Politician's Dilemma: Building State Capacity in Latin America (California Series on Social Choice and Political Economy, No.25): Volume 25 - Softcover

Geddes, Barbara

Softcover

ISBN 10: 0520207629 ISBN 13: 9780520207622

Publisher: University of California Press, 1996

Synopsis

In Latin America as elsewhere, politicians routinely face a painful dilemma: whether to use state resources for national purposes, especially those that foster economic development, or to channel resources to people and projects that will help insure political survival and reelection. While politicians may believe that a competent state bureaucracy is intrinsic to the national good, political realities invariably tempt leaders to reward powerful clients and constituents, undermining long-term competence. Politician's Dilemma explores the ways in which political actors deal with these contradictory pressures and asks the question: when will leaders support reforms that increase state capacity and that establish a more meritocratic and technically competent bureaucracy?

Barbara Geddes brings rational choice theory to her study of Brazil between 1930 and 1964 and shows how state agencies are made more effective when they are protected from partisan pressures and operate through merit-based recruitment and promotion strategies. Looking at administrative reform movements in other Latin American democracies, she traces the incentives offered politicians to either help or hinder the process.

In its balanced insight, wealth of detail, and analytical rigor, Politician's Dilemma provides a powerful key to understanding the conflicts inherent in Latin American politics, and to unlocking possibilities for real political change.

About this title

About the Author

Barbara Geddes is Assistant Professor of Political Science at the University of California, Los Angeles.

Excerpt. © Reprinted by permission. All rights reserved.

Politician's Dilemma

Building State Capacity in Latin AmericaBy Barbara Geddes

University of California Press

Copyright © 1996 Barbara Geddes
All right reserved.
ISBN: 9780520207622
One
The State

The state has been brought back in. It now lumbers through the halls of academe, a great clumsy creature that no one quite knows what to do with. Some have suggested putting it back out, but that, as often happens with strays let in from the cold, has proved difficult to do. Theoretically unwieldy as "the state" has proved to be, the role states1 play in influencing, managing, and sometimes transforming economies and societies compels the attention of social scientists. Increasingly, researchers have endeavored to take into account the use by national leaders of the unparalleled resources and power of the state to change society; to foster economic development; to provide citizens with education, medical care, sanitation, roads, and other amenities of modernity; and even to alter age-old social structures and cultural patterns.

Such state-led development has rendered inadequate old theories, both those that assumed an automatic diffusion of modernization via markets and spontaneous cultural change, and those that viewed political outcomes as the result of class struggle. At one time, comparativists looked primarily to traits of the domestic society and culture or to

The word "state" is used here to refer to regimes, governments, specific administrative agencies, and even individuals who act on behalf of governments. This usage is, I think, consistent with that in much of the literature, despite many efforts to define the state in a more precise manner. Talking about "the state" is a shorthand way of talking about the actions of individuals whose power derives from their positions in government. When dealing with the state as a political actor, one must refer to actions carried out by human beings. Even the state conceived as an expression of class interests "requires an organization which, since it cannot be other than a social network of people, exists in its own right and possesses interests of its own" (Cardoso 1979, 51).

characteristics of the international environment for explanations of economic development and change. They tended to view the state as simply reflecting outcomes caused by the struggle among private interests to secure policies beneficial to themselves. Now, however, most scholars working on these issues, no matter what their theoretical and ideological commitments, believe that governments can and often do act independently of underlying socioeconomic forces. In consequence, observers have recognized the need for theories to explain the state's role in bringing about change.

Scholars working on a remarkably wide spectrum of questions2 have concluded that states are loci of power, resources, and interests. They have shown that state officials sometimes have policy preferences independent of those of major social and economic groups in society, and that these officials can sometimes, by virtue of their positions in government, use state power and resources to pursue their own ideas and interests. In other words, government policies often reflect the interests and economic ideologies of state officials, rather than those of domestic or foreign economic elites or powerful organized middle- and working-class groups. These policies in turn create the incentives that shape the day-to-day choices of individuals in society. And these choices then affect the rate of growth, the distribution of the benefits of growth, and, in sum, the way political, economic, and social systems work.

The State in Economic Development

In the study of developing countries, much of the recent interest in the role of the state has arisen from observations that governments sometimes effect radical shifts in economic policy without the support of important interest groups. Observers see such "state" or government autonomy in the most successful new industrializing countries and also in what used to be considered successful post-revolutionary regimes.

The notion of state autonomy, which draws on theoretical constructs developed in structuralist Marxist and realist international relations theory, emerged in the study of developing countries as part of the effort

For example, on social revolution, Skocpol 1979; on nonrevolutionary socioeconomic transformation, Trimberger 1978; on economic development in Europe, Gerschenkron 1966, North 1979, Eckstein 1958; on development strategies in the Latin American and Asian newly industrializing countries (NICs), Haggard and Cheng 1987, Haggard 1990, Evans 1979, Rueschemeyer and Evans 1986; on squatter settlements, Collier 1976, Ames 1973; on labor, Goldberg 1986, Collier and Collier 1977, 1979, Katznelson 1986; on the failure of agricultural policy in Africa, Bates 1981.

to understand situations in which a conflict paradigmb that is, an assumption that political outcomes are determined by a clash of interests, with the most powerful interest winningb could not explain major policy shifts. In these situations, the interests injured by the new policies were often those that, shortly before a shift, had appeared to be economically and politically dominant in the country (Trimberger 1977, 1978; Hamilton 1982). In any unmediated struggle among interests, standard political theories, whether Marxist or pluralist, would have expected them to win. And yet they lost.

In Latin America, for example, most governments began to implement industrialization policies that systematically disadvantaged the producers of primary product exports at a time when agriculture and mining remained economically dominant.3 The more recent history of the region offers numerous additional instances of policy changes that have injured powerful economic groups. Domestic industries, foreign industries, and the organized working class have all failed on various occasions to protect their own interests, despite their considerable political, organizational, and economic resources. No one believes that these groups are weak or uninfluential, but those who propose a focus on the state point out that they have not proved insurmountable obstacles to governments bent on pursuing policies damaging to them.

The interests that benefited from these policy shifts, in contrast to those that lost, were often unorganized, economically weak, and sometimes even weak in terms of sheer numbers of people. If they subsequently developed into powerful interest groups, it was because of the policy shifts initially undertaken by state elites on their own initiative. Manufacturers of inputs for the auto industry in Brazil, for example, have become politically influential since the 1960s. Prior to the state-sponsored implantation of the auto industry and the passage of strict

Except for Argentina, the more economically advanced Latin American countries had begun to abandon laissez-faire policies and experiment with state intervention in the economy by the 1930s. Conscious, integrated policies to foster industrialization began in Argentina and Chile in the 1940s, in the other more developed countries during the 1950s, and in most of the poorer countries during the 1960s. Measuring sectoral dominance is extremely problematic, but contribution of the industrial sector to GDP gives a crude indication. Of the nineteen Latin American countries for which data are available, only in Uruguay had industrial production outstripped primary product production by the time policy changes disadvantageous to primary product producers began in the thirties. The transition to industrial dominance over agriculture occurred in Argentina and Brazil during the 1940s, and in Chile and Mexico during the 1950s. In the rest of Latin America, the transition had still not occurred by 1970. In other words, in all but Uruguay and possibly Argentina, the initiation of policies favoring the industrial sector occurred before that sector had attained economic power on its own.

domestic content laws, however, this group was small and insignificant both economically and politically. It owes its current status to the government decision in the 1950s to subsidize and otherwise encourage local production of capital goods and consumer durables. Prior to this decision, the producers of nondurable consumer goods dominated Brazil's private domestic manufacturing sector and the interest organizations that represented it. In spite of its economic importance, however, the traditional manufacturing sector reaped relatively few benefits from the state-sponsored industrialization drive during the 1950s.

Policy shifts that confer benefits on groups insufficiently powerful to have effectively demanded them are not common occurrences, but they do happen from time to time, and, when they do, they can have momentous consequences and are therefore worth trying to understand. Since these policy outcomes could not have been predicted on the basis of known preexisting forces, that is, classes or interest groups, observers posited the existence of some other force to account for them. Just as astronomers deduced the existence of Neptune from perturbations in the orbit of Uranus before they could see the new planet, social scientists inferred the existence of an autonomous force located in the state, even though they could not directly observe it. This new force seemed to depend on state resources and to be guided by the preferences of political leaders.

From positing the existence of an autonomous force made possible by state resources to speculating about the possible effects of such a force on policy was a short step. Several scholars have suggested that state autonomy would aid state intervention to foster rapid economic development or structural adjustment, since it would allow government policymakers to ignore interests injured by particular policies. Both development and adjustment inevitably involve shifting resources among sectors of the economy, geographical areas, and social groups. They also involve a temporary shift of resources from consumption to investment. All of these resource movements impose costs and generate opposition.

Since powerful societal interests, that is, those that benefit most from the status quo, would tend to resist radical changes in policy, several observers have credited more autonomous states with greater flexibility in adapting to crises and rapid changes in the international economy (Haggard 1986; Haggard and Kaufman 1989). They note that policies aimed at creating comparative advantage, taking advantage of new opportunities in the international economy, or adapting to international price shocks often involve channeling resources away from customary

uses and into the creation of new productive endeavors. They argue that governments can more readily impose the costs of these transitions if they enjoy some autonomy from the interests being hurt, that is, if their political survival does not depend on support from injured interests.

Autonomy, in this view, frees state elites from the interests vested in earlier stages of development. Thus, autonomy should be especially important when crises precipitate a reexamination of development strategies. Autonomy offers no guarantee that the economic situation will be assessed competently or that decisions in the public interest will be made, but if both do occur, it makes effective action more likely.

The same kind of argument about the utility of autonomy from societal interests has been made about Leninist regimes. Post-revolutionary industrialization drives under Leninist leadership involved an even more radical imposition of costs on large parts of the population. Only governments highly insulated from most societal groups could carry out such policies (Jowitt 1971), and the idea of state autonomy, if not the label, holds no novelty for observers of Leninist political systems.

For scholars interested in explaining why some countries have followed more effective development strategies than others, the notion of state autonomy has seemed a useful starting point, but efforts to develop it have run into snags. State autonomy has not been, and perhaps cannot be, directly observed. Its presence is inferred on the basis of policy outcomes that appear to reflect the preferences of officials, or even of some disembodied national interest (Krasner 1978).4 "State autonomy" thus explains a situation that appears anomalous if one assumes that government policies generally reflect societal interests.

The fact that state autonomy is inferred from its effects rather than directly observed has made it difficult to study and explain. No one, it seems, is quite sure what "it" actually consists of. State autonomy seems at times to refer to the independence of the state itself, the regime, a particular government, some segments or agencies of the government, or even specific leaders. It seems the phrase can refer to any independent force based in the central government.

Although several authors have discussed the subject (Stepan 1978; Hamilton 1982; Haggard 1986; Rueschemeyer and Evans 1986,

Eric Nordlinger (1981, 74-98) makes the point that a state could be autonomous even if the preferences of state actors and important societal groups were identical, as long as the preferences of state actors did not derive from societal preferences. Autonomy can only be clearly perceived, however, when policy outcomes diverge from expressed or assumed societal preferences.

63-65; Waisman 1987), no one has confirmed empirically the existence of political or societal characteristics that allow governments to act autonomously. In the absence of such a set of concrete traits by which to identify state autonomy without recourse to its supposed effects, it is impossible to show empirically that it exists and has effects. It also becomes very difficult to explain its historical development. As a result, analyses tend to focus more on the policy outcomes themselves rather than on the state structure that is hypothesized to produce the outcomes.

Nevertheless, by establishing that officials do at times act on the basis of their own ideologies and preferences, the state autonomy literature has performed a useful service. It has set the agenda for further work in the area. Efforts have begun to explain why states intervene in the economy and society, how and why they choose particular policies as the means of intervention, and why some state interventions result in better outcomes than others. The attempt to answer these questions seems stalled, however, at an early and frustrating stage. Insight and description abound, but explanation remains scarce.

Structural arguments, currently the most cogent strand of the state autonomy literature, describe the features of international and domestic situations that would make it rational for states (i.e., governments) to behave autonomously in order to pursue an interest in, for example, national survival or capitalist development (Trimberger 1977; Skocpol and Trimberger 1977-78, 101-13; Skocpol 1979; Cardoso 1969, 1971; Bamat 1977). Such explanations implicitly treat states as rational unitary actors without building any plausible foundation for doing so.

Analysts who attempt to delve into the concrete details of how states behave usually end up moving away from systematic structural explanations and toward ad hoc political and sociological explanations. If one wants to explain a state's preferences regarding development strategies, for example, one needs to know who has power and what they want and believe. Who has power will depend on the outcome of a political struggle that can be "explained" by particular features of the conjuncture and personalities involved. Such an explanation will have few implications for other cases, however.

Thus within the statist literature, one tends to have a choice between systematic structuralist arguments that lack plausible individual-level foundations and plausible individual-level explanations that lack theoretical reach. There is no doubt that the issues raised by the statist literature are important. But studies carried out to date leave doubts about whether the current statist approach is the best theoretical tool

for trying to understand the role of states in developing countries. For one thing, most arguments about state autonomy were developed in the context of studies of authoritarian regimes.5 Now that most developing countries have taken significant steps toward democratization, states still matter, but old arguments may not.

For another, discussions of the state usually assume that states behave as unitary actors. In reality, they often do not. The problem is not that there is "no there there," but that there are too many theres thereb each having different capacities, intentions, and preferences. Some parts of the state may express independent preferences while others, often the larger parts, reflect societal interests. Discussions of "the" state offer little theoretical leverage for understanding the sources of the competing interests of different actors within the state.

The State as a Collection of Self-Interested Individuals

Observations of outcomes that seemed to have been caused by autonomous state action were initially taken as challenges to traditional paradigms and theories, especially pluralism, modernization, Marxism, and dependency. But by now, the old paradigms are either enfeebled or revised to the point where further challenges seem pointless, and some analysts have come to feel a need for new theories within which autonomous government actions would cease to be anomalous.

This study proposes such an anchor for the statist perspective: an approach to the state based on a model of rationally self-interested political leaders.6 By focusing on individual state actors and the different incentives they confront, it suggests an alternative approach to the state that provides a systematic explanation for the different capacities and apparent preferences of different parts of the state. The keystone of the argument is a simple model of politicians and bureaucrats as rational individuals who attempt to maximize career success.7 The basic idea underlying this model is nicely expressed by Mark Crain and Robert

Nordlinger (1981) is an obvious exception.

Others, for example Douglass North (1979) and Margaret Levi (1988), have used rational actor models to interpret state actions. Their analyses focus on rulers as revenue maximizers. Though the approach taken here does not use this key assumption, it builds on ideas found in their work.

Though novel in the context of the "statist" perspective, the approach used here obviously builds on such works as Mayhew (1974), Fiorina (1977), Arnold (1979), Jacobson and Kernell (1983), Ames (1987), Crain and Tollison (1990), Bates (1989), and Cowhey (1990).

Tollison (1990, 3). "[H]uman behavior in governmental settings is motivated by the same self-interested forces that guide human behavior in private settings. . . . It should be stressed that this is not a dogmatic assertion that politicians and other political actors are self-interested; rather, it is a pragmatic analytical convention"

Since state decisions are made by these human beings, the content of their decisionsb including decisions that contribute to reforming the state itselfb will reflect their interests. Political leaders' interests center on their political careers.

Normally, personal success in politics . . . will imply concentration of the professional kind and relegate a man's other activities to the rank of sidelines or necessary chores. . . . [I]n modern democracies . . . politics will unavoidably be a career. This in turn spells recognition of a distinct professional interest in the individual politician. . . . Many a riddle is solved as soon as we take account of it. Among other things we immediately cease to wonder why it is that politicians so often fail to serve the interest of their class or of the groups with which they are personally connected.
(Schumpeter 1975, 285)

Whatever additional interests they may have, most political leaders want to remain in office. And, generally speaking, their other goals can only be accomplished while in office. Political institutions determine which strategies for staying in office are likely to work.

If one accepts this view of officials, the route to explaining state or government autonomy ceases to be problematic. One can explain much of the behavior of states as the outcome of rational choices by self-interested officials acting within particular institutional and circumstantial contexts. This approach leads to an explanation of state autonomy grounded in the behavior of individual political actors: officials and politicians will behave in ways that result in state autonomy when it serves their own career interests to do so. When their interests can be furthered by representing particular societal interests, they will do this. Using this approach, degrees or kinds of autonomy can be predicted from institutional features of the regime and other characteristics of the political system, rather than inferred on the basis of policy outcomes. This approach thus creates the possibility of falsifying arguments.

The focus on incentives also offers more leverage for explaining policy outcomes than does the undifferentiated notion of state autonomy. If autonomy is viewed as an attribute of a unitary state, no prediction about state behavior can be made unless state interests or preferences can be identified. Although interests are usefully imputed to states

in some schools of thought,8 attempts to deduce domestic behavior from assumed state interests have failed to produce satisfying results.

It has been argued, most notably by Theda Skocpol (1979), that international threats motivate state officials to act autonomously from dominant classes. The approach taken here, in contrast, emphasizes that state officials have great leeway in choosing how to respond to threats. It suggests that the kind of response state actors select will, whatever else it does, reflect their own domestic political interests. This fact was well understood by European colonizers. Britain's famous divide-and-rule strategy of colonization, for example, was effective precisely because elites in soon-to-become colonized areas chose strategies for responding to the British on the basis of how they would affect competition with domestic rivals. Attention to the domestic interests served by particular responses to outside threats thus helps explain why most Third World elites who faced the threat of colonial military incursion failed to carry out the internal administrative, military, and economic changes that might have given them the strength to meet the challenge. Most so far have also failed to initiate the reforms that would enable them to meet international economic threats more successfully.

Arguments that stress the imperatives created by foreign military and economic threats seem to forget that the world is littered with the corpses of potential and former nations that failed to meet the challenge of international competition (for example, Burgundy, Bavaria, Wales), not to mention the battered victims of economic threats that manage to cling to life despite the policy choices of their leaders. The state autonomy literature might lead one to hypothesize that these failures have been caused by a lack of policy autonomy from dominant groups intent on maintaining their own privileges, but evidence often fails to support this hypothesis (Geddes 1990). These failures have not in general been caused by an inability to ignore the interests of powerful societal groups.

Rather, many failures have been caused by the policies political leaders chose when attempting to respond to international threats and at the same time gain an advantage in the struggle with domestic political rivals. In some cases, such as the presidential efforts to increase bureaucratic capacity analyzed in chapter 6, political rivals (rather than economic elites) have realized the threat to their interests and have success-

In some Marxist theory, the (capitalist) state is assumed to have an interest in assuring the preservation of capitalism (Poulantzas 1975, 1976). In much of international relations theory, states acting in the international arena are assumed to maximize power or national wealth.

fully blocked reforms. In cases where policy choices have involved the creation of new political resources (such as state enterprises) rather than the redistribution of preexisting ones, political rivals have generally not opposed them. As in other political situations, it is easier to distribute from an expanding pie than to redistribute. The catastrophic problem with these policies is that they have frequently turned out to be much more effective as vehicles for creating political resources useful to incumbent politicians than as strategies for dealing with international economic or military threats.

Commodity marketing boards set up in a number of developing countries provide especially dramatic examples of policies aimed simultaneously at coping with international economic threats and capturing resources for incumbent political leaders. The marketing boards were designed to counteract the threat of foreign control over the marketing of export crops, while at the same time contributing to economic diversification as a means of reducing vulnerability to external price shocks. Monopolies over the marketing of export crops allowed governments to capture resources from the agricultural sector to use in building the industrial sector. The resources spent on manufacturing and infrastructure projects could easily be directed toward regions and projects that would benefit supporters of the current political leadership and thus help maintain the political machinery of incumbents.

The marketing boards succeeded in the sense that they permitted governments to siphon substantial amounts of capital out of the agricultural sector. They worked far less well, however, as tools for reducing economic vulnerability since they had the unintended but predictable consequence of reducing farmers' incentive to export and hence reducing export earnings (Bates 1981; Smith 1969). In many cases, governments established these boards in opposition to dominant agricultural interests and prior to the emergence of powerful industrial interests. This strategy failed to meet the international economic challenge not because of insufficient policy autonomy from interest groups but because the policy instrument chosen turned out to be economically disastrousb though politically quite useful.

Attention to the interests of individual state officials in competition with other politicians yields systematic explanations of the choice of commodity marketing boards as a strategy for dealing with international economic constraints, and for many other policies that would seem irrational if one assumed that state actors are motivated primarily by the need to respond to international threats. To conceive of politicians

and officials as motivated primarily by interest in domestic political survival and advancement also provides plausible individual-level arguments about how and under what circumstances officials can be expected to behave autonomously from societal interests.

How Institutions Shape Incentives

If one assumes that the interests of most government decision makers, and of aspirants to that role, center on their political careers, one can expect their policy preferences and their behavior to be shaped by the requirements for advancement imposed by their roles in government and their memberships in the parties, professions, and leadership cliques that structure access to government posts. At the regime level, if we classify governments according to the incentives favoring autonomy that they provide officials, they fall somewhere on a continuum anchored at one end by Leninist regimes and at the other, it could be argued, by oligarchical democracies in undiversified economies. In Leninist regimes, the costs and benefits to officials associated with different behaviors are largely controlled by higher level party officials, and, as a result, officials in these regimes exhibit a remarkable degree of autonomy from societal groups. In democracies in which only members of the economically dominant class may participate, politicians will tend to represent dominant interests since reelection depends only on this narrow group. They have little incentive to behave autonomously and, in fact, face great political dangers if they try to do so. The laissez-faire oriented, limited participation democratic governments of Latin America during the late nineteenth and early twentieth centuries fit this pattern. Among Latin American governments, only Uruguay, the sole Latin American country to enjoy both democracy and universal male suffrage before 1930,9 and revolutionary Mexico exhibited any noticeable autonomy before the 1930s.10

The SC!enz PeC1a Law of 1912 established universal male citizen suffrage in Argentina, but the exclusion of non-citizens denied the vote to the majority of the urban working class. As a result of this provision and the systematic electoral fraud practiced during the thirties and early forties, the class base of political participation in Argentina remained limited until 1946. See Waisman (1987).

This discussion is descriptively consistent, up to a point, with those authors who have noted the same change in state behavior and who consequently believe state autonomy to be possible when the mode of production is in transition (Hamilton 1982, ch. 1; Cardoso 1969). In my view, however, their interpretation is an unwarranted generalization from observation of the contrast between state behavior before about 1930, when most Latin American countries' economies depended on primary production of a few crops and minerals and political participation was very limited, and state behavior more re-

cently. They see recent instances in which state policies fail to reflect dominant class interests as evidence of division within the dominant class or lack of hegemony, that can presumably be expected to abate once industrial capitalism is consolidated. In contrast, I see the pre-1930 period as a never-to-be-repeated historical phase during which politicians motivated by electoral concerns would have had no reason to respond to any interests other than those of the narrow group able to vote.

As economies developed, however, and the number of urban, organizable potential voters increased, politicians not supported by established interests had an incentive to extend the franchise to excluded groups as a way of enhancing their own competitiveness against more entrenched politicians. Once the electoral system includes voters with more diverse interests, politicians have considerable leeway when it comes to putting together packages of policies to appeal to different aggregations of voters, but their most effective strategy is rarely a simple defense of the interests of the wealthy few. Consequently, it has become common to see elected politicians initiating policies costly to members of the dominant class.

In other words, modern democratic regimes provide self-interested political leaders with reasons for autonomy from dominant groups. These reasons vary, however, over time, across actors, and across countries, as do reasons for autonomy from other groups that have figured less prominently in the literature. Some of this variation can be explained by institutional differences among democracies. Within modern democratic regimes, particular institutions, such as electoral rules and party procedures, determine what kinds of behavior are most likely to contribute to career advancement. By analyzing the institutionally determined incentives facing political leaders, one can explain state actions that might be unexpected from the perspective of more traditional approaches.

This study deals principally with the motivations and incentives of four types of political actor: presidents,11 legislators, party leaders, and state bureaucrats. Members of each category are assumed to be continually conscious of career interests as they make various policy decisions, but, since actors occupying different institutional roles face different constraints and opportunities, their responses differ.

The career interests of various political actors are discussed in greater detail below, but a sketch of some key assumptions may be useful here. The careers of legislators depend on reelection, so concern with the vote

As this is written, all South and Central American countries have presidential systems, though transitions to parliamentarism are being debated in a few.

in the next election always ranks high in their calculations. Presidents in Latin America, who usually cannot serve consecutive terms and often face a real threat of overthrow, want, first, to complete their terms in office (Ames 1987). Next, they want to maximize their future political power within their parties and their chances of reelection after the constitutionally prescribed time out of office has elapsed. Party leaders pursue career success by maximizing party power and their own power within parties. Bureaucrats prefer, first, security in office, and then, advancement and the material advantages that accompany it. The way these interests impinge on each other and are constrained by other societal forces in specific instances is discussed below.

The time horizons of these actors vary depending on the country and historical period involved, but, on average, Latin American political actors must discount the future fairly heavily because of the high probability that a military intervention will marginalize them. The military ousted 38 percent of all Latin American governments between 1945 and 1982.12 Furthermore, even when the military remains on the sidelines, a change of the party in power may reduce markedly a politician's or bureaucrat's chances of getting ahead, since success depends heavily on patronage, which presidents monopolize. Consequently, discount rates tend to be high. In other words, the unstable political environment forces rational politicians in office to concentrate on activities that lead to quick results and immediate rewards.

The approach to thinking about the state proposed here focuses on how institutions, especially political institutions, shape the incentives of individuals in government, and how, in consequence, these individuals choose policies. It thus bridges the gap between structuralist arguments and intentional, or rationalist, ones. It moves away from considering the state itself as an actor, as having preferences or interests. Instead, it focuses on individuals whose behavior is shaped by their desire to succeed in political roles and whose capacity to influence other individuals depends on control of state resources. This focus on the interests of political leaders seems completely compatible with observations of autonomous state or government actions. Moreover, I would argue, it explains the process through which states sometimes behave autonomously.13

Calculated from information in Ames (1987, 12). This figure would be even higher if Costa Rica, which has no army, and Mexico, which has experienced no military intervention since 1945, had been excluded.

Consequently, I see no point in lumping the statist perspective with other chal-

lenges to the rationalist approach under the rubric of "the new institutionalism" (March and Olsen 1984).

State Capacity to Initiate Change

Up to this point, discussion has focused on state actors as decision makers, but the implementation of decisions is equally important. Although extensively covered in descriptive and prescriptive literature on economic development, state capacity to translate preferences into actions has received surprisingly little attention in the explicitly statist literature, perhaps because of its roots in Marxist and international relations theory, both of which tend to draw on European experience. To observers of less developed countries, however, it is clear that the articulation of independent state preferences, difficult as that may be to achieve, cannot by itself accomplish anything. Preferences matter very little if officials cannot carry out the policies they choose. The capacity to implement state-initiated policies depends on the ability to tax, coerce, shape the incentives facing private actors, and make effective bureaucratic decisions during the course of implementation. All of these abilities depend in turn on the existence of effective bureaucratic organizations. If one wants to understand states as actors, one needs to look at their bureaucratic innards.

Although several authors have noted the importance of competent bureaucratic organizations (Stepan 1978; Haggard and Kaufman 1989; Haggard 1990), little attention has been paid to the politics of achieving such competence. Bureaucratic competence seems, at least at first glance, to be a universally desirable goal. Yet such reforms have been notoriously difficult to carry out, and effective bureaucratic organizations remain rare in the developing world. Work in the statist tradition offers no suggestion for why a trait so obviously useful when governments intervene extensively in the economy has so often failed to develop. The approach taken in this study, in contrast, leads to the identification of the political conditions that cause bureaucratic capacity to develop and flourish, on the one hand, or stagnate, on the other. One of the central arguments of this study is that the development of bureaucratic competence depends on whether it serves the immediate career interests of the politicians who initiate reforms and choose appointment strategies.

The Struggle Over Reforms to Increase State Capacity

Historically, perception of the need for state capacity developed in tandem with perception of the need for state intervention in the economy. Between the end of World War II and the middle 1980s, governments advocating state-sponsored social and economic change achieved power at one time or another in most developing countries.14 These governments came to power committed to moving their societies in new directions, armed with plans and programs for doing so, and supported by substantial segments of society. Many of these programsb for example, road building, irrigation projects, the creation of development banksb had nearly universal support. And, although opposition from threatened interests blocked some policies, in most cases the opposition lacked the power to prevent the proposed programs from being initiated. Moreover, although the military eventually ousted some of these governments, few faced serious hostility to their policies from the military at the beginning of their terms. In most instances, militaries supported extensive state intervention to foster development. Thus, the lackluster performance of these change-oriented governments cannot be attributed solely to implacable hostility from dominant groups or the military. Nevertheless, many of the sweeping programs proposed by these leaders never bore the expected fruit.

If opposition from powerful societal and military interests did not prevent the full implementation of most programs, what did? Observation suggests the following reasons: lack of experience, competence, and expertise among personnel charged with carrying out projects; the diversion of needed funds to other uses; and the subversion of reform goals by politicians and bureaucrats uncommittedb for whatever reasonb to their achievement.15 Taken together, these impediments add up to a lack of state capacity to accomplish the tasks involved in state-led change.

Several excellent studies have described the development of state

In the middle 1980s, this long secular trend toward more state intervention reversed, and governments in many parts of the developing world, under pressure from changes in the international economy and from international agencies, began advocating less state involvement.

No argument is made here that all or even most of these programs and policies were well-conceived by current economic standards or that they would have contributed to long-run economic performance. Rather, the claim is that implementation was often so flawed in the short run that they showed few results at all.

capacity, conceived as expertise in government and the emergence of consensus among political leaders regarding needed changes (Hirsch-man 1973; Moran 1974; Tugwell 1975; Fishlow 1980). In these studies, progress in dealing with public dilemmas takes roughly the following trajectory: identification of a problem (ideological and technical elements both contribute to identification); early failures because of inadequate understanding, lack of political consensus, and insufficient technical capacity; and renewed efforts, which are at least incrementally more successful than the initial effort because knowledge has increased, technical capacity to deal with the problem has improved, and political capacity, in the sense of the formation of a broader coalition favoring reform, has developed. Despite their considerable usefulness, however, these studies do not provide general explanations of when reforms aimed at enhancing state capacity will occur. In addition, although they realize that state capacity to implement decisions is crucial, analysts in this tradition have neglected elements of the argument that would explain bureaucratic or administrative capacity as distinct from technical expertise.16 The focus on expertise creates a deceptive sense that lack of knowledge or know-how is the main impediment to effective state-led change.

These studies leave the impression that increasing competence and persuading political leaders that the national interest does indeed lie in a particular direction will be sufficient to solve problems. And yet many problems with apparently known solutions remain unsolved. Looking at issues on which less progress has been made, or even at the same cases examined by these authors over longer periods of time, brings into focus the political interests that oppose and impede changeb including changes that presidents, legislators, and popular majorities all claim to want.

Much of the literature has treated the development of expertise and the creation of more effective government instruments for managing the economy as primarily technical problemsb a matter of setting up schools to teach skills and figuring out which devices work. But an examination of the history of such innovations shows that political constraints, rather than lack of resources or knowledge, have often hindered the development of these apparently neutral capacities. These political constraints,

A partial exception, Stepan (1978, 238-39) lists a set of state capacities as variables that will help explain outcomes in struggles between multinationals and host country governments, but he makes little attempt to explain them. The same can be said of Cleaves and Pease GarcC-a (1983).

as argued throughout this book, are intelligible in light of the political costs and benefits of potential innovations to the politicians who must initiate them.

Those who may impede widely supported reforms include many whose motives are not at first obvious. A look at the nationalization of Chilean copper mines illustrates this point. By the time the mines were nationalized, support for such a move had become overwhelming in Chile. Not a single vote was cast against it in the Chilean legislature (Moran 1974, 147). Most Chileans saw nationalization as an assertion of national sovereignty and also expected it to contribute to Chilean economic success. Theodore Moran (1974) traces the development of the expertise that gave Chileans the capacity to bargain effectively with the transnationals and to run the mines after nationalization. As he demonstrates, political consensus on taking over the mines existed. So did the technical capacity to run them. Nevertheless, in spite of available expertise, the Chilean government led by Salvador Allende mismanaged the mines after nationalization (Valenzuela 1978, 55).17

For Allende, the need to consolidate political support in the short run overwhelmed the need for adequate economic performance. In consequence, he allocated managerships in the newly nationalized mines and industries on the basis of a quota system in which the parties that made up his coalition were rewarded in proportion to the number of votes they had received in the most recent election (Valenzuela 1978, 64-65; Castillo 1974). No mechanism existed to ensure that these managers were competent or to prevent Allende from turning posts vital to the Chilean economy into political plums. The performance of the copper mines was critical to the success of Allende's economic program and thus to the viability of his government. And yet the benefits of exchanging managerships for political support exceeded the costs in the very short run. Allende and his coalition partners themselves thus contributed to Chile's disastrous decline in economic performance.

The choice of managers for Chile's nationalized mines and industries is a particularly consequential instance of what I call the politician's dilemma. One can analyze it from the point of view of either the governing coalition or the president. To members of the coalition, the choice of managers was a collective action problem. No party in the coalition

No claim is made here that bad management was the only cause of the fall in productivity in the copper mines during Allende's tenure. U.S. refusal to supply spare parts and the atmosphere of revolutionary euphoria that undercut work discipline were at least as important.

wanted to sacrifice the political advantage of having its own managers (and attendant appointment powers), but if all managers were to be chosen for political reasons, the firms would be mismanaged. Mismanagement would increase the probability that the whole coalition would fall, and all members of the coalition would risk losing everything.

Even though all members of the coalition would have been better off if they had all given up some of their political advantage in order to achieve their larger collective goal of remaining in power, in the volatile atmosphere of the early 1970s in Chile, they were unable to do so. Coalition members could not arrive at enforceable agreements that each could trust to prevent others from taking advantage of current political sacrifices in the next election.

From the president's point of view, the choice lay between using the appointment resource as a political or an economic investment. The state would pay the same salary, but the individual who received it, depending on his or her competence and loyalties, might make a greater contribution toward economic performance or toward consolidating political support for the president.18 The president's survival depended on both economic performance and political support. No one had more to lose than Allende himself from the additional economic strain that would be caused by mismanagement, and yet, with his survival in office immediately threatened, he could not afford to disregard short-term political considerations by choosing managers on the basis of technocratic expertise. The president as well as his coalition partners faced a wrenching conflict between their own need for immediate political survival and longer-run collective interests in economic performance and regime stability. This is the politician's dilemma writ large. And this dilemma, as is shown below, creates many of the obstacles to improving state capacity.

Rather than stressing this political dilemma, most attempts to explain why one government is more capable than another emphasize, as noted above, technical and resource requirements: the availability of funds, which depends on the domestic resource base, tax laws and procedures for enforcing them, and access to foreign capital; the existence of instruments and institutions, such as national development banks, through which the government can manipulate the incentives facing key private

This is not a mutually exclusive choice. Some party loyalists also have expert credentials. Parties differ, however, in the extent to which they attract supporters with technical and managerial expertise and experience. The parties in the coalition that supported Allende could call on relatively few such supporters.

actors; consensus among leaders regarding goals; and the skill and expertise of officials.

Although these characteristics may be necessary to state capacity, they are not in themselves sufficient. The advice of experts may be ignored, and, as the Chilean case shows, skilled personnel may come up against insurmountable political obstacles to doing the jobs for which they have been trained. In spite of laws and enforcement procedures, taxes may not be collected or, if collected, may be dissipated in patronage or find their way into Swiss banks rather than being spent to pursue national goals. Development banks may award subsidized loans to political supporters rather than to entrepreneurs willing to invest in targeted sectors of the economy. Political leaders may genuinely support some shared set of national goals but still find that the exigencies of political survival cause them to behave in ways that undermine these goals.

To understand these common situations, one needs to examine the incentives facing the individuals who make decisions about whose advice to take, what projects to fund, to whom to give loans, and so on. Where decision makers can best assure their own personal security and advancement by using funds efficiently to implement programs or by hiring experts and following their advice, they will do so. Where they can best further their own interests by exchanging the resources under their control for political support or by simply appropriating them, they will do this. The incentives that face officials depend on the norms, available career rewards, and methods of enforcing the behavior deemed appropriate that have evolved in the offices, professions, institutions, and parties to which they belong. Later chapters consider how such norms, rewards, and methods of enforcementb in other words, incentivesb might be changed in ways that would result in greater state capacity.

Plan of the Book

Most experts and many political leaders see administrative reform as a key method of increasing state capacity. Chapter 2 reinterprets the struggle over administrative reform as a collective action problem. It discusses solutions to collective action problems that various analysts have proposed and finds that popular movements supporting bureaucratic reform are unlikely to arise spontaneously, even though evidence suggests that most citizens in Latin American countries believe they would benefit

from such reforms. It then considers the role of political entrepreneurs in providing collective goods19 such as administrative reform. It concludes that principal-agent problems20 prevent the provision of public goods from becoming routinized even in democracies, and that bureaucratic reforms are especially unlikely to be provided because of their cost to politicians themselves. This discussion of potential political entrepreneurs sets the stage for later chapters that focus on the particular circumstances that lead democratic politicians to provide bureaucratic reform as a public good.

Chapter 3 is a case study of the struggle to improve bureaucratic competence and honesty in Brazil between 1930 and 1964. The purpose of this case study is to give the reader a detailed and concrete case with which to compare theoretical arguments. Levels of competence, efficiency, and honesty have varied widely within the Brazilian bureaucracy, but some sectors at some times have been as capable as virtually any in the world. The relatively high level of Brazilian bureaucratic performance, especially in some of the agencies that deal with the economy, developed during a decades-long struggle between those seeking reforms and those whose short-run interests would be hurt by them. During some regimes and administrations, bureaucratic reforms made rapid progress; during others, previous reforms were either rolled back or informally disregarded.

The case study of Brazil shows the persistence of political interests opposed to bureaucratic reform and their links with the democratic political system. It then discusses the creation of an insulated segment of the bureaucracy as a functional substitute for politically infeasible universalistic reforms. This insulated segment of the bureaucracy was responsible for the policies leading to Brazil's most impressive economic gains prior to 1964. The analysis follows the vicissitudes of those parts of the bureaucracy charged with economic development tasks and demonstrates the relationship between insulation from day-to-day partisan political concerns and performance.

Theoretically, this chapter makes a contribution toward understanding and redefining state autonomy. It suggests that the type of autonomy

See chapter 2, pp. 27-28, for a detailed discussion of the meaning of collective or public goods.

"Principal-agent" or "agency" problems occur when individuals delegated to carry out some task for someone else who pays their salary, or elects them, have interests at variance with those of their employer (or the voters who elected them) that lead them at times to neglect or distort work on the task. See chapter 2, pp. 39-41, and chapter 3, pp. 47-49, for more detailed discussions of the meaning and implications of these terms.

state policymakers need in order to implement policies effectively may not be autonomy from class-based pressures, but rather autonomy from the tide of particularistic demands that has risen as an unintended consequence of certain kinds of representative institutions.

Chapter 4 develops a game theoretic model to explain the failure of the Brazilian legislature to initiate reforms. This model demonstrates that when patronage is distributed unevenly among the larger parties in the legislatureb as it is most of the time in Latin American countriesb majority party legislators will vote for reforms only under extraordinary circumstances. Consequently, reforms very rarely occur.

If patronage were distributed evenly among the larger parties, however, reforms should occur much more easily. Brazil never experienced equality among the larger parties, and no universal reforms passed the legislature. Thus the model seems consistent with Brazilian evidence, but no general test of the model is possible using only evidence from Brazil.

Chapter 5 turns to the issue of testing the model more generally. It tests the game theoretic model on the South American countries that have experienced substantial periods of democracy since 1930. Case studies of the initiation of reforms in Latin American democracies show that the conclusions of chapter 4 hold in the real world: reforms occur when the larger parties in a political system are approximately equal and thus control approximately equal amounts of patronage. Chapter 5 then goes on to draw several additional conclusions about the effects of different kinds of electoral rulesb which shape party systemsb on the likelihood of reforms.

Reforms may also be initiated by presidents, who have several reasons for wanting to increase the competence of their administrative machinery. In considering whether to initiate reforms, however, presidents must consider the preferences of other political actors as well as their own; they must behave strategically. Chapter 6 extends the analysis to explain why presidents pursue different appointment strategies. On the basis of an examination of all the larger South American countries, it argues that presidents have reasons to favor appointment strategies that contribute to the growth of state capacity, but are often prevented from following the strategy they would prefer because of more immediate political needs. When military intervention threatens presidents' immediate survival in officeb as it did, for example, Salvador Allende'sb they must spend all their patronage resources on survival and cannot afford a longer-term investment in building competence.

When presidents are more secure from military intervention, their strategy for improving bureaucratic capacity will depend on their relationship with other members of their parties or coalitions. Although presidents do not face immediate reelection and might, in consequence, refrain from rewarding supporters (whom they will not need in the immediate future), other members of their parties do need supporters. They can be expected to demand that the president distribute the usual spoils to the party faithful as the price of their support for his or her programs in the legislature and bureaucracy.

Presidents who face little competition from rivals within their parties and who can count on disciplined behavior from legislative members of their parties and coalition allies can generally afford to use some of their appointment resources to hire experts and build competent agencies. Presidents who lack the weapon of party discipline, however, and who have had to make deals with and promises to rivals within their parties in order to secure nomination, may have to exchange most of their appointment resources for the support of political allies that they need in order to govern effectively. In brief, strong and secure presidents are more likely to pursue appointment strategies that contribute to the development of bureaucratic competence.

Chapter 7 then turns to a consideration of some of the implications of the arguments made in earlier chapters. It considers, first of all, which particular democratic institutions are most likely to structure the incentives of legislators and presidents in ways that lead them to contribute to building state capacity.

It then turns to one of the murkier areas of research in the study of developing countries: the relationship between authoritarianism and economic performance. It suggests that authoritarianism tempts elites committed above all else to economic development in part because of the disincentives embedded in electoral politics to carrying out capacity-enhancing reforms. Once in power, however, authoritarian leaders face their own problem of how to create incentives conducive to rapid development which, though different from that experienced in democratic settings, is equally intractable. Consequently, authoritarian governments are as likely as democratic to fail in their efforts to foster rapid growth (Remmer 1990 and 1993; Geddes 1993).

Authoritarian governments do often initiate measures that concentrate technocrats in key administrative agencies and insulate them from outside pressures. Their attempts to build state capacity, however, often

fail to survive the transitions to democracy that follow authoritarian rule.

Frequent alternations between authoritarianism and democracy seem to provide the worst of all possible contexts for building state capacity. Competence tends to suffer during transitions both to and from military rule. Islands of competence nurtured for decades can be destroyed during a few months by presidents engaged in chaotic struggles to defend themselves against overthrow. At the other end of the cycle, redemocratization brings in its wake hungry parties eager for the spoils they have been denied during the years of military rule.

Continues...

Excerpted from Politician's Dilemmaby Barbara Geddes Copyright © 1996 by Barbara Geddes. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

Edition details

• Publisher: University of California Press
• Publication date: 1996
• Language: English
• ISBN 10: 0520207629
• ISBN 13: 9780520207622
• Binding: Paperback
• Number of pages: 262