Scholarly article on topic 'Party Groups and Policy Positions in the European Parliament'

Party Groups and Policy Positions in the European Parliament Academic research paper on "Political Science"

0
0
Share paper
Academic journal
Party Politics
OECD Field of science
Keywords
{""}

Academic research paper on topic "Party Groups and Policy Positions in the European Parliament"

PARTY POLITICS

VOL 13. No.1 pp. 5-28

Copyright © 2007 SAGE Publications London ■ Thousand Oaks ■ New Delhi

www.sagepublications.com

PARTY GROUPS AND POLICY POSITIONS IN THE EUROPEAN PARLIAMENT

Gail McElroy and Kenneth Benoit

ABSTRACT

As the legislative body of the European Union (EU), the European Parliament (EP) comprises 732 elected representatives from over 150 national political parties from 25 member states. At the EP level, these members are affiliated with seven major party groups representing distinct policy positions. Here we provide precise estimates of these policy positions based on expert surveys, in addition to characterizing the dimensionality of policy competition in the EP. Our results suggest not only that party groups have identifiable and differentiated positions on multiple issues of policy, but also that these positions group broadly into two orthogonal dimensions: one consisting of classic left-right social and economic issues, and the other related to the powers and scope of EU institutions.

KEY WORDS ■ European Parliament ■ expert surveys ■ party competition ■ policy positions

Policy and Party Competition in the European Parliament

Legislative activity and decision-making in the European Parliament (EP) is organized around political groups, partisan entities that perform most of the same functions at the transnational level as do parties in national parliaments. Political groups act as the gatekeepers to legislative amendments and the vast bulk of interruptive and procedural motions in the EP. In addition, political group size determines membership of committees and most other key parliamentary decision-making bodies. Despite the important role played by these groups in the functioning of the increasingly powerful EP, however, our understanding and knowledge of the policy space in which they compete remain rudimentary. This article addresses this gap in our knowledge by reporting the results of an expert survey of the policy positions of the political groups in the EP.

1354-0688[DOI: 10.1177/1354068806068593]

From the outset, the EP has been organized in terms of political groupings. Since its inception in 1957, political groups in the EP have been officially recognized in the rules of procedure, and have received financial support for administrative costs from the parliamentary budget. Seven political groups currently exist in the EP, representing over 150 different political parties from the 25 member states. These political groups vary significantly in their degree of institutionalization. Several, such as the European People's Party (EPP) and the Party of European Socialists (PES), have been in existence for over 50 years. Others, such as the European United Left/Nordic Green Left (GUE), are much younger. The largest party, the EPP, has representatives from all 25 member states, comprising a total of 266 representatives from 42 different national political parties. At the other end of the spectrum, the Union for a Europe of Nations (UEN) has just 27 members from only 6 states of the European Union (EU).

Party groups in the EP have become increasingly cohesive and powerful over time. Levels of voting cohesion have been rising across parliamentary sessions, especially for the three largest political groups, despite increases in the size of the EP and in the number of member states (Hix et al., 2005; Raunio, 1997). The increasing importance of the political groups in the EP is not unrelated to the concomitant increase in powers of the Parliament itself. Once a purely consultative institution, the EP is now a co-legislator with the European Council for the roughly 70 percent of legislation that is currently adopted within the co-decision framework. We thus view it as increasingly important to understand better the political space in which these party groups compete and the policy positions they occupy along the dimensions of this space.

In this article we provide the first published estimates of the policy positions of the EP political groups using expert surveys measured just before the European elections of June 2004. Previous attempts to infer these positions have used a variety of indirect methods, such as codings of European election manifestos (Gabel and Hix, 2004), interviews with European elites (Arregui et al., 2004), surveys of MEPs (Thomassen et al., 2004) and analyses of roll-call votes (Hix et al., 2005). Our approach, by contrast, locates the party groups directly by asking experts to place them on predefined policy dimensions. Using the estimates, we also characterize the policy space of EP party competition. Finally, we explore interesting patterns between EP party group positioning and the policy positions of each group's national party members.

In what follows, we discuss different approaches to measuring European policy positions, highlighting the benefits of the expert survey methodology in the EP context. Next, we describe our expert survey and then present and discuss the results. Following that, we use factor analysis to measure the dimensionality and components of the EP policy space, comparing our results to previous findings. Finally, we offer preliminary observations on the interaction of the EP groups and their national affiliates in terms of policy convergence, suggesting promising avenues for emerging research.

Measurement Approaches to European Policy Positions

Previous research has approached the problem of measuring the policy positions of the European party groups in several distinct ways. These approaches can be distinguished along two main dimensions; one associated with the empirical level of measurement and the other with more procedural, methodological approaches to measuring policy positions.

With regard to the level of measurement, previous approaches to estimating European party group positions have either attempted to estimate these directly by focusing on the transnational entities themselves, or instead to estimate them indirectly through measuring the policy positions of their national-level member parties. Direct measures include analysis of roll-call votes (Hix et al., 2005; Noury, 2002), Stokman and Thomson's (2004) expert interviews of political preferences on 66 different Commission proposals, a survey of MEPs (Thomassen et al., 2004), and an analysis of European manifestos (Gabel and Hix, 2004). Indirect measures, on the other hand, rely on estimating the policy positions of party groups in the EP through direct measures of the positions of non-EP actors associated with the EP party groups. After measuring the policy positions of these associated actors, the EP group position is assumed to be the average of their associated groups. For instance, the European Election surveys used by Thomassen and Schmidt (1997) measure the policy positions of mass publics and European Election candidates associated with EP party groups. Alternatively, expert surveys of national party positions, or CMP estimates based on national party manifestos, might be used to estimate EP party group positions. For reasons we highlight later in this article, however, it should not be assumed that EP policy positions are always determined by the central tendency of their national party members or mass public positions. Indeed, we see this degree of convergence as one of the more interesting research questions to submit to empirical testing.

With regard to more procedural issues of measurement methodology, previous approaches to estimating the policy positions of parties have done so in diverse ways. These have included opinion surveys, expert surveys, expert interviews, analysis of party manifestos, and multidimensional scaling from roll-call votes. Each method has advantages and disadvantages. Roll-call votes, for example, provide objective political actions from which parties can be inductively located on policy scales, using statistical techniques of multidimensional scaling. Roll-call votes may suffer selection bias, however, since they may be called selectively and strategically depending on political outcome, as well as for only certain issues, especially in the EP (Kreppel, 2002: 129). In addition, the substantive interpretation of the policy scales which they produce must be interpreted, and these are not always clear (see Hix et al., 2005).

The approach we use here is that of expert surveys: systematic placements of EP groups on numerous predefined policy dimensions, based on

systematically collected judgments of EP political experts. While expert surveys now have a well-established track record in political science for estimating the policy positions of political actors (see for example Benoit and Laver, 2006; Castles and Mair, 1984; Huber and Inglehart, 1995; Laver and Hunt, 1992), none has ever been used thus far to estimate directly the policy positions of party groups in the EP. Expert surveys as a research tool are often chosen for their economy: frequently a survey of experts represents the quickest and least expensive way to gather data on party positions. Given their relative ease of set-up, it is a fairly simple matter to survey experts at any given time point, without the fixed costs of a huge datagathering project, detailed document coding, time-consuming interviews or costly opinion surveys.

Besides the practical virtue of economy, expert surveys also have several compelling substantive advantages in the context of measuring the policy positions of European party groups. A first compelling advantage comes from the explicitly a priori approach to locating policy positions of the expert survey strategy. The underlying assumption is that the key substantive policy dimensions in the EP can be identified in advance of the location of party groups, based on substantive expert understanding of potentially salient EU policy issues. The unknowns which experts are then asked to estimate are the locations of each party group on these a priori dimensions. The estimates of party group positions are then taken to be the statistically aggregated judgment of the experts on each predefined dimension. Unlike factor analytic scorings, constructed scales, or locations in a purely inductive space from multidimension scaling analyses, expert survey summaries eliminate the need for subjective and often ad hoc a posteriori interpretation of results in terms of substantive policy scales.

A second reason for using expert surveys relates to their desirable statistical properties, namely the property that according to well-understood statistical rules we can represent the uncertainty about our estimates of party group positions, based on both the fundamental variability of party positions as measured by differences in expert judgments, as well as the estimation variability that is determined by sample size.

Finally, especially with regard to a rapidly evolving political institution as the EP, we regard experts as the single best source of political information on European party groups. This is because the behavioural benchmarks which might provide observable implications of party group policy positions are not only incompletely understood, but also constantly changing as the roles, powers, size and composition of the EP evolve. As we argue later in this article for instance, there is a potentially complex dynamic which maps policy preferences of constituent national parties into policy preferences of EP party groups, and it cannot be assumed that the latter can always be represented by the mean of the former. Manifesto texts and roll-call votes may suffer from similar problems. Indeed, when trying to resolve which method of estimating party positions is best, we typically fall back on the

expertise and wisdom of political experts. By extension, then, we see systematic collection of judgments of political experts on party locations as the best way to harvest this wisdom systematically, which will take into account all relevant information about a party group's position, including voting behaviour, political speeches, debates, expressed opinions of party leaders, and so on. Even though experts will vary in their judgments, we can combine and summarize these judgments as a substantive indication of a party's likely set of policy locations. In short, our best estimate of European party group positions on policy resides in the collective wisdom of EU experts, available through systematically collected and summarized expert judgments.

An Expert Survey of EP Party Group Policy Positions

Our survey of experts was conducted from April to June 2004, at the time of the historic expansion eastward of the EU to include 10 new member states and just before the June 2004 elections to the newly expanded EP. Our expert survey solicited 36 experts on the EU and the EP drawn from professional directories and citation indices. These experts were largely academic specialists drawn from 32 different institutions in 12 different countries but also included a handful of EP researchers who have published on the topic. Our survey system used individually sent, English-language e-mail solicitations containing a unique URL linking the respondent's solicitation e-mail to our online survey questionnaire website. The questionnaire itself was an interactive, online system linked to a database server which recorded respondent answers. Following an initial solicitation round on 26 April, we sent a second round of request four weeks later to experts who had not yet responded. A total of 14 respondents completed questionnaires in the first round and 10 more in the follow-up round, for a total of 24 respondents and an overall response rate of 67 percent.

As in the Laver-Hunt survey, EP specialists were asked to use their best judgment to locate party groups on substantive policy dimensions. The party groups were the seven political groups existing in the EP at the time of the survey. These groups and their abbreviations are listed in Table 1, along with their share of the total EP seats. For some of the results reported below, we weight by seat share. While previous presentations of expert survey results (e.g. Benoit and Laver, forthcoming; Laver and Hunt, 1992) have been weighted by vote share, the national-based, decentralized nature of EP elections makes computing the vote share of EP party groups a complicated exercise. We have thus opted to represent EP party group size by seat share instead. Given that some variant of proportional representation (PR) is used in all member states, we do not believe this method will significantly distort results.

Each policy dimension was titled in terms of its substantive content and anchored at each end by two short phrases setting out substantive policy

Table 1. Political party groups in European Parliament pre-2004 election

EP party group Label Seat % Seats

European People's Party EPP 37.5 294

Party of the European Socialists PES 29.5 232

European Liberal and Democrat Reform

Party ELDR 8.4 66

European United Left/Nordic Green Left GUE 7 55

Greens Verts 6 47

Union for a Europe of the Nations UEN 3.8 30

Group for a Europe of Democracies and

Diversities EDD 2.2 17

Other Other 5.6 44

Total 100 785

Source: European Parliament official website (http://wwwdb.europarl.eu.int/ep5/owa/p_meps2. repartitiion).

The figures represent the standing just prior to the European Parliament Elections of 2004; the 785 member total is composed of the 626 members from the 15 pre-enlargement states and the temporary members representing the accession states since 1 May prior to the June elections.

positions. The survey provided respondents with a list of the European party groups (see Table 1, excluding 'Other') and asked respondents to locate each group on a 20-point scale for eight different substantive policy dimensions. Substantive policy dimensions covered in the survey included a set of four 'core' dimensions deployed in every country in the Benoit and Laver (2006) study. These were: increase spending versus reduce taxes; 'social' policy; and environmental policy. Also included, on the basis of advice from Parliament watchers, were policy dimensions dealing with, among other matters: immigration, deregulation, EU authority, EU collective security policy and EU federalism. For instance, the question on economic deregulation presented a scale anchored by two opposing endpoints as: (1) Favours high levels of regulation and control of the markets, such as telecommunications, versus (20) Favours deregulation at every opportunity. In a significant extension of the Laver-Hunt approach, we also asked experts to locate all parties on a general left-right dimension. A full list of the question wordings and dimensions is provided in Appendix A.

In addition to locating each politically significant party on each policy dimension, the questionnaire also asked respondents to indicate the relative importance of the issue to each party, again on a 20-point scale (excluding the general left-right dimension). This provides a position-independent measure of the salience of the issue for a particular party group, and may be used along with party group seat share to construct a measure of the overall political salience of a particular policy dimension.

Results: Policy Positioning in the EP

Left-Right Positioning

A full statistical summary of the results of the expert locations of the party groups on each policy dimension is presented in Table 2. The first row gives the mean score, followed by the standard error (SE), the standard deviation

Table 2. Policy positions of European party groups

Party abbreviation

Policy dimension GUE Verts PES ELDR EPP UEN EDD Divergence

Left-Right 3.6 5.1 7.4 11.8 12.6 16.5 17.1 3.8

0.51 0.36 0.30 0.43 0.39 0.58 0.49

2.4 1.7 1.5 2.1 1.9 2.5 2.1

23 23 24 24 24 19 18

Taxes v. Spending 4.9 5.7 6.4 13.1 13.4 13.9 14.1 4.0

1.01 0.72 0.35 0.73 0.60 1.06 1.28

4.4 3.1 1.6 3.3 2.7 4.0 5.1

19 19 20 20 20 14 16

EU Authority 9.5 7.3 6.3 7.5 7.4 17.7 18.9 3.0

0.82 1.00 0.54 0.64 0.63 0.50 0.32

3.8 4.7 2.6 3.1 3.0 2.3 1.4

22 22 23 23 23 21 20

EU Federalism 10.8 8.4 6.1 6.3 7.4 16.6 18.9 3.0

1.07 0.99 0.72 0.85 1.04 1.13 0.56

4.7 4.4 3.3 3.9 4.8 5.0 2.4

19 20 21 21 21 20 19

Deregulation 4.0 6.7 7.4 14.2 13.5 13.0 14.6 3.8

0.62 0.82 0.43 0.89 0.50 1.04 1.02

2.9 3.9 2.1 4.4 2.5 4.0 3.9

22 23 24 24 24 15 15

EU Collective Security 12.2 11.2 5.7 6.3 6.0 16.8 18.5 3.6

0.92 1.12 0.46 0.64 0.77 0.83 0.50

3.9 4.9 2.3 3.1 3.8 3.9 2.3

18 19 24 23 24 22 21

Immigration 6.1 6.5 7.3 7.4 12.0 17.5 17.5 3.4

0.75 1.22 0.63 0.73 0.70 0.50 0.40

3.1 5.3 2.9 3.4 3.3 2.2 1.7

17 19 22 21 22 19 17

Environment 6.9 2.9 8.6 10.9 12.1 12.8 12.9 2.8

0.89 0.83 0.50 0.96 0.55 0.82 0.91

4.2 3.9 2.4 4.6 2.6 3.4 3.8

22 22 23 23 23 17 18

Social 4.3 4.0 5.6 4.4 13.9 15.1 15.1 4.9

0.45 1.05 0.41 0.40 0.58 0.76 0.90

2.0 4.7 1.9 1.8 2.7 3.1 3.5

20 20 21 21 21 17 15

Mean, std. error, std. deviation, N. Dimensions following left-right are ranked by importance as per Table 3 below. Divergence is the weighted standard error of mean party positions on each issue.

and the number of respondents for each party on each dimension. The party groups are presented from left to right according to their mean values on the general left-right dimension, and are ranked following the left-right dimension by descending order of overall salience (see Table 3).

At the far left of the political spectrum is the GUE, with a mean value of 3.6 (SE 0.51), followed by the Greens (Verts) at 5.1 (SE 0.36). The three largest party groups, the PES, the European Liberal and Democrat Reform Party (ELDR) and the EPP, occupied positions on the left-of-centre, centre, and right-of-centre, respectively. The PES scored 7.4 (SE 0.30), the ELDR 11.8 (SE 0.43) and the EPP 12.6 (SE 0.39). On the further right appear the UEN, scoring 16.5 (SE 0.58), and the Group for a Europe of Democracies and Diversities (EDD) at 17.1 (SE 0.49).

We illustrate these positions graphically in Figure 1. Each point represents a party's left-right mean judgment, with the bars representing the 95 percent confidence interval. The left-right space appears to consist of four sets of party groups. First, the GUE and Verts appear on the far left, with a small degree of overlap in their 95 percent confidence intervals. Second, the PES inhabits the moderate centre-left. Third, two political groups, the ELDR and the EPP, form a grouping with a substantial degree of overlap just right-of-centre. Finally, the UEN and the EDD occupy the solid right position, their positions being statistically indistinguishable.

EDD UEN EPP ELDR PES Verts GUE

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Left-right dimension (1 = Left, 20 = Right) 95% confidence intervals

Figure 1. European party groups on the general left-right scale

Specific Policy Results

Looking further at Table 2, we see more specific policy results for the party groupings. On the Taxes versus Spending economic dimension of policy, the rankings are the same from left to right as for the general left-right policy dimension, although there is both more variation (higher standard errors) as well as higher overlap between confidence intervals. On Taxes versus Spending, two groupings of parties whose estimated positions overlap one another appear: the GUE, Verts and PES on the left and the ELDR, EPP, UEN and EDD on the right. (We explore this pattern further below.) This broad grouping into economic left and economic right is matched by the results on Deregulation - a result which we expect given the very high correlation between its mean scores and Taxes versus Spending (0.99).

On issues related to the authority and institutions of the EU, we observe a pattern different from that of the economic left-right. At the pro-integration end of the spectrum, we find the PES and the ELDR most strongly supporting a federal vision of the EU (mean 6.1 and 6.3, respectively), followed by the EPP (7.4) and the Greens (8.4). Positioned right at the centre of this issue is the GUE at 10.8. On the far right, preferring instead a union of nation-states, are the UEN and the EDD (16.6 and 18.9, respectively). On the EU Authority dimension, a similar grouping of pro-European parties emerges. At the left end of the scale favouring increasing the areas in which the EU may set policy, we find the PES, Greens, EPP and the ELDR being basically indistinguishable with mean scores ranging from 6.3 to 7.5. Once again the GUE are nearly at the centre at 9.5 (SE 0.82); the UEN and EDD are found at the Eurosceptic end favouring reducing the range of areas in which the EU may set policy (17.7 and 18.9, respectively). Similar results may be observed for the EU Collective Security dimension, with the PES once again emerging as the most strongly integrative and the EDD the most Eurosceptic.

Finally, we can see in Table 1 several non-economic dimensions of policy, such as Social and moral issues, Immigration and the Environment. On the classic dimension of social liberalism (measured by attitudes toward abortion, homosexuality and euthanasia), we see the EP divided into two distinct groups, one left-of-centre and one to the right. On the left of the spectrum, we find the Greens, GUE, ELDR and the PES, with mean scores ranging from 4.0 to 5.6, respectively. Closest to the centre, but still distinctively right-of-centre, is the EPP with a mean score of 13.9 (SE 0.58). Finally, the UEN and the EDD occupied their usual positions on the right both with mean scores at 15.1. Similar results were observed for Immigration, with these same sets of parties divided into two similar camps, with the EPP once again closest to the centre at 12.0 (SE 0.70).

Finally, on the issue of the Environment, party groups were generally less divided and more centrist. The exception was the Greens on the 'left', meaning that they favoured protecting the environment even at the cost of economic growth, with a mean score of 2.9. Left-of-centre were the GUE

(6.9) and the PES (8.6), followed by the ELDR in the centre at 10.9. The remaining parties EPP, UEN and EDD were slightly to the right-of-centre at essentially the same positions (12.1, 12.8 and 12.9, respectively).

As measured by the Divergence column, we can see that the policy dimensions where party positions were the most dispersed - indicating the policy dimensions with the highest degrees of policy differences between party groups - were first and foremost the classic economic and social left-right positions. Social policy had the highest position divergence, at 4.9, followed by Taxes versus Spending and Deregulation at 4.0 and 3.8, respectively. The three EU dimensions as well as Immigration ranged in the middle region of divergence between 3.0 and 3.6. Interestingly, the least contested policy dimension was the Environment, with a divergence score of just 2.8.

Summarizing the positional information graphically, we can visualize the political parties in two dimensions using a two-dimensional plot of economic left-right versus positions on EU integration, the two most salient policy dimensions of Taxes versus Spending and EU Federalism. Figure 2 portrays the main party groups in this two-dimensional policy space, with each point representing the position mean on the two dimensions. The dashed lines indicate the nearest regions to each party, showing the midpoint lines between each set of adjacent points. This nearest-neighbourhood division of the space is known as a Voronoi tessellation and has been used to represent party policy in Laver and Hunt (1992). Finally, the circles around each party group point are drawn proportional to the seat share of each group.

Figure 2 confirms what we observed earlier, namely that there are two broad camps of economic parties; GUE, Greens and PES on the left, and the ELDR, EPP, UEN and EDD to the right. On EU integration, we can see a grouping of the PES, ELDR, EPP and Greens on the pro-integration side, the GUE in the centre and the UEN and the EDD on the Eurosceptic side. In two dimensions, there appear to be three broad sets of parties, the PES, Greens and GUE on the left and pro-integration, the EPP and ELDR on the centre-right, pro-integration region, and the UEN and EDD in their own policy region of economic right and Euroscepticism.

The EP Issue Space

Relative Issue Salience

Expert respondents were also asked to indicate the relative importance of each policy dimension to each party group. Table 3 presents this information in the same format as in Table 2. In terms of overall importance, the economic (Taxes versus Spending and Deregulation) and EU issues (Federalism, Authority and Collective Security) were the most overall important, as indicated by their average across all parties, weighted by party seat share (scoring between 13.9 and 14.7). Immigration also ranked highly at 13.8. The Environment and Social Liberalism were ranked as the least

CO _ _

CD _ LO _ t _

c £2 — O

to ^ i_ _

to ^ -

-w o _

LH « -

CD —

LO —

CO —

CN —

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 13 14 1 5 1 6 1 7 1 8 1 9 20

Economic Policy

Figure 2. EP party group locations in two-dimensional space

Note: Dashed lines indicate policy neighbourhoods, circle size is proportional to seat share. EU Integration is EU Federalism dimension, Economic Policy is Taxes versus Spending dimension.

EDD •

• UEN

GUE Verts • EPP

• ""•'ELDR

important, at 13.2 and 12.6, respectively. Interestingly, these two dimensions were also the two that turned out to be (from Table 2) the most divisive (Social) and the least divisive (Environment).

A few expected results for individual parties stand out from Table 3, such as the very high importance attached to the environment for the Greens. Also, it is interesting that for the UEN and EDD - the two right, Eurosceptic party groups - the international issues were the most important, with other policy dimensions of only middling importance.

Components of Left and Right in the EP

The results summarized in Figure 2 seem to suggest that two broad dimensions of policy competition are present in the EP. The first represents the classic national policy issues associated with left and right, namely economic

Table 3. Salience of policy dimensions, by European party group

Party abbreviation Overall

Policy dimension GUE Verts PES ELDR EPP UEN EDD Mean SE

Taxes v. Spending 14.6 12.1 15.3 15.8 15.2 10.5 9.9 14.7 1.5

1.23 0.90 0.73 0.66 0.43 1.00 1.37

5.2 3.8 3.2 2.9 1.9 3.9 5.3

18 18 19 19 19 15 15

EU Authority 12.8 14.8 14.3 14.0 14.8 17.9 19.0 14.7 1.2

0.63 0.76 0.65 0.72 0.56 0.53 0.33

2.8 3.4 3.1 3.4 2.6 2.4 1.5

20 20 22 22 22 20 19

EU Federalism 11.4 13.1 14.6 14.6 14.9 17.5 18.7 14.6 1.4

0.76 0.84 0.49 0.65 0.78 0.67 0.58

3.2 3.7 2.2 2.9 3.5 2.9 2.4

18 19 20 20 20 19 18

Deregulation 15.3 14.1 14.5 15.9 14.9 11.3 11.6 14.6 1.0

0.90 0.62 0.57 0.56 0.40 1.18 1.44

4.2 2.9 2.7 2.7 1.9 4.4 5.6

22 22 23 23 23 14 15

EU Collective Security 12.0 11.3 14.0 11.4 14.7 16.8 17.3 13.9 1.6

0.94 0.86 0.62 0.80 0.66 0.63 0.87

3.9 3.6 3.0 3.8 3.2 2.9 3.9

17 18 23 22 23 21 20

Immigration 14.5 13.5 13.3 13.1 13.8 16.7 17.2 13.8 1.0

0.70 1.12 0.80 0.79 0.76 0.65 0.50

2.9 4.8 3.7 3.5 3.5 2.7 2.1

17 18 21 20 21 18 17

Environment 13.0 19.5 14.2 13.0 12.1 8.7 9.1 13.2 2.3

0.95 0.18 0.39 0.53 0.56 0.97 0.94

4.4 0.8 1.8 2.5 2.6 4.0 3.9

21 20 22 22 22 17 17

Social 11.7 14.8 12.5 13.9 12.6 10.1 9.9 12.6 1.1

0.86 1.15 0.78 0.86 0.85 1.14 1.34

3.9 4.9 3.5 3.8 3.8 4.6 5.2

20 18 20 20 20 16 15

Mean, std. error, std. deviation, N; ranked by importance. 'Overall' refers to mean (and std. error) importance of each dimension weighted by party vote share.

and social liberalism, as well as a bundle of relatively newer issues such as immigration and the environment. The second dimension relates to the authority and institutions of the EU itself. Substantively, the question is whether the EU policy space is unidimensional or rather consists of two or possibly more dimensions. The EP policy space has previously been described as unidimensional, with the traditional left-right or 'regulation' dominating (Kreppel and Tsebelis, 1999; Tsebelis and Garrett, 2000) or one-dimensional with geo-political pressures defining the principal axis of competition (Hoffman, 1966; Moravcsik, 1998). Other scholars, however, have described the European policy space as consisting of two dimensions:

a left-right dimension composed of economic and socio-political issues from the domestic arena and an orthogonal dimension of EU integration versus national sovereignty (Hix and Lord, 1997). Variations on the two-dimensional characterization relate to whether positions on EU integration are significantly correlated with left-right (e.g. Gabel and Hix, 2004; Hooghe and Marks, 2001) or whether positioning on the two dimensions is independent.

In Table 4 we have used principal components factor analysis to group and separate the constituent policy dimensional scorings into orthogonal factors. In order to explore the issue of what policy dimensions were grouped with left and right, we also included the general left-right dimension. Two factors clearly emerge (having eigenvalues well above 1.0), together explaining more than 77 percent of the variance in specific policy placements. The last panel in the table provides the varimax-rotated factor

Table 4. Principal components analysis of expert judgments of party positions

European Parliament

Cumulative

Factor Eigenvalue proportion

1 5.38 0.60

2 1.60 0.78

3 0.47 0.83

4 0.38 0.87

5 0.32 0.91

6 0.26 0.94

7 0.23 0.96

8 0.21 0.98

9 0.14 1.00

Factor

Left-Right 0.85 0.28

Taxes v. Spending 0.88 0.15

Deregulation 0.88 0.03

Environment 0.82 0.24

Social 0.73 0.42

Immigration* 0.70 0.55

EU Authority 0.28 0.86

EU Federalism 0.26 0.85

EU collective security 0.04 0.88

N 108

* Nationalism for post-communist countries.

loadings for the eight constituent policy dimensions plus the general dimension of left-right, with the higher loading for each of the two factors highlighted in bold type. The results clearly confirm our earlier interpretation: issue positioning in the EP clusters into two orthogonal component sets, one related to classic left and right issues from national party politics, and the second related clearly to EU issues. Moreover, the general left-right dimension loads very strongly with the first factor, which bundles together the classic socio-economic dimensions of national-level left-right policy. The second factor represents a purely EU dimension, with parties grouping themselves independently into pro-integration stances on the one hand and Eurosceptic positions on the other.1

In the context of previous findings, our results provide strong support for the two-dimensional model of policy competition, based on two orthogonal dimensions consisting on the one hand of classic issues of left-right socioeconomic policy and support for European integration on the other. These two latent factors, moreover, explain more than three-quarters of the variance in party positions on specific policy dimensions. Left-right is positively associated mainly with the first latent factor of socio-economic positioning, but also mildly positively associated with support for European integration. As suggested by Hooge and Marks (2001), we also found an association between socio-economic left-leaning policy and greater support for European integration, although our exploratory analysis would need further investigation and a more structured model before any firmer conclusions can be drawn.

So far, our analysis has looked only at the EP level, yet there are important differences between the policy space at the EU level and the level of national political parties. As we demonstrate in our final section, the EU policy space is not simply a direct mapping of national patterns of party competition from the domestic to the supranational level. In the next section we take a first look at the degree to which policy competition between domestic political parties is congruent with party competition and affiliation at the EP level.

EP Policy Platforms and National Member Parties

One emerging area which scholars of politics in the EU have only recently begun to explore relates to differences in policy competition at the national and EU levels. For instance, how much does agreement between policy positions explain which national parties join which EP party groupings? Do parties in national elections adopt policy positions as a result of policy influences from their EP party group? Or, conversely, do EP party group policy positions directly reflect the policy platforms of their national constituent parties? It is quite possible that the nature and direction of these influences will differ according to policy arena. We view research into this area of the Europeanization of policy among EU political parties as one of

the most promising areas for emerging research into the two overlapping arenas of European party competition.

our preliminary investigation of this issue compares the estimated policy positions of Ep party groups to the distribution of the policy positions of their constituent (member state) national-level political parties. Data on these national-level positions come from the left-right positions from the expert surveys reported in Benoit and Laver (2006). Figure 3 portrays the kernel density estimate of member state party positions for each EP party group, as well as the mean and confidence interval of the EP party group position on each issue. The graphs also indicate how many national member parties were included in each analysis (a full listing is provided in Appendix

The first obvious result is the clear correspondence between the mean EP group position on left-right and the central tendency of the national party left-right positions. By and large, the EP party groups' left-right positions neatly reflect the central tendencies of their constituent parties. In addition, as seen from the shape of their kernel densities, the member parties of these Ep groups clearly have similar national party positions on the left-right dimension, even though some groups include a small number of parties that are out of step with the central group position. It should be recalled that

EPP (42 parties)

ELDR (21 parties)

4 6 8 10 12 14 16 18 20 Left-Right Score

12 4 6 8 10121416 1820 Left-Right Score

PES (26 parties)

12 4 6 8 10 12 14 16 18 20 Left-Right Score

Verts (15 parties)

1 2 4 6 8 Left-

10 12 14 16 18 20 Right Score

GUE (13 parties)

12 4 6 8 10121416 1820 Left-Right Score

UEN (8 parties)

12 4 6 8 101214161 Left-Right Score

Figure 3. EP party group positions and the distribution of national member parties, left-right scores

Source: Benoit and Laver (2005) for national party left-right scores.

the EP and national country results were estimated in completely separate expert surveys - often administered in different languages. The clear correspondence of results suggests not only that expert surveys are an effective method for measuring the policy positions of EP party groups, but also augur well for the issue of the cross-national meaning and stability of expert placements on the general left-right dimension as used in our survey and in Benoit and Laver (2006).

Beyond the clear correspondence of positions, a few interesting individual results can be observed. Several party groups clearly contain national member parties whose positions are out of line with the central tendency of the other party group members. The EPP in particular appears to have adopted a more centrist position than the median position of its constituent national parties. While relatively centrist in the EP, the EPP is made up of mainly right-of-centre national parties. The UEN also contains several parties whose positions are clearly more centrist: the Polish Peasant Party (PSL) at 9.0 and the Lithuanian Union of Peasant and New Democracy Parties (VNDPS) at 7.45. We see similar minor outliers for the UEN, with Ireland's Fianna Fail (FF) and Estonia's Eestimaa Rahvaliit (RL) more centrist than the UEN, with means at 13.3 and 10.4, respectively. Among the Verts, we also see two parties whose positions are clearly more to the right than most other Verts members: the Spanish Partido Nacionalista Vasco (PNV) at 14.5 and the Latvian Green and Farmers' Union (ZZS) at 11.0.3 Another interesting result relates to the variance of member party positions. The distribution of ELDR member parties' positions, for instance, has a wide variance relative to more policy concentrated groups such as the PES, GUE or UEN. While the ELDR is a relatively centrist party grouping, it clearly includes both left-of-centre and right-of-centre national parties.

Overall, these findings suggest interesting patterns between EP policy group positioning and the policy positions of national-level member parties. In particular, it should suggest a note of caution for methods that automatically assume that EP party group positions can be inferred as the central tendency of constituent national party measures. Our preliminary look using left-right positions suggests that this will not always be the case, and further exploration among more specific policy dimensions (not shown here) suggests even greater divergence in national and EU positions. More systematic investigation is certainly warranted, especially taking into account the positions of parties from the new member states; i.e. positions that might not be reflected by the EP party group positions in the same way as the positions of parties from the EU-15. We expect this dynamic between national politics and EU politics to form one of the more interesting topics in the study of party competition for future research.

Discussion

A broad range of empirical and spatial analyses in political science depends on the specification of the policy positions of political parties, covering topics as diverse as coalition formation, political representation, macro-economic policy development and legislative decision-making. In this article we have provided the first measure of such policy positions in the EP using expert surveys. Given the changing political and institutional context of European party groups, we find the use of summaries of expert judgments

- systematic summaries of the collective wisdom of well-informed experts

- to offer compelling advantages over other methods, especially inductive or indirect measurements. While broadly consistent with the findings from recent placements based on roll-call analyses (Hix et al., 2005), for instance, our approach has the benefit of providing precise and direct numerical placements on well-specified a priori dimensions of policy that do not need to be subject to uncertain, inductive interpretation.

Our results indicate that on the two most salient dimensions (Taxes versus Spending and EU Federalism) there appear to be three broad sets of party blocs: the PES, Greens and GUE on the redistributive left and prointegrationist in character; the EPP and ELDR on the centre-right of the redistributive spectrum but broadly pro-integrationist; and finally the UEN and EDD in their own policy region on the economic right and distinctly Eurosceptic on the EU Federalism dimension.

Another central finding of this study is that the first dimension of policy space in the EP strongly bundles with the traditional left-right axis of European party systems, principally socio-economic in nature but also incorporating newer issues such as immigration and the environment. In addition, we found strong evidence of a second axis of policy competition, orthogonal to the first, consisting of positioning on EU integration.

Finally, in order to examine the relationship between national parties and their European party parent groups we compared the estimated policy positions of the EP party groups with those of their domestic affiliates and found that the EP political groups are generally placed at the centre of the distribution of domestic party policy positions. These intriguing preliminary results suggest a remarkably close correspondence between the EP groups and national political parties, and point to interesting possibilities for future research on the dynamic between national and European policy positioning.

Appendix A: Dimension Text Wording

Economic (Taxes versus Spending)

(At the national level) Prefers raising taxes to increase public services (1) (At the national level) Prefers cutting public services to cut taxes (20)

Deregulation

Favours high levels of regulation and control of the markets, such as telecommunications (1)

Favours deregulation at every opportunity (20)

Social

Favours liberal policies on matters such as homosexual law, abortion and euthanasia (1)

Opposes liberal policies on matters such as homosexual law, abortion and euthanasia (20)

Environment

Supports protection of the environment, even at the cost of economic growth (1) Supports economic growth, even at the cost of damage to the environment (20)

EU Authority

Favours increasing the range of areas in which the EU can set policy (1) Favours reducing the range of areas in which the EU can set policy (20)

Immigration

Favours policies designed to help asylum seekers and immigrants integrate into European society (1)

Favours policies designed to restrict access of asylum seekers and immigrants to Europe (20)

EU Federalism

Promotes a federal vision for the EU (1)

Promotes a Europe of nation-states (Europe des Patries) (20)

EU Collective Security

Favours a common defence and security policy for member states (1) Opposes development of common defence and security policy (20)

Left-Right

Please locate each political group on a general left-right dimension, taking all aspects of group policy into account. Left (1). Right (20).

Appendix B. Party group membership of national parties (June 2004)

EP party group Country Party label Party name Total MEPS

EDD FR cpnt Chasse, Pêche, Nature, Traditions 5*

EDD FR RPF Rassemblement pour la France 3

EDD NL CU ChristenUnie 3

EDD NL SGP Staatkundig Gereformeerde Partij 3

EDD UK UKIP UK Independence Party 3*

ELDR BE VLD Flemish Liberals and Democrats 3

ELDR BE Other 1*

ELDR CY DIKO Dimokratikon Komma 1

ELDR CZ Independent 1*

ELDR DK RV Radikale Venstre 1

ELDR DK V Venstre, Danmarks liberale parti 5

ELDR EE Kesk Eesti Keskerakond 1

ELDR EE Ref Eesti Reformierakond 1

ELDR ES Others 2*

ELDR FI KESK Suomen Keskusta 4

ELDR FI SFP Svenska Folkepartiet i Finland 1

ELDR FR UDF Union pour la démocratie française 1

ELDR HU SZDSZ Alliance of Free Democrats 2

ELDR IE Independent 1*

ELDR it It.Val. Lista di Pietro Italia dei Valori 2

ELDR it Others 2*

ELDR it I Democratici 4*

ELDR lt LDP Liberal Democratic Party 1

ELDR lt LiCS Union of Liberals and Center 2

ELDR lt NS/SL New Union-Social Liberals 2

ELDR LU DP Democratic Party 1

ELDR NL D66 Democraten 66 2

ELDR NL VVD Volkspartij voor Vrijheid en Democratie 6

ELDR SE C Centerpartiet 1

ELDR SE FP Folkpartiet Liberalerna 3

ELDR SI LDS Liberalna Demokracija Slovenije 3

ELDR SK ANO New Civic Alliance 1

ELDR UK LD Liberal Democrats 11

GUE CY AKEL Anorthotikon Komma Ergazemenou Laou 2

GUE CZ KSCM Communist Party of Bohemia and Moravia 3

GUE DE PDS Party of Democratic Socialism 6

GUE DE Independent 1*

GUE DK SF Socialistisk Folkeparti 1

GUE DK Den frie Socialdemokrat 1*

GUE DK Folkebevœgelsen 1*

GUE ES IU Izquerda Unida 4

GUE FI VAS Vasemmistoliitto 1

GUE FR PCF Parti Communiste Français 3

GUE FR Others 12*

GUE GR KKE Kommunistiko Koma Ellados 3

GUE GR Others 4*

GUE it PDCI Partito dei Comunisti Italiani 2

Appendix B. Continued

EP party Party Total

group Country label Party name MEPS

GUE IT RC Rifondazione Comunista 4

GUE NL SP Socialistische Partij 1

GUE PT PCP Portuguese Communist Party 2

GUE SE V Vänsterpartiet 3

GUE SK KSS Slovak Communist Party 1

EPP AT OVP Austrian People's Party 7

EPP BE CD&V Christian Democratic & Flemish 2

EPP BE CDH Humanist Democratic Centre 1

EPP BE Others 2*

EPP CY DISI Dimokratikos Sinagermos 2

EPP CZ KDU Christian and Democratic Union-Czechoslovak People 3

EPP CZ ODS Civic Democratic Party 8

EPP CZ US Freedom Union-Democratic Union 1

EPP CZ Independent 1*

EPP DE CDU/C Christian Democratic Union/Christian Social Union 53

EPP DK KF Konservative Folkeparti 1

EPP EE Isam Erakond Isamaaliit 1

EPP EE ResP Ühendus Vabariigi Eest - Res Publica 1

EPP ES PP Partido Popular 27

EPP ES Other 1*

EPP FI KD Suomen Kristillisdemokraatit 1

EPP FI KOK Kansallinen Kokoomus 4

EPP FR UDF Union pour la Democratie Française 7

EPP FR UMP Union pour un Mouvement Populaire 14*

EPP GR ND Nea Dimokratia 9

EPP HU FIDESZ Fidesz - Hungarian Civic Party 9

EPP HU MDF Hungarian Democratic Forum 3

EPP IE FG Fine Gael 4

EPP IE Independent 1*

EPP IT FI Forza Italia 23

EPP IT UDC Unione di Centro 4

EPP IT Others 7*

EPP LT LKD Lithuanian Christian Democrats 1

EPP LT TS Homeland Union 1

EPP LT VNDPS Union of Peasant and New Democracy Parties 1

EPP LU CSV Christian Social People's Party 2

EPP LV JL New Era 2

EPP LV LPP Latvia's First Party 1

EPP LV TP People's Party 2

EPP MT NP Nationalist Party 2

EPP NI UUP Ulster Unionist Party 1

EPP NL CDA Christen Democratisch Appeal 9

EPP PL PO Citizens' Platform 5

EPP PL PSL Polish Peasant Party 5

EPP PL Others 3*

Appendix B. Continued

EP party Party Total

group Country label Party name MEPS

EPP pt PSD Social Democratic Party 9

EPP SE KD Kristdemokraterna 2

EPP SE m Moderata Samlingspartiet 5

EPP SI NSi Nova Slovenija Krscanski Ljudska Stranka 1

EPP SI SDS Socialdemokratska Stranka Slovenije 1

EPP SI SLS Slovenska Ljudska Stranka 1

EPP SK KDH Christian Democratic Movement 2

EPP SK SKDU Slovak Democratic and Christian Union 3

EPP SK smk Party of the Hungarian Coalition 2

EPP UK Con Conservative Party 36

PES at SPO Austrian Social Democratic Party 6

PES BE PS Socialist Party 5

PES BE Other 1*

PES CY EDEK Kinima Sosialdimokraton EDEK 1

PES CZ CSSD Czech Social Democratic Party 7

PES DE SPD Social Democratic Party of Germany 35

PES DK SD Socialdemokratiet i Danmark 2

PES EE Sotsiaaldemokraatlik Erakond 1*

PES ES PSOE Partido Socialista Obrero Español 23

PES ES Other 1*

PES FI SDP Suomen Sosialidemokraattinen Puolue 3

PES FR PS Parti Socialiste 17

PES FR Other 1*

PES GR PASOK Panellinio Sosialistiko Kinima 9

PES HU mszp Hungarian Socialist Party 10

PES IE LB Labour 1

PES it DS Democratici di Sinistra 15

PES it SDI Socialisti Democratici Italiani 1

PES lt LSDP Lithuanian Social Democratic Party 5

PES LU LSAP Luxembourg Socialist Workers' Party 2

PES LV tsp People's Harmony Party 1

PES mt mlp Malta Labour Party 2

PES NI SDLP Social Democratic and Labour Party 1

PES NL PvdA Partij van de Arbeid 6

PES PL SLD Alliance of Democratic Left 22

PES PL UP Labour Union 2

PES PL Socjaldemocracja Polska 3*

PES pt PS Socialist Party 12

PES SE SAP Socialdemokratiska Arbetarepartiet 6

PES SI ZLSD Zdruzena Lista Socialnih Demokratov 1

PES SK Smer Party Direction - Third Way 2

PES UK Lab Labour Party 28

UEN DK DF Dansk Folkeparti 1

UEN EE RL Eestimaa Rahvaliit 1

UEN FR RPF Rassemblement pour la France 2

UEN FR ump Union pour un Mouvement Populaire 2*

UEN IE FF Fianna Fáil 6

Appendix B. Continued

EP party group Country Party label Party name Total MEPS

UEN IT AN Alleanza Nazionale 9

UEN IT Patti Patti Segni 1*

UEN LV TB/LNNK Alliance Fatherland and Freedom-LNNK 1

UEN PL PiS Law and Justice 4

UEN PT CDS/PP People's Party 2

UEN SK LU People's Union 1

Verts AT Gru The Greens 2

Verts BE Eco Ecolo 3

Verts BE Gro! Groen! 1

Verts BE Others 2*

Verts DE GRU Green Party 4

Verts ES PNV Partido Nacionalista Vasco 1

Verts ES Others 4*

Verts FI VIHR Vihrea Liitto 1

Verts FI Independent 1*

Verts FR V Les Verts 9

Verts IE GR Greens 2

Verts IT Green Federazione dei Verdi 2

Verts LU G The Green 1

Verts LV PCTVL For Human Rights in a United Latvia 1

Verts LV ZZS Green and Farmers' Union 1

Verts NL GL Groen Links 4

Verts SE MP Miljopartiet de Grona 2

Verts UK PCy Plaid Cymru 2

Verts UK SNP Scottish National Party 2

Verts UK UKGre Green Party 2*

Indep AT FPO Freedom Party of Austria 3

Indep AT No Affiliation 3*

Indep BE VB Flemish Block 2

Indep BE VLD Vlaamse liberalen en democraten 1

Indep ES EH Euskal Herritarrok 1*

Indep FR FN Front National 5

Indep FR MPF Mouvement pour la France 3

Indep FR No Affiliation 2*

Indep IT LDE Liberali Democratici Europei 1*

Indep IT LN Lega Nord 3

Indep IT Pann Lista Pannella Bonino 7

Indep NI DUP Democratic Unionist Party 1

Indep PL LPR League of Polish Families 3

Indep PL S Self Defence of the Polish Republic 4

Indep PL Others 3*

Indep SK HZDS Movement for a Democratic Slovakia 2

Total 785

Party group affiliations are taken from June 2004 (see Table 1 note). Parties marked with an asterisk are those for which the Benoit-Laver survey did not measure, or (in the case of France) did not directly measure, left-right policy. All non-asterisked parties are included in Figure 3.

Support for this project was provided by the Institute for International Integration

Studies, Trinity College, Dublin and by the European Commission Fifth Framework

(project number SERD-2002-00061). We thank Marina McGale and Alex Baturo

for research support.

1 We have tested the robustness of these results in a variety of additional ways, such as excluding the left-right dimension from the factor analysis, and then regressing the left-right placements on the factor scorings. These results (not shown) strongly confirm those in Table 4.

2 The EDD was excluded from the analysis as we had only three expert surveys at the national level for their constituent parties.

3 Interestingly, in the following Parliament (the 6th), the Partido Nacionalista Vasco joined the Liberal group.

References

Arregui, J., F. Stokman and R. Thomson (2004) 'Bargaining in the European Union and Shifts in Actors' Policy Positions', European Union Politics 5: 47-72.

Benoit, K. and M. Laver (2006) Party Policy in Modern Democracies. London: Routledge.

Castles, F. M. and P. Mair (1984) 'Left-Right Political Scales: Some "Expert" Judgements', European Journal of Political Research 12: 73-88.

Gabel, M. and S. Hix (2004) 'Defining the EU Political Space: an Empirical Study of the European Election Manifestos', in Gary Marks and M. Steenbergen (eds) European Integration and Political Conflict. Cambridge: Cambridge University Press.

Hix, S. and C. Lord (1997) Political Parties in the European Union. Basingstoke: Macmillan.

Hix, S., A. Noury and G. Roland (2005) 'Power to the Parties: Cohesion and Competition in the European Parliament, 1979-2001', British Journal of Political Science 35: 209-34.

Hoffman, S. (1966) 'Obstinate or Obsolete? The Fate of the Nation State and the Case of Western Europe', Daedalus 95: 862-915.

Hooghe, L. and G. Marks (2001) Multi-Level Governance and European Integration. Lanham, MD: Rowman and Littlefield.

Inglehart, R. and J. Huber (1995) 'Expert Interpretations of Party Space and Party Locations', Party Politics 1: 73-112.

Kreppel, A. (2002) The European Parliament and the Supranational Party System: A Study in Institutional Development. Cambridge: Cambridge University Press.

Kreppel, A. and G. Tsebelis (1999) 'Coalition Formation in the European Parliament', Comparative Political Studies 32: 933-66.

Laver, M. and B. Hunt (1992) Policy and Party Competition. New York: Routledge.

Moravcsik, A. (1998) The Choice for Europe, Social Purpose and State Power from Messina to Maastricht. Ithaca, NY: Cornell University Press.

Noury, Abdul (2002) 'Ideology, Nationality and Euro-Parliamentarians', European Union Politics 3: 33-58.

Raunio, T. (1997) The European Perspective: Transnational Party Groups in the 1989-1994 European Parliament. Aldershot: Ashgate.

Stokman, F. and R. Thomson (2004) 'Winners and Losers in the European Union', European Union Politics 5: 5-23.

Thomassen, J. and H. Schmidt (1997) 'Policy Representation', European Journal of Political Research 32: 165-84.

Thomassen, J., A. Noury and E. Voeten (2004) 'Political Competition in the European Parliament: Evidence from Roll Calls and Survey Analysis', in Gary Marks and M. Steenbergen (eds) European Integration and Political Conflict. Cambridge: Cambridge University Press.

Tsebelis, G. and G. Garrett (2000) 'Legislative Politics in the European Union', European Union Politics 1: 9-36.

GAIL McELROY is a Lecturer in the Department of Political Science at Trinity College, Dublin. Her current work focuses on the dynamics of party competition in the European Union. She has recently published on legislative organization and legislative politics in the EU.

KENNETH BENOIT is an Associate Professor of Political Science at Trinity College, Dublin. He is author of a number of journal articles and book chapters on electoral systems and the effects of campaign finance. His current work focuses on surveys of expert judgments, computerized content analysis and the quantitative analysis of political party competition.

ADDRESS: Department of Political Science, Trinity College, Dublin 2, Ireland. [email: mcelroy@tcd.ie and kbenoit@tcd.ie]

Paper submitted 11 March 2005; accepted for publication 19 June 2005.