Jump to ContentJump to Main Navigation
Democratic Accountability in Latin America$

Scott Mainwaring and Christoper Welna

Print publication date: 2003

Print ISBN-13: 9780199256372

Published to Oxford Scholarship Online: April 2005

DOI: 10.1093/0199256373.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 26 February 2017

Introduction: Democratic Accountability in Latin America *

Introduction: Democratic Accountability in Latin America *

(p.3) 1 Introduction: Democratic Accountability in Latin America*
Democratic Accountability in Latin America

Scott Mainwaring (Contributor Webpage)

Oxford University Press

Abstract and Keywords

This introductory chapter begins with a discussion on the challenges of democratic accountability in Latin America. It then presents the three main themes that run throughout the book. These are the controversies about the definition of accountability, the relationship between electoral accountability and mechanisms of control and oversight of public agencies, and the shortcomings and advances of non-electoral forms of accountability in contemporary Latin America.

Keywords:   democratic accountability, Latin America, political accountability, electoral accountability, intrastate accountability, oversight

Scott Mainwaring

This volume on democratic accountability addresses one of the burning issues on the agenda of policy-makers and citizens in contemporary Latin America. Collectively, we hope that the volume enhances understanding of three key issues. First, it enriches understanding of the state of non-electoral forms of democratic accountability in contemporary Latin America. What are some of the major shortcomings in democratic accountability? How can they be addressed? What are some of the major innovations in the efforts to enhance democratic accountability?

A second contribution of the volume is conceptual. Accountability is a key concept in the social sciences, yet its meaning varies widely from one author to the next. The authors in this volume, especially in the first four chapters, explicitly debate how best to define and delimit the concept. Although we cannot claim consensus in our understanding of this concept, we believe that the direct confrontation of ideas will advance the debate.

Finally, the volume also furthers understanding of the interaction between various mechanisms and institutions of accountability. Many of the authors address how electoral accountability (i.e., the accountability of elected officials to voters) interacts with other forms of accountability in which state agencies oversee and sanction public officials. The volume provides a fairly extensive treatment of this important but hitherto underexplored interaction.

Accountability has emerged as one of the key issues in the post-transition period of Latin American politics. In the 1980s, the (p.4) debate about democracy focused primarily on transitions to democracy and on ways of sustaining and consolidating democracy. Since then, Latin America has enjoyed its most democratic period ever. But in much of the region, disenchantment and indeed even cynicism have set in regarding the quality of these elected governments, raising the prospect of a new round of democratic erosion and breakdowns.

One of the important emerging challenges for improving the quality of democracy revolves around how to build more effective mechanisms of accountability. A widespread perception prevails in much of the region that government officials are not sufficiently subject to routinized controls by oversight agencies. Corruption, lack of oversight, impunity for state actors, and improper use of public resources continue to be major problems in most countries of the region. These problems have captured the attention of business, religious, and political leaders, as well as citizens and scholars. Dealing with these issues is paramount to restoring and deepening democratic legitimacy; public opinion surveys show that corruption has tarnished legitimacy. If these problems are not addressed, governmental instability may once again fester in the region, at a minimum through electoral upsets by anti-party outsiders and potentially through extra-institutional means (coup d'etats or popularly supported auto-golpes) as well.

The task facing Latin American democrats is not simply how to build more effective mechanisms of accountability. For while accountability is a desirable feature in political systems, so is governmental effectiveness,1 which often is in tension with accountability. On the one hand, governments cannot deliver what citizens need if they are so hampered by mechanisms of oversight and sanctioning agencies that they cannot undertake new initiatives. Constitutional crises between powers or challenges to policy decisions by oversight bodies, ombudsmen or lawsuits can undermine government effectiveness. Governments that are immobilized by oversight mechanisms are sometimes perceived by voters or power groups as being indecisive, ineffective, or inept. Powerful mechanisms of accountability can impede governments from implementing policies that would benefit most citizens; they can give small minorities, even a few individuals in key institutions (e.g., the Supreme Court) veto power over policy changes. At a minimum, mechanisms of accountability impose modest transaction costs. On the other (p.5) hand, Latin Americans also have extensive experience with lack of governmental accountability. If public officials do not need to account for their actions, they may be unresponsive to citizen needs and to the public good.

Positing the existence of a potential tradeoff between accountability and government performance does not imply a zero sum conflict between these two desiderata. Unaccountable governments can produce terrible policy results, and accountable governments can produce excellent results. Powerful mechanisms of accountability make it easier to block governments from implementing disastrous policies.

The fundamental question in this volume is how democratic leaders in Latin America can improve accountability while simultaneously promoting governmental effectiveness. In this volume, the authors approach this question by focusing on democratic institutions, which are key to accountability. Institutions do not necessarily function as intended, especially in a region known for a gap between the law and political practice, but they are essential for establishing accountability. Without formal institutions designed for this purpose, effective accountability is impossible.

These concerns about combining the dual objectives of accountability and policy results have been at the forefront of the region's policy and scholarly agenda,2 and they are the central concerns of this volume. These issues have acquired urgency in contemporary Latin America because of heightened public concern about corruption and improper governmental actions, on the one hand, and about the potential tradeoff between policy results and accountability, on the other.

Although accountability has become a salient concern in Latin America and other new democracies, only recently has a scholarly debate on non-electoral accountability started to emerge. Much of the bibliography on accountability deals with public administration (e.g., Day and Klein 1987), education, and other issues that fall outside the purview of this book. Przeworski, Stokes, and Manin's (1999) important book focuses on electoral accountability, and a substantial older literature has dealt with the closely related theme of retrospective voting, i.e., voting on the basis of elected officials' actions in office (Fiorina 1981; Key 1966). There is an abundant scholarly literature on some themes related to political accountability (legislatures, judiciaries, oversight agencies, (p.6) corruption control), but little of this literature focuses specifically on accountability. Some books examine a specific slice of accountability; for example, Woodhouse (1994) studies the accountability of ministers to parliament.

Recently, O'Donnell (1994, 1999a, 1999b, this volume) has called attention to how important accountability is in understanding and conceptualizing differences among the post-1978 wave of democracies and differences between these democracies and the advanced industrial polyarchies. Even with O'Donnell's path-breaking contributions, the literature that deals explicitly with non-electoral democratic accountability is limited. Merritt (forthcoming) and Schedler, Diamond, and Plattner (1999) are among the few noteworthy contributions that conceptualize and analyze a broad range of institutions of accountability.

In this introduction, I take up three themes that run throughout the remainder of the volume. First, I discuss some controversies about defining accountability. These controversies run throughout this volume and more broadly throughout the literature on accountability. It is essential to tease out these different understandings of ‘accountability’ and to begin working toward greater conceptual clarity. Second, Moreno et al. (this volume) present an innovative but controversial argument about the relationship between electoral accountability and mechanisms of control and oversight of public agencies. This relationship requires more attention among scholars and democratic practitioners, so I highlight some of the key issues in the controversy. Finally, I present an overview of the debate about the shortcomings and advances in non-electoral forms of accountability in contemporary Latin America.

Defining and Delimiting Political Accountability

Accountability is a far-from-consensual concept. Indeed, the meaning of ‘accountability’ is about as muddled as concepts get in the social sciences. There are sharp differences in usage among authors in this volume and more broadly in the social sciences. One of the contributions that this volume offers is an explicit debate and disagreement especially in the four chapters of Part I regarding how to define and delimit ‘accountability’. In this section, I present my own understanding of the concept. In the next, I survey the main sources of disagreement regarding the concept of accountability because a clear statement about where the disagreements lie can help illuminate the debate.

(p.7) ‘Accountability’ is used in many contexts, ranging from education to business firms, from public administration to politics. This volume is concerned specifically with political accountability, which—as all of the authors in this volume agree—refers to the answerability and responsibility of public officials.3 I delimit the concept of political accountability to relationships that formally give some actor the authority of oversight and/or sanction relative to public officials. Political accountability is thus a formalized relationship of oversight and/or sanctions of public officials by other actors. In a relationship of political accountability, a public official gives a reckoning of the discharge of her public duties to actors that formally (via public law) have the capacity to demand such an accounting and/or to impose sanctions on the official. Thus, my understanding of political accountability hinges on whether an actor is formally ascribed the right to demand answerability of a public official or bureaucracy. When monitoring of public authorities takes place outside an institutionalized framework in which agents are formally charged with this responsibility, it falls outside the scope of my understanding of accountability.

Excluding non-legalized relationships from the concept of ‘accountability’ does not mean that relationships outside this boundary are less important than those that fall within it. The issue is one of conceptual demarcation, not political significance. As conceptualized here, ‘accountability’ implies not only answerability, but also the legal obligation to answer or the institutionalized right of an agent of accountability to impose sanctions on public officials. This focus on legalized authority to request answerability allows for a clearly delimited concept that still includes a range of relationships of answerability.

This definition excludes as agents of accountability the press and civil society organizations that investigate and denounce abuses and wrongdoings by public officials. Smulovitz and Peruzzotti call (p.8) such interactions ‘societal accountability’ in their chapter in this volume (see also Fox and Brown 1998). These exclusions may seem arbitrary in view of the undeniable fact that some social organizations and the press have a major impact, as Smulovitz and Peruzzotti argue, in exposing governmental wrongdoing and creating public oversight of government actions. Yet if we include all forms of public oversight or holding actors responsible, the concept becomes so elastic that it may not be useful.

Within this conception, two kinds of actors can provide political accountability. First, elected public officials are accountable to voters, at least in cases in which re-election is allowed. Second, many state agencies are formally charged with overseeing and/or sanctioning public officials and bureaucracies; I refer to these relationships as intrastate accountability. Examples include legislative committees that investigate possible wrongdoings of public officials, agencies that are created to monitor public officials and bureaucracies with an eye toward possible wrong doing, the legal system when it investigates the possible misdeeds of public officials and bureaucracies, the congress when it initiates a hearing against any public official, and the assembly in a parliamentary system when it makes a political judgment about whether to remove a minister or a cabinet. A third category analyzed by Smulovitz and Peruzzotti in this volume, societal oversight, falls outside the bounds of my definition of accountability, yet with important interactions with the formal network of institutions of accountability. Intrastate accountability and its interactions with electoral accountability and societal oversight are the primary focus of this book.

Democratic accountability refers to the accountability of two broad categories of public officials: elected and non-elected officials. Both of these broad categories should be accountable to different kinds of actors. Elected public officials are accountable first and foremost to the citizens whose votes put them into office. The ways in which elected politicians are accountable through the vote have been analysed at length by political scientists (Ferejohn 1999; Fiorina 1981; Key 1966; Maravall 1999; Powell Jr. 2000; Przeworski et al. 1999). Although this relationship between voters and representatives is not the primary subject of this volume, it is central to democratic politics.

In democracies, many non-elected officials are also presumably accountable to the executive or the legislature. These non-elected officials are in the first instance agents of the principals who appoint them. As with the relationship between voters and elected politicians, however, the relationship between non-elected officials (p.9) and the principal who appointed them (and can in theory remove them) allows ample room for shirking. Agents can acquire considerable autonomy, and mechanisms for overseeing and controlling them are often tenuous.

Competing Conceptions of Accountability

Five areas of conceptual disagreement run throughout works on accountability in this volume (and beyond). All five involve the boundaries of the concept: how broad or narrow should it be?

The most fundamental question is whether all activities that involve holding public officials responsible for their discharge of duties should be included under the very broad rubric of accountability. Many scholars accept such a broad understanding (Day and Klein 1987; Diamond, Plattner, and Schedler 1999; Fox and Brown 1998: 12; Merritt forthcoming; Moncrieffe 1998; Paul 1992). For example, Paul (1992: 1047) states that ‘Accountability means holding individuals and organizations responsible for their performance. Public accountability refers to the spectrum of approaches, mechanisms, and practices used by the stakeholders concerned with public services to ensure a desired level and type of performance.’ Day and Klein (1987: 5) provide another broad definition: ‘To account is to answer for the discharge of a duty or for conduct. It is to provide a reckoning. It is to give a satisfactory reason for or to explain. It is to acknowledge responsibility for one's actions.’ D. Dunn (1999), Keohane (2002), and Oakerson (1989) stipulate that accountability also necessarily involves the capacity of actors to impose sanctions on public officials. Yet their definitions remain broad by virtue of including a wide range of non-institutionalized kinds of answerability and sanctions. For example, Keohane mentions reputational accountability, by which an actor can informally demand answerability and impose reputational costs on a public official.

The authors in this volume who explicitly take a position on this issue (O'Donnell, Kenney, and Moreno et al.) all advocate a concept that is more bounded than simply explaining one's public duties in response to such a request.4 I agree that a more bounded concept is desirable and that it is important to specify what forms of (p.10) ‘answerability’ constitute accountability. A broad conception of accountability as answerability or responsibility makes it difficult to understand where the boundaries of accountability are. While intuitively sensible and in step with quotidian understanding, this definition could be extended to a vast array of different arenas and actions, including some that many analysts would not consider relationships of accountability. For example, a wealthy campaign contributor's request that a politician pursue some action could be seen as a request for answerability with a corresponding potential for a sanction (withdrawing campaign support) in the event of non-compliance. If we considered such informal interactions a relationship of accountability, the concept would encompass an enormous and ill-defined range. Actions as different as a citizen's letter requesting that a public official explain her view on an issue, a court injunction that a public official must obey some regulation, elections, a citizen's law suit against a public official, social mobilization to get public officials to undertake some action, media denunciations of some public policy or actor, and the oversight that one state agency exercises over another could all be construed as demands for public answerability. Or if public opinion in some diffuse way seems to request a change of course on the government's part, this could be considered a request for answerability and hence a relationship of accountability.

It is difficult to impose meaningful boundaries on the concept if every action by which some actor requests or demands answerability of a public official is considered ‘accountability’. Answerability combines many different agents and actions, and it makes the conceptual boundaries murky. My demarcation provides for much greater specificity and clarity than the excessively diffuse notions of answerability or holding actors responsible.

This first issue constitutes the biggest divide in conceptions of accountability. The remaining four sources of disagreement all revolve around how to delimit the quotidian use of ‘accountability’ as answerability.

A second source of disagreement is whether accountability should be restricted to cases of legal transgressions by state actors. Charles Kenney (this volume) and Guillermo O'Donnell (this volume) argue for such a restriction in cases of horizontal accountability. O'Donnell defined horizontal accountability as ‘the existence of state agencies that are legally enabled and empowered, and factually willing and able, to take actions that span from routine oversight to criminal sanctions or impeachment in relation to actions or omissions by other agents or agencies of the state that (p.11) may be qualified as unlawful.…(F)or this kind of accountability to be effective there must exist state agencies that are authorized and willing to oversee, control, redress, and/or sanction unlawful actions of other state agencies’ (1999a: 38, 39). In this volume, O'Donnell writes that horizontal accountability ‘deals exclusively with those (intrastate interactions) that…are undertaken with the explicit purpose of preventing, cancelling, redressing and/or punishing actions (or eventually non-actions) by another state agency that are deemed unlawful, whether on grounds of encroachment or of corruption’ (emphasis added).

Whereas O'Donnell and Kenney limit the notion of horizontal accountability specifically to perceived legal transgressions, most of the literature suggests that accountability also involves oversight (or monitoring) and sanctions related to political disagreements that do not involve such transgressions (e.g., Magaloni, this volume; Moreno et al., this volume; Schedler 1999; Schmitter 1999; Woodhouse 1994). O'Donnell's focus on perceived legal violations excludes institutionalized relationships of answerability when one actor politically disagrees with and calls for an account from a public official.

In my view, the notion of accountability, including what O'Donnell calls horizontal accountability and I call intrastate accountability, should extend beyond issues where the legality of a state actor's behavior is at stake. Intrastate answerability and sanctioning are not limited to perceptions about the legality of a public official's or agency's actions. Public officials and agencies must provide political as well as juridical accountings of the discharge of their public duties. One of the classic forms of intrastate accountability is cabinet (Laver and Shepsle 1999; Strøm 2000) and ministerial accountability (Woodhouse 1994) to the legislature. These refer mainly to political answerability and the possibility of removal from office. In contrast to the situation with impeachment in presidential systems, when in a parliamentary democracy a cabinet or minister answers to the legislature, the sanction of the legislature usually does not rest upon a perceived juridical infraction. The idea that cabinets and ministers are answerable to (and can be removed by) the legislature was one of the earliest uses of ‘accountability’ in politics, and it remains an important one. Consistent with this traditional usage of ‘accountability,’ and in contrast to the boundaries created by O'Donnell's focus on legal transgressions, the political answerability of ministers and cabinets to the legislature is a fundamental aspect of intrastate accountability.

(p.12) Kenney agrees that the relationship of cabinets to parliaments in parliamentary systems is one of accountability that rests on political judgements that are not limited to legal transgressions. He argues that in presidential systems, however, the notion of horizontal accountability should be limited to legal transgressions. This argument, however, raises a different difficulty: why should accountability in presidential systems be restricted to legal transgressions and encroachments when it is not so restricted in parliamentary systems?

As understood traditionally (Day and Klein 1987: 4–31; Woodhouse 1994: 3–39), the conception of accountability includes the answerability of public officials for the range of their public duties. By restricting the notion of horizontal accountability to perceived legal transgressions, Kenney and O'Donnell exclude political issues that in the dominant understanding of accountability should be included. Most chapters in this volume are predicated on this broader conception; they do not limit the analysis to cases of perceived legal transgressions.

This broader conception of accountability (including horizontal or intrastate accountability) is also more consistent with the common notion that elected political representatives are accountable to voters. Elected politicians are accountable to voters for their political actions regardless of whether they are perceived to have transgressed any constitutional norms. O'Donnell used the term ‘vertical’ accountability to refer to the oversight and sanctions that voters, the press, non-governmental organizations (NGOs), and other organizations of civil society exercise over public officials. For O'Donnell, vertical accountability includes political judgements as well as legal transgressions (most citizens vote on the basis of political judgements, not legal issues). It is not clear why one should restrict the concept of horizontal accountability to legal transgressions and encroachments while advocating a concept of electoral accountability and societal oversight that is more expansive and is principally based on political judgements. It seems preferable to have an isomorphic concept: if electoral accountability includes voters' judgements about political matters, so should accountability to state actors.

A third area of contention regarding the definition of accountability is whether the concept necessarily entails the capacity of sanctions on the part of the agent of accountability. D. Dunn (1999: 299), J. Dunn (1999: 337–8), Kenney (this volume), and Moreno et al. argue that a relationship of accountability exists only if the agent of accountability can impose sanctions on a (p.13) transgressor. In contrast, Schedler (1999) (explicitly) and O'Donnell (implicitly)5 argue that some mechanisms of accountability rely exclusively on answerability without necessarily having the capacity to impose sanctions. To disentangle this issue, I distinguish between direct legally ascribed sanctioning power and other sanctions. Accountability cannot exist with no sanctioning power; some capacity to redress wrongdoing by referring a case to other venues (especially the justice system) is critical to systems of accountability. As defined here, however, accountability does not require direct, legally ascribed sanctioning power. Agencies of oversight are expected to refer possible wrongdoings to actors that can impose sanctions; this indirect sanctioning power suffices to characterize a relationship of accountability.

Consider the institution of the ombudsman. The ombudsman is typically charged with investigating citizen complaints about improper state behaviour, ranging from state failure to protect constitutionally guaranteed rights to improper police conduct. The ombudsman fits within my definition of intrastate accountability in which one agent of accountability has a formalized authority of oversight over public officials even though the office cannot impose formal sanctions. Even without direct formal sanctioning power, the ombudsman generates a need for answerability of and accounting by public officials. The ombudsman in most countries is expected not merely to investigate abuses by public officials, but also to take some action to encourage redressing such wrongdoings. For example, the ombudsman may refer the case to the legal system. The failings of this institution (which has achieved important successes in some countries) have more to do with the shortcomings of the legal system in handling the cases referred to it than with the ombudsman's inability to directly prosecute cases. In a similar vein, the Public Prosecution in Brazil lacks direct sanctioning power but is expected to file law suits in cases of wrongdoings by public officials and agencies.

As Kenney and Moreno et al. imply, accountability is different if an agent of oversight also has formal sanctioning power. The difference, however, may be slim because agencies that oversee public officials can bring cases of wrongdoings to the legal system. (p.14) Conversely, many actors that can formally impose direct sanctions are de facto toothless. Thus, although sanctioning potential cannot be entirely absent in a relationship of accountability, it can be indirect (i.e., by referring the wrongdoing to the legal system) as well as direct. By delimiting the notion of accountability to formal institutional relationships, but not delimiting it to relationships that involve direct, legally ascribed sanctioning power, I am advocating a concept considerably narrower than that of answerability but broader than that proposed by Kenney, Moreno et al., and O'Donnell in this volume.

A fourth area of disagreement revolves around whether ‘accountability’ is limited to principal–agent relationships. In their chapter in this volume, Moreno, Crisp, and Shugart argue that ‘accountability’ should be restricted to such relationships, that is, relationships in which some principal commissions an agent to do something for him/her. In this conception, only if the principal can dismiss or decide against renewing the agent is there a relationship of accountability. To quote from their chapter, ‘Accountability means that the principal has the right to withdraw the conditionally delegated authority altogether.’ They argue that ‘when institutions are formally independent of one another—as with the independent branches of presidential systems—they are not accountable to one another.’6 Against this view, Kenney and O'Donnell in this volume argue that not all relationships of accountability involve a principal who can dismiss an agent. Most of the other chapters in this volume also understand relationships of accountability as extending beyond (though including) principal–agent relationships.

Restricting ‘accountability’ to principal–agent relationships calls attention to a substantive point: a relationship of power is different when a principal can dimiss or decide against renewing an agent compared to a context in which no such possibility exists. Under these circumstances, agents have an incentive to be responsible to a principal (though several contributions in Przeworski et al. (1999) question how compelling these incentives are in the voter–representative relationship). The advantages of limiting ‘accountability’ to principal–agent relationships, however, may be offset by other disadvantages—above all, the fact that this definition (p.15) excessively narrows the concept and leaves out some formalized relationships of oversight and potential sanctioning. Whether or not a principal–agent relationship exists, a public official may formally need to answer for her action to some actor and may be subject to sanctions in the event of wrongdoing.

By my definition above, accountability relationships exist whenever a public agency or official is formally (i.e., by law or public decree) answerable to another actor. In this conception, agencies of oversight and the judiciary are parts of the web of accountability because they are formally charged with overseeing and/or sanctioning public officials for the discharge of their public duties. From this perspective, principal–agent relationships are a subset of accountability relationships. Hence, like Kenney and O'Donnell, I diverge from Moreno et al. on this point. The ombudsman and some other agencies in Latin American countries are not principals; they were not responsible for electing or appointing the president and the congress, and they cannot dismiss the president or members of congress. Indeed, the reverse may be true: in Argentina, Ecuador, Guatemala, and Peru, the congress can appoint and dismiss the ombudsman. But the president and legislators may be required to answer to (provide accounting to) the ombudsman, the fiscalía, the controlador, and other mechanisms of oversight. Kenney and Moreno et al. exclude such agents of oversight though on slightly different grounds—Kenney because they have no formal direct sanctioning power, and Moreno et al. for this reason and also more fundamentally because they are not principals in a principal–agent relationship.

In a similar vein, if a legislative committee investigates some putative wrongdoing of the president, that committee is formally charged with the responsibility to review the president's actions. Even though this is not a principal–agent relationship (the president did not ‘hire’ and cannot dismiss the committee as her agent, nor did the committee ‘hire’ the president), in my understanding, there is a relationship of accountability because the president must answer to the committee. For Moreno et al., the judiciary for the most part is not part of the network of accountability because it is neither a principal nor an agent but rather an independent branch of the state. In contrast, in my conception, as well as Kenney's and O'Donnell's, the judiciary is a key institution of accountability; public officials who are accused of a legal transgression need to answer to the courts.

One final doubt about the argument of Moreno et al. on this point: it is questionable whether the concept ‘horizontal exchange’, (p.16) which they coin to refer to relationships of oversight and checks and balances that fall outside principal–agent relationships, serves us optimally. The notion of ‘exchange’ may not quite capture the interaction in a relationship of oversight. ‘Accountability’, ‘checks and balances’, ‘superintendence’, and ‘oversight’ are more specific and precise than ‘exchange’.

A fifth area of disagreement involves which actors can serve as the mechanisms for providing accountability. The relatively unbounded definition of accountability as answerability and informal sanctions would allow almost any actor to be an agent of accountability. In contrast, with the exception of Smulovitz and Peruzzotti, the authors in this volume focus on a narrower range of actors that have formalized responsibilities to oversee public officials. Within this fundamental agreement about a bounded set of actors that can constitute agents of accountability, the authors in this volume nevertheless have some differences of opinion that stem directly from the previous disagreements. Because of their definition of accountability, Moreno et al. limit these actors to voters, to whom elected representatives are accountable; politicians, to whom some bureaucracies are accountable; parliaments, to which cabinets and ministers are responsible in parliamentary democracies; and other principals in principal–agent relationships. They explicitly exclude oversight agencies and institutions that are independent from each other. Kenney also excludes oversight agencies but includes actors that have sanctioning power (the primary example is the judiciary). O'Donnell includes not only principals in principal–agent relationships, but also oversight agencies and the judiciary as agents of horizontal accountability—although all three kinds of actors only when the actor being investigated or sanctioned has committed a legal infraction. I share his list of agents of accountability, but as noted above disagree that they are acting as agents of accountability only in cases of legal transgressions.

These arenas of agreement and disagreement about the concept of accountability are summarized in Table 1.1. In the final row, ‘Agents of Accountability’, the scholars who do not limit accountability to formal authority to oversee and/or sanction public officials have a much less restricted conception than the two possibilities shown in Table 1.1.

Kenney (this volume) raises one other important conceptual issue: the relationship between horizontal accountability and checks and balances. He properly argues that these two concepts should not be confounded or conflated; not all checks and balances (p.17)

Table 1.1. Conceptualizing political accountability

More restricted conception

Less restricted conception

Object of Political Accountability

State actors (all authors in this volume, but for Moreno et al., limited to agents in principal–agent relationships; Merritt forthcoming)

State actors and others (Moncrieffe 1998)

Is accountability limited to formal (legalized) authority to oversee and/or sanction

Yes (Kenny, this volume; Mainwaring; Moreno et al., this volume)

No (Day and Klein 1987; Fox and Brown 1998; Keohane 2002; Merritt forthcoming; Moncrieffe 1998)

Is direct formally ascribed sanctioning power a necessary component of accountability

Yes (Kenney, this volume; Moreno et al., this volume)

No (Mainwaring; O'Donnell, this volume; Schedler 1999; all authors named in the cell immediately above)

Are accountability relationships limited to principal–agent relationships?

Yes (Moreno et al., this volume, Elster 1999; Laver and Shepsle 1999)

No (Kenney, this volume; Mainwaring, Merritt forthcoming; O'Donnell, this volume)

Agents of accountability

Principals in principal–agent relationships (Moreno et al., this volume Elster 1999; Laver and Shepsle 1999)

All actors that formally oversee and/or sanction public officials (Kenney, this volume; Mainwaring, this volume; O'Donnell, this volume)

are a subset of accountability. Political disagreement per se between the legislature and the executive does not necessarily constitute a relationship of accountability. For example, if the legislature refuses to approve an executive-sponsored bill, this does not intrinsically constitute a relationship of accountability because it does not involve an element of answerability or sanctioning. The distinction between accountability and checks and balances, however, is sometimes a fine one; checks and balances can constitute mechanisms of accountability. The key issue according to the definition proposed here is whether an actor has the constitutional/legal capacity to request an accounting of a public official's (or agency's) discharge of duties or to impose sanctions on that official. If so, it is (p.18) a relationship of accountability. When the legislature creates a commission to investigate executive actions, it is a form of oversight and creates a demand for answerability—hence constitutes a mechanism of accountability (see Figueiredo's chapter in this volume). In a similar vein, when the legislature charges a committee thereof to oversee an executive agency, a relationship of accountability exists.

A skeptic might wonder whether the concept of accountability can be rescued in light of the conceptual muddle that is reflected in Table 1.1. I believe that the answer is affirmative. Accountability is too substantively important a subject in the social sciences and in politics to abandon. Nonetheless, the vast differences that are synthesized in Table 1.1 indicate the need for more concerted theorizing about how to define and delimit the concept. Hopefully, this volume will advance thinking toward accomplishing this objective.

Distinguishing Among Types of Accountability

O'Donnell (1994, 1999a, 1999b, this volume) has been a pioneer in conceptualizing distinctions among different kinds of accountability. He developed a distinction between vertical and horizontal accountability. By vertical accountability, O'Donnell meant the accountability of state agents to citizens and to civil society. Elections are an important, indeed ‘arguably the main facet of vertical accountability’, but the notion of vertical accountability also includes actions by civil society and the media to expose ‘apparently wrongful acts of the public authorities’ (1999a: 30).

O'Donnell's distinction has set the agenda for contemporary debates about non-electoral forms of accountability, but it is questionable that his spatial metaphors provide the best way of differentiating among different kinds of accountability. His terminology has two problems. First, the physical metaphor conjured by the notion of vertical accountability is misleading except in principal–agent relationships. Vertical accountability expresses an image of asymmetry of power, that is, a hierarchical relationship. As Moreno et al. argue, the vertical metaphor can reasonably be extended to all principal–agent relationships, even to those in which the principal is in most respects vastly weaker than the agent. For example, they see voters as principals and elected officials as agents. The ‘vertical’ image, however, is inappropriate to describe relationships that do not have a hierarchical component. (p.19) O'Donnell (1999c) includes the monitoring by NGOs over public officials as an element of vertical accountability, but NGOs do not vertically control public officials in any way.

Second, O'Donnell's distinction conflates two different issues that should not be conflated: a physical metaphor that conveys images of independence (horizontality) and hierarchy (verticality), and the location of the agent of accountability (state versus societal actors, respectively). O'Donnell (1999c) stated that vertical accountability refers to voters and societal organizations while horizontal accountability refers to state and regime agencies. In his chapter in this volume, Kenney also views the distinction between horizontal and vertical accountability as resting on whether the agent of accountability is a state or societal actor. Yet the correspondence between horizontal and vertical relationships, on the one hand, and intrastate and electoral accountability (and societal oversight), on the other, is not perfect.

Principal–agent relationships can be seen as vertical: the principal commissions and can ‘fire’ the agent (see Moreno et al., this volume). The accountability of a cabinet to the assembly in a parliamentary system is clearly a case of what I call intrastate accountability, and it is also a principal–agent relationship. By suggesting that accountability is horizontal when state actors must answer to other state actors, Kenney and O'Donnell would consider this an example of horizontal accountability. Yet in constitutional terms, it is a vertical relationship; the cabinet serves at the disposition of the assembly. In a similar vein, relationships between the executive or the assembly and some bureaucracies are vertical in the sense that the executive or assembly creates the bureaucracy and formally has control over it, yet at the same time it is an intrastate relationship. In short, some intrastate relationships are vertical, at least in formal terms. For this reason, it is problematic to equate intrastate accountability with horizontal accountability, and more broadly it is problematic to combine, as Kenney and O'Donnell do, the vertical/horizontal distinction with a distinction based on the location (in society versus in the state) of the agent of accountability.

O'Donnell's categories of horizontal and vertical accountability largely (except in cases of legal transgressions) exclude one of the classic uses of ‘accountability’, namely, the accountability of cabinets and ministers to parliaments in parliamentary systems. According to his definition, only if the parliament removed a minister or cabinet because of legal transgressions would this be an instance of horizontal accountability. Because one state actor (the (p.20) cabinet or minister) must answer to another state actor (the assembly), it is not a case of vertical accountability according to his definition.

Although O'Donnell's notions of horizontal and vertical accountability are problematic, it is useful to differentiate among different kinds of accountability. It is more fruitful to conceptualize these distinctions strictly according to the agent of accountability.7 For the sake of parsimony, I propose a distinction between two main kinds of political accountability: electoral accountability or (synonymously) accountability to voters and intrastate accountability. This distinction parallels O'Donnell's but without the connotations of hierarchy and independence suggested by his metaphor, and without conflating the agent of accountability (state versus societal) with the nature of the relationship (horizontal versus vertical).

Intrastate accountability can be usefully subdivided into three different kinds of relationships. First, in principal–agent relationships, a principal commissions an agent to perform some function and has ultimate control over that agent. Bureaucracies created and ultimately controlled by the executive or legislature are prime examples. Moreno et al. restrict their understanding of ‘accountability’ to this first subcategory. Second, the legal system and on occasions some other state actor (example: a legislature that is trying a president on impeachment hearings) can impose sanctions on public officials accused of wrongdoing. These are not principal–agent relationships, but rather can be thought of as ‘sanctioning actors’. The most common sanctioning actors are within the justice system. The congress becomes a sanctioning actor when it tries the president on impeachment charges. Kenney and O'Donnell include sanctioning actors as well as principal–agent relationships in their rubric of horizontal accountability. Finally, oversight actors have responsibility for monitoring the behavior of state officials and agencies. Kenney excludes oversight actors that do not have legally ascribed, direct sanctioning capacity, and O'Donnell includes them only when a legal transgression is at play.

(p.21) Interactions Between Electoral and Intrastate Accountability

A second key contribution of this volume is an effort, thanks above all to Moreno et al., to advance the debate about interactions among electoral accountability, intrastate accountability, and societal oversight. Rarely have these interactions been discussed in as detailed a manner as they are here—especially by Magaloni, Moreno et al., O'Donnell, and Smulovitz and Peruzzotti. On this point, too, we hope to advance the debate not through consensus but rather through explicit disagreement.

Moreno et al. argue that the linkage between voters and elected representations sets the tone for all other accountability relationships. In their argument, by getting electoral delegation and accountability right, politicians will address the most serious deficiencies in accountability. To quote from their chapter, ‘The deficit of accountability lies in faulty vertical accountability—legislators who do not represent the values and preferences of the broad citizenry. If the accountability of elected officials were working as intended—such that voters could and would punish misdeeds—separate agencies of superintendence would be unnecessary.’ Although none of the other authors take this argument as far, all agree that electoral accountability is important for intrastate accountability as well. But other contributors including O'Donnell argue that electoral and intrastate accountability have strong mutual effects. By this argument, electoral accountability and delegation do less to resolve the problems of intrastate accountability than Moreno et al. argue. Conversely, by this argument, it is possible to make greater gains in intrastate electoral accountability than Moreno et al. argue even without touching electoral rules. For example, Smulovitz and Peruzzotti argue that social mobilization, especially in conjunction with media coverage, can trigger improved intrastate accountability. Sadek and Cavalcanti document the impact of Brazil's Public Prosecution, which has enhanced intrastate accountability despite the lack of major electoral and party reform since 1988.

Although the contribution of Moreno et al. is an innovative and important work, intrastate and electoral accountability are more mutually interactive than they indicate. They overestimate the capacity to improve intrastate accountability by enhancing electoral accountability and understate the capacity to enhance intrastate accountability without changing electoral accountability. (p.22) It is questionable that the primary key to better intrastate accountability is balance in the continuum from a party centric to an individual centric electoral system for the national congress. They place great optimism in the capacity to effect change in the functioning of intrastate accountability through electoral reform.

The ‘accountability deficiency’ can be seen as a product of feeble sanctions. Whereas Moreno et al. locate the inefficacy of sanctions in electoral systems that set up inadequate mechanisms of accountability between voters and elected representatives, perhaps even more important are the failures of sanctions that should be imposed by the justice and penal systems for wrongdoings of public officials. The justice system works reasonably well at the upper echelon in some countries (see Magaloni, this volume), but in most of Latin America it is notoriously bad at the local level, where most cases are tried. In much of the region, at the local level, the ‘justice’ system (‘injustice’ system frequently more appropriately captures the reality) is still under the sway of powerful elites (Dodson and Jackson, this volume; Magaloni, this volume; Pásara 2002). In several countries, at the local level, drug barons with seemingly unlimited resources to buy off and intimidate the police, judges, and witnesses have captured the justice system. When public officials are confident that they can commit wrongdoings without facing penal and legal sanctions, electoral reform is not likely to fix the problem. Greatly compounding this problem is the inefficacy and frequent complicity of police forces in much of Latin America; they, too, are frequently captured and paid off by local elites from drug barons to landowners and traditional oligarchs.

With an important original argument and a wealth of new empirical information, Moreno et al. show that there is great variance in how independent one would expect oversight agencies to be on the basis of institutional arrangements (Figure 4.3 and the Appendix of their chapter). They argue that such agencies are likely to be more independent from political pressures, and hence more effective as mechanisms of oversight, if (a) they are not named directly and exclusively by the legislature and (b) they have long terms of office, such that they need not depend on the ongoing support of politicians for remaining in office. They conclude that many oversight agencies are insufficiently insulated from political pressures to be effective and that the only effective way to enhance accountability is the vertical (electoral) linkage between voters and representatives. But this argument overlooks the possibility of trying to boost the effectiveness of oversight agencies by making them more independent of political pressures. By the logic of their (p.23) argument,8 it should be possible to enhance the effectiveness of oversight agencies by changing the appointment procedures and lengthening the terms of office—a change that is in principle independent of changes in electoral accountability.

Moreno et al. are sanguine about the effects of electoral reform on intrastate accountability. Such optimism assumes that the connection between elected representatives and oversight and sanctioning agencies within the state is fairly tight: intrastate accountability (superintendence in the lexicon of Moreno et al.) will improve because elected politicians find it in their interest to make these institutions work better. This argument probably overstates the extent to which members of congresses would effect improvements in the justice system and oversight agencies even if stronger mechanisms of electoral accountability gave these politicians a bigger stake in national policy outcomes.

The argument also may be excessively sanguine about the effectiveness of electoral accountability between voters and members of national congresses in Latin America. Electoral accountability offers politicians enormous opportunities to shirk (Ferejohn 1999; Maravall 1999; Stokes 1999). Elections occur intermittently, usually every two to eight years, and nothing assures that elected representatives will behave as voters would prefer between elections. Tremendous information asymmetries between elected officials and the average voter give the former ample opportunities to behave with autonomy vis-à-vis the latter (Przeworski et al. 1999). As Schumpeter (1950: 256–64) and Downs (1957) argued long ago, the average voter cannot invest the considerable time needed to closely monitor representatives' actions. These information asymmetries are larger in Latin America, where most voters have limited education and little information about politics, than in the advanced industrial democracies.

Even if voters become disenchanted with a politician or a party in power, they may not believe that other options are better. The ‘supply’ of politicians or parties may appear to voters to be oligopolistic or quasi-monopolistic. This feeling that all politicians or parties are the same diminishes the extent to which elections serve as instruments of accountability: if voters do not believe that the competition is any better, they are less likely to punish (p.24) incumbents by voting them out of office. In sum, as Ferejohn (1999: 137) succinctly points out, ‘Electoral punishment…is a fairly blunt instrument, and incumbent officials will be, at best, only moderately responsive to public wishes.’ (See also J. Dunn 1999: 338–9; Fearon 1999; Manin et al. 1999; Maravall 1999.)

This book does not purport to resolve the important issues related to the interactions among electoral accountability, intrastate accountability, and societal oversight. It does, however, set out a new research agenda on this issue. The authors do not converge in their understanding of how electoral and intrastate accountability interact, but they advance the debate about this important subject.

Intrastate Accountability in Latin America

A third key debate in this book revolves around the quality of intrastate accountability in Latin America. In the past decade, accountability became salient in reflections about democracy among such organizations as the World Bank and the Interamerican Development Bank, as well as a host of NGOs.

One of O'Donnell's central claims is that mechanisms of horizontal accountability are weak in contemporary Latin America. Indeed, he sees the weakness of horizontal accountability as one of the most important differences between the old and wealthy democracies and the post-1978 democracies of Latin America (and elsewhere). His work over the last decade has addressed differences between these older democracies and Latin American regimes and called attention to deficiencies of elected governments in Latin America. In a seminal article (1994), he argued that weak horizontal accountability characterizes most of Latin America's elected governments—in particular, those that he labelled ‘delegative democracies’. ‘The horizontal accountability characteristic of representative democracy is extremely weak or non-existent in delegative democracies’ (61). Subsequent articles (1999a, 1999b) furthered this argument.

Most scholars accept O'Donnell's viewpoint (e.g., Diamond et al. 1999: 1). While diverging from O'Donnell on other points, Moreno et al. (this volume) agree that there is an accountability deficit in the region. In a similar vein, Morgenstern and Manzetti note the deficiencies in mechanisms of oversight and corruption control in contemporary Argentina compared to the United States, Dodson and Jackson underscore the severe deficiencies of the judiciary in (p.25) contemporary El Salvador and Guatemala (see also Pásara 2002). These deficiencies make it difficult for these judiciaries to serve either of the two functions central to O'Donnell's notion of horizontal accountability: preventing corruption and preventing improper state encroachments.

Other authors in this volume, however, point to innovations in intrastate accountability in the region. The authors who focus on innovations show that in some countries, there is potential to redress the accountability deficit. Smulovitz and Peruzzotti note promising innovations in what they call ‘societal accountability’, and I call societal oversight. Civil society has organized to oversee public officials and agencies. Civil society organizations denounce, mobilize, and forge alliances with the independent media to call attention to problems in the discharge of public officials.

Sadek and Cavalcanti (this volume) analyze an old but transformed Brazilian institution whose mandate in the 1988 constitution is broad: the Public Prosecutor's Office (Ministério Público). This institution is charged with undertaking criminal prosecution, defending collective rights and public goods, and defending minority rights. The 1988 constitution devised multiple means to ensure the autonomy of the Public Prosecutor's Office vis-à-vis politicians. Its oversight of and eventually prosecution of state actors (including individuals in the executive, legislative, and judicial branches) falls within the broad domain of accountability. Their chapter underscores the innovative and transformative potential that this institution has manifested. Their portrait of the judiciary and the Public Prosecutor's Office differs sharply from the one that Dodson and Jackson draw of the Salvadorean and Guatemalan judiciaries. It suggests that some mechanisms of accountability in the region are innovative, vibrant, and somewhat powerful. Morgenstern and Manzetti, while underscoring the limitations of mechanisms of oversight in Argentina, also note some promising developments in recent years.

Figueiredo presents a mixed assessment of intrastate accountability in contemporary Brazil. She argues that several potentially important mechanisms of oversight—parliamentary investigative commissions, the Federal Accounting Tribunal (Tribunal de Contas da União), the Joint Congressional Budgetary Committee (Comissão Mista de Orçamento), and a wide range of other tools at the hands of the legislature—have been relatively toothless in practice. As mechanisms of intrastate accountability, they have not been as effective as their legal mandates allow. On the other hand, she argues that they generate information and stimulate (p.26) debate, both of which might be important in triggering mechanisms of electoral accountability.

I generally agree with Moreno et al. and O'Donnell that mechanisms of intrastate accountability tend to be weak in contemporary Latin America. Yet I would qualify this generalization in two ways. First, the quality of intrastate accountability in contemporary Latin America is more promising than many critics suggest, not because a critical assessment is usually wrong but rather because it understates variance. An earlier article of O'Donnell's (1993) properly emphasized gross within country differences in the enforcement of the rule of law (see also Magaloni, this volume). Some Latin American countries (e.g., Brazil) have among the starkest intracountry socioeconomic and political differences in the world. O'Donnell's 1993 article suggested that one key actor of intrastate accountability, the judiciary, is less effective and more patrimonial in the poor ‘brown’ areas of Latin America than in the developed areas (‘the green areas’). By implication, analysts need to be attentive to both the serious deficiencies of accountability with some institutions and in the poor regions and to innovations and promising developments in other institutions, especially in the more developed regions.

Collectively, the essays in the volume strike such a balance, and some chapters, especially Magaloni's, simultaneously call attention to the glaring deficiencies of accountability at some levels and the striking advances at others. Magaloni's argument about intrastate accountability and the courts is two-sided. On the one hand, she notes that the Supreme Court has become more independent vis-à-vis the president and has become a more significant actor in preventing and sanctioning infractions of public officials. This strengthening of the Supreme Court better enables it to exercise the functions associated with intrastate accountability. On the other hand, Magaloni emphasizes the ongoing subordination of local judges and courts to state governors and impunity at the local level. Law-enforcement institutions such as the police and local courts have remained authoritarian enclaves, and they are ineffectual as mechanisms of accountability. The weakness of local courts and law-enforcement agencies means effective impunity for state actors who violate the law—a key aspect of O'Donnell's argument about the weakness of horizontal accountability.

In many countries, Latin American legislatures are important sources of oversight of presidential action. The portrait of delegative democracy that O'Donnell drew in 1994, in which presidents dominated the political systems and mechanisms of horizontal (p.27) accountability were very weak, is exaggerated in many countries. The judiciaries are undertrained and ineffective in most of the region (see Dodson and Jackson, this volume; Pásara 2002), but even seemingly compliant judiciaries sometimes rear their heads and declare presidential initiatives unconstitutional (Helmke 2000, 2002). As Sadek and Cavalcanti (this volume) argue in their chapter on the Ministério Público in Brazil, in some countries, important innovations are surfacing within the judiciary and related institutions. Indeed, it seems likely that some Latin American countries today have stronger mechanisms of intrastate accountability than the United States did in the 19th century. This is striking because most analysts see the Latin American judiciary as woefully inadequate.

Some nontraditional mechanisms of oversight have become visible and effective institutions. For example, under the leadership of Jorge Santiestevan, the Ombudsman (Defensor del Pueblo) in Peru acquired notable autonomy and efficacy even as President Alberto Fujimori crushed most of the rest of Peru's network of accountability institutions. As Kenney notes, Santiestevan's impact was limited by his need to rely on sanctioning agents—the courts, whose independence was undermined by Fujimori. Nonetheless, even under the highly inauspicious conditions of Fujimori's regime, new mechanisms of accountability could work.

In his analysis of horizontal accountability (though not in some of his other work), O'Donnell generalized too much across countries. His 1994 article on delegative democracy carefully distinguished between Chile and Uruguay, on the one hand, and Peru, Argentina, and Brazil, on the other. He noted that Chile and Uruguay have a more solid network of democratic institutions, hence do not fit the category of delegative democracies. But it tended to lump together Argentina, Brazil, and Peru. Peru under Fujimori fits O'Donnell's description of feeble mechanisms of intrastate accountability. Fujimori emasculated the judiciary, the legislature, and other institutions of intrastate accountability.

In the 1990s, mechanisms of intrastate accountability were less emasculated in Argentina than in Fujimori's Peru. Menem's temporary concentration of power in the early 1990s was a reflection more of Peronista majorities in both chambers of congress and of the extraordinary economic crisis of the late 1980s and early 1990s (hyperinflation and steep economic decline) than of the normal functioning of Argentine democracy. After 1993, the judiciary dealt Menem some key setbacks (Helmke 2000, 2002) notwithstanding its failure to address notorious corruption within his administration. (p.28) Mechanisms of corruption control were feeble, but some mechanisms for overseeing the president politically were important.

Brazil also has mechanisms of intrastate accountability more robust than O'Donnell's characterization of delegative democracies suggested. The judiciary has many shortcomings, but it is becoming an activist judiciary that is not readily dominated by the president or by other executive authority, except at the local level (as Magaloni argues occurs in Mexico), especially in the poor regions. Sadek and Cavalcanti (this volume) note the activist nature of the Public Prosecutor's Office in Brazil. The impeachment of President Fernando Collor de Melo in 1992 showed a willingness and capability of the legislature to exercise a dramatic form of accountability. It was followed by presidential impeachments in Venezuela in 1992 and Paraguay in 1999. Pérez-Liñán (2001) has aptly termed these impeachments cases of spasmodic and politicized horizontal accountability.

Second, the contrast that O'Donnell draws between the older democracies, in which mechanisms of horizontal accountability are generally effective, and Latin America's elected regimes, in which they are purportedly much weaker, is too stark if the comparison is intended to illuminate different points in the development of democracy (Schmitter 1999). Latin America's elected regimes have weaker mechanisms of intrastate accountability than the older contemporary wealthy democracies. But it is questionable whether that is the best temporal comparison. If instead we compare today's elected regimes of Latin America with the (now) older democracies during their younger years, then it is less certain whether there is a stark contrast in the efficacy of horizontal accountability. As the chapter by Morgenstern and Manzetti shows, most mechanisms of intrastate accountability were woefully weak in the 19th century United States. The spoils system was pervasive and powerful, and it was predicated on the idea that public agencies would not check executive power. The executive branch generally appointed and could easily remove those who served in public agencies.

These disagreements with O'Donnell's arguments do not take away from their importance. He galvanized a rich discussion on a vital and previously understudied subject. Moreover, if Latin American governments are not able to meet citizen needs, especially in the areas of material well-being and citizen security, there is a grave risk of growing dissatisfaction with democracy. This dissatisfaction could generate support for populists such as Fujimori and Venezuelan President Hugo Chávez (1998–present). These two (p.29) presidents severely damaged mechanisms of democratic accountability, and it is no accident that they came to power in response to particularly severe (Peru) or protracted (Venezuela) economic demise (and a serious terrorist threat as well in Peru). If citizens believe that their physical security and livelihood are at stake, they are more apt to turn an ear to politicians who promise government efficiency even at the expense of greatly diminished accountability. Thus, although several chapters in this volume analyze promising innovations in intrastate accountability and societal oversight, the dismal performance of many Latin American democracies in providing material well-being and citizen security may open the doors to regressions. Such regressions would mean a growing approximation to O'Donnell's portrait of extremely weak intrastate accountability.

The different portraits drawn by those who emphasize innovations in intrastate accountability and those who emphasize ongoing failures suggests the need for a clearer empirical comparative mapping of Latin America, by country, region, and specific agent of accountability. The negative overall portrait is surely largely correct, but it is far from uniform.

The Book

However contestable the boundaries of the concept are, and however difficult it is to assess and measure, ‘accountability’ has for at least a century been a key part of the lexicon of democratic politics. Democracy cannot exist without accountability. The concept is complex and difficult to measure, but turning our backs on what has become a key issue in Latin America and many new democracies elsewhere is ill-advised. Moreover, because of O'Donnell's contributions, the concept has recently taken center stage in theoretical thinking about new democracies. Both because the theme has assumed visibility in Latin America and other new democracies around the world, and because of its growing theoretical importance in understanding the dynamics and deficiencies of democracy, organizing a book around the theme of accountability seemed potentially fruitful.

This volume advances thinking about the web of institutions that together form the mechanisms of accountability, and about the interaction among these institutions and the interaction between electoral accountability, intrastate accountability, and societal oversight. The advantage of cutting across a variety of institutions (p.30) and focusing on accountability, as this volume does, is that institutions interact to form a web of mechanisms of accountability.

The volume deliberately brings together authors from different theoretical perspectives working on the same theme. Moreno et al. work from a political economy approach whereas Kenney and O'Donnell are working from more eclectic approaches. These theoretical differences are reflected in the diverging definitions of accountability. This diversity of meta-theoretical frameworks enriches the volume by allowing for the kind of direct clashes that can advance thinking in the social sciences. It also enhances the volume because both conceptions of accountability and both theoretical frameworks illuminate important dimensions of the problem.

Two mainly conceptual and theoretical chapters about accountability follow. These chapters lay the groundwork for chapters that, although theoretically driven, also have an empirical focus. Part II has three chapters on legislatures, executives, and oversight agencies. Part III has three chapters on the judiciary and related mechanisms of accountability. Finally, Part IV discusses the important innovations in oversight of public officials made possible by organizations in civil society.


Bibliography references:

Day, Patricia, and Rudolf Klein. 1987. Accountabilities: Five Public Services. London and New York: Tavistock.

Diamond, Larry, Marc F. Plattner, and Andreas Schedler. 1999. ‘Introduction’. In Andreas Schedler, Larry Diamond, and Marc F. Plattner, eds., The Self-Restraining State: Power and Accountability in New Democracies, pp. 1–10. Boulder, CO: Lynne Rienner.

Downs, Anthony. 1957. An Economic Theory of Democracy. New York, NY: Harper & Row.

Dunn, Delmer D. 1999. ‘Mixing Elected and Nonelected Officials in Democratic Policy Making: Fundamentals of Accountability and Responsibility’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 297–325. Cambridge: Cambridge University Press.

Dunn, John. 1999. ‘Situating Political Accountability’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 329–44. Cambridge: Cambridge University Press.

(p.31) Elster, Jon. 1999. ‘Accountability in Athenian Politics’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 253–78. Cambridge: Cambridge University Press.

Fearon, James D. 1999. ‘Electoral Accuntability and the Control of Politicians: Selecting Good Types versus Sanctioning Poor Performance’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 55–97. Cambridge: Cambridge University Press.

Ferejohn, John. 1999. ‘Accountability and Authority: Toward a Theory of Political Accountability’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 131–153. Cambridge: Cambridge University Press.

Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven, CT: Yale University Press.

Fox, Jonathan A., and L. David Brown. 1998. ‘Introduction’. In Jonathan A. Fox and L. David Brown, eds., The Struggle for Accountability: The World Bank, NGOs, and Grassroots Movements, pp. 1–47. Cambridge, MA: The MIT Press.

Helmke, Gretchen. 2000. Ruling Against the Rulers: Court-Executive Relations in Argentina Under Dictatorship and Democracy. Ph.D. dissertation, University of Chicago.

—— . 2002. ‘The Logic of Strategic Defection: Court–Executive Relations in Argentina Under Dictatorship and Democracy’. American Political Science Review 96, 2 (June): 291–303.

Keohane, Robert O. 2002. ‘Global Governance and Democratic Accountability’. Unpublished paper.

Key, Vernon O. 1966. The Responsible Electorate. New York: Vintage Books.

Laver, Michael and Kenneth A. Shepsle. 1999. ‘Government Accountability in Parliamentary Democracies’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 279–96. Cambridge: Cambridge University Press.

Manin, Bernard, Adam Przeworski, and Susan C. Stokes. 1999. ‘Elections and Representation’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 29–54. Cambridge: Cambridge University Press.

Maravall, José María. 1999. ‘Accountability and Manipulation’. In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 154–96. Cambridge: Cambridge University Press.

Merritt, Martha. Forthcoming. Accountability in Russian Politics.

Moncrieffe, Joy Marie. 1998. ‘Reconceptualizing Political Accountability’. International Political Science Review 19, 4 (October): 387–406.

Oakerson, Ronald J. 1989. ‘Governance Structures for Enhancing Accountability and Responsiveness’. In James L. Perry, ed., Handbook of Public Administration, pp. 114–30. San Francisco, CA: Jossey-Bass Publishers.

(p.32) O'Donnell, Guillermo. 1993. ‘On the State, Democratization, and Some Conceptual Problems: A Latin American View with Glances at Some Postcommunist Countries’. World Development 21, 8: 1355–69.

—— . 1994. ‘Delegative Democracy’. The Journal of Democracy 5, 1 (January): 55–69.

—— . 1999a. ‘Horizontal Accountability in New Democracies’. In Andreas Schedler, Larry Diamond, and Marc F. Plattner, eds., The Self-Restraining State: Power and Accountability in New Democracies, pp. 29–51. Boulder, CO: Lynne Rienner.

—— . 1999b. ‘Polyarchies and the (Un)Rule of Law in Latin America: A Partial Conclusion’. In Juan Méndez, Guillermo O'Donnell, and Paulo Sérgio Pinheiro, eds., The (Un)Rule of Law and the Underprivileged in Latin America, pp. 303–37. Notre Dame, IN: University of Notre Dame Press.

—— . 1999c. ‘A Response to My Commentators’. In Andreas Schedler, Larry Diamond, and Marc F. Plattner, eds., The Self-Restraining State: Power and Accountability in New Democracies, pp. 68–71. Boulder, CO: Lynne Rienner.

Pásara, Luis. 2002. ‘Justicia y ciudadanía realmente existentes’. Política y gobierno IX, 2 (Second semester): 361–402.

Paul, Samuel. 1992. ‘Accountability in Public Services: Exit, Voice, and Control’. World Development 29, 7 (July): 1047–60.

Pérez-Liñán, Aníbal. 2001. ‘Crisis without Breakdown: Presidential Impeachment in Latin America’. Ph.D. dissertation, University of Notre Dame.

Powell Jr., G. Bingham. 2000. Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven, CT: Yale University Press.

Przeworski, Adam, Susan C. Stokes, and Bernard Manin, eds. 1999. Democracy, Accountability, and Representation. Cambridge: Cambridge University Press.

Schedler, Andreas. 1999. ‘Conceptualizing Accountability’. In Andreas Schedler, Larry Diamond, and Marc F. Plattner, eds., The Self-Restraining State: Power and Accountability in New Democracies, pp. 13–28. Boulder, CO: Lynne Rienner.

Schedler, Andreas, Larry Diamond, and Marc F. Plattner, eds. 1999. The Self-Restraining State: Power and Accountability in New Democracies. Boulder, CO: Lynne Rienner.

Schmitter, Philippe C. 1999. ‘The Limits of Horizontal Accountability’. In Andreas Schedler, Larry Diamond, and Marc F. Plattner, eds., The Self-Restraining State: Power and Accountability in New Democracies, pp. 59–62. Boulder, CO: Lynne Rienner.

Schumpeter, Joseph. 1950. Capitalism, Socialism, and Democracy, 3rd edn. New York, NY: Harper & Row.

Smulovitz, Catalina, and Enrique Peruzzotti. 2000. ‘Societal Accountability in Latin America’. The Journal of Democracy 11, 4 (October): 147–58.

(p.33) Stokes, Susan C. 1999. ‘What Do Policy Switches Tell Us about Democracy?’ In Adam Przeworski, Susan C. Stokes, and Bernard Manin, eds., Democracy, Accountability, and Representation, pp. 98–130. Cambridge: Cambridge University Press.

Strøm, Kaare. 2000. ‘Delegation and Accountability in Parliamentary Democracies’. European Journal of Political Research 37, 3: 261–89.

Woodhouse, Diana. 1994. Ministers and Parliament: Accountability in Theory and Practice. Oxford: Clarendon Press.


(*) I am grateful to Charles Kenney, Luigi Manzetti, Martha Merritt, Guillermo O'Donnell, Matthew Shugart, and Chris Welna for insightful suggestions and to David Altman for research assistance.

(1) By governmental effectiveness, I mean simply the ability of governments to obtain good policy results.

(2) For example, accountability was a prominent theme of the Quebec Summit of the Americas, signed by the presidents of the hemisphere. See http://www.summit- americas.org/List-Summit(new)-Eng.htm

(3) Even this point is not unanimous beyond the authors of this volume; some scholars prefer a less bounded concept. Moncrieffe (1998: 392) argued that ‘(T)he concept of accountability must focus not merely on the responsibilities to the governed of those who are elected to govern but also on the relations among groups or factions and their responsibilities to elected officials’. She avers that citizen willingness to participate in politics, the degree to which citizen rights are effectively enforced and institutionalized, and the extent to which legal mechanisms check the actions of factions and interest groups should be included in measuring accountability (394). She also argues that the notion of democratic accountability should extend beyond the accountability of public officials. I restrict ‘political accountability’ to the answerability of public officials and agencies to other actors.

(4) However, they delimit the concept for different reasons and in different ways. O'Donnell and Kenney limit the notion of horizontal accountability to legal transgressions. Kenney and Moreno et al. limit ‘accountability’ to cases in which the agent of accountability has the capacity to impose sanctions. Moreno et al. limit ‘accountability’ to principal–agent relationships.

(5) O'Donnell (1999a: 30) includes as agents of horizontal accountability agencies such as the ombudsman, fiscalías, controllers, etc., that usually have no direct sanctioning authority. His chapter in this volume likewise does not restrict the notion of accountability to cases in which the agent of accountability can impose direct formal sanctions.

(6) Elster (1999), Fearon (1999), Laver and Shepsle (1999), and Strøm (2000) also analyse accountability in the context of principal–agent relationships, but they do not explicitly rule out the possibility that accountability could be extended to non-principal–agent relationships.

(7) Moreno et al. also avoid distinguishing according to both the agent of accountability and the verticality or horizontality of the relationship, but they advocate the opposite way of resolving this tension. Rather than distinguishing between different types of accountability on the basis of the agent of accountability, they distinguish between vertical (principal–agent) relationships and relationships of horizontal exchange.

(8) In a region where many institutional reforms have produced less than their supporters hoped for (Pásara 2002), we cannot be sure that such reforms of institutional arrangements would really produce the desired effect of enhancing the independence of oversight agencies.