Extra-ordinary in many senses.

The German Federal Election of September 2005 1

Hermann Schmitt & Andreas M. Wst MZES, University of Mannheim 68131 Mannheim Germany

[version of November 17, 2005]

1. Introduction

The German federal election of September 2005 was an extra-ordinary election in many senses. Its premature calling came as a surprise. And, to be sure, its result was a surprise as well not least with reference to the forecasts of German pollsters. They largely failed in predicting the outcome. While this is not really our concern here, we will at the end come back to this issue. This article will address the structural context of the election, it will shed some light on the evolution of public opinion during the legislature and the campaign, and discuss the result and its likely consequences. 2 Before we start doing this, however, we will address some more conceptual, or theoretical, issues that will guide us in defining the variables we are investigating in the present article.3 Note that at the time of writing4 it is too early to engage in analyses of individual-level data of the German Election Study 20055 in a hypothesis-testing manner. This has to wait a little longer, until the different data sets finally become available and analyses can be conducted. Here, we have to do with what we know in terms of individual level evidence. Based upon this, we will relate (more or

An earlier version of this paper was presented at a conference on the German Election of 2005 which was held in Athens, Greece, on October 24, 2005. This conference was organised by the Greek Association of Political Science and supported by the Friedrich-Ebert-Stiftung Athen and the Goethe-Institut Athen. 2 See Roth and Wst (2006) for another account of contextual factors that contributed to the outcome of the 2005 election. 3 All survey information presented in the current article has been gathered and made available by the Forschungsgruppe Wahlen in Mannheim. This is gratefully acknowledged. 4 This article was finished mid November 2005. 5 The German Election Study consists of more than one individual study. Four at least deserve mentioning, three at mass- and one at elite-level: (a) a study on the role of political parties, directed by Oskar Niedermayer and Bettina Westle; (b) a study on the evolution of public opinion and campaigning, directed by Rdiger Schmitt-Beck; (c) a study on the effects of a limited choice set, directed by Berhard Weels and Hermann Schmitt; and (d) a study into attitudes and behaviours of both direct and list candidates for the German Bundestag, directed by Thomas Gschwend, Hermann Schmitt, Andreas Wst and Thomas Zittel. At the time of writing, these data were only available in some preliminary form; the candidates survey was only started.

less) long-term trends on relevant variables to one another and draw our conclusions from it on the future of German electoral politics.

2. Theoretical considerations.

Institutions matter at least somewhat. This is what a series of empirical studies on the effect of institutional arrangements on political behaviour demonstrates (see e.g. the list of publications at www.cses.org). Within the German political system, variations in the institutional context of federal politics are of course limited. The constitution hardly changes, nor do the procedures of federal policy making or the competences of its actors. One might wonder why the institutional context should play a role at all here. This is due to one critical aspect: the power relation between the federal parliament Bundestag and the federal government which depends on it on one hand, and the house of representatives of state governments Bundesrat on the other. Every state parliament election (Landtagswahl) can change this power relation with the ultimate effect that a concordant majority in the two houses changes into a discordant majority, or vice versa. Concordant majorities usually allow the federal parliament to get its legislative projects through. Discordant majorities tend to cause problems because the Bundesrat has to pass two thirds of all federal laws (among them all that involve co-funding of the states).

The consequences of discordant majorities on federal politics are immense. Most obviously they affect the policy output, i.e. the legislation as such. But beyond that, discordant majorities in the two houses have also a moderating effect on the depth of ideological conflicts (see Graph 1 for an illustration of this). The reason is straightforward. When federal government and its constituent parties need the consent of the main party of the opposition to legislate, they cannot afford that deep ideological cleavages pull them apart. They know that, at the end of the day, they will need to compromise and finally agree on less than the preferred. Under those circumstances, consensus-seeking strategies and the ability to co-operate are more useful and hence prevalent than the persistence on competition and conflict (e.g. Lehmbruch 1998, Lijphart 1999). This is why party competition in the German federal system is less ideological than it sometimes appears to be.

Why is this at all relevant for an analysis of the German federal election of 2005? Because ideological conflict is one important source of partisanship (see e.g. Schmitt and Holmberg 1995, Schmitt 2002 and 2005), and partisanship is an important determinant of vote choices and certainly of stable vote choices (e.g. Campbell et al. 1960; Miller and Shanks 1996; Schmitt 2001a). This is all the more relevant in times when traditional social cleavages lost some of their stable impact on political attitudes and behaviours (e.g. Dalton et al., 1984; Franklin et al., 1992; Schmitt 2001b). 2

Graph 1

The Arguments
institutional arrangements ideological conflicts

party attachments

vote

social conflicts short term factors

When the stabilising cues from ideological, partisan and social divisions are weakening, short-term factors are gaining importance (e.g. Dalton 1996) and the timing of vote choices is gradually approaching the election (e.g., Norris 2004). Short-term factors originate in the issues that shape the campaign and in the candidates standing for election (Campbell et al., 1960). There are all sorts of candidates standing for election of course; what increasingly counts in a German parliamentary election are the contenders for the position of head of government (Kanzlerkandidaten; see Klein and Ohr 2000; Schmitt and Ohr 2000). Much the same holds for issues: there are all sorts of issues. The ones that count most in the voting decision are often not very divisive between party camps, i.e. socalled valence issues, and the mechanism that translate their perceived importance is the competence that a voter attributes to a candidate or a party to resolve the issue (Stokes 1966, 1992; Schmitt 1998; 2001a; Clarke et al., 2004).

These claims have been corroborated in a number of empirical analyses and we do not need to reinvestigate them here at the micro-level level. What we will rather do is confronting longitudinal empirical evidence on each of the factors in order to see whether the over-time trends covary. Because of vastly differing time frames, this can only be done in a somewhat impressionistic eye-balling manner. Imperfect as this methodology may appear, we nevertheless claim to support the arguments that German vote choices, due to institutional properties of the federal system of government and the decline of ideological and social cleavages, is increasingly dependent upon the issues that prevail during the election campaign and the image of chancellor candidates that is: short-term factors; that numerous vote choices are only made during the campaign; and that this poses increasing 3

problems on the capability of the polling industry to accurately predict the election result ahead of election day.

3. Ungovernability, ideological centrism, and the decline of core-electorates

The German political system is characterised by two rivalling principles of government: federalism and party government. The first suggests a consensual, the second a competitive mode of policy making (Lehmbruch 1998). Which of the two principles or modes of government prevails is depending upon the majorities in the house of representatives of state governments (the Bundesrat). There are two basic alternatives: either the federal government can rely on a majority of state governments (and their votes) of the same political colours, in which case competitive party government is the prevailing characteristic; or it can not, in which case consensual policy making, package deals, and more generally: politics of the smallest common denominator are the prevailing characteristic.

The evolution of majorities in the Bundesrat during the two terms of chancellor Gerhard Schrder (Schrder I and II) points to a growing trend towards consensual government in the Federal Republic of Germany (Graph 2). Displayed are secure Bundesrat votes for the federal government and the federal opposition (SPD and SPD-Green state governments votes added on the federal government side, and CDU/CSU or CDU-FDP state governments votes added on the federal opposition side).

Graph 2

Deadlock
votes
60 50

federal opposition
40 30 20 10

early election called

federal government
0 1 13 25 37 49 61 73 85
Source: www.bundesrat.de.

months since the commencement of Schrder I

Note: The graph displays the number of secure government and opposition votes in the Bundesrat. The votes of mixed land colaitions, e.g. the SPD-FDP-coalition in Rhineland-Palatinate, are not displayed. In controversial decisions, these Lnder usually abstain. After re-unification the federal states (Bundeslnder) together have 69 votes, 35 being the absolute majority.

Only during the first year of its first term could the federal SPD-Grne coalition government rely on a relative majority of Bundesrat votes. Never did it control an own absolute majority as it is required to pass or block federal legislation. The federal opposition however, during the whole period of the second Schrder government, could count on such an absolute majority and hence obstruct federal legislation at will. The decision to call an early election was taken after the results of the 2005 election of the Landtag of Northrhine-Wesphalia became known that is, at exactly the time when the share of Bundesrat votes of the federal government had declined to zero (0).

Discordant majorities in the two houses are a relatively recent phenomenon in the Federal Republic. They only emerged during the first period of SPD-led governments, between 1972 and 1983, and reappeared again ten years later, during the third and the fourth government of Helmut Kohl. Their potential effect on ideological polarisation should therefore only become manifest during the 1980s and 1990s. Thanks to the Politbarometer surveys of the Forschungsgruppe Wahlen, this is also the period for which we have perceptional data on where the German parties stand in ideological terms. Over the years, altogether ten representative samples of the German population of voting age where asked to place each of the relevant parties on an eleven-point left-right scale. The mean scores of these individual placements are depicted in the following graph (Graph 3).

Graph 3

Ideological Bandwidth
1978 1982
Grne

1986
SPD REP FDP CDU CSU

1990 1994 1998 2002 2006 1

left PDS Die Linke

10

11
right

Quelle: Forschungsgruppe Wahlen Politbarometer

Werte fr 1980-94 face-to-face erhoben, ab 1997 telefonisch

Source: Politbarometer of the Forschungsgruppe Wahlen

If we start from the right-hand side, and in the absence of a relevant far right party, the CSU goes first. Located less than a scale point apart towards the centre is the CDU. The small FDP occupies the middle-ground but still with a clear alignment from 1983 on with the right. The SPD places itself centre-left, and the Greens are more or less left of the SPD. PDS/Die Linke finds itself left of the Greens. In 2005, the two major parties the SPD on the left and the CDU on the right are 2,5 scale 5

points apart; back in 1980, the respective figure was still 3,9. Greens and CSU are now 3,1 scale points apart, as compared to 4,8 scale points in 1980 (and 6,6 points in 1987 when the perceived ideological distance between these parties was largest). Although the parties of the right CDU/CSU and FDP have reoriented themselves somewhat to the right after they had lost government in 1998, the bandwidth of German party competition was probably never narrower than it is at the time being.

One of the consequences of a shrinking ideological spectrum is the opening-up of niches in the electoral market open flanks in military jargon, to be overwhelmed with little effort. One example for this is the formation of a new left party, Die Linke, which integrates in addition to the East Germany-based PDS a sizeable left-wing break-away of the SPD.6 Systematically speaking, there is little reason why similar developments should not also be expected to happen on the right side of the political spectrum.7

Graph 4

Traditional Cleavage Structure

church

CSU CDU labour SPD FDP state

Quelle: Forschungsgruppe Wahlen Politbarometer

capital

Die Linke

Institutional constraints and ideological conflicts are important motivations for political behaviour. But is not social conflict the core of it all? What about the cleavage structure? Are core electorates, defined in social-structural terms, still relevant for our understanding of electoral outcomes? Before we

The most well-known of earlier SPD followers being the former SPD-chairmen and now head (with Gregor Gysi) of the parliamentary group of Die Linke, Oskar Lafontaine. 7 In particular in a situation in which the CDU/CSU is becoming the leading force of a grand coalition which will pull the two parties even further towards the centre and leave the far-right fully unguarded. What could speak against such an expectation is that the extreme right in Germany, in the aftermath of the Nazi regime, has been stigmatised. However, these FRG reflexes to the fascist past might have weakened in re-united Germany where national pride has become more widespread again and acquired a positive connotation while Europe has become less of an auxiliary fatherland than it used to be.

turn to this, we briefly recapitulate the traditional cleavage structure of the German party system. This basic structure has been made up by two cross-cutting cleavages, state vs. church on the one hand, and labour vs. capital on the other (Graph 4).

Within this structure, the SPD used to stand for secular labour while the CDU, and even more so the CSU, used to represent religious business. The FDP has been the party of secular business, thereby balancing a rightist score on the business dimension with a leftist score on the religious dimension and henceforth occupying the middle ground. The new far-left formation Die Linke fits very well in this traditional cleavage structure by being more social-democratic than the original SPD. The same does not apply to Die Grnen, however: for them, a third dimension the new politics dimension was needed to adequately represent their political claims and appeals (e.g. Brklin 1984; Schmitt 1987); would we still classify them within the traditional cleavage system, they would position themselves somewhere between SPD and Die Linke.

Decades of social change have rendered this approach largely inadequate for our understanding of contemporary voting behaviour. While practising catholics and unionised labourers continue to vote predominantly CDU/CSU and SPD, respectively, the size of these social groups has dramatically declined to about ten percent each (see Table 1).

Table 1 Core-Electorates (figures are percentages)

SPD voters unionized among workers election unionised in the SPD year workers electorate

CDU/CSU practising voters among catholics in practising the CDU/CSU catholics electorate

1969 1972 1976 1980 1983 1987 1990 1994 1998 2002 2005

68 62 68 64 58 60 66 55 54

25 19 13 14 19 18 16 13 11 11 9

84 74 78 70 79 74 70 73 72

46 42 36 31 25 29 23 15 15 12 12

Source: Forschungsgruppe Wahlen. West Germany only.

This is not to say that social divisions are now a less powerful predictor of the vote than they were in the early years of the Federal Republic. While the old cleavage characteristics have lost most of their 7

vote structuring power, new social inequalities have become associated with party preference and vote choice, among them age, education, and the regional context (Schmitt 2001b, tables 2 and 3). There is one difference though between new inequalities and old cleavages: the latter are organised which adds stability to their political meaning.

4. Partisanship

Rooted in social oppositions partisanship, once established, lives a life on its own. It is stabilised by the reaffirmation of recurring election campaigns and voting decisions (Converse 1969). It flourishes when ideological conflicts between the electoral contenders are profound, and it weakens in the absence of such conflicts (Schmitt and Holmberg 1995; Schmitt 2002, 2005). In addition to this conflict theory of partisanship, there are of course competing conceptions, most notably the theory that cognitively mobilised citizens would be less in need of partisan cues and thus more independent, not only functionally but also politically (e.g. Dalton 1984, Dalton and Wattenberg 2000). This view, however, could not be confirmed in systematic diachronic tests for six West European countries which compared the likelihood of partisan ties among educated and politically interested people with that same likelihood among other people (e.g. Berlund et al., 2005).

What we should expect, then, is a decline of partisanship in Germany mainly as a result of the declining ideological bandwidth of party competition which, to an important part, is caused by the increasing institutional deadlock between Bundestag and Bundesrat. Graph 5 shows the evolution of the proportion of independents from 1976 on. 8

In 2005, one in three West Germans says no when asked whether they lean towards one of the political parties.9 This proportion has roughly doubled over the past three decades, with the steepest increase in the years immediately following German re-unification in 1990. While the polynomial trend line suggests a slowly continuing increase of independents, the raw data indicate that a plateau of non-partisanship was reached in 1993 and has been maintained since.

We look at independents because this is the most clear-cut information that the German version of party identification produces (see the question wording in the following footnote); gradations of partisanship are much harder to compare between different systems. 9 The German question wording goes as follows: Viele Leute in Deutschland neigen lngere Zeit einer bestimmten Partei zu, obwohl sie auch ab und zu eine andere Partei whlen. Neigen Sie, ganz allgemein gesprochen, einer bestimmten Partei zu? [yes, no] Wenn ja, welcher? [party list] Und wie stark oder schwach neigen Sie alles zusammen genommen dieser Partei zu: sehr stark, ziemlich stark, mig, ziemlich schwach, sehr schwach? Displayed here is the proportion of respondents saying no to the first question.

Graph 5

German Party Identification, 1976-2005

50

election year

40

30

'no'
20

10
1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Source: Forschungsgruppe Wahlen; West Germany only. Note that the yearly figures are arithmetic mean scores of the proportions determined by the monthly Politbarometer surveys .

East German partisanship is not covered in this graph. Because of the changing universe of the samples (West Germany before 1990, re-united Germany thereafter) the trend-line would become blurred. But this of course is not to say that there is no empirical information available. Early in the unification process, it has been established that East Germans were no virgins in terms of affiliation with West German parties many had every opportunity to align themselves virtually with a party of the Federal Republic (Schmitt 1992). In the year after re-unification, East Germans started out with 41 percent independents which however, as a result of dissatisfaction with the material consequences of unification, quickly rose to 52 percent in 1993. From then on, however, non-partisanship in the East declined more or less continuously in order to reach again the 40 percent level from which it started out in 1990 (Politbarometer results, data not shown).

The diminishing of ideological polarisation, the diversification of social oppositions, and the dwindling of partisan ties all of this suggest that short-term factors have become extra-ordinarily important factors in the German voting decision. These short-term factors are what we will look at next.

5. Short-term factors

Short-term factors are the ones that can distort longer-term trends. They come in two species issues and candidates. Both issue and candidate effects on the vote are election-specific which is to say that both can change from election to election. Before we start looking into this, we need to address again the question what the longer-term trends are. Some of the literature suggests electoral cycles being at 9

work. It is proposed that governments support tends to follow a cyclical pattern of the kind that, after some sort of post-electoral euphoria, they usually loose support until after midterm and only then recover again and rise to a level close to the previous result (Tufte 1975; Stimson 1976; Dinkel 1977a, 1977b; Reif and Schmitt 1980). It seems that this cyclical pattern is not found everywhere and does not show all the time, and it is therefore not beyond dispute (e.g. van der Eijk 1987). Research into European Parliament election results, however, has repeatedly provided evidence that can confirm such an electoral cycle in consolidated electoral systems (e.g. Marsh 2005; Schmitt 2005b; Wst and Roth 2005).

Previous research into German electoral cycles has shown that the popularity of SPD-led governments does not follow cyclical developments as smoothly as CDU/CSU-led governments do (Schmitt and Reif 2003). The two Schrder-governments are no exception to the rule (Graph 5). Schrder I in particular does not fit the expected pattern: the government was seen to have a very bad start in its first year, enjoyed high ratings during midterm, but deteriorated in the winter of 2001/2002 only in order to recover again during a spectacular election campaign.10 Satisfaction with Schrder II dropped even more dramatically immediately after the election than it did four years earlier, and remained at a very low level over the first two years. A short increase was followed by another downturn (to the loss of another state election in February 2004), which than marked the start of the 2005 election campaign.11 All in all, these results demonstrate that satisfaction with government has been very volatile in Germany lately. Likely causes, we think, are the shrinking of ideological distances and the complementary decline of partisanship that has been referred to before.

Graph 6

Satisfaction With Government

2,0
Schrder I Schrder II

1,5 1,0 0,5 0,0 -0,5 -1,0 -1,5 -2,0

10/98 01/99 04/99 07/99 10/99 01/00 04/00 07/00 10/00 01/01 04/01 07/01 10/01 01/02 04/02 07/02 10/02 01/03 04/03 07/03 10/03 01/04 04/04 07/04 10/04 01/05 04/05 07/05

Source: Forschungsgruppe Wahlen Politbarometer and other surveys. Respondents are asked how satisfied they are with government (displayed) and opposition (not shown). The graph indicates mean scores on an11 point scale ( +5/-5)..

These midterm heights are attributed to a CDU scandal over illegal donations for that party plus the 9/11-events which favoured as every international crisis does the national government. 11 One could assume that these ups-and-downs of government satisfaction are accompanied by complementary evaluations of the parties of the federal opposition; there is no systematic evidence for this, however.

10

10

5.1. Issues

We distinguish two classes of issues position and valence issues. Position issues are at the core of spatial models of vote choice. The simplest and most popular form of these models assumes that voters have policy preferences on a given issue. Parties position themselves on a policy dimension in order to maximise their vote share, while voters maximise their utility by choosing the party that is closest to their own position on this dimension (Downs 1957). It has been shown that this mechanism is indeed related to vote choices, if only modestly. More important for vote choices are valence issues (Stokes 1966). Valence issues do not involve a decision for or against, or for more or less of, a political measure. Everybody tends to agree on valence issues. What differs between voters is the importance that they assign to different valence issues, and the competence that they attribute to the competing parties to solve these issues/problems. Issue competence is the mechanism that relates the perceived importance of valence issues to actual vote choices (Schmitt 2001).

Unlike other national traditions, the German series of election studies early on concentrated on this saliency/competence approach to issue voting (see the data-file of the European Voter, documented in Thomassen 2005). It is in this tradition that a representative sample of the German voting-age population was asked, shortly ahead of the 2005 election, (a) what the most important problems are, and (b) who is considered to be most competent to deal with these problems.

The results are more than obvious. In the eyes of German voters, the one issue overshadowing anything else is unemployment (85 percent important). Shortage of money i.e. inflation and salaries comes next (14 percent), followed by general references to the state of the economy (13 percent). Welfare state concerns are clearly lagging behind these economic issues: ten percent mention the pension system, nine percent taxation, and eight percent health. At the time of the 2005 election, the state of the economy in general and unemployment in particular were seen as the political problems of utmost importance (Graph 7).

Competence attributions relate saliency perceptions to vote choices. High competence scores in economic matters should translate in a clear electoral advantage. The parties who did best in these fields were CDU and CSU: roughly 40 percent of Germans referred to them as most competent in dealing with unemployment and the economy (Graph 8). The SPD was only mentioned by some 20 percent, and about the same number held the opinion that none of the parties was competent. If it comes to welfare state policies, the clear competence advantage of the CDU/CSU for the economy is not reversed in favour of the SPD, as one could have expected on the basis of the traditional policy

11

profile of this party, but rather turns into a tie between the two big parties. Altogether, then, issue effects on the vote favoured CDU and CSU and a CDU/CSU-led government.12

Graph 7

Agenda.
unemployment inflation, salary state of the economy pensions taxes health care
14 13 10 9 8 85

Source: Forschungsgruppe Wahlen flash survey conducted during the week preceding the election. Figures are percent very important of two possible responses to the agenda question.

Graph 8

Competence.
21 41 22

labour market the economy pensions taxes health care

25 22 29 29 40

25

10 31 29 21

34 35

Source: Forschungsgruppe Wahlen, flash survey conducted during the week preceding the election. Note that light grey bars represent percent SPD responses, dark grey bars represent percent CDU/CSU responses, and white bars represent percent responses none of the parties.

12 Note that both the agenda and the competence measures were taken at the end of a short but intensive election campaign. During the first two months, the SPD lagged far behind. Campaign advantages changed with the growing involvement of the communication champion of the SPD, Gerhard Schrder, but also as a result of apparent leadership deficiencies and mistakes in the CDU/CSU campaign. An example for the latter is Angela Merkels invitation of law professor Kirchhof to join her campaign as shadow finance minister. Instead of promoting the Christian democratic tax concept, however, Kirchhof advertised his own flat tax model. Needless to say that the SPD in concert with the Greens did their best to nurture this conceptual confusion.

12

5.2. Candidates

This is not the case with respect to the candidates standing for office. The SPD-chancellor Gerhard Schrder, not the CDU-candidate Angela Merkel, was the more attractive candidate. If we compare the like-dislike scores of Schrder and the SPD over the eighteen months preceding the election, Schrder was always ahead of his party he was clearly adding to the electoral attractiveness of the SPD (data not shown). The contrary emerges for Merkel and the CDU; she has been clearly less popular than the CDU, and could only during the campaign close the gap and climb up to her partys popularity levels.

A more direct comparison of the two chancellor candidates is presented in Graph 9. It shows which of the two candidates was the most preferred, and the evolution of these preferences during the campaign. Gerhard Schrder is clearly in the lead, here as well. Only in the very early phase of the campaign, when the plan for an early election was just announced, did pro-Merkel preferences outnumber those in favour of Schrder. From mid-July on, Schrder could more or less continuously increase his support-base. On election day, 53 percent of all Germans preferred him as the next chancellor, while only 39 percent spoke out for Angela Merkel.

Graph 9

Chancellor Preference
60

53%
55

Schrder
50

45

40

Merkel
35

39%
30
May Jun I Jun II Jul I Jul II Aug I Aug II Aug III Aug IV Sept I Sept II Sept III

Source: Forschungsgruppe Wahlen Politbarometer

One should not underestimate the impact of these short term factors on vote choices at times when the decision which party to vote for is taken closer and closer to election day. In 2005, one in three voters 13

 more precisely: 31 percent decided on their vote during the election campaign, a few days or a few weeks ahead of the election. The corresponding figure in 2002 was 24 percent, and it was 22 percent in 1998.13

In 2005, these short-term factors were favouring both of the two competing camps differently. On the candidate side, the SPD could profit from the popularity of its chancellor Gerhard Schrder. And with respect to issues, the CDU/CSU was clearly ahead in terms of competence attributions for the most pressing political problems: unemployment and the economy. With a stronger candidate and, one is tempted to add: a better campaign, Christian democrats and Christian socials could have really won the election and formed a Christian-conservative/liberal government coalition as intended.

6. Results and likely consequences.

This however was not the result of the election (Table 2). The two major parties came out about equally strong, with 34,2 and 35,2 percent of the valid vote, respectively. The FDP did remarkably well (9,8 percent), as did Die Linke (8,9 percent) the latter being a new formation combining the Eastern-based PDS and the left-wing SPD breakaway WASG. Die Grnen could about hold their electoral support (8,1 percent). Upon this basis, neither a continuation of the red-green government of SPD and Die Grnen nor the formation of a Christian-liberal government was numerically possible. A broadening of the red-green coalition with the inclusion of Die Linke was not seriously considered, least so perhaps by the SPD and by Die Linke itself. A so-called Ampel-Koalition including SPD, FDP and Grne was ruled-out on policy grounds by the FDP. Another option, a socalled Jamaika-Koalition of CDU/CSU, FDP and Grne was equally energetically ruled-out by Die Grnen and the CSU. The formation of a minority government has no tradition in post-war Germany and was not even considered. The only option that remained to be pursued was the formation of a grand coalition of CDU/CSU and SPD.

What are the likely consequences of this grand coalition? One is that some of the causes of the policy deadlock produced by discordant majorities in Bundestag and Bundesrat might be addressed and eliminated. The grand coalition commands a solid majority in each of the two chambers and should therefore be able to pave the way for a reform of state and federal competences, with the aim of reducing the requirement of dual majorities as much as possible.

The source of these figures is again the flash surveys of the Forschungsgruppe Wahlen during the weak preceding election day. The percentage-base includes all respondents with a valid vote intention, and excludes those who will not vote or do not know which party.

13

14

Table 2 The result (figures are vote percentages and number of seats)

diff 1st vote 2002

diff 2nd vote 2002

diff seats 2002

participation SPD CDU CSU Grne FDP Linke others

77,7 38,4 32,6 8,2 5,4 4,7 8,0 2,7

-1,4 -3,5 0,6 -0,8 -0,3 -1,1 3,6 1,5

77,7 34,2 27,8 7,4 8,1 9,8 8,7 4,0

-1,4 -4,3 -1,7 -1,6 -0,4 2,5 4,7 0,8 222 180 46 51 61 54 -29 -10 -12 -4 14 52 -

Source: Federal Statistical Office.

Another likely consequence of this election is perhaps less positive. It might bring about a further reduction of ideological and policy alternatives in the centre of the German party system. The two major players, CDU/CSU and SPD, will almost certainly move closer together as a result of their cooperation in the grand coalition. This will open up new chances for political entrepreneurs: empty spaces e.g. on the far right might be filled by new or re-vitalised political formations. The NPD is just an example that comes to mind.

As a result of these processes, the grand coalition will certainly contribute to a further reduction of the vote share of the two major parties. CDU/CSU and SPD have already lost a fifth of their 1972 proportion of the valid vote. The current development will lead to further fractionalisation of the German party system and, perhaps in the long run, to a gradual change in the direction of party competition from centripetal to centrifugal tendencies.

Last but not least, this election result might be remembered as another step towards the volatile voter. A further decline of ideological polarisation will contribute to continuing dealignment and the weakening of partisanship. This will strengthen existing tendencies towards vote switching and late vote decisions. German voters will become even less predictable than they were in 2005.

15

References

Berglund, F., S. Holmberg, H. Schmitt and J. Thomassen. 2005. Party Identification. In The European Voter, ed. Jacques Thomassen. Oxford: Oxford University Press, pp. 106-124. Brklin, W. 1984. Grne Politik. Opladen: Westdeutscher Verlag. Campbell, A., Ph. E. Converse, W. E. Miller, and D. E. Stokes. 1960. The American Voter. New York: Wiley. Clarke, H., D. Sanders, M. C. Stewart, and P. Whitley. 2004. Political Choice in Britain. Oxford: Oxford University Press. Converse, Ph. E. 1969. Of time and partisan stability. Comparative Political Studies 2: 139-171. Dalton, R. J. and M. P. Wattenberg, eds. 2000. Parties without Partisans. Oxford: Oxford University Press. Dalton, R. J. 1996. Political Cleavages, Issues, and Electoral Change. In Comparing Democracies, eds. L. Leduc, R. G. Niemi, and P. Norris. Thousand Oaks: Sage. Dalton, R. J. 1984. Cognitive Mobilizationand Partisan Dealignment in Advanved Industrial Democracies. Journal of Politics 46: 264-284. Dalton, R. J., S. Flanagan, and P. E. Beck, eds. 1984. Electoral Change in Advanced Industrial Democracies: Realignment or Dealignment? Princeton: Princeton University Press. Dinkel, R. 1977b. Der Zusammenhang zwischen Bundes- und Landtagswahlergebnissen. In Wahlsoziologie heute. Analysen aus Anla der Bundestagswahl 1976, ed. M. Kaase. Opladen: Westdeutscher Verlag [=PVS Sonderheft 2/3 vom November 1977] Dinkel, R. 1977a. Der Zusammenhang zwischen der konomischen und der politischen Ent-wicklung in einer Demokratie. Berlin: Duncker & Humblodt. Downs, A. 1957. An Economic Theory of Democracy. New York: Harper & Row. Eijk, Cees van der. 1987. Testing theories of electoral cycles: The case of the Netherlands. European Journal of Political Research 15: 253-270. Franklin, M. N., T. Mackie, and H. Valen, eds. 1992. Electoral Change. Cambridge: Cambridge University Press. Klein, M. and D. Ohr. 2000. "Gerhard oder Helmut? 'Unpolitische' Kandidateneigenschaften und ihr Einfluss auf die Wahlentscheidung bei der Bundestagswahl 1998." Politische Vierteljahresschrift 41: 199-224. Lehmbruch, G. 21998. Parteienwettbewerb im Bundesstaat. Opladen: Westdeutscher Verlag. (1. Auflage 1976). Lijphart, A. 1999. Patterns of Democracy. New Haven and London: Yale University Press. Marsh, M. 2005. The results of the 2004 European Parliament elections and the second-order model, In Europawahl 2004, eds. O. Niedermayer and H. Schmitt. Wiesbaden: Verlag fr Sozialwissenschaften, pp. 142-158.

16

Miller, W. E. and J. M. Shanks. 1996. The New American Voter. Cambridge and London: Harvard University Press. Norris, P. 2004. Electoral Engineering. Cambridge: Cambridge University Press. Reif, K. and H. Schmitt. 1980. Nine Second-order National Elections. A Conceptual Framework for the Analysis of European Election Results. European Journal of Political Research 8: pp.344. Roth, D. und A. M. Wst. 2006. Abwahl ohne Machtwechsel. Die Bundestagswahl 2005 im Lichte lngerfristiger Entwicklungen. In Die Bundestagswahl 2005, eds. E. Jesse and R. Sturm. Wiesbaden: VS Verlag [forthcoming]. Schmitt, H. 2005b. The European Parliament elections of June 2004: Still second-order? West European Politics 28: 650-679. Schmitt, H. 2005a. Politische Parteien, Links-Rechts-Orientierungen und die Wahlentscheidung in Frankreich und Deutschland. In Wahlen und Whler. Analysen aus Anlass der Bundestagswahl 2002, eds. J. W. Falter, O. Gabriel, and B. Wessels. Wiesbaden: Verlag fr Sozialwissenschaften, pp. 551-571. Schmitt, H. 2002. Partisanship in Western Europe and the US: Causes and Consequences. Paper presented at the Annual Meeting of the American Political Science Association in Boston. Schmitt, H. 2001a. Politische Reprsentation in Europa. Frankfurt: Campus. Schmitt, H. 2001b.Zur vergleichenden Analyse des Einflusses gesellschaftlicher Faktoren auf das Wahlverhalten: Forschungsfragen, Analysestrategien und einige Ergebnisse. In Wahlen und Whler. Analysen aus Anla der Bundestagswahl 1998, eds. H.-D. Klingemann and M. Kaase. Opladen: Westdeutscher Verlag, pp. 623-45. Schmitt, H. 1998.Issue-Kompetenz oder Policy-Distanz? In Wahlen und Whler, eds. M. Kaase and H.-D. Klingemann. Opladen: Westdeutscher Verlag, pp. 145-72. Schmitt, H. 1992. "So dicht war die Mauer nicht! ber Parteibindungen und cleavages im Osten Deutschlands". In Die Entwicklung der Volksparteien im vereinten Deutschland, ed. G. Hirscher. Mnchen: Verlag Bonn Aktuell, pp. 229-52. Schmitt, H. 1987. Das Parteiensystem der Bundesrepublik Deutschland. Eine Einfhrung aus politiksoziologischer Perspektive. Hagen: Fernuniversitt Hagen (Studienbrief in drei Lehreinheiten). Schmitt, H. and S. Holmberg. 1995. "Political Parties in Decline?" In Citizens and the State, eds. H.-D. Klingemann and D. Fuchs. Oxford: Oxford University Press, pp. 95-133. Schmitt, H. and D. Ohr. 2000. "Are Political Leaders Becoming More Important in German Elections? Leader Effects on the Vote in Germany, 1961-1998." Paper prepared for delivery at the 2000 Annual Meeting of the American Political Science Association, Marriot Wardman Park, Washington DC, August 31-September 3. Schmitt, H. and K. Reif. 2003. Der Hauptwahlzyklus und die Ergebnisse von Nebenwahlen. In Politbarometer, ed. Andreas Wst. Opladen: Leske und Budrich. Stimson, J.A. 1976. Public Support for American Presidents: A Cyclical Model. Public Opinion Quarterly 40: 1-21.

17

Stokes, D. E. 1992. Valence Politics. In Electoral Politics, ed. D. Kavanagh. Oxford: Clarendon Press. Stokes, D. E. 1966. Spatial Models of Party Competition. In Elections and the Political Order, eds. A. Campbell, Ph. E. Converse, W. E. Miller and D. E. Stokes. New York: Wiley. Thomassen, J., ed. 2005. The European Voter. Oxford: Oxford University Press Tufte, E. R. 1975. Determinants of the outcomes of midterm congressional elections. American Political Science Review 67: 540-54. Wst, A. M. and D. Roth. 2005. Parteien, Programme und Wahlverhalten. In Wahl-Kampf um Europa, ed. J. Tenscher. Wiesbaden: Verlag fr Sozialwissenschaften, pp. 56-85.

18