Portal de Periódicos da CAPES

The European Parliament Elections of June 2004: Still Second-Order?

2005; Taylor & Francis; Volume: 28; Issue: 3 Linguagem: Inglês

10.1080/01402380500085962

1743-9655

Hermann Schmitt,

Populism, Right-Wing Movements

Abstract A quarter of a century ago the first series of European Parliament elections were characterised as second-order national elections. Much has changed since, which might have had an impact upon this diagnosis. In this article the central assumptions and predictions of the second-order elections model are restated and evaluated against the outcome of the 2004 European Parliament election and a post-election survey. Surprisingly enough, the findings confirm the persisting second-order nature of EP elections for Western Europe. Matters look very different, however, in the eight new Central and East European member countries. Acknowledgements This paper profited greatly from questions and comments by Bruno Cautrès, John Curtis, André Freire, Thomas Gschwend, Pedro Magalhaes, Renato Mannheimer, Franz U. Pappi, Andrea Römmele, Paolo Segatti, and Andreas M. Wüst, which is all gratefully acknowledged. I would also like to thank the director of European Omnibus Survey (EOS) Gallup Europe, M. Pascal Chelala, who for the present analysis provided early access to the data of an EOS post-election survey (EB Flash 162). Obviously, remaining errors and shortcomings in what follows are mine alone. Notes On multi-level governance see e.g. König et al.(1996 König T Rieger E Schmitt H eds 1996 Das Europäische Mehrebenensystem Frankfurt: Campus [Google Scholar]); Kohler-Koch and Eising (1999 Kohler-Koch B Eising R eds 1999 The Transformation of Governance in the European Union London: Routledge [Crossref] , [Google Scholar]); Hooghe and Marks (2001 Hooghe L Marks G 2001 Multilevel Governance and European Integration Lanham, MD: Rowman & Littlefield [Google Scholar]). We distinguish two sources of government party losses in European Parliament elections. One is vote-switching: some first-order government voters will desert and vote for one of the opposition parties. The other is differential mobilisation: first-order government voters are expected to abstain in greater numbers than first-order opposition voters. We know from the 1999 European Elections Study that differential mobilisation is the stronger source of government parties' losses: many more first-order government voters (41 per cent on average) than opposition voters (29 per cent on average) abstain (see Marsh 2005 Marsh M 2005 The Results of the 2004 European Parliament Elections and the Second-order Model in O. Niedermayer and H. Schmitt (eds.), Die Europawahl 2004, Wiesbaden: VS-Verlag, (forthcoming) [Google Scholar]: Table 3.2). There are at least two different explanations of this, one economic and one political. The economic variant uses the analogy of the business cycle and proposes that governments tend to deal with unpopular legislation (which tends to harm the interests of many voters) early in the period and come up with all sorts of electoral gifts towards the end of it, when the next election draws close (e.g. Kirchgässner 1986 Kirchgässner, G. 1986. Economic Conditions and the Popularity of West German Parties: A Survey. European Journal of Political Research, 14: 421–39. [Crossref], [Web of Science ®] , [Google Scholar]). The political variant emphasises the evolution of electoral mobilisation which reaches a climax at election time only in order to melt down thereafter (e.g. Stimson 1976 Stimson, JA. 1976. Public Support for American Presidents: A Cyclical Model. Public Opinion Quarterly, 40: 1–21. [Crossref], [Web of Science ®] , [Google Scholar]). The European Community of 1979 had already experienced a first enlargement when Denmark, the UK and Ireland joined the original six. On the consolidation of East European party systems and voter alignments see Hofferbert (1998 Hofferbert R ed 1998 Parties and Democracy. Party Structure and Party Performance in Old and New Democracies Oxford: Blackwell Publishers [Google Scholar]); Birch (2001 Birch S 2001 Electoral Systems and Party System Stability in Post-Communist Europe paper prepared for presentation at the 97th Annual Meeting of the American Political Science Association, San Francisco, 30 August–2 September [Google Scholar]); also Mainwaring and Scully (1995 Mainwaring S Scully T 1995 Building Democratic Institutions: Party Systems in Latin America Stanford: Stanford University Press [Google Scholar]) on what they call the 'institutionalisation' of Latin American parties and party systems. Here is the cause of a relatively recent but probably accelerating legitimacy problem on the side of national parliaments. The electoral process that brings them into office hardly recognises – let alone discusses and problematises – that there are limits to the authority of national parliaments of EU member countries (and to the problem-solving capacity of national politics more generally). See Steunenberg and Thomassen (2002 Steunenberg B Thomassen J eds 2002 The European Parliament: Moving Towards Democracy in the EU Lanham, MD: Rowman & Littlefield [Google Scholar]) for a study addressing this kind of questions. International trade regulations are determined by the council alone, without any formal involvement of the parliament. See Maurer (2005 Maurer A 2005 Das Europäische Parlament in O. Niedermayer and H. Schmitt (eds.), Die Europawahl 2004, Wiesbaden: VS-Verlag, (forthcoming) [Google Scholar]) for a recent account addressing these questions. France is the only country for which it is not obvious what the preceding first-order national election actually is. In late spring 2002, the French elected the President of the Republic in two rounds, with Jospin (from the left) defeated in the first round, and Girac (from the right) and Le Pen (extreme right) competing in the second round. Girac was elected with an overwhelming majority stretching from the far left to the (moderate) right. Legislative elections were held shortly thereafter, in which the presidential camp secured a triumphant victory in a sort of honeymoon period after the presidential result. While this essentially characterises this legislative election as 'second-order' – the power question had been answered before – it is still the most logical election with which to compare the European result. For merely technical reasons of the electoral rule applied, it would be odd to compare the distribution of votes between two very uneven candidates in the second round of the presidential election with the European Parliament result. And, after all, the result of the legislative election did only underscore what the true result of the preceding presidential election was (see for more detail Perrineau and Ysmal 2003 Perrineau P Ysmal C eds 2003 Le vote de tous les refus. Les élections présidentielles et législatives de 2002 Paris: Presses de Sciences Po [Google Scholar]; Cautrès and Mayer 2004 Cautrès B Mayer N eds 2004 Le nouveau désordre electoral Paris: Presses de Sciences Po [Crossref] , [Google Scholar]). See Franklin (2001 Franklin, MN. 2001. How Structural Factors Cause Turnout Variations at European Parliament Elections. European Union Politics, 2: 309–28. [Crossref] , [Google Scholar]) for a diachronic analysis along the same lines which covers the elections from 1979 to 1999. While both analyses are based on the same dependent variable, he uses a somewhat different list of predictors. Compulsory voting, last FOE participation, concurrent FOE, FOE cycle, and Sunday voting are identical. The difference is that Franklin introduces an additional dummy variable – whether a particular EP election is the first in a country. We leave this variable out because it is almost identical (exceptions are Cyprus and Malta) and thus highly collinear with our alternative and more powerful predictor 'post-communist member country'. Because self-reported electoral participation chronically suffers from over-reporting, we have recoded the variable in such a way that all 'don't knows' and 'no answers' are considered as non-voters. Note that a standardised logistic regression coefficient is reported which is the odds ratio minus '1' (so that a positive or negative effect has its origin at '0') multiplied by the standard deviation of this particular variable. The advantage of this procedure is that effect sizes can be compared across variables and countries. What this procedure does not do is to produce a fixed range within which the effects could possibly vary; however, our analysis suggests that there is a kind of a ceiling at + / − 3. See for similar strategies for establishing comparability between logistic regression coefficients Kreppel (2002 Kreppel A 2002 The European Parliament and the Supra-National Party System Cambridge: Cambridge University Press [Google Scholar]) and Faas and Rattinger (2004 Faas T Rattinger H 2004 Drei Umfragen, ein Ergebnis? Ergebnisse von Offline- und Online-Umfragen anlässlich der Bundestagswahl 2002 im Vergleich in F. Brettschneider, E. Roller and J. van Deth (eds.), Die Bundestagswahl 2002: Analysen der Wahlergebnisse und des Wahlkampfes, Opladen: Leske + Budrich 277 99 [Crossref] , [Google Scholar]). Among the consolidated party systems, a significant effect of EU scepticism on electoral participation survives all statistical controls only for the German and the Spanish sample. For the German case, this is corroborated by the findings of the European Election Study 2004, which shows that (a) there is a minor but significant effect of EU attitudes on the vote, and (b) that EU attitudes of Germans are chiefly determined by the perspective on Eastwards enlargement of the Union (Schmitt 2005 Schmitt H 2005b Die Wahlbeteiligung der Deutschen bei der Europawahl 2004 in O. Niedermayer and H. Schmitt (eds.), Die Europawahl 2004, Wiesbaden: VS Verlag, (forthcoming) [Google Scholar]b). There could be other reasons why national government parties lose in European Parliament elections, among them the performance of these governments in European Union politics. An alternative hypothesis would therefore state that governments should lose support if the voters are unhappy about how their country is represented in the EU, and vice versa. If we accept the standard Eurobarometer question 'EU membership of [country] – good thing, bad thing, or neither?' as an indicator for popular satisfaction with government performance in EU politics, we can test this hypothesis. The result is discouraging. When we regress government losses on membership approval rates, we find insignificant coefficients. However, while the 'membership' variable is obviously unrelated to government losses, it seems to co-vary substantially, itself, with the first-order electoral cycle. This suggests that there is a relation of EU approval with the national electoral cycle, the nature of which we will explore in future work. Note that Britain and France are the only two member countries which apply some variant of the single-member majority system for the election of national deputies, while members of the European Parliament are elected according to a variant of the PR system – as everywhere else in the European Union. For a Union-wide comparison of the electoral systems applied in the 2004 European Parliament election see Nohlen (2004 Nohlen D 2004a Wahlrecht und Parteiensystem 4th edn, Opladen: Leske + Budrich [Crossref] , [Google Scholar]a) and, with some greater detail, Wüst and Stöver (2005 Wüst A Stöver P 2005 Anhang 2: Die Wahlsysteme einzelner Länder zur Wahl des Europäischen Parlaments in O. Niedermayer and H. Schmitt (eds.), Die Europawahl 2004, Wiesbaden: VS Verlag, (forthcoming) [Google Scholar]). Again, the better showing of small parties in the West could be due to other reasons for strategic voting which have not much to do with the second-order nature of European Parliament elections. Due to the fact that the electoral rules applied differ – in some cases considerably – between first-order national and European Parliament elections, the details of the electoral rules applied are particularly doubtful in that regard. In order to control for this potential influence on the electoral fortunes of parties of different sizes, we have determined the difference in constituency magnitude between European Parliament and first-order national elections (based on Nohlen 2004 Nohlen, D. 2004b. Wie wählt Europa? Das polymorphe Wahlsystem zum Europäischen Parlament. Aus Politik und Zeitgeschichte, B17: 29–37. [Google Scholar]b and Wüst and Stöver 2005 Wüst A Stöver P 2005 Anhang 2: Die Wahlsysteme einzelner Länder zur Wahl des Europäischen Parlaments in O. Niedermayer and H. Schmitt (eds.), Die Europawahl 2004, Wiesbaden: VS Verlag, (forthcoming) [Google Scholar]). It turned out that this difference does not have a significant effect on the relative performance of small parties (more precisely, on the difference in the 'effective number of parties' index between the European Parliament election and the preceding first-order election). This must not be read as a refutation of established knowledge in the electoral systems literature (e.g. Cox 1998 Cox G 1998 Making Votes Count Cambridge: Cambridge University Press [Google Scholar]). The effects originating in the different electoral rules applied could just be comparatively modest compared to the effects that originate in the different nature of first- and second-order elections. Note that our understanding of volatility considers voters to switch if they vote in two consecutive elections for parties with different names, no matter whether these parties are mergers of previously separate parties; compare for a different view Sikk (2001 Sikk A 2001 Stabilisation of Post-Communist Party Systems MA Thesis, University of Tartu, Tartu [Google Scholar]). The latter is of course relevant because it is the candidates of national parties that are campaigning for seats in the European Parliament election. But see Hix (2001 Hix, S. 2001. Legislative Behaviour and Party Competition in the European Parliament: An Application of Nominate to the EP. Journal of Common Market Studies, 39: 663–88. [Crossref], [Web of Science ®] , [Google Scholar]) for an analysis of parliamentary votes which shows that the left–right divide has become more predictive in the voting behaviour of MEPs over the last years, or in other words that the cohesion of group voting has been increasing. Things are moving, though, in the partisan direction. Not only are there 'partisan' caucuses ahead of summit meetings, which are organised by EU party federations and aim at coordinating the policies of council and parliament on a partisan basis (see Hix and Lord 1997 Hix S Lord C 1997 Political Parties in the European Union New York: Saint Martin's Press [Crossref] , [Google Scholar]). The draft treaty on an EU constitution adds the requirement that the president of the European Commission should be able to represent the majority of the European Parliament. I am thinking here of the fact that the current leaders of the two largest political groups in the European Parliament are German (Pöttering, EPP and Schulz, PES), and that neither of them played a significant role in the German EP election campaign. Angela Merkel, president of the CDU and head of the CDU/CSU group in the German Bundestag, and Gerhard Schröder, former chairman of the SPD and chancellor of the Federal Republic of Germany, did not leave them much room in the public debate. See Schmitt (2005 Schmitt H 2005c Political Linkage in the European Union in D. Farrell, P. Ignazi and A. Römmele (eds.), Political Linkage Revisited, New York: Praeger Publishers, (forthcoming) [Google Scholar]c), who argues that EU politics, while quite successful in terms of interest intermediation, is suffering from a defective structure of opinion formation.