Regime Diversity and Electoral Systems in Post-Communism JACK BIELASIAK Regime type has an impact on the design and reform of electoral systems, as revealed by a classification of the 28 former communist states as authoritarian, semi-authoritarian, and democratic regimes and a comparison of their electoral arrangements. From 1989 to the end of 2005, the selection of electoral rules in the post-communist polities conformed to the world-wide trend of more permissive electoral systems over time, although the different regime types embraced diverse electoral design strategies. Democratic and semi-authoritarian regimes adopted election rules in favour of inclusive PR, generous district magnitudes and assembly size. Authoritarian regimes made use of restrictive majoritarian formulas to constrain the political process. As in the rest of the world during the third wave of democratization, 'electoral democracy' emerged as an important legitimating criterion for the post-communist political systems, but the rules for contestation were devised in very different ways to facilitate or constrain political competition. The collapse of the communist system in Eastern Europe in 1989 and the Soviet Union in 1991 ushered in a new era for the politics of the region. The monopoly of power vested in the Communist Party was replaced by a legitimating principle that evolved, in most but not all cases, towards a contested version of politics. For many states the shift signified the embrace of democratic principles associated with the politics of choice, including the design of electoral institutions to invigorate popular participation and political competition. In other countries the changeover fell short of the democratization process, and instead recast politics in new forms of authoritarian rule. None the less, in both the democratizing and the authoritarian systems, there was considerable attention to the engineering of electoral systems. 'Electoral democracy' became the dominant face of regime change through the 1990s, regardless of the elites' commitment to democracy. Jack Bielasiak is Professor of Political Science and Russian-East European Studies at Indiana University, Bloomington. His current research focus is on the development of electoral and party systems in the post-communist states, especially in comparison with other emerging democracies. Journal of Communist Studies and Transition Politics, Vol.22, No.4, December 2006, pp.407^-30 ISSN 1352-3279 print/1743-9116 online DOI: 10.1080/13523270601017569 © 2006 Taylor & Francis 408 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS The process was tied directly to the dominant global ideology of 'democratization' and 'nation building' that demanded at least the semblance of political contestation. The emphasis on elections as the defining characteristic of the new politics culminated in a 'fallacy of electorism' that privileged the formal procedures of choice over other forms of political expression.1 The consequence was that domestic politics were defined not only by electoral rules that led to free and fair elections, but also by authoritarian or illiberal political forms vested in electoral authority. Thus, while many leaders involved in the third wave of democratization committed themselves to an open, pluralist mode of competition, other political entrepreneurs preferred to maintain control while acquiescing to the veneer of democratic choice. In either circumstance, the design of electoral systems became a defining moment for the new regimes. Theorizing Electoral Design The dominant approach in electoral studies has looked to voting procedures as an exogenous variable affecting political outcomes. Much attention is devoted to understanding how electoral institutions shape the structure of party systems, in terms of the number of political parties, the space of competition, or political representation/ The nexus between electoral and party systems has produced critical insights about 'laws' governing political competition, strategic voting or proportionality.4 Both the mechanical effects of the seat-to-vote translation and the strategic calculations of politicians and voters are instrumental in defining how rules affect political competition. The focus is on how election codes provide incentives for politicians to form and align political parties, and condition voters to cast preferences among the parties. The concern is less with the emergence of the 'rules of the game' than with the consequences of these rules for political contestation, so causality is from electoral systems to party systems. As the third wave of democratization ushered in an era of pluralist politics, the question of crafting electoral rules appropriate for competitive politics took on greater urgency and visibility.5 The spotlight shifted to the opposite direction of causality, so as to explain the selection of electoral rules by politicians embarking on a new course. The previous preoccupation with the consequences of electoral systems gave way to ever growing attention to the political origins of electoral laws. Numerous theoretical,6 cross-national,7 8 9 comparative studies of a few cases, and case studies have examined how politicians select and alter the institutions that govern voting choices. However, as is well known, institutions are robust and resistant to change.10 Substantial trauma is often required to initiate significant institutional innovations - such as the 'punctured equilibrium' in development REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 409 witnessed in the collapse of authoritarian and communist regimes around the world in the 1980s and 1990s. In these situations, institutions created de novo are often the result, including the implementation of electoral systems that establish new bases of political authority. In the absence of such dramatic political transformation, more incremental procedural change is likely by adjustments in components of the electoral system, for example district magnitude or threshold of representation.11 I posit that innovation or persistence in post-communist institutional engineering follows this model, with more numerous and large-scale changes at the time of the founding elections, and fewer and less significant alterations after the establishment of post-communist regimes. What motivates political leaders to initiate new systems of competition? The prevailing consensus favours instrumental, partisan interests as the basis for electoral change, rather than normative commitments to pluralist politics. In this scenario, actors' preferences are to maximize self-interest through specific institutional design and protect the standing of established 13 parties in the face of rival groups. " The adaptation and alteration of rules is thus a contested process, driven by the evaluation of potential gains or losses of vote and seat shares. In this view, electoral design is much less a co-operative enterprise concerned with the collective good of the democratic project, the fairness of competition or equal representation than a self-interested move to secure partisan advantage. However, the problem is that the selection of rules is fraught with considerable uncertainty about deviations from expected outcomes.14 This is especially so during episodes of system breakdown and regime transition. The rupture with the past ushers in new actors determined to obtain a stake in the public domain, favouring rules that open access to policy making. Furthermore, the abrupt change creates an institutional vacuum that culminates in the search for alternative decision rules.15 Most significantly, uncertainty about the position of political formations and potential constituencies reinforces the need to secure power. In these circumstances, the strategic consequence of uncertainty is to minimize risk. In electoral engineering, the adaptation of permissive proportional representation systems is the best alternative to ensure survival, rather than restrictive majoritarian systems whose high exclusion thresholds create more evident losers.16 The exception is when there is a clear and known domination of the political scene by one or two actors who can benefit from the majoritarian system that rewards the dominant 17 parties and undermines the access to power of other players. In contrast, in conditions of indeterminate power distribution during the transition, the rational endeavour by political entrepreneurs is to select decision rules that provide at least a minimum guarantee of inclusion in politics. Proportional representation (PR) is the electoral system that best gives 'all participants higher opportunities to obtain a share of power'. Accordingly, preference 410 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS in electoral design is to move away from majoritarian towards PR systems to avoid risk in conditions of uncertainty and institutional breakdown. The post-communist political environment was especially prone to electoral engineering as a result of the discrediting of the previous system's legitimacy - in cases where the opposition assumed a prominent role or in conditions where the former communist elite retained substantial influence. The situation was further complicated by the multiple transitions facing most post-communist states. The economic and social restructuring alongside the political transformation was often an extensive, complex process that left many societal groups unclear about their interests and future payoffs, including potential constituencies uncertain about their own fortunes and preferences. The political transition was often accompanied by the emergence of a multiplicity of long-suppressed voices seeking to enter the political arena through better guarantees of representation. This state of affairs afforded ample opportunities for extensive innovation in electoral design. Indeed, the post-communist period has emerged as one of the most concentrated periods of alteration in systems of competition and governance ever seen. The rapid deployment of new electoral rules and party systems provides a rich opportunity to examine the nature and extent of electoral design and reform. Empirical Data on Electoral Engineering This study undertakes a comprehensive analysis of the electoral systems in all post-communist states, from the time of the initiation of the new regimes in the post-1989 period until the end of 2005. Since the primary interest is the influence of different regime types on the selection and alteration of election procedures, the study includes regimes that are classified across the range of political outcomes, from democratic to semi-authoritarian to authoritarian. In that respect, the approach differs from studies of electoral engineering which concentrate on democratic systems.19 One of the aims here, however, is to uncover whether regimes that were more resistant to the democratic opening were able to resist by instituting electoral procedures that were less inclusive, for example majoritarian rather than proportional representation formulas. For that reason, the study is based on all 28 states emerging out of the communist experience, including East Europe, the former Yugoslavia, the former Soviet Union and Mongolia, allowing for maximum differentiation among regime types of the former communist bloc. The classification of the regimes relies on the combined Freedom House scores of political and civil rights, supplemented by verification on the authoritarian-democratic scale 20 of the Polity IV data. The scheme enables distinction among three types of political systems, depending on the score in both data sets. The Freedom House scale, embracing 'free' states with democratic regimes (1 to 5 score), REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 411 'partially free' states with semi-authoritarian regimes (6 to 10) and 'not free' states with authoritarian regimes (11 to 14) is highly correlated in virtually all cases, with the evaluation of regimes along the authoritarian-democratic continuum in the Polity data set classifying polities on the basis of an additive score of authoritarian and democratic features, with -10 to -4 as authoritarian, -3 to 5 as semi-authoritarian, and 6 to 10 as democratic. While regime type constitutes the independent variable in the study, the dependent variables are the electoral system and its most significant properties. In this regard, I follow Lijphart's approach in examining four dimensions of the electoral system: first, the all-important formulas distinguished along majoritarian, proportional and mixed approaches to the selection method; in addition, the emphasis is on three components of the electoral laws that have a significant bearing on the translation of votes to seats: district magnitude, assembly size and legal threshold. More inclusive electoral procedures are built around higher magnitudes, larger assembly chambers but lower, if any, bars for entry into the legislative arena. The first two elements provide a larger number of available seats per district and the national legislature, thus allowing smaller political parties a better chance at representation. Similarly, a low threshold - or the lack of a threshold - means that there are no artificial barriers for entry into the political space. Lijphart's criterion was to define significant modification along any of these three dimensions as a 20 per cent alteration in magnitude, chamber size or threshold. Such change is deemed sufficiently significant to provide a new structure of incentives for 22 the competing political actors. The same reasoning is applied here to the post-communist states, so that alongside alteration in electoral formula, changes of 20 per cent or more in the electoral properties take on the characteristics of distinct voting systems. The study proceeds by first examining the selection of electoral formulas for the first, founding elections of post-communism - although 'founding' is not meant to convey the adoption of a democratic process but rather a new mechanism of establishing political authority in the aftermath of communism. This is contrasted with the electoral systems existing at the last election before the end of December 2005, so as to determine persistence versus reform in the original design. The approach is twofold. First I examine the selection of electoral designs at the initial stage of the transition, when opportunities for extensive innovation were present to establish a novel institutional structure. Second, the analysis concentrates on the issue of reform to determine the robustness of the original design in the face of political developments associated with either evolving democratization or the consolidation of new forms of authoritarianism. I turn first to an examination of the electoral formulas, followed by the other features in the design of institutional choice, namely magnitude, assembly size and threshold. 412 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS Electoral Design at the First (Founding) Election The exclusive electoral rule operating in the communist states prior to their collapse was the absolute majority run-off system, so all the states entering the new era had a common institutional heritage. - However, precisely for that reason, different political forces diverged substantially on the merits of the single-member district (SMD) system for electoral politics. As noted previously, institutional choice is often governed by the distribution of power at the time of the transition, between forces associated with the previous regime and opposition political actors. So power considerations after the collapse of communism defined the preferences of electoral designs. Where the political opposition was able to amass strong support and launch a democratic challenge, the pressure was for the entry of new voices into the political space, reinforcing the demand for more permissive electoral procedures associated with PR formulas in full or as part of a mixed system. Where the former communist elites were able to maintain a predominant position and establish a new form of authoritarianism, the effort was to maintain more restrictive procedures of competition - the majoritarian type - so as to retain political advantage. Where neither the incumbents nor the opposition were clear about the distribution of forces and potential payoffs, the uncertainty left open the nature of the new regime. In these circumstances, the unknown produced a wider array of options in electoral design, enabling PR or SMD formulas as well as mixed systems incorporating both majoritarian and proportional representation electoral rules. Accordingly, the selection of the post-communist institutional mechanisms depended substantially on the power of political actors and the emerging regime types. The former communist elite, whether in its reconstructed version or not, saw its advantage in the organization of the former ruling party and in the personalization of politics around well-known candidates, thus favouring an election procedure based in single-member districts that placed the spotlight on individual candidates and on organizational capacity while de-emphasizing party-list formulas that accentuated the 'communist' label and its negative heritage. In contrast, the democratic opposition could not count on an organizational or personality advantage, and instead sought to highlight its contribution as a force undermining communist power. This political lineage was a benefit in the evolving competitive environment, and culminated in the opposition's preference for a system where voter choice focused on party lists through proportional representation. The procedure reinforced the conduct of politics around the emerging parties and provided greater assurance for political inclusion. In summary, the distinctive pulls towards different electoral systems lead to the following hypothesized relationships: regimes experiencing an open REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 413 process of transition and that are 'democratic' will opt for a permissive system of political contestation, namely a PR electoral formula, and regimes where change is forced upon a recalcitrant elite and that are classified as 'authoritarian' will opt for a restrictive process and favour majoritarian formulas. 'Semi-authoritarian' regimes defined by opposition pressures but also with elite continuity face the option of either permissive or restrictive systems, depending on the distribution of forces, or settle on a compromise version of a mixed system utilizing both majoritarian and proportional elements. The distribution of these possible design outcomes by country is summarized in Tables 1 and 2. Table 1 provides a typology of regime type by electoral formula that supports the expected relationships. The distribution of cases is clearly in the proper cells, with formulas reflecting the predicted distribution by regime type. Six out of the 28 states fell under the free, democratic rubric at the time of the founding elections, and all conducted elections through the PR method in full or as part of a mixed electoral arrangement. Moreover, the states that embarked on the path of democratization by means of full PR elections - the Czech Republic, Poland, Slovakia and Slovenia - all opted for the more permissive type of proportional representation through the largest remainder Hare or Droop methods of converting votes to seats. The same PR formulas were introduced in the mixed systems of Hungary and Lithuania, where the proportional representation tier coexisted with a majoritarian electoral formula, with the portion of PR accounting for 54 per cent and 50 per cent of seats. In all the democratic regimes at their establishment, then, the choice in the proportional representation design TABLE 1 FIRST ELECTIONS: REGIME DIVERSITY AND ELECTORAL SYSTEMS Majoritarian Mixed PR Democratic Hungary Lithuania Czech Republic Poland Slovakia Slovenia Semi-authoritarian Albania Belarus Kyrgyzstan Macedonia Mongolia Ukraine Armenia Bulgaria Croatia Georgia Russia Bosnia Estonia Latvia Moldova Romania Authoritarian Kazakhstan Tajikistan Turkmenistan Uzbekistan Azerbaijan Yugoslavia 414 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS featured a conversion method that did not penalize smaller parties, unlike the PR d'Hondt formula that favours larger political parties over smaller ones. In that sense, the choice among these six cases was for the more permissive method of vote-to-seat calculation within the PR family, a process more favourable to proportionality and inclusion. In contrast, the electoral mechanism for the regimes defined as authoritarian is primarily in the majoritarian column, with two cases adopting a mixed electoral structure. The latter two differ in both the share of the vote and the selection method of the majoritarian component: in Azerbaijan, 80 per cent of the seats were determined by the run-off formula; while in Yugoslavia 44 per cent of the seats were selected on the basis of first-past-the-post plurality. This distinction supports the assessment that while the latter was an authoritarian regime aiming to constrain the opening of the political process, it was not able to foreclose opposition preferences to the same extent as the authoritarian elites in Central Asia.26 Indeed, the four cases combining authoritarianism with majoritarian electoral choices all come from this region, so authoritarian rulers' preferences were towards restrictive electoral institutions to enable the dominant political force to garner the majority of seats in the legislature. In distinction to the predominant choice of electoral design by the democratic and authoritarian regimes which fell respectively towards PR and majoritarian formulas, the semi-authoritarian regimes depicted in Table 1 conform to the predicted typology by straddling the choices across all three electoral forms: majoritarian, mixed and proportional representation. The distribution of cases is equivalent across these three choices, with six regimes opting for majoritarian formulas, five for mixed systems, and five for PR mechanisms; confirming that semi-authoritarianism is subject to greater uncertainty about the distribution of political forces, reflected in the diversity of choices in electoral design. In general, then, the evidence at the initiation of systemic transformation reaffirms the predicted relationships between regime and electoral design. The regimes that replaced the communist authority took on different political shapes that reallocated power to a different degree between the heirs and the opponents of the former system; these new structures of power were charged in turn with adopting new decision rules. The design of electoral institutions was a factor of these power relations: at one end of the spectrum, the new authoritarian regimes favoured more restrictive rules while, at the other, democratic regimes installed more permissive selection procedures. Electoral Rules at the Last Elections How did the initial array of competitive rules fare during the evolution of the post-communist systems? The status of the electoral designs at the time of the REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 415 last elections before December 2005 demonstrates that the pattern of choice at the founding moment is carried over to the subsequent periods of electoral design (Table 2). The evidence showcases the link between regime and electoral design, as the typology reaffirms the nexus between democratic regimes and proportional representation, authoritarianism and majoritarian formulas, and semi-authoritarianism and mixed or PR election rules. At the same time, there is a clear movement in the direction of the proportional formulas, as 13 states operate according to full PR systems and nine according to mixed systems that include a significant share of electoral choice by the PR method. The pattern is due to two developments. One is linked to the evolution of regime types from semi-authoritarian to democratic regimes while maintaining the PR system (Estonia, Latvia and Romania). The other is due to a shift to proportional representation procedures (Bulgaria, Croatia) that accompanies movement from semi-authoritarian to democratic regimes. Together, these trends explain the predominance of countries with PR election mechanisms at the last election cycle. The association between regime and election rule is due to the evolution of post-communist states towards more entrenched pluralist politics, substantiating the predicted norm of democratic regimes favouring more inclusive voting procedures. This tendency, however, is not uniform. Hungary and Lithuania, for example, demonstrate stability in TABLE 2 LAST ELECTIONS: REGIME DIVERSITY AND ELECTORAL SYSTEMS Majoritarian Mixed PR Democratic Mongolia Hungary Lithuania Bulgaria Croatia Czech Republic Estonia Latvia Poland Romania Slovakia Slovenia Semi-authoritarian Albania Armenia Georgia Russia Ukraine Bosnia Macedonia Moldova Yugoslavia Authoritarian Azerbaijan Belarus Kyrgyzstan Turkmenistan Uzbekistan Kazakhstan Tajikistan 416 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS democratic regime and mixed election design. These cases are the only democratic regimes that opted for a mixed system at the initiation of pluralist politics, and continuity in their original design demonstrates the persistence of institutions during periods of political stability. Since there were no significant political disequilibria in these instances, there was little incentive to change the existing rules; instead, the mixed systems remained in place. In contrast, those countries whose political conditions altered in favour of full democratic practices adapt to the transformation by embracing elements of the PR selection process to assure more inclusive politics, as happened early in the transition in Bulgaria and later on in Croatia. Mongolia is an exception as well, as the shift from a semi-authoritarian to a democratic regime remains bound to the majoritarian voting process. In this case, the persistence of the original majoritarian formula is tied to the dominance of the two large political actors in domestic politics, representing the former communists and the democratic opposition, who have little incentive to open up the electoral system to wider representation by an alternative selection process. The Mongolian case is important for its affirmation that majoritarian electoral provisions can serve the cause of democratization even in conditions of political transition. This is not the case in the other instances of majoritarian electoral institutions, which are all found in states with the worst record of pluralist politics in the former communist world. Table 2 reveals that the most authoritarian regimes in the region, such as Belarus, Turkmenistan or Uzbekistan, opted for an exclusively majoritarian form of elections, while the other two countries in the 'authoritarian' category have a mixed system but with predominant majoritarian tiers, 87 per cent in Kazakhstan and 65 per cent in Tajikistan. Obviously, authoritarian regimes prefer to rely on the winner-takes-all mode of elections to secure their political domination. With the exception of Mongolia, no regime classified as democratic or semi-authoritarian has chosen to implement a purely majoritarian system of voting. A comparison of Tables 1 and 2 reinforces the point, as there is an obvious shift over time among semi-authoritarian regimes away from majoritarian formulas, so that the semi-authoritarian-majoritarian group consisting in the first election of six cases is devoid of cases at the last election. This is due either to the migration of regimes with majoritarian systems from semi-authoritarian to full authoritarian status, as for example Belarus and Kyrgyzstan, or to the more common path of moving away from majoritarian towards PR, either as part of a mixed system (Albania, Ukraine) or to full PR (Macedonia, Yugoslavia). Both paths, the descending one of semi-authoritarianism to authoritarianism and the ascending one to more open politics, confirm that authoritarian regimes select majoritarian systems, while democratic polities prefer proportional representation rules. REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 417 The distribution of cases among the three principal variants of electoral systems underscores the strong migration over time towards more inclusive election rules. The states opting for the PR system are concentrated in Eastern Europe, where the transformation from communism was led more evidently by opposition forces or gave rise to mass movements. The transition here was infused with the demand to open the political process to a multiplicity of voices, thereby strengthening the propensity to design election procedures favouring broad representation. By the last election cycle, the East European states had all shed the majority electoral systems inherited from the communist period, in most instances moving to full PR procedures. There has been no reverse trajectory from proportional to majoritarian selection, although a number of attempts in that direction were proposed by emerging dominant forces in the new democracies. The failure of these reforms affirms once more the robust nature of institutions. Beyond that, it offers evidence that the opening of the political process to numerous actors is difficult to restrain subsequently by reforming the election formula, as these actors fear that a shift away from permissive rules of competition will endanger their presence in the legislative arena. In this, developments in the post-communist camp reflect a world-wide trend that guards against the uncertainty of outcomes by means of a shift from majoritarian systems, which divide the political actors into winners and losers, to proportional representation systems, which provide 27 a better opportunity for political access to more segments of the polity. Reform of Electoral Rules The comparison of electoral formulas at the start and end points of the post-communist period demonstrates that, while there is alteration at the election system level, there is also considerable endurance in the original design. This raises the corollary issue of whether change is more evident in the components of the electoral system, namely district magnitude, legal threshold and assembly size, since the expectation is that modifications within the system are easier to initiate than transformation of the formula. For that reason, reforms of the various properties of the electoral structure are likely to be more evident than changes in formulas. To recall, change is driven by the uncertainty of competitive politics, so the prediction is that reforms are most likely where politics is highly contested and voters' preferences are unpredictable. In contrast, where the distribution of power is known and entrenched, there is little incentive to amend the system that already works to the benefit of the political winners. Authoritarian regimes, therefore, will have the least reason to institute change, while democratic regimes may seek adjustments in the electoral process to remedy miscalculations induced by uncertainty. Among the 418 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS latter, there is considerable evidence for the saliency of uncertainty in post-communist politics. The abrupt transition from closed to more open politics produced calculations without firm grounding in experience - a leap into the future that required readjustment after the first result of electoral contestation. Beyond that, transition conditions created opportunities for many new political forces to try their fortunes in the political game, often producing extensive fragmentation and high volatility in the party systems that under- 28 mined political stability. In turn, these developments ushered in concerns to ensure stability in the political process by means of greater efficacy and better governance, through a reduction of the divisions in the political space. In that way, the electoral system was subject to two distinct pulls, representation and efficiency. The first demand favoured a more permissive electoral process allowing for multiple political parties; the second favoured more restrictive rules to control the entry of political players. Reform of the electoral design in the post-communist period was subject to these two contradictory incentives. Uncertainty favoured electoral adjustments towards assurances of representation by means of inclusive electoral procedures, for example PR with large district magnitude, with low thresholds and large legislative chambers. Stability required electoral changes that favoured larger parties and effective government by foreclosing opportunities for entry by instituting majoritarian systems, or low magnitude districts in PR systems, or higher thresholds for representation in the legislature. Table 3 offers a synopsis of reforms for the electoral formula and its component properties for the 112 elections in the region between 1990 and 2005. At the system level, the changes reaffirm the direction of change by regime type between the first and the last elections. Changes to different formulas favour the incorporation of PR among semi-authoritarian regimes, with few reforms at the formula level among democratic and authoritarian regimes. Some exceptions occur as in Azerbaijan 2003 and in Kyrgyzstan 2005, when the authoritarian regimes reverted from mixed systems with some proportional representation to full majoritarian systems, as expected on the part of entrenched elites who favour more restrictive procedures to consolidate power. Among democratic regimes that had committed themselves to exclusive PR or mixed systems with significant proportional representation, there are no changes from one formula type to another - although there were alterations in the counting rules within each formula type, to the benefit of the larger parties. Clearly, while retaining the commitment of permissive representation, these reforms were a nod in the direction of stability by the introduction of rules making the penetration of the legislative space by smaller or new political parties more difficult. Evidence on the other components of the electoral system confirms significant alterations (that is, changes of more than 20 per cent) in magnitude, TABLE 3 PROCESS OF ELECTORAL SYSTEM DESIGN Number of changes in Status at No. of System PR Majorit. Mixed District Assembly first election First election Last election elections formula type type type magnitude Threshold size Democratic Czech Rep. PR PR 5 1 1 Hungary Mix, 54 per cent PR Mix, 54 per cent PR 4 1 Lithuania Mix, 50 per cent PR Mix, 50 per cent PR 4 1 1 Poland PR PR 5 2 2 1 Slovakia PR PR 5 1 1 1 Slovenia PR PR 4 1 1 Sub-total 27 0 5 0 1 4 5 0 Semi-authoritarian Albania Majoritarian Mix, 29 per cent PR 6 1 1 2 2 1 Armenia Mix, 21 per cent PR Mix, 43 per cent PR 3 1 1 Belarus Majoritarian Majoritarian 3 1 Bosnia PR PR 4 1 Bulgaria Mix, 50 per cent PR PR 6 1 1 Croatia Mix, 48 per cent PR PR 4 1 2 1 Estonia PR PR 4 Georgia Mix, 64 per cent PR Mix, 64 per cent PR 4 1 1 2 Kyrgyzstan Majoritarian Majoritarian 3 2 2 Latvia PR PR 4 1 Macedonia Majoritarian PR 3 2 1 1 Moldova PR PR 4 1 Mongolia Majoritarian Majoritarian 5 2 2 Romania PR PR 5 2 O m oo > Ö m r M o H O > r oo oo H M oo O oo H O o oo (Continued) ^ O TABLE 3 CONTINUED O a z; > r o O o z; 00 H 00 H Ö M 00 > Ö H > 00 H O Z ^ O r H o 00 Number of changes in Status at No. of System PR Majorit. Mixed District Assembly first election First election Last election elections formula type type type magnitude Threshold size Russia Mix, 50 per cent PR Mix, 50 per cent PR 4 Ukraine Majoritarian Mix, 50 per cent PR 3 1 Sub-total 65 8 0 2 2 10 10 6 Authoritarian Azerbaijan Mix, 20 per cent PR Majoritarian 3 1 1 Kazakhstan Majoritarian Mix, 13 per cent PR 4 1 1 1 Tajikistan Majoritarian Mix, 35 per cent PR 3 1 1 Turkmenistan Majoritarian Majoritarian 3 Uzbekistan Majoritarian Majoritarian 3 1 Yugoslavia Mix, 60 per cent PR PR 4 1 3 Sub-total 20 4 0 1 0 3 1 3 Total 112 12 5 3 3 17 16 9 REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 421 threshold and assembly size that together outweigh the formula reform (Table 3). These changes are not as strongly unidirectional as the alteration in formulas favouring proportional representation. Furthermore, the direction of reforms on these electoral properties is more towards restrictive rather than permissive decision rules, in effect constraining some of the benefits of opening the political process by the shift from majoritarian to PR systems. These reforms conform to the fact that the extensive fragmentation and volatility of party systems in many former communist states create pressures to trim the permissive nature of the formula by imposing constraints through other procedures. This trend is especially evident in threshold and assembly size reforms that restrict access, in comparison with alterations in district magnitude that operate in both exclusory and inclusory directions. The most frequent change is district magnitude, a trend expected as magnitude is thought to have the greatest impact on the structuring of party systems. Yet there is no uniform direction here, the data show that the 17 reforms at times constrain but more often facilitate political competition, so that variations in district size are the most ambiguous of the reforms undertaken in the electoral systems of post-communism (Tables 3 and 4). On average, the reforms are implemented with the same frequency (one alteration for about six elections) in all three regime types, although in absolute terms the number is concentrated among semi-authoritarian regimes as this sort is the most numerous at the start of the transition. Among the seven states in this category, three revised the district size twice, some reverting to the original magnitude after an interlude of reform. Mongolia, for example, went from a single-member district in 1990 to multi-member districts of two to four members by plurality in 1992, only to restore single-member districts for the 1996 election. In a similar back and forth process, Albania revised its single PR district from a magnitude of 40 to 25 to 40 again for the 1992, 1996 and 1997 elections. In these cases, the change was only a temporary abridgement of a more permanent design. In Albania the effort towards a more restrictive decision rule was due to the attempt by the elites to consolidate their position, a manoeuvre which failed thanks to a strong backlash that questioned the fairness of the new rules. In Croatia, the two reforms in PR magnitude accompanied a change from a mixed to a fully proportional system, shifting from a single PR district of 80 to ten districts with 14 seats each. While this appears as an important exclusory move, the elimination of the SMD component of the mixed electoral system to full PR redressed the effect in favour of an inclusive process. That being the case, the reform was integral to the opening of Croatian politics at the end of the Tudjman era. The overall trend in reforms of district magnitude among the semi-authoritarian regimes is in the direction of more inclusive politics, as several of these regimes moved to more pluralist forms of politics. Most 422 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS pronounced is the change in Georgia for the 1999 election, from ten districts with 15 seats each to a single nation-wide district of 150 seats, a clear reflection of the pressures on the Georgian regime to open the process to opposition forces - a development that culminated in the coloured revolution of 2003. A similar trend occurred in Yugoslavia, where the changes in district magnitude for the three elections between 1992 and 2000 are directly related to attempts to manipulate access to the site of power, for example the imposition by Slobodan Milosevic of lower district magnitude for the 1996 election in order to constrain the political opposition. Yugoslavia was the only case among regimes defined initially as authoritarian that sought to manipulate the political process by adjusting electoral districts, since it was operating according to mixed rules that included a large PR component. All other authoritarian regimes were heavily reliant on the majoritarian system as the most restrictive form of electoral politics, and have adhered to single-member districts throughout the post-communist period. The three democratic regimes that have altered their district magnitude have moved towards both restrictive and permissive electoral procedures. Poland has altered its districts twice, first for the 1993 elections by reducing the average magnitude from 10.6 to 7.5 while maintaining the same magnitude of 69 for the national district, then for the 2001 election by increasing average magnitude to 11.2 at the lower tier and eliminating the upper one. These changes sought to curb the multipartism and volatility of the Polish party system, one of the most unwieldy in Eastern Europe/ The change in the Czech Republic for the 2002 election resulted in an important decrease in district magnitude, from 25 to 14, reflecting a restrictive effort on the part of 'winners' in a more consolidated Czech party system than its Polish equivalent. In these instances, the aim of the magnitude shifts has been to introduce greater effectiveness or to enforce stability in the political system. An opposite trend has been in evidence in Slovakia, which adopted a new electoral law for the 1998 legislative vote that altered the four districts with an average magnitude of 37.5 into a single nation-wide PR district of 150 seats. This provided a significant increase of opportunities for political entry by new and smaller parties, and was a reflection of the change in the political situation of the republic at the end of the Meciar era.31 The evidence for reform in the democratic regimes, at least in relation to district magnitude, testifies to a variety of motivations to affect the electoral process, and underscores the crucial import of political context in institutional design. Among all the components of electoral systems, magnitude is the property with the greatest influence on voting results. The fact that alteration in district magnitude is most evident in the semi-authoritarian regimes that are more prone to political evolution, and that persistence in district magnitude is most common among the more stable authoritarian and democratic regimes, reflects prevailing political conditions. REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 423 As for the other elements of the electoral systems, there are only half as many changes in assembly size as in magnitude among the electoral systems in Eastern Europe and the former Soviet Union (Tables 3 and 4). In this case, however, the modifications are virtually uniformly in the direction of smaller legislative chambers. The effect is a more restrictive electoral process since there are fewer available seats for the contending parties. The reforms on this dimension are heavily concentrated among the semi-authoritarian and authoritarian regimes. Bulgaria in 1991 is the only democratic regime to have instituted a smaller legislative chamber: there the shift is from the exceptional Constituent Assembly of 1990, devised to create a new constitutional order, to a smaller working parliament. In the other countries, the number of seats declines significantly to foster a restrictive electoral environment. The one exception is Kyrgyzstan, where the initial number of legislative seats is unusually low - 35 in 1995, and the two reforms increased the assembly size, first to 60 in 2000, then to 75 for the 2005 election (subsequently invalidated). While this does make the electoral system more inclusive over time, even in conditions of authoritarianism, the parliament in Kyrgyzstan remained among the smallest chambers in the post-communist universe, and thus still highly restrictive in comparative terms. In all the other cases of reform of the assembly size among authoritarian and semi-authoritarian regimes, the direction of change is towards a more exclusory electoral process. As many of these regimes have successfully entrenched the power of the incumbent elite during the post-communist era, so they have constrained the legislative arena to the entry of opposition forces. The restrictive trend is most pronounced among the strictest regimes in the region: for example, in Belarus from 260 to 110 seats between the 1995 and 2000 elections, in Tajikistan from 181 to 63 seats in the same period, in Uzbekistan from 250 to 120 seats between 1999 and 2004, and in Kazakhstan from 177 to 77 seats from 1994 to 1999. There is thus no doubt that revisions in assembly size have been used in these instances to foreclose political openings and ensure the malleability of the legislative chambers. The legal threshold as a minimum requirement for legislative representation is another mechanism for exclusion. In fact, the legal bar has been extensively used in the post-communist PR electoral systems to prevent small political parties from cluttering the legislative arena. The practice already occurred in the original electoral design immediately after the collapse of communism, as well as through a series of reforms in states that have adopted or raised the threshold during the subsequent period (see Table 4). Since the device applies only to PR systems, it is not encountered among the authoritarian regimes that employ majoritarian electoral formulas. Thus, the particularly high eight per cent threshold used in Azerbaijan's PR tier had disappeared by the last election, since the electoral system reverted 424 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS TABLE 4 CHANGES IN DISTRICT MAGNITUDE, THRESHOLD AND ASSEMBLY SIZE" Assembly Magnitude Threshold size Status at PR percentage Number of last election at last election First Last First Last First Last elections Democratic Bulgaria 400 240 6 Croatia 1 4 60 14 3 5 Czech Republic 25 14.28 5 5* 5 Estonia 4 Hungary 54 4 5 4 Latvia 4 5 4 Lithuania 50 4 5* 4 Mongolia 5 Poland 10.57 11.22 0 5* 5 69 Romania 0 5* 5 Slovakia 37.5 150 3 5* 5 Slovenia 3 seats 4 4 Semi-authoritarian Albania 1 1 250 140 6 29 40 2.5* Armenia 1 1 190 131 3 43 40 56 Bosnia 21 3.75 4 12 Georgia 1 1 4 64 15 150 0 7 Macedonia 1 3 20 0 Moldova 4 6* 4 Russia 50 4 Ukraine 1 1 3 50 225 4 Yugoslavia 1 4 8.4 5.11 5 Dist. 5 Dist. Authoritarian Azerbaijan 1 1 3 25 8 Belarus 260 110 3 Kazakhstan 1 1 177 77 4 13 10 7 Kyrgyzstan 35 75 3 Tajikistan 1 1 181 63 3 35 22 5 Turkmenistan 3 Uzbekistan 250 120 3 *Threshold requirement for single party; higher requirements for coalitions. ** Where relevant the changes for mixed or tiered PR systems are provided on separate lines for each tier. REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 425 from a mixed to a full majoritarian procedure. The threshold continues to be present in the two mixed systems in this cohort, with Kazakhstan utilizing an unusually high restrictive minimum of seven per cent, and Tajikistan employing the more common five per cent threshold. Even in these two countries, where PR is but a small part of the mixed electoral design, the proportional representation tier enforces additional restrictive demands on political actors to help sustain the dominance of the party in power. A similar tendency is evident among the semi-authoritarian systems, where the threshold has been introduced and increased on a number of occasions to constrain entry into parliament. For instance, the requirement to clear the representation bar was increased significantly in Georgia prior to that nation's coloured revolution, first from zero to five per cent and then to seven per cent in the three elections of the period 1992-99. In Ukraine, as well, a four per cent threshold was put into operation as the country reformed its electoral formula from majoritarian to mixed tiers in 1998. In these contexts, the manoeuvre confirmed the desire of the authoritarian elite to minimize political challenges by foreclosing options for representation. In contrast, in Albania and Macedonia the threshold was lowered, introducing more permissive electoral contestation. In both states, the reform of electoral processes coincided with a relaxation of the political climate from an entrenched semi-authoritarianism towards the more pluralist end of the regime continuum. The opposite trend has taken place in polities with already well-established open politics, where increases in the minimum threshold of representation have been used to stabilize volatile party systems. The democratic regimes in the post-communist region employ legal bars to preclude the further dispersion of the political space, with virtually all thresholds at the last elections requiring a five per cent vote for single parties as a condition for entry into parliament (Table 4). In several states, including Bulgaria, Lithuania, Poland, Romania and Slovakia, the minimum is even higher for party coalitions, an additional electoral barrier to prevent the circumvention of the threshold by political alliances. The intention is to curb the fragmentation of the party systems by excluding smaller actors from political access. In that way, the more demanding minimums for representation among the post-communist democratic regimes are implanted to assure the stability of pluralist politics. Most democratic regimes in the region have increased the threshold at some point in the aftermath of the founding elections. Among these steps was the imposition of a threshold where there was none previously, to stem the proliferation of parties that undermined political stability. The most notorious example is that of Poland in the aftermath of the 1991 election, when the absence of any legal bar gave rise to dozens of parties competing for a chance 426 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS to enter parliament, and culminating in 29 parties succeeding in the endeavour. In large part as a reaction to the consequent instability, a threshold of five per cent for parties and eight per cent for coalitions was introduced before the premature 1993 elections. In Romania, as well, the lack of a minimum requirement at the first post-communist election gave way to a three per cent threshold for the 1992 and 1996 elections, and a subsequent increase to five per cent starting with the 2000 contest, with higher requirements for party coalitions. In the other democratic regimes, there was a threshold requirement in place for the founding election, but only Bulgaria and Estonia did not tamper subsequently with the legal bar. In the remaining cases, there was an increase in the minimum requirement for representation in the legislative chamber, generally through a rise from four per cent to five per cent, but also larger increments in Croatia and Slovakia or more stringent requirements for coalitions, as in the Czech Republic or Lithuania. These data point to the universal use of formal thresholds of representation among the democratic regimes, and also reveal consistent adjustments over time to raise the bar for entry into parliament. The threshold is employed among the open political regimes in the post-communist region to prevent small parties and new aspirants from cluttering the political system. The legal bar is more prevalent than the other measures of exclusion, such as low district magnitude or assembly size. The threshold is also an electoral rule that is employed much more frequently and at higher levels than in other democratizing regimes, whether in West Europe or Latin America/ The practice is an especially significant innovation in electoral design applied in the post-communist states. It can be surmised that this mechanism was a deliberate attempt on the part of political elites and electoral designers to reduce the radical fragmentation of the political space with the blossoming of pluralist politics. The explosion of political actors at the start of the transition accentuated the uncertainty of the system, as voters experienced considerable co-ordination problems in electoral choices, given the absence of historical ties and partisan identifications between the electorate 33 and parties." In order to curtail the disruptive effect of too many actors, the choice in electoral design was to stabilize the system by reducing the fragmentation through the imposition of minimum thresholds of representation. In effect, the threshold was a signal to political actors and to voters to discourage choices for contenders unable to clear the minimum requirement, thereby forcing more strategic actions by elites and voters. From the point of view of electoral design, the utility of the threshold over other forms of electoral management is its transparency. Simply put, minimum bars are easier to comprehend as regards their effect on party fortunes than adjustments in district magnitude or rules for calculating the vote-to-seat conversion. By first introducing and then raising thresholds, the new democracies in the REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 427 post-communist world were signalling their concern for stability and effectiveness, alongside their commitment to representation by PR systems. Conclusion Electoral engineering in post-communism has followed a predictable pattern that emulates the dominant world-wide trend towards enhanced provisions of proportional representation, through the selection and institutionalization of PR formulas over time, to the detriment of majoritarian electoral systems. Regimes successful at democratization are especially likely to turn to proportional representation as a primary means of determining voters' preferences in political contestation. While the finding may appear tautological, the fact remains that many established democracies around the world, such as the Anglo-Saxon countries and other states influenced by that political tradition, continue to embrace majoritarian formulas as the preferred selection method for political candidates/ None the less, despite this permissive trend, a dualism persists in the electoral engineering of the post-communist states. Alongside the turn to more permissive, inclusive PR systems among democratic regimes, the penchant for the more restrictive, exclusory majoritarian systems persists among the authoritarian regimes and to a lesser extent the semi-authoritarian regimes of the region. The introduction of electoral choices in these conditions demonstrates a preference among the power elites for decision rules that heed the prevailing global ideology of 'electoral democracy', but constrain representation by instituting winner-takes-all voting procedures that limit the opposition's access to the political process. The divide in the design of electoral rules reflects strongly a geopolitical and historical division between the states of Eastern Europe and those of the former Soviet Union. Among the former, including the Baltic States, the selected designs have been the PR electoral formulas or mixed formulas with a significant PR tier. Albania remains the only exception in this part of the region, testifying to the problematic nature of transition politics in that country. In contrast, the post-Soviet states are squarely aligned along an axis of majoritarian or mixed systems with prevailing SMD tiers. All these patterns demonstrate a strong connection between regime type and electoral design: proportional representation is associated with an open, pluralist political process that helps with the consolidation of democracy; by contrast, most authoritarian regimes cling to majoritarian formulas that restrict competition. While it is difficult to isolate the causal mechanism between regime type and electoral engineering, recent political events in the post-communist region demonstrate the importance of election rules for political development. The success of the recent 'coloured revolutions' turned on the opposition's ability in Yugoslavia, Georgia or Ukraine to challenge closed or fraudulent 428 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS elections, and then initiate an opening of decision rules and practices that facilitated a 'second regime' transformation towards pluralist politics/ In contrast, where the opposition was unable to challenge the conduct of the political contest effectively and undermine the existing election procedures, in, say, Azerbaijan or Uzbekistan, the authoritarian regimes continue to cling to exclusory forms of selection by a majoritarian formula. At the very least, these events show the influence that political regimes have on the design of electoral systems and point to the need to transform regimes in order to enable the reform of voting rules. Another dualism characterizes the trajectory of reforms after the founding elections, expressed on the one hand by the adaptation of more inclusive electoral formulas and on the other hand by more restrictive decision rules within the formulas. The predominant direction of change during the post-communist era has occurred at the system level with the movement away from majoritarian to proportional and mixed systems, so as to accommodate the diversity and multiplicity of political actors emerging at the end of communism. Naturally this embrace of PR as a means of translating votes into seats occurs among the most democratic regimes in the region and among those regimes that move away from semi-authoritarianism to install a more open political process. The selection is driven primarily by the concern of the newly energized political forces to preserve their access to legislative politics; so, even though the design of electoral systems is a function of self-interest, the choice of more inclusive rules helps to nurture the development of democratic tendencies. At the same time, the opening of the political process in the wake of the collapse of communism often leads to an overload of competitors, an uncertain political space, and voter confusion, contributing to an inchoate party structure. To remedy this thrust produced by PR electoral designs, the reform agenda of the democratic regimes concentrates on additional steps to contain the political instability and excessive fragmentation. In these cases, the effort is to stem extreme dispersion of political forces by instituting procedures that enhance the effectiveness of the new systems. This is achieved by maintaining the PR electoral formulas to safeguard access, but at the same time employing electoral mechanisms that enhance political order. The main method of attaining these goals is by reducing opportunities for small and new political actors and by encouraging voters to behave strategically in favour of existing and larger political actors. As a result, most of the reforms focus on a reduction in the number of available parliamentary seats and an increase in the threshold of representation. In this manner, the design and reform of the electoral process in the democratic regimes of post-communism have worked to foster both representation and effectiveness. The collapse of communism produced an expansive period of institutional engineering, matched by few other periods in history. In the new political REGIMES AND ELECTORAL SYSTEMS IN POST-COMMUNISM 429 context, 'electoral democracy' emerged as an important legitimating criterion for the emerging systems, although the development of the electoral structure was clearly defined by the nature of the successor regimes. Where the political opposition had an advantage and democratic regimes were established, the design and reform of electoral rules stressed the values of inclusion through the adaptation of permissive proportional representation formulas in the new competitive environment. In contrast, where incumbent political elites remained in control and new types of authoritarian regimes arose, there was little incentive to reform the inherited majoritarian electoral structure. Here the original design relegated other political preferences to minority status without the capacity to penetrate the political space effectively, and as a result there were few reforms to alter the system of selection that already favoured the dominant political force. NOTES 1. Terry Lynn Karl, 'Imposing Consent: Electoralism and Democratization in El Salvador', in Paul W. Drake and Eduardo Silva (eds.), Elections and Democratization in Latin America, 1980-1985 (La Jolla, CA: University of California-San Diego, Center for International Studies, 1986), pp.9-38. 2. Farreed Zakaria, 'The Rise of Illiberal Democracy', Foreign Affairs, Vol.76, No.6 (1997), pp.22-43; Steven Levitsky and Lucan A. Way, 'The Rise of Competitive Authoritarianism', Journal of Democracy, Vol.13, No.2 (2002), pp.51-65; Larry Diamond, 'Thinking about Hybrid Regimes', Journal of Democracy, Vol.13, No.2 (2002), pp.22-35. 3. Maurice Duverger, Political Parties: Their Organization and Activity in the Modern State (New York: Wiley, 1954); Giovanni Sartori, Parties and Party Systems: A Framework for Analysis (New York: Cambridge University Press, 1976); Douglas W. Rae, The Political Consequences of Electoral Laws (New Haven, CT: Yale University Press, 1967); Arend Lijphart, Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies 1945-1990 (New York: Oxford University Press, 1994); G. Bingham Powell, Jr., Elections as Instruments of Democracy (New Haven, CT: Yale University Press, 2000). 4. William Riker, 'The Two-Party System and Duverger's Law: An Essay on the History of Political Science', American Political Science Review, Vol.76, No.4 (1982), pp.753-66; Gary Cox, Making Votes Count: Strategic Coordination in the World's Electoral Systems (Cambridge: Cambridge University Press, 1997); Rein Taagepera and Matthew Shugart, Seats and Votes: The Effects and Determinants of Electoral Systems (New Haven, CT: Yale University Press, 1989). 5. Giuseppe di Palma, To Craft Democracies (Berkeley, CA: University of California Press, 1990); and Jon Elster, Claus Offe and U.K. Preuss, Institutional Design in Post-communist Societies (Cambridge: Cambridge University Press, 1998). 6. Kenneth Benoit, 'Models of Electoral System Change', Electoral Studies, Vol.23, No.3 (2004), pp.363-89. 7. Carles Boix, 'Setting the Rules of the Game: The Choice of Electoral Systems in Advanced Democracies', American Political Science Review, Vol.93, No.3 (1999), pp.609-24; and Joseph T Andrews and Robert W. Jackman, 'Strategic Fools: Electoral Rule Choice under Extreme Uncertainty', Electoral Studies, Vo.24, No.l (2005), pp.65-84. For an overview of election rules in the post-communist states, see Sarah Birch, Frances Millard, Marina Popescu and Kieran Williams, Embodying Democracy: Electoral System Design in Post-Communist Europe (New York: Palgrave, 2002). 430 JOURNAL OF COMMUNIST STUDIES AND TRANSITION POLITICS 8. Arend Lijphart, 'Democratization and Constitutional Choices in Czecho-Slovakia, Hungary and Poland 1989-91', Journal of Theoretical Politics, Vol.4, No.2 (1992), pp.207-23; and Barbara Geddes, 'Initiation of New Democratic Institutions in Eastern Europe and Latin America', in Arend Lijphart and C.H. Waisman (eds.), Institutional Design in New Democracies: Eastern Europe and Latin America (Boulder, CO: Westview, 1996), pp. 14-52. 9. Thomas F. Remington and S.S. Smith, 'Political Goals, Institutional Context and the Choice of an Electoral System: the Russian Parliamentary Election Law', American Journal of Political Science, Vol.40, No.4 (1996), pp.1253-79; and Kenneth Benoit and J.W. Schiemann, 'Institutional Choice in New Democracies: Bargaining over Hungary's 1989 Electoral Law', Journal of Theoretical Politics, Vol.13, No.2 (2001), pp.159-88. 10. Stephen Krasner, 'Approaches to the State: Alternative Conceptions and Historical Dynamics', Comparative Politics, Vol.16, No.2 (1984), pp.223-46. 11. Lijphart, Electoral Systems and Party Systems, pp.78-94. Magnitude refers to the number of seats per district. 12. Benoit, 'Models of Electoral System Change'; and Josep M. Colomer, 'It's Parties that Choose Electoral Systems (or Duverger's Laws Upside Down)', Political Studies, Vol.53, No.l (2005), pp. 1-21. 13. Stein Rokkan, Citizens, Elections, Parties (Oslo: Universitetsforlaget, 1970). 14. Andrews and Jackman, 'Strategic Fools'; and Regina Smyth, 'Engineering Consolidation: Explaining Electoral System Innovation in Post-Communist Democracies', paper presented at the American Political Science Association Convention, Washington, DC, Sept. 2005. 15. Benoit, 'Models of Electoral System Change', pp.366-9. 16. Colomer, 'It's Parties that Choose', pp.2-3. 17. Adam Przeworski, Democracy and the Market (New York: Cambridge University Press, 1991). 18. Colomer, 'It's Parties that Choose', p.4. 19. For example, see Pippa Norris, Electoral Engineering: Voting Rules and Political Behavior (New York: Cambridge University Press, 2004). 20. The rankings of Freedom House are available at and the Polity IV data at . 21. Lijphart, Electoral Systems and Party Systems, pp.10-12. 22. Ibid., p. 13. 23. Robert K. Furtak, Elections in Socialist States (New York: St. Martin's Press, 1990). 24. Przeworski, Democracy and the Market; Philip G. Roeder, 'Varieties of Post-Soviet Authoritarian Regimes', Post-Soviet Affairs, Vol.10, No.l (1994), pp.61-101; Pauline Jones Luong, 'After the Break Up: Institutional Design in Postcommunist States', Comparative Political Studies, Vol.35, No.5 (2000), pp.563-92. 25. Benoit and Scheimann, 'Institutional Choice in New Democracies', p.165. 26. Laslo Sekelj, 'Parties and Elections: The Federal Republic of Yugoslavia - Change without Transformation', Europe-Asia Studies, Vol.52, No.l (2000), pp.57-75. 27. Colomer, 'It's Parties that Choose', p.2 28. Jack Bielasiak, 'Party Competition in Emerging Democracies: Representation and Effectiveness in Post-communism and Beyond', Democratization, Vol.12, No.3 (2005), pp.331-56. 29. CeSID, 'Guide through Electoral Controversies in Serbia' (Belgrade: CeSID, 2000). 30. Jack Bielasiak, 'The Institutionalization of Electoral and Party Systems in Post-Communist States', Comparative Politics, Vol.34, No.2 (2002), pp. 189-210. 31. Kevin Deegan Krause, 'Slovakia's Second Transition', Journal of Democracy, Vol.14, No.2 (2003), pp.65-79. 32. Sarah Birch, 'Electoral Systems and Party Systems in Europe East and West', Perspectives on European Politics and Society, Vol.2, No.3 (2001), pp.35-77. 33. Olga Shvetsova, 'Endogenous Selection of Institutions and Their Exogenous Effects', Constitutional Political Economy, Vol.14, No.3 (2003), pp.191-212. 34. Matt Golder, 'Democratic Electoral Systems Around the World, 1946-2000', Electoral Studies, Vol.24, No.l (2005), pp. 103-21. 35. Sarah Birch, 'The 2000 Elections in Yugoslavia: The "Bulldozer Revolution'", Electoral Studies, Vol.21, No.3 (2002), pp.499-511 (p.505).