Download Political Institutions and Development in Latin America: An Empirical Analysis and more Exams Comparative Law and Politics in PDF only on Docsity!
Political Institutions,
Accountability and
Government Performance
Chapter
Latin America’s troubled history of economic devel- opment has often been blamed on its political lead- ers. Most accounts gravitate between two seemingly contradictory caricatures: elitism and populism. Ac- cording to the first, Latin America has been ruled by a wealthy elite that has ruthlessly used the power of the state to enrich itself and maintain its privileges. According to the second, Latin America has been ruled by reckless populists who have sacrificed prosperity in pursuit of redistributive fantasies. This chapter examines the connection be- tween politics and development, but with an empha- sis on the impact of political institutions instead of political personalities. In particular, it focuses on the inherent difficulties of the political process in demo- cratic societies, and on how these difficulties often prevent democracies from increasing economic growth and advancing social justice. This approach leaves out some key aspects of politics. It does not address the role of ideas, ideolo- gies or political leaders, some of whom have the power to lead their countries toward prosperity or ruin. However, it does consider the role of culture and other exogenous factors, including political participation and ethnic and geographic divisions. In other words, the emphasis is not only on political institutions but also on political practices and behaviors, all in the context of democratic societies. Most Latin American democracies find them- selves today at a crucial crossroad. The initial enthu- siasm that accompanied the wave of democratization that swept Latin America over a decade ago has be- gun to erode. It has been replaced in many cases by dissatisfaction and cynicism. Further, there is a grow- ing consensus that far-reaching institutional reforms
are needed to enhance economic efficiency and social equity. But unlike many of the previous reforms, which involved mostly technical matters, these reforms can- not be conceived in a political vacuum. Bluntly put, any attempt to advance what have been called “sec- ond-generation reforms” is doomed to fail if it does not take politics into account. Given that these re- forms stand on the horizon, politics and political in- stitutions are bound to take on preeminent impor- tance in the years to come. This chapter first describes the evolution of democracy in Latin America and presents some evi- dence on levels of satisfaction with democracy and democratic institutions in the region. A simple ana- lytical framework is then put forth that seeks to un- derstand why democracies do not always function properly. That framework is used to evaluate the sta- tus of political institutions in Latin America and to empirically explore a few connections between poli- tics and development. Finally, the evidence from this analysis is used to explore some of the most pressing policy issues in the political realm.
The Latin American Democratic Wave
The last quarter of the final century in the millen- nium brought a new wave of democratization across the world that was unprecedented in its magnitude, geographical breadth and durability.^1 Figure 4.1 shows that the percentage of countries classified as “free” according to Freedom House went from 20 in 1978
(^1) Huntington (1991).
164 Chapter 4
to 38 percent in 1998.^2 Unlike the previous demo- cratic wave brought about by the Allied occupation of the vanquished Axis powers after World War II, this latest surge of democracy has been widespread, citizen driven, and spontaneous. Latin America has been on the crest of this wave of democracy. Governments led or supervised by the armed forces relinquished control throughout the late 1970s and 1980s to civilian rule, and today almost all countries in the region have elected and constitutionally-bound governments. There is better protection of civil liberties, and increased decentrali- zation of governmental authority has enhanced the decisionmaking power of citizens regarding salient local issues. Finally, many of what were once only nominally democratic institutions have become more democratic in practice. Political parties in several countries, for example, have opened the process of selecting candidates for the presidency and other po- litical offices. Yet, democracy in the region is still in the pro- cess of being consolidated. It may have given people more freedom to criticize unresponsive politicians and bureaucracies, inefficient spending, and other politi- cal maladies. But it has not necessarily solved these problems. Opinion surveys consistently show that citi- zens are not entirely satisfied with the performance of their governments. Apathy toward politics and a seeming acceptance of political leaders who arbitrarily bend the rules has led some observers to fear that Latin
America will move to a kind of “delegative democ- racy,” where citizens elect leaders but then summarily relinquish all policy control to them.^3
Level of Satisfaction with Democracy in Latin America
The expansion of democratic freedoms and the rou- tine use of free and competitive elections to choose political leaders have brought considerable advances in terms of protecting basic human rights and mak- ing governments more accountable to citizens. Still, the fluctuations in the Freedom House ratings for Latin America presented above show that establish- ing a stable and comprehensive democratic political system takes time, and is not an inevitable result of opening up public offices to electoral competition. Indeed, if democratic regimes do not maintain the broad support of citizens, they will eventually become vulnerable to individuals or groups that prefer to cir- cumvent democratic procedures and limit citizen rights.
Following the Democratic Wave: Percentages of "Free" Countries
Figure 4.
10
20
30
40
50
60
70
1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 World Latin America Source: Freedom House (1999).
(^2) The evolution of democracy around the world has been consistently monitored by Freedom House, a broad-based nonpartisan organiza- tion led by eminent politicians, business leaders and scholars (see Box 4.1). (^3) O’Donnell (1994).
Citizens Very or Partly Satisfied with Democracy (In percent)
Figure 4.
0 10 20 30 40 50 60 70
Paraguay Peru Brazil Mexico ColombiaBolivia
Chile Ecuador Venezuela
HondurasPanama El Salvador Nicaragua Argentina Costa RicaUruguay
Commonwealth Independent States Latin America Central and Eastern Europe European Union (15 countries)
Source: Latinobarómetro, various years, Eurobarometer (1999), Central and Eastern Eurobarometer (1998).
166 Chapter 4
(^4) The main features of Latinobarómetro are described in the Techni- cal Appendix. The survey includes 17 countries in South and Central America and has been conducted regularly since 1995. Throughout the chapter, we use the average of the three most recent rounds of the sur- vey carried out in 1996, 1997 and 1998. Similar public opinion surveys are regularly administered in the European Community (Eurobarometer) and in Central and Eastern Europe (Central and Eastern Eurobarometer). Since the degree of overlap in the questionnaires cor- responding to the different surveys is considerable, interregional com- parisons are possible for several of the topics of interest in this chapter.
democratic principles.^4 To the extent that it is pos- sible, we use other surveys to compare Latin America with other regions of the world. Those surveyed by Latinobarómetro were regularly asked the following question: “On the whole, are you (1) very satisfied, (2) fairly satisfied, (3) not very satisfied, or (4) not at all satisfied with the way democ- racy works in your country?” Figure 4.2 shows the per-
As might be expected, a larger share of Latin American citi- zens are dissatisfied with democracy than one finds in the high-income democracies. A key to assessing the prospects for firmly establishing and deepening democracy in Latin America is to identify the types of people who are more likely to be dissatisfied with democracy. As pointed out in the text, the survey question deal- ing with the level of satisfaction with democracy is some- what vague. “Dissatisfaction with democracy” can indicate unhappiness with the particular government in power, with the state of the economy, with the degree of income inequal- ity, with other aspects of democratic government perfor- mance (such as the ability to control crime or corruption), or with the quality of democratic institutions and procedures. Or it can indicate a more general dislike of democracy as a system of government. For any given person, a response of dissatisfaction could reflect a combination of any number of these sentiments. Consequently, identifying dissatisfied citizens must be approached from two different perspectives. First, taking the response to the question at face value, it is necessary to examine whether people’s socioeconomic or demographic attributes are related to their likelihood of expressing dis- satisfaction with democracy. Second, it must be determined which of the sentiments mentioned in the previous paragraph are more strongly associated with those feelings of dissatis- faction. The level of dissatisfaction with democracy does not vary substantially according to the age, sex, wealth or edu- cational attainment of the respondent. Women and men appear to have the same disposition toward the state of de- mocracy. There is a tendency for older individuals to be more satisfied with democracy, but the effect is small: every 20 year interval increases the probability of an individual being sat- isfied by just 1 percent (Figure 1). And while the number of years of education appears to make no difference, higher economic status produces a slightly greater chance that an individual is content with democracy. On the whole, how- ever, dissatisfaction with democracy would appear to be a
product of social and political attitudes not closely connected with socioeconomic or demographic attributes. With respect to the second perspective, it is clear that dissatisfaction with democracy can involve any one of the sentiments mentioned above. The data show that people who express confidence in democratic institutions and be- lieve that electoral procedures are fair are considerably more likely to be satisfied with democracy. Positive perceptions of current government performance regarding socioeconomic problems are also associated with higher levels of satisfac- tion with democracy. Among all the attitudes considered, those regarding the degree of income inequality stand out for their association with a person being content with de- mocracy. Those who perceive current income distribution as unjust are, overall, about 20 percent less likely to express satisfaction with democracy. And, as expected, people who express only lukewarm support for democracy as a system of government are less likely to be satisfied with the way de- mocracy is working.
Which Citizens Are Dissatisfied with Democracy in Latin America?
Figure 1. Percent Satisfied With Democracy, by Age Group (In percent)
Age group Source: Latinobarómetro, various years.
20
25
30
35
40
45
50
18-25 26-35 36-45 46-55 56-65 66-
Box 4.
Political Institutions, Accountability and Government Performance (^167)
centage of respondents in Latin American countries and three European regions who report being either “very satisfied” or “fairly satisfied” with the way de- mocracy works. Satisfaction with democracy varies widely across countries. While over 60 percent of the respondents in Uruguay and Costa Rica are satisfied with the way democracy is working, less than 20 per- cent of the respondents in Paraguay have a similar opin- ion. Large cross-country variation is also pervasive in other regions. In the European Community, for ex- ample, the range of variation is even broader, from 84 percent in Denmark to 28 percent in Italy. There are also differences between regions, but they are less pronounced. While 47 percent of the respondents reported being satisfied with democ- racy in the European Community, only 35 percent did so in Latin America. Further, only in Uruguay and Costa Rica is satisfaction with democracy above the average level in the European Community. Latin American citizens, however, are almost as satisfied with democracy as are citizens in Central and Eastern Eu- rope, and much more satisfied than citizens in the former Soviet republics. As a measure of broad support for democracy these figures must be viewed with caution. “Low sat- isfaction” with democracy does not necessarily imply weak support for democratic principles (see Box 4.2). Indeed, low satisfaction could mean that citizens be- lieve that democratic rights are still substantially re- stricted or democratic institutions are underdevel- oped, or it could mean that citizens are not happy
with the performance of democratic government even though they support the regime itself. The evidence shown in Figure 4.3 is roughly consistent with the former interpretation. Latin American countries with the highest ratings of political rights and civil liber- ties also have the highest levels of public satisfaction with democracy, while countries with the lowest rat- ings have the lowest satisfaction levels. Survey respondents were also directly que- ried about their support for democratic ideals. Spe- cifically, they were asked the following question: “With which of the following statements do you agree the most: (1) democracy is preferable to whatever other form of government; (2) in some circumstances an authoritarian government can be preferable to a democratic government; (3) for someone like me, a democratic or a nondemocratic regime makes no dif- ference.” Figure 4.4 shows that, as before, there is considerable variation across countries. People are most supportive of democratic ideals in Uruguay, Costa Rica and Argentina and less supportive in Para- guay, Brazil and Mexico. The range in the propor- tion of respondents preferring democracy to any alternative is substantial—from a maximum of 83 per- cent in Uruguay to a minimum of about 51 percent in Brazil. On the other hand, the percentage of those who believe that authoritarianism is sometimes pref- erable does not vary much across countries, and in most cases is below 20 percent. When compared with similar surveys taken in Europe, Figure 4.4 also shows that there is a large
Democratic Rating and Satisfaction with Democracy
Figure 4.
10
20
30
40
50
60
70
Percent satisfied with democracy
3.5 4.0 4.5 5.0 5.5 6.0 6. Freedom House rating
Argentina
Bolivia Brazil
Chile Colombia
Costa Rica
Ecuador
Honduras
Mexico
NicaraguaPanama
Paraguay
Peru
Uruguay
Venezuela
Source: Latinobarómetro, various years, and Freedom House (1999).
El Salvador
Support for Democracy as a System of Government (In percent)
Figure 4.
40 20 0 20 40 60 80 100 Uruguay Costa Rica Argentina Panama Nicaragua
Bolivia El Salvador Peru Venezuela Colombia Chile
Honduras Guatemala Mexico Paraguay Ecuador Brazil
Latin America Europe
Source: Latinobarómetro, various years.
Autocracy Democracy
Political Institutions, Accountability and Government Performance (^169)
Problems of bias in representation occur if a political minority is regularly able to sway political outcomes in its favor. There are two main elements here. On the one hand, political minorities demand special prerogatives either by forming pressure groups or by utilizing their special command of resources (money, knowledge or status) on a more individual- ized basis. On the other, politicians supply special pre- rogatives in the form of targeted services and favor- able bills. An important consideration is the extent to which electoral systems give politicians more or less incentives to respond to demands for special treat- ment. While some systems encourage politicians to be overly responsive to specific interests, others en- courage them to heed broad national interests. Problems of agency occur if politicians pur- sue their own goals instead of those delegated to them by their constituencies. Delegation is difficult because citizens have limited means to ensure that politicians abide by their promises. Enforceable agreements, for example, are impossible in the political realm, and elections, which can be used by citizens to oust those who have not fulfilled their promises, occur only ev- ery few years. Obviously, some institutional arrange- ments are better than others for coping with agency problems. An informed and involved electorate, de- veloped political parties, and the presence of checks and balances will reduce the extent to which politi- cians can ignore their constituents. Problems of aggregation occur if political rep- resentatives, once elected, are unable to reconcile the diverse interests they claim to represent. Here we em- phasize the institutional elements that make some po- litical systems particularly liable to gridlock. Some sys- tems, for example, give voice to so many disparate interests that even policies that will benefit most people cannot rise above the ensuing cacophony. Other systems produce stalemate by empowering many actors with the ability to veto each other’s ini- tiatives. Each of these potential sources of democratic failure can cause political outcomes to deviate from the preferences of the majority, and each may lead to poor development outcomes in the ways explained below. The connections between politics and devel- opment are, of course, many and complex. Politics determines the level of government services and who
benefits from them, which in turn may affect indi- vidual and regional inequalities. Politics also deter- mines what types of formal institutions are adopted and how they perform, which in turn may affect the efficiency not only of government but of the private sector as well. Finally, politics determines the extent to which democracies succeed in translating citizen preferences into effective and fair policies. The role of politics in the creation, mainte- nance and performance of formal institutions must be underscored. Politics has its greatest impact on development through its effect on institutions. The logic is clear: if politics matters for institutions, and institutions matters for development, politics must matter for development. The following sections look more closely at the links between politics and development by focus- ing separately on each of the different problems iden- tified above as potential sources of political failure.
Bias in Representation
Democratic constitutions usually proclaim that all citi- zens are equal in political terms. In practice, how- ever, some groups—either because they are better organized, more politically engaged, or more knowl- edgeable, wealthy or socially prominent—are able to gain greater attention from elected politicians than others. The extent of such bias in representation var- ies across democratic systems depending on such fac- tors as the levels of political involvement by different groups of citizens, and the characteristics of electoral and other political institutions. Bias in representation toward wealthier or better-educated citizens can clearly promote inequali- ties in the distribution of income and opportunities for social advancement. Thus, university education might be unduly subsidized at the expense of primary or secondary education; public works might be di- rected toward better-off regions or neighborhoods at the expense of disadvantaged ones; and taxes may be too low to pay for the social investments needed by the majority. Such policies will not only exacerbate inequalities, but can also produce economic inefficien- cies, thus affecting economic growth and impeding social development. Bias in representation toward narrow orga- nized interests can lead to other sources of economic
170 Chapter 4
inefficiencies. Organized interests are an important element of the political game in democracies. Some organizations represent a fairly broad cross-section of citizens, such as farmer associations, labor unions, industrial associations or consumer groups. Others represent narrower interests, such as coffee growers, textile laborers or government employees. If repre- sentation is heavily biased toward narrow interests, economic policies (including taxation, public invest- ment, pricing, trade and exchange rate policies) will likely be inefficient, and overall economic growth will suffer as a result. Indeed, the disproportionate influ- ence of relatively narrow interests on public policy has often been seen as a powerful element underlying the economic decline of once prosperous nations and the poor performance of developing countries, includ- ing those in Latin America. 8
Agency Problems
Political representation involves a complex transac- tion: people exchange votes for a catalogue of prom- ises. The “contracts” that regulate this transaction have two types of problems. First, they are impos- sible to enforce if only because no courts have juris- diction over them. Second, they are incomplete in the sense that they specify only general guidelines, leav- ing much open to interpretation and making it diffi- cult to determine and verify when a breach of con- tract has occurred. Although political institutions can ameliorate these problems, they can never completely solve them, meaning that politicians will always en- joy some leeway to pursue their own agendas and to extract rents. If the public lacks the means to punish unre- sponsive and corrupt politicians, as will happen when agency problems are pervasive, government perfor- mance will suffer. Corruption will flourish as politi- cians pilfer public funds and freely engage in the busi- ness of auctioning regulations and laws. This in turn will hamper the ability of the government to provide public services, and will increase the costs of doing business and the level of uncertainty for investors. Worse yet, corruption and government inefficiency may cause many disenchanted people to withdraw from politics, giving politicians more leeway, and thus compounding the original problems. In sum, a vicious circle of agency problems and government corrup-
tion is a distinct and disturbing possibility in demo- cratic regimes. Moreover, the existence of agency problems impedes one of the main channels through which democracy can spur development. One of the main advantages of democracies, at least from an efficiency point of view, is that they empower people to take part in formulating solutions to their own problems. One of the main difficulties of democracy, on the other hand, is that people do not participate directly in most public matters but through their elected rep- resentatives. So if the views of citizens are lost in that process—or, more precisely, if their views are replaced with the narrow views of their representa- tives—a wealth of valuable local knowledge will be lost and democracy will lose much of its appeal. Agency problems can undermine the whole idea of democracy. Simply stated, if political delega- tion does not work, democracy does not work. There- fore, solving agency problems not only will diminish corruption, increase government efficiency and facili- tate growth, but also will restore the democratic ideal of government of the people, by the people, and for the people.
Aggregation Problems
Politics in democratic societies requires considerable give and take. There are often so many interests to reconcile and so many factions to please that the pro- cess is not always expeditious. Compromises are not always worked out, bargains not made, and conflicts not settled, which means that changing the status quo, even if favorable to the majority, can become an exer- cise in futility. This difficult and laborious process of recon- ciling divergent interests and opinions frequently be- comes a source of public disenchantment with poli- tics. At the same time, the very essence of democracy is that the people, in all their diversity, be represented. Thus, for the democratic game of compromise to pro- duce expeditious and fair responses to collective prob- lems, a balance must be struck between the broad and
(^8) See Olson (1982) and Bates (1981).
172 Chapter 4
friends, trying to convince others in political matters, and working for a political candidate.^10 Table 4.1 shows average patterns of political participation across five education groups: primary or less, some secondary, secondary, some college (includ- ing technical training), and college. Differences in participation are small for voting but large for the other forms of participation. College educated indi- viduals are twice as likely as those with only primary education to talk about politics with friends, but only slightly more likely to vote. In sum, substantial in- equality of political participation across education groups is only apparent as one moves from voting to more involved forms of political engagement. The patterns of political participation by in- come quintile are shown in Table 4.2. As expected, the results closely mirror the previous table; there is no discernible trend for voting and there is a steady increase in participation across quintiles for the other types of political participation. The demographics of political participation are shown in Table 4.3. Sur-
prisingly, participation changes very little across age cohorts, though it is slightly smaller for the youngest and oldest cohorts. This pattern is very similar for all the types of participation considered here, except that it holds less clearly in the case of paying attention to political news. Are the patterns of political participation in Latin America any different from those of other re- gions of the world? A tentative answer to this question is no. The available evidence shows, for example, that the patterns of political participation across education and income categories are, if anything, flatter in Latin America than in the United States.^11 Similarly, politi- cal participation in the United States varies much more with age than it does in either Latin America as a whole or in any Latin American country individually.
(^10) A detailed analysis of the patterns of political participation in Latin America is presented in Gaviria, Panizza and Seddon (1999). (^11) See Wolfinger and Rosenstone (1980) for a comprehensive analysis of patterns of political participation in the United States.
Political Participation by Education
Participation by years of schooling (%) Ratio Primary Some Some College Primary/ Political activity or less secondary Secondary college or more College
Vote 66.1 68.6 68.0 67.4 70.0 1. Pay attention to news 43.4 47.4 52.1 59.6 69.6 1. Talk about politics 24.0 27.2 31.1 37.6 46.4 1. Try to influence others 14.2 14.3 16.4 18.1 20.8 1. Work for a candidate 9.7 9.0 10.7 13.6 15.1 1.
Source: Latinobarómetro, various years.
Table 4.
Political Participation by Income
Participation by income quintile (%) Ratio Political activity First Second Third Fourth Fifth 5th/1st
Vote 64.1 67.2 69.5 68.7 66.9 1. Pay attention to news 40.1 44.8 51.1 54.1 58.0 1. Talk about politics 21.6 25.5 29.3 32.2 36.8 1. Try to influence others 12.8 13.7 15.8 16.3 18.4 1. Work for a candidate 10.2 8.9 8.9 11.0 13.2 1.
Source: Latinobarómetro, various years.
Table 4.
Political Institutions, Accountability and Government Performance (^173)
The World Values Survey also offers some evidence indicat- ing that political participation in Latin America is not particularly unequal by international stan- dards.^12 The data show that for the European countries as a whole, in- dividuals in the top quintile of the income distribution are 1.87 times more likely to talk about politics with friends than those in the bot- tom quintile. The corresponding figure for Latin America is 1.88. Similar numbers are obtained for the category of working for a political candidate— 2.24 for Europe and 2.18 for Latin America. To what extent can policy outcomes in Latin America be explained by participation differentials? Although answering this question is difficult, we can establish the extent to which the observed higher rates of participation among more wealthy individuals can potentially bias social decisions in their favor. A natu- ral way to address this question is by computing the location of the median participant in politics—that is, the percentile of the person with the median in- come among those who participate in the political activity under consideration. Obviously, if political participation is evenly distributed, the median par- ticipant will be located in the 50th percentile, mean- ing that the income of the median participant would
match that of the median citizen. Higher values indi- cate that the median participant is richer than the median citizen and lower values the reverse. The locations of the median participant for three distinct forms of participation and 17 countries are depicted in Figure 4.6. If voting is all that mat- ters, the distortions caused by inequality of participa- tion will be minimal. However, if other types of po- litical engagement matter as well (as would be the case if better informed and more vocal individuals are more successful in attracting the attention of politicians), inequality of participation could introduce significant distortions in some countries (including, in particu- lar, Peru, El Salvador and Panama). However, defini- tive conclusions are impossible without information about the relative efficacy of the different forms of political participation. No discussion of political representation is complete without mention of its organizational as- pects, since participation is often a collective endeavor. Individuals with common interests usually find it to their advantage to join forces in their quest for repre- sentation. Collective participation is difficult to sus- tain, however, because individuals have strong incen- tives to free ride on the effort of others. When and how political organizations can overcome these in- centives is still an open question, but the microeconomics of group formation unambiguously indicate that small groups will, all things being equal, be more successful than large ones. This advantage
Political Participation by Age (In percent)
Age Political activity 18-29 30-44 44-60 >
Vote 66.8 67.0 68.3 67. Pay attention to news 47.0 51.1 51.4 52. Talk about politics 27.5 31.2 31.2 27. Try to influence others 15.5 16.2 16.2 13. Work for a candidate 9.5 11.4 12.0 9.
Source: Latinobarómetro, various years.
Table 4.
Location of the Median Participant in Political Activities
Figure 4.
Argentina Bolivia Brazil
Colombia
Costa Rica
Chile
Ecuador El Salvador Honduras Guatemala
Mexico
Nicaragua
Panama
Paraguay
Peru
Uruguay
Vote Political news Talk politics
Source: Latinobarómetro, various years.
Venezuela
0
10
20
30
40
50
60
70
80%
(^12) This survey includes 17 European countries and four Latin Ameri- can countries (Argentina, Brazil, Chile and Mexico).
Political Institutions, Accountability and Government Performance (^175)
Some electoral systems give politicians incen- tives to cultivate personal followings, while others give strong incentives to adhere to their parties’ directives. In the first case, the careers of politicians will hinge on whether they are able to establish strong links with their constituents. In the second, their careers will depend on whether they remain on good terms with the party leadership. In principle, politicians who con- cern themselves more with their personal relation- ship with voters will be more likely to fall prey to geo- graphically concentrated interests. Conversely, poli- ticians who care mostly about their parties will be more likely to respond to broader national interests, assum- ing that parties have a strong national following and party leadership is relatively centralized. However, electoral systems that dispose poli- ticians to care mostly about parties may also under- mine effective representation. Politicians in these sys- tems, worried about staying on good terms with the party bosses, will have weak incentives to find out what voters want. As a result, the ties between politicians and voters will loosen, which may in turn allow nar- row class or sectoral (but not necessarily geographic) interests to gain undue representation. In sum, while heavily party-centered political systems can reduce biases of representation toward geographically con- centrated interests, they can actually exacerbate bi- ases in representation toward other narrow nongeographical interests. Box 4.3 presents an index to measure the dif- ferences across countries in the extent to which elec- toral institutions give politicians more or less incen- tives to cultivate personal relationships with voters.^14 A high score in the index indicates a high proclivity among politicians to cultivate a personal following with voters, and, in our interpretation, a high pro- clivity to respond to narrow geographic interests. A low score indicates a high proclivity among politicians to follow the directives of their party leaders, and, in our interpretation, a higher proclivity to respond to national interests. Figure 4.7 compares Latin America to other regions of the world in terms of the index. The evi- dence suggests that politicians in Latin America have more incentives to be on good terms with party lead- ers than do politicians anywhere else in the world. So Latin American parties are strong in this particular sense—they hold the keys to political power. In our
interpretation, this implies that electoral systems in Latin America provide relatively weaker incentives to respond to geographically based demands for special treatment. There are substantial differences within the region regarding the extent to which electoral systems entice politicians to establish personal links with vot- ers. Table 4.4 shows the scores of the index and its com- ponents for most countries in Latin America.^15 Closed lists are the most common ballot structure in the re- gion, which explains why most countries have a ballot score of 0. Chile’s open lists and Colombia’s multiple party lists are the main exceptions. Most countries have proportional systems in which votes are pooled across the entire party. The main exception is Colombia, which has a peculiar system in that parties present multiple lists and votes are pooled only across candi- dates within a single party list. Similarly, most coun- tries have systems in which citizens cast single votes for parties. The main exceptions are Mexico, where a large portion of the legislators are elected in single member districts, and Brazil, where voters can choose individual candidates within party lists. Do Latin American democracies suffer from unusual problems of bias in representation? The evi- dence presented above is not conclusive. On the one
Index of Incentives to Cultivate Personal Relations with Voters, by Region
Figure 4.
0 0.2 0.4 0.6 0.8 1
Latin America
Eastern Europe
Developed
Africa
East Asia
Asia
Middle East
Source: Parline Online and IDB calculations.
(^14) The index is based on the theoretical work of Carey and Shugart (1995) and Shugart (1999). (^15) The English-speaking Caribbean countries were not included, as one house of the bicameral legislatures is usually appointed and the lines of accountability for these legislators are not clear.
176 Chapter 4
hand, inequality of political participation is, if any- thing, less serious a problem in Latin America than it is in advanced industrial democracies. Further, electoral systems in Latin America do not appear to provide large incentives for politicians to respond to geographically concentrated interests. On the other hand, political systems in some Latin Ameri- can countries are heavily centered around parties, which may allow narrow interests (especially those that can successfully court party leaders) to gain un- due representation. Further, campaign financial regulations (and electoral regulations in general) are not only lax in many Latin America countries, but are also rarely enforced, which may also allow orga- nized interests to buy political influence.
Agency Problems
The extent of agency problems depends on many fac- tors. Four are considered here: levels of political par- ticipation, the freedom of the press, the strength of political parties, and the presence of institutional checks and balances.
Elections are the main mechanism through which citizens can enforce political contracts in demo- cratic societies. It goes almost without saying that elec- tions are more effective when people participate in politics and are better informed about all that politi- cians have said and done. In general, greater political participation means that politicians face more scrutiny and enjoy less leeway to pursue their own agendas.
Political Participation
How do the levels of political participation in Latin America compare with those in other regions of the world? While the previous section looked at inequal- ity in participation—which can bias social choices to- ward the more politically active citizens—this section examines differences in levels of participation, which can influence the ability of citizens to monitor politi- cians. Figure 4.8 shows that voter turnout in Latin America is midway between the high levels of Euro- pean countries and the low voter turnouts of African countries. (Differences in voter turnout are discussed
Index of Incentives to Cultivate Personal Relations with Voters
Components of the index
Country Pool Ballot Vote Index
Argentina 0 0 0 0. Bolivia 0 0.3 0.5 0. Brazil 1 2 1.5 1. Chile 0 1 0 0. Colombia 1 1 0 0. Costa Rica 0 0 0 0. Dominican Republic 1 0.5 0.5 0. Ecuador 0 0 0 0. El Salvador 0 0 0 0. Guatemala 1.6 0 1.6 1. Haiti 1 1 1 1. Honduras 0 0 0 0. Mexico 0 0 1.1 0. Nicaragua 0 0 0 0. Panama 0 0 0 0. Paraguay 0 0 0 0. Peru 0 0 0 0. Uruguay 0 1 0 0. Venezuela 1 0 0 0 0.
(^1) Venezuela has up to three (out of 57) senators elected at the national level to compensate for disproportional seat/vote shares. In the lower house, up to five house seats (out of 206) per party are distributed according to national vote totals and assigned to underrepresented constituencies. Source: Parline Online.
Table 4.
178 Chapter 4
Other measures of political participation also show large differences within the region. In terms of people who self-report paying attention to political news and talking about politics with friends, Ecuador and Para- guay are at the top and Chile and Guatemala at the bottom (see Figure 4.10). Surprisingly, there is no association between voter turnout and these more sophisticated forms of political engagement.
Press Freedom
When trying to use elections to punish or reward poli- ticians, voters can face some serious information prob- lems. Voters observe some general outcomes and have some vague ideas about policy. But they are usually
very uncertain about how outcomes relate to policies, and they often have little information about the track record of politicians running for office. The media has traditionally played an impor- tant role in providing information about political matters. An inquisitive media can provide vital infor- mation to help voters evaluate the extent to which the actions of politicians conform to their electoral promises. Similarly, the media can uncover and pub- licize corrupt practices by politicians and their asso- ciates. In sum, a free and independent media can di- minish the ability of politicians to breach electoral contracts and to extract rents. Freedom House has recently put together an index to measure the extent to which journalists are
Freedom House Rating of Press Freedom, by Region
Figure 4.
0 20 40 60 80 100 Index of press freedom
Middle East
Africa
Asia
Commonwealth Independent States
World
Latin America
Developed countries
Source: Freedom House (1999).
Figure 4.
Africa
Asia
Latin America
Middle East
Eastern Europe
East Asia
Developed countries
Circulation of Daily Newspapers per 1,000 Inhabitants
Source: World Bank (1998).
0 50 100 150 200 250 300 350
Figure 4.
0 20 40 60 80 100
Guatemala ColombiaHaiti
ParaguayJamaica VenezuelaBolivia
MexicoPeru
El SalvadorHonduras
EcuadorGuyana Trinidad and TobagoPanama
NicaraguaBrazil
Argentina Dominican RepublicChile
Costa RicaUruguay
Voter Turnout in Latin America, 1990- (In percent)
Source: IDEA (1997).
Figure 4.
0 10 20 30 40 50 60 70
Chile
Guatemala
El Salvador
Venezuela
Honduras
Costa Rica
Brazil
Mexico
Colombia
Argentina
Nicaragua
Peru
Bolivia
Uruguay
Panama
Paraguay
Ecuador
Political news Talk about politics
Alternative Forms of Political Participation in Latin America (In percent)
Source: Latinobarómetro, various years.
Political Institutions, Accountability and Government Performance (^179)
free to follow their leads and report their findings without being harassed by the government or by other powerful elements of society. 17 Although one can quarrel endlessly about the ranking of this country or that, the index provides a useful comparison of press freedom around the world. As Figure 4.11 shows, Latin America as a whole has a relatively free press compared to other regions of the word—less free than more developed countries, but more so than other developing regions. Press freedom is not the only important vari- able in terms of the availability of information for voters. Two societies can differ substantially in terms of the ability of the media to monitor politicians, even if they have similar constraints on press freedom. The key element in this respect is the degree of inquisi- tiveness of the society in question. More inquisitive societies can be expected to have more and better in- formation outlets, and therefore will find it easier to monitor politicians. Figure 4.12 shows, for example, that circulation of daily newspapers in Latin America is well below the level one would expect given the region’s level of development and its higher degree of press freedom.
Political Parties
Another factor that can affect the enforceability of political contracts is the strength of political parties.^18 If parties are reasonably cohesive and disciplined and have fairly deep roots in society, they can improve the effectiveness of democratic delegation by reducing the information costs of voting, making it easier for citi- zens with little time and political information to par- ticipate in politics. In contrast, if political parties are weakly organized and undisciplined, voters will not trust them as conveyors of reliable political informa- tion and hence will face the daunting task of becom- ing informed about the policy positions and records of each individual candidate.^19 Political parties can also ameliorate agency problems by constraining the actions of elected poli- ticians. The power of parties in this regard increases to the extent politicians have something to gain by not deviating from their parties’ general directives. If parties lack continuity and cohesion, politicians will be more prone to deviate from party ideals in order to satisfy personal political ambitions. Thus, relatively
institutionalized political parties can provide a check on excessive political entrepreneurialism. There is a subtle but important distinction between the role of parties emphasized here and that which was underlined earlier. The previous section argued that political parties are important because they are more likely than individual candidates to articulate the wishes of the majority. This section argues that parties are important because they con- vey information about complex political issues and monitor elected politicians. The problem there was bias in representation; the problem here is agency. The emphasis there was on the electoral laws that give parties power over individual politicians; the emphasis here is on the extent to which parties are cohesive and rooted in society. The attributes of political parties that contrib- ute to effective democratic delegation are, for the most part, encapsulated by three conditions: (1) party sup- port is relatively stable over time; (2) parties have rela- tively solid and stable roots in society; and (3) parties are perceived as central to determining who governs and as indispensable for the progress of the country. 20 The first dimension of the strength of party systems can be measured by an index of volatility in electoral support for parties from one election to the next. This index is computed by adding the net change in the percentage of seats (or votes) gained or lost by each party from one election to the next, and then dividing by two. 21 Table 4.5 shows a wide variation in electoral volatility for congressional elections (lower
(^17) In constructing this index, Freedom House considers not just the formal laws and the constitution but also current and practical con- straints on the press. If journalistic freedom is impeded by threats from armed groups or criminals, or by illegal government-led acts of intimi- dation, then this counts against press freedom potentially as much as legal restrictions. (^18) Studies on political parties in Latin America include Mainwaring and Scully (1995) and Hagopian (1998). (^19) See Lupia and McCubbins (1998) for insightful information on the cognitive dimension of political participation. (^20) See Mainwaring and Scully (1995). (^21) Although large shifts in voter support for parties and changes in the identity of the major parties may complicate agency problems, they could be a positive sign. A relatively high degree of volatility could reflect an efficient response to the emergence of new critical issues dividing the electorate, an opening of the political system to greater competition through electoral reforms, or a broad rejection of tradi- tional parties perceived as ineffective or corrupt.
Political Institutions, Accountability and Government Performance (^181)
1970s, attachment to parties is still comparatively low in Latin America. With the sole exception of Uru- guay, Latin American countries fall below the Euro- pean average with respect to the percentage of citi- zens who identify with a political party. The potentially positive benefits of parties with strong societal links must be tempered by con- sidering the mechanisms by which such ties with citi- zens are established and maintained. As happened in many U.S. cities in the late 19th^ and early 20 th^ centu- ries, many Latin American parties have established
loyal followings by distributing state jobs, providing favors, and otherwise directly or indirectly buying votes. If the political contract between citizens and politicians becomes an exchange of votes for the de- livery of an individual benefit (such as a job for a fam- ily member or a special benefit for a family business), elections do not serve to transmit voters’ preferences for the provision of collective goods or to ensure that politicians follow these expressed preferences. Al- though data are not available to measure the com- parative importance of such clientelistic practices, it is probable that free market reforms and the reduc- tion in the scope of the state in most countries have reduced the amount of resources available for pur- chasing political support. Further, less clientelism could partially account for the apparent decline in party loyalty and the increase in electoral volatility experienced in some countries during the 1980s and 1990s. The third dimension proposed to measure the strength of political parties is the degree to which citi- zens and societal groups perceive that parties are cen- tral actors in determining who governs and in shap- ing the country’s policy direction and rate of progress. One obvious measure of the centrality of parties is the degree to which citizens place confidence in them. If parties are distrusted or viewed unfavorably, citi- zens, organized interests and politicians will tend to eschew them as intermediaries in the political pro- cess. Figure 4.14 shows that confidence in political parties is closely related to the degree to which citi- zens identify with them. According to both indica- tors, parties seem particularly relevant in Uruguay, Paraguay, Nicaragua and Honduras, and less impor- tant in Peru and Brazil. Figure 4.15 shows the percentage of people in selected Latin American countries who consider parties as indispensable to the progress of the coun- try. The results show a familiar pattern. While rela- tively few respondents mentioned parties as essential in Brazil, Ecuador, Peru and Bolivia, the opposite was true in Uruguay, Honduras, Costa Rica, and Chile. Mexico and Paraguay, however, deviate significantly from their previous positions. Even though Mexican citizens do not appear to identify closely with parties and have only moderate confidence in them, they do see parties as indispensable to the progress of the coun- try. In Paraguay, the opposite pattern is observed.
Percentage of People Who Feel Very Close or Fairly Close to a Political Party
Figure 4.
0 10 20 30 40 50
BrazilPeru
MexicoBolivia
GuatemalaColombia
VenezuelaEcuador ArgentinaCosta Rica
LATIN AMERICAChile El SalvadorPanama
NicaraguaHonduras ParaguayUruguay
PortugalSpain
BelgiumIreland
EUROPEFrance LuxembourgGermany
United KingdomDenmark
NetherlandsGreece
Italy
Source: Latinobarómetro, various years, and Eurobarometer (1991).
Figure 4.
15
20
25
30
35
40
Confidence in parties
0 10 20 30 40 50 Identifying with parties
Argentina
Chile
Colombia
Costa Rica
Ecuador
Guatemala
Honduras
Nicaragua
Panama
Peru
Paraguay
El Salvador Uruguay
Venezuela
Party Identification and Confidence in Parties in Latin America (In percent)
Source: Latinobarómetro, various years.
Mexico
Bolivia Brazil
182 Chapter 4
While citizens appear to trust and feel close to par- ties, they do not see them as indispensable to progress. How strong, then, are political parties in Latin America? The overall picture that emerges from this analysis is that parties appear to be especially strong in Uruguay and still relatively strong, but perhaps weakening, in Chile, Costa Rica, Honduras, Mexico and Paraguay. Parties also seem to organize public opinion and be relatively well valued in El Salvador and Nicaragua. On the other end of the spectrum, parties are relatively weak and distrusted in Bolivia, Brazil, Ecuador and Peru.
Institutional Checks and Balances
Political contracts involve a commitment by elected officials to fulfill electoral pledges, follow the law and respect the constitution. This is important because, once elected, politicians have an incentive to try to change the nature of the political game in order to augment their power or to increase their chance of retaining this power in the future. Politicians may also be tempted to use their positions of power for per- sonal enrichment. Democracy thus requires institu- tions that protect the extant rules and safeguard pub- lic interests from unchecked politicians. If politicians are to be prevented from ignor- ing or modifying the rules of the game in order to suit their needs, the constitution must define the di- vision of responsibilities between the different branches of government and specify fairly stringent
procedures through which the constitution can be changed. In addition, if the constitution and other laws of the land are to be more than just words, there must be an independent and effective judicial system that enforces them. Finally, abuses of authority, misman- agement or deception are not likely to be exposed unless there is a legislature in which opposition par- ties can scrutinize the conduct of government offi- cials, openly question and criticize government per- formance, and launch criminal investigations. The agency problems presented by relatively weak legislative institutions are compounded by the relative weakness of the judicial branch. Constitutions provide for the independence of the judiciary in all Latin American countries, yet the independence of the courts has not always been guaranteed. This is evidenced by the frequent abrogation of judicial in- dependence, dismissal, transfer and reassignment of judges, and denial of enforcement of judicial decisions. This traditional lack of judicial independence in Latin America, along with the perception of ineffi- ciency, has contributed to the low level of trust in the judiciary. Figure 4.16 shows that trust in the judicial system in Latin America is below that of Europe.^24 Although levels of confidence in the judiciary do vary
(^24) The information in Figure 4.16 dates to the mid-1980s because re- cent Eurobarometer surveys have not included the question on confi- dence in the judiciary.
Percentage of Respondents Who View Political Parties as Institutions Indispensable to National Progress
Figure 4.
0 20 40 60 80
Paraguay
Brazil
Ecuador
Peru
Bolivia
Guatemala
Argentina
Panama
Colombia
Chile
Nicaragua
Venezuela
El Salvador
Costa Rica
Honduras
Uruguay
Mexico
Source: Latinobarómetro, various years.
Confidence in the Judiciary in Latin America vs. Western Europe, 1985 (In percent)
Figure 4.
0 20 40 60 80
Peru Argentina Mexico
EcuadorBolivia
VenezuelaPanama Guatemala LATIN AMERICAColombia
Paraguay Chile NicaraguaBrazil
El SalvadorHonduras
Costa Rica Uruguay
FranceSpain Italy EUROPEBelgium
United Kingdom
West GermanyNetherlands
Source: Latinobarómetro, various years, and Eurobarometer (1986).
