The Czech and Slovak Republics: the Surprising Resilience of Proportional Representation The Czechoslovak federation broke up in 1992 because, on some accounts, its two constituent republics were too divergent in their political cultures and policy preferences. While there may be some truth to this, at various times after the split the two independent states experienced very similar developments. One of these was the attempt to escape from the proportional representation system set out hurriedly in early 1990, shortly before the founding free election, which was held partly or largely responsible for the countries' alleged dependence on government by shaky coalition. In both states, there were attempts to shift toward a less proportional regime that would facilitate one-party or at most two-party government. In both countries, however, such efforts were thwarted or subverted. This chapter explains the initial selection of proportional representation, the reforms that were proposed and enacted, and the surprising resilience of the federation's institutional legacy. It argues that the initial choice of electoral system was the product of party preferences derived endogenously through bargaining and in anticipation of expected results, but that these preferences were not solely self-interested. Like many statutory institutions during transitions to democracy, the new electoral system was seen not as locked in, but open to revision as its effects were clarified and party preferences shifted, again endogenously. Shocks exogenous to the electoral system were used to justify reform, but in fact the motor was the long-standing aim of the emerging beneficiaries of the new competitive politics to push the electoral systems in a win-concentrating direction. Although parties threatened by these revisions were able to mobilize the resources to stop, overturn or 67 68 Embodying Democracy neutralize them, these episodes confirmed that statutory institutions do not account independently for the consolidation of democracy. Instead, they remain contested as long as parties distrust each others' intentions and the stakes of office are high.1 Initial choice of electoral regime Following Shugart's rational-choice model of institutional design in new democracies, we would expect a country such as Czechoslovakia to opt initially for a party-dominated rather than candidate-centred electoral system. The model starts from the fact that the Czechoslovak transition was 'provisionary', in that the removal of a frozen post-totalitarian regime was induced by mass protests and round-table talks at the end of 1989, followed by the investiture of an interim government led by outsiders, who in turn set the pace in agreeing the rules for the founding of free elections. In these conditions, the rational interest of politicians was in an electoral system that enhanced the development of strong parties. The logic behind this prediction lies in the relative anonymity of most of the new democratic activists, whose careers would be best served by trading on the name of an uncompromised new party rather than on their personal reputations.2 Since proportional representation (PR) usually empowers parties, some form of PR would be the expected first electoral system. This model is supported to a considerable degree by what happened in Czechoslovakia in 1990, although it is impossible to prove that the motivation behind legislators' choice of electoral system was indeed personal career advancement. Also, the new democratic elite's preference for PR was not as immediate and obvious as the model would lead us to expect; it first had to withstand serious challenges from within the largest movements, Civic Forum in Bohemia and Moravia and Public against Violence in Slovakia. Polls in January 1990 showed support for the two movements at almost 30 per cent of the electorate, three times more than for any other contestant; their strategists were understandably tempted to rout rival parties under majoritarian rules.3 One of the most talented, Civic Forum's Josef Vavroušek, acknowledged the appeal and rationality of that option: If Civic Forum in the Czech lands and Public against Violence in Slovakia behaved like classic political parties striving for the greatest share of state power, they would try to push through an electoral system that would allow them to use the spontaneous support of broad The Czech and Slovak Republics 69 swathes of the population based on rejection of the totalitarian system run in this country by the Communist Party. Both movements would most likely propose holding elections at the soonest possible date (for example, in February [1990]), on the basis of the majoritarian system used in past decades. The elections would then probably turn into a plebiscite, in which the movements Civic Forum and Public against Violence would receive a majority in most districts at the expense of the Communist Party. [...] But for Civic Forum and Public against Violence, there is more at issue than just electoral victory.4 As Vavrousek explained, several arguments steered the movements' leaders toward PR. First, men and women who until recently had been dissidents felt bound by their long-standing normative commitment to pluralism. A false unity had been imposed on Czechoslovak society for four decades and it was felt that the new federal legislature ought to be a more faithful reflection of its variety.5 Civic Forum had actually embedded this objective in its organizational plan in December 1989, defining itself as 'a movement of citizens for political pluralism and the holding of free elections'.6 As one Civic Forum leader, Petr Pithart, explained when an American political scientist advised him to go for a majoritarian system and trounce the Communists, 'We do not want to replace one dominant political force with another, so we shall not tailor the electoral system to benefit us and only us.'7 Second, it was recognized that parties were essential organizing elements of a democracy, and the system had to encourage their institutionalization before some day shifting the focus onto individual candidates linked to particular constituencies. Third, and more practically, Civic Forum and Public against Violence were constrained by the fact that a diverse multi-party system already existed, the combined inheritance of the sham pluralism of the National Front (the communist-controlled bloc that coordinated four satellite parties, the trade unions and other social organizations) and the genuine pluralism of dissident society. The consent of these other parties had to be secured in two special round-table talks held on 5 and 11 January 1990, in a cross-party experts' meeting on 19 January, and then in 18 legislative committees before a floor vote in the bicameral Federal Assembly on 27 February.8 Fourth, as they acquired greater appreciation of the state of the economy and the perils of its restructuring, Civic Forum and Public against Violence realized that they would do well to avoid sole responsibility for 70 Embodying Democracy the hardships lying ahead, so a broader coalition government would be preferable and PR would facilitate that.9 Finally, some strategists warned that the movements' national ratings in opinion polls might not actually convert into many victories in single-member districts, where locally respected independents or other parties could attract and aggregate support.10 The broad outlines of a PR system were quickly thrashed out by a working group and hotly debated in Civic Forum's supreme assembly on 6 January 1990. Although several of the movement's most celebrated figures argued for a single-member system that would allow independents greater opportunity to compete, they were gently but firmly overruled for the reasons enumerated above. The following day, Vavrousek distilled the assembly's discussion into three possibilities, which he directly linked to larger schemes for the restructuring of the federation: 1. PR in 12 multi-member districts, with a 4 or 5 per cent threshold, and an opportunity for voters to indicate preferences for particular candidates. 2. The West German additional-member system, with a 4 or 5 per cent threshold. 3. The use of option 1 for election of only one of the Federal Assembly's houses, the Diet of the People, in which Czechs and Slovaks were represented on a 2:1 ratio according to population size, while the Diet of Nations, divided evenly between the two republics, would be elected through single-member districts.11 Without consulting its own assembly or the public, the inner core of Civic Forum selected the first option. The second was rejected as too demanding of voters, while the third was quickly torpedoed by Slovakia's Public against Violence on the grounds that it would require a major constitutional settlement, which only freely elected legislatures would have the legitimacy to undertake.12 Subsequent talks with other parties swiftly led to general agreement on Vavrousek's first option, but the fine detail of the law remained to be written. Acute time pressure then became a powerful factor, as founding elections were scheduled for early June 1990. To expedite the bill's drafting, one of the former satellite parties, the Czechoslovak Socialist Party, dusted off the law used in the last largely free election, in 1946.13 That law had been in turn a hasty revival (again, under time pressure) of the electoral system introduced for the new state in 1920, which in itself was a conscious reaction against the ethnically biased majoritarian The Czech and Slovak Republics 71 system of the Habsburg monarchy.14 The 1946 law had been superseded after the Communist seizure of power by a 1954 act that replaced PR with single-member districts; voters had the choice only of approving or rejecting the sole candidate fielded by the National Front.15 The return to PR in 1990 was thus a replay of the break, seventy years previously with the perceived injustice of majoritarianism. The result of bargaining in early 1990 over a version of the 1920/1946 electoral law was a PR system using party lists in 12 multi-member districts (eight in the Czech Republic, four in Slovakia). • To start with, electoral commissions would ascertain which parties had received at least 5 per cent of the vote in either the Czech or Slovak republic.16 • The Hagenbach-Bischoff formula (Votes/Seats + 1) would be applied to the total votes of qualifying parties in each of the 12 districts, to generate the quota for allocating seats.17 • After this initial district scrutiny, it was expected 10-20 per cent of the seats would still be unfilled, as no party within a given district would have enough votes remaining to meet the quota. The unused votes would be aggregated at the level of each republic, a new republic quota would be generated and the outstanding seats filled accordingly. The main departure from the 1946 law was that Hagenbach-Bischoff (H-B) was used for the first and second quotas; in 1946, the first was generated by the Hare formula. It was hoped that the use of H-B would result in fewer seats to be filled on the second scrutiny, since party headquarters would have greater discretion in drawing up republic-level lists of candidates and centralized party power was an unpopular concept in 1990.18 (A separate law governing elections to the Slovak republic legislature followed the 1946 model, but this was changed for the 1992 election to bring it into line with federal and Czech practice.19) Many speakers in the parliamentary debate on the bill reported that they saw strict PR as a short-term, transitional measure that should quickly yield to at least a mixed system. The expectation was thus present in Czech and Slovak politics from the very beginning that at some point soon a majoritarian element would enter the equation; polls in the autumn of 1991 suggested that a majority of voters would have welcomed it.20 In fact, there was evidence of a long-standing public aversion to pure PR. During the liberalization of 1968, one survey found dissatisfaction with the practice of the single candidate, but not with 72 Embodying Democracy Table 4.1 Summary of Czechoslovak electoral laws Election year Electoral District Seat allocation Threshold system type structure formula 1990 PR 14 multi-member Hagenbach- 5% districts Bischoff 2 republic Hagenbach- districts for Bischoff unfilled seats 1992 PR 14 multi-member Hagenbach- 5% for single districts Bischoff parties; 7% for 2 republic Hagenbach- alliances of districts for Bischoff two or three unfilled seats parties; 10% for four or more single-member districts (SMDs). The most popular reform in 1968, preferred by three-fifths of respondents, would have been to retain SMDs while introducing true competition between candidates of the five parties of the National Front. Around one half, however, would also have welcomed multi-member districts with a mix of list and nominal voting. What united the most favoured options was the opportunity to choose individual candidates and not just a party. The least popular system was straight list-voting, with mandates allocated to candidates according to their rank on the ballot; more respondents preferred the unreformed status quo over a switch to that.21 The key features of the Czechoslovak electoral law are summarized in Table 4.1. However, sensing a general appetite for electoral reform, in late 1991 President Vaclav Havel submitted a bold initiative as part of a larger package of bills to rescue the federation from the constitutional crisis that beset it soon after the end of Communist rule. Havel had been one of the last leaders in Civic Forum to accept PR in 1990; as he explained a year later. I consider it one of the mistakes I have made in office that before the last elections I was not far more persistent in my struggle for a different electoral law. [...] To put it simply, it was a battle between a proportional and majoritarian electoral system. I clearly prefer the majoritarian. But I would be grateful also for a mixed system.22 The alternative proposed by Havel in 1991 would have introduced the supplementary vote, akin to that used for election of the Queensland The Czech and Slovak Republics 73 Legislative Assembly between 1892 and 1942, and more recently the president of Sri Lanka and mayor of London.23 It allows voters to indicate first and second preferences among candidates standing in single-member districts. Should no one win an outright majority after first preferences have been counted, the weakest candidates would be eliminated stepwise and their second preferences assigned until one contestant had accumulated more than 40 per cent of the vote and won the seat. Should no candidate reach even that sum, each party's votes would be pooled with those from any other district in the Czech or Slovak Republic that had similarly failed to return a representative, the d'Hondt divisor would be applied, and the vacant seats awarded to parties proportionally. The presidential bill had several attractive features. First, it introduced a stronger constituency link and accountability through the SMD and granted voters more power, since the 1990 law had contained only the narrowly circumscribed opportunity of approval voting.24 Second, it allowed voters to signal what sort of coalition they might welcome after the election, as second choices could be read as acceptable partners. Third, it increased the likelihood of winners enjoying the backing of a majority of their constituents without second-round run-offs. Finally it allowed local independents to compete while reassuring the larger parties of their overall pre-eminence, thus preserving the country from what the bill's backers called 'the Polish effect' (fissile governments held hostage by tyrannous micro-parties).25 The bill, however, died along with the rest of the constitutional rescue package in early 1992, primarily because the president failed to convince the parties that their place in the sun was not threatened by his plan. The pressure of time was again invoked to justify institutional conservatism, since the designation of SMDs and other technical demands might not be manageable with an election already fast approaching (the 1990 legislature, like that of 1946, was elected for only a two-year term). Admittedly, there was one sizeable group interested in electoral reform: Václav Klaus's Civic Democratic Party (Občanská demokratická strana, ODS), the largest successor to the now-defunct Civic Forum. ODS, however, liked the French system of second-round run-offs; once it sensed that it could not build a majority for its preference, it shifted its energy into campaigning at least for the adoption of the d'Hondt algorithm, known to favour bigger parties, in place of the Hagenbach-Bischoff quota.26 This effort likewise failed to find sympathy among the 15 diminutive parties in the federal legislature.27 74 Embodying Democracy The only substantial changes agreed in 1992 were the introduction of differentiated thresholds for electoral alliances (a coalition of two or three parties would need 7 per cent of the vote between them to qualify for seats, while a coalition of four or more would need 10 per cent), and a slight expansion in the opportunity for preference voting.28 The threshold change compounded the disproportionality already generated by the 5 per cent cut-off, as shown in Table 4.2, with deviation rates comparable to those in countries with single-member plurality systems.29 Turnovec argues that this increased disproportionality primarily punished parties that favoured the federation's continuation, and manufactured a premium for the two parties that would unravel the union in the coming months: Klaus's ODS and Vladimir Meciar's Movement for a Democratic Slovakia (Hnutie za demokratické Slovensko, HZDS).30 The electoral law may also have contributed to the federation's demise by applying thresholds and quotas only through the constituent republics and not at the federal level; parties could therefore attract votes by Table 4.2 Wasted votes and deviation from proportionality (D) in the Czechoslovak 1990 and 1992 elections Assembly and year Wasted votes (as % of all votes cast in Czech Republic) D (Czech Republic) Wasted votes (as % of all votes cast in Slovakia) D (Slovakia) Diet of the 16.8 People 1990 Diet of Nations 18.4 1990 Republic legislature 18.8 1990 Diet of the 25.9 People 1992 Diet of Nations 26.8 1992 Republic legislature 19.1 1992 11.2 15.1 7.2 13.1 12.7 6.6 11.6 7.6* 3.6* 13.0 26.3 12.5 13.8 20.5 9.6 8.6 23.8 11.2 * The threshold for seats in the 1990 Slovak legislature was 3 per cent. Source: Author's calculations from election results in Statistická ročenka České a Slovenské federativní republiky (Prague: SNTL, 1991), pp. 629-30 and Statistická ročenka České republiky (Prague: Český spisovatel, 1993), pp. 437^1. Disproportionality (D) is derived according to the least-squares index proposed in Michael Gallagher, 'Proportionality, Disproportionality and Electoral Systems', Electoral Studies, vol. 10, no. 1 (1991), p. 40: the square root of V2 [2 (vi~si)2] where v,- is each party's share of the vote, and s; each party's share of the seats. The Czech and Slovak Republics 75 catering to particular interests at the expense of the general, so no nationwide party system developed.31 Czechoslovakia failed to emulate the crucial leap made by West Germany in 1953, when it moved from the original threshold of 5 per cent in one Land to the more demanding and integrative 5 per cent nationwide (or three district seats).32 Consequently, the independent Czech and Slovak Republics came into being in January 1993 still under the PR regime hurriedly introduced without enthusiasm three years before. The unintended consequences of the pursuit of stable government Slovakia Slovakia was the first to alter its system. Prime Minister Mečiar raised the possibility after the 1994 elections, which took place early owing to the collapse of his second government, and his third government's programme in January 1995 contained a vague promise of amendments to the law. A serious campaign, however, was not launched until March 1996. At that time, Mečiar told his party's congress that the lessons of 1994 (the fall of his government and the difficulty of forming a new one after early elections) showed that the 5 per cent threshold was not enough to ensure a manageable number of parties: We would therefore accept if the development of political parties' integration went towards the creation of another strong political subject so that in competition of the two (of course upon existence of all the others) the political system could stabilize. We can see that it will be necessary to change the system, to abandon the present system of relative [sic] representation - also our friends in Italy left the system that we have taken over and have been practising until now. And to switch to the majority system or a combination of the majority and relative [proportional] systems.33 Mečiar went into more detail during his concluding remarks to a closed session of the congress: 150 single-member districts, whose winner-take-all nature, he explained, offered 'a certain political hope'.34 Only three months later, however, Mečiar was reminded both of the reason for seeking this change and of the obstacles to its realization when a bitter feud erupted between HZDS and its junior cabinet colleagues, the Slovak National Party (Slovenská národná strana, SNS) and Association of Workers of Slovakia (Združenie robotníkov Slovenska, ZRS), 76 Embodying Democracy over privatization spoils. Although the coalition survived, Mečiar learned that he could not expect his partners altruistically to support actions that would damage their own interests. As the two junior parties would be hurt by a law introducing severe disproportionality and HZDS held only 61 of the 150 legislative seats, coalition arithmetic forced Mečiar to retreat from majoritarianism.35 The bill's drafting was assigned to the Interior Ministry but after almost two years had passed, a group of parliamentarians including Meciar's legal counsellor, Ján Cuper, assumed the task. That long delay, which may have been deliberate to deprive the opposition of reaction time, meant that the bill was not submitted until March 1998, only six months before elections were due. Containing a number of unconstitutional clauses,36 the bill introduced three significant changes. First, the four existing electoral districts were to be collapsed into one, containing all 150 members. (The Association of Workers preferred eight multimember districts, but was overruled by its coalition partners.) This huge rise in district magnitude would normally be welcomed as a boost to proportionality, but opposition parties interpreted it as a device to rob them of the campaign advantages of regional lists (which could be headed by locally popular figures), forcing them into a nationwide showdown with the charismatic Mečiar. Secondly, preference voting, though formally untouched, was neutralized by the single district, as the application of the existing formula to such a large area made it practically impossible for any but the already most prominent figures to move up the candidate list.37 Preference voting's impact was also lessened by a change in the procedure for filling a vacated seat - a common event, since deputies must suspend their mandates if they become ministers. Previously, it went to the recipient of the largest number of preferential votes who had not earned enough to move up the list into electable range; now the party leadership would have a free hand in selecting the replacement, to ensure that someone who had since become outspoken or even quit the party did not obtain a mandate. Thirdly, the Hagenbach-Bischoff quota and 5 per cent threshold were retained but applied to each party, including the constituent members of an electoral alliance. (Electoral commissions would be able to determine the exact percentages garnered because allied parties would operate separate candidate lists.38) The amendment in effect rendered alliances redundant, as weaker parties could no longer piggyback on the stronger. (Apparentement, however, was not altogether abandoned: thanks to a motion tabled by the opposition, it was agreed that the The Czech and Slovak Republics 77 votes of parties competing together successfully would still be pooled for the allocation of seats.) The reform was clearly directed at opposition moves to cluster small parties around the Democratic Union (formed in 1994 by defectors from Mečiar's HZDS and the Slovak National Party) and the Christian Democratic Movement and thereby minimize vote-wastage. The day-long debate on 20 May 1998 consisted primarily of the governing coalition rejecting more than 200 proposed amendments from the opposition before enacting the bill largely as it had been originally drafted. The changes were clearly intended to disadvantage the diffuse opposition parties. Instead, they resulted in the merger of five of them into a shell party, the Slovak Democratic Coalition {Slovenská demokratická koalícia, SDK), with a single candidate list.39 The three parties representing the Hungarian minority likewise coalesced into a single electoral entity. This arrangement removed the issue of thresholds while minimizing vote-wasting. In the September 1998 election, on a turnout of 84 per cent, only 5.5 per cent of the vote went to parties that did not win seats, a marked improvement on 1994 (13 per cent). More than half of the vote went to just two parties, Mečiar's HZDS and the SDK. The least-squares disproportionality index, 2.9, halved the 1994 score of 5.9, and was lower even than the rate from 1990 (3.6) when only a 3 per cent threshold applied. This ultra-proportional outcome left Mečiar completely unable to assemble a new majority coalition, let alone govern by himself. Instead of a government of one or at most two parties, the reform resulted in a new ruling partnership of ten.40 Even before the election was held, the opposition had filed with the constitutional court a case against the revisions of the electoral law. The court's ruling in March 1999, six months after the elections, struck down five contested sections, including the stipulations about filling vacated seats and candidate lists that made coalitions redundant.41 As before, the initiative to prepare new legislation fell to a group of parliamentarians from the governing parties, this time to erase the defining features of the Mečiar reform. Going beyond the requirements of the court's finding, the amendment passed in August 1999 reinstated the threshold scale of 5 per cent for single parties, 7 per cent for alliances of two or three parties, and 10 per cent for alliances of four or more.42 The single electoral district, however, remained. This partial restoration of the status quo ante infuriated the opposition, now led by Mečiar's HZDS. Claiming that the country's new rulers were recondemning it to political instability, Ján Cuper, one of the authors of the 1998 law, even denounced multi-partism as unpatriotic 78 Embodying Democracy because of the alleged harm it inflicted on the young state. In reply, the 1999 law's authors freely confessed that it was merely a stop-gap prelude to a systematic rethinking of the electoral system. Displaying the same lack of passion for their product as the designers of the original 1990 PR regime, they promised that the government would submit a major bill in the year 2000 but offered no hints of its content.43 The following year elapsed without the promised legislation; the government's calendar for 2001 set a September target date but the bill was not a priority and remained off the crowded agenda. In the absence of executive leadership, the legislature itself produced two proposals for specific amendments: • A working group of parliamentarians convened in the first half of 2000 and recommended a set of core changes, including the restoration of the four multi-member districts (although following boundaries different from those used in 1990-4), and easier conditions for preference voting.44 Opposed by HZDS and backed by only half of the ruling coalition's legislators, the bill was voted down on its first reading in September 2001. • In January 2001, a private member's bill was submitted by Robert Fico, leader of a new movement, Smer (literally, 'direction' or 'way'). Fico's bill aimed to raise the thresholds to 7 per cent for a single party, 14 per cent for an alliance of two or three, and 21 per cent for a coalition of four or more. Clearly intended to simplify the party landscape and award premiums to the successful (his own party was polling well at the time), the bill suffered the fate of most solo initiatives and died on first reading.45 With the next general election fast approaching, Slovakia had not undergone the thorough revamping of the electoral system that its political class deemed necessary (see Table 4.3). The reform introduced by Meciar was largely undone, but with it went the powerful centripetal incentive of the 5 per cent threshold. Consequently, the SDK reverted to its constituent parts, which in turn were sundered by factional and personal rancour. Meanwhile, all of the ruling parties were damaged in public opinion by the painful austerity programme introduced to repair the fiscal disaster bequeathed by Meciar in 1998. Ironically, the beneficiary of the reversal of the 1998 reform may be precisely the main sponsor of that reform: Meciar's HZDS. If the 2002 election sees a rise in vote-wastage and disproportionality, it stands to gain, as the spell in opposition saw its support well above that for any governing party. The Czech and Slovak Republics 79 Table 4.3 Summary of Slovak electoral laws Election year Electoral District Seat allocation Threshold system type structure formula 1990 PR 4 multi-member Hare 3% districts 1 national district Hagenbach- for unfilled seats Bischoff 1992 PR 4 multi-member Hagenbach- 5% for single districts Bischoff parties; 1 national district Hagenbach- 7% for for unfilled seats Bischoff alliances of two or three parties; 10% for four or more 1994 PR 4 multi-member Hagenbach- 5% for single districts Bischoff parties; 1 national district Hagenbach- 7% for for unfilled seats Bischoff alliances of two or three parties; 10% for four or more 1998 PR 1 national Hagenbach- 5% for all multi-member Bischoff parties, district whether standing alone or in alliance 2002 PR 1 national Hagenbach- 5% for single multi-member Bischoff parties; district 7% for alliances of two or three parties; 10% for four or more The Czech Republic Meciar's challenge to the 1990 PR model appears modest when compared to that attempted in the Czech Republic. As in Slovakia, the catalyst was a government crisis, which erupted in November 1997 and ended Prime Minister Vaclav Klaus's five years at the helm. As in 80 Embodying Democracy Slovakia, the causes of the crisis were the premier's style of rule, discontent within and defections from the largest party (ODS), the departure of coalition partners, economic woes and privatization scandals. A scapegoat, however, was found in the electoral system, because the 1996 election had returned Klaus's coalition to power two seats short of a majority. Although ad hoc deals with independent deputies kept the coalition alive, it was a precarious existence and encouraged brinkmanship by the junior partners. Whereas in 1995 Klaus had seemed content with the status quo,46 in 1998 he targeted PR for replacement. This choice of culprit was largely unfair. The 1998 election results showed that the existing system could sideline difficult extremist parties (the far-right Republicans, whose vote fell below 5 per cent) and produce a possible coalition of ODS, the Christian Democrats and the Freedom Union (recently formed by ODS defectors). Such a coalition would have had both a minimum-winning majority (102 of 200 seats) and policy contiguity. That option, however, was wrecked by intractable arguments over the premiership and ratio of portfolio allocations. Unlike the 1996 election, that of 1998 also produced an alternative minimum-winning coalition, of the Social Democrats (the largest party in the legislature), the Christian Democrats and the Freedom Union, which would have held 113 of 200 seats. Even though the Social Democrats offered huge concessions, going so far as to offer the premiership to the Christian Democrats, Freedom Union refused to join on personal and policy grounds.47 The impasse, caused not by the electoral system but by various parties' bluffing and stalling,48 was broken in July 1998 when the Social Democrats and ODS came to a historic compromise: the former would be allowed to govern alone in a minority by the latter, under a set of conditions enshrined in a 'Pact on the creation of a stable political environment'. Article VII of this pact committed the two parties to writing within 12 months a bill that 'in accordance with the constitutional principles of the Czech Republic will increase the significance of the results of political parties' competition'.49 This clause was widely interpreted as portending an electoral reform to benefit the larger contestants. The constitutional proviso of Article VII was critical. The Czech constitution follows the example set in 1920 of marking the parameters of the electoral system. Furthermore, also as in 1920, it stipulates a bicameral legislature, with a Diet of Deputies (poslanecká snimovna, to which the government answers) and a Senate. According to Article 18, the Diet of Deputies is elected by proportional representation, and the Senate by a majoritarian method. The intention was to give the country the best The Czech and Slovak Republics 81 of both worlds without directly employing a mixed system.50 The specifics of elections to the Senate were not agreed until 1995, due to a protracted intra-cabinet clash between ODS (again favouring the two-ballot majority system) and the Christian Democrats (arguing for the Australian alternative-vote system). The ODS eventually bullied the rest of the coalition into accepting its preference, using the illiterate claim that the alternative vote would not be sufficiently majoritarian to satisfy the constitution.51 Discussion of electoral reform in 1998, therefore, had to respect the requirement of PR for the Diet of Deputies or - and it was arithmetically possible at the time - include provision for constitutional amendment. Neither the ODS nor Social Democrats was internally united on the best approach: Klaus's lieutenants were divided between proponents of a one-ballot plurality method and forms of PR with naturally high thresholds,52 while the Social Democrats debated whether any reform was wise when economic recession was eroding their own standing in opinion polls.53 As the governing party dithered, ODS united on a plan inspired by a political-science undergraduate dissertation. They suggested fewer deputies (162 instead of 200) to be elected in 35 multi-member districts with the Imperiáli divisor and a 5 per cent threshold without a second scrutiny.54 While still technically a PR system, its disproportionality effect would be enormous, as an average district magnitude of 4.6 would create an extremely high effective threshold. The Social Democrat prime minister, Miloš Zeman, signalled a sympathy for the ODS plan, but soon encountered resistance from his strongest in-house critic, Petra Buzková. Deputy chair of the party and of the Diet of Deputies, the very popular Buzková counter-proposed 14 multi-member districts based on the newly designated regions and the d'Hondt divisor.55 Out of this confrontation within the Social Democrat leadership came a compromise position: 200 deputies elected in up to 36 districts but with d'Hondt rather than Imperiáli.56 An exasperated ODS put the country through a three-month crisis in the autumn of 1999, holding up passage of the budget and threatening to terminate the pact that kept the government in power, while sending out feelers to the parties with which it had failed to coalesce the year before. The Christian Democrats and Freedom Union, however, were now interested only in the highly successful 'Quad-coalition' (4K) they had formed with two mini-parties for the 1998 Senate elections.57 In this charged environment, ODS and the Social Democrats sealed a new pact, dubbed in characteristically regal style by Klaus a 'patent of tolerance', in January 2000. It committed the signatories to 'finding an 82 Embodying Democracy electoral system that would significantly facilitate the formation of a functional majority government comprised of at most two political subjects'.58 The outlines were close to those accepted by the Social Democrats in the summer of 1999, with 200 deputies to be returned from 35 multi-member districts. To accommodate the ODS demand for greater disproportionality, the Social Democrats invented a modification of the d'Hondt sequence, whereby the first divisor would be the square root of 2 (rounded up to 1.42) rather than l.59 The potential impact of this innovation is shown by Table 4.4, which compares the actual allocation of seats from 1998 using the Hagenbach-Bischoff quota with those that would have occurred under modified d'Hondt as well as the two formulas originally preferred by ODS (Imperiáli) and the Social Democrats (unmodified d'Hondt). In the course of the tempestuous second reading of the government's bill in May 2000, a final major change was introduced by the ODS club: in a more subtle version of Mečiar's reform, the differentiated threshold would be replaced by the requirement that electoral alliances of two parties should win 10 per cent of the vote, alliances of three 15 per cent, and alliances of four or more 20 per cent.60 The rule skilfully avoided the accusation that it made coalitions redundant, since a two-party coalition could satisfy it by combining the support of 7 and 3 per cent of the electorate. In combination with an average district magnitude of 5.7, however, it would reinforce the new system's potential effective threshold of 14-17 per cent.61 Had the 1998 elections been held under these rules, the least-squares disproportionality score would have been an enormous 20.2 instead of the actual rate, 5.8 (up slightly from 5.6 in 1996). Such a figure, which admittedly cannot take into account the strategic responses that small parties and voters would adopt to the new Table 4.4 Allocations of seats according to formula, using the 1998 vote Formula Social ODS Communist Christian Freedom Democrats Party Democrats Union Hagenbach- 74 63 24 20 19 Bischoff d'Hondt 88 71 22 11 8 modified 101 88 4 6 1 d'Hondt Imperiáli 108 83 4 5 0 Source: Tomáš Lebeda, 'Přiblížení vybraných aspektů reformy volebního systému', PoHtologický časopis, vol. 7, no. 3 (2000), p. 246. The Czech and Slovak Republics 83 rules, is comparable to the average for France under its two-round majority system, and nowhere near those for 'reinforced' PR countries such as Spain (average D of 8.15) and Greece (D = 8.08).62 The bill passed the Diet on 26 May 2000 and then the Senate a month later only with the support of ODS and the Social Democrats; Prime Minister Zeman threatened rebellious Social Democrat senators with expulsion from the party, while ODS had to drag three from their hospital beds to ensure ratification.63 President Havel, once an ardent critic of PR but now opposed to a change so explicitly intended to skew the distribution of power, exercised his suspensive veto; the Diet overturned it on 10 July 2000 and the law was published in the official gazette.64 Along with senators from the Quad-coalition (4K), Havel asked the constitutional court to strike down the changes, even though the 4K, like the Slovak Democratic Coalition in 1998, stood to benefit enormously from them if its recent surge in opinion polls were translated into seats (see Table 4.5). This paradox can be explained by the reluctance of the 4K's components to submerge their distinct identities, resources and leadership profiles permanently in an artifice they accepted only as a temporary survival mechanism. The court had been involved in electoral system disputes thrice before: first regarding the 5 per cent threshold, then the introduction in 1995 of deposits refundable only to parties that won seats, and lastly the requirement that a party win at least 3 per cent of the vote to qualify for remuneration of campaign costs. In the first case, the court found that modest disproportionality was an acceptable externality of any PR system. In the second, it upheld the principle of deposits to deter frivolous campaigns. In the third, however, it objected to the reimbursement threshold as an excessive obstacle to political competition and Table 4.5 How poll ratings would have translated into seats under the new Czech electoral system had an election been held in June 2000 Social ODS Communist Quad-coalition Democrats Party Actual 1998 74 63 24 39 result (pre-reform) Hypothetical 27 56 46 71 2000 result (post-reform) Source: Miroslav Korecký, 'Průzkum: Na volební systém doplatí ČSSD i ODS', Lidové noviny, 19 June 2000, p. 1. 84 Embodying Democracy recommended its reduction to 1 per cent. The court's track record thus offered precedents for both sides: Havel and the senators could depict the new system as a limitation of political free play analogous to the rule on remuneration, while the Diet's team (led by its speaker, Klaus) could invoke the recognition of PR's unavoidable deviations and the value of 'integrative stimuli'.65 The court's ruling of 24 January 2001 found almost entirely in Havel's favour. • While accepting that low district magnitude and the modified d'Hondt divisor individually were unobjectionable instruments, in combination they would give rise to 'some sort of hybrid [electoral system]' that conformed to neither of the general types - PR and majoritarianism - recognized by the constitution.66 The supporting parties, moreover, had failed to supply a convincing explanation of their modification of the conventional d'Hondt sequence. • Departing from precedent, the court overturned the requirement of a deposit of 40,000 crowns (equivalent to $1066/€1139 at the time of the court's ruling) in each electoral district. Most of the justices now viewed it as biased against poorer parties, redundant in the presence of a challenging threshold scale and rare in European states with PR.67 The court repeated its earlier objection to the threshold for remuneration of campaign costs, now set at 2 per cent instead of the 1 per cent recommended previously by the court, especially as the amount of money awarded was to be drastically reduced and state financing directed more heavily just to parliamentary parties through an annual grant based on legislative seats. (The court struck down that change in a separate case in February 2001.) • The one victory for Klaus came when a majority of the justices, in accordance with their earlier ruling on thresholds and the variety of policies and jurisprudence in other European states, upheld the new threshold scale for coalitions. They viewed it as politically motivated, but not unconstitutional.68 The verdict deeply divided the political elite and the political-science community: those who privileged stable (i.e. one-party) government condemned the court for discounting precedents of disproportional PR such as Greece and Spain, while those who deplored the bastardizing of PR rejoiced.69 The Social Democrat government quickly fell back on the scheme originally championed by its deputy chair Buzkova (who had meanwhile resigned her party post in protest at the renewed pact with ODS): a straight d'Hondt divisor and 14 multi-member districts corresponding The Czech and Slovak Republics 85 to the new regional tier of administration. The government's bill, which received its first Diet reading in June 2001, reduced the remuneration threshold to 1.5 per cent but audaciously replaced the deposit struck down by the court with a mandatory, non-refundable 'contribution to the cost of holding elections' of 15,000 crowns (around $400/€442) per district. As the government's bill retained the challenging thresholds of the 2000 law, senators from the Quad-coalition introduced their own to restore the old scale of 5, 7, 9 and 11 per cent. Thanks to their triumph in the 2000 Senate elections, the Quad-coalition needed the support of only two independents to assemble a majority in the upper chamber for their rival bill in August 2001. In the ensuing intercameral game of 'chicken', as on so many previous occasions, the pressure of time proved decisive. A general election was due by June 2002, and the Social Democrats and ODS chose to call the Senate's bluff by pushing the government's bill through the Diet largely unchanged. They calculated that the Quad-coalition senators would yield, lest the latter's obstruction of the bill jeopardize the very holding of the election, an embarrassment that no country approaching the final stage of EU accession could afford to suffer.70 In January 2002 the Senate grudgingly passed the Diet's bill, and President Havel signed it into law (see Table 4.6). Table 4.6 Summary of Czech electoral laws Election year Electoral District Seat allocation Threshold system structure formula type 1990 PR 8 multi-member Hagenbach- 5% districts Bischoff 1 national district Hagenbach- for unfilled seats Bischoff 1992 PR 8 multi-member Hagenbach- 5% for a party; districts Bischoff 7% for alliances 1 national district Hagenbach- of two parties; for unfilled seats Bischoff 9% for three; 11% for four or more 1996 PR 8 multi-member Hagenbach- 5% for a party; districts Bischoff 7% for alliances 1 national district Hagenbach- of two parties; for unfilled seats Bischoff 9% for three; 11% for four or more 86 Embodying Democracy Table 4.6 (continued) Election year Electoral District Seat allocation Threshold system structure formula type 1998 PR 8 multi-member Hagenbach- 5% for a party; districts Bischoff 7% for alliances 1 national district Hagenbach- of two parties; for unfilled seats Bischoff 9% for three; 11% for four or more 2002 PR 14 multi-member d'Hondt 5% for a party; districts 10% for alliances of two; 15% for alliances of three; 20% for alliances of four or more Explaining electoral reform in the successor states The pretext for electoral reform, as mentioned above, was the crises of government duration and formation in both countries between 1994 and 1998. Before we accept this motive at face value, we should consider alternative explanations in light of the experiences of other countries in recent years. Elsewhere, electoral reform has been attributed to a range of causes: • Voters and elites enjoy greater freedom with the end of the Cold War to criticize the shortcomings of liberal democracy, so imperfect institutions untouchable in a more adversarial age can now be questioned. • Countries have become more open to learning from each other's experiences, with possible gravitation toward mixed systems. • Electorates no longer put up with the stranglehold on office of corrupt or unrepresentative parties and, in PR systems, wish to make individual elected officials more accountable.71 All of these explanations view electoral reform as at least partly the product of public pressure, not just of elite interests and bargaining; in Italy and New Zealand, for example, referendums played an important part in kicking off the process. This line of explanation does not apply so well to Czechoslovakia and its successor states. A shift to a more The Czech and Slovak Republics 87 majoritarian system would indeed have been popular in 1991, but by the second half of the decade, especially in Slovakia, polls showed respondents preferring to stay within the bounds of PR. A 1997 survey a year before electoral reform took place, found that 49 per cent of Slovaks wanted to stick with PR while 24 per cent favoured a shift to majoritarianism; the remaining quarter was undecided.72 Once the outlines of the new electoral law became known in spring 1998, again only 24 per cent of respondents endorsed it while 500,000 people signed a petition against it.73 Similarly, in the Czech Republic, polls in March and July 1998 (before and after that year's elections) found only 17 per cent favouring a majoritarian system, with commitment to proportional representation rising between the two surveys from 26.6 to 36 per cent. One-quarter expressed interest in a mixed system.74 Single-member districts may remain an abstractly attractive concept for alienated citizens, but any shift to them could exaggerate and freeze the existing balance of power between parties; few voters feel moved to award bonuses on such a scale to the current incumbents. Office-seeking provides the best explanation of moves to reform the electoral system in the Czech Republic and Slovakia. In each case, a major party had a selfish interest in maximizing its share of legislative seats in order to escape the constraints and exposure of multi-party governance. A key moment in both cases was the downfall of the preeminent political figure in each country - Vladimir Meciar in 1994 and Vaclav Klaus in 1997 - owing to defections from his own party and coalition. (In both cases, their removal from the premier's office eventually led to early elections, which is unusual in post-communist countries.) These crises, however, did not change preferences so much as provide an opportunity to parties that ever since their foundation in 1991 had favoured majoritarian or mixed systems, and had always resented the PR regime that forced them to share power, resources and kickbacks with smaller, ever-obstreperous parties. Having survived the ordeal of their downfall, both leaders rallied, determined to minimize the number of parliamentary parties so that such power-sharing could be discarded. It should also be noted that the underlying assumption of an inverse relationship between the number of parliamentary parties and cabinet survival, an axiom voiced not only by interested politicians but also (in the Czech case) by sympathetic political scientists,75 is largely unsubstantiated. In fact, a bivariate correlation of 29 cases from all ten post-communist states applying for EU membership shows no relationship between the effective number of parties sitting in each newly elected 88 Embodying Democracy legislature and the duration of the first post-election cabinet.76 As in other new democracies, and some older ones, the causes of post-communist coalitions' early ends lie not in an excess of parties, but in the constraints on partner selection imposed by the presence of 'uncoalitionable' extremists, sudden events such as corruption scandals and economic crises, ego storms and the greater propensity of new parties to splinter rather than accommodate dissent.77 Unwilling to accept those awkward facts, the strongest parties in both countries scapegoated overpopulated parliaments and sought a solution in electoral reform. Conclusions After a decade of competitive politics, elections in the Czech and Slovak republics remained governed by rules very similar to those agreed, quickly and with little commitment, at the start of the transition. Attempts at reform were stopped, overturned or diluted, but not before they produced a range of unintended consequences. Not least of these was harm, in the short run, to ODS and HZDS, the very parties that sought what their antecedents Civic Forum and Public against Violence forswore in 1990: the design of institutions to award seats and power well out of proportion to the victors' popularity. In both cases the mechanical effects of such manipulation were anticipated and neutralized by other parties' adroit mergers and voters' calculations (as shown in the 1998 Slovak elections, the 1998 and 2000 Czech Senate elections and the 2000 Czech regional elections78). In both cases the judicial branch intervened, decisively in the Czech Republic. The experience of the two republics suggests two conclusions regarding electoral reform and the resilience of institutions. First, the Czech and Slovak party systems were already stabilizing at the time electoral reform was attempted, with the effective number of parliamentary parties falling to 3.7 and 4.4, respectively, comparable to the means for older democracies such as France and Belgium. Voter volatility and party replacement rates had also been declining, in line with the trend throughout Central Europe.79 These cases thus confirmed that deliberate shifts toward greater disproportionality tend to be initiated in PR systems not in order to counteract fragmentation, but after a party system has begun to shrink and the fittest seek to push electoral systems in win-concentrating directions.80 Second, the creation of new electoral systems and their modification were the product of intense endogenous bargaining and horse-trading, and the parties involved proved capable of both self-interested and The Czech and Slovak Republics 89 sociotropic motives. The Czech and Slovak cases thus tally with other findings in this volume, as well as studies of post-fascist Germany and post-Franco Spain that stress the political nature of major institutional choices, with some features selected to serve the public good and others to benefit only certain competitors.81 The resilience of electoral institutions introduced hurriedly in 1990 was not an equilibrium outcome but an initially contingent process that generated uneven and unpredictable benefits which no one had the power to redirect even after new preferences were derived.82 The perceived stakes of electoral institutions remained high as long as parties large and small, left and right, feared that victorious opponents would aggrandize their power and pervert public policy.