Distribution:
Johan Martinsson
Department of Political Science
University of Gothenburg
Box 711 y 405 30 Gothenburg y Sweden
E-mail: johan.martinsson@pol.gu.se
© 2009 Johan Martinsson
Printed by Geson, Göteborg, 2009.
ISBN: 978-91-89246-41-6 y ISSN: 0346-5942
EAN: 9789189246416
This dissertation is included as number 115 in the series
Gothenburg Studies in Politics, edited by Bo Rothstein,
Department of Political Science, University of Gothenburg.
Distribution:
Johan Martinsson
Department of Political Science
University of Gothenburg
Box 711 y 405 30 Gothenburg y Sweden
E-mail: johan.martinsson@pol.gu.se
© 2009 Johan Martinsson
Printed by Geson, Göteborg, 2009.
ISBN: 978-91-89246-41-6 y ISSN: 0346-5942
EAN: 9789189246416
Contents
1.INTRODUCTION 1
THE INTEGRATED MODEL OF ECONOMIC VOTING AND ISSUE OWNERSHIP 7
WHAT NEEDS TO BE DONE 16
2.THE CASE OF SWEDEN 27
THE CONTEXT OF ECONOMIC VOTING 27
THE ECONOMY AND THE LABOUR MARKET 32
3.ECONOMIC VOTING IN SWEDEN 41
THE VOTE FUNCTION – THE ECONOMY AND ELECTORAL RESULTS 43
THE POPULARITY FUNCTION – THE ECONOMY AND THE POLLS 55
CONCLUSIONS 65
4.THE ECONOMY AND THE PUBLIC 67
ECONOMIC PERFORMANCE AND PUBLIC EVALUATIONS 80
THE DEVELOPMENT OF UNEMPLOYMENT AND THE AGENDA 99
CONCLUSIONS 106
5.ISSUE OWNERSHIP 109
THE FIRST CRITERION – BEING ASSOCIATED WITH THE ISSUE 124
THE SECOND CRITERION – BEING PERCEIVED AS COMPETENT 128
DETERMINANTS OF PARTY COMPETENCE 135
CONCLUSIONS 145
6.EXPERIENCING UNEMPLOYMENT 149
EFFECTS ON ISSUE SALIENCE OF UNEMPLOYMENT 151
EFFECTS ON ECONOMIC EVALUATIONS 157
EFFECTS ON LEFT-RIGHT IDEOLOGY 163
EFFECTS ON PARTY SYMPATHY AND VOTING BEHAVIOUR 167
CONCLUSIONS 173
Contents
1.INTRODUCTION 1
THE INTEGRATED MODEL OF ECONOMIC VOTING AND ISSUE OWNERSHIP 7
WHAT NEEDS TO BE DONE 16
2.THE CASE OF SWEDEN 27
THE CONTEXT OF ECONOMIC VOTING 27
THE ECONOMY AND THE LABOUR MARKET 32
3.ECONOMIC VOTING IN SWEDEN 41
THE VOTE FUNCTION – THE ECONOMY AND ELECTORAL RESULTS 43
THE POPULARITY FUNCTION – THE ECONOMY AND THE POLLS 55
CONCLUSIONS 65
4.THE ECONOMY AND THE PUBLIC 67
ECONOMIC PERFORMANCE AND PUBLIC EVALUATIONS 80
THE DEVELOPMENT OF UNEMPLOYMENT AND THE AGENDA 99
CONCLUSIONS 106
5.ISSUE OWNERSHIP 109
THE FIRST CRITERION – BEING ASSOCIATED WITH THE ISSUE 124
THE SECOND CRITERION – BEING PERCEIVED AS COMPETENT 128
DETERMINANTS OF PARTY COMPETENCE 135
CONCLUSIONS 145
6.EXPERIENCING UNEMPLOYMENT 149
EFFECTS ON ISSUE SALIENCE OF UNEMPLOYMENT 151
EFFECTS ON ECONOMIC EVALUATIONS 157
EFFECTS ON LEFT-RIGHT IDEOLOGY 163
EFFECTS ON PARTY SYMPATHY AND VOTING BEHAVIOUR 167
7.ELECTORAL CONSEQUENCES 175
TESTING ISSUE OWNERSHIP OF UNEMPLOYMENT 182
EFFECTS ON GOVERNMENT SUPPORT 183
IMPLICATIONS FOR ELECTION OUTCOMES 187
CONCLUSIONS 193
8.THE 2006 ELECTION 195
PUBLIC OPINION AND ECONOMIC CHANGES 197
ISSUE OWNERSHIP 202
VOTING 207
CONCLUSIONS 215
9.CONCLUSIONS 217
TWO SIMULTANEOUSLY OPERATING MECHANISMS 218
THE TARDINESS OF POLITICS AND THE EXISTENCE OF ISSUE OWNERSHIP 219 MECHANISMS THAT COUNTERACT OR REINFORCE EACH OTHER 220 CONSEQUENCES FOR UNDERSTANDING ELECTION OUTCOMES 221
WHEN DOES ISSUE OWNERSHIP MATTER FOR ECONOMIC VOTING? 222
LIMITATIONS AND GENERALITY 225
IMPLICATIONS FOR ECONOMIC VOTING THEORY 230
IMPLICATIONS FOR ISSUE OWNERSHIP THEORY 231
APPENDIX A 235
APPENDIX B 269
REFERENCES 273
7.ELECTORAL CONSEQUENCES 175
TESTING ISSUE OWNERSHIP OF UNEMPLOYMENT 182
EFFECTS ON GOVERNMENT SUPPORT 183
IMPLICATIONS FOR ELECTION OUTCOMES 187
CONCLUSIONS 193
8.THE 2006 ELECTION 195
PUBLIC OPINION AND ECONOMIC CHANGES 197
ISSUE OWNERSHIP 202
VOTING 207
CONCLUSIONS 215
9.CONCLUSIONS 217
TWO SIMULTANEOUSLY OPERATING MECHANISMS 218
THE TARDINESS OF POLITICS AND THE EXISTENCE OF ISSUE OWNERSHIP 219 MECHANISMS THAT COUNTERACT OR REINFORCE EACH OTHER 220 CONSEQUENCES FOR UNDERSTANDING ELECTION OUTCOMES 221
WHEN DOES ISSUE OWNERSHIP MATTER FOR ECONOMIC VOTING? 222
LIMITATIONS AND GENERALITY 225
IMPLICATIONS FOR ECONOMIC VOTING THEORY 230
IMPLICATIONS FOR ISSUE OWNERSHIP THEORY 231
APPENDIX A 235
APPENDIX B 269
Tables
Table 3.1 Specific macro economic results for different governments in Sweden 1973-2002 45
Table 3.2 General macro economic results for different governments in Sweden 1973-2002 47
Table 3.3 Change in comparative economic performance of Swedish governments 1973-2002 (deviation change)
48
Table 3.4 Correlations between change in incumbent electoral support and economic indicators 1973-2002
52
Table 3.5 Correlations between government popularity and various economic indicators 61
Table 3.6 Economic effects on incumbent support 1967-2002 63
Table 3.7 Economic effects on incumbent support for different periods in Sweden 1967-2002 64
Table 4.1 Error correction model of unemployment expectations 92
Table 5.1 Party profiles over time. Share of voters saying parties emphasized the specified issues during that year’s election campaign (percent)
126
Table 5.2 Perceived party competence concerning unemployment 1982-2002 (percent) 129
Table 5.3 Fluctuations in perceived party competence 1985-2002 (AAD) 131
Table 5.4 Does performance influence unemployment party competence? 143
Table 6.1 Change in salience of unemployment among unemployed and others from one election to the next 1979-2002 (percent)
155
Table 6.2 Differences in prospective and retrospective personal economic evaluations between unemployed and others (percentage point differences)
158
Table 6.3 Differences in prospective and retrospective national economic evaluations between unemployed and others (percentage point differences)
161
Table 6.4 Government vote transitions from one election to the next among unemployed and others 1979-2002
172
Table 7.1 Share of Social Democrats by unemployment salience (1988-2002, percent) 182
Table 7.2 Models of government support 1988-2002 (binary logistic regression) 185
Table 7.3 Exploring the electoral consequences of simultaneous changes in economic evaluations and issue salience of unemployment (change in incumbent vote share)
191
Table 8.1 Share of voters saying that different parties emphasized the issue of unemployment during the election campaign in 2006 (percent)
205
Table 8.2 Effects on government support (binary logistic regression) 210
Table 8.3 Exploring the consequences of simultaneous changes in economic evaluations and issue salience of unemployment (change in incumbent vote share)
213
Table 9.1 Potential effects of issue ownership for different issues in 2006 227
Table 9.2 Potential effects of issue ownership of unemployment in different countries 229
Tables
Table 3.1 Specific macro economic results for different governments in Sweden 1973-2002 45
Table 3.2 General macro economic results for different governments in Sweden 1973-2002 47
Table 3.3 Change in comparative economic performance of Swedish governments 1973-2002 (deviation change)
48
Table 3.4 Correlations between change in incumbent electoral support and economic indicators 1973-2002
52
Table 3.5 Correlations between government popularity and various economic indicators 61
Table 3.6 Economic effects on incumbent support 1967-2002 63
Table 3.7 Economic effects on incumbent support for different periods in Sweden 1967-2002 64
Table 4.1 Error correction model of unemployment expectations 92
Table 5.1 Party profiles over time. Share of voters saying parties emphasized the specified issues during that year’s election campaign (percent)
126
Table 5.2 Perceived party competence concerning unemployment 1982-2002 (percent) 129
Table 5.3 Fluctuations in perceived party competence 1985-2002 (AAD) 131
Table 5.4 Does performance influence unemployment party competence? 143
Table 6.1 Change in salience of unemployment among unemployed and others from one election to the next 1979-2002 (percent)
155
Table 6.2 Differences in prospective and retrospective personal economic evaluations between unemployed and others (percentage point differences)
158
Table 6.3 Differences in prospective and retrospective national economic evaluations between unemployed and others (percentage point differences)
161
Table 6.4 Government vote transitions from one election to the next among unemployed and others 1979-2002
172
Table 7.1 Share of Social Democrats by unemployment salience (1988-2002, percent) 182
Table 7.2 Models of government support 1988-2002 (binary logistic regression) 185
Table 7.3 Exploring the electoral consequences of simultaneous changes in economic evaluations and issue salience of unemployment (change in incumbent vote share)
191
Table 8.1 Share of voters saying that different parties emphasized the issue of unemployment during the election campaign in 2006 (percent)
205
Table 8.2 Effects on government support (binary logistic regression) 210
Table 8.3 Exploring the consequences of simultaneous changes in economic evaluations and issue salience of unemployment (change in incumbent vote share)
213
Table 9.1 Potential effects of issue ownership for different issues in 2006 227
Figures
Figure 1.1 Two pathways of influence for economic changes – the integrated model of economic voting and issue ownership
8
Figure 1.2 Consequences of economic development and issue ownership for incumbent support according to the integrated model of economic voting and issue ownership
10
Figure 2.1 The Swedish economy 1970-2002. Some important macro economic indicators (percent)
33
Figure 2.2 The Swedish economy 1970-2002 in relation to European Union means (deviations, percentage points)
34
Figure 2.3 Unemployment in Sweden 1980-2002 (percent) 36
Figure 3.1 Government popularity 1967-2002 (percent) 56
Figure 3.2 Scatterplots of government popularity and various economic indicators 59
Figure 4.1 Retrospective economic evaluations and actual economic development 1993-2002 81
Figure 4.2 Prospective and retrospective economic evaluations 84
Figure 4.3 General economic expectations and current economic situation 1976-2002 85
Figure 4.4 Perspectives on unemployment in Sweden 1976-2002 89
Figure 4.5 Expectations and unemployment change 1976-2002 90
Figure 4.6 Real inflation and perceived inflation 1979-2002 94
Figure 4.7 Perceived prospective inflation and actual future inflation 97
Figure 4.8 Salience of unemployment and actual unemployment levels 1987-2002 (percent) 100
Figure 4.9 TV news coverage of unemployment 1985-2002 103
Figure 4.10 Issue salience, media and unemployment 1987-2002 104
Figure 5.1 Issue ownership’s dependency on government performance. Four alternative views. 117
Figure 5.2 Parties emphasising unemployment. Share of voters saying the party emphasized unemployment during that years’ election campaign (percent)
125
Figure 5.3 Perceived party competence concerning unemployment for three parties 1984-2002 (percent)
133
Figure 5.4 General party competence and competence specific to unemployment for the Social Democratic party (z scores)
141
Figure 5.5 Unemployment, party support and issue competence of Social Democrats 1984-2002 142
Figure 6.1 Salience of the issue of unemployment among unemployed and non-unemployed 1987-2002 (percent)
153
Figure 6.2 Left-right ideological self placement among unemployed and others 1986-2004 (mean 1-5)
165
Figure 6.3 Differences in party sympathy between unemployed and others 1986-2004 (-10 to +10)
169
Figure 7.1 A theoretical model of the effects of unemployment on voting 180
Figure 8.1 Share of voters saying that unemployment is an important issue for party choice 2002-2006 (percent)
199
Figure 8.2 Issue salience and unemployment level 1986-2006 (percent) 200
Figures
Figure 1.1 Two pathways of influence for economic changes – the integrated model of economic voting and issue ownership
8
Figure 1.2 Consequences of economic development and issue ownership for incumbent support according to the integrated model of economic voting and issue ownership
10
Figure 2.1 The Swedish economy 1970-2002. Some important macro economic indicators (percent)
33
Figure 2.2 The Swedish economy 1970-2002 in relation to European Union means (deviations, percentage points)
34
Figure 2.3 Unemployment in Sweden 1980-2002 (percent) 36
Figure 3.1 Government popularity 1967-2002 (percent) 56
Figure 3.2 Scatterplots of government popularity and various economic indicators 59
Figure 4.1 Retrospective economic evaluations and actual economic development 1993-2002 81
Figure 4.2 Prospective and retrospective economic evaluations 84
Figure 4.3 General economic expectations and current economic situation 1976-2002 85
Figure 4.4 Perspectives on unemployment in Sweden 1976-2002 89
Figure 4.5 Expectations and unemployment change 1976-2002 90
Figure 4.6 Real inflation and perceived inflation 1979-2002 94
Figure 4.7 Perceived prospective inflation and actual future inflation 97
Figure 4.8 Salience of unemployment and actual unemployment levels 1987-2002 (percent) 100
Figure 4.9 TV news coverage of unemployment 1985-2002 103
Figure 4.10 Issue salience, media and unemployment 1987-2002 104
Figure 5.1 Issue ownership’s dependency on government performance. Four alternative views. 117
Figure 5.2 Parties emphasising unemployment. Share of voters saying the party emphasized unemployment during that years’ election campaign (percent)
125
Figure 5.3 Perceived party competence concerning unemployment for three parties 1984-2002 (percent)
133
Figure 5.4 General party competence and competence specific to unemployment for the Social Democratic party (z scores)
141
Figure 5.5 Unemployment, party support and issue competence of Social Democrats 1984-2002 142
Figure 6.1 Salience of the issue of unemployment among unemployed and non-unemployed 1987-2002 (percent)
153
Figure 6.2 Left-right ideological self placement among unemployed and others 1986-2004 (mean 1-5)
165
Figure 6.3 Differences in party sympathy between unemployed and others 1986-2004 (-10 to +10)
169
Figure 7.1 A theoretical model of the effects of unemployment on voting 180
Figure 8.1 Share of voters saying that unemployment is an important issue for party choice 2002-2006 (percent)
199
Acknowledgements
During the latter half of the decade that I have spent as a PhD student, I have come to realize that my goal never was to complete a PhD thesis. Rather, when I applied to the PhD program in the end of the 1990s, I just wanted to become a PhD student. That was it.
Ten years older, I am very grateful that I knew so little about what to expect when I started out. Countless are the hours and the evenings I have spent in my office during this time. But in no way do I feel this was a waste. Because countless are also all the beautiful sunsets I have briefly witnessed in passing from within the Department of Political Science at Sprängkullsgatan. Just think about what I would have missed if I had not been working late so many days in the past years.
First and foremost I would like to thank my supervisors during the writing of this thesis. My main supervisor, Sören Holmberg, has been very important dur-ing the whole project and has always supported me in many ways. Sören, I am very grateful for your encouragement, inspiration and support ever since I started on my Master thesis in 1998. Although I did not always immediately agree with your recommendations, thanks for not letting me get bogged down with the details. And my second supervisor, Henrik Oscarsson, deserves special thanks for his careful reading of my drafts and for good judgement, as well as for being available for all sorts of questions. No topic too small or unimportant. Many thanks also to Bo Rothstein and Maria Oskarson for luring me into the project I started in together with Helena Rohdén: Arbetsmarknadens politik.
Financial support is also needed to complete a PhD. For this I offer my thanks to Rådet för arbetslivsforskning, Adlerbertska stipendiestiftelsen, Petrus Hed-lunds studiefond, Stiftelsen Paul och Marie Berghaus donationsfond and the TMR network on Political Representation and Electoral Behaviour in the Euro-pean Union.
From the very beginning of the PhD program I have also enjoyed the moral “support” of the lost generation.i
To Helena, Elin, Andreas, Lina and to Birgitta: thank you all for accompanying me; and do you still remember how long four years felt when we started out in 1999? I would especially like to thank Helena for always caring about me and for being such great company, and Birgitta for pointing out that sometimes it might be a good idea to have some fun instead of working, and Lina for being such a pleasant and stimulating office mate and for actually trying, though clearly in vain, to make me finish my PhD too.
During the last few years, Sverker Jagers has been especially important for my professional well being. Starting at a crayfish party at Professor Rothsteins house many years ago, Sverker had the questionable judgment to somehow come to think that I was a suitable future research collaborator. My collabora-tion with Sverker has immensely enriched my work during the last years. Great
i
For a previous discussion of this notion, see Bågenholm (2008).
Acknowledgements
During the latter half of the decade that I have spent as a PhD student, I have come to realize that my goal never was to complete a PhD thesis. Rather, when I applied to the PhD program in the end of the 1990s, I just wanted to become a PhD student. That was it.
Ten years older, I am very grateful that I knew so little about what to expect when I started out. Countless are the hours and the evenings I have spent in my office during this time. But in no way do I feel this was a waste. Because countless are also all the beautiful sunsets I have briefly witnessed in passing from within the Department of Political Science at Sprängkullsgatan. Just think about what I would have missed if I had not been working late so many days in the past years.
First and foremost I would like to thank my supervisors during the writing of this thesis. My main supervisor, Sören Holmberg, has been very important dur-ing the whole project and has always supported me in many ways. Sören, I am very grateful for your encouragement, inspiration and support ever since I started on my Master thesis in 1998. Although I did not always immediately agree with your recommendations, thanks for not letting me get bogged down with the details. And my second supervisor, Henrik Oscarsson, deserves special thanks for his careful reading of my drafts and for good judgement, as well as for being available for all sorts of questions. No topic too small or unimportant. Many thanks also to Bo Rothstein and Maria Oskarson for luring me into the project I started in together with Helena Rohdén: Arbetsmarknadens politik.
Financial support is also needed to complete a PhD. For this I offer my thanks to Rådet för arbetslivsforskning, Adlerbertska stipendiestiftelsen, Petrus Hed-lunds studiefond, Stiftelsen Paul och Marie Berghaus donationsfond and the TMR network on Political Representation and Electoral Behaviour in the Euro-pean Union.
From the very beginning of the PhD program I have also enjoyed the moral “support” of the lost generation.i
To Helena, Elin, Andreas, Lina and to Birgitta: thank you all for accompanying me; and do you still remember how long four years felt when we started out in 1999? I would especially like to thank Helena for always caring about me and for being such great company, and Birgitta for pointing out that sometimes it might be a good idea to have some fun instead of working, and Lina for being such a pleasant and stimulating office mate and for actually trying, though clearly in vain, to make me finish my PhD too.
During the last few years, Sverker Jagers has been especially important for my professional well being. Starting at a crayfish party at Professor Rothsteins house many years ago, Sverker had the questionable judgment to somehow come to think that I was a suitable future research collaborator. My collabora-tion with Sverker has immensely enriched my work during the last years. Great
i
thanks for bringing me into the spinach, and for all your innumerable ideas and never ending enthusiasm.
During my time as a PhD student I also had the privilege to spend a year at the Mannheimer Zentrum für Europäische Sozialforschung at the University of Mannheim. Thanks to Hemann Schmitt for receiving me and providing such a good working environment. For me, Mannheim was truly a blessed place that helped me evolve in many ways. So a big thank you to all you people in the cel-lar: Hermann, Andreas, Andrea, Tanja, Edda and ofcourse my closest colleagues at that time: Alice, Raül and Laura. Special thanks to my office mate Alice Lud-vig I for giving me the courage to write my thesis in English.
A big collective thanks also to all the fantastic people I meet at the various TMR meetings in places such as Amsterdam, Leuven, Paris and Cadenabbia, and ofcourse, in Ann Arbor and in Essex. You are way too many too be men-tioned. But you know who you are. You all helped making my process of being transformed into a political scientist so much more fun and enjoyable.
The importance of the SOM institute during my first years as a PhD student and the inspiring and pleasant environment it provides, especially during the annual evaluation conferences, cannot be overestimated. I am very grateful for the opportunity to work with the people at the SOM institute at the University of Gothenburg. Especially I would like to thank Åsa Nilsson for sharing her knowledge so generously and for being such a good “data buddy”.
There are too many people that deserves to be mentioned, but I will try to re-member those most important: thanks to Marcia Grimes for linguistic emer-gency rescues, to Mikael Gilljam and Folke Johansson for carefully reading the entire manuscript at a late stage, to Stefan Dahlbergh for sharing my enthusiasm for Stata programming, to all the PhD students who back in the days had ten o’clock coffee in the west wing – you certainly made me look forward to going to work at that time. Thanks to Martin Brothén for all the encouragement dur-ing my first years at the Department, to Peter Esaiasson for helpful critique at the VOD seminar, to Per Hedberg for excellent and patient assistance with ex-plaining the election studies data sets, to Anna Brodin and Maria Oskarson for trustingly helping to shape my professional identity as a university teacher, to Douglas McCarthy and Keith Flint for keeping me alert when working late, to Henrik Friberg-Fernros for stimulating collaboration and to Staffan Kumlin and Mette Anthonsen for inspiration and good advice.
All together, I am grateful for having the opportunity to work at the Depart-ment of Political Science at the University of Gothenburg. Every week of the year, at least at some point, I feel that doing research here is simply one of the best jobs I can imagine.
And finally, to Caroline, thanks for all your patience with my work, for all your efforts when I finally finished this book, and for teaching me to appreciate a perfect cup of tea. To Eira, my daughter, thank you for all the joy, for teaching me how to sleep less, and to appreciate the simple things in life, like flowers, books with pictures and porridge. I look forward to spending more time with you both.
Johan Martinsson Göteborg, April 2009
thanks for bringing me into the spinach, and for all your innumerable ideas and never ending enthusiasm.
During my time as a PhD student I also had the privilege to spend a year at the Mannheimer Zentrum für Europäische Sozialforschung at the University of Mannheim. Thanks to Hemann Schmitt for receiving me and providing such a good working environment. For me, Mannheim was truly a blessed place that helped me evolve in many ways. So a big thank you to all you people in the cel-lar: Hermann, Andreas, Andrea, Tanja, Edda and ofcourse my closest colleagues at that time: Alice, Raül and Laura. Special thanks to my office mate Alice Lud-vig I for giving me the courage to write my thesis in English.
A big collective thanks also to all the fantastic people I meet at the various TMR meetings in places such as Amsterdam, Leuven, Paris and Cadenabbia, and ofcourse, in Ann Arbor and in Essex. You are way too many too be men-tioned. But you know who you are. You all helped making my process of being transformed into a political scientist so much more fun and enjoyable.
The importance of the SOM institute during my first years as a PhD student and the inspiring and pleasant environment it provides, especially during the annual evaluation conferences, cannot be overestimated. I am very grateful for the opportunity to work with the people at the SOM institute at the University of Gothenburg. Especially I would like to thank Åsa Nilsson for sharing her knowledge so generously and for being such a good “data buddy”.
There are too many people that deserves to be mentioned, but I will try to re-member those most important: thanks to Marcia Grimes for linguistic emer-gency rescues, to Mikael Gilljam and Folke Johansson for carefully reading the entire manuscript at a late stage, to Stefan Dahlbergh for sharing my enthusiasm for Stata programming, to all the PhD students who back in the days had ten o’clock coffee in the west wing – you certainly made me look forward to going to work at that time. Thanks to Martin Brothén for all the encouragement dur-ing my first years at the Department, to Peter Esaiasson for helpful critique at the VOD seminar, to Per Hedberg for excellent and patient assistance with ex-plaining the election studies data sets, to Anna Brodin and Maria Oskarson for trustingly helping to shape my professional identity as a university teacher, to Douglas McCarthy and Keith Flint for keeping me alert when working late, to Henrik Friberg-Fernros for stimulating collaboration and to Staffan Kumlin and Mette Anthonsen for inspiration and good advice.
All together, I am grateful for having the opportunity to work at the Depart-ment of Political Science at the University of Gothenburg. Every week of the year, at least at some point, I feel that doing research here is simply one of the best jobs I can imagine.
And finally, to Caroline, thanks for all your patience with my work, for all your efforts when I finally finished this book, and for teaching me to appreciate a perfect cup of tea. To Eira, my daughter, thank you for all the joy, for teaching me how to sleep less, and to appreciate the simple things in life, like flowers, books with pictures and porridge. I look forward to spending more time with you both.
Chapter 1
Introduction
Unemployment began to rise rapidly in Sweden at the beginning of the 1990s. The age of Swedish exceptionalism in relation to the European problem of per-sistent high unemployment came to an end. In the space of a few years, from 1990 to 1993, the official unemployment figures increased from a mere 1.5 per-cent to a startling 8 perper-cent. This would have been unimaginable in the public debate only a few years earlier.
The Swedish economy had previously managed to combine substantial growth with low to moderate unemployment for some time. Unemployment never ex-ceeded 4 percent during the 1980s and the average annual growth was slightly above 2 percent. In contrast, average unemployment in continental Europe in-creased from 3 percent in the mid 1970s to about 10 percent less than a decade later. Consequently, the “Swedish model” or the “third way” was often called a success. But the severe downturn in the economy and rapid escalation of unem-ployment of the early 1990s changed all that. Apart from the widely discussed unemployment level, the national budget deficit and the national debt became distressing topics for the nation.
All these sudden changes had very tangible consequences for many citizens. In fact, in the year 2000, 35 percent of Sweden’s adult population reported having been unemployed at some point during the previous ten years. If family members and friends are included the share of the population with direct or indirect ex-perience of unemployment during the previous ten years becomes as high as 81 percent (Gustafson 2003). Such thorough changes in the economic situation do not go unnoticed in the minds of ordinary citizens.
The new and changing economic situation could for example affect people’s per-sonal living conditions because of a reduction of disposable income, it could affect their perceptions of the national economic situation in general or they could ob-serve relatives or their children going through a troublesome time in the labour market. These kinds of observations of the economic situation that citizens make directly in their everyday lives or indirectly via media reporting all have the po-tential to influence how they think about politics as well (Kinder & Kiewiet 1979; Kramer 1983; Mutz 1998). Events such as economic decline or increasing unemployment can alter what people think about, what their main concerns are, and potentially also how they think or feel about certain political issues or con-cerns. For example, if unemployment rises, more citizens than before are likely to consider fighting unemployment a political priority. In addition, such a change in the labour market might also contribute to changing citizens’ view of the national economic situation for the worse. The influence that economic changes have on public opinion provides a link between politics and the economy.
Chapter 1
Introduction
Unemployment began to rise rapidly in Sweden at the beginning of the 1990s. The age of Swedish exceptionalism in relation to the European problem of per-sistent high unemployment came to an end. In the space of a few years, from 1990 to 1993, the official unemployment figures increased from a mere 1.5 per-cent to a startling 8 perper-cent. This would have been unimaginable in the public debate only a few years earlier.
The Swedish economy had previously managed to combine substantial growth with low to moderate unemployment for some time. Unemployment never ex-ceeded 4 percent during the 1980s and the average annual growth was slightly above 2 percent. In contrast, average unemployment in continental Europe in-creased from 3 percent in the mid 1970s to about 10 percent less than a decade later. Consequently, the “Swedish model” or the “third way” was often called a success. But the severe downturn in the economy and rapid escalation of unem-ployment of the early 1990s changed all that. Apart from the widely discussed unemployment level, the national budget deficit and the national debt became distressing topics for the nation.
All these sudden changes had very tangible consequences for many citizens. In fact, in the year 2000, 35 percent of Sweden’s adult population reported having been unemployed at some point during the previous ten years. If family members and friends are included the share of the population with direct or indirect ex-perience of unemployment during the previous ten years becomes as high as 81 percent (Gustafson 2003). Such thorough changes in the economic situation do not go unnoticed in the minds of ordinary citizens.
Chapter 1 2
What then are the political consequences of unemployment? If we restrict this question to electoral consequences or closely related phenomena such as the popularity of different political parties, this topic has primarily been studied in the longstanding research tradition of economic voting.1
Such studies have ex-amined how economic changes, for instance recessions, have influenced the out-come of elections (Lewis-Beck & Paldam 2000). The basic reward-punishment hypothesis of this research tradition was inspired by Key (1964) and simply states that voters tend to support the government if the economy is doing well and to vote against it otherwise. Since the seminal works of Goodhart and Bhansali (1970) and Kramer (1971) up to the recent and increasingly technically sophisticated debate (see for example Lewis-Beck 2006; Evans & Andersen 2006), most of these studies have concluded that people do tend to punish in-cumbent governments or presidents to some extent for economic downturns. In general, governments seem to be held accountable for the economic develop-ment as part of their governdevelop-mental performance.
Already at this early stage I want to point out that, although governmental performance and objective economic development are clearly different notions, when I use the term government’s economic performance I refer to the actual economic development under a specific government. The reason for this is to comply with the practice in the international literature on economic voting. This does not mean that I consider governments to be fully in control of the economic development. Empirically, a good economic development might occur despite poor governmental performance, and good governmental performance, where the government has done everything it can be expected to do, might still be fol-lowed by negative economic development. In this thesis governmental perform-ance should be regarded as a neutral term denoting actual policy outcome under a specific government, and economic performance simply refers to the actual eco-nomic development.
The link between the economy and electoral outcomes is, however, more com-plicated than the original reward-punishment hypothesis suggests. Although the general conclusion is that the reward-punishment hypothesis is supported and that the economic development seems to be an influential predictor of how well the incumbent government will do in the next election, it has become increas-ingly clear that there is a good deal of variation in the magnitude of this elec-toral punishment and in the economic vote. The empirical results of economic voting studies suffer from “instability” (Lewis-Beck & Paldam 2000:119). It has also been found that the importance of the economic development for electoral outcomes varies greatly both within countries over time (Sanders & Carey 2002; Stevenson 2002a; 2002b) and between countries (Anderson 1995; 2000; Powell & Whitten 1993; Sanders & Carey 2002; Stevenson 2002b). The future task for research on economic voting must now be to explain and understand why the influence of the economy on elections and incumbent popularity varies.
1
When it comes to less direct political consequences of unemployment many sociological studies exist. One of the earliest and best known of these is the famous study of the unemployed in the Austrian vil-lage of Marienthal led by one of the founders of survey based electoral research, Paul Lazarsfeldt (Jahoda et al. 1974[1933]). For an overview of recent studies of social consequences of unemploy-ment, such as effects on family life, self-esteem, health and more, see for example Ström (2002).
Chapter 1 2
What then are the political consequences of unemployment? If we restrict this question to electoral consequences or closely related phenomena such as the popularity of different political parties, this topic has primarily been studied in the longstanding research tradition of economic voting.1
Such studies have ex-amined how economic changes, for instance recessions, have influenced the out-come of elections (Lewis-Beck & Paldam 2000). The basic reward-punishment hypothesis of this research tradition was inspired by Key (1964) and simply states that voters tend to support the government if the economy is doing well and to vote against it otherwise. Since the seminal works of Goodhart and Bhansali (1970) and Kramer (1971) up to the recent and increasingly technically sophisticated debate (see for example Lewis-Beck 2006; Evans & Andersen 2006), most of these studies have concluded that people do tend to punish in-cumbent governments or presidents to some extent for economic downturns. In general, governments seem to be held accountable for the economic develop-ment as part of their governdevelop-mental performance.
Already at this early stage I want to point out that, although governmental performance and objective economic development are clearly different notions, when I use the term government’s economic performance I refer to the actual economic development under a specific government. The reason for this is to comply with the practice in the international literature on economic voting. This does not mean that I consider governments to be fully in control of the economic development. Empirically, a good economic development might occur despite poor governmental performance, and good governmental performance, where the government has done everything it can be expected to do, might still be fol-lowed by negative economic development. In this thesis governmental perform-ance should be regarded as a neutral term denoting actual policy outcome under a specific government, and economic performance simply refers to the actual eco-nomic development.
The link between the economy and electoral outcomes is, however, more com-plicated than the original reward-punishment hypothesis suggests. Although the general conclusion is that the reward-punishment hypothesis is supported and that the economic development seems to be an influential predictor of how well the incumbent government will do in the next election, it has become increas-ingly clear that there is a good deal of variation in the magnitude of this elec-toral punishment and in the economic vote. The empirical results of economic voting studies suffer from “instability” (Lewis-Beck & Paldam 2000:119). It has also been found that the importance of the economic development for electoral outcomes varies greatly both within countries over time (Sanders & Carey 2002; Stevenson 2002a; 2002b) and between countries (Anderson 1995; 2000; Powell & Whitten 1993; Sanders & Carey 2002; Stevenson 2002b). The future task for research on economic voting must now be to explain and understand why the influence of the economy on elections and incumbent popularity varies.
1
Introduction 3 In this thesis I argue that in order to better understand the effects of economic changes on election outcomes and party popularity, the focus of economic vot-ing studies must be broadened. Part of this hard-to-explain variation in the strength of the economic vote might stem from the omission of another theoreti-cal tradition in electoral research, namely issue ownership theory and issue-priority models of voting.2
This chapter will explain why this may possibly be so and what we are missing as a result of the relatively narrow focus of tradi-tional economic voting studies.
Fundamentally, this study is about the electoral consequences of unemploy-ment in Sweden. To examine this I take the research tradition of economic vot-ing as my point of departure. To gain a more comprehensive understandvot-ing of the consequences for electoral outcomes and incumbent popularity of unem-ployment in Sweden than the traditional economic voting framework offers, I will develop and apply a joint model of the electoral consequences of unem-ployment by combining economic voting with issue ownership theory.
Economic voting and partisan effects
The links between the economy and voting have been explored for more than half a century (early examples include Tibbits 1931; Berelson et al. 1954; Good-hart & Bhansali 1970; while on Sweden we find early studies by Jonung & Wa-densjö 1979; Åkerman 1946; Rydé 1950; Kramer 1971; Madsen 1980; Holm-berg 1984). Systematic quantitative studies of the influence of the economy on public support for the ruling party or president and on the election results of incumbents have been conducted en masse in the past decades. This field of re-search clearly continues to flourish.3
But this is not just another study of eco-nomic voting. The aim of this study is instead to integrate ecoeco-nomic voting and another theoretical tradition that takes political factors into account to a larger extent. Most economic voting studies seem to wear a kind of political blinkers, making their focus too narrow to gain a proper understanding of the conse-quences of economic changes for elections or incumbent popularity. Despite the vast amount of research, many questions still remain in the realm of economic influences on elections and voting.
The main approach to explaining variations in the strength of the economic vote has been the institutional approach set off by Powell and Whitten in 1993. The central notion in this research tradition is that of clarity of responsibility. Political contexts with a high degree of clarity of responsibility are expected to facilitate the retrospective performance-based voting of the economic voter. On the other hand, unclear patterns of political responsibility are thought to make it more difficult for voters to hold the incumbent accountable for the economic performance (Powell & Whitten 1993). Among the factors examined in these studies we find government composition (e.g. single party vs multi-party gov-ernments), bicameralism and the committee system.
2
Even at this early point, I wish to make clear that I do not intend to provide a general explanation for the variation in economic voting. My goal is rather to outline and test one specific and new such potential explanation on the case of Sweden: namely the integrated model of economic voting and issue ownership presented in this chapter.
3
A quick search in the Social Sciences Citation Index (SSCI) shows that over 100 articles have been published since the year 2000 on the topic of “economic voting” alone.
Introduction 3
In this thesis I argue that in order to better understand the effects of economic changes on election outcomes and party popularity, the focus of economic vot-ing studies must be broadened. Part of this hard-to-explain variation in the strength of the economic vote might stem from the omission of another theoreti-cal tradition in electoral research, namely issue ownership theory and issue-priority models of voting.2
This chapter will explain why this may possibly be so and what we are missing as a result of the relatively narrow focus of tradi-tional economic voting studies.
Fundamentally, this study is about the electoral consequences of unemploy-ment in Sweden. To examine this I take the research tradition of economic vot-ing as my point of departure. To gain a more comprehensive understandvot-ing of the consequences for electoral outcomes and incumbent popularity of unem-ployment in Sweden than the traditional economic voting framework offers, I will develop and apply a joint model of the electoral consequences of unem-ployment by combining economic voting with issue ownership theory.
Economic voting and partisan effects
The links between the economy and voting have been explored for more than half a century (early examples include Tibbits 1931; Berelson et al. 1954; Good-hart & Bhansali 1970; while on Sweden we find early studies by Jonung & Wa-densjö 1979; Åkerman 1946; Rydé 1950; Kramer 1971; Madsen 1980; Holm-berg 1984). Systematic quantitative studies of the influence of the economy on public support for the ruling party or president and on the election results of incumbents have been conducted en masse in the past decades. This field of re-search clearly continues to flourish.3
But this is not just another study of eco-nomic voting. The aim of this study is instead to integrate ecoeco-nomic voting and another theoretical tradition that takes political factors into account to a larger extent. Most economic voting studies seem to wear a kind of political blinkers, making their focus too narrow to gain a proper understanding of the conse-quences of economic changes for elections or incumbent popularity. Despite the vast amount of research, many questions still remain in the realm of economic influences on elections and voting.
The main approach to explaining variations in the strength of the economic vote has been the institutional approach set off by Powell and Whitten in 1993. The central notion in this research tradition is that of clarity of responsibility. Political contexts with a high degree of clarity of responsibility are expected to facilitate the retrospective performance-based voting of the economic voter. On the other hand, unclear patterns of political responsibility are thought to make it more difficult for voters to hold the incumbent accountable for the economic performance (Powell & Whitten 1993). Among the factors examined in these studies we find government composition (e.g. single party vs multi-party gov-ernments), bicameralism and the committee system.
2
Even at this early point, I wish to make clear that I do not intend to provide a general explanation for the variation in economic voting. My goal is rather to outline and test one specific and new such potential explanation on the case of Sweden: namely the integrated model of economic voting and issue ownership presented in this chapter.
3
Chapter 1 4
This is not a country comparative study, however. The study’s empirical scope is limited to the case of Sweden and will instead depart from another sub-branch of economic voting called partisan effects. Previous studies have suggested that the public holds incumbent parties or presidents accountable for the economic development to a different extent, or even in different ways, depending on the in-cumbents’ partisan colour or position in the ideological spectrum (Powell & Whitten 1993; Anderson 1995; Carlsen 2000; van der Brug et al. 2007). Conse-quently, this is often referred to as partisan effects in research on economic voting.
One such partisan difference found in some American studies is that the elec-torate does not seem to hold Democratic presidents accountable for rising un-employment, but that it works as expected for Republican ones (Swank 1993). Subsequent research has later elaborated somewhat on the partisan effects in economic voting and specified two contradicting hypotheses: the clientele hy-pothesis and the salient goal hyhy-pothesis (Carlsen 2000). According to the clien-tele hypothesis the electorate does not punish left governments for unemploy-ment because they assume there is no one better at dealing with the issue of un-employment despite the failure of the incumbent (leftist) government. In con-trast, the salient goal hypothesis claims that a left government would instead be electorally punished to an even greater extent than other governments since the issue of unemployment is seen as the home turf of left governments and they are therefore expected to handle it especially well. Although the amount of research on this topic is still very limited, some support has been found for the general idea of partisan effects in economic voting. No firm conclusions concerning the direction of partisan effects have been reached, however. The study by Carlsen (2000) found support for the clientele hypothesis in the US and in Canada (as did Swank 1993), while the findings for the UK were inconclusive and the sali-ent goal hypothesis was instead supported in Australia (Carlsen 2000:148). Fur-ther, Carlsen’s results are not always clear-cut for individual countries eiFur-ther, but sometimes vary between different governments in the same country. Sander’s (2000:288f) results also mainly supported the clientele hypothesis in the UK. On the other hand, Powell and Whitten’s (1993:404f) comparative study instead found evidence favouring the salient goal hypothesis.
Studies on partisan effects in economic voting have usually tested the hypothe-ses sketched above in a straightforward manner, for example via including in-teraction terms for the ideological flavour of the government in their regression models or by examining different governments separately. But previous studies have never been particularly clear about what causes these differences and how the individual voting decisions are influenced by whether the incumbent party under consideration is a party to the left or to the right on the political spectrum in the country studied. In my view, previous research has not theoretically speci-fied the causal mechanism in a satisfying manner.
It is also imperative to note that it cannot be taken for granted that left and right wing governments have certain properties per se, nor in general, and neither per-taining to specific issues as the one in focus here: unemployment. Rather, this should be regarded as a topic for empirical inquiry. All left governments are not the same, and the same naturally goes for right governments. The theory and study of partisan effects need to rely less on assumptions and generalizations of what characterizes left and right party incumbents, both in time and in space. The relation between a leftist party and the issue of unemployment for example is not
Chapter 1 4
This is not a country comparative study, however. The study’s empirical scope is limited to the case of Sweden and will instead depart from another sub-branch of economic voting called partisan effects. Previous studies have suggested that the public holds incumbent parties or presidents accountable for the economic development to a different extent, or even in different ways, depending on the in-cumbents’ partisan colour or position in the ideological spectrum (Powell & Whitten 1993; Anderson 1995; Carlsen 2000; van der Brug et al. 2007). Conse-quently, this is often referred to as partisan effects in research on economic voting.
One such partisan difference found in some American studies is that the elec-torate does not seem to hold Democratic presidents accountable for rising un-employment, but that it works as expected for Republican ones (Swank 1993). Subsequent research has later elaborated somewhat on the partisan effects in economic voting and specified two contradicting hypotheses: the clientele hy-pothesis and the salient goal hyhy-pothesis (Carlsen 2000). According to the clien-tele hypothesis the electorate does not punish left governments for unemploy-ment because they assume there is no one better at dealing with the issue of un-employment despite the failure of the incumbent (leftist) government. In con-trast, the salient goal hypothesis claims that a left government would instead be electorally punished to an even greater extent than other governments since the issue of unemployment is seen as the home turf of left governments and they are therefore expected to handle it especially well. Although the amount of research on this topic is still very limited, some support has been found for the general idea of partisan effects in economic voting. No firm conclusions concerning the direction of partisan effects have been reached, however. The study by Carlsen (2000) found support for the clientele hypothesis in the US and in Canada (as did Swank 1993), while the findings for the UK were inconclusive and the sali-ent goal hypothesis was instead supported in Australia (Carlsen 2000:148). Fur-ther, Carlsen’s results are not always clear-cut for individual countries eiFur-ther, but sometimes vary between different governments in the same country. Sander’s (2000:288f) results also mainly supported the clientele hypothesis in the UK. On the other hand, Powell and Whitten’s (1993:404f) comparative study instead found evidence favouring the salient goal hypothesis.
Studies on partisan effects in economic voting have usually tested the hypothe-ses sketched above in a straightforward manner, for example via including in-teraction terms for the ideological flavour of the government in their regression models or by examining different governments separately. But previous studies have never been particularly clear about what causes these differences and how the individual voting decisions are influenced by whether the incumbent party under consideration is a party to the left or to the right on the political spectrum in the country studied. In my view, previous research has not theoretically speci-fied the causal mechanism in a satisfying manner.
Introduction 5 necessarily the same in all countries or at all points in time in any one country. This might perhaps be the explanation for the difficulties of previous research on parti-san effects in economic voting studies to reach agreement in their conclusions.4
I propose that the integration of two research traditions within the study of electoral behaviour that have been unfortunately insulated from each other, economic voting and issue ownership theory, will serve the aim of this thesis: to advance our knowledge of partisan effects in economic voting and to examine the electoral consequences of the issue of unemployment in Sweden.
The idea of partisan effects in economic voting provides a good basis for a theo-retical link between retrospective voting models, such as economic voting, and issue-priority models (or party competence models, see Schmitt 2001) such as issue ownership. It does so by pointing out the possibility that all governments are not necessarily held equally accountable for the economic development that goes on during their terms. In fact, it is possible that both the magnitude and the direction of the retrospective vote vary according to the partisan composition of the incumbent government. If this is so, however, a good explanation is needed for why and when the magnitude or direction of the retrospective vote varies between political parties. Previous studies have failed to provide such an expla-nation. I suggest that one important mechanism that creates these partisan dif-ferences in economic voting is party issue ownership. In fact, we do not need a new theory or an add-on to the partisan theory of economic voting.5
To improve our understanding of the electoral consequences of economic changes, we need to combine the explanations and predictions of two different but already estab-lished traditions in research on electoral behaviour.
The electoral consequences of unemployment
Economic voting studies have usually overlooked the possibility that certain economic issues on the public agenda might benefit certain political parties through their ownership of the issue at hand. Likewise, studies of issue owner-ship have usually overlooked the possibility that although certain issues might benefit the party owning the issue, by means of their being salient, voters might still simultaneously take retrospective considerations of governmental perform-ance in the same issues into account when deciding whom to vote for.
In addition, I claim that it is not only desirable to take both economic voting and issue ownership into consideration because they complement each other, but that this is sometimes necessary in order to correctly understand the toral consequences of economic changes. This is so because predictions of elec-toral outcomes based on both models will be affected by the same factor –
4
It might also be added that, even if all left parties or all right parties did have the same relation to a certain issue, we should be conscious that in principle it is not the actual properties or policy of political parties, but voters’ perceptions of these properties or policies, that eventually matter to individual electoral behaviour.
5
Especially the clientele hypothesis in the partisan theory of economic voting is particularly close to issue ownership theory. In fact, it is questionable whether electoral behaviour in accordance with the clientele hypothesis can be said to constitute “economic voting” in the traditional sense at all since it states that left (right) parties gain from high unemployment (inflation), both when in government and in opposition (Carlsen 2000:142). This sounds very much like issue ownership, minus the part where the issues have to become salient to the public in order to have an effect on party support.
Introduction 5
necessarily the same in all countries or at all points in time in any one country. This might perhaps be the explanation for the difficulties of previous research on parti-san effects in economic voting studies to reach agreement in their conclusions.4
I propose that the integration of two research traditions within the study of electoral behaviour that have been unfortunately insulated from each other, economic voting and issue ownership theory, will serve the aim of this thesis: to advance our knowledge of partisan effects in economic voting and to examine the electoral consequences of the issue of unemployment in Sweden.
The idea of partisan effects in economic voting provides a good basis for a theo-retical link between retrospective voting models, such as economic voting, and issue-priority models (or party competence models, see Schmitt 2001) such as issue ownership. It does so by pointing out the possibility that all governments are not necessarily held equally accountable for the economic development that goes on during their terms. In fact, it is possible that both the magnitude and the direction of the retrospective vote vary according to the partisan composition of the incumbent government. If this is so, however, a good explanation is needed for why and when the magnitude or direction of the retrospective vote varies between political parties. Previous studies have failed to provide such an expla-nation. I suggest that one important mechanism that creates these partisan dif-ferences in economic voting is party issue ownership. In fact, we do not need a new theory or an add-on to the partisan theory of economic voting.5
To improve our understanding of the electoral consequences of economic changes, we need to combine the explanations and predictions of two different but already estab-lished traditions in research on electoral behaviour.
The electoral consequences of unemployment
Economic voting studies have usually overlooked the possibility that certain economic issues on the public agenda might benefit certain political parties through their ownership of the issue at hand. Likewise, studies of issue owner-ship have usually overlooked the possibility that although certain issues might benefit the party owning the issue, by means of their being salient, voters might still simultaneously take retrospective considerations of governmental perform-ance in the same issues into account when deciding whom to vote for.
In addition, I claim that it is not only desirable to take both economic voting and issue ownership into consideration because they complement each other, but that this is sometimes necessary in order to correctly understand the toral consequences of economic changes. This is so because predictions of elec-toral outcomes based on both models will be affected by the same factor –
4
It might also be added that, even if all left parties or all right parties did have the same relation to a certain issue, we should be conscious that in principle it is not the actual properties or policy of political parties, but voters’ perceptions of these properties or policies, that eventually matter to individual electoral behaviour.
5
Chapter 1 6
changes in economic performance. Interestingly, the two models might under certain circumstances yield completely contradictory predictions of the effect on electoral outcomes for the same change in economic performance.
For example, let us consider what happens when unemployment is on the rise. From the viewpoint of economic voting, the most important change in public opinion is that rising unemployment is likely to alter people’s evaluations of the country’s general economic situation. Unemployment is a prominent feature of the economy and one of the best known macro economic indicators. Further, this change in economic evaluation is likely to be for the worse since unem-ployment is generally seen as something unwanted. In short, if unemunem-ployment rises, citizens will take notice and their view of the economy will become more negative (or less positive) than before. In turn, such an evolution might lead to greater public dissatisfaction with the way the government is handling the econ-omy and eventually electoral punishment with fewer voters prepared to support the government in the next election.
From an electoral point of view, however, this is not the only important change in public opinion that rising unemployment might cause. What happens more when unemployment is rising is that the issue of unemployment itself moves up on the agenda. When the objective situation on the labour market is becoming worse, not only will the public’s view of the labour market situation become less positive, but public interest in issues related to the labour market will increase too. If unemployment itself rises, the issue of unemployment is also likely to rise on the media’s agenda as well as on the public agenda.
There are many ways that citizens can notice rising unemployment. Many citi-zens primarily receive their information about the changes in the labour market indirectly, for example via the media. There are real people behind the unem-ployment figures, however, and many citizens also experience unemunem-ployment in a more direct way, either personally or among relatives or friends.6
Either way, citi-zens are likely to adjust their view on the world and their political priorities and concerns to the information they receive. Thus the issue will become more salient. But why is issue salience – the importance of an issue – in itself electorally consequential? This is where issue ownership theory comes into play. The basic idea of issue ownership is that certain political issues are advantageous to cer-tain political parties. This means that the mere fact that a particular issue is be-ing discussed or considered is in itself electorally advantageous to a certain party, compared to a situation where that issue is not being discussed. Different political issues are advantageous or disadvantageous to different political parties in such a way that certain parties gain support while others lose support when an issue steps out of the shadows and becomes salient (Holmberg 1981; Budge & Farlie 1983)7
. This is sometimes, for the party that gains an advantage by the issue, referred to as issue ownership.
6
Although personal experiences are often regarded as politically inconsequential, it is clearly possible for citizens to use personal experiences for political conclusions, perhaps especially if these experi-ences, as in the case of rising unemployment, correspond to societal trends and collective experiences (Mutz 1998; Kumlin 2002).
7
Issue ownership theory has more recently been slightly reshaped and renewed, see for example Petrocik (1996) or Petrocik et al. (2003). More on this in Chapter 5.
Chapter 1 6
changes in economic performance. Interestingly, the two models might under certain circumstances yield completely contradictory predictions of the effect on electoral outcomes for the same change in economic performance.
For example, let us consider what happens when unemployment is on the rise. From the viewpoint of economic voting, the most important change in public opinion is that rising unemployment is likely to alter people’s evaluations of the country’s general economic situation. Unemployment is a prominent feature of the economy and one of the best known macro economic indicators. Further, this change in economic evaluation is likely to be for the worse since unem-ployment is generally seen as something unwanted. In short, if unemunem-ployment rises, citizens will take notice and their view of the economy will become more negative (or less positive) than before. In turn, such an evolution might lead to greater public dissatisfaction with the way the government is handling the econ-omy and eventually electoral punishment with fewer voters prepared to support the government in the next election.
From an electoral point of view, however, this is not the only important change in public opinion that rising unemployment might cause. What happens more when unemployment is rising is that the issue of unemployment itself moves up on the agenda. When the objective situation on the labour market is becoming worse, not only will the public’s view of the labour market situation become less positive, but public interest in issues related to the labour market will increase too. If unemployment itself rises, the issue of unemployment is also likely to rise on the media’s agenda as well as on the public agenda.
There are many ways that citizens can notice rising unemployment. Many citi-zens primarily receive their information about the changes in the labour market indirectly, for example via the media. There are real people behind the unem-ployment figures, however, and many citizens also experience unemunem-ployment in a more direct way, either personally or among relatives or friends.6
Either way, citi-zens are likely to adjust their view on the world and their political priorities and concerns to the information they receive. Thus the issue will become more salient. But why is issue salience – the importance of an issue – in itself electorally consequential? This is where issue ownership theory comes into play. The basic idea of issue ownership is that certain political issues are advantageous to cer-tain political parties. This means that the mere fact that a particular issue is be-ing discussed or considered is in itself electorally advantageous to a certain party, compared to a situation where that issue is not being discussed. Different political issues are advantageous or disadvantageous to different political parties in such a way that certain parties gain support while others lose support when an issue steps out of the shadows and becomes salient (Holmberg 1981; Budge & Farlie 1983)7
. This is sometimes, for the party that gains an advantage by the issue, referred to as issue ownership.
6
Although personal experiences are often regarded as politically inconsequential, it is clearly possible for citizens to use personal experiences for political conclusions, perhaps especially if these experi-ences, as in the case of rising unemployment, correspond to societal trends and collective experiences (Mutz 1998; Kumlin 2002).
7
