The Austrian Voter 9783737001663, 9783847101666, 9783847001669

Citation preview

Sylvia Kritzinger / Eva Zeglovits / Michael S. Lewis-Beck / Richard Nadeau

The Austrian Voter

V& R unipress Vienna University Press

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. ISBN 978-3-8471-0166-6 ISBN 978-3-8470-0166-9 (E-Book) Publications of Vienna University Press are published by V& R unipress GmbH. Printed with the kind support of the Rectorate of the University of Vienna. Ó Copyright 2013 by V& R unipress GmbH, D-37079 Goettingen All rights reserved, including those of translation into foreign languages. No part of this work may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, microfilm and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Printing and binding: CPI Buch Bücher.de GmbH, Birkach Printed in Germany

Inhalt

Entfernen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

11 11 13 13 14 15 15 21 24 26 27

2 Chapter 2: Region, Age, Gender . . . . . . . . . . . . . . . . . . 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Region, Residence and Degree of Urbanization . . . . . . . . 2.2.1 Size of Municipality and the Vote: 1986 – 2008 . . . . . . 2.2.2 Electoral Constituencies and the Vote in 2008 . . . . . . 2.3 Gender . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Gender and the Vote 1986 – 2006 . . . . . . . . . . . . . 2.4 Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Age and the Vote 1986 – 2006 . . . . . . . . . . . . . . . . 2.5 Region-Age-Gender and Vote Choice in Austria: Multivariate Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1 The period 1990 – 2006 . . . . . . . . . . . . . . . . . . . 2.6 Region, Age and Gender and the Vote 2008 . . . . . . . . . . 2.6.1 Gender and the Vote 2008 . . . . . . . . . . . . . . . . . 2.6.2 Age and the Vote 2008 . . . . . . . . . . . . . . . . . . . 2.6.3 The 2008 election: Multivariate Analysis . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

31 31 31 33 35 39 40 42 43

. . . . . .

. . . . . .

. . . . . .

45 45 48 48 49 49

1 Chapter 1: The Austrian Case: an Overview . . . . . . . . . . 1.1 Aim of the Book . . . . . . . . . . . . . . . . . . . . . . . 1.2 A Bit of History – from World War II till Today . . . . . . . 1.2.1 The Electoral System – Changes over Time . . . . . . 1.2.2 Voter Turnout over Time . . . . . . . . . . . . . . . . 1.3 Party Politics and Electoral Behaviour in Austria . . . . . 1.3.1 The Austrian Party System over Time . . . . . . . . 1.3.2 Voters’ Characteristics and Party Choice over Time . 1.3.3 The 2008 Election: the Context . . . . . . . . . . . . 1.4 Theoretical Considerations . . . . . . . . . . . . . . . . . 1.5 Data and Methodological Approach . . . . . . . . . . . .

. . . . . . . . . . .

6

Inhalt

2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

. . . . . . . . . . . .

53 53 54 55 57 59 63 64 71 71 72 73

. . . . . . . . . .

73 75

4 Chapter 4: Social-Psychological Factors . . . . . . . . . . . . . . . . . 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Group Identification: The Political Party and Union Membership . 4.2.1 Union Membership and the Vote: 1990 to 2006 . . . . . . . . 4.2.2 Party Identification and the Vote in Austria: 1954 – 2006 . . . 4.3 Ideological Orientations: the Left-Right Dimension . . . . . . . . . 4.3.1 The Left-Right Dimension and the Vote in Austria: neglected in earlier research? . . . . . . . . . . . . . . . . . . . . . . . 4.4 Union Membership, Party Identification and Ideology in the 2008 Election . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 Union Membership and the Vote in the 2008 Election . . . . 4.4.2 Party Identification and the Vote in the 2008 Election . . . . 4.4.3 The Left-Right Dimension and the Vote in the 2008 election . 4.4.4 Multivariate Analysis: the Austrian National Election 2008 . . 4.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79 79 80 81 83 84

87 87 88 89 92 95

5 Chapter 5: Short-Term Indicators: Issue Positions . . . . . . . . . . . 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Issue Voting in Austria . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Issue Voting and the 2008 Election: an Array of Possible Issues . . 5.4 Issue Voting and the 2008 Election: Issues Regarded as Important . 5.5 Key Issues in the 2008 Election: A Summary . . . . . . . . . . . .

97 97 97 100 102 104

3 Chapter 3: The Main Cleavages: Religion and Social Class . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Religion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Religion and the Vote: 1990 to 2006 . . . . . . . . . 3.3 Social Class . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Occupation and the Vote: 1990 – 2006 . . . . . . . . 3.3.2 Education and the Vote: 1990 – 2006 . . . . . . . . . 3.4 Multivariate Analyses: 1990 – 2006 . . . . . . . . . . . . . 3.5 The Main Cleavages and the 2008 Vote . . . . . . . . . . . 3.5.1 Religion and Vote in 2008 . . . . . . . . . . . . . . . 3.5.2 Occupation and the Vote: the 2008 Election . . . . . 3.5.3 Education and the Vote: the 2008 election . . . . . 3.5.4 Social Structure and the Vote in 2008 : Multivariate Analysis . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

85

7

Inhalt

5.6 Multivariate Analysis: Issues and Vote Choice . . . . . . . . . . . . 5.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

105 108

. . . . . .

111 112

. . .

112

. . .

113

. . .

115

. . .

116

. . .

116

. . . . . .

118 120

7 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1 Data 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Data 1986 – 2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123 123 131

8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

137

6 Chapter 6: Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 6.1 The Impact of Long Term Variables on Vote Choice . . . . . . 6.1.1 Austrian Electoral Behaviour Over Time – the Impact of Demographic Variables . . . . . . . . . . . . . . . . . . 6.1.2 Austrian Electoral Behaviour Over Time – the Impact of Social Structure . . . . . . . . . . . . . . . . . . . . . . 6.1.3 Austrian Electoral Behaviour – the Impact of Social-Psychological Variables . . . . . . . . . . . . . . 6.1.4 Long-Term Determinants of Vote Choice in Austria – a Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Austrian Electoral Behaviour – the Impact of Short-Term Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Austria – a Peculiar Case? Austrian Electoral Behaviour in a European Perspective . . . . . . . . . . . . . . . . . . . . . . 6.4 Austrian Electoral Behaviour – a Glance at the Future . . . .

Entfernen

We dedicate this book to Paul F. Lazarsfeld and Philip E. Converse, two truly inspiring scholars for electoral researchers.

This research was carried out under the auspices of the Austrian National Election Study (AUTNES), a National Research Network (NFN) sponsored by the Austrian Research Fund (FWF) (S10903-G11). Sylvia Kritzinger’s research was also supported by a Grant from the Netherlands Institute for Advanced Studies in the Humanities and Social Sciences (NIAS). We would like to thank Patricia Oberluggauer, Maria Schlechter and Kristian Dunn for their research assistance.

1 Chapter 1: The Austrian Case: an Overview

1.1

Aim of the Book

When reviewing the comparative political science literature on electoral behaviour, one can observe that Austria is rarely considered either theoretically or empirically. This state of the literature is rather surprising as Austria is part of the ‘old’ West European democracies, acting as an important bridge to the former ‘communist block’ and featuring interesting party characteristics. One of the reasons for Austria’s absence in the comparative literature is due to how data on electoral behaviour were collected in Austria. Unlike in other countries, data on national electoral behaviour were mostly not collected by academics; rather, polling institutes collected the data on behalf of political parties interested in their electoral fortune. Access to electoral data was thus to a great extent dependent on political parties’ goodwill in sharing their data and hardly ever made it into the public domain or into comparative international datasets. Apart from data access problems, researchers who were granted access to the data were faced with severe constrains in terms of which variables were available as the data collection process was not driven by academic interests. Rather, the data was collected to meet the operational needs of the political parties and was unsuitable for sophisticated models on electoral behaviour. This problem became even more evident the more fine-grained and sophisticated the theoretical models in electoral research became. The situation changed substantially in 2009 when for the first time an academic election study was carried out in Austria, namely AUTNES. AUTNES, an acronym for the Austrian National Election Study, financed by the Austrian Science Fund and led by Wolfgang C. Müller, is the first academic election study to serve the needs of the academic community rather than the political parties. Thus, it includes for the first time a wide range of indicators necessary to test and extend the various theoretical models in electoral behaviour. Further, taking full advantage of this freely available new data source (AUTNES, 2009; Kritzinger et al., 2011) allows us to gain new insight into the Austrian case. This book has overall two main objectives: first, we ‘bring Austria in’, we extend

12

The Austrian Case: an Overview

comparative electoral research by adding the Austrian case. Second, we provide an international audience with empirical knowledge about electoral behaviour in Austria. We analyse contemporary electoral behaviour in Austria in the context of Austrian electoral history. As such, we capture the continuities and changes in Austrian electoral behaviour. Building on common theoretical approaches used in the literature on electoral behaviour, we test the impact of sociological, ideological and issue variables on vote choice, taking a time dimension into account, where possible. Thus, we provide empirical evidence on the Austrian voter – both from an historical and present-day perspective. Apart from adding a case study of Austria to comparative electoral research, Austria is an interesting case in and of itself; both in terms of its electoral rules and electoral outcomes. Since 2007, Austria is the only country amongst established democracies which allows voting at the age of 16 in all national elections and referenda. In terms of electoral outcome, Austria was known for its stability in vote choice in the past. In fact, there was a duopoly of the two mainstream parties – the Social Democratic Party of Austria (SPÖ) and the Austrian People’s Party (ÖVP) – dominating political, economic and cultural life till the 1990s. At the mass level, this confrontation found expression in two main ideological Lagers, also called ‘camps’ or ‘political subcultures’ (e. g. Luther and Müller, 1992), in the Austrian electorate with one Lager having a left-oriented mentality and the second Lager oriented towards the right. Recently, Austria is more well-known for its strong radical-right parties. The rise of the Austrian Freedom Party (FPÖ) started in 1986 when Jörg Haider took over as chairman of the party and culminated in 1999 with the FPÖ gaining nearly 27 per cent of the national vote share: as a result, the FPÖ became the junior partner in the coalition government with the ÖVP. After some internal conflicts, in 2005 the FPÖ split; next to the FPÖ the Alliance for the Future of Austria (BZÖ), a minor radical-right party, was established. In the 2008 national election, which is the focus of this book, these two radical-right parties jointly received 28.2 per cent of the vote share. The book is structured as follows. In this introduction, we provide a short overview of the Austrian case drawing on past descriptions of Austrian scholars.1 We put the Austrian case and our analysis in context: we briefly present the historical background of the Austrian political system and the election of 2008 and describe our data in detail before we start our analysis on Austrian electoral behaviour. In the following chapters, we summarize previous empirical findings and test various theoretical models against the data of the 2008 election. 1 In our short overview, we focus on the period after World War II. For more information on electoral behaviour during the time of the Habsburg monarchy or in the First Republic see e. g. Heinisch, 2002; Rathkolb, 2008.

A Bit of History – from World War II till Today

1.2

13

A Bit of History – from World War II till Today

Interpreting and understanding current electoral behaviour cannot be done without putting it into a historical context. On the one hand, the institutional context such as the electoral and the party system is of importance, on the other hand, former electoral behaviour stands in close relationship with current individual voting decisions. Both continuities and changes in electoral behaviour need to be nested in past behaviour to unfold its meanings but also its repercussions for the current political system. In the following, we present the context of Austrian politics, first, by explaining the electoral system and the changes it experienced over time, and second, by describing the Austrian party system, its changes and its current status quo.

1.2.1 The Electoral System – Changes over Time As a federal country, Austria has a bi-cameral parliament: a Federal Council (Bundesrat) with limited power and a National Council (Nationalrat). Members of the Federal Council are delegates appointed by the nine regional parliaments, while members of the National Council are chosen by popular vote. Seats in the National Council are assigned to parties via proportional formulae. Since World War II (WWII) several reforms of the electoral system have taken place. The general trend in established democracies to lower the voting age from usually 21 to 18 after the mid-1960s (Franklin, 2004) could be observed in Austria as well: in 1968 voting age was lowered from 21 to 19. The reform of 1970 then increased the number of seats in the National Council from 165 to 183, and reduced the number of constituencies from 25 to 9, next to some changes in seat assignment and the introduction of the preference vote in the 9 constituencies (Ucakar, 1995). The next major reform took place in 1992 and introduced three important changes: First, voting age was further lowered from 19 to 18. Second, the number of constituencies was again increased, namely from 9 to 43, and also seat assignment again changed: most importantly, parties must either exceed the general threshold of 4 per cent of nationwide vote share to gain seats in the National Council, or win a direct mandate at one of the regional constituencies. Third, the reform considerably strengthened the preference vote: the closed list system came thus to an end (Müller and Scheucher, 1995). The last reform of the electoral system took place in 2007. This most recent reform facilitated absentee voting and most importantly enfranchised adolescents. The reform introduced a general voting age of 16 for all nationwide elections, including federal elections, presidential elections and elections for the European Parliament, as well as referenda and all forms of plebiscites (Hofer et al., 2008).

14

The Austrian Case: an Overview

Austria is thus the only western democracy that lowered the voting age to 16 in national elections. Moreover, the legislative period was extended from four to five years.2

1.2.2 Voter Turnout over Time After the end of World War II, Austria was occupied by the Allies until 1955. Austria, like Berlin, was divided into a Soviet, an American, a French and a British sector. Nevertheless, the first national elections were held in November 1945. Over 93 per cent of Austrian citizens eligible to vote3 turned out in this first election (see Table 1). Table 1: Turnout in Austria: elections for the National Council 1945 – 2008

Election 1945 1949 1953 1956 1959 1962 1966 1970 1971 1975 1979 1983 1986 1990 1994 1995 1999 2002 2006 2008

Turnout 94.3 96.8 95.8 96.0 94.2 93.8 93.8 91.8 92.4 92.9 92.2 92.6 90.5 86.1 81.9 86.0 80.4 84.3 78.5 78.8

Source: Federal Ministry of Internal Affairs. Note: Turnout also includes invalid votes.

In the following election in 1949, this rate increased to over 95 per cent. Though this high turnout rate of the 1940s was never reached again, turnout remains high in Austria. Only as late as the 1980s, did turnout rates fall below 90 per cent, 2 For a more detailed overview on changes in the Austrian electoral system see e. g. Ucakar and Gschiegl, 2010. 3 Noteworthy in this first election, former Nazi party members and members of some military units were not allowed to vote (Bundesministerium für Inneres, 2012). This is the main reason why the number of eligible voters increased from 3.5 million in 1945 to 4.3 million in 1949.

Party Politics and Electoral Behaviour in Austria

15

decreasing to approximately 80 per cent in the 1990s before increasing again slightly in 2002. However, in the following elections, turnout fell below 80 per cent for the first time, with ‘only’ 78.8 per cent of citizens turning out to vote in 2008. Thus, Austria does not differ from other countries in exhibiting a decrease in turnout rates; a trend Franklin (2004) partly traced back to the lowering of the voting age to 19 or 18 after the mid-1960s. Nevertheless, compared to other democracies, turnout rates in Austria remain relatively high, and thus, Austria falls into the category of European democracies with very high turnout.

1.3

Party Politics and Electoral Behaviour in Austria

1.3.1 The Austrian Party System over Time The result of the first election in 1945 was already a reflection of the electoral trends to come in the following years. Two parties were able to obtain 94.4 per cent of the votes : The two dominating parties were the Austrian People’s Party (ÖVP), obtaining 49.8 per cent of the vote share, and the Socialist Party of Austria (SPÖ), with 44.6 per cent. Only one other party was able to win seats for the National Council, the Communist Party of Austria (KPÖ), which received 5.4 per cent of the votes. In the election of 1949, a fourth party made it to the National Council ; the Verband der Unabhängigkeiten (VdU) obtained 11.7 per cent of the votes primarily at the expense of the two dominating parties which each dropped about 5 per cent in their vote share. The VdU aimed to attract war returnees and former Nazis (Luther, 2000), and is the predecessor of the Freedom Party of Austria (FPÖ).4 With the election of 1959, the electoral distribution largely returned to the initial pattern of 1945. The KPÖ lost most of its vote share and did not make it into the National Council. The FPÖ only obtained 7.7 per cent of the votes, and the SPÖ and ÖVP each received about 44 per cent of the votes. For the next 25 years this was the outcome one could observe when looking at Austrian electoral results : SPÖ and ÖVP together obtaining around 90 per cent of the votes and a third party (the FPÖ) with a steady vote share of around 5.5 per cent. Electoral success and defeat were thus considered exceptional once they exceed a variation of 2 percentage points (Kienzl, 1964). The SPÖ and the ÖVP were the two main political actors in the Austrian 4 As a result of internal frictions and slight vote losses in 1953, the VdU was absorbed by the newly founded FPÖ in 1956.

16

The Austrian Case: an Overview

political scene for years to come. Their overwhelming and stable electoral success was the reflection of societal division into two so-called Lagers. This division worked mainly along two main cleavages: class and religion (Lipset, 1983 ; Lipset and Rokkan, 1967). To a smaller degree a region cleavage was also reflected in this Lager division, as the catholic-conservative Lager was operating foremost in rural areas and small towns, although its initial political strength was also found in Vienna. The third historic Lager, nationalist, anticlerical and anti-socialist was – distinct from the other two Lagers – not reflected in a mass party5 (Luther and Müller, 1992). While the SPÖ represented the blue-collar worker with low religious ties living in urban areas and with low education, the ÖVP was the typical Christian-democratic party, representing the interests of enterprises and farmers as well as voters with strong Catholic ties living mostly in rural areas. This clear-cut societal division remained in place for many decades after WWII. In the literature, Austria is therefore often described as a typical example of a political system with Lager-mentalities dominating the political scene and discourse. The time between 1947 and 1966 was described as the “‘classic’ phase of Austrian consociationalism” (Luther and Müller, 1992, p.10) characterized by high partisan attachment and stable electoral behaviour. Until the early 1970s, party competition was thus limited (Plasser and Ulram, 2008). The constellation of two dominating parties with a third small player complicated government formation. In the immediate aftermath of WWII a so-called concentration government, including all parties represented in the National Council, was formed. After the election in 1949 the government was formed as a grand-coalition between SPÖ and ÖVP. Neither of the two dominant parties was able to gain a majority of seats in the National Council and the third party, the VdU, was not considered a viable option due to its historical connection with National Socialism. Unlike in other countries, the third party was thus not a pivotal player in government formation.

5 Some scholars talk about another salient cleavage in the Austrian electorate; next to class, religion and region, nationalism was considered a salient cleavage (e. g. Plasser and Ulram., 1992). This cleavage reflects mostly the electorate of the FPÖ which was often considered a vessel for former Nazis. Due to the negligible electoral success of the FPÖ until the 1980s we do not consider nationalism a salient cleavage. At most, we think it can be considered a latent cleavage.

Party Politics and Electoral Behaviour in Austria

17

Table 2: Government coalitions 1945 – 2013 Legislative period Government Chancellor (name) Chancellor (party) 1945 – 1945 ÖVP/SPÖ/KPÖ Karl Renner SPÖ 1945 – 1949 ÖVP/SPÖ Leopold Figl ÖVP 1949 – 1952 ÖVP/SPÖ Leopold Figl ÖVP 1952 – 1953 ÖVP/SPÖ Leopold Figl ÖVP 1953 – 1956 ÖVP/SPÖ Julius Raab ÖVP 1956 – 1959 ÖVP/SPÖ Julius Raab ÖVP 1959 – 1960 ÖVP/SPÖ Julius Raab ÖVP 1960 – 1961 ÖVP/SPÖ Julius Raab ÖVP 1961 – 1963 ÖVP/SPÖ Alfons Gorbach ÖVP 1963 – 1964 ÖVP/SPÖ Alfons Gorbach ÖVP 1964 – 1966 ÖVP/SPÖ Josef Klaus ÖVP 1966 – 1970 ÖVP Josef Klaus ÖVP 1970 – 1971 SPÖ Bruno Kreisky SPÖ 1971 – 1975 SPÖ Bruno Kreisky SPÖ 1975 – 1979 SPÖ Bruno Kreisky SPÖ 1979 – 1983 SPÖ Bruno Kreisky SPÖ 1983 – 1986 SPÖ/FPÖ Fred Sinowatz SPÖ 1986 – 1987 SPÖ/FPÖ Franz Vranitzky SPÖ 1987 – 1990 SPÖ/ÖVP Franz Vranitzky SPÖ 1990 – 1994 SPÖ/ÖVP Franz Vranitzky SPÖ 1994 – 1996 SPÖ/ÖVP Franz Vranitzky SPÖ 1996 – 1997 SPÖ/ÖVP Franz Vranitzky SPÖ 1997 – 2000 SPÖ/ÖVP Viktor Klima SPÖ 2000 – 2003 ÖVP/FPÖ Wolfgang Schüssel ÖVP 2003 – 2007 ÖVP/FPÖ(BZÖ) Wolfgang Schüssel ÖVP 2007 – 2008 SPÖ/ÖVP Alfred Gusenbauer SPÖ 2008 – 2013 SPÖ/ÖVP Werner Fayman SPÖ Source: Austrian Federal Chancellery Note: Dates of government formation can deviate from election dates due to coalition bargaining or cabinet reshuffle.

The phenomenon of forming grand-coalitions can be observed for the next two decades, until 1966. With the SPÖ and the ÖVP being forced to work together, over 90 per cent of the electorate was represented in government; in other words, the government took care of the interests of both of the main cleavage groups. It is therefore not surprising that Austria developed into a consensual democracy (Lijphart, 1999). Proporz6 (i. e. proportionality) and corporatism were established in order to guarantee the cohesion necessary for political and economic stability after the war (Heinisch, 2002). In Table 2, we see the composition of government coalitions across the entire period from World War II, including the party and chancellor names7. 6 Proporz describes the power sharing mechanism that divided “virtually all public and quasipublic institutions” (Heinisch, 2002: p.3) proportionally between SPÖ and ÖVP. 7 Although the ÖVP did not always have the majority of the votes (e. g. in 1953 and 1959, see Table 3), it nevertheless had the majority of seats in the National Council due to the electoral

18

The Austrian Case: an Overview

The 1966 election brought the governing alliance between the ÖVP and the SPÖ to a halt for the next 20 years: In the 1966 election, the ÖVP gained enough votes to secure an absolute majority in seats in the Austrian National Council. In the 1970 election, the SPÖ won 48.4 per cent of the votes but did not win a majority in seats, while the ÖVP obtained ‘only’ 44.7 per cent of the vote share. Instead of again forming a grand-coalition, this time for the first time with an SPÖ chancellor, the SPÖ opted for a minority government supported by the FPÖ. It lasted for about one and half years, and in the election of 1971 the SPÖ not only gained an absolute majority in votes but also in seats. From this point until 1983, Austria was governed by a single party government formed by the SPÖ. In 1983 the SPÖ lost its absolute majority and formed a coalition with the FPÖ which at that time presented itself as a liberal rather than a nationalistic party (Luther, 1987). The 1980s brought quite a lot of changes to the ‘stable’ Austrian party system. In 1983 for the first time, two green parties ran in an electoral campaign. Together, they gained 3.4 per cent of the vote share but due to the threshold of 4 per cent they did not obtain any seats in the National Council. In 1986, Jörg Haider became the leader of the FPÖ and changed the party’s profile substantially. While under the former leader, Norbert Steger, the FPÖ tried to present itself as a liberal party which guaranteed government participation with the SPÖ between 1983 and 1986, under Haider the FPÖ developed into a right-wing party focusing on anti-establishment and nationalistic issues and using anti-immigration rhetoric. As a consequence of this policy change, the SPÖ terminated its coalition with the FPÖ. The following election marked a major upheaval in Austrian politics. The Greens managed to exceed the 4 per cent threshold and were represented with 8 members of parliament (MPs) in the Austrian National Council. The FPÖ under its new leader experienced a major electoral boost increasing their vote share from 5 per cent to 9.7 per cent. The SPÖ and ÖVP together ‘only’ received 84.4 per cent of the vote share – a novelty in Austrian politics. Thus, 1986 presents a major change in Austria both in terms of party politics and electoral behaviour: a fourth party was able to obtain representation in the National Council and the first signs emerged that the old Lagers were losing cohesion as the vote share both for SPÖ and ÖVP decreased substantially. It also marked the beginning of the success of the radical right parties in Austria (e. g. Luther, 1987; Müller, 2000). These changes made the remake of the grand-coalition between SPÖ and ÖVP necessary, but this time with the SPÖ nominating the chancellor. In the 1990, 1994, 1995 and 1999 elections these two main parties continued to lose formulae used to convert votes into parliamentary seats. Thus it always nominated the chancellor in the grand-coalition with the SPÖ between 1945 and 1966.

Party Politics and Electoral Behaviour in Austria

19

votes. While in 1990 they jointly gained 74.9 per cent of the votes, in 1999 they were only left with 60.1 per cent. In the meantime, the Greens, to a very modest degree, and the FPÖ, to a rather substantial degree, increased their vote shares. The FPÖ successfully introduced issues in Austria that united its own voters but split the SPÖ and ÖVP from certain segments of their voters (Müller, 2000). However, the new policy positions of the FPÖ were not accepted by all MPs of the FPÖ. The more liberal wing split from the FPÖ in 1993 and founded the party ‘Liberal Forum’. While it was able to pass the 4 per cent threshold in the 1994 and 1995 elections, and thus increased the number of parties in the National Council from 4 to 5, it lost the following election in 1999 and has yet to return to the National Council since. After experiencing major electoral declines, the two main parties finally reacted and tried to reposition themselves. As a reaction to the gains of the FPÖ, the SPÖ at first tried to shift the political agenda to their own key issues, and thus followed a strategy that can be described as a “mixture of hold and defuse” until the mid-1990s (Bale et al., 2010, p.419). However, as this strategy proved unsuccessful, they later tried to adopt some elements of the issues of the FPÖ, e. g. tougher stances on immigration. The ÖVP on the other hand put less emphasis on religious issues due to a sharp decrease of Catholics amongst their core voters, re-oriented their policy positions towards their other core electorate – farmers, employers and bureaucrats – and tried to attract new groups that were emerging, in particular white collar employees (Duncan, 2006). After the major defeat in the 1990 election, the ÖVP worked on a new party manifesto, which can be described as a shift from liberal market-oriented positions to more traditional conservatism (Heinisch, 2002). Though both the SPÖ and ÖVP made many efforts to re-position themselves, they never returned to their former strength but rather continued to lose electoral support from one election to the next. In 1999 the FPÖ reached 26.9 per cent of the votes, their highest vote share to date, even surpassing the ÖVP by just several hundred votes to become the second strongest political party in Austria after the SPÖ. This electoral upheaval ended the grand-coalition between SPÖ and ÖVP. It was no longer possible for them to “keep a lid on the head of steam that has been building up since the mid1980s” (Luther, 1998, p.147). A coalition government between the ÖVP and FPÖ was formed in the beginning of 2000. The ÖVP terminated the isolation of the FPÖ strategically aiming at “co-optation and castration” (Luther, 2003, p.150). This first coalition with a radical-right party caused a large international outcry and the other EU member states imposed sanctions on Austria. Rejecting this international disapproval, the two parties continued their coalition till 2006. In 2002, due to internal party frictions in the FPÖ, early elections were called by the ÖVP. The FPÖ lost a considerable amount of its voters and was only able to secure 10 per cent the vote. Both internal party dissent and poor government

20

The Austrian Case: an Overview

performance were a heavy burden for the FPÖ (Heinisch, 2003). The winner was the ÖVP which was able to increase its vote share by over 15 per cent and thus, the coalition between the ÖVP and FPÖ could be continued. Table 3: Election Results for the National Council in Austria, 1945 – 2008 Election SPÖ ÖVP FPÖ Greens Liberals BZÖ Other 1945 44.6 49.8 5.6 1949 38.7 44.0 11.7 5.6 1953 42.1 41.3 10.9 5.7 1956 43.0 46.0 6.5 4.5 1959 44.8 44.2 7.7 3.3 1962 44.0 45.4 7.0 3.6 1966 42.6 48.4 5.4 3.7 1970 48.4 44.7 5.5 0.1 1971 50.0 43.1 5.5 1.4 1975 50.4 43.0 5.4 1.2 1979 51.0 41.9 6.1 1.0 1983 47.7 43.2 5.0 4.2 1986 43.1 41.3 9.7 4.8 1.0 1990 42.8 32.1 16.6 4.8 3.7 1994 34.9 27.7 22.5 7.3 6.0 1.6 1995 38.1 28.3 21.9 4.8 5.5 1.4 1999 33.2 26.9 26.9 7.4 3.7 2.0 2002 36.5 42.3 10.0 9.5 1.0 0.7 2006 35.3 34.3 11.0 11.1 4.1 4.1 2008 29.3 26.0 17.5 10.4 2.1 10.7 4.0 Source: Federal Ministry of Internal Affairs. Note: FPÖ includes the results of its predecessor VdU (1949 and 1953); former NSDAP members were excluded from voting in 1945; there have been several Green parties in particular in the mid-1980s – we only present the vote share of the most popular one; vote shares for other parties are all the included in the ‘other’ category.

Due to continuing internal friction the FPÖ split again in 2005, with the former leader Haider founding the new party ‘Alliance for the Future of Austria’ (BZÖ). This new party was then the coalition partner of the ÖVP until the 2006 election. Running for the first time in a national election, the BZÖ was able to gain 4.1 per cent of the vote share. But the BZÖ was not a feasible coalition partner of the ÖVP any longer, as in this election the ÖVP lost a considerable amount of votes: the two parties could no longer form a stable government majority. Rather, the SPÖ won the election and again a grand-coalition between SPÖ and ÖVP was installed. This grand-coalition government did not last very long and early elections were called in 2008, again by the ÖVP. The outcome of this election has been unique for Austria: five parties made it to the National Council and all five parties obtained more than 10 per cent of the vote share. The dominance of the SPÖ and ÖVP was definitely over. The BZÖ was able to improve its electoral

Party Politics and Electoral Behaviour in Austria

21

result to 10.7 per cent while the Greens lost a small amount but still reached 10.4 per cent. The three main players were once again the SPÖ (29.3 per cent), the ÖVP (26 per cent) and the FPÖ (17.5 per cent). Given this electoral result the only numerically possible and viable coalition was again a grand-coalition, which is still in place at the moment of writing. Compared to the political constellation in 1945 and in the years up to the 1980s, Austria has changed quite substantially : from a very stable political system with two main parties dominating the political scene, in the last 30 years, and in particular in the last 10 years, Austria has produced a more volatile system with five political parties possessing more evenly distributed vote shares.

1.3.2 Voters’ Characteristics and Party Choice over Time Of course, the development of the Austrian party system is the reflection of the development of the electoral choices of Austrian voters. Here, we present the characteristics of Austrian voters in more detail and over time. In particular, we show the de-alignment processes affecting the electorate which dissolved the stable Lager mentalities in the aftermath of WWII. Taking Lipset and Rokkan’s (1967) categorization into account, two main cleavages were salient in Austria: class and religion. To a minor extent region was also important. Thus, in Austria several cleavages were at work (Müller, 1997): the class cleavages reflected the antagonism between workers (labour) and industry/enterprise/farming, while the religious cleavage represented the antagonism between being Catholic and being secular. The two cleavages overlapped as workers were more likely to be secular while religious attachments were more likely to be found amongst citizens in the farming area or owning their own enterprise. Moreover, also the regional cleavage overlapped with the other two cleavages: while the former group could be found in urban areas, the latter group was mainly distributed in rural areas. The SPÖ represented the workers/secular/ urban group and the ÖVP considered itself representative of the second group. The division of the society into two Lagers, which found its expression in the two main political parties, the SPÖ and ÖVP, could be found in all parts of life. Austria’s “‘pillarised’ political culture” (Plasser et al., 1992, p. 18) organized and structured education, occupation, cultural activities, leisure time, etc. Each Lager was characterized by a strong and uniform ideology being entirely distinctive from the ideology of the other Lager. This highly politicized structure of daily life led to a high percentage of citizens holding a party membership. The SPÖ had more than 700,000 members in the beginning of the 1980s and the ÖVP even more (Urban and Zeidner, 1983), meaning that more than a quarter of all eligible voters were members of a party.

22

The Austrian Case: an Overview

Party membership allowed access to networks which were important for career developments, schooling, etc. Moreover, a lot of interest groups and even leisure associations were affiliated to the parties (e. g. Pulzer, 1998). Nowadays Austria still has one of the highest levels of party membership in Europe (Mair and van Biezen, 2001; Treib, 2012; van Biezen et al., 2012) even though the number of members in both political parties and interest groups has decreased since the 1970s and in particular in the 1990s (Plasser and Seeber, 2000). Based on these observations, vote choice was considered to be closely connected to the social position of the voters (e. g. Berelson et al., 1954; Lazarsfeld et al., 1960). The occupation, educational background, religious affiliation, regional attachments etc. of voters were indeed strong indicators of citizens’ electoral choices (for Austria see e. g. Plasser and Ulram, 1995). Like in other West European democracies, societal changes started in the 1960s and the 1970s at the latest. Austria moved from an agricultural and industrial country to a country with a growing service sector and therefore experienced a significant decline in the number of manual workers (Nieuwbeerta and de Graaf, 1999). The clear class division started to vanish with more people working in the service sector and fewer people working in factories and in the farming sector (see Table 4). Table 4: Employment in sectors 1951 – 2001 (in percentage points) Year Agriculture Industry 1951 32.6 37.6 1961 23.0 41.4 1971 13.9 43.2 1981 8.5 41.0 1991 5.8 35.6 2001 3.9 29.5 Source: Statistik Austria.

Service sector 29.8 35.6 42.9 50.5 58.6 66.6

A growing number of well-educated citizens also accelerated the societal changes. For instance, in the 1970s the SPÖ government allowed for free university education and thus provided support for better educated populous. It paid off as the new middle class of young, educated, urban voters became supporters of the SPÖ, and the ÖVP lost its status as the preferred party of whitecollar workers in Austria. Church ties also became weaker. Despite the number of Catholics in Austria remaining quite high, a growing number developed weaker religious attachments. However, each year a considerable number of people left the Catholic Church. Nevertheless, Austrian society remains to this day nearly homogenously Catholic (see Table 5). Of course, other religions can be found as well, such as

23

Party Politics and Electoral Behaviour in Austria

Protestants or Muslims, but these constitute small minorities. However, the share of both Muslims and those who do not belong to any specific religious denomination has been rising continuously. Table 5: Religious denominations in Austria 1971 – 2001 (in percentage points) Year of census 1971 1981 1991 Roman Catholic 87 84 78 Protestant/Lutheran 6 6 5 Muslim/Islamic 0 1 2 Other 1 2 3 no religious denomination 4 6 9 Refused 1 1 3 Total 100 100 100 Source: Statistik Austria, own calculations.

2001 74 5 4 3 12 2 100

The two main parties failed to notice and interpret these changes in the electorate correctly and did not react. Their ingrained belief that electoral losses come as electoral gains for the respective other main party made them blind to the on-going changes in society. As Müller et al. put it “[…] the major parties were mistaken, at least concerning the magnitude of their capacity to structure electoral competition in the 1990s” (2004, p. 153). Not only societal changes left an imprint on the voters, the main parties also contributed with their behaviour. Life itself was structured by political parties and voters’ party affiliation. The Proporz was an important regulation mechanism in public life as well and extended the reach of the party system by far (e. g. Heinisch, 2002; Luther, 1998). For instance, the allocation of positions in the public sector was regulated according to the so-called Proporzregelung: if one position was allocated to a person affiliated with the ÖVP, another position was automatically allocated to a person affiliated with the SPÖ. The entirety of public life was structured according to this principle, thus, providing positions for both Lagers. More and more citizens were dissatisfied with this handling of public positions and were angry and disappointed with the two parties SPÖ and ÖVP (Müller et al., 2004). These developments mark the moment when social structure variables lost a considerable amount of power in explaining party choice in Austria. In their place, citizens’ issues positions – so-called short-term factors – became more important (Müller, 2000). However, the dividing line along socio-economic issues was no longer the only game in town. New policy issues have entered the political discourse since the mid-1980s such as the environment and immigration issues. With Austria joining the European Union in 1995, issues regarding

24

The Austrian Case: an Overview

the European integration process also became important; though these issues were not occupied or owned by the two main parties (Palme, 2000). Therefore, other parties became an attractive option to voters. Due to their environmental policy position, the Greens appealed to both the middle class and the farmers. The blue-collar worker started to discover a ‘new’ home in the FPÖ who promoted welfare state policies only for nationals. Indeed, Müller et al. (2004) point out that as early as 1994 the FPÖ were gaining electoral support amongst so-called modernization-losers, typically found amongst blue-collar workers.8 Party membership eroded, voters became less partisan (e. g. Müller et al., 2004), and thus, the relationship between the party and its voters became weaker, less stable and more volatile. Once the electoral options became more numerous, the number of late-deciders and the degree of volatility increased (Plasser et al., 1992). In particular, the new middle-class was rather ‘unpredictable’ in terms of vote choice. Thus, the duopoly (see Härpfer, 1985) of SPÖ and ÖVP came to an end. In the 1990s, the structural basis of FPÖ support became more obvious. As with other radical-right parties it appealed to young, male voters, but it also gained electoral support amongst low educated voters including low qualified blue-collar workers and low income pensioners, people alienated by the status quo, and susceptible to conservative values and anti-immigration policy positions (Müller et al., 2004; McGann and Kitschelt, 2005). Thus, concerning the socio-structural factors the FPÖ electorate resembled the core electorate of the SPÖ (Aichholzer et al., forthcoming). In the next section we describe the 2008 election in more detail. Since this election provides the core focus for our in-depth analysis of Austrian electoral behaviour, it is important that its features are highlighted.

1.3.3 The 2008 Election: the Context Although, the 2008 elections were called early, they did not come unexpectedly. On the one hand, after several severe defeats of the SPÖ in regional elections, the SPÖ leader and chancellor Alfred Gusenbauer was exposed to scorching criticism. On the other hand, after its electoral defeat in the 2006 election the ÖVP, being the senior partner in the coalition with the FPÖ, was downgraded to the junior partner in the coalition with the SPÖ. In spring 2008, public opinion polls 8 Müller et al. (2004) point out that commentators regard the election in 1994 as the moment when a new cleavage emerged; particularly when considering the electoral changes amongst workers. According to these commentators, the new cleavage distinguishes amongst modernization winners and modernization losers.

Party Politics and Electoral Behaviour in Austria

25

indicated that the ÖVP might surpass the SPÖ in electoral support (Hofer and Týth, 2008). Indeed, in June 2008 polls saw the ÖVP 5 per cent ahead of the SPÖ. Within this context, Chancellor Gusenbauer announced that his minister for infrastructure, Werner Faymann, would replace him as the SPÖ leader. Together, they unexpectedly switched the SPÖ’s hitherto rather positive position on the issue of EU integration to a more sceptical one. Consequently, the ÖVP, a very pro-European party, declared its resignation from the coalition with the SPÖ in July 2008. At this stage the coalition was only in place for one and a half years. It is supposed that the ÖVP took these internal conflicts as an opportunity to call for early elections and become the strongest party in the election that followed. New elections were called for September 28, 2008, and due to the summer months the electoral campaign period was rather short. In essence, only the month of September could be used for electoral campaigning: July and August are holiday months with the National Council closed and political life almost coming to a standstill. For the ÖVP, its party leader and the Vice Chancellor Wilhelm Molterer was the leading candidate, while the SPÖ ran with its newly designated leader Werner Faymann. The FPÖ was led by its party leader HeinzChristian Strache into the electoral campaign (as in 2006), while Jörg Haider was the new front runner for the BZÖ. Alexander van der Bellen continued as the leading candidate for the Greens. Of the parties represented in the National Council not one had a female leading candidate. The 2008 campaign was neither technologically nor substantively innovative. One of the major topics of the electoral campaign was inflation which was comparatively high by Austrian standards. Therefore, the ÖVP propagated budget discipline to control inflation rates and to reduce taxes in general. Meanwhile, the SPÖ presented a programme listing 5 main points that were meant to reduce the consequences of high inflation for the people: a reduction in VAT on food products, an increase in family allowances, an increase in nursing allowances, an abolition of university fees and early retirement plans for workers in heavy industrial sectors. They presented these points to the National Council on September 24, a mere 4 days before the election: four out of these five points passed the vote with floating coalitions, something that is quite unusual in Austria. Only the reduction in VAT was unsuccessful. Thus, important pledges from the electoral programme of the SPÖ were already passed into laws before the elections even took place. The electoral campaigns of both the BZÖ and FPÖ focused on immigration issues though the BZÖ was less negative in tone than the FPÖ. The Greens presented topics such as alternative forms of energy, education, equality and human rights (e. g. Hofer, 2008). In multiparty systems such as Austria a main topic during the electoral campaign are potential coalition agreements. While the SPÖ excluded a coalition

26

The Austrian Case: an Overview

with the two right-wing parties FPÖ and BZÖ, the ÖVP kept a low profile: it did not explicitly exclude a coalition with the FPÖ and the BZÖ, and it was open for a coalition with the Greens. For the Greens both the SPÖ and ÖVP were potential coalition partners, while the FPÖ excluded any collaboration with the two main parties as long as they were headed by the two party leaders Faymann and Molterer. After the election, which resulted in substantial losses for both the SPÖ and ÖVP, a new edition of the grand coalition continued working – this time though with much less seats than in the previous legislative periods. The SPÖ nominated the Chancellor (Werner Faymann) and the ÖVP became the junior partner again, however with a new party leader, Josef Pröll. The Greens also changed their party leader as the electoral result fell short of their expectations. Though the FPÖ gained votes, the biggest surprise was the BZÖ: under the leadership of Jörg Haider the party was able to obtain over 10 per cent of the vote share and to become the fourth strongest political party in the National Council. In the following, we now describe the theoretical considerations that lead our analyses, before presenting in detail our data pool.

1.4

Theoretical Considerations

As mentioned, most of the scholarly work analyses electoral behaviour and its changes in Austria from a descriptive perspective, an important contribution in itself. There are only a few contributions that study the topic from an analytical perspective. Most of this analytical work on Austria focuses on explaining the electoral success of the right-wing party FPÖ but this does not serve as an overarching explanation of electoral behaviour in Austria. In this book we aim to fill this gap in the literature and to provide analyses of electoral behaviour in Austria and its changes over time. We analyse various explanations for these changes in a systematic way. As described in the literature, changes in the social structure of Austrian society, weaker partisan attachments and an increase in issue voting may account for these changes. Our aim is to establish to what degree these changes explain electoral behaviour in the past and the explanatory power they still have today. Thus, we add to the past descriptive analyses of voting behaviour in Austria analytical leverage demonstrating the explanatory power of the different indicators. As voting decisions are complex, the theoretical set-up of the book draws on the various schools of electoral choice to explain party choice in Austria, with special emphasis on the contributions of Lazarsfeld et al. (1960) and the sociological school, and Campbell et al. (1960) and the Michigan school. In principal, we organize the work around the funnel of causality, first introduced

Data and Methodological Approach

27

by Campbell (1960), later adapted by other scholars (Dalton, 2002; Lewis-Beck et al., 2008; Miller and Shanks, 1996). The funnel of causality brings the various factors of vote choice into a conceptual order. We study long-term (Columbia and Michigan school) as well as short-term factors (Rochester school) and divide between sociological and social-psychological theories (long-term) and issue-voting theories (short-term). We start our inquiry with an analysis of the impact of demographics such as age, gender and region (as the size of citizens’ places of residence) on vote choice. Then, we continue with the sociological approach, posing the question which impact social structure has on vote choice and how its impact has changed over time. ‘Social structure’ refers to the two overlapping cleavages in Austria, namely class and religion. As such we include as class indicators education and occupation as explanatory variables, while for religion we look at religious denomination. We continue to look at ideology in terms of left-right positioning and union membership and its impact on party choice, before we conclude by analysing short-term factors such as issue positions on immigration, European integration, state intervention, inequality and environment – having arrived at the end of the funnel of causality. All the empirical chapters are structured as follows: first, we present the theoretical considerations followed by the descriptive findings of former research, then we present our model specification before we finally discuss our results. Finally, we summarize our results and position Austria in comparative research on electoral behaviour.

1.5

Data and Methodological Approach

To conduct our analyses we rely on several data sets. We conduct the overtime analyses using datasets commissioned by political parties between 1986 and 2006. As outlined, the variables of interest are gender, age, region, education, occupation, religion, left-right ideology, union membership and different issue positions. While the variables gender, age, region, education, occupation and religion are available in all datasets, this is not the case for the variables left-right ideology, union membership and issue positions. These latter variables can only be found in the AUTNES dataset capturing the election of 2008 (AUTNES, 2009; Kritzinger et al., 2011). As the funnel of causality leads our theoretical model, we will first and foremost analyse our theoretical models against the data of the most recent election 2008. Thus, the AUTNES-project 2008 dataset constitutes the main source for our analyses. Where possible, we put our 2008 results into perspective with former party choices. We do so by replicating our models on past elections dating back to 1986; at least where the data is available. Conducting overtime analysis is always a rather tricky business. Data col-

28

The Austrian Case: an Overview

lection modes, and question wordings might have changed as may item lists. Our data sets are no exception to this common problem, particularly as different parties interested in different research objectives commissioned the surveys. Above all, the data collection modes are quite different. The datasets we use for elections up to 2006 are ‘exit polls’, conducted on election day all over Austria. Meanwhile, the 2008 survey is a post-election survey conducted several months after the election. Luckily we have almost identical measures for the variables religion and education over time which allows us to construct the same categories for all elections. For education we have five categories running from ‘primary, incomplete secondary’ education to ‘post-secondary, college, university degree’ and for religion we can separate respondents in three categories: the first category includes citizens which are neither part of a church nor attend services; the second category focuses on citizens declaring themselves as either Catholics or Protestants but without frequently attending church services; and the last category includes both declared Catholics and Protestants who are also regular church goers. However, for occupation we do not have similar measures in the various surveys at hand. While in the 2008 data set we are able to assign respondents to an occupation they have participated in at an earlier point of their life, in earlier datasets we cannot. For instance, in 2008 retired people or the unemployed can be linked to their former occupation, and thus, we can capture the influence of occupation over the entire life span of a citizen. Unfortunately, we cannot conduct the same coding for the other data sets as this information is simply not provided. Therefore, we have to keep the distinction between currently employed and currently not employed people in the analysis (which fits with earlier studies, e. g. Luther, 1998). This has several implications for the forthcoming analyses and the interpretations of our results. Comparisons over time become difficult though not impossible. As our aim is to trace changes in the explanatory power of different variables over time, we are mostly interested in studying main tendencies over the period under investigation. Thus, we are less dependent on exact question and item wording. This said, though it is not possible to draw direct comparison between the results resting on data of the 1990 to 2006 elections and the results of the 2008 election, we are still able to capture the overall trend in the explanatory power. However, the reader should be aware that the differences between former datasets and the 2008 dataset exist, and one should refrain from making direct comparisons over time. Concerning the variables of importance at the second (social-psychological) and third step (issue-voting) of the funnel of causality, we do not face these problems as no data for earlier elections is available. Another common problem in data handling is how to deal with the categories

Data and Methodological Approach

29

‘don’t know’ and ‘no answer’. In our case, this is relevant for the variables capturing issue attitudes. Respondents opting for ‘don’t know’ or ‘no answer’ might nevertheless have an opinion on the issue, but do not take the effort to retrieve this information when asked, or choose not to volunteer their real opinion (e. g. Beatty et al., 1998). To avoid a non-randomly distributed loss of many cases, we use the mean value on each of the issue variables to replace missing values based on ‘don’t know’ and ‘no answers’. We do so as, in regression models, this solution can be assumed to have the least impact on the results while keeping the sample size high. An exact list of the variables, on how they have been coded, which means have been used for imputation and how the distribution of missing variables looks like can be found in the Appendix. As we aim to explain vote choice, the variable capturing party choice is our dependent variable. There has been quite a lot of discussion on how to deal with the variable ‘party choice’ in electoral research. For instance, some scholars use the variable vote choice in a stacked data matrix to conduct analysis for all parties simultaneously (van der Eijk et al., 2006); others, however, work with logistic regression models where the electoral fortune of one party or candidate is compared to those of all other parties or candidates (e. g. Lewis-Beck et al., 2012). In this book we follow this latter approach and use logistic regression analyses to capture vote choice dynamics. As we are interested in how party choice for single parties develops in comparison to the context they act in (e. g. the other parties), we believe that this approach provides those meaningful outcomes we try to explore. Our strategy will be to run a series of binomial logistic regressions, where the dependent variables are party choice (one party versus all others), yielding for each election a set of four equations (i. e. Social Democrats v. others, People’s Party v. others, Greens v. others, radical right parties v. others) While for the period between 1986 and 2002 our dependent variable includes four parties represented in the National Council (SPÖ, ÖVP, FPÖ and Greens), in the elections of 2006 and 2008 one would have to include the BZÖ as a fifth party as well. However, we face the problem that in both of these latter elections only a few survey respondents declared openly that they have voted for the BZÖ – a common pattern in radical right voting. Therefore, we grouped the few BZÖ voters together with the FPÖ voters to a category entitled ‘radical right’. Most of the voters included in this category are FPÖ voters, nevertheless, with adding the BZÖ electorate we obtain an even better understanding of the patterns of radical right voting in Austria. We conduct our analyses using the Stata (version 11.2) software package. To obtain a first impression on the indicators of our models in each chapter we first run bivariate analyses between vote choice and the variable of interest. As a poststratification weight is provided for 2008, we use it for these analyses with

30

The Austrian Case: an Overview

descriptive character. We do not use a weight for the other data sets. For our multivariate analysis we always use un-weighted data: the weights in the dataset are not design weights but rather post-stratification weights that balance out the distributions of some socio-demographic variables (e. g. age, gender, region). As we use all these variables as controls in the models, we do not need to use the weights for multivariate analysis.

2 Chapter 2: Region, Age, Gender

2.1

Introduction

Before we start our analyses around the funnel of causality, in this chapter, we elucidate the case of Austria by presenting and analysing those variables which normally are only considered as ‘control variables’. Thus, we start with the impact of demographics on vote choice: we present the factors region, gender and age in more detail and describe their importance for the Austrian context (for a general overview of the potential impact of demographics see Van der Brug and Kritzinger, 2012). This is necessary to provide a comprehensive picture of vote choice in Austria as these indicators are informative in general and help predict voting patterns in particular. We describe the relationship between both age groups and gender and electoral results. To capture regional differences, we present electoral results on the level of municipalities as this is not only the best available data source but also serves our goal to describe the Austrian context. As pointed out in the previous chapter, for our descriptive analyses we use unweighted data for the period 1986 – 2006, but weighted data for the 2008 election (post-stratification weights). Thus, as a reminder, we do not directly compare results from our 2008 analyses with those from earlier elections. Using unweighted data for 1986 – 2006 also implies that we do not compare levels of vote shares across parties, as Greens will be overrepresented in unweighted survey data and the radical right will be underrepresented, but compare vote shares of one party only across different demographic groups.

2.2

Region, Residence and Degree of Urbanization

Lipset and Rokkan (1967) define the centre-periphery-cleavage as the diverging interests of the central nation building culture and the culture of the rural population in the peripheries. Simply put, the urban-rural divide reflects the conflict between the commercial and industrial interests of the urban centre and

32

Region, Age, Gender

traditional rural interests. The underlying question is how to balance the different regional needs. The primary sector, mostly located in the rural areas, had distinctive interests from the urban population buying the products of the primary sector (e. g. Knutsen, 2007). Alongside diverging interests, disagreement over values and attitudes also characterize this cleavage (e. g. Gallagher et al., 2006). Indeed, Knutsen (2007) points out that in many cases the economic regional interests were superposed by the religious dispositions of the rural citizens. Thus, region has an impact on territorial identities and value orientations, though the latter is also affected by other structural variables such as religion and social class (Knutsen, 2010). This can also be observed in Austria, where the urban-rural divide was always considered coterminous to the religious and class divide. For instance, in Austria the degree of urbanization was negatively correlated with the degree of religiosity (Härpfer and Gehmacher, 1984). Therefore, some researchers even go as far as to deny the existence of a rural-urban cleavage in Austria. They state that the rural-urban division is simply the division one can find when looking at social class (e. g. Härpfer and Gehmacher, 1984) and religion. Knutsen (2010) on the other hand, assigns some explanatory power to the regional cleavage in Austria, over and above social class and religion. Though a large part of the regional differences in vote choice can be explained by socio-structural disparities in Austria, there remain some differences which can be traced to the urban-rural divide. In the decades after WWII, urbanized voters supported the SPÖ to a great extent, while most of the rural citizens voted for the ÖVP. This clear division was embedded in a high degree of immobility among Austrian voters. Indeed, Austria had rather little interregional migration (Sauber, 1981). Regional belonging was thus used as an indicator in explaining voting behaviour (Härpfer and Gehmacher, 1984). For instance, while the East of the country, with the capital Vienna and its surroundings, representing the urban centres was mostly populated by SPÖ voters, the opposite is true for the West of Austria: in the rural areas of Austria it was the ÖVP which derived large vote shares. This changed in the 1980s when Austrians became more mobile. The first trends could already be observed in the 1970s when the SPÖ was able to gain vote shares in rural areas, and the ÖVP could attract voters from the new middle class in urbanized areas. Nevertheless, the ÖVP remained the most successful party in the rural areas in the 1990s, and while the SPÖ still gained its highest vote shares in industrialized areas and in larger cities, it faced competition both from the Greens and the FPÖ.

33

Region, Residence and Degree of Urbanization

2.2.1 Size of Municipality and the Vote: 1986 – 2008 To get a grasp on the effect of the urban-rural division in Austria, we analyse the relationship between vote choice and the size of municipality voters live in which we use as an indicator of urbanization. In Table 6 we present electoral results by the size of municipalities.9 As the analysis is not based on survey data, but on electoral results, it is possible to use a more fine grained distinction than would be possible from survey data and to present the results for the small parties as well. Table 6: Region and vote choice 1986 – 2008 (percentage points – based on electoral results)

1986

SPÖ ÖVP FPÖ Greens others

-2,000 35.63 51.68 9.26 3.01 0.42

1990

SPÖ ÖVP FPÖ Greens others

-2,000 36.59 43.51 14.81 4.01 1.07

1994

SPÖ ÖVP FPÖ Greens Liberals others

-2,000 30.88 38.30 21.03 4.95 3.40 1.44

2,001 – 5,000 43.35 40.79 10.95 4.32 0.59 2,001 – 5,000 43.00 33.12 16.81 5.67 1.41 2,001 – 5,000 35.77 29.17 22.39 6.37 4.70 1.61

5,001 – 10,000 46.86 35.62 11.56 4.99 0.97 5,001 – 10,000 46.81 26.70 17.99 6.55 1.94 5,001 – 10,000 39.07 23.35 23.26 7.04 5.50 1.78

10,001 – 50,000 49.44 33.03 11.11 5.31 1.11 10,001 – 50,000 48.45 23.43 18.44 7.15 2.53 10,001 – 50,000 40.09 20.09 24.04 7.52 6.55 1.71

50,001 – 1,000,000 43.12 35.43 13.36 7.10 0.99 50,001 – 1,000,000 42.25 24.26 20.84 10.21 2.43 50,001 – 1,000,000 33.60 19.85 26.25 10.32 8.09 1.89

1,000,000 – 52.35 33.23 5.76 6.75 1.91 1,000,000 – 50.70 21.09 15.73 9.79 2.69 1,000,000 – 38.48 17.14 22.72 9.76 10.08 1.82

9 Absentee voters are excluded from the analysis, as their votes are allocated to the electoral constituency and not to the municipality.

34

Region, Age, Gender

2,001 – 5,001 – 10,001 – 50,001 – 5,000 10,000 50,000 1,000,000 37.96 41.41 42.88 38.15 1995 SPÖ 29.41 24.37 21.47 22.01 ÖVP 22.52 22.82 23.53 24.48 FPÖ 4.24 4.62 4.86 6.45 Greens 4.47 5.29 5.85 7.35 Liberals others 1.40 1.49 1.41 1.56 2,001 – 5,001 – 10,001 – 50,001 – -2,000 5,000 10,000 50,000 1,000,000 29.36 33.42 36.21 37.08 31.62 1999 SPÖ 36.47 28.31 23.40 20.01 19.84 ÖVP 26.12 27.68 27.96 29.53 30.39 FPÖ 4.82 6.37 7.36 7.73 10.54 Greens 2.54 3.03 3.66 5.24 Liberals 1.79 others 1.44 1.68 2.03 2.00 2.36 2,001 – 5,001 – 10,001 – 50,001 – -2,000 5,000 10,000 50,000 1,000,000 31.27 36.23 39.80 41.74 35.58 2002 SPÖ 52.60 44.37 38.71 34.69 36.19 ÖVP 9.68 10.59 10.80 11.91 11.81 FPÖ 5.28 7.39 8.93 9.75 13.97 Greens 0.91 1.08 1.15 1.37 Liberals 0.69 others 0.49 0.52 0.69 0.76 1.09 2,001 – 5,001 – 10,001 – 50,001 – -2,000 5,000 10,000 50,000 1,000,000 31.27 34.93 38.66 39.99 33.63 2006 SPÖ 45.16 36.70 30.20 26.72 27.55 ÖVP 9.75 10.72 11.23 11.37 11.63 FPÖ 6.25 8.94 10.40 11.24 16.59 Greens 3.94 4.56 4.78 5.89 5.87 BZÖ others 3.63 4.16 4.73 4.78 4.72 2,001 – 5,001 – 10,001 – 50,001 – -2,000 5,000 10,000 50,000 1,000,000 25.97 28.83 31.58 32.48 27.38 2008 SPÖ 34.74 27.60 22.68 19.69 19.23 ÖVP 17.00 17.78 17.82 17.75 16.15 FPÖ 5.89 8.23 9.64 10.74 15.95 Greens 11.44 12.11 12.13 13.58 13.46 BZÖ others 4.97 5.46 6.16 5.77 7.83 Note: based on electoral results, own calculations, absentee voters excluded. -2,000 33.05 37.70 21.41 3.34 3.23 1.28

1,000,000 – 43.98 19.55 20.07 6.04 8.63 1.72 1,000,000 – 37.85 16.99 24.77 10.33 6.96 3.10 1,000,000 – 43.77 30.65 7.97 15.14 1.18 1.30 1,000,000 – 41.04 21.83 13.86 17.38 1.82 4.07 1,000,000 – 34.79 16.72 20.43 15.96 4.74 7.35

Looking at electoral results over time and using simple descriptive analyses, we can observe that the trends from the 1980s and 1990s still hold in the most recent elections: small differences according to the size of the region are still evident,

Region, Residence and Degree of Urbanization

35

though the differences are sometimes marginal. In general, in small areas voters are more likely to support the ÖVP, while in bigger areas (50,000 inhabitants and more) the SPÖ has more supporters. Indeed, in 1999, when the ÖVP obtained less votes than the FPÖ and was thus only the third largest party in Austria, the smallest municipalities with less than 2,000 inhabitants were the only areas where the ÖVP was still more successful than the FPÖ. This supports the proposition that the ÖVP draws its core voters from rural areas and is by far the strongest party in these small municipalities. For the Greens, the bigger the city, the higher the vote share. In the three most recent elections (2002, 2006 and 2008), the vote share for the Greens was about three times larger in Vienna than it was in small municipalities with less than 2,000 inhabitants. The FPÖ is known to be the least regionalized party of all Austrian political parties (Sommer, 1996). Recently, FPÖ supporters are more or less evenly distributed across the different regional sizes, with slightly better results in the middle categories. This phenomenon can be observed both in the 1999 election, when the FPÖ gained 27 per cent of the vote share, and also in the 2002 election, when the FPÖ had to face dramatic losses and gained only 10 per cent of the vote share. To sum up, the distinction between the SPÖ and ÖVP on the rural-urban divide is still visible, though it diminished over time. Other parties exhibit different success rates depending on the size of the area. While the Greens find their voters in larger cities and areas surrounding these cities and are not successful in the rural areas, the FPÖ is able to find voters across all regional categories. Thus, we have to some extent and for certain parties an indication that a rural-urban divide exists, yet, this is not the case for the FPÖ.

2.2.2 Electoral Constituencies and the Vote in 2008 There are several other strategies to define regional distinctions. Besides using the size of a municipality, we can also use geographic units that are defined by the electoral system (e. g. Knutsen, 2010). Since the electoral reform of 1992, Austria has 9 regions (Länder) divided into 43 electoral constituencies. The city of Vienna, where nearly one fifth of eligible voters live, is itself divided in seven constituencies. The electoral constituencies do not necessarily reflect an urbanrural divide as some constituencies hold municipalities of very different sizes. Of course, some constituencies are definitely urban (as the constituencies within Vienna, but also the constituencies of the cities of Graz, Salzburg and Innsbruck), and some are mostly urban, as, for instance, the constituency of Linz and its surrounding area. Others are mainly rural with no larger municipality, while

36

Region, Age, Gender

most combine medium sized towns and villages. One could argue that constituencies reflect territorial identities and economic structures.

Figure 1. SPÖ vote share in regional constituencies

Figure 2. ÖVP vote share in regional constituencies

Figures 1 to 5 show the vote shares of the parties represented in the National Council in the different electoral constituencies in Austria in 2008, with the

Region, Residence and Degree of Urbanization

37

constituencies of Vienna being displayed separately on the top left of the three Figures. The differences in vote shares in Vienna are particularly remarkable and reflect nicely the socio-economic differences of its inhabitants by constituency (see also Chapter 3). It must be noted, however, that the 2008 election was unique regarding regional considerations as Jörg Haider, the popular frontrunner of the BZÖ, was the ‘beloved’ governor of the region of Carinthia, this being partly responsible for the major regional differences we observe in 2008. In Figures 1 and 2 we look at the two governing parties, the ÖVP and SPÖ. For the SPÖ, the 2008 election illustrates the same pattern as in former elections: there is a clear East-West divide with very low vote shares in the Western constituencies of the country and higher shares in the East, most notably in the constituencies of the region of Burgenland (on the Eastern border to Hungary), in the upper Styrian constituencies, where an economic cluster of industry and mining was historically linked to high vote shares for the Social Democrats, in the city of Linz, and in most parts of Vienna – however, not in other large cities such as Graz or Salzburg. The ÖVP (Figure 2), in contrast, is better characterized by high vote shares in very rural constituencies (as Waldviertel, Mostviertel or Tiroler Oberland). It is generally weak in the large cities, in particular in Vienna, and in the entire region of Carinthia.

Figure 3. FPÖ vote share in regional constituencies

The FPÖ can be characterized as a party that is more or less equally successful across most regions in Austria, as Figure 3 reveals. Thus, the 2008 results confirm

38

Region, Age, Gender

Figure 4. BZÖ vote share in regional constituencies

Figure 5. Green vote share in regional constituencies

the trend that had already been identified in the 1990s (Sommer, 1996) – with the exception of Carinthia. The FPÖ acquired between 15 and 20 per cent of the vote share in most constituencies. As a result of the former FPÖ leader Jörg Haider being governor of Carinthia and at the same time the frontrunner for the BZÖ in the 2008 campaign, the FPÖ is least successful (less than 15 per cent) in all

39

Gender

constituencies of Carinthia. On the other hand, the FPÖ was most successful in some districts of Vienna. Figures 4 and 5 present the vote shares of the smaller opposition parties. Obviously, there is a North-South gap for the BZÖ: In the Carinthian constituencies in the South of the country, the BZÖ gained nearly 40 per cent of the votes. In contrast, it was very weak in the Northern and Eastern part of the country, particularly in most districts of Vienna where the BZÖ gained less than 5 per cent of the votes. Lastly, the Greens are generally stronger in the larger cities (Vienna, Graz, Linz, Salzburg and Innsbruck) and the constituencies surrounding these cities (Figure 5). The Greens gained more than 20 per cent of the votes in the centre of Vienna. Outside the larger cities the Greens gained a high percentage of the vote share (between 15 and 20 per cent) only in the two constituencies of Vorarlberg, in the very West of the country. In all other constituencies the Greens have only weak support (between 5 to 10 per cent) and almost no support in some constituencies of Carinthia and Styria.

2.3

Gender

Early studies on electoral behaviour revealed that gender also predicts vote choice. Women tended to vote more conservatively and, thus, for parties on the right. The explanation for this is that women are more family oriented and have a higher degree of religiosity, both of which predict voting for right parties (e. g. Harrop and Miller, 1987). In the 1970s the first changes in the impact of gender on vote choice can be observed. Women now tended to vote more for left parties, so the predictive power of this variable changed. Reasons for these developments could be found in the changing role of women in society : unlike in former times, women now had a more important role in the workforce which made them less conservative and their religious views changed into more liberal views (e. g. Howell and Day, 2000; Inglehart and Norris, 2003; O’Neill, 2001; Vaus & McAllister, 1989; Wagner and Kritzinger, 2012). Discussing the reasons for a gender gap in electoral behaviour more recently, Box-Steffenmeister et al. (2004) offer different theories. They argue that women are assumed to be economically more vulnerable, and thus more likely to vote for parties that support the welfare state. They find that changes in the gender gap can be explained by “the interaction of societal conditions and politics” (Box-Steffensmeier et al., 2004, p. 527). The rise of the radical right in a number of Western democracies also raised questions about the reasons for the gender gap observed in its voting patterns: women were less likely to vote for radical

40

Region, Age, Gender

right parties (e. g. Betz, 1994; Gidengil et al., 2005). Gidengil et al. (2004) state that situational and structural factors cannot explain the gender gap. Indeed, the gender gap disappears when values and beliefs are taken into account. However, in most countries the gender gap seems to be much less relevant in explaining party choice than are social class and religious affiliations (Brooks et al., 2006), with Austria not being an exception. Some differences between the genders can be observed in Austria in the past: as in other countries, women also used to vote more conservatively. In the 1920s and 1930s, when some electoral results could be analysed separately for men and women due to differently coloured ballot papers, women were more likely to vote for the Christian Democrats, while men were more likely to support the Social Democrats and other parties, including Communists and radical-right parties (Hofinger and Ogris, 1996). After WWII, data from the 1962 election reveal that the ÖVP electorate is comprised of 55 per cent women and 45 per cent men, while within the SPÖ electorate the opposite can be observed: 47 per cent support among females versus 53 per cent among males (Kienzl, 1964). Analyses of the 1970s demonstrate a different picture. As Härpfer and Gehmacher (1984) show, 54 per cent of the SPÖ electorate was female and 46 per cent was male, but the same picture as previous elections still applied to the ÖVP (57 per cent female, 43 per cent male). The FPÖ already showed signs of being a male dominated party, though in the mid-1980s gender differences could no longer be observed. In the 1995 election, women were more supportive of the left parties SPÖ and the Greens, while the gender gap was particularly visible for the FPÖ with more men supporting it (Hofinger and Ogris, 1996). These gender differences stood in contrast to the European trend. This can be explained by the rise of the FPÖ which was more attractive to men than to women, and the emergence of new leftish and liberal parties, such as the Greens, which women found more appealing (Hofinger and Ogris, 1996). Most recently, Wagner and Kritzinger (2012) showed that young women in Austria are very likely to support the Greens but could not detect any other gender differences.

2.3.1 Gender and the Vote 1986 – 2006 Analysing the overall gender gap in vote choice models for the period of 1986 – 2006 (Table 7), our data reveal that indeed more women than men voted for the SPÖ, and only in 2006 was this no longer the case. For the ÖVP the vote share of men and women was nearly balanced in the period of interest. As anticipated, men found the FPÖ more attractive than women, which is particularly visible in the elections where the FPÖ was most successful. In 2002 and 2006, when the

41

Gender

FPÖ was comparatively weak, the gender gap is less visible. Gender is rather balanced in 2006 even for the Greens, although they attracted more women than men in most elections. To sum up, the gender gap has never been particularly strong in Austria in the period 1986 – 2006. It was highest in the 1990s when the radical right FPÖ attracted more men than women, but it seems to have been diminished in the 2000s and is hardly observable in the 2006 election.

Table 7: Gender and vote choice 1986 – 2006 (percentage points) Women 1986 SPÖ 40 ÖVP 44 FPÖ 6 Green 10 Total 100 N 2037 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.09 Women 1990 SPÖ 46 ÖVP 33 FPÖ 12 Green 9 Total 100 N 2093 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.10 Women 1994 SPÖ 37 ÖVP 32 FPÖ 16 Green 15 Total 100 N 1968 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.14

Men 40 41 11 8 100

Men 39 34 19 8 100

Men 33 30 27 10 100

42

Region, Age, Gender

Women 1995 SPÖ 41 ÖVP 34 FPÖ 15 Green 10 Total 100 N 2074 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.12 Women 1999 SPÖ 35 ÖVP 29 FPÖ 20 Green 16 Total 100 N 1932 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.14 Women 2002 SPÖ 40 ÖVP 36 FPÖ 8 Green 16 Total 100 N 2108 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.11 Women 2006 SPÖ 35 ÖVP 38 FPÖ+BZÖ 9 Green 18 Total 100 N 1877 P-Value Pearson’s w2 0.00 Cram¦r’s V 0.07 Note: unweighted data.

2.4

Men 36 34 23 7 100

Men 32 29 29 10 100

Men 33 41 13 13 100

Men 33 37 13 17 100

Age

As with gender, age is often considered an important explanatory factor in party choice models (for an overview see Jennings and Niemi, 1981; Van der Brug and Kritzinger, 2012). Voters are subject to different societal influences over their life span which might impact their political behaviour. In general, the literature distinguishes between cohort, life-cycle and period effects (e. g. Goerres, 2009). Cohort effects point to the possibility that individuals who went through the same processes

Age

43

of socialization and habituation in their youth (e. g. Butler and Stokes, 1974; Campbell et al., 1960) may behave differently than individuals whose experiences stem from a different point in time. The political orientation will remain rather stable over time (Goerres, 2009). Life-cycle effects mean that voters’ attitudes and decisions are influenced by the stage of life the individual is going through. For instance, needs, interests and demands of an older, retired person might be different from those of a younger citizen. This is then reflected in different ideological orientations and then, ultimately, in the voting decisions of the individual (e. g. Konzelmann et al., 2012). Finally, period effects are those which result from extraordinary events which affect all individuals in the same manner (e. g. Lutz et al., 2006). Though we do not specifically test for period, cohort or life-cycle effects in this book, it is important to point out these differences as they might be helpful in the interpretation of potential age group differences amongst parties in Austria. Apart from these effects, scholars have also revealed other explanations for age differences in electoral behaviour. For instance, van der Brug (2010) has shown that for older voters structural voting is important, while young voters put more emphasis on ideological voting. In Austria, generational divisions did not play an important role until the 1980s (Härpfer and Gehmacher, 1984). As early as the 1970s it was obvious that young voters did not hold a strong preference for any of the established parties, and in the 1980s they oriented themselves towards the new political force, namely the Greens. The increase in voters’ volatility could thus already be sensed. Wagner and Kritzinger (2012) found for the European Parliament election of 2009 that the overall age group differences are rather small in Austria. However, there are some generational differences which explain the relationship between issues attitudes and vote choice. 2.4.1 Age and the Vote 1986 – 2006 The results stemming from the data covering the period from 1986 to 2006 show that both the SPÖ and the ÖVP had always been more successful among older voters, while the FPÖ and the Greens gained the most votes from younger age groups (see Table 8). Differences due to age are mostly visible for the Greens who have hardly ever gained more than 5 per cent of their vote share among voters aged 60 years and older. For each period under examination, the share of Green votes decreased with age. Generational replacement might change this in the long-run. The 18 – 29 year olds of 1986 were 38 – 49 in 2006, and the Greens are growing stronger and stronger in the middle age groups. The tendency is not so clear for the FPÖ. Only in their most successful election in 1999 can we state that their success among the youngest cohorts is obvious. Again, we need to remind the reader of the peculiarities of our data sets to be able to put the results into context. We use unweighted data for our descriptive analysis:

44

Region, Age, Gender

this is why Green voters are overrepresented while voters of the FPÖ are underrepresented.10 As we compare across age-groups for each party and not between parties, this does thus not affect our findings. Table 8: Age and vote choice 1990 – 2006 (percentage points)

1986

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

1991

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

1994

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

1995

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

18 – 29 35 34 11 20 100 2037 0.00 0.16 18 – 29 37 27 18 18 100 2093 0.00 0.14 18 – 29 32 23 24 21 100 1968 0.00 0.14 18 – 29 33 25 23 19 100 2074 0.00 0.15

30 – 39 37 40 11 12 100

40 – 49 44 43 8 5 100

50 – 59 40 53 5 2 100

60 – 69 45 46 8 1 100

70+ 45 44 9 2 100

30 – 39 39 34 13 14 100

40 – 49 42 37 15 6 100

50 – 59 45 36 16 3 100

60 – 69 52 34 13 1 100

70+ 42 37 19 2 100

30 – 39 33 27 21 19 100

40 – 49 33 34 20 13 100

50 – 59 32 37 24 7 100

60 – 69 44 37 16 3 100

70+ 37 34 27 2 100

30 – 39 38 30 20 12 100

40 – 49 38 37 17 8 100

50 – 59 38 39 19 4 100

60 – 69 44 42 13 1 100

70+ 47 37 15 1 100

10 Therefore, we present very high vote shares for the Greens amongst younger age groups and in more recent elections in Table 3a.

45

Region-Age-Gender and Vote Choice in Austria: Multivariate Analysis

1999

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

2002

SPÖ ÖVP FPÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V

SPÖ ÖVP FPÖ and BZÖ Green Total N P-Value Pearson’s w2 Cram¦r’s V Note: unweighted data.

2006

2.5

18 – 29 25 19 34 22 100 1932 0.00 0.16 18 – 29 28 28 13 31 100 2108 0.00 0.17 18 – 29 27 29 10 34 100 1877 0.00 0.16

30 – 39 30 25 27 18 100

40 – 49 35 31 23 11 100

50 – 59 37 38 16 9 100

60 – 69 40 39 19 2 100

70+ 39 35 22 4 100

30 – 39 36 38 11 15 100

40 – 49 36 40 12 12 100

50 – 59 37 42 10 11 100

60 – 69 45 44 7 4 100

70+ 43 49 7 1 100

30 – 39 25 37 13 25 100

40 – 49 31 35 9 25 100

50 – 59 39 35 9 17 100

60 – 69 39 42 12 7 100

70+ 42 45 11 2 100

Region-Age-Gender and Vote Choice in Austria: Multivariate Analysis

We now analyse how the vote choice of the Austrian voter can be explained by simply using region, gender and age as independent variables. This simple model thus assumes: Party choice = a + b1age + b2gender + b3region +e

All multivariate models are run with unweighted data in Stata.

2.5.1 The period 1990 – 2006 For the period 1990 – 2006, we only include those years where all three variables are available, i. e. 1990, 1994, 1999 and 2002. In all, the models do explain some

46

Region, Age, Gender

variation in vote choice but only to a small extent (see Tables 9 and 10). The Pseudo R2 s, which indicate goodness of fit, are generally very low, with the highest values associated with those models where the vote choice for the Greens is analysed. For the two mainstream parties, the SPÖ and ÖVP, age has a consistent positive effect on the chances to vote for one of these parties, while the effect was consistently negative for the FPÖ and the Greens. The age gap found in the bivariate analysis is thus also apparent in the multivariate models. Although the gender gap is not always significant, we can generally say that in Austria in the period between 1990 and 2006, men were more likely to vote for the FPÖ and, to a smaller extent, for the ÖVP, while women were more likely to vote for the Greens and, again to a smaller extent, for the SPÖ. The parties on the right were thus more attractive to men, while the parties on the left attracted more women. Regional differences have the greatest impact on voting for the ÖVP and the Greens. The results reveal that in more populated areas voters are less likely to vote for the ÖVP. The results for the SPÖ are more nuanced: in 1999, region had no significant influence at all, and for the other elections we see very few significant differences: Voters are more likely to support the SPÖ if they live in areas with 2,000 to 50,000 inhabitants than if they live in areas up to 2,000 inhabitants. Most surprisingly, no significant differences can be found between the largest urban area (more than 1 million inhabitants) and the smallest area. This is rather surprising as for decades the SPÖ had its stronghold in the region of Vienna which falls into this category. For the Greens the regional effects are quite strong: for all categories bigger than the smallest reference group we see significant positive effects. In 1990 only the two largest categories show positive effects but later in our observation period all three categories feature significant positive effects. Unlike the SPÖ, for the Greens we find significant effects in the largest urban area. Finally, for the FPÖ we find nearly no regional effects, confirming the assumption that the FPÖ is the most regionally homogenous party. Only in 1999 and 2002, for the category 50,000 to 1 million, do we find significant effects, indicating that voters from medium sized towns and cities were more likely to vote for the FPÖ than voters from small villages. To sum up, for both the ÖVP and the Greens region, namely population size, matters. The ÖVP obtains more support in rural areas while the Greens are more successful in urban areas. These trends are stable over time.

47

Region-Age-Gender and Vote Choice in Austria: Multivariate Analysis

Table 9: Multivariate logistic regression: age, gender and region: SPÖ and ÖVP

1990 0.94*** -0.26**

SPÖ 1994 1999 1.10*** 1.43*** -0.14 -0.08

Age Gender Region: reference group = max. 2,000 inhabitants 2,–50,000 0.28* 0.37** 0.17 inhabitants 0.25 0.36* -0.03 50,000 – 1,000,000 inhabitants more than 0.28* 0.04 0.29 1,000,000 Constant -0.67*** -1.24*** -1.35*** N 2093 2160 2146 Pseudo R2 0.01 0.01 0.01 Note: *p< 0.05, **p