What Kind of Democracy Is Emerging in Eastern Europe? Andrew Roberts1

Abstract: An expert on political institutions analyzes the patterns of democracy emerging in Eastern Europe by extending Lijphart’s study of democratic patterns to 10 new democracies. Analyzing established democracies, Lijphart found that 10 variables capturing the type and functioning of institutions cluster on two clearly separate dimensions. Distribution of these variables is used to characterize the new democracies of Eastern Europe. The causes of institutional choice and the reasons institutions produce divergent effects are examined in the context of these Eastern European patterns. The results also challenge the use of certain measures of institutions without consideration of the party system.

A

s more nations become democratic, increasing attention has been paid to the ways in which their democracies function. Indeed, it appears that just as there are varieties of capitalism (Hall and Soskice, 2001), there are also varieties of democracy. Democracies differ both in terms of formal institutional features such as electoral laws and central banks as well as in less formal features like party systems and cabinet types. Much of the literature on these varieties, however, has focused on the established democracies of Western Europe, North America, and the Antipodes with only occasional ventures to newer democracies. In this article, I consider what happens when we apply this literature to the new democracies of Eastern Europe. In particular, I ask how well Lijphart’s (1984, 1999) widely used distinction between majoritarian and consensus democracies explains institutional patterns and functioning in 10 newly democratic countries in Eastern Europe.

1 Assistant Professor, Department of Political Science, Northwestern University, Evanston, IL 60208; [email protected]

37 Post-Soviet Affairs, 2006, 22, 1, pp. 37–64. Copyright © 2006 by V. H. Winston & Son, Inc. All rights reserved.

38

ANDREW ROBERTS

I find that Eastern Europe presents several anomalies to Lijphart’s typology. Lijphart finds consistent “patterns of democracy” with countries exhibiting specific constellations of institutions, which have predictable consequences for their style of politics. His patterns, however, fit Eastern European realities only partially. Not only do these countries appear to be hybrids—mixing Lijphart’s consensus and majoritarian categories—but it is uncertain whether they produce the sort of politics that Lijphart hypothesizes. These anomalies point us toward a greater focus on the context in which institutions operate. In the first place, they lead us to ask where institutions come from and why they work the way they do. The hybrids of Eastern Europe appear to have much to do with the historically specific circumstances under which they democratized. Second, these new cases raise questions about whether conventional measures of democratic patterns capture the distinctiveness of Eastern European democracies. In particular, unconsolidated party systems in the region may undermine some of the advantages of a consensus design. The article is organized as follows. In the first half, I ask how well Lijphart’s typology fits Eastern European democracies. In the second half, I consider the implications of Eastern European patterns for our evaluation of both Lijphart’s typology and institutional theories more generally.

DEMOCRATIC INSTITUTIONS IN EASTERN EUROPE In this part of the article, I consider how well 10 Eastern European countries fit in Lijphart’s widely cited typology of majoritarian and consensual democracies.2 These types represent two polar designs in allowing the people to rule. As the name suggests, majoritarian democracies allow bare majorities to rule. In their pure form, they concentrate all power in a single-party majority government and are characterized by exclusiveness and winner-take-all competition. By contrast, consensus democracies try to maximize the number of actors involved in decision-making. They divide power among many institutions, tend to be inclusive, and emphasize bargaining and compromise among multiple parties. Lijphart measures the degree to which 36 established democracies fit into the majoritarian or consensus camp by looking at 10 variables. These variables measure both institutional design–for example, degree of central bank independence—and the effects of institutions—for example, the number of parties.3 His main conclusion is that these 10 variables cluster along two separate dimensions, which he calls the executive-parties dimension and the 2 As a measure of the influence of these works, the Social Sciences Citation Index reports 191 journal articles citing them. 3 Taagepera (2003) distinguishes between institutions that are subject to direct constitutional design and those that emerge indirectly from the interaction between institutions and other factors.

DEMOCRACY IN EASTERN EUROPE

39

federal-unitary dimension. The first dimension refers to whether power is dispersed within the central government (e.g., through coalition governments), while the second refers to whether power is dispersed among separate political institutions (e.g., second chambers, independent central banks). Lijphart finds that variables are strongly related to each other on each dimension and only weakly related across the two dimensions. To take two examples, a country with few parties is also likely to have minimum-winning-coalition cabinets, disproportional electoral results, and pluralist interest group relations, while one with a federal system should also have strong bicameralism, judicial review, and a rigid constitution. In what follows, I determine whether these patterns hold in Eastern Europe. I measure all 10 of Lijphart’s variables for 10 new countries for the period from the first competitive election after the transition up to the most recent election.4 The countries are Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, and Slovenia. These 10 countries were chosen because they have clearly met democratic norms for almost the entire transition period and thus provide the best comparators to Lijphart’s established democracies.5 The main finding in this part of the article is that Lijphart’s patterns do not fit well in Eastern Europe.

Executive-Parties Dimension Lijphart’s executive-parties dimension consists of five variables that measure the degree of consensus or majoritarianism in the main policymaking branch of government. A consensus government is characterized by a multiparty system, the predominance of minority or surplus majority cabinets, a rough balance between the executive and legislature, a low level of disproportionality in electoral results, and strong corporatism. Majoritarianism, by contrast, is exemplified by a two-party system, minimumwinning-coalition or single-party cabinets, executive dominance over the legislature, manufactured majorities in election results, and interest group pluralism. 4 One caveat to this study is that the 10 countries considered here have been democracies for only a short time. While the newest democracies in Lijphart’s dataset had at least 20 years of democracy under their belts at the time of his study, the oldest democracies in this sample have been democracies for only 15 years. In this sense, this study is more a preliminary evaluation than the final word. Given the stickiness of institutions, however, it is likely the findings here will have some durability. 5 While Freedom House rated all the democracies in Lijphart’s sample as free—with the exception of four that slipped to “partly free” in the 1990s—my sample includes several less democratic countries. Specifically, four countries had short spells of partial freedom of the following length: Estonia (1 year), Latvia (2 years), Slovakia (3 years), and Romania (6 years). Yet, over time, all of these countries have become fully democratic. Indeed, eight of the countries have entered the EU and the remaining two are expected to enter in the near future.

40

ANDREW ROBERTS

Table 1. Executive-Parties Dimensiona,b Parties

Cabinets

Executive Disproportionality Corporatism

Score Rank Score Rank Score Rank

Score

Rank

Score Rank

Bulgaria

2.98

19

70.3

17

1.97

31

7.22

23

3.0

12–15

Czech Rep.

4.08

32

39.3

32

2.78

20

6.43

25

2.2

29

26

1.57

36

6.05

26

2.8

17–18

45–46 3.50

16

10.34

15

1.9

32

Estonia

5.06

39

45.2

Hungary

3.33

24

0

Latvia

5.79

44

13.9

41

0.94

46

5.57

29

2.6

22–23

Lithuania

4.24

34

45.4

25

2.25

26

8.27

20

2.6

22–23

Poland

5.11

41

43.7

28

1.48

39

9.09

19

2.7

19–20

Romania

3.57

28

29.4

34

1.46

40

5.67

28

3.2

9

Slovakia

4.62

36

25.8

35

1.62

34

6.63

24

2.7

19–20

Slovenia

4.99

38

24.0

36

1.58

35

4.10

31

1.6

34–35

East European mean

4.38

33.7

1.92

6.94

2.5

Lijphart mean

3.20

60.2

3.32

8.02

2.2

a

Sources: See footnotes 6, 8, 10, 12, and 13 and Lijphart (1999). The ordinal rankings range from 1 (most majoritarian) to 46 (most consensus).

b

How do Eastern European countries rank on these variables? Table 1 presents scores equivalent to Lijphart’s for all 10 Eastern European countries as well as means for the Eastern European sample and for Lijphart’s 36 democracies. For help in comparison, Table 1 also includes ordinal rankings of the Eastern European countries in the entire set of 46 countries (Lijphart’s 36 plus the 10 considered here) where a ranking of 1 equals most majoritarian and 46 equals most consensus. On the first variable, party system type, Eastern Europe tends to be highly consensus. As measured by Laakso and Taagepera’s effective number of parties (ENP) index, these countries have greater numbers of parties than most established democracies (Laakso and Taagepera, 1979).6 The 6

The ENP is calculated by squaring the seat percentages of each party, summing them and taking the reciprocal of this sum. It thus weights large parties more heavily than small parties. ENPs for most countries were calculated with election data from the Project on Political Transformation and the Electoral Process in Post-Communist Europe at the University of Essex. These data can be found at www.essex.ac.uk/elections. For Slovenia, which is not included in this project, data come from Rose and Munro (2003). I do not consider free elections held in these countries while Czechoslovakia, the Soviet Union, and Yugoslavia still existed.

DEMOCRACY IN EASTERN EUROPE

41

average country in Eastern Europe has considerably more than four parties versus a little over three in Lijphart’s sample. When compared with Lijphart’s 36 democracies, all but one of the Eastern European countries fall in the upper half of the index and several, such as Estonia, Latvia, Poland, and Slovenia, are at the very top. Though there is a declining trend over time among these countries, even the regional mean from the latest elections puts Eastern Europe in the top quarter of advanced democracies.7 On measures of cabinet type, Lijphart’s second variable, Eastern Europe is likewise at the highly consensus end of the scale. Lijphart measures cabinet type by averaging the percentage of time a country is governed by a minimum winning coalition (MWC) and by a single-party government. Both of these indicate majoritarian rule because a limited number of political forces exercise power. By contrast, minority and surplus majority governments are evidence of more widely shared power. In Lijphart’s sample, 20 countries (55 percent) were governed by a MWC or single-party government more than half of the time and 26 (69 percent) more than 40 percent of the time. In Eastern Europe, by contrast, only one of 10 has these forms of government more than half the time and only four of 10 more than 40 percent of the time.8 At the extreme, Hungary was never governed by either a MWC or single-party government over the first 12 years of transition.9 In general, most cabinets in Eastern Europe are of the minority or surplus majority variety, and thus, according to Lijphart, consensus types. To measure the third variable, degree of executive dominance, Lijphart looks at the durability of cabinets. Long-lasting cabinets indicate executive dominance—a majoritarian trait—because the legislature is unable to unseat the executive; short-lived cabinets, by contrast, indicate a rough balance of power. Again, Eastern European countries are quite consensus.10 7

On the other hand, there is a good chance that we are underestimating the number of effective parties. Lijphart notes that it is not always clear what counts as a party. Some parties are closely allied and might be better treated as a single party, while other parties are heavily factionalized and could reasonably be considered two or more parties. It is the latter problem that is more severe in Eastern Europe. Parties have had little time to develop strong organizations and membership bases. Further, due to thresholds necessary to enter parliament, parties have frequently joined together to contest elections under a single banner and split apart after entering parliament. In dealing with factionalized parties, Lijphart determines the ENP by averaging two values: the ENP under the assumption that the party is a single party and the ENP under the assumption that the party is actually two equally-sized parties. In this article I have not adopted this coding for two reasons. First, factionalization and the declining trend in the ENP may simply cancel each other out. Second, I have not yet gathered definitive evidence on which parties are factionalized. Thus, in Table 1, I compare my data to Lijphart’s data without his calculation for allied or factionalized parties. The same is true for the calculations of cabinet types and disproportionality. 8 Scores were calculated with data from Müller-Rommel et al. (2004), Keesing’s Contemporary Archives (1990–2005), Blondel and Müller-Rommel (2001), Katz and Van Biezen (2004, 2005). 9 Since 2002, Hungary has been governed by a minimum winning coalition. 10 Durations were calculated with data from Müller-Rommel et al. (2004), Keesing’s Contemporary Archives (1990–2005), Blondel and Müller-Rommel (2001), Katz and Van Biezen (2004, 2005).

42

ANDREW ROBERTS

With the exception of the Czech Republic, Hungary, and Lithuania, executives in the region tend to serve less than half of their four-year terms before they are unseated. This compares to only 11 countries in Lijphart’s sample of 36 with averages below two. In Eastern Europe, few governments serve out their entire term and almost none are re-elected intact. In Lijphart’s view, this means that there is a rough balance between legislative and executive power, a sign of consensus government. When we turn to the fourth variable of electoral systems, however, Eastern Europe begins to deviate from this pattern. Lijphart measures the relative majoritarian-consensus bias of electoral systems with Gallagher’s index of disproportionality. This index ranges from 0—where vote and seat shares are equal—to 100—where a party receiving no votes holds all seats in parliament.11 Though most Eastern European countries have proportional electoral systems, which should lead to low levels of disproportionality, in fact most have quite high levels.12 While countries with PR systems in Lijphart’s sample usually have Gallagher indices under five, all but one country in Eastern Europe have an average of above five. Most have levels more characteristic of majoritarian countries. Electoral systems in Eastern Europe, despite their consensus design, seem to act in a majoritarian manner, with large parties receiving a considerable bonus in seat allocation. Eastern Europe also diverges from the consensus pattern in Lijphart’s fifth variable, interest group relations. Here Lijphart uses a measure of corporatism developed by Alan Siaroff (1999) that summarizes such characteristics of corporatism as strike levels, the nature and goals of trade unions, and means of dispute resolution. Though it is difficult to produce equivalent measures for Eastern Europe, I have attempted to place these countries approximately on Siaroff’s scale, where 0.44 is the consensus extreme of Sweden and 3.56 is the majoritarian extreme of Canada.13 By my measures, almost all of the Eastern European countries are in the pluralist or majoritarian half of Lijphart’s continuum.14 Though most did establish tripartite institutions after the transition, they are a far cry from their Western European counterparts. In an article on what he calls “illusory corporatism,” Ost (2000, p. 515) shows that there is a “consistent pattern of belittling of unions, non-binding agreements, restrictions to the 11

The index is calculated by first taking the difference between vote share and seat share for all parties and then squaring this number. These values are summed and divided by two. The index is the square root of this number. 12 Disproportionality was calculated with data from the Project on Political Transformation and the Electoral Process in Post-Communist Europe at the University of Essex, which can be found at www.essex.ac.uk/elections, and Rose and Munro (2003). In making the calculations for the two mixed electoral systems—Hungary and Lithuania—I used the vote share from the list portion and the total number of seats. 13 In developing these ratings, I relied on the following comparative sources as well as a number of individual country studies: Casale (1999); Crowley (2003); Crowley and Ost (2001); Rychly and Pritzer (2003); and Thirkell et al. (1995). The ratings of the less studied Eastern European countries—Bulgaria, Romania, and Slovenia—are subject to greater error than the others.

DEMOCRACY IN EASTERN EUROPE

43

state sector and general tripartite impotence.” To add to this is the general weakness of labor unions and business organizations, the primary bargaining partners in corporatist systems. In short, Eastern European countries tend to have quite high levels of consensus behavior in terms of party systems, cabinet types, and executivelegislative relations. On the variables of disproportionality and interest group relations, however, they exhibit more majoritarian tendencies.

Federal-Unitary Dimension While the five institutions discussed above refer to the dispersion or concentration of power within the central government, Lijphart’s second dimension—he calls it the federal-unitary dimension—focuses on the dispersion of power to separate institutions. Here, a consensus state would be federal and bicameral and feature a rigid constitution, strong judicial review, and an independent central bank. In short, institutions outside the central government exercise influence on policy. By contrast, a majoritarian state is unitary and unicameral and has no constitution, no judicial review, and a subordinate central bank. The central government rules largely unimpeded by other institutions. Scores and rankings of Eastern European countries on all of these variables as well as means can be found in Table 2. The first variable on Lijphart’s federal-unitary dimension is federalism itself. Lijphart ranks states from one to four, where four represents the consensus extreme of a federal and decentralized system and one is a unitary and centralized system. All of the Eastern European countries, however, fall on the majoritarian end of the scale.15 All are unitary systems and most are quite centralized. Though they have retreated from communist centralization by instituting local self-government, “in only a few cases has real decision-making authority been devolved” (Wetzel, 2001, p. 1). Most significantly, local governments have very little control of their own revenue and are highly dependent on non-transparent transfers from the center. This majoritarian tendency also shows up in the cameral structure of legislatures, the second variable on this dimension. Lijphart characterizes 14

Several commentators have challenged this conclusion. They argue that corporatist institutions have played an important role in ensuring social peace and economic reforms during the transition. I would note that they do not claim that they approximate corporatism in countries like Germany and Austria. See, for example, Heinisch (1999), Iankova (1998), and Orenstein and Hale (2001). As evidence in favor of weak corporatism, consider the chapter titles from a recent edited volume (Crowley and Ost, 2001) on workers in post-communist Europe: “The Failure of Social-Democratic Unionism in Hungary”; “Neocorporatism in Slovakia: Formalizing Labor Weakness in a (Re)democratizing State”; “The Weakness of Symbolic Strength: Labor and Union Identity in Poland”; “Winning the Battles, Losing the War: Contradictions of Romanian Labor in the Postcommunist Transformation”; and “Bulgarian Trade Unions in Transition: The Taming of the Hedgehog.” 15 Rankings are derived from constitutional provisions and the following sources: Bird, Ebel, and Wallich (1995); Ebel and Yilmaz (2002); Gibson and Hanson (1996); Horvath (2001); Kimball (1999); Kirchner (1999); Rose and Traut (2001); and Wetzel (2001).

44

ANDREW ROBERTS

Table 2. Federal-Unitary Dimensiona,b Federalism Bicameralism Constitution

Judiciary

Bank

Score Rank Score Rank Score Rank Score Rank Score Rank Bulgaria

1.0

1–25

1.0

1–17

4.0

40–46

3.0

28–40 0.55

37

Czech Rep.

1.0

1–25

2.0

21–29

2.5

17

3.0

28–40 0.73

45

Estonia

1.0

1–25

1.0

1–17

2.0

8–16

3.0

28–40 0.78

46

Hungary

1.0

1–25

1.0

1–17

3.0

18–36

4.0

42–46 0.67

42

Latvia

1.0

1–25

1.0

1–17

3.0

18–36

3.0

28–40 0.49

33

Lithuania

1.0

1–25

1.0

1–17

3.0

18–36

3.0

28–40 0.53

36

Poland

1.0

1–25

2.0

21–29

3.5

37–39

4.0

42–46 0.68

43

Romania

1.0

1–25

3.0

32–41

4.0

40–46

2.0

8–25

15

0.34

Slovakia

1.0

1–25

1.0

1–17

2.0

8–16

3.0

28–40 0.62

39

Slovenia

1.0

1–25

2.0

21–29

3.5

37–39

3.0

28–40 0.63

41

East European mean

1.0

1.5

3.1

3.1

0.60

Lijphart mean

2.9

2.2

2.6

2.2

0.38

a

Sources: See footnotes 15, 16, 17 as well as Cukierman et al. (2002) and Lijphart (1999). The ordinal rankings range from 1 (most majoritarian) to 46 (most consensus).

b

cameral structure on a one-to-four scale from (1) unicameral through (2) weak and (3) medium-strength bicameral to (4) strong bicameral. Only four of 10 Eastern European countries have a second chamber at all, and in three of them the second chamber is quite weak.16 In the Czech Republic, Poland, and Slovenia, the second chamber has no veto on legislation and can only delay the work of the lower house. In all three cases, an absolute majority of the lower house can override the upper house’s veto. Slovenia’s upper chamber even lacks the legitimacy of a direct election. Only in Romania does the second chamber have powers equal to those of the first chamber. All of the other Eastern European states (six of 10) have unicameral legislatures and fall at the extreme majoritarian end of the scale. This contrasts with Lijphart’s sample, where only 11 of 36 countries (31 percent) are unicameral. When we move to constitutions, judicial review, and central banks, Eastern European countries start to manifest more consensus traits. On 16

Data were gathered from each country’s constitution (Blaustein and Flanz, 1971).

DEMOCRACY IN EASTERN EUROPE

45

constitutional rigidity, the third variable, Lijphart divides countries into four categories depending on the size of the majority necessary to pass constitutional amendments: (1) simple majorities, (2) between one-half and two-thirds, (3) two-thirds, and (4) greater than two-thirds. Most Eastern European countries have opted for moderately difficult-to-amend constitutions. No country allows ordinary majorities to amend the constitution, and only three countries—the Czech Republic, Estonia, and Slovakia— have three-fifths majorities. Most of the remainder require a two-thirds majority. Several add extra provisions to this two-thirds: in Slovenia a small minority of MPs can request a referendum; in the Czech Republic, Poland, and Romania the approval of the Senate is required as well; Romania and Latvia also require referendums. Bulgaria is unique in requiring a threefourths majority. Several countries also require more than one vote separated in time. The modal outcome then is moderately difficult-to-amend constitutions. Similarly, all of the countries in the region have judicial review—a consensus trait —as a check on legislative majorities. On the basis of impressionistic judgments, Lijphart ranks countries on a four-point scale of (1) non-existent, (2) weak, (3) medium-strength, and (4) strong judicial review. For my own measures, I rely on several first-hand accounts of these courts.17 In Eastern Europe, typically, the scope for review is wide. With the exception of Estonia, all of these countries adopted the European rather than the American form of review. Their constitutional courts are set up specifically to resolve constitutional questions, which can be brought to them by many sorts of plaintiffs, including legislators and public interest groups, even without concrete injured parties. The actual performance of these courts has been diverse. As Schwartz (1998, p. 100) writes, “The record of these postcommunist courts has by no means been unblemished, and some of them have become mere tools of the regime. Most, however, have done remarkably well.” Hungary and 17

I find this evidence in Schwartz (1998, 2000), the American Bar Association’s assessment of the judicial system in these countries (www.ceelijri.org), and a survey of the Constitutional Courts in the region conducted by the Venice Commission (www.venice.coe.int), as well as single country monographs. Smithey and Ishiyama (2000) have attempted to measure judicial independence in these countries through an analysis of formal provisions of review. Their ranking, however, does not square with those of closer observers of the region’s courts. There appears to be a wide gap between formal provisions and actual independence. Smithey and Ishiyama find, for example, that Romania has the most powerful court in the region, while Poland’s is the weakest. Here, by contrast, is the comment of a first-hand observer of many of the region’s courts: “in Romania, where the Iliescu regime was semi-authoritarian from the start, the court was subservient to the president from the start” (Schwartz, 2000, p. 226). Poland’s court meanwhile has performed “surprisingly well,” overruling 40 of 52 statutes presented to it between 1990 and 1994 (Schwartz, 2000, p. 56). Another counterexample is Hungary, which Smithey and Ishiyama place toward the middle of their rankings. According to Schwartz, the Hungarian court “quickly became one of the most powerful tribunals of its kind anywhere in the world” (Schwartz, 2000, p. 9). Since Lijphart explicitly states that it is actual practice rather than formal independence that determines his ranking, I discount their results in my ranking here.

46

ANDREW ROBERTS

Poland have some of the strongest courts in the world, while Bulgaria, the Czech Republic, Estonia, Latvia, Lithuania, Slovakia, and Slovenia have courts that are not hesitant to overrule parliamentary legislation. By contrast, the court in Romania appears to be more subservient to other branches of government. In short, judicial review in these countries tends to place limits on governing majorities, in line with the consensus philosophy. Finally, on the fifth variable of this dimension, central bank independence, these countries are on the extreme consensus end of the scale. Lijphart measures bank independence with two indices that summarize the provisions of central bank charters in such areas as authority over monetary policy, limitations on lending to government, and procedures for the appointment and dismissal of bank governors.18 His index ranks countries from a minimum of independence at 0 to a maximum of 1. Cukierman et al. (2002) have prepared an identical index for the countries of Eastern Europe. They find that these countries have almost unanimously opted for independent central banks. In Estonia and the Czech Republic, bank independence is even greater than that of the highly independent Bundesbank. Lithuania and Poland meanwhile revised their central bank laws to put their levels of independence above the Bundesbank as well.19 Altogether, nine of the 10 countries are among the top one-third with the most independent banks in the democratic world. Only Romania falls in the majoritarian half of the scale. To summarize this dimension, these countries are unequivocally majoritarian in their centralization and unequivocally consensus in the independence of central banks. While cameral structure tends toward majoritarianism, both constitutions and judicial review tend toward consensus.

Patterns of Democracy in Eastern Europe One of the main conclusions of Lijphart’s work on advanced democracies is that these variables cluster on two separate dimensions. The executives-parties dimension includes the first five variables considered here, while the federal-unitary dimension encompasses the last five. Using correlational analysis, Lijphart finds that all the correlations of variables within the dimensions are statistically significant, while none of those between dimensions are significant. Factor analysis confirms that the variables can be divided into two encompassing and exclusive factors. In short, countries choose between consensus and majoritarian systems on two dimensions of democracy. To test these conclusions I repeat Lijphart’s correlational and factor analysis for the entire set of 46 countries (36 from his data set and 10 from 18 A third index, used only for developing countries, measures the rate of turnover of bank governors. 19 Their scores here are averages of the levels before and after the changes.

DEMOCRACY IN EASTERN EUROPE

47

Table 3. Correlations between the Variables (Results for Total Sample of 10 Eastern European and 36 Established Democracies)a,b (1)

(2)

(3)

(4)

(5)

(6)

(7)

(1)

1.00

(2)

–.81c

1.00

(3)

–.63c

.62c

(4)

–.56

c

.57

c

(5)

–.47

c

.58

c

(6)

.07

–.03

–.03

–.11

–.31d

(7)

–.13

.14

.15

.12

.04

.64c c

(8)

(9)

1.00 .31d –.22

1.00 .55c

1.00 1.00 1.00

(8)

.08

–.09

–.10

–.04

–.07

.38

.32d

(9)

.03

–.06

–.11

.15

.17

.23

.20

.37d

.00

d

(10)

.26

e

–.25

e

(10)

–.19

–.14

–.03

.13

1.00 .30

1.00 .53c

1.00

a

Calculated from data in Tables 1 and 2 and Lijphart (1999). Numbers in parentheses in the heading row and the first column refer to Lijphart’s 10 variables in the order in which they are presented in the text and in Tables 1 and 2. (1) refers to party systems; (2) is cabinet types; (3) is executive-legislative relations; (4) is disproportionality; (5) is corporatism; (6) is federalism; (7) is bicameralism; (8) is constitutions; (9) is judicial review; and (10) is central banks. c Statistically significant at the 1 percent level. d Statistically significant at the 5 percent level. e Statistically significant at the 10 percent level. b

mine).20 Table 3 presents the correlation matrix for all of these variables.21 It reveals much weaker relations between the variables than in Lijphart’s original study. Of the 10 significant correlations on each dimension in Lijphart’s sample, nine are significant on the first dimension and six on the second dimension in the expanded sample.22 Moreover, the size of the correlation coefficients is reduced in 17 of the 20 cases, often considerably. As we found above, the most distinctive variables can be found on the second dimension. Neither central bank independence nor judicial review is correlated with federalism and bicameralism as they should be. The 20

Lijphart provides data both for the 1945–1996 and 1971–1996 periods as well as alternative measures of several variables without his coding of allied or factionalized parties. I tested all possible codings and found similar results for all of them. The results listed here are for the most comparable of Lijphart’s datasets: i.e., the 1971–1996 period without his coding of allied and factionalized parties. These values are presented in all of the tables. 21 This table can be compared with Lijphart’s (1999, p. 244, table 14.1). 22 Given the larger sample size, one would expect significance levels to rise, ceteris paribus.

48

ANDREW ROBERTS

coefficient on each of these four correlations drops by more than half from Lijphart’s sample. This confirms our finding of a mixed second dimension. Bank independence is also correlated, though only at the .10 level, with party systems and cabinet types where it should not be. On the first dimension, corporatism does not correlate with executive dominance as it should—in fact, the correlation is negative rather than positive. Corporatism also appears to be related to federalism where it should not be. The substantive coefficients drop most strongly for corporatism, off by between 14 percent and 42 percent for the first three variables. Again, this confirms our finding above of a distinctive corporatism variable. Factor analysis reveals that there are not two but three factors with eigenvalues greater than one. While the first five variables load on a single factor as in Lijphart, the federal-unitary dimension breaks down into one factor for federalism and bicameralism and another for judicial review and central bank independence, with constitutional amendment provisions falling in between. To further verify that Eastern Europe is distinctive, I add an Eastern European dummy variable to the complete dataset. This variable is significantly correlated with six of Lijphart’s 10 variables at the .05 level. Compared to established democracies, the countries of Eastern Europe have higher numbers of parties, a lower percentage of time governed by minimum winning coalitions/single-party governments, lower government durations, lower levels of decentralization, and higher levels of judicial review and central bank independence. Rerunning the factor analysis with this dummy variable puts the Eastern European variable in a common factor with judicial review and central bank independence, suggesting that these are the most distinctive institutions in the region. Taken together, these findings suggest that Eastern Europe follows different dynamics than Lijphart’s established democracies. We should keep in mind that Lijphart’s cases are overrepresented in the dataset by 3.6 to 1 and thus exercise considerably greater influence on the results. Thus, it is also worth looking at the Eastern European dataset by itself. The correlation matrix for Eastern Europe is not presented because the small number of cases limits the statistical power of the exercise. Nevertheless, very few of the expected correlations are significant (even using a more generous 0.10 standard because of the small sample) and in the correct direction: only two out of 10 on the first dimension and one out of six on the second dimension.23 Eight of the 16 correlations expected by Lijphart carry the wrong sign. Factor analysis yields three principal factors, though they do not correspond to Lijphart.24 23

Since there was no variance on the federalism variable, it was not included in this analysis. The first factor includes the central bank, the constitution, and the negative of bicameralism; the second dimension is the party system, executive dominance, and disproportionality; the third combines cabinet types and corporatism.

24

DEMOCRACY IN EASTERN EUROPE

49

If we probe these admittedly weak results, we find that on the first dimension, higher numbers of parties and greater disproportionality are both correlated with longer government durations as we would expect. The cabinet types variable, however, behaves unusually—it is negatively correlated with duration and disproportionality, though not significantly so. As we will see below, bargaining failures may explain the unexpected behavior on cabinet types. This explanation bears noting because Lijphart believes this variable is most strongly correlated with the first dimension. Corporatism similarly is negatively correlated with duration and not correlated at all with disproportionality. On the second dimension, judicial review and central bank independence are correctly and significantly correlated, but this is virtually the only “normal” correlation. Confounding expectations, both bicameralism and constitutions are negatively correlated with judiciaries and banks, in three of the four cases at substantively large levels. This result corroborates some of the new patterns we found earlier. Again, these results should be taken with a grain of salt because of the small sample size. To summarize, it appears that though the countries have generally followed the consensus model on the executive-parties dimension and the majoritarian model on the federal-unitary dimension, several institutions do not fit the model. On the executive-parties dimension, interest group organization stands out like a sore thumb. Despite consensus behavior on most variables, interest group relations in these countries are quite pluralist. On the federal-unitary dimension, constitutional amendment procedures, judicial review, and central banks break with an initially majoritarian design. Given their unitary, centralized states, we would expect these countries to have flexible constitutions (there is no need to entrench the powers of provinces), weak judicial review (it is not necessary to resolve intergovernmental conflicts), and weak central banks. In fact, just the opposite is true. Lijphart’s patterns travel only partially to Eastern Europe.

EASTERN EUROPE AND THEORIES OF INSTITUTIONS: INSTITUTIONAL ORIGINS What do these results tell us about Eastern European politics and theories of institutions? In the first place, they lead us to ask why Eastern Europe is anomalous. Lijphart’s theory powerfully explains the relations among institutional forms in 36 advanced democracies. Why does Eastern Europe not fit with these results? Lijphart is not entirely clear on why he finds the patterns that he does. At times it appears that democracies simply have a choice between two philosophies of governing. Later he shows that an English colonial heritage is correlated with majoritarianism and ethnic diversity with consensus, but this does not fully explain his main result: the correlations he finds within each dimension.25 I will try to answer this question, looking separately at

50

ANDREW ROBERTS

Lijphart’s two dimensions. I find that Eastern Europe’s distinctiveness is related to both the domestic and the international context of its democratization.26

Executive-Parties Dimension Taagepera (2003) has come up with a more persuasive explanation for Lijphart’s patterns. He notes that on the first dimension, electoral laws drive disproportionality and the party system, which, in turn, affect coalition types and government duration. This observation does get us part of the way to understanding Eastern Europe. Most of these countries have chosen proportional electoral systems that typically produce low disproportionality, high numbers of parties, and, in turn, minority and supermajority cabinets and short-lasting governments.27 Indeed, the correlations between these elements hold up in the total sample and to a lesser extent within Eastern Europe. Of course, it is worth noting that Eastern European countries are not just consensus, but, on the variables of party system, cabinet types, and government duration, at the very consensus end of Lijphart’s scale. These extreme values are worth explaining. The high numbers of parties are probably due not just to the choice of PR,28 but to the undeveloped party systems in the region (Moser, 1999). Citizens have weak connections to parties, with the result that their vote choices are quite unstable. The high numbers of parties then have knock-on effects on cabinet types—with so many parties, one-party cabinets are unlikely to form—and duration: multiparty cabinets tend to be of shorter duration than singleparty cabinets (Warwick, 1994). The nature of the transition may also affect these two variables. The high costs of governing during massive recessions could lead parties to avoid government and could lead existing governments to have short durations. As I suggest below, bargaining problems, particularly between supporters and opponents of the old regime, may play a role in producing minority or oversized cabinets and short-lived governments. The bigger anomaly, however, appears when we look at disproportionality. While number of parties, cabinet types, and duration all are highly consensus, disproportionality appears quite majoritarian. To further 25

While a British influence might lead a country to reproduce Britain’s majoritarian institutions, it does not explain why Lijphart finds two separate dimensions of patterns. For general works on institutional design in Eastern Europe, see, among others, Elster et al. (1998) and Zielonka (2001). 27 Two countries have mixed electoral systems: Hungary and Lithuania. There is an extensive literature on the choice of electoral systems in Eastern Europe. See, for example, Benoit and Hayden (2004), Benoit and Schiemann (2001), Colomer (1995), and Geddes (1996). For a critical view, see Bernhard (2000). 28 For evidence that PR produces a higher number of parties than majoritarian systems in Eastern Europe, see Filippov et al. (1999) and Kostadinova (2002). For a dissenting view, see Moser (1999). 26

DEMOCRACY IN EASTERN EUROPE

51

confuse the picture, disproportionality should have a causal influence on the other variables because it is a direct consequence of the electoral system. I will pursue this point in the following part of the article, but here it is enough to note that a major reason why disproportionality is so high is that voters waste a high percentage of their votes. They often vote for parties that do not pass the threshold—typically five percent—for representation in parliament. For example, in Poland’s second free election in 1993, over 30 percent of votes went to parties that did not enter parliament. In general, voters and elites appear to have difficulty coordinating their actions. We can find the cause of this anomaly in the very unconsolidated party systems of the region. Communism had so thoroughly destroyed historical parties—along with the civil society that might have kept any parties strong—that virtually all parties are brand new and have sunk only shallow roots into society. We find an even larger anomaly when we move to corporatism, where almost all of the Eastern European countries are majoritarian. Anomalous values on this variable make sense in Taagepera’s rendering. Corporatism is a separate and distinct choice from the electoral system, though it does embody a similar principle—bringing additional groups into the governing process. Indeed, this principle seems to explain the correlation between the two institutions in advanced democracies.29 Why then did Eastern European countries choose PR along with majoritarian interest group relations? First off, we should note that in some sense they did not. All of them did set up tripartite bodies. But generally these bodies failed to work as in Western Europe. Two factors can explain why they did not work well (Ost, 2000). First, communism had eliminated almost all traces of civil and economic society, the source of the bargaining partners for corporatism. Labor unions were mere tools of the state and thus suffered serious legitimacy problems after the revolution. At the same time, employers’ organizations were missing altogether and, because of pervasive state-ownership, have been slow to develop. Rapid and comprehensive ownership changes meanwhile hindered the development of stable negotiating relationships. Further, as Kubicek (1999) notes, some of the same factors that have weakened labor unions and corporatism in the West have come to the fore in the East. In particular, globalization and capital flows have made corporatism harder to construct. In order to attract foreign capital, Eastern European countries need flexible labor markets. The shift from manufacturing to services and the smaller scale of new businesses has similarly weakened unionization. Only the most advanced of the Eastern European states have strong enough societies to demand a weakened form of corporatism and strong enough economies to afford it. 29 Iversen (2005) explains the correlation between PR and corporatism as stemming from the fact that PR systems tend to produce leftist governments and majoritarian systems rightist governments.

52

ANDREW ROBERTS

On the first dimension, then, coordination problems—on the one hand among voters and parties and on the other among economic actors—have prevented these countries from exhibiting the full-blown consensus system. In both cases, these coordination problems can be traced to the communist regime and to a lesser extent the international economy.

Federal-Unitary Dimension The anomalies grow stronger when we move to the second dimension of Lijphart’s typology, the federal-unitary dimension. Here, Eastern European countries are almost uniformly unitary, centralized, and unicameral—all majoritarian traits—but they combine these institutions with consensus-oriented constitutional amendment procedures, judicial review, and, especially, central banks. Taagepera (2003) again has an explanation for the weaker cohesion on this dimension. He argues that federalism logically implies bicameralism, a rigid constitution, and strong judicial review in order to preserve local power and resolve conflicts between the center and the provinces. Nonfederal states, however, are free to choose or not choose these ancillary institutions. Lijphart’s correlations, he suggests, are driven by the federal countries in the sample. This helps to explain why the correlations are even weaker here: federalism is diluted further with the addition of 10 nonfederal states. But how do we explain the consistent, albeit hybrid, patterns we see in Eastern Europe? It is easy to explain the choice of unitary, centralized systems. The Eastern Europe states in this sample are all small, mostly ethnically homogeneous states; federalism, by contrast, is chosen mainly in large or heterogeneous states. Centralization can be explained both by the volatile economic situation in these countries and by communist legacies. Severe recessions and the necessity of major economic reforms meant that power had to be concentrated in the center. More importantly, communism left a legacy of heavy centralization. The result is extremely low administrative capacities in local government. Indeed, a whole set of institutions necessary for decentralization—for example, land registers needed for property tax assessment—had to be built up from scratch. Unicameralism is likewise a communist legacy. In their drive to centralize power and disenfranchise powerful groups, communist regimes tended to abolish upper houses. While newly democratic countries might have resurrected a bicameral system, Elster (1996) argues that constituent assemblies followed the inherited cameral structure and through selfinterest tended to preserve it. Thus, “[b]icameral assemblies in Poland and Romania created bicameral constitutions; unicameral assemblies in Hungary, Slovakia, and Bulgaria created unicameral constitutions.” Similar influences can be seen in other states.30 Unicameral countries had no reason 30

For details on apparent anomalies, see Elster (1996).

DEMOCRACY IN EASTERN EUROPE

53

to choose bicameralism—there were no particular interests for them to represent—and self-interest led elites to protect their positions. Countries that for whatever reason found themselves with a second chamber tended to preserve bicameralism; otherwise, legislators would have been eliminating their own jobs. While communist legacies influenced these two institutions, international pressures appear to be equally strong in causing amendment procedures, judicial review, and central banks to be put in place.31 Rule of law and what Tate and Vallinder (1995) call the “global expansion of judicial power” were watchwords in academic and aid circles when these countries wrote their constitutions. International advice proffered by constitutional experts was almost unanimously in favor of rigid constitutions and strong courts.32 The international consensus was even stronger on central bank independence. In this last case, there were also powerful networks of influence that led countries to adopt independent banks (Johnson, 2002). Indeed, Cukierman et al. (2002) find that countries in the first wave of accession to the EU tended to have higher levels of independence. In all three of these cases, domestic forces also played a role. In particular, there was a backlash against the institutional legacies of communism. Communist constitutions were not worth the paper they were written on and communist courts had little independence or respect for law. Rigid constitutions and strong judicial review were seen as ways to overcome this legacy and create a law-based society. Bank independence would similarly provide insurance against the sort of predatory behavior common under the old regime. New democratic elites, particularly when they were connected with the opposition, thus reacted against inherited institutions. Probing more deeply, one notices the importance of the unpopularity of the communist system of government in these countries. The decision to choose institutions dramatically different from those of the communist regime was taken partially because of dissatisfaction with that regime.33

The Context of Democratization What this all shows is that the choice of institutions and the way they function depend heavily on the context of democratization. “Context” here means both the nature of the pre-democratic regime and the international environment. As far as the preceding regime goes, communism was distinctive in destroying all political parties and all legal restraints on government power. On the one hand this led, as I show in the next section, to unconsolidated party systems and the corresponding coordination failures 31

For more on international influences, see Zielonka (2001). For a dissenting view, see Holmes and Sunstein (1995). 33 One sees the same phenomenon with respect to electoral systems. Electoral systems in communist regimes typically consisted of single-member districts. 32

54

ANDREW ROBERTS

between voters and parties. On the other hand, it led to a desire for institutions that would limit the power of governments. The international environment exercises influence through the global economy—which made new corporatist arrangements difficult to introduce and independent central banks almost a necessity—and ideational pressures, which influenced constitutions and judicial review. Lijphart’s patterns may thus be an artifact of the context in which his countries democratized. Most of them democratized at a time when international advice on institutional design was unknown and in reaction against ancien regimes that respected the rule of law and allowed political parties and markets to function. I should add that past democratizations also took place in an economic era far more conducive to corporatism. It is likely then that the strong correlations Lijphart observes within each dimension are an artifact that emerges only under certain conditions. Given present trends of increasing international influence and the fact that the remaining non-democracies have highly undeveloped civil, political, and economic societies,34 it is unlikely that new democracies will manifest the cohesion of Lijphart’s democracies. More likely, they will be the type of hybrids we see in Eastern Europe, mixing and matching government forms as circumstances dictate.

PATTERNS AND POLITICS Another question is the consequence of these choices. Lijphart (1999) argues that particular types of institutions tell us something important about how a country is governed.35 Countries that score high on consensus variables are ones characterized by representation, compromise, and negotiation, while those with majoritarian traits are home to efficiency, conflict, and adversarial politics. Here, I consider how well these connections hold in Eastern Europe. I focus on Lijphart’s first dimension where most national policy making takes place. While most of the findings in this section are speculative, they do raise some doubts about whether institutions determine politics. The Eastern European countries considered here have mainly adopted consensus institutions on the first dimension—they have PR systems, many parties, non– minimum winning cabinets, and short-lived governments. If we standardize all five first-dimension variables and average them to produce a summary measure of government types, we find that nine out of 10 of the Eastern European countries (all except Bulgaria) are on the consensus side of the map.36 34

On these terms, see Linz and Stepan (1996). He tests these relations, however, not at the level of politics, but only at the level of social outcomes. 36 This conclusion does not contradict the lack of patterns found above. By averaging all of the variables, the lack of patterns is obscured. 35

DEMOCRACY IN EASTERN EUROPE

55

What is not clear is whether politics in these countries takes on a particularly consensus cast. Rather, there are preliminary indications that they may be characterized more by lack of representation, conflict, and bargaining failures.37 Lijphart’s institutional measures might thus be masking a different style of politics than that found in the consensus democracies that he studies.

Representation The initial impetus for consensus institutions like PR was a belief that they would better represent the diversity of society. In one sense this applies to Eastern Europe as well. These countries do feature a large number of parties, representing a variety of ideologies. But does this mean that citizens are being represented? One problem is wasted votes. Voters in Eastern Europe often choose parties that do not enter parliament. In a typical election, the number of wasted votes—those that go toward parties that do not win any seats—is about 12 percent, double the percentage of Western Europe. The real difference may even be greater because in Western Europe many wasted votes are consciously wasted, while those in Eastern Europe are probably wasted because of miscalculations. In one basic sense, then, consensus institutions are not doing their job. Poland is a prime example. Its next-to-last election, in 2001, the fourth since the fall of communism, featured the complete disappearance of the Solidarity and Freedom Union parties, which had governed the country from 1997 to 2001, and the emergence of four new parties that had never before been represented in parliament. Only two of the six parties entering parliament had a history of representation, and one of these had almost collapsed by the end of its term. The roots of this representational failure are likely in unstable party systems. In established democracies, the same parties compete over and over again in elections. This helps voters to coordinate because they are continually choosing from the same slate of parties. In Eastern Europe, by contrast, there are dramatic changes in the competitors between elections. If we look more generally at levels of volatility—changes in votes for parties between elections—we find values about three times as high as Western Europe and twice as high as Western European countries at the time they became democratic (Rose and Munro, 2003). Most of this volatility is caused by votes for brand new parties (Rose and Munro, 2003; Birch, 2003). This fact goes a long way toward explaining the coordination prob37

This is not to imply that representation and accommodation are the most important values in a political system. There is a tradeoff between these values and others such as accountability and efficiency, which are better represented by a majoritarian system. There are good reasons to prefer either one of these. The following sections ask only whether the predominantly consensus institutions in Eastern Europe produce typically consensus politics. For a good discussion of these issues, see Powell (2000).

56

ANDREW ROBERTS

lem of voters. At each election, voters are presented with a different slate of parties. This makes it difficult for voters to coordinate on electorally viable parties and produces the unusual combination of high numbers of parties and high levels of disproportionality. These phenomena might be seen as typical of new democracies, and indeed their like can be found in Latin America. However, one would expect them to decline over time as parties establish themselves and citizens learn to recognize them. In fact, levels of volatility have not dropped appreciably in Eastern Europe. Average volatility in the region rose between the third and fourth elections, and even at the fifth election the number of new parties appearing exceeded the number of parties disappearing (Rose and Munro, 2003).38 The problem of wasted votes and high volatility might be mitigated if the parties that do enter parliament represent their electorates—that is, if they act on behalf of their constituents and enact the policies they have pledged to enact. If this happened, we could still say that these countries are characterized by representation. While few studies have tackled this issue, many scholars have expressed doubts about the ability and desire of parties to represent their voters. These doubts again flow from a consideration of party systems in the region. In the first place, the transience of parties means they will find it difficult to build close links with voters. Exacerbating this tendency, even established parties appear to be elite cliques with few connections to their voters (Kostelecky, 2002; Lewis, 1996, 2000). On the side of the voters, party membership in the region is quite low and most citizens are not involved in politics (Howard, 2003). Complicating party-constituent links even further is the very high degree of party switching in post-communist assemblies (Shabad and Slomczynski, 2004; Kreuzer and Pettai, 2003). All of these factors, however, are circumstantial. They would appear to reduce representation, but they do not show that representation is lacking. The one study to look systematically at the issue has, in fact, shown just the opposite. Kitschelt et al. (1999) persuasively demonstrate that there is a close correspondence between party programs and the policy preferences of their voters. Parties do take definite positions on issues, and these positions are discernable to voters. Tworzecki (2003) similarly finds 38

Why do party systems in the region remain so unstable? Mair (1997) finds an answer in the nature of both parties and the electorate. In the first place, a pre-existing party system is lacking. Communism had destroyed most historical parties so completely that very few re-emerged after 1989. The new parties that were created, moreover, were mostly top-down organizations formed within parliament (Kostelecky, 2002; Lewis 1996, 2000; for a review, see Kreuzer and Pettai, 2004). As a result, their links with voters were not strong. At the same time, cleavages in society were weak and rapidly changing, and many citizens inherited a suspicion of mass participation from communist times (Bielasiak, 2002). Severe transformational recessions and scandals connected with privatization likewise tested citizens’ attachments to any party.

DEMOCRACY IN EASTERN EUROPE

57

evidence that voters do know what policies individual parties favor and support them at least partially on this basis. The degree to which parties and governments represent voters remains one of the key questions in the study of post-communist democracies. It gets directly at the issue of how well democracy is working in the region. We would expect high-quality democracies to have high levels of representation. To date, findings have been ambiguous, with studies of the context of policy making finding low representation and those looking directly at the issue of representation producing more positive results. Still in short supply, but particularly valuable, are studies that relate actual policies (rather than programs) with public preferences.

Compromise and Accommodation The other main advantage of consensus government—its higher degree of compromise and accommodation—is purported to flow from the consistency of party representation. As Lijphart (1977) showed in his work on consociationalism, it is the long-term interaction of multiple parties that produces tendencies of compromise and accommodation.39 As the same parties face the same situations over and over again they develop formal and informal modes of living and working together. Given that parties come and go in Eastern Europe, one might expect compromise and accommodation also to be in short supply in the region.40 Mair (1997) shows other ways in which the superficial consensualism of multiple, but ephemeral, parties may lead to conflict rather than compromise in Eastern Europe. The very open market for votes means that parties will see each other as life-and-death adversaries. According to Mair, only where there is some political certainty can cooperation and accommodation arise. The fact that the stakes of elections are already high in Eastern Europe because of the enormous number of political and economic reforms that governments must undertake only adds to the tendency for conflict. The non-institutionalized party systems of the region may thus undermine many of the supposed benefits of a multi-party consensus system. Evidence on these matters to date has been anecdotal—compromise and negotiation are not easy concepts to measure—but a new strand of research has looked for evidence of these trends in coalition government. The study of coalition government has the dual advantages that we know a considerable amount about how coalitions form in Western Europe (Laver and Schofield, 1990) and many of the outcomes of coalition bargaining can be measured. According to Lijphart, the large number of consensus-type cabinets in the region and their short duration are evidence of compromise and cooperation. But there is also emerging evidence that this may not be the 39

For a review of this literature see Andeweg (2000). It is not clear, though, whether elites turn over as quickly as parties. If elites are more constant, compromise may still be possible. 40

58

ANDREW ROBERTS

case. These minority and surplus majority governments may instead arise as a result of failures in bargaining and a desire not to accommodate rivals. For example, recent research by Grzymala-Busse (2001) and Druckman and Roberts (2004, 2005) finds that coalition formation in the region is heavily influenced by the divide between parties connected with the old regime and parties connected with the former opposition. Grzymala-Busse (2001), for example, finds that in certain countries coalitions are unlikely to form between what Tucker (2006) calls new regime and old regime parties. Druckman and Roberts (2004) add that communist successor parties— those that are the organizational successors to the former ruling parties— are discriminated against in almost all aspects of coalition governance. They tend not to be invited to join coalitions, and when they do join, they are awarded fewer ministries than other parties. These results hold even when we control for ideology: parties may share programmatic commitments, but refuse to cooperate because of bad blood over the old regime. It appears that residual hostility towards communism still prevents cooperation between political parties. Even when coalitions with communists do form, it is usually out of necessity rather than from a desire to cooperate. As with representation, more research on these issues is necessary. There have been few detailed studies to date about how parties actually govern—whether they compromise and negotiate or govern in a majoritarian fashion. While initial results suggest a distinctive sort of bargaining in Eastern Europe, we do not know how far this extends to the creation of real policies. We might speculate on one final consequence of these superficial consensus forms for post-communist politics. One of the puzzles of Eastern European politics is its high level of corruption. Though many expected that corruption would decline in these countries as they became consolidated democracies, in fact, corruption remains quite high. I would speculate that corruption is the grease that allows governments to function. Indeed, what else but exchanges of political favors can explain how countries with large numbers of ephemeral parties without strong links to society have managed to remain democratic without crippling conflict or paralysis. And while we expect corruption to decrease as citizens hold politicians accountable for their corruption, in these countries unstable party systems may be standing in the way of accountability. What this all suggests is not that Lijphart’s consensus category is useless, but that his measures of consensus government need to be supplemented by other contextual ones. Multiparty systems, minority or surplus majority cabinets, and short duration governments are good indications of consensus-style politics in Western Europe because these countries have strong and stable party systems linked to their societies. This leads to repeated interactions and the creation of traditions of representation, compromise, and accommodation. While Eastern Europe shares with the West high numbers of parties and consensus cabinets, its non-institutionalized party systems mean that

DEMOCRACY IN EASTERN EUROPE

59

they may not produce the effects that Lijphart predicts. Instead, there may be failures in representation as well as considerably more conflict than we might expect. Indeed, institutionalized party systems might be seen as an essential precondition of consensus government.

CONCLUSION Since it was first published in 1984, Lijphart’s distinction between consensus and majoritarian government has been one of the most influential theories on institutional forms and their effects in political science. Yet, as this article shows, some of its key findings may not travel well to new democracies. In the first place, cohesion between the 10 variables, one of the strongest and most surprising findings in his work, does not apply to Eastern European democracies. Their patterns are hybrids. According to this article, the key causes of these hybrids are in the context of democratization— specifically, the presence of communist legacies and international pressures. Especially noteworthy is the influence of the old regime. By wiping out most of political, economic, and civil society, communism made it that much more difficult for countries to adopt consistent institutional patterns in areas like disproportionality, corporatism, and bicameralism. These differences between Eastern Europe and established democracies correspond with the different contexts of Huntington’s (1991) waves of democracy. Lijphart’s countries are mainly first- and second-wave democracies that typically emerged after a long history of party development and gradual liberalization and in a less interconnected global environment. By contrast, the third-wave democracies of Eastern Europe emerged from much more severe tyrannies, and they emerged into an international system less respectful of national sovereignty. They may thus represent a new variety of democracy that will become more widespread. Second, though here evidence is more speculative, some of the potential advantages of consensus democracies may not emerge in these new democracies, at least not yet. Representation, which should follow from the choice of PR, does not necessarily emerge where there are unstable party systems. Compromise and negotiation also appear to be in short supply, though supporting evidence is more anecdotal. These results suggest that traditional measures of institutions and their effects may not be revealing the same information that they do in established democracies. More subtle measures that consider context and particularly a country’s party system may be more valuable than the traditional focus on electoral laws and coalition types alone. We may thus tentatively say that the Eastern European countries represent a new type of governing pattern, at least for the time being. They are characterized largely by consensus government but they lack the party system stability which produces such consensus traits as representation and compromise. We might call them consensus governments in the mak-

60

ANDREW ROBERTS

ing. As party systems stabilize and communist legacies recede, they are likely to become more consistently consensus. Three further questions should be explored in future research. First, does this new type of democracy apply only to Eastern Europe or can we find similar trends in other new democracies? Lijphart’s sample includes only three democracies in Latin America, so this region is a logical place to add new data. The results from this article predict that newly emerging democracies are more likely to be hybrids. The international pressures and authoritarian legacies that affected Eastern Europe are likely to affect newly emerging democracies in Africa and Asia as well. This prediction needs to be tested on new data. Second, do these outwardly consensus systems actually behave in a consensus manner? I suggested earlier that one can raise doubts about the degree to which their political systems manifest the characteristics of representation and accommodation, but that hard conclusions are still lacking. More research on these issues will tell us not only about the applicability of Lijphart’s theory, but also the quality of democracy in the region. Third, do these patterns matter? Lijphart finds that consensus institutions produce a kinder and gentler democracy. The key question for future research is to determine whether the patterns in Eastern Europe correlate with substantive policy outcomes there as well. Some have argued that consensus institutions are related to positive economic reform outcomes (Commander and Frye, 1999; Hellman, 1998). Fortin (2005) has likewise attempted to relate Lijphart’s distinction to a variety of societal outcomes. These preliminary results need to be supplemented both with new dependent variables and more rigorous control variables. These tests constitute the next frontier of research.

REFERENCES Andeweg, Rudy B., “Consociational Democracy,” Annual Review of Political Science, 3:509–536, June 2000. Benoit, Kenneth and Jacqueline Hayden, “Institutional Change and Persistence: The Evolution of Poland’s Electoral System, 1989–2001,” Journal of Politics, 66, 2:396–427, May 2004. Benoit, Kenneth and John W. Schiemann, “Institutional Choice in New Democracies: Bargaining over Hungary’s 1989 Electoral Law,” Journal of Theoretical Politics, 13, 2:153–182, April 2001. Bernhard, Michael, “Institutional Choice after Communism: A Critique of TheoryBuilding in an Empirical Wasteland,” East European Politics and Societies, 14, 2:316– 347, Spring 2000. Bielasiak, Jack, “The Institutionalization of Electoral and Party Systems in Postcommunist States,” Comparative Politics, 34, 2:189–210, January 2002. Birch, Sarah, Electoral Systems and Political Transformation in Post-Communist Europe. New York, NY: Palgrave Macmillan, 2003.

DEMOCRACY IN EASTERN EUROPE

61

Bird, Richard M., Robert D. Ebel, and Christine I. Wallich, eds., Decentralization of the Socialist State: Intergovernmental Finance in Transition Economies. Washington, DC: International Bank for Reconstruction and Development/The World Bank, 1995. Blaustein, Albert P. and Gisbert H. Flanz, Constitutions of the Countries of the World. Dobbs Ferry, NY: Oceana Publications, 1971–present. Blondel, Jean and Ferdinand Müller-Rommel, eds., Cabinets in Eastern Europe. New York, NY: Palgrave, 2001. Casale, Giuseppe, ed., Social Dialogue in Central and Eastern Europe. Geneva: International Labour Office, 1999. Colomer, Josep M., “Strategies and Outcomes in Eastern Europe,” Journal of Democracy, 6, 2:74–85, April 1995. Commander, Simon and Timothy Frye, “The Politics of Economic Reform,” in Transition Report 1999: Ten Years of Transition. London, UK: European Bank for Reconstruction and Development, 1999. Crowley, Stephen, “Explaining Labor Quiescence in Post-Communist Europe: Historical Legacies and Comparative Perspective,” Central and Eastern European Working Paper No. 56, Minda de Gunzburg Center for European Studies, Harvard University, Cambridge, MA, 2003. Crowley, Stephen and David Ost, eds., Workers after Workers’ States: Labor and Politics in Postcommunist Eastern Europe. Lanham, MD: Rowman and Littlefield, 2001. Cukierman, Alex, Geoffrey P. Miller, and Bilin Neyapti, “Central Bank Reform, Liberalization, and Inflation in Transition Economies—An International Perspective,” Journal of Monetary Economics, 49, 2:237–264, March 2002. Druckman, James and Andrew Roberts, “Communist Successor Parties and Coalition Formation,” paper presented at the annual meeting of the American Political Science Association, Chicago, IL, September 2–5, 2004. Druckman, James N. and Andrew Roberts, “Context and Coalition-Bargaining: Comparing Portfolio Allocation in Eastern and Western Europe,” Party Politics, 11, 5:535– 555, September 2005. Ebel, Robert D. and Serdar Yilmaz, “On the Measurement and Impact of Fiscal Decentralization,” Policy Research Working Paper Series, WPS 2809, Economic Policy and Poverty Reduction Division, World Bank Institute, World Bank, Washington, DC, March 2002. Elster, Jon, “The Role of Institutional Interest in East European Constitution-Making— Explaining Legislative Dominance,” East European Constitutional Review, 5, 6:63–65, Winter 1996. Elster, Jon, Claus Offe, and Ulrich K. Preuss, Institutional Design in Post-Communist Societies: Rebuilding the Ship at Sea. Cambridge, UK: Cambridge University Press, 1998. Filippov, Mikhail G., Peter C. Ordeshook, and Olga V. Shvetsova, “Party Fragmentation and Presidential Elections in Post-Communist Democracies,” Constitutional Political Economy, 10, 1:3–26, March 1999. Fortin, Jessica, “Patterns of Democracy? Counter-Evidence from Nineteen Post-Communist Countries,” paper presented at the European Consortium for Political Research Joint Sessions, Grenada, April 14–19, 2005. Geddes, Barbara, “Initiation of New Democratic Institutions in Eastern Europe and Latin America,” in Arend Lijphart and Carlos H. Waisman, eds., Institutional Design in New Democracies: Eastern Europe and Latin America. Boulder, CO: Westview, 1996.

62

ANDREW ROBERTS

Gibson, John and Philip Hanson, eds., Transformation from Below: Local Power and the Political Economy of Post-Communist Transitions. Cheltenham, UK: Edward Elgar, 1996. Grzymala-Busse, Anna, “Coalition Formation and the Regime Divide in New Democracies: East Central Europe,” Comparative Politics, 34, 1:85–104, October 2001. Hall, Peter A. and David Soskice, eds., Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford, UK: Oxford University Press, 2001. Heinisch, Reinhard, “The State of Corporatism in a Central Europe in Transition,” in Irwin Collier, Herwig Roggemann, Oliver Scholz, and Horst Tomann, eds., Welfare States in Transition: East and West. New York, NY: St. Martin’s Press, 1999. Hellman, Joel S., “Winners Take All: The Politics of Partial Reform in Postcommunist Transitions,” World Politics, 50, 2:203–234, January 1998. Holmes, Stephen and Cass Sunstein, “The Politics of Constitutional Revision in Eastern Europe,” in Sanford Levinson, ed., Responding to Imperfection: The Theory and Practice of Constitutional Amendment. Princeton, NJ: Princeton University Press, 1995. Horvath, Tamas M., ed., Decentralization: Experiments and Reforms. Budapest: Local Government and Public Services Reform Initiative, 2001. Howard, Marc Morje, The Weakness of Civil Society in Post-Communist Europe. Cambridge, UK: Cambridge University Press, 2003. Huntington, Samuel P., The Third Wave: Democratization in the Late Twentieth Century. Norman, OK: University of Oklahoma Press, 1991. Iankova, Elena A., “The Transformative Corporatism of Eastern Europe,” East European Politics and Societies, 12, 2:222–264, Spring 1998. Iversen, Torben, Capitalism, Democracy, and Welfare. New York, NY: Cambridge University Press, 2005. Johnson, Juliet, “Agents of Transformation: The Role of the West in Post-Communist Central Bank Development,” Studies in Public Policy No. 361, Centre for the Study of Public Policy, University of Strathclyde, Glasgow, 2002. Katz, Richard S. and Ingrid Van Biezen, eds., Political Data Yearbook 2003, a special issue of the European Journal of Political Research, 43, 7-8:919–1175, December 2004. Katz, Richard S. and Ingrid Van Biezen, eds., Political Data Yearbook 2004, a special issue of the European Journal of Political Research, 44, 7-8:919–1230, December 2005. Kimball, Jonathan D., ed., The Transfer of Power: Decentralization in Central and Eastern Europe. Budapest: Local Government and Public Services Reform Initiative, 1999. Kirchner, Emil J., ed., Decentralization and Transition in the Visegrad: Poland, Hungary, the Czech Republic and Slovakia. New York, NY: St. Martin’s Press, 1999. Kitschelt, Herbert, Zdenka Mansfeldova, Radoslaw Markowski, and Gabor Toka, Post-Communist Party Systems: Competition, Representation, and Inter-Party Cooperation. New York, NY: Cambridge University Press, 1999. Kostadinova, Tatiana, “Do Mixed Electoral Systems Matter? A Cross-National Analysis of Their Effects in Eastern Europe,” Electoral Studies, 21, 1:23–34, March 2002. Kostelecky, Tomas, Political Parties after Communism: Developments in East Central Europe. Washington, DC: Woodrow Wilson Center Press, 2002. Kreuzer, Marcus and Vello Pettai, “Patterns of Political Instability: Affiliation Patterns of Politicians and Voters in Post-Communist Estonia, Latvia, and Lithuania,” Studies in Comparative International Development, 38, 2:76–98, Summer 2003. Kreuzer, Marcus and Vello Pettai, “Political Parties and the Study of Political Development: New Insights from the Postcommunist Democracies,” World Politics, 56, 4:608–633, July 2004.

DEMOCRACY IN EASTERN EUROPE

63

Kubicek, Paul, “Organized Labor in Postcommunist States: Will the Western Sun Set on It, Too?,” Comparative Politics, 32, 1:83–102, October 1999. Laakso, Markku and Rein Taagepera, “ ‘Effective’ Number of Parties: A Measure with Application to West Europe,” Comparative Political Studies, 12, 1:3–27, April 1979. Laver, Michael and Norman Schofield, Multiparty Government: The Politics of Coalition in Europe. Oxford, UK: Oxford University Press, 1990. Lewis, Paul G., ed., Party Structure and Organization in East-Central Europe. Cheltenham, UK: Edward Elgar, 1996. Lewis, Paul G., Political Parties in Post-Communist Eastern Europe. London, UK: Routledge, 2000. Lijphart, Arend, Democracy in Plural Societies: A Comparative Exploration. New Haven, CT: Yale University Press, 1977. Lijphart, Arend, Democracies: Patterns of Majoritarian and Consensus Government in Twenty-One Countries. New Haven, CT: Yale University Press, 1984. Lijphart, Arend, Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries. New Haven, CT: Yale University Press, 1999. Linz, Juan J. and Alfred Stepan, Problems of Democratic Transition and Consolidation: Southern Europe, South America, and Post-Communist Europe. Baltimore, MD: The Johns Hopkins University Press, 1996. Mair, Peter, Party System Change: Approaches and Interpretations. Oxford, UK: Oxford University Press, 1997. Moser, Robert G., “Electoral Systems and the Number of Parties in Postcommunist States,” World Politics, 51, 3:359–384, April 1999. Müller-Rommel, Ferdinand, Katja Fettelschoss, and Philipp Harfst, “Party Government in Central Eastern European Democracies: A Data Collection (1990–2003),” European Journal of Political Research, 43, 6:869–893, October 2004. Orenstein, Mitchell and Lisa E. Hale, “Corporatist Renaissance,” in Christopher Candland and Rudra Sil, eds., The Politics of Labor in a Global Age: Continuity and Change in Late-Developing and Post-Socialist Economies. Oxford, UK: Oxford University Press, 2001. Ost, David, “Illusory Corporatism in Eastern Europe: Neoliberal Tripartism and Postcommunist Class Identities,” Politics and Society 28, 4:503–530, December 2000. Powell, G. Bingham, Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven, CT: Yale University Press, 2000. Rose, Jürgen and Johannes Ch. Traut, eds., Federalism and Decentralization: Perspectives for the Transformation Process in Eastern and Central Europe. Münster: Lit Verlag, 2001. Rose, Richard and Neil Munro, Elections and Parties in New European Democracies. Washington, DC: CQ Press, 2003. Rychly, Ludek and Rainer Pritzer, “Social Dialogue at National Level in the EU Accession Countries,” Working Paper, InFocus Programme on Social Dialogue, Labour Law and Labour Administration, International Labour Office, Geneva, February 2003, available online at www.ilo.org/public/english/region/eurpro/ geneva/conf/malta2003/soc_nat_dial.pdf. Shabad, Goldie and Kamizierz M. Slomczynski, “Inter-Party Mobility among Parliamentary Candidates in Post-Communist East Central Europe,” Party Politics, 10, 2:151–176, March 2004. Schwartz, Herman, “Eastern Europe's Constitutional Courts,” Journal of Democracy, 9, 4:100–114, October 1998.

64

ANDREW ROBERTS

Schwartz, Herman, The Struggle for Constitutional Justice in Post-Communist Europe. Chicago, IL: University of Chicago Press, 2000. Siaroff, Alan, “Corporatism in 24 Industrial Democracies: Meaning and Measurement,” European Journal of Political Research, 36, 2:175–205, October 1999. Smithey, Shannon I. and John Ishiyama, “Judicious Choices: Designing Courts in PostCommunist Politics,” Communist and Post-Communist Studies, 33, 2:163–182, June 2000. Taagepera, Rein, “Arend Lijphart’s Dimensions of Democracy: Logical Connections and Institutional Design,” Political Studies, 51, 1:1–19, March 2003. Tate, C. Neal and Torbjorn Vallinder, eds., The Global Expansion of Judicial Power. New York, NY: New York University Press, 1995. Thirkell, John, Richard Scase, and Sarah Vickerstaff, eds., Labor Relations and Political Change in Eastern Europe: A Comparative Perspective. Ithaca, NY: ILR Press, 1995. Tucker, Joshua A., Regional Economic Voting: Russia, Poland, Hungary, Slovakia and the Czech Republic, 1990–1999. Cambridge, UK: Cambridge University Press, 2006. Tworzecki, Hubert, Learning to Choose: Electoral Politics in East-Central Europe. Stanford, CA: Stanford University Press, 2003. Warwick, Paul, Government Survival in Parliamentary Democracies. Cambridge, UK: Cambridge University Press, 1994. Wetzel, Deborah, “Decentralization in the Transition Economies: Challenges and the Road Ahead,” Poverty Reduction and Economic Management Unit—Europe and Central Asia, World Bank, Washington, DC, June 2001, available online at www1.worldbank.org/wbiep/decentralization/library1/ decentralization_transitionecon.pdf. Zielonka, Jan, ed., Democratic Consolidation in Eastern Europe. Oxford, UK: Oxford University Press, 2001.