Michael Coppedge

Kellogg Institute, Hesburgh Center

University of Notre Dame

Notre Dame, IN 46556

tel 219/631-7036

fax 219/631-6717

e-mail coppedge.1@nd.edu
 
 

Final (4/11/99) draft of chapter for a book to be edited by Larry Diamond, Richard Gunther, and Marc Plattner, based on papers presented at the conference on "Political Parties and Democracy," International Forum for Democratic Studies, National Endowment for Democracy, Washington, D.C., November 18-19, 1996. I am grateful to the editors, Charles Kenney, Marcelo Leiras, Scott Mainwaring, Pierre Ostiguy, and Gabor Toka for their many valuable contributions to this analysis, and to Miguel Basáñez for permission to use the 1995 LatinoBarómetro data.

Are parties in decline? In many ways, this is like asking whether musicians' performances have declined. Even in classical music or the relatively stable and legitimate party systems of Western Europe, it is a difficult question to answer, simply because styles and tastes change. For this reason the editors of this volume rightly ask whether the perception of party decline might be the result of judging contemporary parties by old-fashioned models. They therefore suggest that we propose new models for describing parties and that we evaluate parties according to their ability to perform certain basic functions.

The question becomes even harder to answer when the focus shifts to Latin American parties. To continue the musical analogy, we are no longer talking about Mozart and Beethoven; they are more like pop musicians in several respects. First, there has been tremendous turnover: there are a handful of perennial favorites, a modest number that last a decade or so, and quite a few "one-hit wonders" or fenómenos. Second, most parties, like most bands, have never succeeded in winning much of a following. Third, there has always been a wide range of quality. Just as there are great bands and awful ones, there have been parties that perform their basic functions well and others that are dysfunctional in various ways and to varying degrees. This is true across the countries of the region and within most of the countries as well. Finally, we must ask whether a classical model rooted in Western Europe was ever a useful standard. For all these reasons, the question must be reformulated before it can be answered for Latin American parties.

I reformulate the question in several ways. First, given the accelerated pace of change in Latin America, what is the nature of change over the past 10-15 years rather than over many decades? Second, are there any systematic differences between the parties that were important 30 years ago and those that were important at the end of the 1990s? Third, because high turnover makes it impractical to compare the same parties over time, have party systems become more functional or less so in recent years? And finally, do the forces that drive the process of turnover favor the survival of parties with certain characteristics?

There are only a few general tendencies across these cases. In the long run, population growth, urbanization, and the spread of the mass media have modified some of the ways in which the best organized parties operate. Although they continue to mobilize supporters during election campaigns, they rely less on routine mobilization and socialization through party-affiliated social organizations. Those that once hoped to finance themselves through dues collection have become dependent on outside campaign financing instead. And all parties that can afford it rely extensively on polling and the mass media to tap public opinion and get out their message.

In all other respects, individual parties have experienced very little organizational change. However, many parties have been in effect replaced by other parties with different organizational characteristics. This process of replacement can be understood as "political Darwinism": the survival of the parties best adapted to the political environment of the "Lost Decade" of austerity and economic stagnation from 1982 until the early 1990s. This environment tended to select in favor of right or center-right governing parties, and personalistic or center-left opposition parties. Differences in initial conditions, economic performance, and skill at adaptation meant that different party systems evolved in different directions. Surprisingly, however, these changes rendered party systems less functional only in Brazil, Ecuador, and Peru, among the major countries. This article illustrates these tendencies with brief case histories of Acción Democrática in Venezuela, APRA in Peru, and the Peronist party in Argentina.

In some Latin American countries it is normal for parties to shrink or disappear. In 166 20th-century legislative elections in 11 Latin American countries, approximately 1,200 parties competed.⁽¹⁾ Of these, only 15 participated in all the elections held in their country, and only 3 contested as many as 20 elections. More than 80 percent ran in just one election before becoming defunct.

Another way to measure the degree of change after 1982 is to calculate volatility rates using the election closest to 1982 as a baseline and the most recent election as an endpoint.⁽²⁾ Table 1 reports these rates of change for the 11 Latin American countries with the most electoral experience. Table 1shows that the initial party systems of Peru, Brazil, Ecuador and Bolivia sustained severe damage, with less than half of their party systems remaining in the same form. The Peruvian party system of 1980 was nearly wiped out. Major parties in Venezuela, Argentina, and Mexico also lost a great deal of support. Only in Chile, Colombia, Costa Rica, and Uruguay did the pre-1982 party systems remain more or less intact. In these countries much of the change consisted of some splintering and name changes, in addition to some healthy vote fluctuation.

The generally high rate of turnover makes it impossible to trace long-term changes in the characteristics of all parties. However, some generalizations are possible with respect to the small number of parties that managed to survive several decades. Some of these parties were similar to major parties in Western Europe. There were parties of the left, such as the Chilean Socialist and Communist parties, that tended to have the most elaborate and coherent ideology and sought a working-class base of support (not always with success) and a mass-based organization. There were also "radical" parties, such as the Chilean Radicals and the Unión Cívica Radical in Argentina, that promoted classic political liberalism, targeted a predominantly middle-class base of support, and were organized more on the caucus party model. There were also some prominent Christian democratic parties, such as the Partido Demócrata Cristiano in Chile and COPEI in Venezuela, that possessed mass-based organizations and were inspired by Catholic social doctrine but moved toward the catch-all party model. These parties were not carbon copies of their European cousins, however, in that they tended to rely more on clientelistic mobilization techniques, contain more important personalistic cliques, rest on a more diverse social base, and (except for the parties of the left) be less concerned with ideological purity.

Some other important parties have been associated with a distinctly Latin American type, the "national revolutionary" or "Aprista" party. This family included the Alianza Popular Revolucionaria de América (APRA) in Peru, Acción Democrática (AD) in Venezuela, Liberación Nacional (PLN) in Costa Rica, and for some scholars, Mexico's Institutional Revolutionary Party (PRI) and Bolivia's Movimiento Nacional Revolucionario (MNR). The national revolutionary parties attempted to unite the middle class, workers, and peasants behind a diffuse nationalistic and anti-oligarchical platform. They were structured around sectoral organizations designed to mobilize unions, peasants, students, and some middle-class occupational groups, and were heavy practitioners of clientelistic mobilization.

Several other historically important parties do not fit any type well. The Blancos and Colorados of Uruguay have been considered parties for some purposes and alliances of factions for others. They had caucuses at the local level, but these local groupings were divided among well-institutionalized regional and national fractions that represented considerable ideological diversity. Neither party had close ties to unions or other social organizations, aside from the Nacional and Peñarol soccer clubs, but both were buoyed by deep party identification and extensive use of patronage. The Liberals and Conservatives of Colombia were alliances of personal and local clientelistic machines, practically indistinguishable in terms of ideology, but bound together by the electoral value of their parties' historic labels. And the Justicialista (Peronist) Party in Argentina was the most unique of all: an organization strongly backed by both neighborhood caucuses and most blue-collar unions, united more by a series of personalistic leaders than by any common ideology, all open practitioners of clientelism and nepotism. This party's structure was kept exceptionally loose, as its most passionate defenders preferred to call it a movement of the whole Argentine pueblo ("the people," in the populistic sense) rather than a party.

Has the nature of these important parties changed over the decades? In four respects, yes. First, most of these parties, if they existed before the 1950s, used to mobilize and socialize their membership through party-affiliated sports clubs, dining clubs, literacy workshops, and discussion groups, much as mass-based parties in Europe did in the 1930s. All of these activities have greatly declined as a consequence of population growth, urbanization, rising standards of living, greater personal autonomy, and higher levels of education. Most have fallen into disuse if they exist at all. Second, many of these parties initially attempted to be self-financing through dues collection; this effort was never all that successful, so now all of them have become dependent on outside public and private donations. Third, technological change has led all important parties in Latin America to make extensive use of polling, mass-media campaign advertising, and increasingly sophisticated U.S.-style campaign techniques. Fourth, more and more parties are experimenting with primaries for the selection of their candidates for president, congress, and governorships.

These secular changes might suggest that major Latin American parties have turned away from personal contact and face-to-face, patron-client relationships as socialization and mobilization techniques. There is some change in this direction, but it should not be exaggerated, because the new techniques often supplement the old ones rather than replacing them. For example, almost all of these parties continue their mass mobilization efforts during campaigns, partly because bringing lots of supporters to rallies is one of the few ways local brokers can demonstrate that they can get out the vote and are valuable to the party leadership. As long as clientelism is rampant in Latin American politics, the mass mobilization will continue. And as long as poverty and deep inequalities plague Latin American societies, clientelism will remain a favorite political tool.

Because the parties mentioned so far are only a small fraction of all the parties, and are not representative of the others, it is necessary to examine changes in a country's whole system of parties. This can be done by identifying certain basic functions that party systems should perform and certain structural characteristics of party systems that are functional according to these criteria, and analyzing indicators of these characteristics. Party systems matter for democracy in two principal ways. First, they are the chief vehicles for representation, and therefore affect the quality of democracy. And second, they affect governability, especially in the legislative arena, and therefore indirectly affect the stability of a democratic regime in the long run. Both good representation and governability require legitimacy and are enhanced when there is strong party identification. And in the long run, good representation promotes governability.⁽³⁾ In a more immediate sense, however, what is good for democracy is not always good for governability, or vice versa, and for this reason we have conflicting notions about what sort of party system would be "best" for stable democracy. Because many people focus exclusively on the long-term compatibility between representation and governability, it is instructive to compare party systems at the extremes of fragmentation and polarization in order to highlight the tradeoffs that present themselves in the short term.

If our goal were perfect representation, we would want a large number of parties, to represent all possible combinations of positions on all the relevant issues; and rigid parties that resist compromising on the mandate received from the voters and have sharp issue differences with other parties. Such a party system would be highly representative in the most pure sense of the term. But it would also tend to be divisive, polarized, and indecisive, and therefore dangerously inclined toward ungovernability in the legislative arena in the short term. If our goal were perfect governability, we would want just one party, to guarantee and mobilize full support for whatever the government decides to do; and a highly pragmatic party that strives for consensus and whose activists are always willing to compromise to achieve it. Such a party would be wonderful for governability (unless it alienated a substantial body of citizens and allowed them to organize an anti-system force), but also the very antithesis of democracy.

In the real world, we are willing to sacrifice some democracy and some governability in order to achieve as much as possible of both. Thinking about these tradeoffs is a useful way to identify a standard for evaluating how functional a party system is for stable democracy. With respect to fragmentation, there should be enough parties for meaningful competition but not so many that it becomes difficult to form governments and make decisions. A minimum of two parties is a reasonable lower limit. If there are, in effect, fewer than two parties (in the sense that there is one party that is much larger than any other), then the system is insufficiently competitive because the largest party is expected to win control of the national executive all the time. The Laakso-Taagepera index of the Effective Number of Parties, which ranges from 1.00 to infinity, counts parties after weighting them by their shares of the votes or seats, producing an "effective" number of parties that can be expressed in fractional terms, such as "2.63 parties."⁽⁴⁾ This index has become the standard indicator of party-system fragmentation in political science.

Meaningful competition can also occur among more than 2 parties, but at some point further fragmentation undermines governability by making it harder to form a working majority in the legislature. Many cutoffs for this perilously large number of parties are conceivable, but one conservative standard is the number at which it becomes impossible for just 2 parties to form a majority. This number depends on the size of the largest (and presumably governing) party. If the largest party controls more than half the seats, a coalition is unnecessary; if it control less than a quarter, any majority coalition must include at least 3 parties. Between one-quarter and one-half of the seats, this threshold is generally 4.0 to 4.5 parties, reaching a maximum of 4.57 parties when the largest party controls 37.5 percent of the seats.

With respect to polarization, the parties should take positions that are distinctive enough to provide the voters with a meaningful choice, but these positions should not be so far apart that they interfere with the construction of majorities for legislation and governing. One indicator of polarization is the dispersion of the vote away from the relative center of the party system, which takes on values between zero (all of the vote in one ideological bloc) and 100 (half of the vote at each of the ideological extremes).⁽⁵⁾ A minimum functional level of polarization would be 25, which is the lowest level that guarantees that no bloc wins more than half the vote and therefore ensures some competition between blocs. The maximum functional level of polarization can be set at 60, which corresponds to a perfectly even distribution of voters among all blocs. (This "flat" distribution marks the threshold between the "unimodal" or "single-peaked" distributions that are more concentrated than dispersed and the "bimodal" distributions that are more dispersed than concentrated.)

Figure 1 depicts changes in the levels of party-system fragmentation and left-right polarization in the 11 countries surveyed here using these two indicators. Lines in the figure define the functional ranges of fragmentation and polarization. As shown here, functional party systems have between 2 and 4.57 effective parties (calculated on the basis of seats) and are from 25 to 60 percent polarized. The figure also displays a line for each country that begins at the fragmentation and polarization levels at the "initial" elections as defined in Table 1 and ends at the corresponding values for the "final" election in the period of analysis. Overall during this period there was an increase in fragmentation and polarization (as well as an association between these two tendencies). But in order to evaluate these changes it is essential to consider each party system's starting point and the magnitude of the change, for these facts determine whether fragmentation and polarization made these party systems more functional or less so.

Brazil and Ecuador suffered severe fragmentation and polarization that moved their party systems into a zone of great dysfunctionality. Mexico and Colombia also experienced growing fragmentation and polarization, but because they began with too little of each, these changes were in the direction of more meaningful and competitive elections. Venezuela, Uruguay, and Costa Rica experienced some changes, but not of sufficient magnitude to move them out of the functional zone.

Partial changes occurred in other countries. In Bolivia the party system remained fragmented (but not excessively so: two-party coalitions could still be formed, in theory and in practice) during this period, but the level of polarization declined dramatically by 1993 (after a surge in 1985), to a borderline low level. Similarly, Peru maintained a manageable number of parties, but leaped from excessively high polarization in 1980 to excessively low polarization in 1995, due to the dominance of Fujimori's depoliticizing Cambio 90/Nueva Mayoría. In Chile, left-right polarization was lower in the 1990s, although some intense conflict remained over constitutional issues. Fragmentation increased to a potentially dysfunctional level, but the decline in polarization more than compensated for it, as solid multiparty coalitions sustained the first two democratic governments after Pinochet. The Argentine party system was not too fragmented in the 1980s, but in the base year (1973), it was dysfunctionally polarized. IP does not reflect polarization well in Argentina because the principal dimension of competition was historically between Peronists and everyone else rather than between left and right.⁽⁶⁾ By 1995, competition was perhaps insufficiently meaningful in left-right terms, but still quite meaningful in Peronist-anti-Peronist terms.

In summary, there were significant improvements in levels of fragmentation and polarization in Mexico; a partial improvement in Argentina, Chile, and Colombia; no net change in Bolivia, Costa Rica, Uruguay, and Venezuela; a partial decline in Peru; and disaster in Ecuador and probably Brazil. Eight of the 11 systems functioned as well as before or better, and only 3 deteriorated overall. The number of improvements is rather surprising, in view of the difficult economic times the region was going through. However, the claim that some Latin American party systems became more functional does not mean that they were functioning well, and still less that Latin American democracy in general was thriving. Many of these party systems started from great dysfunctionality; delegative democracy was on the rise in Argentina, Peru, and Venezuela at least; the drug trade was spreading corruption in the Andean region and Mexico; indigenous peoples were underrepresented in the same countries; the military was incompletely subordinated to civilian leadership in Bolivia, Chile, and Peru; the courts and the rule of law were weak everywhere except Costa Rica and Chile; and human rights were violated far too often.

These conditions make it easier to understand why Latin Americans held their parties and other institutions in such low esteem in the mid-1990s. As Table 5 shows for 7 countries, most Latin American respondents in the LatinoBarómetro survey of 1995 claimed to have little or no confidence in parties and, except for Uruguayans and perhaps Mexicans, claimed not to be close to any party. These subjective evaluations of parties correspond closely with the objective classifications of parties in Figure 1. Two of the countries (excluding Costa Rica, which was not included in the 1995 LatinoBarómetro survey) that ended up on the zone of functionality--Uruguay and Mexico--had less negative evaluations of parties than all the rest, and the country with the most dysfunctional party system, Brazil, also had the most negative evaluations. (Chile is the only outlier here, probably due to the unusual degree of cooperation among the parties in the governing coalition, which has moderated the problems associated with fragmentation and polarization.) These poor evaluations did not affect Latin Americans' majority preference for democracy in principle (Table 5); and they had only a light relation to the predominant dissatisfaction with actual democratic performance. These data suggest that political parties do not help legitimate the political system in these countries. If anything, the legitimacy of democracy may induce in the population a reluctant toleration of parties.

It should be clear by now that party politics in Latin America is a harsh struggle for survival with few survivors. It could be aptly called "political Darwinism" because there are several parallels between the evolution of party systems and the evolution of natural species. Both natural selection and the more artificial selection of political parties by voters involve competition for limited resources, whether votes or food; the winners of this competition grow (in popular support or numbers) while the losers decline and eventually become extinct; the survivors tend to reproduce themselves more or less faithfully for the next round of competition, although with some capacity for innovation; and the best adaptations to the environment are favored for growth and survival in future rounds.⁽⁷⁾

This view of the process suggests that four basic conditions shape the evolution of parties and party systems: (1) the amount of stress to which the party system is subjected; (2) the nature of the stress, which determines which party characteristics are rewarded and which are punished; (3) the vulnerability of the parties to this kind of stress; and (4) the parties' capacity to adapt appropriately. How did each of these conditions apply to Latin America after 1982? Did they combine to select in favor of parties with certain characteristics?

Table 2 ranks the biggest winners and losers in elections during the Lost Decade and provides some clues about the selection criteria of the Darwinian struggle for survival. First, all of the biggest losers were governing parties at some time during this period. Apparently incumbency was frequently very costly. However, this was not always the case, as some of the biggest winners in the region were also governing parties, such as the PSC in Ecuador, the MNR in Bolivia, and Alberto Fujimori's Cambio 90/Nueva Mayoría. This suggests that the real impact of incumbency was to raise the stakes, because governing parties were held most directly responsible for government performance. A second tendency was that among the parties that predated the initial election, the biggest winners were usually on the right or center-right.⁽⁸⁾ Only one of the six was on the left, and the one that is sometimes considered center-left--Peronism's Justicialista party--shifted decisively to the center or even center-right by the late 1980s under the leadership of Carlos Menem. Third, among the new parties, the biggest winners were all either personalist or center-left parties. The implications of these last two tendencies are cloudy until the parties' governing status is also taken into account; then it seems likely that there were three types of parties that tended to do especially well after 1982: personalist parties, governing parties to the right of center, and left-of-center parties in the opposition.⁽⁹⁾

All of these observations can be tested more rigorously with a larger sample of cases and data that are less aggregated. Such a test is reported below, and it confirms these ideas and adds some others. But first the theory must be fleshed out more.

The Amount and Type of Stress

Parties are pressed to adapt whenever their environment changes in ways that affect the voters' beliefs and priorities. Many aspects of the politically relevant environment have changed rapidly in Latin America, creating (to a different degree in each country) regime change, rapid urbanization, economic boom-and-bust cycles, rising drug trafficking and related crime, high and wildly varying inflation, deepening social inequalities, terrorism, guerrilla war, and economic liberalization. Any party would be sorely challenged to adapt well to any of these conditions, so no one should be surprised that the stressful environments in Latin America coincide with volatile party systems. On average, the average party-system volatility in Latin America is 29.3 in the 11 countries examined here, compared to an average of 8.6 percent in Western Europe from 1885 to 1985.⁽¹⁰⁾ Any one of the kinds of change listed above could be cited to account for the large volatility gap between these two regions.

One type of stress was particularly common and particularly severe in Latin America in the period after 1982--economic stress. The 1980s were such rough times for Latin America that they are referred to as "the lost decade." Only Colombia and Chile registered significant improvements in per capita GDP in this period; it fell in 5% in Venezuela, 8% in Argentina, 20% in Bolivia, and 27% in Peru.⁽¹¹⁾ As a result, between 1980 and 1992, inequality increased in Argentina, Bolivia, Brazil, Chile, Costa Rica, Mexico, and Venezuela; it decreased only in Colombia and Uruguay.⁽¹²⁾ Inflation was generally higher than levels that would be tolerated in Western Europe and the United States, and it reached the nightmarish rates of 2,938% in Brazil (1990), 3,087% in Argentina (1989), 7,482% in Peru (1990), and 11,749% in Bolivia (1985).⁽¹³⁾ A plausible working hypothesis, therefore, is that voters tended to reward parties that could claim credit for taming inflation, restoring economic growth, and improving the standard of living. Similarly, one may suppose that they tended to punish parties that made the economic situation worse in these respects.

Parties' Vulnerability to Stress and Capacity to Adapt

Voters do not treat all parties the same, however: in some countries voters are reluctant to question their party identification even if their party wrecks the economy or someone else's party produces a boom. In other countries, voters are far more generous in their rewards and more harsh in their punishments. The former type of party has a "solid" base of support; the latter, a "fluid" base. The strength of party identification therefore mediates the impact of economic performance on the vote. Students of Latin American parties consider identification with the major parties to be fairly strong in Argentina, Chile, Colombia, Costa Rica, Uruguay, and Venezuela, but fairly or extremely weak in Ecuador, Brazil, and Bolivia.⁽¹⁴⁾ However, the strength of identification can vary among parties in the same country, and can change over time.⁽¹⁵⁾

Whether a party's base of support is solid or fluid, the impact of economic performance also depends on whether the party is in the government or in the opposition. Incumbents may be credited or blamed for economic performance, but the opposition cannot. Therefore, we would expect governing parties that reduce inflation to grow, and those that make it worse, to decline. These effects should be greater for parties with a fluid base than for parties with a solid one; and the tendencies should be more clearly defined for governing parties than for opposition parties.⁽¹⁶⁾

Table 3 shows how the fates of selected governing and opposition parties varied, depending on the strength of identification, incumbency or opposition, and government success in fighting inflation. These estimates are based on regression analysis of 132 cases--the 23 most extreme cases reported in the table, plus 109 cases in which inflation rates were much lower. (For details, see the appendix to this chapter.) All of the expected relationships hold true, but only at the extremes experienced by the cases included in the table.

The tendencies observed are as follows: (1) Parties with a solid base of support found it electorally costly to govern and electorally beneficial to be in the opposition. (2) Economic performance had no significant impact on this tendency among parties with a solid base: governing parties were not more harshly punished if they raised inflation than if they lowered it, and opposition parties did not profit from worsening inflation more than they profited from price stabilization.⁽¹⁷⁾ For parties with a fluid base of support, however, the changes were much larger and both governing and economic performance mattered: (3) Opposition parties with a fluid base were hurt a little by falling inflation and helped a little by rising inflation, while (4) governing parties with a fluid base were helped a lot by falling inflation and hurt a lot by rising inflation. These relationships are summarized in Table 4. As Table 3 indicates, there are some notable exceptions, because elections are not simply referendums on economic performance. The explanation offered here merely traces some general tendencies that account for 38 percent of the variance; 62 percent must be explained by other factors. These factors could be characteristics of the competitive environment such as noneconomic issues or economic issues besides inflation; or other forms of adaptation, such as leadership, campaign styles and tactics, and alliances and boycotts.

A Note on Emerging Parties

As noted above, parties change by replacement as well as adaptation. If there are any commonalities among the emerging parties in Latin America it would be good to identify them. It cannot be said, however, that the new parties that emerged to completely or partially replace old major parties were necessarily to the right or to the left, or even that they tended to be personalistic. There were examples of all three. But two generalizations can be made.

First, emerging parties tended to be reactions against some major party that failed to adapt, and therefore they tended to be its opposite in some respects. Some characteristics of emerging parties differed greatly from country to country, therefore, depending on what sort of party they were replacing. La Causa R in Venezuela was opposed, on principle, to requiring its activists to toe any party line, in reaction to the iron discipline of AD and COPEI. But in Brazil, where most parties, and especially PMDB, were notoriously uncohesive, one of the most successful emerging parties was the PT, which achieved the tightest discipline of any party in the system. In Mexico both PAN and PRD were committed to political democracy and voluntary participation, in contrast to the mobilizational authoritarian techniques of the dominant PRI. Peru presents perhaps the most extreme case, in which Fujimori's Cambio 90/Nueva Mayoría reacted against the legacies of the previous AP and APRA governments by trying not to be a party at all.

Second, the new parties favored by voters are those that had a credible chance of winning, which in turn was a function of two qualities. One was experience in governing at the regional or local level. In Mexico, PAN was the opposition party that had won the largest number of local elections before 1988 and the largest number of governorships before 1994. In Venezuela, La Causa R first became nationally prominent through the governorship of its leader, Andrés Velásquez, in the eastern state of Bolívar. And in Brazil, the PT won quite a few municipal elections in major cities before its presidential candidate, Luis Inácio "Lula" da Silva, made it to the presidential runoff in 1989. The other token of credibility was being a splinter away from one of the old major parties. This probably gave an advantage to the PRD in Mexico (led by Porfirio Muñoz Ledo and Cuauhtémoc Cárdenas, both formerly of the PRI); to the PPD in Chile (led by Ricardo Lagos of the Socialist Party); to Convergencia Nacional in Venezuela (led by Rafael Caldera, the founder of COPEI); to the Frente Grande and FREPASO in Argentina (led by the Group of Eight and José Octavio Bordón, all ex-Peronists); and to the Partido Roldosista Ecuatoriano in Ecuador, which split away from the CFP in the early 1980s.

The rewards for economic stabilization in the post-debt-crisis environment explain why most of the biggest winners among established parties were on the right or center-right: fighting inflation was a natural part of their agenda, for which little adaptation was required. For the governing parties to the left of center, survival required a wrenching adaptation. For them, implementing stabilization and structural adjustment meant reversing many of the policies they had championed for years--expansion of the state sector, aggressive regulation of the private sector, extensive state intervention setting wages and prices--and postponing attempts to reduce poverty and inequality. Often it also meant recruiting economic advisers from pro-market institutes and turning a deaf ear to demands coming from unionized workers who traditionally had been a source of strong support.⁽¹⁸⁾ For these parties, the 1982-95 period was a Mephistophelian environment that presented them with a Faustian bargain: surrender your soul and you can live forever; otherwise, you will die. Most left-of-center parties either would not or could not keep such a bargain. Of the 10 elections held with center-left parties as incumbents, only 3 took place while inflation was falling. By contrast, inflation was falling during 53 percent of the elections with center incumbents, 73 percent with center-right incumbents, and 63 percent with incumbents on the right.

Among the qualitative changes in parties with which this volume is concerned--changes in recruitment, electoral mobilization, issue structuration, and societal representation--the most interesting examples were adaptations by center-left parties to the conservative environment. Case studies of three historically center-left parties will illustrate the dimensions of change and their consequences. Each was well-established, began with strong identifiers in the electorate, mobilized trade unions, and governed during part of this period. But these three parties responded to the challenges of the time differently. APRA in Peru misadapted-zigging to the left when it should have zagged right--and brought about its own destruction.⁽¹⁹⁾ Acción Democrática in Venezuela resisted adaptation, by withdrawing support from its own leaders who turned toward the market, and lost 40 percent of its voters.⁽²⁰⁾ The Justicialista party in Argentina, however, followed Carlos Menem to the right and won more votes, but underwent profound internal reorganization in the process.⁽²¹⁾

Before beginning, it should be noted that many characteristics of these three parties did not change significantly during this period. All three had long suffered from a poverty of practical policy ideas, which forced them to depend heavily on outside advisers for policy guidance in government. All three parties practiced very tight discipline in congress, and in Peru and Venezuela this discipline extended beyond the legislature. Other practices varied across the parties, but nevertheless remained constant over time. All mobilized voters with a mix of clientelism, mass meetings, and mass media, but better access to campaign funding enabled Acción Democrática (AD) to do all of these more intensively. Leadership was more personalized in APRA and the Partido Justicialista (PJ) than it was in AD, and procedures for recruitment and promotion were less institutionalized in the PJ. There are probably two reasons for the lack of change in these respects. The first is that every large party creates its own organizational subculture early in its existence, and this subculture tends to reproduce itself faithfully.⁽²²⁾ The second is that these parties had few incentives to adapt these aspects of their organizational life. Rather, the kind of adaptation required by the political environment after 1982 concerned the parties' ideological positioning and relations with labor and business. Therefore, whatever changes occurred in the ways parties recruited leaders, mobilized voters, and formed or sustained governments were the result of the replacement of old parties by new ones that performed these functions differently rather than the result of adaptation.

APRA: Misadaptation and Replacement

The Alianza Popular Revolucionaria de América (APRA) was founded by Víctor Raúl Haya de la Torre as an alternative to communism that would unite workers, peasants, and the middle class behind an anti-oligarchical program. From the 1930s to the mid-1960s it enjoyed a cult-like adherence among a substantial proportion of the urban middle class and workers in industry and export agriculture.⁽²³⁾ By the mid-1960s it exercised paternalistic control over 75 percent of the trade unions in Peru. It also informally sponsored paramilitary squads called búfalos that took violent action against rivals. At its peak, the party had an extensive grass-roots organization that involved activists in a wide range of activities, including rallies, lectures, cooperatives, and soccer clubs. In 1945-48 and 1956-58, however, APRA briefly backed two relatively conservative governments; and during the 1963-68 presidency of Fernando Belaunde, Haya joined the conservative former dictator Manuel Odría to block attempts at land reform. In the increasingly pro-reform context of the time, such actions eroded APRA's support base, especially among organized labor.

The transition to democracy after 1978 coincided with three challenges to APRA in addition to the debt crisis, which hit Peru early. One was the radicalization of many voters by the leftist military government of General Juan Velasco Alvarado (1968-75), which had nationalized major industries, set up agricultural cooperatives, established close ties to Cuba, and placed former Socialist and Communist politicians in charge of some urban social programs.⁽²⁴⁾ During the 1970s and afterwards most unions affiliated with the leftist Confederación General de Trabajadores Peruanos (CGTP) rather than APRA's Confederación de Trabajadores Peruanos (CTP). The youth wing, Juventud Aprista Peruana (JAP), was even more radicalized, to the point of being openly and explicitly Marxist. A second challenge was the need to renovate the leadership, as most of the well-known leaders were quite old and discredited by Haya's past compromises. Finally, the most serious challenge of all was the death of Haya himself in 1979. Although he had taken on several leaders in their 20s and 30s as proteges, he left neither a clear successor nor a means for choosing one, so the party languished in a leadership vacuum for the next 4 years.⁽²⁵⁾

The leader who eventually emerged, as the winner of the presidential candidacy nomination in late 1983, was 34-year-old Alan García. Before long activists treated him as the unquestioned leader much as they had treated Haya, and it fell to him to guide APRA's adaptation to the post-1982 political environment. At first he planted the party firmly in the center, where it was abundantly rewarded. Then he unexpectedly swerved to the left, which cost the party dearly.

García's 1983-85 presidential campaign was designed to reshape APRA in order to attract votes outside its traditional base. One promise was targeted at radicalized workers and students: an announcement that Peru would use no more than 20 percent of its export earnings for servicing the foreign debt. (He lowered this figure to 10 percent on inauguration day.) This promise was also appealing to many voters who were not otherwise very radical, however, and most of his other appeals were directed toward less radical groups. He encouraged the formation of middle-class "Civic Communities" to endorse his candidacy; he initiated consensus-building talks with the armed forces and the Church; he publicized the most technical policy studies Apristas had to offer (which were still not as detailed or rigorous as those of their rivals); he criticized unionized workers for being a labor elite less deserving of state support than the un- and under-employed; he repudiated some of Haya's anticapitalist utterings; he distanced himself from the party's violent past; and he advocated a "social pact" with business.⁽²⁶⁾ The overall effect was to cast APRA in the role of a pragmatic, reasonable, moderate reformist alternative to the growing left. This strategy led to the party's greatest electoral success ever: García won 47.8 percent of the vote in 1985, compared to less than 36 percent in all previous elections in which the party had run its own candidate.

At first García was an extremely popular president, with approval ratings over 80 percent in his first year and over 60 percent in his second.⁽²⁷⁾ But in July 1987 he suddenly shifted to the left by nationalizing all remaining private domestic banks, insurance companies, and finance corporations. This move instantly alienated the international financial community (which already distrusted García's heterodox policies), Peruvian business leaders, opposition leaders on the right, and even most leaders of APRA--none of whom had been consulted before the decision was made--without winning the support of the parties on the left.⁽²⁸⁾ The economy shuddered, then collapsed, and from this point on APRA suffered an unbroken series of disasters: García's approval rating immediately plummeted to 30 percent; within months, a rival wrested from him control of the APRA congressional delegation; some CTP unions joined general strikes in 1988 and 1989; JAP youth began defecting to the Movimiento Revolucionario Tupac Amaru (MRTA, the same terrorist group that held hostages in the Japanese embassy for 4 months in 1997); and the party's share of the vote shrank to 25 percent in 1990. In 1995, further discredited by Alberto Fujimori's economic success and his virulently anti-party rhetoric, APRA's share shrank to 6.5 percent--yielding just 6.7 percent of the seats in congress and 4 percent of the presidential vote--not enough to maintain its registration as a national political party. García himself went into exile in Colombia to escape corruption charges.⁽²⁹⁾

Peru is a clear-cut case of a change in the nature of parties due to replacement rather than adaptation. The major party in Peru after 1990 was Cambio 90/Nueva Mayoría, the personalist vehicle of Alberto Fujimori, which controlled 67 of the 120 seats in congress as a result of the 1995 election. Its only ideology was to back whatever Fujimori wanted to do, without question. It had no grass-roots presence. It screened candidates in focus groups and took polls to set priorities. It went farther in the direction of the unmediated electronic executive than any other party in the region, and perhaps the world. Conaghan has argued that "Peru's party system has ceased to exist in any meaningful sense."⁽³⁰⁾

Acción Democrática: Resistance and Decline

Acción Democrática (AD), like APRA, was founded to bring together the middle class, workers, and peasants into a nationalistic anti-oligarchical alliance. Unlike APRA, it was very successful at winning power. After seizing power jointly with military conspirators in 1945, it won Venezuela's first fair presidential election in 1948 with more than 70 percent of the vote, and won 5 of the 8 presidential elections after the restoration of democracy in 1958. Although AD was fairly radical for its time in the beginning, by 1958 it had become a slightly center-left catchall party. It won the votes of Venezuelans from all classes, occupations, and regions, and in this respect was indistinguishable from its principal rival, the Social Christian party COPEI. Unlike many catchall parties, however, AD had a Leninist organizational structure explicitly based on democratic centralism and actively enforced party discipline among its legislators, leaders, and militants at the national, state, and local levels. As late as 1995 it was still expelling hundreds of members for ignoring the party line in local elections.

In addition to being tightly disciplined, AD aggressively penetrated most organizations in civil society aside from the Church and private businesses.⁽³¹⁾ Through infiltration, cooptation, and the creation of parallel organizations, AD and the other major parties succeeded in placing party members as leaders of student governments, professional associations, and unions. These leaders then mobilized their organizations in support of their party during and between elections. AD was much more successful than other parties at gaining control over labor unions and federations, and in winning the most important offices at the head of the peak labor confederation, the Confederación de Trabajadores de Venezuela (CTV). The labor wing of AD was therefore a large and valuable component of the party. However, AD's labor leaders habitually deferred to the party line in exchange for party support for labor demands in the long run. And to the extent that the labor wing had influence over party affairs, it was counterbalanced by the influence of the state party bosses, who used their regional patronage machines to deliver large blocs of votes at national party meetings. In general elections, both state bosses and labor leaders turned out large numbers of activists for open-air rallies, caravans, and processions, all festooned with party posters, banners, T-shirts, hats, and other paraphernalia. Since about 1973, parties have also made heavy use of high-tech polling, campaign consultants, and slick television ads, but not at the expense of this old-fashioned type of electoral mobilization.

During the long intervals between major national party meetings, however, a much smaller group of leaders made party decisions. Formally, that body was the National Executive Committee (CEN); informally, it was a group of 5-7 leaders known as the cogollo. And when the president came from AD, the cogollo normally did its best to reach agreement with the president privately so that president and party could appear united in public. When the national leadership was united, the cogollo had the power to dictate the party line to the Parliamentary Fraction and state and local party leaders. But when the national leadership was divided, as often occurred during the selection of the presidential candidate, de facto leadership reverted to state party bosses and labor leaders.

When the debt crisis hit Venezuela, AD resisted adaptation every step of the way. One president temporized with heterodoxy with the full support of the party; the next veered toward neoliberalism, but the party blocked his way; and when the next presidential candidate also endorsed the turn to the market, he and his supporters were marginalized, leaving the organization in the hands of an extremely pragmatic general secretary who supported minimal liberalizations only reluctantly and semi-publicly. Jaime Lusinchi (1984-89) followed heterodox policies throughout his term, so there was not much of an attempt at adaptation for the party to resist. Lusinchi's policies were supported by a majority of the CEN, the Parliamentary Fraction, and the labor wing.⁽³²⁾ Lusinchi's successor was Carlos Andrés Pérez, also from AD, who surprised everyone by announcing a shock economic liberalization package at his inauguration. Due to the concentration of policymaking authority in the executive branch in Venezuela, he managed to implement the easier parts of his program, but he was a very unpopular president throughout his term. As time passed, AD became less willing to support further reform; and after the two unsuccessful coup attempts in 1992, the party left its president without support and Pérez's economic liberalization stalled. A minister in the Pérez cabinet wrote that "Pérez's own party, Acción Democrática, having spent most of the 1980s profiting from the many opportunities to serve as broker between society and the state, adamantly opposed any changes resulting in reduced government intervention."⁽³³⁾ AD leaders were not happy when Pérez was impeached in 1993; but they expelled him from the party anyway while he awaited sentencing on corruption charges.

It could be argued that AD adapted in other ways. In 1988 and 1989, for example, the party voted for two electoral reforms that provided for the direct election of governors and mayors and the election of half of the national deputies in single-member districts. These reforms were designed to make public officials more responsible to their own constituents and less responsible to national party leaders. It may appear that AD was adapting to the environment either by lessening partidocracia (the distortion of democracy by excessively strong parties) or by creating a diversion away from the economic situation. In reality, the party per se was not enthusiastic about political reform, either. The electoral reform was pushed by Pérez during the campaign itself, which was the time of his greatest influence over the party; as soon as the law went into effect, other party leaders took steps to minimize its decentralizing effects by, for example, exerting tight control over nominations for governor, mayor, and national deputy. When grass-roots demands arose for further constitutional reform in 1992, AD took the lead in delaying, watering down, and ultimately shelving the major reform bill. In spite of these efforts to resist adaptation, several governors or mayors with a regional base of support began to challenge the national leadership. One of these, Claudio Fermín, won the presidential nomination in a primary. He was very much the candidate of renovation, with calls for thorough economic liberalization and more openness and participation within the parties. But in the 1993 election, Fermín won only 23.6 percent of the vote, the worst performance ever for an AD candidate. After the election, hundreds of his supporters were purged from the party and Fermín himself was treated so coldly by other party leaders that he took refuge in the United States for several years.

Although AD was not hurt as badly as APRA, both it and COPEI lost considerable support. From 1973 to 1988 they never won less than a combined 74 percent of the legislative vote; in 1993, they won only 46 percent, and in 1998, 36 percent. In presidential races, they customarily shared 90 percent of the vote until 1988; but in 1993 their combined share was 54 percent. In 1998, COPEI backed an independent candidate rather than attempt to win on its own, and at the last minute both parties threw their support to a different independent and together contributed only 11 percent to his vote total. In 1993 two parties filled the breach.

The largest new parties that have filled the vacuum left by AD and COPEI are all either personalist or center-left, and all are in some ways reactions against the traditional parties. In 1993 one was Convergencia Nacional, the personalist vehicle of Rafael Caldera, who abandoned COPEI and won the presidential contest. The other that year was La Causa R, a center-left party that explicitly campaigned against the AD-COPEI "establishment." It was committed to being the opposite of Acción Democrática in several ways: it was responsive to the union rank-and-file rather than cooptative and controlling; pluralistic rather than disciplined; and respectful of the autonomy of new social movements.⁽³⁴⁾ In 1998 independents dominated the field. For much of the long campaign, the frontrunner was Irene Sáez Conde, a former Miss Universe who earned a reputation as an efficient and honest mayor. Her star fell, however, when she accepted COPEI's endorsement, which voters interpreted as "the kiss of death." Much of her support then went to Henrique Salas Römer, a state governor who also had a reputation for efficiency and honesty. But the big winner in 1998 was the candidate who combined all of these characteristics. Hugo Chávez Frías was the charismatic leader of a personalistic movement, the Movimiento V República (Fifth Republic Movement); he employed populist rhetoric, even if he was not certifiably on the center-left, and was allied with the center-left Movement Toward Socialism (MAS) and the Causa R splinter Patria Para Todos; and he vowed to wipe out corruption, which he associated with the traditional parties. Furthermore, it was clear that he stood for radical change, as he was the leader of a nearly successful coup attempt in February 1992. No one could have personified better the tendency of new parties to be reactions against the old ones.

The Peronists: Adaptation and Growth

Even more than APRA or AD, the Justicialist party (PJ) in Argentina depended organizationally on the support of trade unions, monopolized union support, and gave the unions a prominent leadership role in the party. Before the late 1980s the unions were customarily allocated one third of the positions on the National Council; 1983 presidential candidate Italo Luder was backed strongly, if customarily, by the unions; 35 of the 115 PJ deputies were associated with the unions; and the head of the metalworkers union, Lorenzo Miguel, served as party president in 1983-84.⁽³⁵⁾ The PJ was not exclusively a labor party, however, for three reasons. First, there was a strong element of personalism toward Perón before his death in 1974, and after 1987 to Carlos Menem. (In the intervening years the party struggled to produce a new unifying leader.) Second, because of its apparently all-encompassing diversity, the PJ for decades claimed not even to be a party, which would represent only a part of society, but a movement of the entire Argentine pueblo.⁽³⁶⁾ And third, the party enjoyed vibrant grass-roots organization structured around neighborhoods and provinces, not just sectoral organizations such as unions. Because identification with Peronism was so strong, there were self-mobilizing, unofficial unidades de base (base units) in practically every neighborhood, often more in working-class neighborhoods; these were grouped together into local clientelistic factions called agrupaciones by brokers called punteros. The agrupaciones fed informally into the provincial factions that constituted the provincial parties, and the national party encompassed most of the provincial Peronist parties. Nevertheless, while not strictly a labor party, the PJ was more of a working-class party than AD or APRA, as there was a very significant differentiation of Peronist and anti-Peronist vote by class.

In spite of the strong working-class base of support, it is less clear that the PJ possessed a center-left ideology. Its prominent leaders spanned the entire left-right spectrum, from the Montoneros on the extreme left, the Peronist Youth of the 1970s on the left, to Herminio Iglesias, López Rega, and the Guardia de Hierro ("Iron Guard") on the right or far right.⁽³⁷⁾ Moreover, the corporatist labor interests that were dominant in the party before 1985 were often labeled "conservative." However, this conservatism consisted of nationalism, clericalism, and sympathy for authoritarianism rather than support for economic liberalization. Before 1989, most of the party's prominent leaders and fractions shared an opposition to many of the pro-market reforms that were being prescribed for Argentina and other Latin American countries emerging from the debt crisis. Privatization came in for especially vehement opposition due to its perceived negative impact on the Peronist unions. Even the anti-union "Renewalist" faction that won control of the party in 1985 was on the center-left in economic terms; its attack on the unions was directed only at the unions' power within the party organization and parliamentary caucus.

Menem himself adapted to the 1982-1995 environment by shifting to the right. In 1985-1987, as governor of La Rioja province, he had backed the "Renewalist" faction⁽³⁸⁾ led by Antonio Cafiero that called for internal democratization of the party, institutionalization of the party organization, and a return to what it viewed as the party's ideological heritage on the center-left. Once the corporatist union leadership was defeated by the Renewalists, however, Menem broke with Cafiero to launch his own candidacy for the presidential nomination. During the campaign, his economic program was vague, and many feared (or hoped) that he would turn out to be a populist. Once in office, however, his shift to the right became clear. With the pro-business Unión del Centro Democrático as a coalition partner, and with the help of ministers recruited from prominent business groups and think tanks, Menem aggressively decontrolled prices, liberalized trade, sold off state enterprises, and cut the budget. By 1993 inflation was down to 10.6 percent and still falling, while investment poured in and growth returned.

What is amazing is that Menem managed to keep the support of most of his party while carrying out this shift to the right. It helped considerably that the marginalization of the old "orthodox" labor wing had already been accomplished by the Renewalists. Led by Antonio Cafiero, this faction amended the party charter to reduce labor representation in the National Council from one-third to 17 positions out of 110 (15 percent). It also promoted primaries for the selection of legislative candidates, which reduced the number of labor legislators from 35 to 6. Even more importantly, the Renewalists sidelined all of the old labor representatives. Ever since the 1960s, an informal labor confederation known as the "62 organizations" had been the de facto representative of labor within Peronism. Cafiero simply refused to recognize this body, choosing to deal instead with a more cooperative set of labor leaders, although without granting them a formal role in the party. Menem encountered union opposition, especially from public employees opposed to privatization, but managed to hold onto most union and non-union support within the party. For example, the peak labor confederation CGT organized only one general strike during Menem's first term, compared to 13 during the 1983-89 government of Raúl Alfonsín. Some union leaders were brutally repressed; others were coopted; a few made their peace with market capitalism; and still others grumbled but stayed within the party.⁽³⁹⁾ But almost all union leaders backed Menem for reelection in 1995.

Non-union opposition within the party was also skillfully marginalized. Due to the weak institutionalization of the party organization, Menem had little trouble in arranging for the lateral entry of non-politicians into party leadership positions, displacing politicians whose entire careers had been spent in Peronism.⁽⁴⁰⁾ His one setback was the departure of the "Group of Eight" deputies, who defected in protest against Menem's shift to the right and his authoritarian style. Overall, however, this defection was not very costly, because Menem maintained the PJ's share of the vote despite the defection of the Group of Eight and Governor José Octavio Bordón, who ran as the presidential candidate of the Frente Para un País Solidario (FREPASO) and won a startling 29 percent of the vote. Such was the success of Menem's adaptation that even though Bordón had been a Peronist, his party drew votes away from other parties, principally the Radicals.⁽⁴¹⁾

Latin America's parties and party systems are too diverse and dynamic to provide a simple answer to questions about the decline of parties.⁽⁴²⁾ They have changed in the past two decades, but then, they have always been changing; some parties have moved away from the mass-party model, but they were not that common in the region to begin with; and some (Costa Rica, Uruguay) have not changed very much.

However, we can identify certain tendencies in the nature of this change. First, the nature of individual party organizations-centralization, discipline, cohesion, recruitment, mobilization, socialization, financing-seems to change very slowly, if at all. In the meantime, it is more likely that the party will be sidelined in the volatile electoral environment. Therefore the primary mechanism of change in parties is replacement rather than internal reform. Second, this evolutionary process, akin to Darwin's principle of natural selection, tends to favor the survival of parties that are well adapted to the political environment. Parties are not the passive objects of the process, as they possess some capacity to adapt. Those that adapt well survive; those that stubbornly refuse to adapt, or misadapt, lose votes and move toward extinction. Third, in the Lost Decade of approximately 1982 to 1995, the environment selected in favor of governing parties of the center-right or right and opposition parties that were either personalist or left of center. But the environment has probably changed already, so we can expect to see different sorts of parties favored in future elections.

Evaluating the consequences of these changes is a separate question, and its answers are not simple, either. Changes that improve the quality of representation sometimes weaken governability, and vice versa. Whether or not a change takes a party system into or away from a zone of functionality (defined by a happy medium of fragmentation, polarization, and other characteristics) depends on the starting point and the nature, magnitude, and direction of the change. In Latin America, the party system became less functional for representation and governability in Brazil and Ecuador but more functional in both respects in Colombia and Mexico. It became less functional for representation in Peru and possibly Argentina, and less governable in Venezuela (Figure 1 again). In the other four major countries there was either no significant change or a gain in one respect that was offset by a loss in another. This summary assessment is not a cause for celebration, but it should temper the much more negative impression left by a handful of dramatic electoral upsets (by Collor in Brazil, Fujimori in Peru, ADM-19 in Colombia, Caldera and Chávez in Venezuela) that were the focus of disproportionate media coverage and scholarly discussion. In view of the grim economic environment, we should have expected party-system change to be much worse. Perhaps a quiet, sober celebration would be in order after all.

The dependent variable for this model is the change in the percentage of the vote won by a party from one legislative election to the next. The sample consists of 132 such changes for parties that had been, were, or would become major parties in 11 Latin American countries from 1978 to 1995. The sample therefore includes all the presidential parties and all major opposition parties in this period in Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Peru, Uruguay, and Venezuela.

Two of the independent variables are control variables. The first is simply the percentage of the vote won by the party in the previous election. This is included because the percentage of the vote that is potentially available to add to a party's share is 100 minus its previous share. Parties that are already large may shrink easily, but they find it difficult to grow beyond a certain point. Small parties have a greater potential for growth, but they can lose only so many votes before they are eliminated. The negative coefficient of the lagged vote correctly specifies these different constraints on large and small parties. The second control variable is the percentage of the vote won previously by parties that merged with the party of interest (a positive number) and the percentage of the vote currently won by parties that split away from it (a negative number). When these splits and mergers occur (only 10 instances are included in this sample), they have an obvious direct impact on vote shares. Controlling for them makes it possible to estimate the impact of economic performance with less bias.

The only indicator of economic performance used here is the change in ln(inflation) from the last year of the previous government to the last year of the current government. Lower inflation is represented by positive numbers and higher inflation by negative numbers. The model reported below contains interactions between economic performance as measured in this way, on the one hand, and incumbency and the strength of party identification, on the other. (See note 10 on party ID.) ln(inflation)_{incumbents with fluid base} represents the change in logged inflation for all incumbent parties with a fluid base of support. ln(rising inflation)_{opposition with fluid base} applies to all opposition parties with a fluid base, but only if inflation was rising. The model also includes a dummy variable for all parties classified as having a fluid base of support. The OLS regression estimates are:
 

vote = 11.03 -0.32*lagged vote -7.30*weak ID +0.46*splits and mergers

(2.00)** (0.05)** (1.74)** (0.19)*
 

+3.32*ln(inflation)_{incumbent with fluid base} -2.14*ln(rising inflation)_{opposition with fluid base}

(0.76)** (0.90)*
 

N=132, Adjusted R²=.381

Standard errors are in parentheses. *=significant at the .02 level; **=significant at the .0001 level

Source: author's data.
 

SOURCE: Miguel Basáñez, Marta Lagos, and Tatiana Beltrán, Reporte 1995: Encuesta Latino Barómetro: Opiniones y actitudes en Latinoamérica; Economía, sociedad, política y asuntos internacionales (May 1996), pp. 91, 103, 50, and 51.

1. I define parties simply as organizations that compete in legislative elections. This definition necessarily includes many ephemeral organizations that managed to get on the ballot, and excludes some political organizations that were never willing or able to compete in elections.

2. The volatility index is the sum of the absolute values of the differences the shares won by all parties from one election to the next, halved to adjust for double-counting. This index was first defined in Mogens Pedersen, "The Dynamics of European Party Systems: Changing Patterns of Electoral Volatility," European Journal of Political Research 7:1 (1979): 1-26. Although volatility is normally calculated for pairs of consecutive elections, it may apply to any pair of elections. The index reported in table 1 applies to the first and last elections in the periods listed. This method better reflects cumulative change over a long span of time because it measures all the changes that last, and only changes that last. The more commonly reported average volatility in consecutive elections counts changes that are subsequently undone. For example, if two parties' shares in 4 elections were 65-35, 45-55, 55-45, and 60-40, the average volatility would be (20+10+5)/3 = 11.7, while cumulative volatility would be only 5.0, better reflecting a system that is very similar at the beginning and the end of the period.

3. Arend Lijphart, Democracies (New Haven: Yale University Press, 1984) pp. 106-114.

4. The effective number of parties is the reciprocal of the sum of squared party shares. Its value is 2.0 for a 50-50 party system and 4.0 for the 25-25-25-25 party system. When parties are unequal in size, it usually takes on fractional values. This indicator was originally described in Murkuu Laakso and Rein Taagepera, "Effective Number of Parties: A Measure with application to Western Europe," Comparative Political Studies 12 (1979): 3-27.

5. The relative center can be farther to the right or the left than the absolute center as defined in the classification criteria, and is operationalized here as the Mean Left-Right Position of all parties in the system (MLRP). MLRP measures the how far to the left or the right the average party was in each election, based on the left-right positions of all the parties and their shares of the vote. This indicator assumes that all parties classified left (whether Christian or secular) are approximately twice as far from the center as parties classified center-left; and right parties are twice as far to the right as the parties of the center-right. This assumption permits the calculation of MLRP as:
 

The formula for left-right polarization is therefore:

where mlrp = MLRP/100. The index can reach its maximum only when half of the vote goes to the right and half to the left; if all of the vote went to just one extreme, polarization would be zero because the relative center would be at the extreme as well and there would be no dispersion. It is important to remember that this is an indicator of left-right polarization only, and does not reflect the intense personal, ethnic, ins-outs, or other rivalries that sometimes exist between parties that are relatively close in left-right terms.

6. See note 31 for elaboration.

7. This conception of party-system evolution as the product of both the environment and choices by leaders was first developed at length in Seymour Martin Lipset and Stein Rokkan, "Cleavage Structures, Party Systems, and Voter Alignments," in Lipset and Rokkan, eds., Party Systems and Voter Alignments: Cross-National Perspectives, pp. 1-64 (New York: The Free Press, 1967). It has been further developed more recently in Angelo Panebianco, Political Parties: Organisation and Power (Cambridge: Cambridge UP, 1988) and Herbert Kitschelt, The Transformation of European Social Democracy (Cambridge: Cambridge UP, 1994).

8. All left-right ideological labels applied to parties in this chapter come from a comprehensive classification of 97 percent of the parties in the major countries of Latin America. I drafted a classification myself, obtained feedback from 53 country specialists, then did my best to reconcile their suggested corrections consistently over time and across countries. This project is fully documented in Michael Coppedge, "A Classification of Latin American Political Parties," Kellogg Institute Working Paper No. 244 (November 1997).

9. Methodological sophisticates will object that I have selected on the dependent variable for all the arguments in this paragraph. They are correct, but I have done so only to generate hypotheses that are tested with a larger and far more appropriate sample below.

10. Michael Coppedge, "Freezing in the Tropics: Explaining Party-System Volatility in Latin America," paper presented at the 1995 Meeting of the Midwest Political Science Association, Chicago, April 6-8; and Stefano Bartolini and Peter Mair, Identity, Competition, and Electoral Availability: The Stabilization of European Electorates, 1885-1985 (Cambridge: Cambridge UP, 1990). The figures reported are both average, not cumulative, volatility, using the same counting rules.

11. For Peru and Ecuador, IMF, International Financial Statistics 1992; for other countries, United Nations Economic Commission for Latin America and the Caribbean, Social Panorama of Latin America 1994 (Santiago, Chile: ECLAC, 1994), tables 1 and 19. Dates are 1980-92 in Argentina, Bolivia, Chile, and Colombia; 1981-92 in Costa Rica, Uruguay, and Venezuela; 1979-82 in Brazil; 1980-90 in Peru, 1979-90 in Ecuador, and 1984-92 in Mexico.

12. United Nations Economic Commission for Latin America and the Caribbean, Social Panorama of Latin America 1994 (Santiago, Chile: ECLAC, 1994), pp. 35-45. Comparable data were unavailable for Ecuador and Peru.

13. Peak inflation 1980-85: Anuario estadístico de América Latina y el Caribe 1988 (Santiago, Chile: CEPAL, 1988), p. 94-95; Peak inflation 1986-95: Inter-American Development Bank, "Average Annual Growth of Consumer Prices" <http:\\iadb6000.iadb.org/int_data/pri2.html>.

14. No cross-nationally comparable data on the strength of party identification in Latin America yet exist. One study tracks the stability of self-reported party preferences in Argentina, Chile, Mexico, and Venezuela, but the survey questions are not strictly comparable and cover only four cases. (Lauro Mercado Gasca, "Visiting Party Loyalties in Latin America," paper prepared for the XX International Congress of the Latin American Studies Association, Guadalajara, Mexico, April 17-19, 1997.) Comparable instruments were used in the 1995 and 1996 Latinobarómetro surveys in eight countries, but a one-year span is not long enough to justify inferences about the stability of party loyalties, and the data have not been shared with the scholarly community. An independent assessment based on the best data available is Scott Mainwaring and Timothy Scully, "Introduction: Party Systems in Latin America," in Mainwaring and Scully, eds., Building Democratic Institutions: Party Systems in Latin America (Stanford: Stanford UP, 1995), pp. 17-21. The coding of my data coincides with Mainwaring and Scully's conclusions.

15. Here PRI and PAN in Mexico, PT in Brazil, and the UCR and PJ in Argentina are treated as having a solid base of support, while PRD in Mexico, PMDB and PFL in Brazil, and Frente Grande and FREPASO in Argentina are treated as having a fluid base. In Peru identification with APRA was strong before 1968 but weak afterwards.

16. Robert Dix, "Incumbency and Electoral Turnover in Latin America," Journal of Interamerican Studies and World Affairs 26 (1984): 435-48; George A. Quattrone and Amos Tversky, "Contrasting Rational and Psychological Analyses of Political Choice," American Political Science Review 82:3 (1988): 719-36; Martin Paldam, "How Robust Is the Vote Function? A Study of Seventeen Nations over Four Decades," in Helmut Norputh, Michael Lewis-Beck, and Jean Dominique Lafay, eds., Economics and Politics: The Calculus of Support (London: Sage, 1991).

17. There is no way to know whether the insignificance of economic performance for solid parties is due to their strong ID or to the fact that none of them happened to experience inflation changes of 3,000 percent or greater. Also, it is possible that the association between strong ID and inflation is no accident, as solid governing parties may be less likely to allow inflation to surge to such extraordinarily high levels.

18. Catherine Conaghan, James Malloy, and Luis Abugattas, "Business and the 'Boys': The Politics of Neoliberalism in the Central Andes," Latin American Research Review 25 (1990): 3-30.

19. Other cases include Izquierda Democrática and the CFP in Ecuador and PMDB in Brazil.

20. Similar cases of resistance and decline are the Radicals in Argentina, Acción Popular in Peru, and the Blancos and Colorados in Uruguay.

21. Other parties that adapted and grew were the Christian Democrats and the Socialists in Chile. More conservative parties prospered electorally without having to adapt as drastically. Examples are the PUSC in Costa Rica, the Social Christian Party in Ecuador, and MNR in Bolivia which, despite having led the 1952 Revolution, was already right of center by 1985.

22. Panebianco, Political Parties, chapter 4.

23. Grant Hilliker, The Politics of Reform in Peru: The Aprista and Other Mass Parties of Latin America (Baltimore: Johns Hopkins UP, 1971), pp. 74-82.

24. The best study of the impact of this government on voters is Susan C. Stokes, Cultures in Conflict: Social Movements and the State in Peru (Berkeley: U California P, 1995).

25. Carol Graham, Peru's APRA: Parties, Politics, and the Elusive Quest for Democracy (Boulder, Colo.: Lynne Rienner, 1992).

26. All of this evidence comes from Graham, Peru's APRA, pp. 84-91.

27. Martín Tanaka, "Los espejos y espejismos de la democracia y el colapso de un sistema de partidos: Perú, 1980-1995, en perspectiva comparada," paper prepared for delivery at the 1997 meeting of the Latin American Studies Association, Guadalajara, Mexico, April 17-19, 1997, p. 26.

28. Graham, Peru's APRA, pp. 111-21.

29. In 1995 Fujimori raised the signature requirements for registering a national party from 100,000 to about 490,000 signatures, which the remnant of APRA failed to meet. It nevertheless ran many candidates as independents in the November 1995 mayoral races, and privately claimed to have elected about one thousand. Private communication from Charles Kenney, May 29, 1997.

30. Catherine Conaghan, "The Irrelevant Right: Alberto Fujimori and the New Politics of Pragmatic Peru," paper prepared for the conference on "Conservative Parties, Democratization, and Neoliberalism in Latin America: Mexico in Comparative Perspective," Center for U.S.-Mexican Studies, University of California, San Diego, May 31-June 1, 1996.

31. Michael Coppedge, Strong Parties and Lame Ducks: Presidential Partyarchy and Factionalism in Venezuela (Stanford: Stanford UP, 1994), chapter 2.

32. Lusinchi was not able to impose his choice for the presidential nomination, but this was because most state bosses and national leaders considered the alternative more electable, not because of any disapproval of Lusinchi's policies. Coppedge, Strong Parties and Lame Ducks, chapters 4 and 6.

33. Moisés Naím, Paper Tigers and Minotaurs: The Politics of Venezuela's Economic Reforms (Washington, D.C.: Carnegie Endowment, 1993, p. 47.

34. Margarita López Maya, "El ascenso en Venezuela de la Causa R," paper presented at the XVIII International Congress of the Latin American Studies Association, Atlanta, 9-13 March 1994.

35. I am deeply indebted to Steven Levitsky for his detailed descriptions and insights into Justicialist party organization, as analyzed in Steven Levitsky, "Crisis, Party Adaptation, and Regime Stability in Argentina: The Case of Peronism, 1989-1995," paper prepared for delivery at the 1997 meeting of the Latin American Studies Association, Guadalajara, Mexico, April 17-19, 1997.

36. James McGuire, "Political Parties and Democracy in Argentina," in Mainwaring and Scully, Building Democratic Institutions, pp. 200-46.

37. I am also deeply indebted to Pierre Ostiguy for his "Peronism and Anti-Peronism: Social-Cultural Bases of Political Identity in Argentina," paper prepared for delivery at the 1997 meeting of the Latin American Studies Association, Guadalajara, Mexico, April 17-19, 1997. Ostiguy argues that the diversity within the Peronist and anti-Peronist blocs obscures the fact that they are consistently divided by what he calls the "high-low" cleavage. Those at the high end of this divide are "refined," cosmopolitan, and inclined to favor formal procedures; those at the low end are "crude," localist, and personalist or caudillista. It is a class cleavage defined in socio-cultural rather than economic terms. In this scheme the Peronists, Montoneros, and Carapintadas are "low" and the Radicals, Socialists, UCeDé, and FREPASO are "high."

38. I am again heavily indebted to Levitsky, "Crisis, Party Adaptation, and Regime Stability in Argentina," for these arguments.

39. Steven Levitsky, "Populism is Dead! Long Live the Populist Party! Party Adaptation Through Coalitional Realignment in Peronist Argentina," paper prepared for delivery at the 1995 meeting of the Latin American Studies Association, Washington, DC, September 28-30, 1995, pp. 20-24.

40. Levitsky, "Crisis, Party Adaptation, and Regime Stability in Argentina."

41. Levitsky, op. cit., pp. 4-5.

42. Michael Coppedge, "The Dynamic Diversity of Latin American Party Systems," Party Politics 4:4 (1998): 547-68.