Article contents
Patrimonial Economic Voting and Asset Value – New Evidence from Taxation Register Data
Published online by Cambridge University Press: 08 November 2016
Abstract
Recent research on economic voting has moved beyond the traditional reward–punishment hypothesis, according to which the economy is merely considered a valence issue. Instead, patrimonial economic voting research looks at voters as property owners within the economic system. These studies have relied on survey items that measure whether individuals own different kinds of property to test the patrimonial dimension. This study emphasizes the importance of a surprisingly neglected aspect: the value of assets. It uses official register data files from the Swedish Tax Agency on the value of individuals’ assets merged with survey data from the 2006 Swedish National Election Study. The study finds that the relationship between patrimony and voting behavior in Sweden is similar to that found in other countries, but only when it is tested in a similar way as in these studies – that is, only when it is coded as whether voters own different assets. This study brings three important contributions to the debate. First, it offers a new empirically based categorization of the dimensionality of asset ownership and shows that the previous distinction between low- and high-risk assets is insufficient. Secondly, it shows that merely having assets or not, which is what previous studies have measured, is not what primarily matters; the relevant factor is the value of the assets. And thirdly, it demonstrates that only the value of some kinds of assets matters (especially stocks and real estate properties), while other assets (savings in bonds and funds) do not affect voting behavior or political opinions.
- Type
- Articles
- Information
- Copyright
- © Cambridge University Press 2016
Footnotes
Department of Political Science, University of Gothenburg, Sweden (emails: [email protected], [email protected]). Persson’s work in this project was supported by ‘Stiftelsen Lars Hiertas Minne’. The authors wish to thank participants for useful comments from presentations at APSA 2013 in Chicago and a seminar at the University of Gothenburg. In order to access the register data used for this article, the authors had to sign a contract stating that they were not allowed to make the data publicly available. The authors were only allowed to analyze the data using Statistics Sweden’s remote server system MONA. The dataset and Stata do-files used to carry out the analyses are stored at Statistics Sweden, Stockholm. Instructions on how to obtain the data replication files for the analyses are available at http://dataverse.harvard.edu/dataverse/BJPolS. Online appendices are available at http://dx.doi.org/doi: 10.1017/S0007123416000181.
References
- 18
- Cited by