170–71, conceptual map of, 112f 171b constraints and, 314b informal economy, 143 defined, 42, 115–16 informal institutions, 118–21, 121f, 127–31, as equilibria, 116, 117 132t failure of, 140–44 informal organizations, 128 formal, 118–26, 121f, 123–26t information asymmetries, 207–32 game theory and, 144–45 collective action and, 85, 87–88, incentives and, 113–15, 122, 123–26t, 224–29 131, 132t, 313–14b conceptual map of, 209f informal, 118–21, 121f, 127–31, 130b, correctives for, 216, 218–21 132t defined, 85, 87, 187, 207 information asymmetries effect on, development and, 228b 217–18 equilibria and, 217, 228b intersection of, 121, 121f implications of, 217–18 monitoring and sanctioning in, institutions and, 217–18 117–18 media influence on, 216, 217b, 218 as norms, 116, 118 monitoring and sanctioning of, 216 organizations distinguished from, overview, 207–10 116–17 in political-economy analyses, origins of, 131–33 210–12 overview, 111–13 in principal-agent relationships, 187, path dependence and, 139 216 public goods and, 253–54 technology and, 221, 221b resilience of, 139, 140 transparency and, 222–23, 223f role of, 286–87 352 Understanding Policy Change as rules, 115, 117–18 health care in, 119b stability and instability of, 140 indigenous laws in, 129 weak enforcement, 141–44 tax agency reform in, 193 interest groups leadership, 261–65 defined, 7 adverse selection and, 87 incentives for development and in agenda setting, 169, 261b collective action, 125t collective action problems and, number of, 93 263–64 interest rates, 23–24 defined, 293 Interior Department (U.S.), 64 game theory and, 264–65 international “best practiceâ€? loyalty, importance of, 257 recommendations, 308 in political-economy analyses, International Development Department 292–93 (U.K.), 44 transitions from autocracy to International Monetary Fund, 138b, 308 democracy, 135 Iran trust and, 261b budgetary decision making in, 62b legislative systems, incentives for collective action in, 260 development and collective action, government structure of, 62b 124t Iraq, pension system in, 186b Liberia, leadership and collective action Ireland, structural adjustment policies in, 263 in, 308 Lincoln, Abraham, 171b literacy rates, 60 Japan lobbying and resource allocation, 69 financial liberalization in, 284–85 logic of collective action, 88, 90–91, 92b, informal rules of, 131 93, 108n1 tax policies and rebellions in, 118, 214 low-level corruption, 63 Lusinchi, Jaime, 127 Kenya collective action in, 225 majoritarian systems, 178n3 corruption in, 64 Malawi, corruption in, 65, 68–69 elections in, 35 managerial accountability, 57, 58, 58f. See elites in, 140b also principal-agent relationships procurement reform in, 165, 227–29 mandate responsiveness, 52 technology interventions in, 221b manipulation in agenda setting, 162b, 169, women in leadership roles in, 263 171–72, 172b, 178n1 kleptocracy, 66 Mapping for Results program, 221b market failure, 19 Laar, Mart, 138b media influence labor relations, 92b, 131, 133b. See also accountability and, 218–20 unions on bureaucracies, 217b large-scale corruption, 71 on famine and flood relief, 219b latent interests, 284 information asymmetries and, 216, Latin America. See also specific countries 217b, 218 agenda setting in, 169 on public opinion, 306 civil service reform in, 310 Mexico economic reform in, 307 institutional enforcement in, 142 electoral laws in, 27 poverty alleviation programs in, 33 Index 353 unions in, 251 Nigeria water conservation in, 89 budget issues in, 227 mobilization of public goods, 90–91, decentralization in, 25 249–54 hometown associations in, 22, 254 in democratic settings, 61 monitoring and sanctioning in, group size and, 90–91, 250–51 255–56 institutional arrangements and, transparency in, 225, 227 253–54 women’s cooperatives in, 255–56 modes of, 251–53 nondemocratic settings. See also in nondemocratic settings, 61 authoritarian regimes mobilized interests, 284 accountability in, 58–61 Mobutu Sese Seko, 66 checks and balances in, 53, 60, 61 monetary stakes, 286 mobilization of public goods in, 61 monitoring and sanctioning overview, 58–59 accountability and, 55, 57, 61 single-party regimes, 59–60 of collective action, 254–58 nonexcludable goods, 82, 83, 83t credible commitment and, 236 nonexhaustible goods, 83 ex ante checks, 190, 192 nongovernmental organizations (NGOs). ex post checks, 190, 192 See civil society organizations independent enforcement, 236–37 nonrivalrous goods, 82 of information asymmetries, 216 norms institutional enforcement, 141–44 institutions as, 116, 118 in institutions, 117–18 reputational, 239–40 in principal-agent relationships, North, Douglass, 137 190–93 selective incentives and, 216, 257–58 Odyssey (Homer), 235 technology used in, 221, 221b Olson, Mancur, 108n1 top-down, 253 one-party regimes. See authoritarian monocentric governance systems, 253 regimes Montesinos, Vladimir, 10 one-vote systems, 164 moral hazards, 87, 213–15 open-list system for ballots, 34, 127b mortality rates. See infant mortality rates open voting rules, 167–68 motivational problems in collective ordered open voting rules, 167 action, 85–86, 255 Organisation for Economic Co-operation multi-issue organizations, 285 and Development countries, 305 multiparty democracies, 53 organizations. See also civil society organizations; political parties; unions Namibia, elites in, 140b agenda setting for, 160–61 Nash equilibria, 95, 97, 194, 228b, 244, 321 coalitions, 94, 252 neo-patrimonialism, 139, 140b collective action and, 251–53 Nepal informal, 128 education spending in, 31 institutions distinguished from, leadership solutions in, 22 116–17 nepotism, 64 multi-issue, 285 “network externalities,â€? 139 single-issue, 285 new public management approach, 73, tribal, 128–29, 132t, 262b 73b Ostrom, Elinor, 6, 81 354 Understanding Policy Change Ottoman Empire, 214 Pitkin, Hannah, 74n2 outcome-based incentives, 189 “police patrolâ€? oversight mechanisms, oversight mechanisms. See monitoring 190, 191b and sanctioning policy change and reform, 22–40, 303–16 civil service, 24–25, 310 PAC (public account committees), 216 collective action in, 227–29 pacing reform, 309–10 complementarity vs. trade-offs, 310 Papua New Guinea decentralization, 25, 34–35, 135, 219 education in, 32 decision making. See agenda setting political system in, 88 financial management, 26–27 paradox of voting, 178n7 interest rates, 23–24 Paraguay, information asymmetries in, overview, 303–4 207 pacing and sequencing in, 309–10 parchment institutions, 122 poverty, 28–40 parliamentary systems administrative constraints on, coalitions in, 252 31–33, 39–40 delegation in, 185 economic constraints on, 30–31, incentives for development and 39–40 collective action, 123t implementation gaps in legislative power and, 26 spending and, 28, 29f monitoring and sanctioning in, 192 political constraints on, 33–40 public spending in, 44n2 positions on, 28–29 partisan veto players, 239b procurement. See Philippines Pascua, Laura, 327 procurement reform; public path dependence, 139 procurement patronage. See clientelism stages of, 7 payoffs, 285, 286, 301n1, 318, 319t timing of, 305–9 PEA. See political-economy analyses domestic factors, 305–7 personality types in bureaucracies, 213b economic factors, 307–8 Peru international factors, 308–9 collective action in, 225, 226–27 veto players and, 238–39, 239b institutional enforcement in, 142 political accountability, 57, 58, 58f legislature in, 225, 226–27 political business cycles, 24 political violence in, 10 political clientelism. See clientelism Philippines procurement reform, 323–32 political competition, 137, 219b agenda setting in, 11, 163–64, 165–66, political economy, defined, 18–20, 20f 167, 327, 330 political-economy analyses (PEA), 17–47, collective action in, 247–48 281–301 first round of legislation, 325–27, agenda ownership and, 292 328f characteristics of, 21–22 identification of problem, 324 checklist, 311–15b overview, 13–14, 323–24 conceptual map of, 282f reform teams and, 7 constraints in, 287–88 second round of legislation, 327–32, corruption and, 72 331f descriptive, 289, 290 stakeholder mapping of, 329, 329t inferential, 289, 290 timeline of, 324, 325f information asymmetries in, 210–12 Index 355 institutional context, 121, 121f constitutional limits on, 237–38 leadership in, 292–93 endowments, 134–35 limitations of, 40–44 prebendal predation, 65, 65b methods and instruments for, 18–22, preferential voting systems, 164 288–91, 291b presidential systems modes of, 289–90 delegation in, 185 overview, 281–82 incentives for development and reform. See policy change and collective action, 123t reform legislative power and, 26 stakeholder mapping and, 283–85, monitoring and sanctioning in, 192 289 public spending in, 44n2 political participation, 261b presidential veto, 170b political parties PRI (Institutional Revolutionary Party, as collective action problem solvers, Mexico), 33 95b primogeniture, 134–35 elections and, 88 principal-agent relationships, 181–206 functions of, 7, 94, 251 accountability and, 53 incentives for development and assumptions of, 186 collective action, 126t collective action and, 194b programmatic, 194b compliance in, 189–93 political representation and conceptual map of, 182f accountability, 52, 74n2 corruption in, 63 political uncertainty, 184, 192 delegation and, 182, 183–86, 186b political violence features of, 187–89 clientelism and, 71 game theory and, 188, 193–94 collective action and, 9, 10, 114–15 incentives in, 189 education and, 39 information asymmetries in, 187, 216 infant mortality rates and, 39 inherent problems in, 187 politics, defined, 18, 19 monitoring and sanctioning in, polycentric governance systems, 253–54 190–93 pork-barrel spending, 23 overview, 181–83 Portugal, structural adjustment policies probability of detection and, 190 in, 308 prisoner’s dilemma, 88–89, 95–98 Potemkin village phenomenon, 142 private goods, 83, 83t, 84 poverty private investment, 71, 241 defined, 86 privatization of public policy, 64 in democratic settings, 33 probability of detection, 190 reform, 28–40 procurement. See Philippines administrative constraints on, procurement reform; public 31–33, 39–40 procurement economic constraints on, 30–31, programmatic linkage between voters 39–40 and politicians, 57 implementation gaps in programmatic political parties, 194b spending and, 28, 29f PRONASOL (Mexico), 33 political constraints on, 33–40 property rights, 136, 138 positions on, 28–29 proportional representation, 122, 127b power public account committees (PACs), 216 in agenda setting, 169 public goods 356 Understanding Policy Change club goods vs., 83, 83t reform opponents, 285 defined, 82–85 reform teams, 7, 94. See also policy electoral laws affecting, 126, 127b change and reform incentives for development and Regional Transport Committee (RTC, collective action, 122, 123–26t, 126 Bangladesh), 68 incentives impact on, 82, 82f reliability and accountability, 52 mobilization and production of, reputational norms, 239–40 90–91, 249–54 resource allocation and corruption, in democratic settings, 61 69–70, 194b group size and, 90–91, 250–51 responsibility and accountability, 52 institutional arrangements and, responsiveness and accountability, 52 253–54 restrictive voting rules, 167 modes of, 251–53 rhetoric, 169 in nondemocratic settings, 61 Riker, William, 162b monitoring of, 221, 221b risk asymmetries, 187–88 private goods vs., 83, 83t rivalry of use, 83 social dilemmas of delivering, 88–91 RTC (Regional Transport Committee, logic of collective action, 88, Bangladesh), 68 90–91, 92b, 93, 108n1 rules prisoner’s dilemma, 88–89, amendment, 164–66 95–98 inheritance, 134 tragedy of the commons, 27, 88, initiative, 163 89–90 institutions as, 115, 117–18 public health. See health care provision of public goods and public policy. See policy change and collective action for reform, 122, reform 123–26t public procurement. See also Philippines voting, 164, 167–68 procurement reform Rules Committee (U.S. Congress), 159, agenda setting and, 165 167 corruption and, 13, 22–23, 324 Russia, corruption in, 64–65 of health care, 38 Rwanda, demographic changes in, 135–36 public sector employment, 72, 213b public spending salaries and incentives, 216, 258 on education, 28, 29f, 30–31 sanctioning. See monitoring and on health care, 28, 29f, 30–31 sanctioning implementation gaps in, 28, 29f São Tomé and Príncipe, political pork-barrel, 23 participation in, 261b structure of, 31 Saudi Arabia, legislature in, 10 scenario analyses, 289 QWERTY typewriter keyboard, 139, schools. See education 152n8 Seip, Jens Arup, 237 selective incentives, 216, 257–58 ranked voting systems, 164 self-enforcing equilibria, 117, 152n2 rated voting systems, 164 self-help associations, 22, 254 rational players and actors, 317 self-selection, 212–13, 213b rational utility maximization, 44n1 Sen, Amartya, 53, 54b, 284 recruitment screening, 190 sequencing reform, 309–10 reform champions, 285 shirking, 187, 191 Index 357 Singapore state capture, 64–65, 66 development in, 259b strategic communication. See incentive pay in, 216, 258 communication single-issue organizations, 285 strategic voting, 169 single-member district plurality votes, strategies of game theory, 265, 318, 320, 122 321 single-party regimes, 59–60. See also strictly dominated strategies, 320, 321 authoritarian regimes structural adjustment policies, 308 slavery and agenda setting, 171b structure-induced equilibria, 163 slippage, 187 subgame perfect Nash equilibria, 194, 244 Slovakia, technology interventions in, Sub-Saharan Africa. See also specific 221b countries small-scale corruption, 70–71 agricultural policies in, 92b social accountability, 57, 58, 58f corruption in, 63 social capital, 259, 259b decentralization in, 219 social dilemmas, 88–91 education in, 31 defined, 81 foreign aid to, 309 logic of collective action, 88, 90–91, water access in, 119b 92b, 93, 108n1 substitution effect, 192 prisoner’s dilemma, 88–89, 95–98 subtractability of use, 83 tragedy of the commons, 27, 88, supply-side approaches, 23, 30, 131 89–90 systemic corruption, 69 social movement organizations, 7, 94 social optimums, 162b Tanzania, self-help associations in, 22, South Africa 254 collective action in, 225–26 tariffs, 69 decentralization in, 135, 219 Tatad, Francisco, 327 health care in, 38–39, 119b tax policies, 138b, 143, 214 HIV/AIDS in, 243b technology institutional enforcement in, 142–43 information asymmetries and, 221, water access in, 118, 119b, 225–26 221b South African National AIDS Council, institutional changes and, 136 243b monitoring public goods using, 221, Soviet Union 221b Communist system in, 59 Thailand, information asymmetries in, 218 dismemberment of, 304–5 time horizons spending. See public spending for agenda setting, 166–67 stakeholder mapping, 283–85, 289, 329, collective action and, 252, 265 329t for performance evaluation, 56, 56t stakeholders stakeholders and, 286 collective action and, 314–15b time inconsistency, 234–35 defined, 283 timing of reform, 305–9 instruments for identifying, 290–91, domestic factors, 305–7 291b economic factors, 307–8 mapping. See stakeholder mapping international factors, 308–9 preferences of, 311–12b toll goods, 83, 83t time horizons and, 286 top-down monitoring, 253 stakes, 260, 285–86, 301n1 trade-offs of reform areas, 310 358 Understanding Policy Change trade unions. See unions foreign aid and, 309 tragedy of the commons, 27, 88, 89–90, 253 information asymmetries in, 207–8 transaction costs on development, 259b media influence in, 306 transparency. See also accountability slavery in, 171b collective action and, 226–27 veto players in, 239b information asymmetries and, Uruguay, agenda setting in, 170b 222–23, 223f U.S. Agency for International Treatment Action Campaign Case (2002, Development (USAID), 325 South Africa), 38–39 triangulation, 289 Venecia, Jose de, 330 tribal organizations, 128–29, 132t, 262b Venezuela, República Bolivariana de Trujillo, Rafael, 242b electoral laws in, 126, 127b trust hostage mechanisms in, 238 collective action and, 258–61, 265 information asymmetries in, 207 defined, 258 veto, presidential, 170b game theory and, 265 veto players, 238–39, 239b leadership and, 261b vicious circles, 71, 140 social capital and, 259, 259b violence. See political violence sources and types of, 258–60 voting rules, 164, 167–68. See also elections tools for building, 260–61 value of, 260 water Tunisia access to, 37, 118, 119b, 225–26 Arab Spring and, 114–15, 247 conservation and management, 89, institutional enforcement in, 143 120b, 121, 128 Turkey, fishing in, 89–90, 256 constitutional rights to, 119b weakly dominated strategies, 321 UDF (Malawi), 68–69 women. See gender differences Uganda women’s cooperatives, 22, 136, 254, 255, accountability in, 218–19 263 decentralization in, 219 World Bank education spending in, 32, 36, 218–19 budgeting process, 308 public procurement in, 23 civil service reform projects, 24 self-help associations in, 22, 254 Iraqi pension law implementation, Ukraine, corruption in, 64–65 186b ultimatum bargaining, 188 Mapping for Results, 221b unions, 9, 33, 69, 116, 131, 133b, 222, 251, political-economy analysis, 44 256 World Development Report Making unitary states, 237b Services Work for the Poor (World United Kingdom Bank), 28, 29f credible commitment in, 240–41 veto players in, 239b Yemen, institutional networks in, 129 United Nations Development Programme, 44 Zambia United States education in, 31 agenda setting in, 163, 165, 167 health care in, 32 Constitution, 141b Zimbabwe corruption in, 64 elites in, 140b elections in, 171b water management in, 120b, 121, 128 Index 359 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The • 9 trees Office of the Publisher has chosen to print •4 million Btu of Understanding Policy Change on re­cycled total energy paper with 30 percent postconsumer fiber in •895 lbs. of net accordance with the recommended standards greenhouse gases for paper usage set by the Green Press Initia- •4,034 gal. of waste tive, a nonprofit program supporting publishers water in using fiber that is not sourced from endan- •256 lbs. of solid gered forests. For more information, visit waste www.greenpressinitiative.org.