Why Elections Deteriorate. Sarah Birch Department of Government University of Essex Wivenhoe Park Colchester CO4 3SQ

Transcription

1 Why Elections Deteriorate Sarah Birch Department of Government University of Essex Wivenhoe Park Colchester CO4 3SQ Paper prepared for presentation at the conference on Electoral Fraud, Vote Buying and Clientelism, Juan March Institute, Madrid, 29 June 1 July 2011 Work in Progress please do not cite without permission 1

2 Why Elections Deteriorate As more and more countries begin to hold competitive elections, there is an urgent need better to understand the dynamics of electoral malpractice. The Arab Spring heralds a new wave of possible democratization. At the same time, experience from the third wave of democratic transition suggests that movement toward democracy is in many cases non-linear, and we cannot always assume that countries setting out along this path will reach democratic consolidation (Carothers, 1997). Though there have been a number of noteworthy advances in recent years that have helped us to comprehend partial and failed transitions to democracy (e.g. Gandhi and Przeworski, 2007; Levitsky and Way, 2002; 2010; Linz and Stepan, 1996; O Donnell, 1997; Ottoway, 2003; Schedler, 2006; Svolik, 2008), we are still a long way from having a full understanding of how democratic processes can go wrong. The aim of this paper is to contribute to this goal by seeking to explain why and under what circumstances electoral processes in semi-democratic, semi-authoritarian states can deteriorate in quality. The rise of electoral assistance and observation since the end of the Cold War has sparked recent academic interest in electoral fraud and manipulation (e.g. Alvarez, Hyde and Hall, 2008; Beaulieu and Hyde, 2008; Birch, 2007; 2008; 2011; forthcoming; Bjornlund, 2004; Donno and Roussias, forthcoming; Elklit and Reynolds, 2005; Hartlyn and McCoy and Mustillo, 2008; Hyde, 2007; Mozaffar and Schedler, 2002; Schedler, 2006). Scholars have sought to identify and explain aspects of electoral misconduct; there has also been some attention devoted to the results of fraudulent elections. The most sustained debate to date in this as yet fragmentary literature has, however, been on the topic of whether elections are a motor of democratic change or whether, on the contrary, fraudulent elections can help to bolster the power base of authoritarian leaders. Some scholars have argued that elections tend, under the right conditions, 2

3 to lead eventually to democracy (Eisenstadt, 2004; Howard and Roessler, 2006; Lindberg, 2006; 2009; Roessler and Howard, 2009; Schedler, 2002; 2009), while others emphasize the role of elections in sustaining authoritarianism (Brownlee, 2007; Gandhi and Przeworski, 2007; Gandhi and Lust-Okar, 2009; Geddes, 2006; Lust-Okar, 2006; 2009; McCoy and Hartlyn, 2009; Magaloni, 2008; Stokes, 2005). Yet if elections do not always improve in quality over time as the first group of scholars suggest, nor do they always help authoritarian regimes retain a steady-state level of manipulation, as the second group would have it. There are many cases where elections exhibit dramatic changes for the worse, including Russia since the late 1990s, Ukraine in 2004, Peru in 2000 and Venezuela in 2005, to take a few well-known examples. In some cases, the deterioration of election quality leads to large-scale repression of the population (as in Algeria in 1992, Kenya in 2007, or Iran in 2009). In other cases, dramatic downturns in election quality spark protests which lead rulers to back down, and even exit from power (as in the Philippines in 1986 or Cote d Ivoire in 2000). Indeed, if a deterioration in election quality is by no means a sufficient condition for a popular uprising to bring about democratic change, one might be tempted to say it is a common condition. The burgeoning literature on the coloured revolutions in central and Eastern Europe documents how in case after case, worsening electoral irregularities led groups of citizens to take to the streets in protest (Beachain and Polese, 2010; Beissinger, 2007; Bunce and Wolchik, 2010; Kalandadze and Orenstein, 2009; Lane and White, 2010; Tucker, 2007). These protests were less than successful in countries such as Azerbaijan and Belarus, but in Serbia, Georgia and Ukraine, anti-fraud movements led to wholesale changes of power that improved the quality of electoral conduct (at least for a time). These considerations suggest that we need an understanding not 3

4 only of why and under what circumstances elections improve in quality, but also why and under what circumstances they deteriorate. This is what the present paper seeks to initiate. To this end, the paper deploys prospect theory to help explain why, in some contexts, leaders may hold dirtier elections than rational considerations would dictate. In brief, the theory is that when leaders are acting in a domain of loss, they are prepared to engage in electoral strategies that represent a considerable risk to their legitimacy. The dirty elections held under these circumstances can represent opportunities for popular mobilization in opposition to fraud, but they can also result in brutal repression of the population. The analysis will be rolled out in three parts. The first section develops a theoretical account of the circumstances under which elections deteriorate in quality over time. The second section examines two cases which provide evidence to support the account put forward in the first part. A final section discusses the results and concludes. I The Dictator s Electoral Calculus and the Limits of Rationality This analysis starts from the very basic observation that leaders want to win elections, but this is not all they can be expected to want. More than winning elections, they should want to remain in power, and even more than remaining in power, they should want to remain alive. The question then arises as to why leaders would hold elections so fraudulent that the ensuing backlash threatens to remove them from power and in some cases even puts their physical well-being in danger. The argument to be developed here is that when the leaders of electoral authoritarian states perceive they are losing power, they are likely to engage in excessively risky forms of electoral malpractice. 4

5 In some senses this argument is not new. Larry Diamond has already noted that leaders will engage in the greatest amounts of fraud and repression when they are vulnerable: When longtime authoritarian rulers face serious challenges (as in Malaysia and Zimbabwe recently), they may turn to their nastiest levels of repression, deploying levels of violence and intimidation that are unnecessary when political domination can be more subtly secured at the ballot box. (2002: 33) What I intend to demonstrate, however, is why in such situations leaders are likely to employ unnecessary levels of intimidation which can often in retrospect been seen the result of miscalculation. This argument is based on two key assertions: that some forms and amounts of electoral malpractice are riskier than others, and that under strain, leaders will select strategies that are riskier than rational strategic considerations dictate. My understanding of the riskiness of different forms of electoral malpractice is based on previous work (Birch, 2009; forthcoming) in which I divide the varieties of electoral malpractice into three basic types: the manipulation of electoral rules, the manipulation of voter choice, and the manipulation of the voting act (including electoral administrative procedures). The manipulation of electoral rules includes the passage of electoral laws that are at considerable variance with international standards of best practice, as well as the design of electoral institutions with the express purpose of biasing the electoral playing field. The manipulation of voter choice includes the illicit manipulation of campaign activities and the dissemination of information through which voters make up their minds, as well as the manipulation of the vote choice itself through particularistic rewards and/or punishments (vote-buying, clientelism and voter intimidation). Finally the manipulation of the voting act includes the manipulation of the 5

6 administrative procedures governing electoral registration, polling provisions, voting, vote counting, vote tabulation and the announcement of electoral results. Of the three types of electoral malpractice, the manipulation of voting opens leaders to the greatest threat of domestic and international backlash, in as much as it involves observable procedures that affect electoral actions at the retail level. There have been relatively few popular uprising against the passage of biased laws or vote-buying; virtually all large-scale protests against undemocratic electoral practices result from violations of democratic norms of electoral administrative procedures connecting with the voting act. Likewise, the most severe international sanctions tend to be reserved for blatant acts of fraud and other forms of electoral administrative manipulation, rather than forms of misconduct such as illicit campaign activities or malapportionment. This suggests that leaders are faced with the greatest risk to their power if they engage in electoral fraud, while they risk less by manipulating electoral systems and vote choice. The question then arises as to why leaders would ever run the risks associated with the most blatant forms of fraud, especially when international observers are present to document it. Leaders ought, in theory, to elect packages of electoral manipulation in which the greatest emphasis is put on up-stream forms of manipulation such as passing laws which advantage them in electoral terms and cajoling, deceiving or coercing voters into choosing the incumbent or incumbent-backed electoral option. The short answer is that in most cases this is precisely what leaders do. A study of electoral malpractice in Eastern Europe and the former Soviet Union, Latin America and the Caribbean and Sub-Saharan Africa over the period found that the manipulation of electoral administrative procedures was less common than either the manipulation of the legislative framework subtending elections or the manipulation of voter 6

7 choice (Birch, forthcoming: Chap. 3). Leaders also generally design packages of electoral malpractice that enable them to win elections peacefully and to retain power. However, this is not always the choice leaders make. They do sometimes miscalculate and opt for more overall electoral malpractice and riskier forms of electoral manipulation than the population is prepared to withstand. The result then typically takes the form of a popular uprising that ends either in the leader s downfall or in repression. 1 The way in which popular protests against electoral fraud play out has been the subject of considerable recent research (Beissinger, 2007; Bunce and Wolchick, 2010; Cox, 2009; Thompson and Kuntz, 2006; Lehoucq, 1995; Magaloni, 2006; 2010; Schedler, 2002). The current analysis is concerned not so much with the ultimate fate of leaders in this situation as with the decision to undertake electoral fraud of a magnitude sufficient to provoke a popular backlash. The answer that I propose is that the decisions leaders make under risk are affected by well-known cognitive biases which cause them to choose strategies that are sub-optimal. In a recent paper Gary Cox (2009) argues that the leaders of authoritarian states risk holding elections so as to gain information about the strength of their rivals and thereby reduce the chances of violent removal from office. It follows that over time, elections ought to become increasingly democratic as leaders will, if they behave strategically, refrain from the most blatant forms of electoral abuse in order to prevent threats of reaction from increasingly powerful opponents that might put them at risk of violent removal from office. This implication of Cox s analysis is articulated in the closing paragraph of the paper: 1 Occasionally, electoral malpractice also results in foreign intervention or sanctions with result in the leader losing power. This is rare, however, and the focus of this paper will be on domestic protests. 7

8 The model suggests a path-dependent story about democratization. Once an electoral process is established, each new generation of competitors for power have an incentive to invest in the ability to mobilize popular support and win elections. As the military power of the rival(s) increases, the incumbent offers fairer and fairer elections, and is more and more likely to step down after an electoral defeat. (Cox, 2009: 29) A similar argument is articulated by Staffan Lindberg, who details what he describes as democratization by elections : [I]terative, multiparty elections change the cost of both repression and toleration and are thus key events that affect the cost-benefit analysis for the incumbent as well as for reformers. [...] If large numbers of the electorate turn out for elections and vote for the opposition, the cost of electoral manipulation and intimidation goes up. A more massive scheme of fraud and use of force needs to be put in place and shielded from outsiders, the risk of adverse reactions from the international community increases, the risk of postelectoral protest and violence goes up, and so on. At the same time, the benefits for the opposition elites of daring to mobilize citizens for protest increase. (Lindberg, 2009: 325-6). Lindberg acknowledges that in some cases elections can serve to reproduce and strengthen authoritarian regimes, but he argues that over time the cost of repeated repression at election time goes up, leading many authoritarian rulers to democratize eventually (2009: ). This explains why leaders sometimes hold elections that are too free from the point of view of their 8

9 narrow electoral interests, causing them to lose office and, under the right conditions, bringing to an end a period of authoritarianism. The problem with both Cox and Lindberg s arguments is that although they can undoubtedly help to explain certain individual cases, it is unclear that that they help us to explain aggregate patterns over time, as there is little evidence of a trend in recent years in the quality of elections in semi-authoritarian states, once the democratization of Central Europe in the 1990s is taken into consideration (Birch, forthcoming: chapter 2). Something else must be going on to counteract cases where elections serve as a motor of democratization. And the fact remains than in many contexts, leaders of semi-authoritarian states appear to make choices that are too risky in that they engage in forms and amounts of electoral malpractice that exceed the levels dictated by strategic considerations alone. There have been numerous anecdotal reports of authoritarian leaders in risky electoral situations miscalculating and over-using fraud (Thompson and Kuntz, 2006: ), but we have as yet a poor understanding of why this might happen. It is at this point that the results of social-psychological research on risk can help us to understand such decisions. Prospect theory, as developed by Daniel Kahneman and Amos Tversky (Kahneman and Tversky, 2000; Tversky and Kahneman, 2000), is a theory of decisionmaking under risk which predicts that people are loss averse in the sense that when they face the prospect of losing (when they are in the domain of loss ), they will seek to minimize their losses by engaging in behaviour that is riskier than what purely rational considerations based on classical expected utility theory would dictate. Kahneman won a Nobel Prize for his work on prospect theory in 2002, and it has become increasingly influential in the social sciences in recent years, as behavioural economists and international relations scholars have integrated 9

10 knowledge about cognitive biases into their understanding of human reasoning. Indeed, prospect theory has recently been described as the most influential behavioural theory of choice in the social sciences (Mercer, 2005: 1). Prospect theory has received several applications in comparative political science (Mercer, 2005; McDermott, Fowler and Smirnov, 2008), including in the study of semi-democratic systems (Weyland, 2002). Recent work on prospect theory suggests that the tendency it describes has its origins in the evolution of survival strategies, and that it may be particularly difficult to overcome for those whose survival is threatened, including dictators (McDermott, Fowler and Smirnov, 2008: 336). Applied to decisions over electoral strategies, prospect theory can be used to predict the risks leaders will be prepared to take. The basic argument put forward here is thus that when the leaders of undemocratic and semi-democratic states perceive that their legitimacy is ebbing, they are in the domain of loss. Under these circumstances they will be willing to take irrational risks in order to protect their position. In some cases these risks will be non-electoral e.g. undertaking risky military ventures or engaging in human rights abuses likely to incur the wrath of trading partners. But in as much as elections are the principal mechanism through which leaders maintain power and legitimacy, prospect theory ought to apply also to the strategies adopted by leaders at election time. In devising the ideal package of electoral strategies, leaders must balance the desire to win elections against the risk of losing legitimacy should they engage in electoral malpractice. The risk of losing legitimacy has two components: external and internal; external legitimacy will be lost if the international community perceives that elections are have deteriorated and judge a country to be moving away from democracy. In the present context it is, however, internal legitimacy that in most cases represents the most immediate threat to a ruler s ability to maintain 10

11 power. If an election is too poorly run, and fraud is too blatant, the leader risks provoking a legitimacy crisis and being faced with protests that could potentially unseat him or her. Loss averse leaders will, however, be prepared to take such risks in order to avoid power slipping away. The protests which follow blatantly rigged elections may unseat the leader, though in most cases it will be possible to use force to repress them. In any case, large-scale popular protests represent a non-negligible risk that the leader will be deposed in a disorderly and possibly violent fashion. As Cox (2009) notes, such an outcome is one that fully rational leaders ought to avoid at any cost, but it is a risk that loss averse leaders may be prepared to run. To summarize: when leaders perceive that they are losing legitimacy, prospect theory suggests that they should be prepared to take greater risks than strategic considerations alone would warrant. These risks will include amounts and forms of electoral manipulation that could lead to a popular uprising and violent conflict, including the most blatant forms of fraud and manipulation of electoral results. Inasmuch as many popular uprisings are successfully repressed, the risky behaviour exhibited by leaders under these circumstances can help to account for why elections do not always become increasingly democratic over time and why in some contexts they deteriorate in quality. II Case Studies Gathering evidence to test the proposition articulated above poses methodological challenges, as it is difficult to see how perceived loss of leader legitimacy could easily be operationalized in such a way as to enable conventional methods of large-n analysis to be employed. The approach adopted here is therefore that of a series of illustrative case studies, designed to flesh out the 11

12 theoretical expectations outlined above. (Ultimately, of course, a more robust means of testing this hypothesis will need to be identified). The examples chosen to illustrate the argument are the 1988 election in Mexico and the 2000 election in Serbia. In both the Mexican and Serbian cases the ruling party faced declining support and reacted by resorting to blatant forms of fraud. These acts of electoral desperation were ultimately the downfall of the ruling parties in both cases, as popular backlashes resulted in reforms that saw them lose power. In the Mexican case the reform process was protracted and relatively orderly, and the ruling PRI lost power electorally, while in Serbia the whole process was quicker and more irregular, with popular demonstrations in the final days of 2000 resulting in the precipitous downfall of Milosevic and his regime. Mexico The 1988 presidential election in Mexico stands out for the blatant electoral malpractice that led to 50.4 per cent of the vote being declared for Carlos Salinas (Nohlen, 2005). The level of fraud committed by the dominant Institutional Revolutionary Party (PRI) and the widespread perception that challenger Cuauhtémoc Cárdenas had been robbed of victory sparked a protracted series of post-electoral demonstrations which eventually resulted in electoral reforms (Eisenstadt, 2004; Magaloni, 2006; Nohlen, 2005). The reforms introduced between 1990 and 1996 brought to an end decades of coercive electoral mobilization by the PRI and resulted in the completion of Mexico s democratic transition when the PRI lost control of the lower house of the legislature in 1997 and of the presidency in In this sense, the fraudulent 1988 elections were the beginning of the end for the PRI. 12

13 The question to be answered is therefore: why did the PRI use blatant fraud in the 1988 elections? The argument I propose is that the party s declining support base placed it in the domain of loss, which caused it to miscalculate and engage in more electoral malpractice than the population and the opposition would accept. The PRI maintained its hegemonic position in Mexican politics for 67 years through a combination of tactics, including moderate amounts of electoral malpractice (Knight, 1996; Magaloni, 2006). But the main basis of the party s support was an extensive system of electoral clientelism which was particularly strong in rural areas (Camp, 1993: ; Knight, 1996; Magaloni, 2006; Magloni, Diaz-Cayeros, and Estévez, 2007; Middlebrook, 1986: 142). Over time, economic development and social change eroded the PRI s clientelist support base; at the same time, there was less pork to distribute to voters due to the government s decision to adopt neoliberal economic policies following the 1982 debt crisis (Eisenstadt, 2006; Greene, 2008; Magaloni, 2006: 84). During the 1980s, the party therefore relied increasingly on electoral fraud to win elections (Levitsky and Way, 2010: 153; Magaloni, 2006: 5-6). This shift from relatively safe clientelism to riskier fraud was thus driven by a decline in the true popularity of the party, which placed the party firmly in the domain of loss. Moreover, Mexico s economy had suffered considerably in the 1980s, such that by 1988 the inflation rate was 52 per cent and the minimum salary in real terms in Mexico city was 46 per cent lower than it had been in 1980 (Domínguez and McCann, 1995: 36). The PRI clearly had grounds to be concerned for its electoral fortunes. The 1988 elections pitted PRI stalwart Salinas against renegade Cárdenas, who had broken from the PRI the previous year. It remains unclear what the PRI s level of genuine support was in this election, but commentators agree that fraud was used to push Salinas s vote over the 50 13

14 per cent mark required for a majority in the Electoral College (Magaloni, 2006; Levitsky and Way, 2010: 153). The fraud included the traditional tactics of multiple voting and ballot-box stuffing, but these proved insufficient in the face of the unexpectedly strong showing of Cárdenas. Though the party had been divided on whether to falsify the results, those in favour of the move ultimately won the debate (Magaloni, 2006: ). On the night of the election, an announcement was made that the electronic counting system designed to tally the votes had stopped working; at this point the true results were altered to the benefit of the PRI (as admitted by outgoing president Miguel de la Madrid in his memoires (Thompson, 2004)). The high-risk fraud strategy prompted a considerable backlash both at home and abroad (Levitsky and Way, 2010: 155). Though Cárdenas eventually backed down in exchange for concessions, the perceptions of widespread fraud in the 1988 election was a key turning point that initiated a cycle of post-electoral protests which had the effect of weakening the traditionalists in the PRI (Eisenstadt, 2004; Magaloni, 2006). Reform eventually came about as election results were repeatedly denounced by opposition parties, who regularly took to the streets to demand concessions from the PRI. Over time, these demands strengthened the opposition, which was then in a position to demand more far-reaching changes to the electoral regime (Eisenstadt, 2004; 2006). In retrospect, the over-use of fraud in 1988 can be seen as a significant miscalculation, arguably resulting from the loss aversion of the sort predicted by prospect theory. Serbia The so-called bulldozer revolution following the 2000 pre-term general elections in Serbia marked the end of the Milosevic era. This momentous event, in which thousands of protesters 14

15 took to the streets, was the first in a wave of so-called coloured revolutions which swept through post-communist Europe in the first decade of the new century. The protests were sparked by a marked increase in the level of electoral fraud and abuse perpetrated by Slobodan Milosevic and his regime. What concerns us here is not so much the protest itself as the decision by Milosevic to step up electoral malpractice in the face of an increasingly organized civil society and disaffected opposition. Slobodan Milosevic had been at the helm of both Serbia and Yugoslavia since the late 1980s. Though his formal title had changed several times over the years, he and his wife Mira Markovic had led the dwindling Balkan state as if by right (Thompson and Kuntz, 2006: 125). Milosevic maintained power through a variety of means, including the effective division of the opposition, restrictions on the media and varying degrees of electoral malpractice (Birch, 2002; OSCE, 2000; Thomson and Kuntz, 2004). At the same time, Milosevic s power exhibited a gradual decline over the 1990s as his margin of electoral victory fell and he failed in his not inconsiderable efforts to stand up to international condemnation of his political tactics. The Serbian economy had suffered considerably from Western trade sanctions; by one report, half the population was wearing shoes over ten years old at the time of the 2000 elections (Bujosevic and Radovanovic, 2003: 4). Meanwhile, the NATO-led invasion of Kosovo in 1999 and Milosevic s indictment for war crimes had made it clear that his powers on the larger political stage were limited. There is also evidence of tensions at the time within the regime forces, which consisted largely of Milosevic s own Socialist Party and the United Left party of his wife (International Crisis Group, 2000: 11-12), as well as doubts on the part of some members of the armed forces over the war crimes they had committed in Kosovo (International Crisis Group, 2000; RFE/RL Balkan Report 3.27, 29 15

16 September 2000). To add to his problems, Milosevic had to contend with a decision by the federal republic of Montenegro to boycott the election following Milosevic s introduction of a constitutional change to allow the direct election of the Yugoslav federal president. To make matters worse for Milosevic, Serbia s fractious opposition had finally managed to unify around a common front headed by Vojislav Kostunica (Birch, 2002; Lamont, 2010; Thompson and Kuntz, 2004), a factor that is widely-held to be key to democratic change in authoritarian states (Howard and Roessler, 2006; van de Walle, 2006). Civil society groups had also received considerable technical assistance in domestic election observation from abroad (Birch, 2002; Bunce and Wolchick, 2010). It is thus clear that Milosevic s legitimacy was in decline and that he risked considerable losses in the 2000 elections. What interests us is why Milosevic chose to step up electoral fraud and risk confrontation with opposition/civil society forces rather than exit power gracefully and negotiate a safe retirement in exile. Fear of prosecution for war crimes and fear of the Serbian criminal underworld have been cited as reasons for this decision, as have economic interests and sheer obstinacy borne out of Milosevic s belief in his right to rule Serbia (Thompson and Kuntz, 2006: ). Yet ultimately Milosevic was ejected from power and sent to the Hague where he died during his trial; it is thus fair to say that he suffered more than he might have had he chosen an alternative path. Although electoral fraud and manipulation had been a feature of Serbian/Yugoslav elections for many years (Thomson and Kuntz, 2004), there is considerable evidence of increased electoral malpractice at the time of the 2000 election. New restrictions on the media seriously limited the opposition s ability to campaign (OSCE, 2000: 3-5); the opposition was denied an opportunity to inspect voter registers (OSCE, 2000: 6); an order, issued on the eve of 16

17 polling, required polling commission members to inspect marked ballots before voters deposited them in ballot boxes (OSCE, 2000: 8); opposition party polling commission members were expelled from polling stations on polling day (OSCE, 2000: 80); and all opposition members were expelled from the Federal Election Commission session and premises on the night of the election (OSCE, 2000: 8). When even these extreme measures only yielded Milosevic 39 per cent of the popular vote to Kostunica s officially-reported 49 per cent, the Federal Election Commission declared that a run-off would be held. But Kostunica, convinced that he had won an absolute majority, refused to accept a runoff, and eventually Milosevic stepped down when it became clear that he had lost the support of the military. Though it is impossible to say for sure why Slobodan Milosevic stepped up electoral abuse in the face of declining support, it is clear that he engaged in a highly risky strategy which ultimately failed to pay off. This scenario accords with what we would expect of a loss-averse leader and suggests that Milosevic s miscalculation could well have been the result of the type of cognitive behaviour described by prospect theory. III Discussion and Conclusion Reports of flawed elections frequently include passing comments that the leadership miscalculated and went too far in using fraudulent and intimidatory tactics against the population. Yet we have little understanding of why this sort of miscalculation takes place. Certainly leaders in semi-authoritarian states face formidable problems gathering accurate information (Brownlee, 2007; Cox, 2009; Egorov, Guriev and Sonin, 2009; Wintrobe, 1998), but this paper has suggested that the over-use of fraud may also result from loss aversion associated with the psychological tendency to over-value assets held and engage in excessively risky 17

18 behaviour to protect them. When leaders perceive that they are losing legitimacy, they tend to increase the amount of fraud they employ; such increases are often greater than strategic considerations alone warrant. We have considered the examples of Serbia and Mexico; others that might be seen to fit the general pattern described here are, inter alia, Albania in 1996, Peru in 2000, Zimbabwe in 2002 and Iran in There are of course other factors which may be held to account for increases in electoral misconduct. In cases such as Kyrgyzstan and Kenya, we observe brief openings in which electoral practices improve, followed by crackdowns as leaders tighten their stranglehold over politics and electoral malpractice increases once again. In other cases, international factors play a role, including rises in oil and gas exports (and prices) in countries such as Russia, Azerbaijan and Venezuela, where governments could increasingly afford to defy Western concerns with electoral integrity given the strategic role they played in supplying the world with fossil fuels. Integrating considerations of international legitimacy into the argument outlined here is an important direction for future research. Further empirical work is also needed to evaluate the hypothesis set out in this paper. The examples of Serbia and Mexico are illustrative, but by no means are they confirmatory. The next step in this research project will therefore be to develop a rigorous approach to testing the role of loss aversion in shaping leaders electoral strategies in semi-authoritarian states. 18

19 References Alvarez, Michael R., Thad E. Hall and Susan D. Hyde (eds.) Election Fraud: Detecting and Deterring Electoral Manipulation. Washington, DC: Brookings Institution Press. Beachain, Donnacha O. and Abel Polese (eds.) The Colour Revolutions in the Former Soviet Republics, New York and London: Routledge. Beaulieu, Emily and Susan D. Hyde In the Shadow of Democracy Promotion; Strategic Manipulation, International Observers and Election Boycotts, Comparative Political Studies 42.3, pp Beissinger, Mark R Structure and Example in Modular Political Phenomena: The Diffusion of Bulldozer/Rose/Orange/Tulip Revolutions, Perspectives on Politics 5.2, pp Birch, Sarah The 2000 Elections in Yugoslavia: The Bulldozer Revolution, Electoral Studies 21.3, pp Electoral Systems and Electoral Misconduct, Comparative Political Studies, 40.12, pp Electoral Institutions and Popular Confidence in Electoral Processes: A Cross- National Analysis, Electoral Studies 27.2, pp Electoral Corruption, in Todd Landman and Neil Robinson (eds.), Handbook of Comparative Politics, London and Thousand Oaks, CA: Sage Perceptions of Electoral Fairness and Voter Turnout, Comparative Political Studies, 43.12, pp Forthcoming. Electoral Malpractice. Oxford: Oxford University Press. 19

20 Bjornlund, Eric C Beyond Free and Fair: Monitoring Elections and Building Democracy, Baltimore and London: John Hopkins University Press. Brownlee, Jason Authoritarianism in an Age of Democratization, Cambridge: Cambridge University Press. Bujosevic, Dragan and Ivan Radovanovic The Fall of Milosevic: The October 5 th Revolution, Basingstoke: Palgrave Macmillan. Bunce, Valerie J. and Wolchik, Sharon L Defeating Dictators: Electoral Change and Stability in Competitive Authoritarian Regimes, World Politics, 62, 1, pp Camp, Roderic A Politics in Mexico. Oxford: Oxford University Press. Carothers, Thomas The Observers Observed, Journal of Democracy 8.3, pp Cox, Gary Authoritarian Elections and Leadership Succession, , Paper presented at the Annual Political Science Association conference, Toronto. Diamond, Larry Thinking about Hybrid Regimes. Journal of Democracy 13.2, pp Domínguez, Jorge I. and James A. McCann Shaping Mexico's Electoral Arena: The Construction of Partisan Cleavages in the 1988 and 1991 National Elections, American Political Science Review 89.1, pp Donno, Daniela and Athanassios Roussias. Forthcoming. Does Cheating Pay? The Impact of Electoral Manipulation on Party Systems, Comparative Political Studies. Eisenstadt, Todd Courting Democracy in Mexico: Party Strategies and Electoral Institutions. Cambridge: Cambridge University Press Mexico s Postelectoral Concertacesiones: The Rise and Demise of a Substitutive Informal Institution, in Gretchen Helmke and Steven Levitsky (eds.), Informal 20

21 Institutions and Democracy: Lessons from Latin America. Baltimore, MD: Johns Hopkins University Press, pp Egorov, Georgy, Sergei Guriev and Konstantin Sonin Why Resource-poor Dictators Allow Freer Media: A Theory and Evidence from Panel Data, American Political Science Review Elklit, Jorgen and Andrew Reynolds A Framework for the Systematic Study of Election Quality, Democratization, 12.2, pp Gandhi, Jennifer and Adam Przeworski Authoritarian Institutions and the Survival of Autocrats, Comparative Political Studies 40.11, pp Geddes, Barbara Why Parties and Elections in Authoritarian Regimes?, presented at the Annual Meeting of the American Political Science Association, Washington DC. Greene, Kenneth F Dominant Party Strategy and Democratization, American Journal of Political Science 52.1, pp Hartlyn, J., J. McCoy and T. M. Mustillo Electoral Governance Matters: Explaining Electoral Outcomes in Contemporary Latin America, Comparative Political Studies 41.1, pp Howard, Marc M. and Phillip G. Roessler Liberalizing Electoral Outcomes in Competitive Authoritarian Regimes, American Journal of Political Science 50.2, pp Hyde, Susan D The Observer Effect in International Politics: Evidence from a Natural Experiment, World Politics 60.1, pp International Crisis Group Yugoslavia s Presidential Election: The Serbian People s Moment of Truth, Report No. 102, 19 September. 21

22 Kahneman, Daniel and Amos Tversky Prospect Theory: An Analysis of Decision under Risk, in Daniel Kahneman and Amos Tversky (eds.), Choices, Values and Frames, Cambridge: Russell Sage Foundation and Cambridge University Press, pp Kalandadze, Katya, and Mitchell A. Orenstein Electoral Protests and Democratization Beyond the Color Revolutions, Comparative Political Studies, 42.11, Knight, Alan Corruption in Twentieth Century Mexico in Walter Little and Eduardo Posada-Carbó (eds.), Political Corruption in Europe and Latin America, Basingstoke: Macmillan, pp Lamont, Christopher Contested Sovereignty: The International Politics of Regime Change in the Federal Republic of Yugoslavia, in David Lane and Stephen White (eds.), Rethinking the Coloured Revolutions, New York and London: Routledge, pp Lane, David and Stephen White (eds.) Rethinking the Coloured Revolutions, New York and London: Routledge. Lehoucq, Fabrice Edouard Institutional Change and Political Conflict: Evaluating Alternative Explanations of Electoral Reform in Costa Rica, Electoral Studies 14.1, pp Levitsky, Stevan and Lucan Way Elections about Democracy: The Rise of Competitive Authoritarianism, Journal of Democracy 13.2 (2002) International Linkage and Democratization, Journal of Democracy Competitive Authoritarianism: Hybrid Regimes after the Cold War. Cambridge: Cambridge University Press. Lindberg, Staffan I Democracy and Elections in Africa, Baltimore, MD: Johns Hopkins University Press. 22

23 A Theory of Elections as a Mode of Transition, in Staffan Lindberg (ed.) Democratization by Elections: A New Mode of Transition?, Baltimore: Johns Hopkins University Press, pp Linz, Juan J. and Alfred Stepan Problems of Democratic Transition and Consolidation, Baltimore and London: Johns Hopkins University Press. Lust-Okar, Ellen Elections under Authoritarianism: Preliminary Lessons from Jordan, Democratization 13.3, pp Legislative Elections in Hegemonic Authoritarian Regimes: Competitive Clientelism and Resistance to Democratization, in Staffan I. Lindberg (ed.), Democratization by Elections: A New Mode of Transition, Baltimore: Johns Hopkins University Press, pp McDermott, Rose, James H. Fowler and Oleg Smirnov On the Evolutionary Origin of Prospect Theory Preferences, Journal of Politics 70.2, pp Magaloni, Beatriz Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico, Cambridge: Cambridge University Press Credible Power-Sharing and the Longevity of Authoritarian Rule, Comparative Political Studies 41.4/5, pp The Game of Electoral Fraud and the Ousting of Authoritarian Rule, American Journal of Political Science 54.3, pp Magaloni, Beatriz, Alberto Díaz-Cayeros and Federico Estévez Clientelism and Portfolio Diversification: A Model of Electoral Investment with Applications to Mexico, in Herbert Kitschelt and Steven I. Wilkinson (eds.), Patrons, Clients and Policies: Patterns 23

24 of Democratic Accountability and Political Competition, Cambridge: Cambridge University Press, pp Mercer, Jonathan Prospect Theory and Political Science. Annual Review of Political Science 8, pp Middlebrook, Kevin J Political Liberalization in an Authoritarian Regime: The Case of Mexico, in Guillermo O Donnell, Philippe C. Schmitter and Laurence Whitehead (eds.), Transitions from Authoritarian Rule: Latin America, Baltimore, MD: Johns Hopkins University Press, pp Mozaffar, Shaheen and Andreas Schedler The Comparative Study of Electoral Governance An Introduction, International Political Science Review, 23.2, pp O Donnell, Guillermo Illusions about Consolidation, in L. Diamond et al (eds.), Consolidating the Third Wave Democracies: Themes and Perspectives, Baltimore and London: Johns Hopkins University Press. Organization for Security and Co-operation in Europe [OSCE] Elections in the Federal Republic of Yugoslavia 24 September 2000: Preliminary Findings and Conclusions, Warsaw: OSCE. Ottoway, Marina Democracy Challenged: The Rise of Semi-Authoritarianism, Washington DC: Carnegie Endowment for International Peace. Schedler, Andreas The Nested Game of Democratization by Elections, International Political Science Review 23.1, pp (ed.) Electoral Authoritarianism: The Dynamics of Unfree Competition, Boulder, CO and London: Lynne Rienner. 24

25 The Contingent Power of Authoritarian Elections in in Staffan Lindberg (ed.) Democratization by Elections: A New Mode of Transition?, Baltimore: Johns Hopkins University Press, pp Stokes, Susan C Perverse Accountability; a Formal Model of Machine Politics with Evidence from Argentina, American Political Science Review 99.3, pp Svolik, Milan Authoritarian Reversals and Democratic Consolidation, American Political Science Review 102.2, pp, Thomson, Ginger, Ex-President in Mexico Casts New Light on Rigged 1988 Election, New York Times, 9 March Thompson, Mark R., and Philipp Kuntz Stolen Elections: The Case of the Serbian October, Journal of Democracy 15.4, pp After Defeat? When Do Rulers Steal Elections? in A. Schedler (ed.), Electoral Authoritarianism: The Dynamics of Unfree Competition, Boulder and London: Lynne Rienner, pp Tucker, Joshua Enough! Electoral Fraud, Collective Action Problems, and Post- Communist Colored Revolutions, Perspectives on Politics 5. Tversky, Amos and Daniel Kahneman Advances in Prospect Theory: Cumulative Representation of Uncertainty, in Daniel Kahneman and Amos Tversky (eds.), Choices, Values and Frames, Cambridge: Russell Sage Foundation and Cambridge University Press, pp van de Walle, Nicolas Tipping Games: When Do Opposition Parties Coalesce?, in Andreas Schedler (ed.), Electoral Authoritarianism: The Dynamics of Unfree Competition, Boulder, CO: Lynne Rienner, pp

26 Weyland, Kurt The Politics of Market Reform in Fragile Democracies: Argentina, Brazil, Peru and Venezuela. Princeton and Oxford: Princeton University Press. Wintrobe, Ronald The Political Economy of Dictatorship, Cambridge: Cambridge University Press. 26