American Political Science Review Page 1 of 22 August 2012 doi:10.1017/S0003055412000287 Inequality and Regime Change: Democratic Transitions and the Stability of Democratic Rule STEPHAN HAGGARD University of California at San Diego ROBERT R. KAUFMAN Rutgers University ecent work by Carles Boix and Daron Acemoglu and James Robinson has focused on the role of inequality and distributive conflict in transitions to and from democratic rule. We assess these R claims through causal process observation, using an original qualitative dataset on democratic transitions and reversions during the “third wave” from 1980 to 2000. We show that distributive conflict, a key causal mechanism in these theories, is present in just over half of all transition cases. Against theoretical expectations, a substantial number of these transitions occur in countries with high levels of inequality. Less than a third of all reversions are driven by distributive conflicts between elites and masses. We suggest a variety of alternative causal pathways to both transitions and reversions.
re inequality and distributive conflicts a driving wider the income disparities in society, the more elites force in the transition to democratic rule? Are have to fear from the transition to democratic rule and Aunequal democracies more likely to revert to the greater the incentives to repress challenges from authoritarianism? These questions have a long pedi- below. Given this potential indeterminacy, theoretical gree in in the analysis of the transition to democratic models have hinged on a variety of other parameters, rule in Europe (Lipset 1960; Marshall 1963; Moore such as the cost of repression or the mobility of assets. 1966), and have been raised again in newer compar- Even with these refinements, attempts to demon- ative historical work on democratization (Collier 1999; strate the relationship between inequality and regime Rueschemeyer, Stephens, and Stephens 1992). More type have yielded only mixed results. In cross-section, recently, an influential line of theory has attempted there is a relationship between income distribution and to ground the politics of inequality on rationalist as- the level of democracy: Ceteris paribus, more equal so- sumptions about citizens’ preferences over institutions cieties are more democratic. Yet the causal relationship (Acemoglu and Robinson 2000; 2001; 2006; Boix 2003; between inequality and either transitions to democratic 2008; Przeworski 2009). These distributive conflict ap- rule or reversions from it is much less robust. proaches conceptualize authoritarian rule as an insti- We focus on regime change during the “third wave” tutional means through which unequal class or group of democratic transitions from 1980–2000. This period relations are sustained by limiting the franchise and the was marked by the spread of democracy to a wide range ability of social groups to organize. The rise and fall of of developing and postsocialist countries. These in- democratic rule thus reflect deeper conflicts between cluded not only middle-income nations in Latin Amer- elites and masses over the distribution of wealth and ica, Eastern Europe, and East and Southeast Asia but income. also a substantial number of lower income countries, in- Despite its logic, there are several theoretical and cluding in Africa (Bratton and van de Walle 1997). Al- empirical reasons to question the expectations of these though democratic transitions outnumber reversions new distributive conflict models. Socioeconomic in- from democratic rule, the period also saw a number of equality plays a central role in these models, but has transitions to authoritarian rule. cross-cutting effects. The more unequal a society, the Not only does this temporal focus on the third wave greater the incentives for disadvantaged groups to capture a wide-ranging sample of regime changes but it press for more open and competitive politics. Yet the also overlaps with important changes in international context. During the Cold War era, both right- and left- Stephan Haggard is Lawrence and Sallye Krause Distinguished Pro- wing dictators could exploit great power rivalries to win fessor, Graduate School of International Relations and Pacific Stud- support from external patrons. During the 1980s and ies, University of California at San Diego, 9500 Gilman Drive, La 1990s, the decline and ultimate collapse of the Soviet Jolla, CA 92093 ([email protected]). Union created a much more permissive international Robert R. Kaufman is Professor, Department of Political Science, Rutgers University, 89 George Street, New Brunswick, NJ 08901 environment for democratic rule (Boix 2011). ([email protected]). Using an extremely generous definition of “distribu- The authors thank Carles Boix, Michael Bratton, T.J. Cheng, tive conflict” transitions, we find that between 55% and Ruth Collier, Javier Corrales, Ellen Commisso, Sharon Crasnow, 58% of the democratic transitions during this period Anna Grzymala-Busse, Allan Hicken, Jan Kubik, James Long, Ir- fan Noorudin, Grigore Pop-Eleches, Celeste Raymond, Andrew conformed—even very loosely—to the causal mecha- Schrank, and Nic VanDewalle for comments on earlier drafts, includ- nisms specified in the distributive conflict models. Thus, ing on the construction of the dataset. We received useful feedback even with an expansive definition of distributive con- from a presentation at the Watson Institute, Brown University, and flict, more than 40% did not conform at all. Moreover, comments from Ronald Rogowski and anonymous reviewers of the a substantial number of the distributive conflict tran- APSR. Particular thanks as well to Christian Houle for making his dataset available. We also thank Vincent Greco, Terence Teo, and sitions occurred under conditions of high inequality, Steve Weymouth for research assistance. a result that is at odds with the expectations of the
1 Inequality and Regime Change August 2012 theory. Approximately 30% of all transitions occurred is limited, permitting more intensive treatment of the in countries that ranked in the top tercile in terms of in- relevant cases and thus more robust inference. equality, and a substantial majority of these transitions We begin in the first section by reviewing distribu- resulted from distributive conflict; this finding is robust tive conflict models of regime change, focusing on the to alternative measures of inequality. These findings do contributions by Boix (2003; 2008), Acemoglu and not necessarily overturn distributive conflict theories, Robinson (2000; 2001; 2006), and Przeworski (2009). but suggest that they are underspecified with respect The next section discusses methodological issues. The to scope conditions and only operate under very par- remainder of the article is structured around a con- ticular circumstances. sideration of transitions to democracy and reversions Given the substantial incidence of nondistributive to authoritarian rule. Our causal process observations conflict transitions, we find several alternative causal show not only that transitions occur across cases with pathways to democratic rule. External actors were de- very different levels of inequality—as the null findings cisive in some cases. In many cases, however, other do- in econometric models already attest—but also that a mestic causal factors induced incumbents to relinquish large number of democratic transitions and reversions power in the absence of strong challenges from below. occur in the absence of significant redistributive conflict Elite incumbents were sometimes challenged by elite altogether. outgroups or defectors from the ruling coalition who The returns from this exercise are both substan- saw gains from democratic openings. In other cases, tive and methodological. First, the findings cast doubt elite incumbents ceded power in the absence of mass on the prevalence of the core causal mechanisms at pressure because they believed they could control the work in the underlying model, including the relation- design of democratic institutions in ways that protected ship between inequality and particular types of elite their material interests. and mass behavior. In the conclusion, we raise ques- An even smaller percentage of reversions—less than tions about alternative approaches and suggest several a third—conformed to the elite-mass dynamics postu- ways in which the theory might be modified: There lated in the theory, and once again, we found little rela- may be other channels through which inequality can tionship between the incidence of these transitions and destabilize democratic rule, and there might be other socioeconomic inequality. However, we did find several economic and institutional factors that condition the alternative causal mechanisms. In several cases, incum- capacity of low-income groups to engage in collective bent democratic governments were overthrown not by action. Second, our methodological contribution raises socioeconomic elites seeking to block redistribution, important questions about the validity of reduced-form but by authoritarian populist leaders promising more panel designs, including with respect to the coding of redistribution. Even more commonly, however, rever- regime type itself. More positively, it suggests a fruitful sions were driven by conflicts that either cut across class way of combining quantitative and qualitative methods lines or arose from purely intra-elite conflicts, particu- that focuses attention on alternative transition paths larly conflicts in which factions of the military staged rather than the partial-equilibrium treatment effects coups against incumbent office holders. of favored variables. Our analysis is motivated by methodological as well as substantive concerns. In contrast to quantitative THEORY tests of the relationship between inequality and regime change, we have constructed a qualitative dataset Adam Przeworski (2009, 291) poses the puzzle of of within-case causal process observations (Haggard, democratic transitions in the clearest terms: “Why Kaufman, and Teo 2012). Our approach differs from would people who monopolize political power ever other such designs in that it examines all discrete decide to put their interests or values at risk by shar- country-years that have been coded as transitions or ing it with others? Specifically, why would those who reversions in two prominent datasets: Polity IV and the hold political rights in the form of suffrage decide to dichotomous coding scheme developed by Przeworski extend these rights to anyone else?” The seminal work et al. (2000) and extended by Cheibub, Ghandi, and of Meltzer and Richard (1981) provides the point of Vreeland (2010). departure for all current distributive conflict models of Critics of “medium-N” designs have argued that regime change.1 The Meltzer-Richard model posits that such designs lack both the detail of individual case the distribution of productivity and income is skewed studies or smaller-N designs and the precision of well- to the right, with most citizens falling at the lower and specified larger-N econometric models. Yet we argue middle range of the distribution and a smaller tail con- that they are particularly useful for evaluating whether stituting the rich; the mean income exceeds the median. the causal mechanisms stipulated in formal models— Where voting rules result in appeals to the median which typically involve complex sequences of strate- voter, the wider the divergence between the median gic interactions—are in fact present in the cases. The and mean income, the more is to be gained from re- approach is particularly useful for testing theories of distribution. Put differently, in countries with more relatively rare events, such as democratic transitions skewed income distributions, the poor have more to and reversions, civil wars, genocides, financial crises, gain from redistribution and should have more gener- and famines. In cross-national quantitative models of ous tax and transfer programs as a result. these phenomena, the number of country-years in the panel is large, but the number of cases to be explained 1 See also Romer (1975) and Roberts (1977).
2 American Political Science Review
In the distributive conflict theories of regime change, have less to lose from competitive politics. Elites have most notably in the work of Boix (2003) and Acemoglu stronger incentives to repress as inequality increases and Robinson (2000; 2001; 2006), these expectations and when assets are fixed. are modified and expanded to endogenize the very Although broadly similar in spirit, Acemoglu and existence of democratic governments. These models Robinson’s (2006) Economic Origins of Dictatorship differ in ways we explicate later, but both rest on com- and Democracy introduces several innovations. Ace- plex causal chains including both structural and game- moglu and Robinson concur with Boix that regime type theoretic components: inequality, distributive conflict, is a function of the balance of power between high- and strategic interactions between incumbents and op- and low-income groups. Although elites monopolize positions over the nature of political institutions. In de jure power, masses potentially wield de facto power models of democratic transitions, low-income groups— through their capacity to mobilize against the regime. sometimes in coalition with middle-class forces—mobi- Like Boix, Acemoglu and Robinson elaborate more lize in favor of redistribution and against the authoritar- complex “three-class” models in which the pressure on ian institutions that sustain inequalities. These theories elites comes from coalitions of middle- and low-income are vague about how collective action problems are groups. However, the establishment of mass democracy solved, but posit that they can be overcome by changes presupposes the engagement of low-income sectors be- in information with respect to the solidity of incum- cause of their sheer weight. Because they constitute the bent power (Boix) or by increasing returns from mobi- majority, masses can sometimes “challenge the system, lization as inequality rises (Acemoglu and Robinson). create significant social unrest and turbulence, or even Faced with the threat of being displaced by force—in ef- pose a serious revolutionary threat” (25). fect, through revolution—elites calculate the net cost of High inequality increases the incentives for au- repression vs. concession, including institutional ones. thoritarian elites to repress these political demands At very high levels of inequality, the threats posed by for redistribution. To this observation, Acemoglu and democratization are too high to accept and they choose Robinson add an important point about credible com- to repress. Yet at low or medium levels of inequality, mitments. When elites are confronted by mobilization redistributive demands can be managed through class from below, they can make short-run economic conces- compromises over institutions and policy that permit sions to diffuse the threat. Yet politically and econom- democratic transitions. ically excluded groups are aware that elites can renege Carles Boix’s (2003) Democracy and Redistribution on these concessions when pressures from below sub- is a significant exemplar of this broad approach. Boix side. Because there is a cost to subsequently reversing defines a right-wing authoritarian regime as one in democracy after a transition has occurred, democratic which the political exclusion of the poor sustains exist- institutions provide a means for elites to credibly com- ing economic inequalities. According to Boix (2003, 37) mit to a more equal distribution of resources not only “a more unequal distribution of wealth increases the re- in the present but into the future as well. distributive demands of the population.... [However] Acemoglu and Robinson agree with Boix that, al- as the potential level of transfers becomes larger, though inequality increases the incentive for excluded the authoritarian inclinations of the wealthy increase groups to press for democracy, it also increases elite and the probabilities of democratization and demo- incentives to repress. High inequality is inauspicious cratic stability decline steadily.” The translation of for democracy. However, Acemoglu and Robinson ar- these demands into a change in institutions hinges on gue that democratization is also unlikely to occur in the balance of power between the wealthy and the authoritarian governments with low levels of inequal- poor. Boix offers an informational model in which ity because the demand for it is also attenuated; de- regime changes are triggered by exogenous shocks that spite political restrictions, excluded groups nonethe- weaken the elite or reveal its weakness (28–30). A less share in the distribution of societal income. They necessary (although not sufficient) mechanism driving conclude that the relationship between inequality and regime change is pressure from below: “As the least democratic transitions should exhibit an inverted-U well off overcome their collective action problems, that pattern, with transitions to democratic rule most likely is, as they mobilize and organize in unions and political to occur at intermediate levels of inequality. parties, the repression cost incurred by the wealthy It is important to emphasize that the theory is not rise[s],” forcing elites to make institutional compro- simply a structural one but operates through strategic mises (13). interactions between elites and masses: incentives for Boix also emphasizes the role played by capital mo- collective action on the part of the masses and repres- bility in mitigating this relationship (see also Freeman sion or concessions on the part of elites. At middle lev- and Quinn 2012). High levels of capital mobility en- els of inequality, grievances are sufficient to motivate hance the bargaining power of elites. Fixed assets, by the disenfranchised to mobilize, but not threatening contrast, limit the options of the wealthy and make enough to invite repression (see also Burkhart 1997; them vulnerable to democratic redistribution and thus Epstein et al. 2006). more resistant to it. Given the decision of the poor to Boix (2003) and Acemoglu and Robinson (2006) mobilize, the incentives of upper-class incumbents to extend their arguments to a consideration of the sta- repress are a function of the level of inequality and bility of democratic rule and reversion to autocracy mobility of assets. Transitions are most likely when as well. Implicit in the theory is the assumption that inequality is low, asset mobility is high, and elites high-inequality democracies are rare; for that reason,
3 Inequality and Regime Change August 2012 less attention is given to the relationship between in- Kaufman, and Teo 2012). Selecting on the dependent equality and democratic breakdown. Nonetheless Boix variable is a central feature of this approach, which postulates a direct linear relation between the degree is designed to test a particular theory and thus rests of inequality and the likelihood of reversion to dicta- on identification of the causal mechanism leading to torship. In high-inequality democracies, redistributive regime change. In contrast to the more common prac- pressures from lower-class groups will be more intense, tice of purposeful (Gerring 2006; 2007a; 2007b) or ran- motivating elites to deploy force against incumbents in dom (Fearon and Laitin 2011) selection of cases for order to reimpose authoritarian rule. Although Ace- more intensive analysis, our approach is to select all moglu and Robinson posit an inverted-U shaped re- transition and reversion cases in the relevant sample lation between inequality and democratic transitions, period (1980–2000). The cases included in our dataset they agree that countries that do manage to democ- come from the dichotomous coding of transitions and ratize at high levels of inequality “do not consolidate reversions in Cheibub, Ghandi, and Vreeland (CGV; because coups are attractive” (38). The costs for elites 2010) and from Polity IV. For the continuous Polity IV of mobilizing against democratic rule are less than metric, we use a cutoff of 6 to indicate a transition, a the losses arising from redistribution under democratic benchmark used in the dataset itself. We have, however, rule. examined alternative cutoff points of 7 and 8 and find that as the bar is raised, the percentage of distributive METHOD: conflict transitions in our sample actually declines to CAUSAL PROCESS OBSERVATIONS 47.9 and 42.1%, respectively, suggesting that the results are in fact robust. At first glance, these theories appear amenable to rel- Within-case causal process observation involves the atively straightforward tests. Is the level of inequality reconstruction of an empirical sequence of actor deci- associated with transitions to and from democratic rule sions, ultimately strategic in form, that are postulated or not? Yet empirical tests are complicated by the fact by the theory to yield the given outcome. Within-case that different measures of inequality capture different analysis codes whether and to what extent individual socioeconomic cleavages and the quality of the data is cases conform with the stipulated causal logic. This cod- notoriously poor. We also find that measures of democ- ing can then be aggregated in a second stage to consider racy commonly used in panel designs leave much to be characteristics of the whole population or subsets of it. desired. In constructing the dataset of causal process obser- The problems of testing these theories are not limited vations on regime change, we begin with the stipulated to the constraints posed by the data: They are also re- causal mechanisms that run from inequality through lated to the reduced-form nature of most cross-national the following elements: the mobilization of distribu- panel designs. These quantitative models typically omit tive grievances by the poor or—more commonly—by the intervening causal processes and focus directly on coalitions of low- and middle-income groups; elite cal- the relationship between some antecedent condition— culations about the costs of repressing these challenges in this case, levels of inequality—and the outcome vari- or offering political concessions; the iterated strategic able, regime change in this instance. However, as the lit- response of the masses to those elite decisions; and the erature on process-tracing and causal process observa- ultimate outcome of regime maintenance or change tion has pointed out,2 the empirical question is not only (see particularly Boix 2003, 27–36, and Acemoglu and whether antecedent conditions are linked statistically Robinson 2006, 181–220, for explication of the basic to the outcome but whether they also do so through the models). In the first instance, we seek to establish stipulated causal mechanisms. In this case, we want to whether distributive conflict is present or not and, if so, know not only whether inequality is associated with whether and how it affects the decisions that result in or regime change but also whether its effects operate constitute regime change. For democratic transitions, through the particular causal mechanisms postulated we first identify the decisions made by authoritarian in distributive conflict theory. leaders to make political concessions or withdraw al- Our method of causal process observation includes together. For reversions, we identify actions taken by two stages: (1) within-case analysis and coding and (2) challengers within or outside the government that re- aggregation across the population of cases (Haggard, sult in the overthrow of democratic rule. For each tran- sition and reversion, we then provide a narrative that 2 The concept of causal process observation (Collier, Brady, and reconstructs the causal process and assesses whether Seawright 2010) grew out of an earlier stream of methodological the key political decisions in question were a result of work on process-tracing initiated by Alexander George (Bennett and distributive conflicts. We then provide a justification of George 2005; George and McKeown 1985) and subsequently joined the coding and references used to make the decision.3 by work on the empirical testing of formal models, including through The selection of all cases for a given time period has “analytic narratives” (Bates et. al. 1998). Although Collier, Brady, and Seawright distinguish between causal process observation and the advantage of permitting what we call “stage two” process-tracing, we see them as essentially the same. However, we prefer the term “causal process observation” because it underscores the link to the testing of a particular theory; we suggest later the 3 We personally researched all cases cited in the dataset and con- particular way in which this approach can be used to leverage causal sulted closely with each other on each coding decision and consis- inference. A related strand of work is associated with the “mecha- tency across cases. Country and regional experts also reviewed cod- nism” approach to causation (Falletti and Lynch 2010; Gerring 2007b; ing decisions, particularly in ambiguous cases (see Haggard, Kauf- 2010; Hedstrom and Ylikoski 2010). man, and Teo 2012).
4 American Political Science Review analysis: the aggregation of the individual causal pro- In principle, multistage models can be constructed cess observations to permit analysis of the population that work from structural causes through intervening as a whole or relevant subsets of it. For example, we pay behaviors to institutional effects (King, Keohane, and particular attention to high- and low-inequality cases Verba 1994, 85–87). Some critics of the mechanisms because the theory has particular expectations about approach have argued that mechanisms may be nothing how such cases should behave. more than such chains of intervening variables (Beck The method of causal process observation has sev- 2006; 2010; Gerring 2007b; 2010; Hafner-Burton and eral advantages that can enrich the testing of formal Ron 2009). Although possible in principle, the con- theories through quantitative empirical designs; we see tinued reliance on reduced-form specification suggests it as a complement to such approaches, not a substitute. that this problem is in fact not addressed, in part be- In a quantitative model, the effects of either structural cause of the labor intensity of recoding existing datasets variables, such as inequality, or behavioral ones such to conform more precisely with the theory being tested. as protest are estimated across a heterogeneous set In each of the remaining sections on democratic of cases, some of which transition as a result of the transitions and reversions, we begin with a review of stipulated causal mechanism and some of which do the quantitative findings on the relationship between not. The focus on average treatment effects masks inequality and regime change and then present both the heterogeneity of transition paths; the variable in aggregate and select case study findings from the causal question is either significant or not. By contrast, causal process observations in our dataset. We show that the process observations do not ask whether the variable in support for the distributive conflict model of regime question is significant, but whether the transition path change is weak, even under highly generous coding in the cases conforms with the causal process stipulated rules. When these rules are tightened, the evidence is in the theoretical model. weaker still. As we see later, the quantitative work on inequal- ity and regime change is highly inconclusive at best and is even more limited for the third wave tran- TRANSITIONS TO DEMOCRATIC RULE sitions. Nonetheless, causal process observations can Acemoglu and Robinson (2006) do not present sys- complement quantitative analysis in two ways that can tematic empirical evidence in support of their claims.4 strengthen causal inference. First, if causal process ob- Much of their book is taken up with a discussion of servations showed that elite-mass conflicts did drive the underlying intuition of the theory (1–47, 80–87) transitions in a significant number of cases, it could and the presentation of a family of formal models of reopen null statistical findings. The causal process ob- democratic and nondemocratic regimes (89–172) and servations would suggest, for example, the need for of regime change (173–320). Acemoglu and Robinson better specification of the quantitative model or more do present scatterplots showing a positive relationship appropriate measures of inequality. However, if regime between equality and the level of democracy across a change was not driven by such conflicts in a signifi- global sample of countries (58–61) and provide short cant number of cases, the finding could be considered case studies of Great Britain, Argentina, South Africa, disconfirmatory. More importantly, the finding could and Singapore (1–14). Yet these correlations and cases be disconfirmatory even if inequality were statistically are illustrative at most. significant in the quantitative analysis; this would occur In his analysis of democratic transitions over the very if causal process observation showed that the effects long run (1850–1980), Boix (2003) finds that the distri- of inequality work through causal channels not posited bution of land, proxied by the share of family farms, has by the game-theoretic models. an effect on the transition to democratic rule. More un- In addition to its advantages in more closely test- equal societies are both less likely to make a transition ing the actual mechanisms specified in causal mod- to democracy and less stable when they do (90–97). els, causal process observations also address a second Boix also explores a highly uneven panel of countries important problem in standard quantitative panel de- for the 1950–90 period (only 587 observations), includ- signs: the mismatch between the temporal framework ing developed ones (71–88). Using a Gini index as his of a stipulated causal process and the constraints of measure of inequality, Boix finds some evidence that country-year coding of cases. In cross-national panels, increases in the level of inequality reduce the likeli- each country-year is coded as a transition or nontransi- hood of a democratic transition, but the findings are tion year; these codings constitute the dependent vari- not altogether robust (see for example, 79: Model 2A). able. The causal covariates are similarly either contem- More recently, other quantitative studies have taken poraneous or antecedent with some lag structure. Yet up the challenge raised by Boix (2003) and Ace- the causal sequence of actor choices associated with moglu and Robinson (2006), but with mixed results. transitions and reversions may be more compressed or extended, not constant across cases, and thus not well captured by the artifact of the country-year coding 4 Earlier work (Acemoglu and Robinson 2000; 2001) was motivated constraint typical of the panel design. As we see later, by experiences in nineteenth-century Europe and early twentieth- many cases that are coded as transitions prove to be century Latin America. However, in those articles, as well as in the dubious when a more extended but variable temporal later book, the formal theory is cast in general terms, without specify- ing scope conditions that might apply to third wave transitions. In the context is taken into account, a point emphasized more book, moreover, the illustrations from South Africa and Singapore generally in the work of Pierson (2004). rely on much more recent developments.
5 Inequality and Regime Change August 2012
Like Boix, Ansell, and Samuels (2010) consider both sequence (mass mobilization followed by authori- long-historical and postwar samples (1850–1993, 1955– tarian withdrawal). 2004). They find that land concentration makes de- mocratization less likely, but that increases in income inequality make it more likely. They argue that increas- In coding the cases, we were deliberately permissive, ing income inequality reflects the emergence of a new writing coding rules that gave the benefit of the doubt capitalist class that challenges landed elites, a dynamic to the theory (Haggard, Kaufman, and Teo 2012). Our consistent with Boix’s (2003, 47–59) and Acemoglu and coding allowed us to consider a variety of distributive Robinson’s (2006, 266–86) extension of their models conflicts that may not be captured by any single in- into three-class variants. equality measure, from urban class conflicts to ethnic The limited number of other tests in the literature and regional ones. Yet such conflicts must be fought generally fail to find a relationship between inequality around distinctive and identifiable inequalities. The and democratic transitions. A cross-sectional design by economically disadvantaged or the organizations rep- Dutt and Mitra (2008) finds a relationship between resenting them need not be the only ones mobilized in inequality measured by the Gini coefficient and “polit- opposition to the existing regime. Although mass mobi- ical instability,” but fails to find a relationship between lization must partly reflect demands for redistribution, inequality and transitions to democratic rule. Chris- it can be motivated by other grievances as well. tian Houle (2009) creates a dataset using an alterna- An important coding issue is the question of “poten- tive measure of inequality: capital’s share of income tial” threats in the absence of actual mobilization. As in the manufacturing sector. Using the dichotomous Acemoglu and Robinson (2006) note in distinguishing coding scheme developed by Przeworski et al. (2000) between de jure and de facto power, the poor can be and Cheibub and Ghandi (2004), Houle shows that in- considered a potential threat in virtually every case. equality bears no systematic relationship to democratic However, the strategic basis of reforms aimed at pre- transitions over the 1960–2000 period, but is a signifi- empting potential long-term threats rests on probabil- cant predictor of reversions to authoritarian rule. In a ity estimates and time horizons on the part of elites that wide-ranging study of the determinants of democrati- differ quite substantially from those that drive elite re- zation, Teorell (2010, 60) also fails to find a relationship sponses to more immediate challenges. Moreover, we between a Gini coefficient and democratic transitions. are also wary of the coding challenge: Virtually any case could be coded as one in which there was a “potential” challenge from below, with a corresponding decline in analytic leverage. However, we do take potential Distributive Conflict and threats into account where there has been a recent Nondistributive Conflict Transitions history of mass mobilization demanding democratic In sum, the quantitative work on inequality and regime reforms. change is highly inconclusive at best, and even more We coded all cases in which such threats from below limited for the third wave transitions. Many of these did not occur at all or appeared to play only a marginal tests do not empirically model the underlying causal causal role as “nondistributive transitions.” Why, in the processes stipulated in the most significant formal absence of significant pressure from below, would elites models. Therefore, to undertake causal process ob- withdraw or make institutional compromises that risk servations, we need to interpret the underlying causal the redistribution of assets and income not only in the mechanisms at work in the theory. Two mechanisms present but also into the indefinite future? As others appear central. First, elites must confront political- have argued (Huntington 1991; Linz and Stepan 1996; cum-distributive pressure from below, or a “clear and Collier 1999; O’Donnell, Schmitter, and Whitehead present danger” of it. In the absence of such pressures, 1986), there are a variety of routes from closed political it is not clear why elites would be motivated to cede systems to democracy. We identify three: those driven power at all, as Przeworski’s trenchant question sug- by international pressures, those involving intra-elite gests. Second, there must be some evidence—minimally conflicts and defections, and those in which incumbent in the temporal sequence of events—that the repression authoritarian elites withdraw in the belief that they can of these challenges appears too costly and that elites control the post-transition democratic order in ways make institutional compromises as a result. that limit democracy’s redistributive impact. We therefore code “distributive conflict” transitions International factors played a decisive role in a as ones in which both of the following occurred: number of third wave transitions (Boix 2011; White- head 1996). In a handful of cases—including Grenada • The mobilization of redistributive grievances on (1984), Panama (1989), and Haiti (1994)—outside in- the part of economically disadvantaged groups or tervention took a military form. Yet particularly in the representatives of such groups (parties, unions, wake of the end of the Cold War, aid donors—both NGOs) posed a threat to the incumbency of ruling multilateral and bilateral—became less tolerant of un- elites. democratic regimes that appeared guilty of economic • And the rising costs of repressing these demands mismanagement and outright corruption. Threats or appear to have motivated elites to make politi- withdrawal of aid played an important role in transi- cal compromises or exit in favor of democratic tions in a group of low-income African countries in challengers, typically indicated by a clear temporal particular.
6 American Political Science Review
Even if we set aside the role of international pres- of countries ranked in terms of land distribution. More sures, threats from below are by no means the only problematic, and against theoretical expectations in domestic pressures that can cause elites to acquiesce to both Boix (2003) and Acemoglu and Robinson (2006), democratizing institutional changes. A common cause there was a substantial incidence of distributive conflict of transition in the nondistributive conflict cases is transitions among the high-inequality cases. When in- intra-elite rivalries. These rivalries may stem from com- equality is measured using the Gini, about 75% of high- petition among the political, military, and economic inequality transitions were distributive conflict transi- elites that constitute the authoritarian coalition—for tions; the incidence of such transitions is 60% using the example, when factions within the regime seek to dis- land inequality measure, and 57% using capital’s share place incumbents—or from elite challenges from out- of income. side the regime altogether (Slater and Smith 2012). In Table 2 offers additional insight into the causal role a number of cases, we found that concessions to elites of distributive conflict. Columns 2 and 3 divide the rather than mass challenges appear as important as CGV transitions into distributive and nondistributive distributive conflicts pitting rich against poor. types; columns 4 and 5 replicate the exercise for Polity Even when elites remain relatively unified, they may transitions. We also identify the non-overlapping cases. still acquiesce to—or even lead—democratic reform if The last column shows the average Polity score from they believe they can retain leverage over the political the time of the transition through either the end of the process while reducing the costs of repression. Incum- sample period or until an outright reversion to author- bent elites can do this in several ways, including through itarian rule. the design of political institutions that give them effec- The information contained in Table 2 raises serious tive vetoes or through the organization of political par- questions about the validity of the coding of democratic ties that exploit other cleavages to dampen distributive transitions in these two major datasets and, as a result, conflicts. Dominant parties provide incumbent political casts doubt on the inferences that have been drawn in elites particular organizational advantages that can be the quantitative work that employs them. Only 55.4% redeployed in a more competitive context. of the CGV transitions are also Polity cases, and 21 of Note that each of the alternative domestic causal the 65 CGV transitions had Polity scores of less than 6. mechanisms we have sketched—intra-elite conflict or Even where the two datasets are in agreement, more- defection and authoritarian elites ceding office because over, our examination of the cases raises questions of confidence in their post-transition chances—may in about the validity of the coding process. Insiders and fact be related precisely to the weakness of immediate elites repressed opposition and/or exercised dispropor- threats from below. Where such threats are limited, tionate control over them in the nominallydemocratic elites are more likely to control the transition. Societies cases of Croatia, Niger, and Thailand. Transitions in in which the poor are not mobilized through program- Guatemala (1986) and Honduras (1982) empowered matic parties, unions, or other organizations may be nominally democratic governments that actually inten- especially prone to vote buying, patronage, and other sified repression of social movements that had redis- forms of clientelistic control that would guarantee elite tributive objectives. Death squads continued to terror- control of politics, even in nominally democratic set- ize the opposition in El Salvador after the transition tings (Kitschelt and Wilkinson 2006). in 1984. In at least four cases—Ghana under Rawlings; Kenya under Moi; Malawi, where an “insider” won the transitional election; and Romania—the military or Inequality and the Incidence of incumbent elites continued to exercise disproportion- Distributive Conflict Transitions ate influence over the allocation of resources after the transition. In all of these cases, the transitions appear to Table 1 shows the distributive and nondistributive tran- conform more closely to what Levitsky and Way (2010) sitions, using the definition of transitions in the CGV call “competitive authoritarianism” than to democracy. dataset; in the text, we also report the distribution of Because we seek to engage the quantitative analysis these types of cases based on Polity transition cod- that deploys such data, however, we do not discard ing. The cases are arrayed according to three mea- or reclassify cases identified as transitions in the two sures of inequality: Christian Houle’s (2009) measure datasets. of capital’s share of income in the manufacturing sec- What about the theoretical expectations of the role tor (capshare), a Gini coefficient from the University of distributive conflict in democratic transitions? We of Texas Inequality Project’s Estimated Household In- found that distributive conflict played some causal role come Inequality (EHII) dataset (2008), and the Van- in propelling transitions in about 55% of CGV and hanen (2003) measure of land inequality. We divide 58% of Polity transition cases. These are substantial, the sample of all developing countries into terciles of but by no means overwhelming shares of the cases. In high-, medium-, and low-inequality cases and identify combination with the findings in Table 1, the large per- the transitions that fall into each tercile. centage of nondistributive transitions suggests strongly The table shows that transitions occurred at all lev- that the link between inequality and distributive con- els of inequality, regardless of which measure is used. flict transitions is conditional at best. Twenty-nine percent of transitions occurred in the up- Yet even these findings need to be tempered by the per third of countries ranked by capshare and Gini generosity of our coding rules. Although pressure from inequality, and about 34% occurred in the top tercile below did play an unambiguously significant role in a
7 Inequality and Regime Change August 2012 Uruguay (1985) Inequality Measures Malawi (1994)Mongolia (1990)Nepal (1990)Romania (1990) Mongolia Peru (1990) (1980) Ukraine Romania (1991) (1990) Capital Share of Income in Manufacturing Sector Gini coefficient (Texas Inequality dataset) Share of Family Farms (Vanhanen) Albania (1991) Bangladesh (1986) Argentina (1983) Central African Republic Armenia (1991) Central African Republic Bolivia (1982) Chile (1990) Armenia (1991) Ghana (1993) Albania (1992) Belarus (1991) The Philippines (1986) Poland (1989) South Korea (1988) Sudan (1986) Uruguay (1985) 57.1Argentina (1983)Benin (1991)Bulgaria (1990) 42.9El Salvador (1984) CroatiaGuatemala (1991) (1986)Kenya (1998)Latvia (1991) Panama Hungary Pakistan (1989) (1988) (1990)Madagascar (1993) Paraguay (1989)Malawi (1994)Nepal (1990) El Salvador (1984) Senegal (2000) Turkey Indonesia Fiji Peru (1983) (1999) (1980) (1992) 75.0 The Philippines (1986) (1993) Panama (1989) Honduras Sri (1982) Lanka (1989) Suriname Pakistan (1988) (1988) Chile (1990) Fiji (1992) Senegal Congo (2000) (1992) 25.0 Suriname Thailand Argentina (1991) (1992) (1983) El Salvador (1984) Uruguay (1985) Benin Mexico Kenya (1991) (2000) (1998) Senegal Turkey (1983) Pakistan (2000) Malawi (1988) (1993) (1994) Uganda (1980) Comoros 60.0 (1990) The Philippines (1986) Sudan (1986) 40.0% Brazil (1985)Burundi (1993)Indonesia (1999)Nigeria (1999)Peru (1980)Sri Lanka Ghana (1989) Mexico (1993) Nicaragua (2000)Thailand (1984) (1992) Sierra Leone (1996) Sierra Leone Brazil (1998) (1985) Benin Bolivia Burundi (1991) (1982) (1993) Congo (1992) Sierra Leone (1998) Sierra Leone Paraguay (1996) (1989) Bulgaria (1990) Brazil (1985) Guatemala (1986) Bolivia (1982) Kenya (1991) Honduras (1982) Czechoslovakia (1989) Chile (1990) Estonia (1991) Guatemala (1986) Hungary (1990) Nicaragua (1984) Latvia (1991) Lithuania (1991) Panama (1989) Paraguay (1989) distributive and nondistributive conflict cases TABLE 1. Distributive and Nondistributive Transitions by Level of Inequality, 1980–2000 Level of InequalityHigh Distributive NondistributivePercentage of DistributiveMedium Nondistributive Distributive Nondistributive
8 American Political Science Review ity of Texas Inequality Project Suriname (1991) Thailand (1992) (1991) Serbia (2000)Taiwan (1996) South Korea (1988) Sri Lanka (1989) Turkey (1983) Uganda (1980) Inequality Measures Poland (1989)South Korea (1988)Ukraine (1991) Mexico (2000) Macedonia (1991) Nicaragua (1984) Niger (2000) Nigeria (1999) Poland (1989) Sierra Sierra Leone Leone (1998) (1996) Taiwan (1996) Macedonia (1991)Uganda (1980) Madagascar (1993) Nigeria (1999) Czech Republic (1989) Nepal (1990) Hungary (1990) Macedonia (1991) Niger (1993) Serbia (2000) Serbia (2000) Suriname (1991) Taiwan (1996) Capital Share of Income in Manufacturing Sector Gini coefficient (Texas Inequality dataset) Share of Family Farms (Vanhanen) Armenia (1991) Belarus (1991) Estonia (1991) Belarus (1991) Suriname (1988) Cape Verde (1991) Fiji (1992) Central African Republic Albania (1992) Bangladesh (1986) Burundi (1993) Bangladesh (1986) Congo (1992)Estonia (1991)Lithuania (1991)Mali (1992) Cape Verde (1990) Comoros (1990) Czechoslovakia (1989) Niger (2000) Mali Grenada (1992) (1984) Niger (1993) Guinea-Bissau (2000) Sudan (1986) Comoros (1990) Grenada (1984) Sao Tome and Principe Sao Tome and Principle (1991) Cyprus (1983) Grenada (1984) 66.7Niger (1993)Niger (2000)Romania (1990) 33.3 Honduras (1982) (1993) Cyprus (1983) Lithuania (1991) 52.644.4% Latvia (1991) Bulgaria (1990) Cyprus (1989) Croatia (1991) Cape Verde (1990) 55.6% Mali 47.4 (1992) Indonesia (1999) Madagascar (1993) Ghana (1993) Croatia (1991) Guinea-Bissau (2000) 45.0% 68.8 55.0% 31.2% 52.2% 47.8% Mongolia (1990)Suriname (1988)Ukraine (1991) Guinea-Bissau (2000) Sao Tome and Principe (1991) Transitions: Cheibub, Ghandi, and Vreeland (2010); transition types: Haggard, Kaufman, and Teo (2012); capital share: Houle (2009); Gini: Univers distributive and nondistributive conflict cases distributive and nondistributive conflict cases TABLE 1. Continued. Level of InequalityPercentage of DistributiveLow Nondistributive DistributivePercentage of NondistributiveMissing Data Distributive Nondistributive Sources: (2008); share of family farms: Vanhanen (2003).
9 Inequality and Regime Change August 2012
TABLE 2. Distributive and Nondistributive Transitions, 1980–2000
CGV Transitions Polity Transitions
Country/Year Distributive Nondistributive Distributive Nondistributive Polity score
Albania 1991 X Not a Polity transition 4.1 Argentina 1983 X X 7.4 Armenia 1991 X X 7.0 Bangladesh 1986 X Not a Polity transition 2.3 Bangladesh 1991 Not a CGV transition X 6.0 Belarus 1991 X X 7.0 Benin 1991 X X 6.0 Bolivia 1982 X X 8.8 Brazil 1985 X X 7.8 Bulgaria 1990 X X 8.0 Burundi 1993 X Not a Polity transition −1.6 Cape Verde 1990 XX7.1 (CGV), 1991 (Polity) Central African X Not a Polity transition 5.0 Republic 1993 Chile 1990 (CGV), XX8.1 1989 (Polity) Comoros 1990 X Not a Polity transition 2.6 Congo 1992 X Not a Polity transition 5.0 Croatia 1991 X Not a Polity transition −2.0 Croatia 2000 Not a CGV transition X 8.0 Cyprus 1983 X Not a Polity transition 10.0 Czechoslovakia 1989 XX8.2 (CGV), 1990 (Polity) Dominican Republic Not a CGV transition X 8.0 1996 El Salvador 1984 X X 6.6 Estonia 1991 X X 6.0 Fiji 1992 X Not a Polity transition 5.1 Fiji 1999 Not a CGV transition X 5.5 Ghana 1993 X Not a Polity transition 1.0 Grenada 1984 X Not a Polity transition N.A. Guatemala 1986 X Not a Polity transition 4.7 Guatemala 1996 Not a CGV transition X 8.0 Guinea-Bissau 2000 X Not a Polity transition 5.0 Guyana 1992 Not a CGV transition X 6.0 Haiti 1990 Not a CGV transition X 7.0 Haiti 1994 Not a CGV transition X 7.0 Honduras 1982 X X 6.0 Honduras 1989 Not a CGV transition X 6.2 Hungary 1990 X X 10.0 Indonesia 1999 X X 6.0 Kenya 1998 X Not a Polity transition −2.0 Latvia 1991 X X 8.0 Lesotho 1993 Not a CGV transition X 8.0 Lithuania 1991 X X 10.0 Macedonia 1991 X X 6.0 Madagascar 1992 X X 8.2 Malawi 1994 X X 6.0 Mali 1992 X X 6.5 Mexico 1997 Not a CGV transition X 6.5 Mexico 2000 X Not a Polity transition 8.0 Moldova 1993 Not a CGV transition X 7.0 Mongolia 1990(CGV), XX8.2 1992 (Polity) Nepal 1990 X Not a Polity transition 5.2 Nepal 1999 Not a CGV transition X 6.0 Nicaragua 1984 X Not a Polity transition 4.2 Nicaragua 1990 Not a CGV transition X 7.1 Niger 1993 (CGV), XX8.0 1992 (Polity)
10 American Political Science Review
TABLE 2. Continued.
CGV Transitions Polity Transitions
Country/Year Distributive Nondistributive Distributive Nondistributive Polity score
Niger 2000 X Not a Polity transition 5.0 Nigeria 1999 X Not a Polity transition 4.0 Pakistan 1988 X X 7.8 Panama 1989 X X 8.6 Paraguay 1989 X Not a Polity transition 5.7 Paraguay 1992 Not a CGV transition X 6.9 Peru 1980 X X 7.2 The Philippines 1986 XX7.5 (CGV), 1987 (Polity) Poland 1989 (CGV), XX8.0 1991(Polity) Romania 1990 X Not a Polity transition 6.4 Romania 1996 Not a CGV transition X 8.0 Russia 2000 Not a CGV transition X 6.0 Sao Tome and X Not a Polity transition N.A. Principe 1991 Senegal 2000 X X 8.0 Serbia 2000 X X 7.0 Sierra Leone 1996 X Not a Polity transition 4.0 Sierra Leone 1998 X Not a Polity transition 0.0 South Africa 1992 Not a CGV transition X 8.6 South Korea 1988 X X 6.5 Sri Lanka 1989 X Not a Polity transition 5.0 Sudan 1986 X X 7.0 Suriname 1988 X Not a Polity transition N.A. Suriname 1991 X Not a Polity transition N.A. Taiwan 1992 Not a CGV transition X 8.0 Taiwan 1996 X Not a Polity transition 8.8 Thailand 1992 X X 9.0 Turkey 1983 X X 7.7 Uganda 1980 X Not a Polity transition 2.5 Ukraine 1991 X X 6.0 Ukraine 1994 Not a CGV transition X 6.9 Uruguay 1985 X X 9.8 Zambia 1991 Not a CGV transition X 6.0 N/% 36/55.4% 29/44.6% 33/57.9% 24/42.1% 6.3
Note: In the dataset, we treat any transitions that are coded within a two-year window as the same case (for example, the CGV coding of the Philippines transition occurring in 1986, the Polity coding as 1987). Outside of this two-year window (for example, Paraguay) or where there is an intervening reversion (Sierra Leone), we treat them as separate cases. There are no Polity scores for Grenada, Sao Tome, and Suriname. Sources: CGV transitions from Cheibub, Ghandi, and Vreeland (2010); Polity transitions from Marshall, Gurr, and Jaggers (2010); transition types from Haggard, Kaufman, and Teo (2012). number of middle-income countries such as Argentina, groups, calling into question the class dynamics of South Korea, and South Africa, our expansive coding the model even if we allow for cross-class coalitions rules also necessitated the classification of cases as dis- including the poor. tributive conflict where there was considerable ambi- 3. A third source of ambiguity involved judgments guity about its causal weight. The ambiguity in specific about the role played by redistributive grievances cases stemmed from one or more of three factors. in opposition demands; in many instances, it was difficult to separate redistributive demands from grievances that focused on a defense of privileged 1. First, some distributive conflict transitions occurred positions, generalized dissatisfaction with authori- in small open economies that were highly vulnera- tarian incumbents, or nationalist claims. ble to pressure from donors or other international actors, and this pressure may have been decisive. 2. The class basis of protest constituted a second Table 3 lists the cases in the dataset where interna- source of ambiguity; in many cases, protest was tional pressures, the class composition of the protestors, dominated by middle- or even upper-middle-class or the nature of their redistributive grievances made
11 Inequality and Regime Change August 2012
TABLE 3. Ambiguous Cases of Distributive Conflict Transitions
CGV Dataset Polity Dataset
Country Source of Ambiguity Country Source of Ambiguity
Armenia Grievance Armenia Grievance Benin Class Benin Class Bulgaria Grievance Bulgaria Grievance Congo Class El Salvador International El Salvador International Estonia Class/Grievance Estonia Class/Grievance Fiji International Kenya International Lesotho Class/International Latvia Class/Grievance Latvia Class/Grievance Lithuania Class/Grievance Lithuania Class/Grievance Malawi Class/International Malawi Class/International Mali Class Mali Class Mongolia Class/Grievance Mongolia Class/Grievance Niger Class/Grievance/International Niger Class/Grievance/International Sri Lanka Grievance Suriname International Ukraine Class/Grievance Ukraine Class/Grievance Total 17 13 Percent of 26.2 22.8 total transitions the coding of the case ambiguous. Of particular signifi- cause of the presence of mobilization “from below” cance is the coding of several African “distributive con- that affected the transition, protest was primarily lim- flict” transitions, in which incumbent regimes—in the ited to civil servants, students, and other sectors of the midst of severe economic recessions—were vulnerable urban middle class. Moreover, several African cases both to intense donor pressure and the protest of rela- (Lesotho, Kenya, Malawi, and Niger) were also am- tively well-off public employees and student groups. biguous with respect to the role of international pres- Niger provides an example. The pivotal decision in sures. this case was an agreement by the military strongman, The nature of the grievances associated with the se- General Ali Saibou, to convene a National Confer- cession from Yugoslavia and the former Soviet Union ence, which then assumed the role of a transitional also warrants special mention. In three such cases— government and organized competitive elections. Dis- Croatia, Macedonia, and Belarus—the coding was un- tributive protests played a role in Saibou’s decision ambiguously nondistributive because mass mobiliza- to yield authority. Yet the opposition came primarily tion on distributive lines was altogether absent or there from the Nigerien Workers Union, which represented is strong evidence that the political process of indepen- Niger’s 39,000 civil servants, and the Union of Nigerian dence occurred as a result of intra-elite processes. Yet Scholars, which represented about 6% of the coun- in the Baltic cases, as well as in Ukraine, Mongolia, try’s school-aged population (Gervais 1997, 93). Both and Armenia, there is ambiguity as to the nature of groups bitterly opposed tough adjustment programs the claims made by groups engaged in mass mobiliza- demanded by the International Monetary Fund, but tion. Several regional specialists whom we consulted in the conflicts did not appear to engage the poor. As constructing our coding objected that these cases did Gervais (1997, 105) writes, “the political stakes raised not fall easily into the distributive conflict category and by ... adjustment policies tended to compromise the should be seen as the outcome of cross-class secession- benefits of the organized groups of the modern sector ist or nationalist movements and the resulting collapse as much as the privileges of the traditional political of multinational empires. In these cases, we believed class.” Notwithstanding our generous coding decision, that the evidence of conflicts within the polity between it is ambiguous at best to claim that the transition indigenous populations and the Russians warranted a process mapped directly to the underlying Meltzer- “distributive conflict” coding, but it is important to ac- Richard model in which the interests of the poor or knowledge the pivotal importance of strong nationalist even middle classes are pitted against the rich. aspirations that cut across class lines.5 If we were to Similar questions can be raised about the class com- position of protest in other African cases, includ- ing Benin, Congo, Lesotho, Malawi, and Mali. Even 5 These cases also posed a second coding problem: whether they though all of these cases meet our coding rules be- should be treated as democratic transitions at all given that they
12 American Political Science Review shift all of the ambiguous cases in Table 3 from the dis- political system to limit its redistributive impact. The tributive to nondistributive categories, the incidence of Turkish military transferred power to a new civilian distributive conflict transitions would fall to only 29.2% government in 1983, but only after crushing violent of the CGV transitions and to about 31% of the Polity left and right factions that had been a feature of Turk- transitions. ish politics in the late 1970s. As it gradually reopened Even with the expansive coding of distributive con- the political space in early 1983, the military vetoed flict transitions, we found a large share of cases—44.6% most of the new parties that had formed around estab- using the CGV measure and 42.1% of Polity cases—in lished politicians and designed institutions that gave it which distributive conflict played only a marginal role veto power over crucial areas of policy. Although the in the transition process. These cases followed the al- military elite was surprised by the victory of the one ternative causal pathways we identified earlier: transi- opposition party it had allowed to function, there is no tions driven by international pressures or by intra-elite indication that threats of mass mobilization influenced conflicts, and elite-led transitions in which incumbents the decision to allow the elections or to permit the believed they could control the democratic process to results to stand. limit its redistributive impact. We see similar processes of reform in Chile, where We have already noted that, in several of the “am- outgoing governments built in quite specific mech- biguous cases” discussed earlier, popular protest un- anisms through which the military would continue folded in the context of severe international pressure. to exercise oversight and supporters of the outgoing However, in other cases, protest was weak or entirely government would be overrepresented (Haggard and absent, and outside intervention was unambiguously Kaufman 1995). These mechanisms included the es- decisive. Transitions in Grenada (1984) and Panama tablishment of national security councils with a veto (1989) hinged almost entirely on U.S. military opera- role for the military establishment, constitutional and tions. In Haiti (1994), the military ruler negotiated his judicial guarantees limiting the authority of incoming exit as an international force of 21,000 troops prepared governments, and the allocation of Senate seats to be to land on the island. External political and economic filled by the head of the outgoing regime. In Kenya, pressures from donors or great power patrons were also Mexico, and Taiwan, incumbents ceded power gradu- decisive in Comoros (1990), Cape Verde (1990), the ally while competing aggressively and successfully in Central African Republic (1993), and Cyprus (1983). the newly liberalized environment. Several communist Intra-elite conflicts appear significant in a number transitions, including Hungary and Mongolia, also fit of nondistributive conflict cases. The 1989 transition in this pattern. Paraguay provides an illustration. The key decision was Two conclusions emerge from our discussion of a palace coup that ousted the aging dictator Alfredo democratic transitions. First, although certainly some Stroessner and initiated a process of constitutional re- democratic transitions are driven by distributive con- form and competitive presidential elections. The coup flict in ways that conform with the theory, these cases do was led by General Andres Rodriguez, Stroessner’s not appear to be related in any systematic way with the second in command, and by a faction of the ruling Col- level of inequality, as the lack of quantitative findings orado party that hoped to extend one-party rule by en- already suggests. Second, the assumption that elites gineering a “nonpersonalist” transition. Mass protest do not yield power in the absence of mass pressure did not pose a serious threat to the regime (Lambert from below is called into question by the high inci- 2000). dence of alternative transition paths. Taken together, In Mexico the ruling PRI was challenged primarily these conclusions indicate that the theory is, at best, by business elites and an opposition party (PAN) that underspecified and needs to delineate more explicitly was outside the regime and its ruling coalition and the conditions in which redistributive conflicts emerge. wanted less rather than more redistribution. Popular We return to these issues in the conclusion. protest over alleged fraud in local elections strength- ened the bargaining leverage of the PAN in its negoti- ations with the ruling party, but the political left played The Collapse of Democratic Rule: only a marginal role in pushing the regime out of power. Causal Process Observations Among other cases in which elite concessions to other elites appeared significant are the military’s acquies- Although Acemoglu and Robinson (2006) and Boix cence in the assumption of power by parliamentary (2003) offer diverging predictions about transitions politicians in Pakistan, the Thai military’s accommo- to democracy, they agree that when democracies do dation of emerging political-economic elites from the emerge at high levels of inequality, they are more Northern part of the country, and the Kuomintang’s likely to revert to authoritarian rule. As Acemoglu and accommodation of native Taiwanese elites. Robinson put it succinctly, “in democracy, the elites are Finally, in a number of cases incumbent authoritarian unhappy because of the high degree of redistribution elites opened politics under the assumption—justified and, in consequence, may undertake coups against the or mistaken—that they could effectively control the democratic regime” (222). This view comports with an earlier generation of theory on “bureaucratic authori- are entirely new countries. We chose to include them in the dataset tarian” installations in the Southern Cone (O’Donnell because they cross standard thresholds (Polity) or appear as new 1973; for critiques: see Collier 1979; Linz and Stepan democracies (CGV); see Haggard, Kaufman, and Teo (2012). 1978; Valenzuela 1978): Brazil (1964), Argentina (1966,
13 Inequality and Regime Change August 2012 and 1976), and Chile and Uruguay (both 1973), with other armed forces against democratic rule. We called extensions to other regions as well (for example, Im these nondistributive conflict cases “cross-class” and 1987 on Korea). “intra-elite” reversions, respectively.7 Unlike the quantitative evidence on transitions, Table 4 reports the incidence of distributive conflict there is somewhat stronger cross-national evidence and nondistributive conflict reversions by level of in- that inequality is incompatible with democratic sta- equality. The distributive conflict column aggregates bility (e.g., Dutt and Mitra 2008; Reenock, Bernhard, both elite-reaction and populist reversions; the nondis- and Sobek 2007). Houle (2009) deploys his innova- tributive conflict column aggregates both cross-class tive measure of capital share to test the relationship and intra-elite reversions. Table 5 shows the types of and finds that high return to capital relative to labor reversion, Polity scores of the deposed regimes, and significantly undermined democratic stability between economic circumstances surrounding the change. Al- 1960 and 2000. We replicated his model using the Gini though there is surprisingly little overlap between the and Vanhanen index of land inequality. Although land ranking of cases on the three measures of inequality, shows no effects, the Gini was a significant determinant reversions do cluster at the middle and high levels of of democratic breakdowns, both for the entire 1960– inequality; relatively few took place at the lowest levels. 2000 period and for the third wave between 1980 and However, we find only a minority of cases that con- 2000.6 form with the distributive conflict model. In the CGV Do these findings hold up when subjected to closer dataset, four cases (Bolivia 1980, Burundi 1996, Fiji qualitative scrutiny? To what extent do the causal pro- 2000, and Turkey 1980) or 21% of the cases are elite- cess observations comport with the expectations of reaction reversions. Three cases (16% of the sample)— distributive conflict theories? As with the transition Ecuador (2000), Ghana (1981), and Suriname (1980)— cases, we considered whether political pressures for are populist reversions. A substantial majority (63%) redistribution drove regime change, in this case the of the reversions are nondistributive. breakdown of democratic rule. We identified a category In the 20 reversions identified in the Polity measure called “elite-reaction” reversions that conform with (not shown here), 8 of the cases—40%—are classified the distributive conflict model. In these cases, elites as elite-reaction reversions,8 and there are 2 populist undermine democracy either by (a) seeking to oust reversions.9 Half the cases, however, are classified as incumbent governments that rely on the political sup- nondistributive. Missing data play more of a constraint port of lower class or excluded groups and are actively in allocating the Polity cases across levels of inequality, committed to the redistribution of assets and income but they are somewhat less concentrated at higher lev- or by (b) imposing restraints on political competition in els of inequality, and there is no evidence that higher order to prevent coalitions with explicitly redistributive inequality cases are more likely to be distributive. Four- aims from taking office. In these cases, distributive con- teen Polity reversions fall into the high-inequality ter- flicts are in evidence and elites are acting against gov- cile on one or more of the three measures of inequality; ernments, parties, and organized social forces that are if each case is counted only once, only five are distribu- actively committed to greater redistribution through tive conflict reversions.10 the democratic process. To elaborate the implications of these findings, we We also identified a second type of distributive con- focus on the high-inequality cases using the capital flict reversion in which the incumbent democratic gov- share measure; as the distribution of cases across dif- ernment is overthrown by authoritarian populist lead- ferent inequality terciles suggests, very similar results ers. These types of reversion do not comport with our would be obtained by using different income inequality expectation that reversions are driven by the right, but measures.11 According to distributive conflict models, they clearly involve redistributive conflict and are given these cases are most likely to revert as a result of elite some attention in Boix (2006, 18, 214–19). Whereas reactions to distributive demands from below. Given in elite-reaction reversions, challengers to democratic the low correlation between measures of inequality, rule appeal to elite interests and target the masses for alternative measures would show a different set of repression, in “populist reversions,” authoritarian chal- high-inequality cases. However, as can be seen from lengers appeal to the masses and target the elite. Table 4 no measure of inequality generates a distri- Finally, “nondistributive” reversions are unambigu- bution of reversions that conforms with theoretical ous instances of the null hypothesis, but we dis- expectations for a clustering of distributive conflict tinguished two alternative subtypes. In some cases, reversions among high-inequality cases; selection of support for a reversion cuts across class lines: Au- thoritarian challengers exploit wide disaffection with 7 Precise coding rules are available in Haggard, Kaufman, and Teo the performance of democratic incumbents and invoke (2012). broad valence issues, such as economic performance 8 Armenia (1995), Dominican Republic (1994), Fiji (1987 and 2000), and corruption, that cut across distributive cleavages. Haiti (1991), Turkey (1991), Ukraine (1993), and Zambia (1996). In other cases, purely intra-elite conflicts cause rever- 9 Ghana (1981) and Haiti (1999). sions . The military—or factions within it—might stage 10 Ghana (1981) was a populist reversion; Armenia (1995), the Do- a coup against incumbent office holders, or compet- minican Republic (1994), the Ukraine (1993) and Zambia (1991) were elite-reaction reversions. ing economic elites might mobilize military, militia, or 11 One case, Sierra Leone, is identified as high inequality on the capshare measure, but was not included by Houle (2009) in the 6 Results available on request from the authors. regressions because of a subsequent change in coding of the case.
14 American Political Science Review ty (1982) (1997) Comoros (1995) Guatemala Suriname (1990) Congo (1997) Niger (1996) Nigeria (1983) Sierra Leone Thailand (1991) Uganda(1985) P P E P E E E Populist) Non-distributive Distributive Elite/ ( Turkey (1980) (1997) Pakistan (1999) Ecuador (2000) Congo (1997) Bolivia (1980) Peru (1990) Fiji (2000) Guatemala (1982) Peru (1990) Suriname (1990) Pakistan (1999) Thailand (1991) Sudan (1989) Sierra Leone Uganda (1985) P P E P E E Gini coefficient E Inequality Measures Populist) Nondistributive (Texas Inequality dataset) Share of Family Farms (Vanhanen) Distributive Elite/ ( Ghana (1981) Suriname (1980) (1997) Comoros (1995) Comoros (1995) Suriname (1980) Guatemala (1982) Ecuador (2000) Niger (1996) Nigeria (1983) Burundi (1996) Pakistan (1999) Fiji (2000) Nigeria (1983) Bolivia (1980) Peru (1990) Burundi (1996) Sierra Leone Congo (1997)Suriname (1990) Sudan Niger (1989) (1996) Thailand (1991) Sudan (1989) Turkey(1980) Uganda (1985) Ghana (1981) P P E P E E E Populist) Nondistributive Capital Share of Income in Distributive Manufacturing Sector (capshare) Elite/ ( Bolivia (1980) Suriname (1980) Ecuador (2000) Fiji (2000) Burundi (1996) Turkey (1980) Ghana (1981) : Reversions: Cheibub, Ghandi, and Vreeland (2010); reversion types: Haggard, Kaufman, and Teo (2012); capital share: Houle (2009); Gini: Universi Data of Texas Inequality Project (2008); family farms: Vanhanen (2003). TABLE 4. Distributive and Nondistributive Reversions by Level of Inequality, 1980–2000 Level of Inequality High Missing Medium Low Sources
15 Inequality and Regime Change August 2012
TABLE 5. Distributive and Nondistributive CGV Reversions, Polity Scores, Prior Coups, Per Capita GDP, and GDP Growth, 1980–2000
Distributive Nondistributive Country/Year Reversions Reversions Polity Score Prior Coups GDP/ capita ($) Growth
Bolivia (1980) E −4 3 1070 −1.4 Burundi (1996) E 0 2 113 −8.0 Comoros (1995) X 4 1 386 3.6 Congo (1997) X 5 0 104 −5.6 Ecuador (2000) P 9 0 1295 2.8 Fiji (2000) E 6 1 2075 −1.7 Ghana (1981) P 6 4 224 −3.5 Guatemala (1982) X −5 1 1556 −3.5 Niger (1996) X 8 1 168 3.4 Nigeria (1983) X 7 2 319 −5.3 Pakistan (1999) X 7 0 526 3.7 Peru (1990) X 7 0 1657 −5.1 Sierra Leone (1997) X 4 3 168 −16.7 Sudan (1989) X 7 2 282 8.9 Suriname (1980) P − - 2536 −5.3 Suriname (1990) X − - 2049 −0.5 Thailand (1991) X 3 1 1500 8.6 Turkey (1980) E 9 1 2427 −2.4 Uganda (1985) X 3 2 170 −3.3 % or average 36.8% 63.2% 4.5 1.4 980 −1.6
Notes and sources: E, elite reversion; P, populist reversion. Polity scores are the country’s score the year preceding the reversion (Marshall, Gurr, and Jaggers 2010). Prior coups is the number of coups (excluding attempted coups or plots) in the 10 years prior to the reversion (Marshall and Marshall 2010; McGowan 2007). GDP/capita and GDP growth refer to the values of these variables in the year of the reversion (World Bank 2010). Reversion types are from Haggard, Kaufman, and Teo (2012). cases based on a different inequality indicator would and clearly less well-off majority, producing a highly therefore yield similar results. fraught political environment (Lemarchand 1996). Be- tween 1966 and 1996, the country experienced no fewer than 11 coups and attempted coups (McGowan 2007), Distributive Conflict I: with periodic episodes of wider violence. The deposed Elite Reactions in Bolivia and Burundi democratic government was led by moderate Hutu Bolivia and Burundi are the only high-inequality cases politician Melchior Ndadaye, but was extremely frag- to revert to authoritarian rule through the causal pro- ile; the coding of the transition to democracy is 1993 cess stipulated by the theory. In Bolivia, a right-wing is dubious. Ndadave died in an unsuccessful coup at- military faction led by General Luis Garcia Meza tempt in 1994, and his successor, Cyprien Ntaryamira, deposed acting President Lidia Gueiler on July 17, was killed in a suspicious plane crash in the same year. 1980, following the victory of leftist Hernan Siles in After a massacre of more than three hundred Tutsis by an election held earlier that year. The coup occurred radical Hutu rebels in 1996, a military coup by former in the context of severe, ongoing conflicts between president Pierre Beyoya restored the Tutsis to power. militant miners’ unions and more conservative polit- ical and economic forces after the breakdown of the Distributive Conflict II: long-standing Banzer dictatorship in 1978. The Meza Populist Reversion in Ghana dictatorship was in turn ousted only two years later by working-class protests that forced new elections. Jerry Rawlings’ coup in Ghana constitutes a clear ex- Significantly, severe distributive conflicts continued to ample of a populist reversion, although once in office threaten the stability of the new democratic regime his military government shifted sharply to the right. In and ended only in 1985, when the elected government 1981, Rawlings overthrew the feckless constitutional harshly repressed union opposition and implemented government of Hilla Limann with the backing of mili- an aggressive structural adjustment program. tant student organizations, unions, and left social move- In Burundi, inequality is by no means correctly cap- ments. By the time of the coup, the economy had deteri- tured by the capshare or other inequality measures; orated badly, and the Limann government faced strikes much more significant are the deep ethnic cleavages and confrontations with workers over back pay and a that divide the country. A Tutsi minority (about 15% of tough austerity program. On seizing power, Rawlings the population) had long dominated the military, civil actively solicited the support of these forces by plac- service, and the economy. Hutus constituted a large ing representatives of radical left organizations on the
16 American Political Science Review military’s Provisional National Defense Council and Nigeria. As in Peru, the 1983 coup in Nigeria oc- creating a raft of populist consultative organizations curred in the context of severe economic deteriora- (Graham 1985; Hutchful 1997). Rawlings’ populism tion and a widespread loss of public confidence in the only aggravated Ghana’s economic problems, and the government. The leader of the coup, Major General military regime ultimately reversed course entirely and Muhaamadu Buhari, was—like his predecessors—tied vigorously embraced the “Washington consensus.” Yet closely to the Muslim north and had held a high po- the initial overthrow of the democratic regime clearly sition within the deposed government. Yet there are appealed to, and mobilized support from, populist and no indications that the takeover was motivated by leftist social forces. class or ethnic demands on the state, nor by the sig- nificant involvement of civil society. Nor is there evi- dence that factional rivalries within the military were The Null Cases: Nondistributive Reversions connected with broader social conflicts that could be modeled in elite-mass terms, whether engaging class, The other high-inequality reversions are Peru, Nigeria, ethnic, or regional interests. The most consequential Thailand, and Sierra Leone. Some of these involved divisions were within the elites, most notably, the mil- broad appeals that cut across class lines, whereas oth- itary, clientelistic politicians, and the business class. ers resulted primarily from conflicts within the elite When oil revenues collapsed, the ruling coalition frag- itself. However, in none of the cases were redistribu- mented under competing claims for patronage. The tive cleavages between elites and masses central to inability of the hegemonic party to reconcile these the reversion, and in several the specific political pres- conflicting interests, argues Augustine Udo (1985, 337), sures stipulated by the theory—redistributive demo- came to a head in a blatantly corrupt election in 1983 cratic governments or social movements—were alto- that exposed “unprecedented corruption, intimidation, gether absent. and flagrant abuse of electoral privilege by all par- ties.” The coup was a response to these democratic Peru. Alberto Fujimori’s decision to close congress failures. and rule by decree in April 1992 drew support from a broad cross-section of Peruvian society. Military back- Thailand 1991. The 1991 coup in Thailand was un- ing was, of course, essential and was motivated in part dertaken by a military faction that bridled under both by the desire for a free hand to confront the Shining the existing military leadership and the efforts of the Path, an insurgency that had pretenses of representing elected assembly to exercise greater control over mili- disadvantaged peasants in some highland areas of the tary spending and prerogatives (Baker and Phongpai- country. Yet in other important ways, the case does not chit 2002). Elected officials were concerned, among correspond with the theory. First, the “self-coup” ini- other things, with channeling patronage resources to tially met opposition from international and some local disadvantaged parts of the country, but they were business sectors—in short, from economic elites—who linked closely to upcountry business interests. Al- were concerned that an outright dictatorship would though the distribution of income had deteriorated in have adverse economic consequences. Although these Thailand during the economic reforms of the 1980s, left sectors eventually warmed to the regime after Fujimori parties remained confined to the fringes of political life, agreed to a fac¸ade of constitutionalism, they were by and a long-standing rural insurgency had long since pe- no means drivers or even supporters of the coup. tered out. The coup had the effect of galvanizing mass At the same time, Fujimori enjoyed surprisingly wide opposition, including groups explicitly representing the popular support, visible in his overwhelming victory in poor, and this opposition subsequently played a role an early referendum on a new constitution that would in the transition back to democratic rule. Yet there is cement his hold on power. The unions and the political no evidence that the coup either responded to popu- left did oppose the coup, but their organizations had lar pressures for redistribution or reflected populist- been decimated by the hyperinflation and economic authoritarian dissatisfaction with democracy’s failure collapse of the late 1980s, and they themselves enjoyed to redress redistributive grievances. little popular support. The large majority of the Peru- vian poor were attracted by a leader who promised to Weak Democracy Syndrome deal with a strong hand with the economic crisis and the insurgency. One 1992 survey showed that almost 76% We do not seek to elaborate an alternative theory of of low-income people supported Fujimori’s plan for democratic instability during the third wave, but our constitutional reform (Rubio 1992, 7; cited in Weyland analysis suggests a “weak democracy” syndrome that 1996, fn 16). While undertaking economic reforms, Fu- comports with a growing body of literature on demo- jimori also strengthened his electoral base through the cratic vulnerability (Diamond 2008; Levitsky and Way expansion of clientelistic antipoverty programs (Wey- 2010). Before turning to this issue, however, we should land 1996). In the late 1990s, as the economy once underscore that at least some of the reversions may be again slowed and corruption scandals surfaced, Fuji- artifacts of coding rules governing these two influential mori’s popularity waned, and he was eventually forced datasets. Table 5 shows that 8 of the 19 cases coded as to withdraw from power. Until that time, however, his reversions in the CGV dataset did not rise above the government rested on a surprisingly broad cross-class standard Polity cutoff score of 6 in the year preced- coalition. ing their collapse (Bolivia, Burundi, Comoros, Congo,
17 Inequality and Regime Change August 2012
Guatemala, Sierra Leone, Thailand, and Uganda). An- for challengers to act with the acquiescence or even other six cases (Fiji, Ghana, Nigeria, Pakistan, Peru, support from disaffected publics. and Sudan) barely make that threshold with scores of In sum, a close examination of the causal mech- either 6 or 7. The average Polity score for all of the anisms driving reversal during the third wave sug- CGV reversion countries in the year preceding the gests a more deep-seated syndrome in which distribu- collapse of democratic rule is only 4.5. “Reversions” tive conflict plays a surprisingly minor role. Struc- are occurring against democracies that are marginally tural constraints such as low per capita income and democratic at best. weak institutions combined with short-run crises seem Yet the weakness of the distributive conflict theory to be major factors in the breakdown of these weak of regime change is not simply an artifact of the coding democracies. rules; the causal mechanisms stipulated in the theory do not appear to operate either. Electoral competition in Thailand, Nigeria, Pakistan, Honduras, Ecuador, CONCLUSION Ghana, and Guatemala was dominated by patronage parties with close ties to economic elites or the military Viewed over the long run, the emergence of democracy establishment. In none of these cases do we see a sig- in the advanced industrial states resulted in part from nificant presence of parties, interest groups, or social fundamental changes in class structures. Demands on movements representing the interests of the poor that the state from new social classes—first the emergent could serve as the basis for distributive conflict that bourgeoisie and then the urban working class—played would in turn trigger elite intervention. a role in the gradual extension of the franchise. These Rather, conflicts within the political elite—between stylized facts played an important role in the new dis- ins and outs—was more likely to pose a challenge to tributive conflict models of regime change. democratic rule, with the military playing a pivotal Yet these models do not appear to travel well to role. In the 11 cases in which distributive conflicts the very different international, political, and socioeco- were implicated in the collapse of democratic rule, nomic conditions that prevailed during the third wave the military could plausibly be seen as an agent of ei- of democratization. Standard panel designs have found ther elites (elite-reaction reversions) or excluded social at best limited evidence for the inequality-transition forces (populist reversions). However, in many of the logic of the distributive conflict models, and the causal other cases, the military entered politics largely on its process observations reported here show that it does own behalf. Such intervention was more likely to oc- not appear to operate even in cases in which it should. cur where prior military intervention had established Although more refined measures of inequality may ul- a precedent. Cross-national quantitative work on both timately capture ethnic or regional inequalities that Latin America and Africa finds that the likelihood of we are underestimating, our causal process observa- a military coup is strongly affected by the previous his- tions are designed to capture at least the overt political tory of coups (Collier and Hoeffler 2005; Lehoucq and manifestations of a wide array of different distributive Perez-Linan 2009). The data presented in Table 5 are cleavages. It therefore seems likely that the problems consistent with these findings. Thirteen of 19 reversions lie with theory as well as measurement. came in countries that had already experienced at least How should we respond to such findings? Distribu- one prior coup, and in 7 of these cases, the military was tive conflict theories may simply be weaker than their a repeat offender. proponents suggest, and we later highlight several al- The data in Table 4 also highlight the poverty and ternative approaches to regime change. However, the poor economic performance of the countries experi- core insight of distributive conflict theories is intuitively encing reversion. As Londregan and Poole (1990) and appealing, and we are inclined to look for avenues for Przeworski et al. (2000) have shown convincingly, the refinement. One avenue would be to consider whether probability that democratic governments will survive is inequality influences the stability of democratic rule strongly affected by the level of development. Average through channels other than those postulated by the GDP per capita for the reversion cases at the time of the distributive conflict theorists. High inequality may be collapse of democratic rule was only $980, way below a determinant of the “weak democracy” syndrome, for the thresholds for consolidated democracies. Among example by contributing to low growth and poverty the non-African cases, only Thailand, Ecuador, and (Persson and Tabellini 1994), which are in turn re- Peru are middle-income countries. lated to political instability and weak, ineffective states The relationship between short-run economic per- (Londregan and Poole 1990). This causal path may formance and reversions has also been explored in well help explain an important class of low-income some detail (Gasiorowsksi 1995; Haggard and Kauf- cases, as we argued in the conclusion to our discus- man 1995; Kricheli and Livne 2011; Teorell 2010). Prze- sion of reversions; we return to this group of countries worski et al. (2000) show that the odds of democratic later. survival decrease substantially after three consecutive Yet the collective protest of citizens against elites is a years of negative economic growth. On average, the core causal mechanism in distributive conflict theories, economies of the reversion countries declined by 1.6% and such an approach would abandon that insight alto- in the year of the reversion, and a number were in the gether. The incentives and capacity to mobilize such midst of full-blown economic crises (Table 5). Both low protest are central to the theory, yet are either as- per capita income and slow growth provided openings sumed to be a function of levels of inequality or ignored
18 American Political Science Review altogether.12 The free-rider problem highlighted by Ol- Walle 1997). Both transitions and reversion in these son decades ago (1965) problematizes the assumption cases often came at best in response to generalized that shared interests in redistribution will enable large protest against poor economic conditions waged by groups to overcome barriers to collective action. The relatively better-off urban forces, many with close ties question of how to solve this problem has become to the state apparatus. the cornerstone of the literature not only on regime Institutional approaches represent another point of change but also on revolution, collective violence, and departure for explaining collective action. Prior experi- contentious politics. ence with democracy or institutionalized opportunities In the absence of a capacity to overcome barriers to for collective action in semi-authoritarian regimes may collective action, transitions both to and from demo- be important for understanding how collective chal- cratic rule are more likely to reflect narrow, intra-elite lenges are subsequently mobilized. In Latin America, conflicts. Although such conflicts certainly have a dis- corporatist unions, which had initially been financed tributive component—and indeed a highly conflictual and sponsored by the state (Collier and Collier 1991; one—it is harder to root them in the class-conflict logic Schmitter 1974), subsequently formed a core compo- of the underlying Meltzer-Richard model. Although nent of protests against authoritarian incumbents in we found a surprising number of distributive conflict Peru, Bolivia, and Argentina, as did a number of labor- transitions in the high-inequality cases, the concentra- based political parties with close links to the state. tion of income and assets in such settings may also At a more general level, differences in authoritarian empower elites to shape the course of regime change; institutions might shape both the actors and cleavages the effects of transitions on the distribution of income that lead to the establishment or reversal of democracy. could as well be regressive as progressive. Collier (1999) has shown this empirically with earlier We suspect that distributive conflict theories may democratic transitions in Europe and Latin America, ultimately prove to be conditional in form; that is, and an exploding literature on varieties of authori- they are dependent on incentives and capacities for tarian rule raises the possibility for the postwar pe- collective action that are not in fact given by the level riod as well (Geddes 1999; Levitsky and Way 2010; of inequality. What are these additional factors that Magaloni and Kricheli 2010). The effects of the organi- might enable subaltern groups to overcome barriers zational spaces provided by such regimes on collective to collective action? We identify at least three lines of demands for democracy remain a subject of ongoing research, each of which is potentially complementary research. For instance, controlled competition under to the political economy approaches discussed in this semi-competitive regimes might provide opportunities article but may also represent competing approaches for mobilization that subsequently spill over into chal- to regime change. lenges to the regime itself. Yet it is also possible that It may not be necessary to reach beyond a political- controlled opening may yield advantages for incum- economy framework to clarify conditions in which dis- bents by establishing organized channels for recruit- tributive conflict becomes more likely to affect regime ment of supporters, opportunities for control, and the change. One such condition is economic development. revelation of politically useful information, such as the At various points, both Boix (2003) and Acemoglu identity and strength of the opposition. and Robinson (2006) suggest that capacities for col- Finally, the social movement and “contentious pol- lective action are likely to be greater in relatively de- itics” literature provides the starkest alternative to veloped countries where industrialization and urban- political-economy approaches. Work in this area em- ization provide a social basis for organization. Our phasizes the significance of political opportunities, re- case studies also suggest a contrast between middle- sources, and cultural framing, typically casting the ap- income countries with substantial concentrations of in- proach in opposition both to strictly rationalist expla- dustrial labor and poorer countries where low-income nations for collective action and theories that stress groups are concentrated in the agricultural and urban underlying structural conditions such as inequality (see informal sectors and face greater barriers to collec- McAdam, Tarrow, and Tilly 2003). As noted earlier, tive action. In relatively industrialized countries such much of the analysis of nationalist and ethnic move- as Argentina, Brazil, Poland, South Africa, and South ments in Eastern Europe and the former Soviet Union Korea, distributive conflict transitions involved or were focuses on the factors emphasized in the contentious even led by workers’ movements with a relatively long politics literature. Kubik (1994), for example, provides history of political mobilization and collective action an important account of how the Solidarity movement (Collier 1999; Drake 1998). Conversely, in a number of mobilized around a protest discourse that emerged in the poorer African transitions we examined, political the wake of Pope John Paul II’s return to Poland in parties and civil society groups representing the poor 1979. More broadly, Beissinger’s (2002) seminal work were often too weak to check the predatory tenden- on anti-regime protest in the Soviet Union emphasizes cies of state elites and of other more privileged social nationalism and ethnic identities, rather than socioe- forces; as a result regime change was better understood conomic grievances, as the principal spur to protest in terms of intra-elite processes (Bratton and van de against Soviet authority. As we argued in our discussion of the coding, such protest can be viewed as a reaction 12 See Green and Shapiro (1994) for a general critique of rational against other forms of inequality. Yet it can also be choice theory and its inability to deal persuasively with collective viewed as an alternative to the structural and rationalist action problems. foundations of distributive conflict approaches.
19 Inequality and Regime Change August 2012
The relationship between these analyses of the as regime change, revolution, financial crisis, war, or sources of collective action and distributive conflict the- famine—and thus amenable to intensive qualitative ories is not straightforward. Economic development, scrutiny. Such an approach combines within-case anal- political institutions, and even the dynamics of con- ysis that is sensitive to context and sequence with tests tentious politics may simply mediate the effects of of the underlying theories and causal mechanisms that inequality emphasized in the distributive conflict ap- are often only implicit in larger-N designs. However, proaches. However, given the agnostic nature of our this approach rests on a willingness to open up exist- findings, they might also prove to be contending ex- ing datasets and recode them in line with theoretical planations that move away from an emphasis on un- expectations to maximize inferential leverage. derlying inequalities altogether. An important line of research is to exploit opportunities to distinguish be- tween competing theories that may appear observa- REFERENCES tionally equivalent. A final theoretical note concerns an important con- Acemoglu, Daron, and James A. Robinson. 2000. “Democratization ditioning factor in our own analysis, which is limited to or Repression?” European Economics Review 44 (4–6): 683–93. Acemoglu, Daron, and James A. Robinson. 2001. “A Theory of the third wave era. As noted in the introduction, this Political Transitions.” American Economic Review 91(4): 938–63. period encompasses important international systemic Acemoglu, Daron, and James A. Robinson. 2006. Economic Origins changes, and a number of our cases had a significant of Dictatorship and Democracy. New York: Cambridge University international dimension. The winding down of the Cold Press. Ansell, Ben, and David Samuels. 2010. “Inequality and Democrati- War eliminated opportunities for authoritarian rulers zation: A Contractarian Approach.” Comparative Political Studies to secure support from great power patrons and thus 43 (12): 1543–74. altered the resources and calculus of domestic politi- Baker, Chris, and Pasuk Phongpaichit. 2002. Thailand: Economy and cal actors, including disadvantaged groups. With more Politics. 2nd ed. Oxford: Oxford University Press. permissive international conditions, democracy could Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosen- thal, and Barry R. Weingast. 1998. Analytic Narratives. Princeton, spread to countries with vastly different class structures NJ: Princeton University Press. and degrees of inequality. Beck, Nathaniel. 2006. “Is Causal-process Observation an Oxy- Boix (2011) has recently explored the mediating moron?” Political Analysis 14 (3): 347–52. effects of the international order and finds that eco- Beck, Nathaniel. 2010. “Causal Process ‘Observation’: Oxymoron or (Fine) Old Wine.” Political Analysis 18 (4): 499–505. nomic development is more likely to drive democrati- Beissinger, Mark 2002. Nationalist Mobilization and the Collapse of zation when the international order is dominated by a the Soviet State. Cambridge: Cambridge University Press. democratic hegemon rather than by rivalries between Boix, Carles. 2003. Democracy and Redistribution. New York: Cam- democratic and authoritarian powers. Against this ar- bridge University Press. gument, our examination of democratization during Boix, Carles. 2008. “Economic Roots and Civil Wars and Revolu- tions in the Contemporary World.” World Politics 60 (3): 390– this period emphasizes the spread of democracy—at 437. least temporarily—to poorer countries as well, partly Boix, Carles 2011. “Democracy, Development, and the International as a consequence of pressure and encouragement from System.” American Political Science Review 105 (4): 809–28. international donors. However, we also found that it Brady, Henry E., and David Collier. 2004. Rethinking Social Inquiry, Diverse Tools, Shared Standards. Lanham, MD: Rowman and Lit- was precisely in the poorer countries where transitions tlefield. facilitated by positive international conditions were Bratton, Michael, and Nicolas van de Walle. 1997. Democratic Exper- most likely to be reversed, suggesting that the longer iments in Africa: Regime Transitions in Comparative Perspective. run structural forces outlined earlier are likely at work. New York: Cambridge University Press. Distributive conflict models have formalized and Burkhart, Ross E. 1997. “Comparative Democracy and Income Dis- tribution: Shape and Direction of the Causal Arrow.” Journal of thus given a sharper edge to class-conflict models of Politics 59 (1): 148–64. political change that have long been a part of the Cheibub, Jose Antonio, and Jennifer Gandhi. 2004. “Classifying Po- social science canon. They have done so in part by litical Regimes: A Six-fold Measure of Democracies and Dictator- etching more sharply the causal mechanisms that drive ships.” Presented at the Annual Meeting of the American Political Science Association, Chicago. regime change. Yet in doing so, it is incumbent on these Cheibub, Jose Antonio, Jennifer Gandhi, and James Raymond theories to provide compelling empirical tests of their Vreeland. 2010. “Democracy and Dictatorship Revisited.” Public claims, not only with respect to the relationship be- Choice 143 (1/2): 67–101. tween inequality and regime change but with respect Collier, David. 1979. The New Authoritarianism in Latin America. to the postulated causal processes as well. Although an Princeton, NJ: Princeton University Press. Collier, David, Henry E. Brady, and Jason Seawright. 2010. “Sources earlier generation of case study scholarship on demo- of Leverage in Causal Inference: Toward an Alternative View of cratic transitions faced significant selection problems, Methodology.” In Rethinking Social Inquiry: Diverse Tools, Shared large-N cross-national designs have their own disabil- Standards, eds. Henry E. Brady and David Collier. Lanham, MD: ities related to measurement error and reduced-form Rowman and Littlefield, 161–200. Collier, Paul, and Anke Hoeffler. 2005. “Resource Rents, Gover- designs that do not actually test for stipulated causal nance, and Conflict.” Journal of Conflict Resolution 49 (4): 625–33. mechanisms. Collier, Ruth Berins. 1999. Paths toward Democracy; Working Class The tradeoffs between small- and large-N designs are and Elites in Western Europe and South America. Cambridge: real. Yet medium-N designs based on causal process Cambridge University Press. observations provide an additional means of squar- Collier, Ruth Berins, and David Collier. 1991. Shaping the Politi- cal Arena: Critical Junctures, the Labor Movement, and Regime ing the methodological circle, particularly where the Dynamics in Latin America. Princeton, NJ: Princeton University phenomena being explained are relatively rare—such Press.
20 American Political Science Review
Diamond, Larry. 2008. The Spirit of Democracy: The Struggle to Kricheli, Ruth, and Yair Livne. 2011. “Mass Revolutions vs. Build Free Societies. New York: Times Books. Elite Coups.” Presented at the Annual Meeting of the Amer- Drake, Paul W. 1998. Labor Movements and Dictatorships: The ican Political Science Association, Toronto. http://ssrn.com/ Southern Cone in Comparative Perspective. Baltimore: Johns Hop- abstract=1449852 (accessed June 1, 2012). kins University Press. Kubik, Jan 1994. The Power of Symbols against the Symbols of Power: Dutt, Pushan, and Devashish Mitra. 2008. “Inequality and the Insta- The Rise of Solidarity and the Fall of State Socialism in Poland. bility of Polity and Policy.” Economic Journal 118 (531): 1285–1314. University Park: Pennsylvania University Press. Epstein, David L., Robert Bates, Jack Goldstone, Ida Kristensen, Lambert, Peter. 2000. “A Decade of Electoral Democracy: Continu- and Sharyn O’Halloran. 2006. “Democratic Transitions.” Ameri- ity, Change, and Crisis in Paraguay.” Bulletin of Latin American can Journal of Political Science 50 (3): 551–69. Research 19 (3): 379–96. Falleti, Tulia G., and Julia F. Lynch. 2010. “Context and Causal Mech- Lehoucq, Fabrice, and Anibal Perez-Linan. 2009. “Regimes, anisms in Political Analysis.” Comparative Political Studies 42 (9): Competition, and Military Coups in Latin America.” Pre- 1143–66. sented at the Annual Meeting of the American Political Fearon, James D., and David D. Laitin. 2008. “Integrating Quantita- Science Association, Toronto. http://papers.ssrn.com/sol3/papers. tive and Qualitative Methods.” In Oxford Handbook of Political cfm?abstract_id=1449035 (accessed June 1, 2012). Methodology, eds. Janet M. Box-Steffensmeier, Henry E. Brady, Lemarchand, Renee. 1996. Burundi: Ethnic Conflict and Genocide. and David Collier. Oxford: Oxford University Press, 756—78. Washington, DC: Woodrow Wilson Press. Freeman, John R., and Dennis P. Quinn. 2012. “The Economic Ori- Levitsky, Steven, and Lucan A. Way. 2010. Competitive Authoritari- gins of Democracy Reconsidered.” American Political Science Re- anism: Hybrid Regimes after the Cold War. New York: Cambridge view 106 (1): 58–80. University Press. Gasiorowski, Mark J. 1995. “Economic Crisis and Political Regime Linz, Juan, and Alfred Stepan. 1978. The Breakdown of Democratic Change: An Event History Analysis.” American Political Science Regimes. Baltimore: Johns Hopkins University Press. Review 89 (4): 882–97. Linz, Juan, and Alfred Stepan. 1996. Problems of Democratic Tran- Geddes, Barbara 1999. “What Do We Know about Democratic Tran- sition and Consolidation. Baltimore: Johns Hopkins University sitions after Twenty Years?” Annual Review of Political Science 2: Press. 115–44. Lipset, Seymour M. 1960. Political Man: The Social Bases of Politics. George, Alexander L., and Andrew Bennett. 2005. Case Studies and Baltimore: Johns Hopkins University Press. Theory Development in the Social Sciences. Cambridge, MA: MIT Londregan, John B., and Keith T. Poole. 1990. “Poverty, the Coup Press. Trap, and the Seizure of Executive Power.” World Politics 42 (2): George, Alexander L., and Timothy J. McKeown. 1985. “Case Stud- 151–83. ies and Theories of Organizational Decision Making.” Advances Magaloni, Beatriz, and Ruth Kricheli. 2010. “Political Order and in Information Processing in Organizations 2: 21–58. One-party Rule.” Annual Review of Political Science 13: 123–43. Gerring, John. 2006. “Single-outcome Studies: A Methodological Mainwaring, Scott, Daniel Brinks, and Anibal Perez-Linan. 2001. Primer.” International Sociology 21 (5): 707–34. “Classifying Political Regimes in Latin America, 1945–1999.” Gerring, John. 2007a. Case Study Research: Principles and Practices. Studies in Comparative International Development 36 (1): 37–65. New York: Cambridge University Press. Marshall, Thomas Humphrey. 1963. Class, Citizenship, and Social Gerring, John. 2007b. “The Mechanismic Worldview: Thinking inside Development: Essays. Chicago: University of Chicago Press. the Box.” British Journal of Political Science 38 (1): 161–79. Marshall, Monty G., Ted Robert Gurr, and Keith Jaggers. 2010. Gerring, John. 2010. “Causal Mechanisms: Yes, but....” Comparative Polity IV Project: Political Regime Characteristics and Tran- Political Studies 42 (3): 327–59. sitions, 1800–2009. Vienna, VA: Center for Systemic Peace. Gervais, Myriam. 1997. “Niger: Regime Change, Economic Crisis http://www.systemicpeace.org/polity/polity4.htm (accessed June 1, and the Perpetuation of Privilege.” In Political Reform in Franco- 2012). phone Africa, eds. John F. Clark and David Gardinier. Boulder, Marshall, Monty G., and Donna Ramsey Marshall. 2010. Coup d’etat´ CO: Westview Press, 86–108. Events, 1946–2009. Vienna, VA: Center for Systemic Peace. Graham, Yao. 1985. “The Politics of Crisis in Ghana: Class Struggle McAdam, Doug, Sidney Tarrow, and Charles Tilly 2003. Dynamics and Organization, 1981–1984.” Review of African Political Econ- of Contention. Cambridge: Cambridge University Press. omy 12 (34): 54–68. McGowan, Patrick K. 2007. “African Military Intervention Events, Green, Donald P., and Ian Shapiro. 1994. Pathologies of Rational January 1, 1955, to December 31, 2006.” Arizona State University. Choice Theory. New Haven, CT: Yale University Press. Dataset. Hafner-Burton, Emilie, and James Ron. 2009. “Seeing Double: Hu- Meltzer, Allan H., and Scott F. Richard. 1981. “A Rational Theory man Rights Impact through Qualitative and Quantitative Eyes?” of the Size of Government.” Journal of Political Economy 89 (5): World Politics 61 (2): 360–401. 917–27. Haggard, Stephan, and Robert R. Kaufman. 1995. The Political Moore, Barrington, Jr. 1966. Social Origins of Dictatorship and Economy of Democratic Transitions. Princeton, NJ: Princeton Democracy. Boston: Beacon Press. University Press. O’Donnell, Guillermo A. 1973. Modernization and Bureaucratic- Haggard, Stephan, Robert R. Kaufman, and TerenceTeo. 2012. “Dis- authoritarianism: Studies in South American Politics. Berkeley: tributive Conflict and Regime Change: A Qualitative Dataset.” University of California Press. http://hdl.handle.net/1902.1/18276 V1 (accessed June 1, 2012). O’Donnell, Guillermo A., Philippe Schmitter, and Laurence Hedstrom, Peter, and Petri Ylikoski. 2010. “Causal Mechanisms in Whitehead, eds. 1986. Transitions from Authoritarian Rule: Tenta- the Social Sciences.” Annual Review of Sociology 36: 49–67. tive Conclusions about Uncertain Democracies. Baltimore: Johns Houle, Christian 2009. “Inequality and Democracy: Why Inequal- Hopkins University Press. ity Harms Consolidation but Does Not Affect Democratization,” Olson, Mancur. 1965. The Logic of Collective Action: Public Goods World Politics 61 (4): 589–622. and the Theory of Groups. Cambridge, MA: Harvard University Huntington, Samuel P.1991. The Third Wave: Democratization in the Press. Late Twentieth Century. Norman: University of Oklahoma Press. Persson, Torsten, and Guido Tabellini. 1994. “Is Inequality Harmful Hutchful, Eboe. 1997. “Military Policy and Reform in Ghana.” Jour- for Economic Growth?” American Economic Review 84: 600–21. nal of Modern African Studies 35 (2): 251–78. Pierson, Paul. 2004. Politics in Time: History, Institutions, and Social Im, Hyung Baeg. 1987. “The Rise of Bureaucratic Authoritarianism Analysis. Princeton, NJ: Princeton University Press. in South Korea.” World Politics 39 (2): 231–57. Przeworski, Adam. 2009. “Conquered or Granted? A History of King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Suffrage Extensions.” British Journal of Political Science 39 (2): Social Inquiry: Scientific Inference in Qualitative Research.Prince- 291–321. ton, NJ: Princeton University Press. Przeworski, Adam, Michael Alvarez, Jose Antonio Cheibub, and Kitschelt, Herbert, and Steven Wilkinson, eds. 2006. Patrons or Poli- Fernando Limongi. 2000. Democracy and Development: Political cies? Patterns of Democratic Accountability and Political Compe- Institutions and Well-being in the World, 1950–1990. Cambridge: tition. New York: Cambridge University Press. Cambridge University Press.
21 Inequality and Regime Change August 2012
Reenock, Christopher, Michael Bernhard, and David Sobek. 2007. Udo, Augustine. 1985. “Class, Party Politics, and the 1983 Coup in “Regressive Socioeconomic Distribution and Democratic Sur- Nigeria.” Africa Spectrum 20 (3): 327–38. vival.” International Studies Quarterly 51 (3): 677–99. University of Texas Inequality Project. 2008. Estimated House- Roberts, Kevin W. S. 1977. “Voting over Income Tax Schedules.” hold Income Inequality (EHII) Dataset. http://utip.gov.utexas. Journal of Public Economics 8 (3): 329–40. edu/data.html (accessed June 1, 2012). Romer, Thomas. 1975. “Individual Welfare, Majority Voting, and the Valenzuela, Arturo 1978. The Breakdown of Democratic Regimes: Properties of a Linear Income Tax.” Journal of Public Economics Chile. Baltimore: Johns Hopkins University Press. 4 (2): 163–85. Vanhanen, Tatu. 2003. “Democratization and Power Resources Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John 1850–2000.” FSD1216, version 1.0 (2003-03-10). University Stephens. 1992. Capitalist Development and Democracy. Chicago: of Tampere, Department of Political Science. Finnish So- University of Chicago Press. cial Science Data Archive. http://www.fsd.uta.fi/english/ Schmitter, Philippe C. 1974. “Still the Century of Corporatism?” data/catalogue/FSD1216/meF1216e.html (accessed June 1, Review of Politics 36 (1): 85–131. 2012). Slater, Dan, and Benjamin Smith. 2012. “Economic Origins of Demo- Weyland, Kurt. 1996. “Neoliberalism and Neopopulism in Latin cratic Breakdown? The Redistributive Model and the Postcolonial America: Unexpected Affinities.” Studies in Comparative Inter- State.” University of Chicago. Unpublished manuscript. national Development 31 (3): 3–31. Stepan, Alfred 2001. Arguing Comparative Politics. Oxford: Oxford Whitehead, Lawrence, ed. 1996. The International Dimensions of University Press. Democratization: Europe and the Americas. Oxford: Oxford Uni- Teorell, Jan. 2010. Determinants of Democratization: Explaining versity Press. Regime Change in the World, 1972–2006. Cambridge: Cambridge World Bank. 2010. World Development Indicators 2010. Washington University Press. DC: World Bank.
22 The American Political Science Association APSA Volume 11, No. 3 Comparative Democratization October 2013 I! T"#$ I$$%& CD INEQUALITY AND REGIME CHANGE: THE ROLE OF “Inequality and DISTRIBUTIVE CONFLICT Democratization: Stephan Haggard, University of California, San Diego Robert Kaufman, Rutgers University What Do We Terence Teo, Rutgers University Know?” In a recent article in the American Political Science Review, we 1 Obituary for Juan Linz Alfred Stepan and Je! Miley attempted to test what we call “distributive conflict” models of regime change using a qualitative data set of transitions to and 1 Inequality and Regime Change 1 Stephan Haggard, Robert from democracy from 1980 through 2000. These models, pioneered Kaufman, and Terence Teo by Carles Boix (2003) and Daron Acemoglu and James Robinson (2006)2 rest on complex 1 Rethinking Inequality and Democratization causal chains including both structural and game-theoretic components: inequality, Ben Ansell and David Samuels strategic interactions between incumbents and oppositions over the nature of political 2 RMDs institutions, and the ever-present threat of repression from above and violence from below. Carles Boix 2 Democracy, Public Policy, and Inequality 1. Stephan Haggard and Robert R. Kaufman, “Inequality and Regime Change: Democratic Transitions and the Stability of Daron Acemoglu, Suresh Naidu, Democratic Rule,” American Political Science Review 106 (August 2012): 495–516; Stephan Haggard, Robert Kaufman, and Pascual Restrepo and James A. Terence K. Teo, Distributive Con$ict and Regime Change: A Qualitative Dataset, 1980-2008, 2012. Robinson 3 Inequality, Democratization, and 2. Carles Boix, Democracy and Redistribution (New York: Cambridge University Press, 2003); Daron Acemoglu and James Democratic Consolidation Robinson, Economic Origins of Dictatorship and Democracy (New York: Cambridge University Press, 2006). Christian Houle (continued on page 4) 26 Section News 33 New Research 40 Editorial Committee RETHINKING INEQUALITY AND DEMOCRATIZATION: HOW INEQUALITY DIVIDES ELITES AND UNDERPINS REGIME OBITUARY FOR CHANGE UAN INZ J L Ben Ansell, Oxford University With the sad news of Professor David Samuels, University of Minnesota Juan Linz passing away on 1 Tuesday, October 1, 2013, we Despite the implications of Przeworski et al. , the search for factors that might felt no need to issue an editor’s drive “endogenous” democratization is alive and well. However, scholarship on note in this issue. We instead the political consequences of economic change has shifted from the hypothesized asked Professor Linz’s disciples impact of economic growth to the question of the political consequences of and friends Je! Miley and di'erent patterns of equal or unequal growth. We owe this ‘redistributivist’ turn - which draws Alfred Stepan to write an attention to a purported tension between democracy and property - to the in(uence of Daron 2 obituary. "ere is poetry and Acemoglu and James Robinson and Carles Boix. )ese studies vary in how they formalize the meaning in that this obituary 1. Adam Przeworski, Michael E. Alvarez, Jose Antonio Cheibub, and Fernando Limongi, Democracy and Development is being written by one of his (New York: Cambridge University Press, 2000). #rst, and one of his last PhD 2. Daron Acemoglu and James Robinson, “A )eory of Political Transitions,” American Economic Review 91 (September candidates, both of whom 2001): 938–963; Daron Acemoglu and James Robinson, Economic Origins of Dictatorship and Democracy (New York: Cambridge University Press, 2006); Carles Boix, Democracy and Redistribution (New York: Cambridge University Press, learned from him to his #nal 2003). days, and like all his students, (continued on page 8) (continued on page 3) Vol. 11, No. 3 Comparative Democratization Oct. 2013
Articles
RMDS Carles Boix, Princeton University
Redistributive models of democracy (RMD), to use Haggard and Kaufman’s expression, have been criticized on several counts: (1) their empirical performance is weak; (2) they make unconditional predictions about the relationship between structural variables (inequality, asset speci*city, organizational and information parameters) and political transitions; and (3) the parameters of the models are either too narrow and stylized or simply wrong – particularly (a) the assumption of rational, self-interested actors motivated by material interests, (b) the de*nition of ‘classes’, (c) the sequence of the political decision process, and (d) the tax setting model. After examining these critiques brie(y here, I conclude that, broadly speaking, the idea of democracy as an equilibrium (given by the material payo's of relevant social and economic actors) is: (1) relatively robust and (2) the best point of departure (or, in Lakatos’ terms, a core) from which to progressively build a satisfactory theory of political transitions.
Empirical Performance of the !eory Several important empirical tests on RMD *nd that the association between economic inequality, asset speci*city and political transitions either does not exist, is highly unstable or is restricted to democratic breakdowns. Houle (2009) concludes that inequality makes democratic breakdowns more likely but does not a'ect democratic transitions after 1960. Ansell and Samuels (2010) * nd that land inequality explains democratic transitions since the mid-19th century but that income inequality has the opposite e'ect. Haggard and Kaufman (2012) claim that almost half of all political transitions since 1980 are unrelated to distributive con(ict.
As I have insisted elsewhere,1 the examination of the covariates of political transitions has to be systematic to the point of including all the 1. Christian Houle, “Inequality and Democracy: Why Inequality Harms Consolidation but Does not A'ect Democratization,” World Politics 61 (October 2009): 589-622; Ben Ansell and David Samuels, “Inequality and Democratization: A Contractarian Approach,” Comparative Political Studies 43 (December 2010): 1543-1574; Stephan Haggard and Robert R. Kaufman, “Inequality and Regime Change: Democratic Transitions and the Stability of Democratic Rule,” American Political Science Review 106
(continued on page 12)
DEMOCRACY, PUBLIC POLICY AND INEQUALITY Daron Acemoglu, Massachusetts Institute of Technology Suresh Naidu, Columbia University Pascual Restrepo, Universidad de los Andes James A. Robinson, Harvard University )e relationship between inequality and democracy has been theorized since at least Aristotle, but in the last decade it has been subject to intense theoretical and empirical investigation. )e *rst formal models of democratic transitions by Acemoglu and Robinson (2000, 2001) suggested that there would be an inverse U-shaped relationship between inequality and democratization. Autocracies that were too equal would not democratize because there would not be enough social con(ict to create an e'ective demand for changes in political institutions. Autocracies that were too unequal would not democratize either because democratization would be very costly for non-democratic elites who would attempt to stay in power via repression. )ese models also predicted that democratization itself ought to reduce inequality as the newly enfranchised would vote for redistribution and more active government policy.
)ese theoretical results were obviously conditional on key modeling decisions. For one, political con(ict was conceived of as rich/elite versus poor/citizen with autocracy being associated with rule by the elite and democratization being associated with a transfer of power from rich to poor with a resulting change in policy from pro-elite to pro-poor. )ough this set-up has a parsimonious appeal, the comparative statics are conditional on some very simple models of both types of political regime. For example, Acemoglu and Robinson (2006) showed that once one relaxed the simple poor versus rich nature of political con(ict in their original models as well as the restriction of policy instruments, the nature of the comparative statics with respect to inequality in the basic model changed.1 Put simply, if the groups in con(ict were not
1. Daron Acemoglu, and James A. Robinson, “Why Did the West Extend the Franchise? Growth, Inequality and Democracy in Historical Perspective,” Quarterly Journal of Economics 115 (2000): 1167-1199; Daron Acemoglu and James A. Robinson, “A )eory of Political Transitions,” American Economic Review 91 (September 2001): 938- 963; Daron Acemoglu and James A. Robinson, Economic Origins of Dictatorship and Democracy (New York: Cambridge University Press, 2006). (continued on page 16)
2 Vol. 11, No. 3 Comparative Democratization Oct. 2013
Articles
INEQUALITY, DEMOCRATIZATION, AND DEMOCRATIC CONSOLIDATION Christian Houle, Michigan State University Does inequality a'ect democracy? Recently a large literature has argued that inequality in(uences both the likelihood of transition to and away from democracy, often through similar mechanisms. In this note, I argue that it is necessary to clearly distinguish between the e'ects of inequality on democratization and on democratic consolidation. As demonstrated by Przeworski et al. regarding economic development, for example, some factors may have very di'erent implications for these two transition processes.
Building on my previous work, I argue that inequality harms the consolidation of democracies but does not a'ect the likelihood of transition to democracy itself. In other words,1 unequal countries are not more or less likely to transition to democracy, but once they democratize they are less likely to remain democratic. I extend my previous analysis in three ways. First, my previous analysis used a single measure of inequality: the capital shares of the value added in production. In this note, I show that my results are robust to the use of Gini indexes. Second, I tackle the issue of endogeneity between inequality and democracy by using a novel instrumental variable strategy.
)ird, the capital shares dataset I used in my previous article ended in 2000 and about seventy countries were excluded from the analysis because of the lack of inequality data. Other recent empirical studies typically have an even larger proportion of missing observations. I use the extended version of the capital shares dataset I introduced in Houle.2 It covers 183 countries between 1960 and 2008, and contains more 1. Adam Przeworski, Michael E. Alvarez, Jose Antonio Cheibub, and Fernando Limongi, Democracy and Development (New York: Cambridge University Press, 2000); see, in particular: Christian Houle, “Inequality and Democracy: Why Inequality Harms Consolidation but Does not A'ect Democratization,” World Politics 61 (October 2009): 589-622.
2. Christian Houle, “Does Inequality Harm Economic Development and Democracy? Evidence from a Complete and Comparable Data Set on Inequality,” in Carol Lancaster and Nicolas van de Walle, eds., Oxford University Press Handbook on the Politics of Development (Oxford: Oxford University Press, forthcoming). See Houle, “Does Inequality Harm,” for more information on the imputation technique and the extent to which the dataset satis*es the basic criteria necessary for using such (continued on page 21 )
Obituary for Juan Linz, continued (continued from page 1) love and miss him terribly. works were on inequality and political and therefore help avoid repeating such paralysis in the United States, and on “state collective tragedies. His work on democratic Obituary for Juan J. Linz nations” in countries like India where the breakdowns especially so, motivated as it was On Tuesday October 1, 2013, Juan José e'ort to impose a “nation state” would be in by a sentiment well expressed by Meinecke, Linz Storch de Gracia died at the age of tension with an inclusionary democracy and the great German historian whose reaction 86. Professor Linz was undoubtedly one internal peace. to Hitler’s appointment as chancellor was the *nest political sociologists in the world. one that Linz was particularly fond of Legendary for the encyclopedic breadth of Linz’s undying passion for such diverse but quoting – namely, “)is was not necessary.” his knowledge, his ideas and writings deeply intertwined subjects was largely a product in(uenced debates surrounding a vast array of his traumatic experience growing up Linz came to New York in 1950 to pursue of the century’s most important political in interwar Europe. Born in the Weimar a doctoral degree in Sociology at Columbia problems. Republic to a Spanish mother and German University, an institution with which he father, Linz would witness *rst-hand over would remain a+liated for nearly two Linz’s empirical and theoretical the course of his childhood and adolescence a decades until 1969, when he moved to contributions to scholarly research and sequence of tragic social and political events: Yale where he would stay for the rest of his literature were legion. He contributed *rst in Germany, the economic crisis of the life. Upon his arrival at Columbia, he soon with path-breaking work on regime types, Weimar Reublic, its subsequent breakdown, gained a reputation for his extraordinary the dynamics of democratic breakdowns, and the rise to power and domination of the erudition and unparalleled command of transitions to democracy, democratic Nazis; then, after moving with his mother to comparative European history as well as institutional design, presidentialism versus Spain in the Spring of 1936, the breakdown social and political thought. Having already parliamentarism, parties and party systems, of the country’s Second Republic and been mentored in Spain by Javier Conde, political and business elites, federalism, its bloody Civil War. Linz’s work would he took classes and worked very closely at nationalism, and fascism. His most recent be consistently concerned to understand Columbia with Robert K. Merton, Paul (continued on page 25) 3 Vol. 11, No. 3 Comparative Democratization Oct. 2013
Haggard, Kaufman, and Teo
HAGGARD, KAUFMAN, AND TEO, CONTINUED (continued from page 1)
We argued that both theoretical and from distributive conflict transitions for authoritarian elites to repress methodological progress could be made appear more robust than those that political demands for redistribution. by undertaking detailed process tracing occur through a non-distributive They also note—contrary to Boix— of the components of these models. route. that at low levels of inequality there is We examined not only the reduced- little demand for democratization. Boix form relationship between inequality Distributive Conflict Models thus sees the prospects for democratic and regime change—on which there The work of both Boix (2003) and transitions to be inversely correlated has been surprisingly little supportive Acemoglu and Robinson (A&R, 2006) with inequality. A&R by contrast evidence for the theory (Acemoglu builds on the seminal Meltzer-Richard conclude that the relationship between et. al., this symposium)—but also the (MR) model (1981).3 MR provide a inequality and democratic transitions postulated mechanisms through which formal model of redistribution under should exhibit an inverted-U pattern, inequality translated into pressures for democratic rule, and thus a baseline with transitions to democratic rule authoritarian or democratic elites to for how the distribution of income most likely to occur at intermediate yield power. would change as a result of a transition levels of inequality. from authoritarian to democratic We distinguished in particular governance. Boix (p. 37) captures A&R add another layer of complexity between distributive conflict and non- the general spirit of these models: “a by considering credible commitment distributive conflict transitions. In the more unequal distribution of wealth problems; these issues are directly former, pressures from below appeared increases the redistributive demands germane to the controversial question to directly influence decisions by elites of the population…. [However] as the of how these models treat collective to make democratic concessions. In the potential level of transfers becomes action. In addition to the possibility latter, pressures from below did not larger, the authoritarian inclinations of repressing outright, A&R note that play a decisive role; transitions resulted of the wealthy increase and the elites can maintain power by making from incumbent initiatives, intra-elite probabilities of democratization and short-run economic concessions conflicts, and/or external pressures. democratic stability decline steadily.” to defuse threats from below. Yet How this strategic interaction between politically and economically excluded In this note, we revisit the theoretical elites and masses plays out depends groups are aware that elites can renege issue of how inequality generates regime on the level of inequality, the capacity on these concessions when pressures change, and the role of distributive to repress and other parameters such from below subside. Lower class groups conflict in particular. We summarize as capital mobility. Nonetheless, the are thus likely to press their advantage new results based on an updated challenge to the authoritarian status during windows when collective action version of our dataset that includes all quo emanates from what Acemoglu problems are temporarily resolved. democratic transitions through 2008. and Robinson call de facto as opposed The results strengthen our earlier to de jure political power: the ability These credible commitment problems finding that a large share of transitions of lower class groups to challenge elite can generate a counterintuitive result. occur in the absence of significant incumbents through mass mobilization, It might seem that transitions would be pressure from below, suggesting that strikes, demonstrations, riots and other more likely when lower class groups are distributive conflict models are at best physical threats to elite security. well-organized. Yet A&R argue that this subject to unspecified scope limitations, is not necessarily the case “because with including the capacity of subordinated While the basic insight of these a frequent revolutionary threat, future groups to overcome barriers to collective distributive conflict models is intuitive, redistribution becomes credible.”4 As action. the details are not. This can be seen an historical example, they cite the fact in differences in the treatment of that Germany—the country with the We conclude with some preliminary inequality, the central causal factor in most developed socialist movement— findings on how the nature of the these models. A&R agree with Boix that created novel welfare institutions transition to democratic rule may affect high inequality increases the incentives without extending the franchise while
the prospects for consolidation. We 3. Allan Meltzer and Scott Richard, “A Rational political elites in Britain and France find that the democracies that emerge )eory of the Size of Government,” Journal of 4. Acemoglu and Robinson, Economic Origins, 161, Political Economy 89 (October 1981): 914-927. 200.
4 Vol. 11, No. 3 Comparative Democratization Oct. 2013
Haggard, Kaufman, and Teo
were forced to extend the franchise as a on the effects of inequality and other contrast, were ones in which these result of pressures from below. structural variables. But if distributive elements were missing. Elite withdrawal We are hard pressed, however, to think conflict models are correct, we would was motivated by international of contemporary examples in which a expect to see democratic transitions pressures, intra elite conflicts, or what high capacity for collective action on preceded by mass mobilization that we call “pre-emptive” motives, in which the part of the poor was responsible for threatens authoritarian incumbents and elites initiated regime change in the stable, redistributive authoritarian rule. forces them to withdraw. belief that they could remain in office The primary focus of Economic Origins or effectively veto their democratic is on situations in which sporadic—if The qualitative data set that provided successors. unexplained—collective action drives the empirical base for the APSR regime change. The basic game on paper5 looks directly at this causal In coding the cases, we were deliberately which all others build distinguishes mechanism. Our data set assessed permissive, writing coding rules that between a low threat situation in which the role of distributive conflict in all gave the benefit of the doubt to the there are high costs for citizens to transitions indicated in the Polity theory. Unlike the extant inequality solve collective action problems and a IV (n=57) and Cheibub, Ghandi and data, our coding allowed us to consider high threat situation in which “citizens Vreeland (hereafter CGV; n=65) a variety of distributive conflicts that are able to solve the collective action datasets between 1980 and 2000.6 We may not be captured by any single problem relatively costlessly and/or drew a simple dichotomous distinction inequality measure, from urban class elites are not well organized in their between distributive and non- conflicts to ethnic, regional and sectoral defense…” (p. 145). To what extent do distributive conflict transitions. We ones. The economically disadvantaged contemporary transitions comport with coded “distributive conflict” transitions or the organizations representing them this distinction between “high threat” as ones in which both of the following need not be the only ones mobilized and “low threat” environments? occurred: in opposition to the existing regime. Although mass mobilization must partly Some simple tests R5 " 5 ')#.#)(5 ) 5 , #-.,#/.#0 5 reflect demands for redistribution, it Despite their differences, these grievances on the part of economically can be motivated by other grievances distributive conflict theories share disadvantaged groups or representatives as well. Yet mobilization must arise two important assumptions that are of such groups (parties, unions, NGOs) around distinctive and identifiable amenable to empirical observation. posed a threat—a “clear and present inequalities at least to some extent. First, although there are disagreements danger”—to the incumbency of ruling about the political dynamics of low and elites, and Even with a very permissive coding, intermediate levels of inequality, there is we found a large share of cases (44.6 agreement that democratic transitions R5" 5 ,#-#(!5 )-.-5 ) 5 , *, --#(!5 ." - 5 percent of the CVG transitions and are unlikely at high levels of inequality. demands appear to have motivated 42.1 percent of the Polity cases) in Second—and more important for elites to make political compromises or which distributive conflict played only our purposes—it is assumed that exit in favor of democratic challengers. a marginal role. Using three separate democracy is likely to occur when The presence of this causal mechanisms measures of inequality (capital’s share lower class groups are able to overcome was indicated at a minimum by a clear of income in the manufacturing sector, barriers to collective action—even if temporal sequence—mass mobilization a Gini coefficient from the Estimated only temporarily—and mobilize “de followed by authoritarian withdrawal— Household Income Inequality Data facto power” in favor of democracy. but where possible we drew on other Set and the Vanhanen measure of land The assumptions about collective evidence as well, including elite inequality) we also found that between action receive only limited attention in statements. 29 and 34 percent of all transitions the two books (Boix, this symposium); occurred in countries ranked in the in fact, A&R explicitly assume the Non-distributive transitions, by upper tercile of these measures; a high problem away by treating “citizens” as a share of transitions were taking place 5. Haggard and Kaufman, “Inequality and Regime 7 unitary actor in the formal models. And Change”. in high-inequality settings. Moreover, the role of mass mobilization is almost 7. Christian Houle, “Inequality and Democracy: entirely ignored in the econometric 6. José Antonio Cheibub, Jennifer Gandhi, and Why Inequality Harms Consolidation but Does Not James R. Vreeland, “Democracy and Dictatorship A'ect Democratization,” World Politics 61(October literature, which focuses more directly Revisited,” Public Choice 143(April 2010): 67-101. 2009): 589-622; University of Texas Inequality
5 Vol. 11, No. 3 Comparative Democratization Oct. 2013
Haggard, Kaufman, and Teo
a high proportion of these were heterogeneity. For the Third Wave of a negative role in non-distributive ones. distributive conflict transitions. Using recent democratization—when there the Gini as the measure of inequality, was in fact substantial divergence in We also find that the type of about 75 percent of the high-inequality political developments across cases- authoritarian regime appears to have transitions were characterized by -a large share of transitions simply a differential effect on the likelihood distributive conflict; the incidence of do not reflect the causal mechanisms of distributive and non-distributive such high-inequality transitions was stipulated in the theory, either with transitions. Challenges from below are 60 percent using the land inequality respect to the role of inequality or less likely under authoritarian regimes measure and 57 percent using capital’s distributive conflict. with multiparty legislatures—perhaps share of manufacturing income. We because of their capacity to coopt drew two conclusions: that inequality Extensions opposition—and more likely under did not appear to have the stipulated Despite these findings, the distributive military regimes that did not typically effect on the likelihood of transitions; conflict approach reopens the debate provide such channels of representation. and that distributive conflict was not a about the causes and consequences of On the other hand, the distinction uniform driver of democratization. At different transition paths. Do these between military and multiparty best, the effect of inequality worked paths arise from different causal regimes was not consequential in non- under scope conditions that were not roots? And more importantly, does the distributive transitions, which were clearly specified in the theory. distinction between distributive and driven primarily by elite actors who non-distributive conflict transitions were either tolerated by incumbent We have subsequently extended the have any enduring effect on the nature rulers or parts of the ruling circle itself. Haggard, Kaufman, and Teo data of democratic rule? We report some set through 2008, adding 14 cases to preliminary findings here. The likelihood of non-distributive the CGV transitions (n=79) and 16 transitions was, however, affected by cases to the Polity ones (n=73). The To explore the first question, we ran economic and international factors results remain essentially the same; if separate rare event logit estimates proxied in the regressions. Low or anything, they are even less favorable with country-clustered robust standard negative growth consistently predicted to the distributive conflict approach. errors and cubic time polynomials on non-distributive as well as distributive Between 34 and 45 percent of all the likelihood of each type of transition. transitions, presumably by intensifying transitions were in the most unequal Given space limitations the regressions elite struggles over rents or diminishing countries—again measured by the top are not presented here but are available their capacity to manipulate electoral terciles—and of these, between 37.5 from the authors on request. support. Non-distributive transitions and 55.6 percent were distributive (but not distributive ones) were affected conflict transitions. The percentage of As noted, we are particularly interested as well by the incidence of neighboring distributive conflict transitions among in the capacity of mass groups to democracies, an indication of the the CGV coding fell from 54.4 to 53.2 overcome barriers to collective action. relative importance of diffusion effects percent; Polity transitions conforming One factor – industrialization – has and other forms of external pressure. to the distributive conflict model long been viewed as a foundation for fell from 57.9 to only 49.3 percent. mobilization along class lines. In the Again, inequality had no effect on Boix (this symposium) argues that regressions, we use the size of the either type of transition. valid tests of the model must include manufacturing sector to proxy for this the full historical record to capture potential. Of course, the role played The Effects of Transition Paths the initial divergence associated with in collective action by non-economic The implicit question raised by democratization in the advanced factors such as ethnicity or religion the discussion in the preceding industrial states. However, this approach also require examination. Nevertheless, section is whether “non-distributive” makes strong assumptions about the it is noteworthy that manufacturing transitions—dominated by external ability to control for incredible panel –a basis for worker coordination and influences and intra-elite politics—are Project, Estimated Household Income Inequality organization–does have a consistently less likely to result in full democracies (EHII) Dataset, Available at http://utip.gov.utexas. edu/data.html; Tatu Vanhanen, Democratization significant impact on distributive than ones driven at least in part by and Power Resources, 1850-2000, 2003, Available at transitions and an insignificant or even pressures from below. Distributive http://www.fsd.uta.*/en/data/catalogue/FSD1216/.
6 Vol. 11, No. 3 Comparative Democratization Oct. 2013 Haggard, Kaufman, and Teo
Table 1. Regression Estimates of the E"ects of CGV Distributive and Non- Distributive Transitions on Polity Score in the Year Following the Transition, be associated with regime change in 1980-2008 a straightforward way, as Acemoglu et. al. note in this symposium. Core theoretical assumptions about the causal