How political parties’ issue ownerships reduce to spatial proximity

Abstract

Scholarly interest in issue ownership is growing rapidly. Although originally introduced as a competence-oriented, alternative concept to the predominantly spatial understanding of voting and party behaviour, parties’ policy positions are an inescapable aspect of issue ownership. Using data for multiple issues in several countries over time, this article shows that the party with issue ownership sides with the median voter. A party earns issue ownership by taking up a position as close to as many voters as possible. Moreover, the analysis indicates that a party’s issue emphasis only matters to issue ownership insofar as it is used as a device to make its position credible to voters. Hence, to have issue ownership is to have a credible position, and in that sense, issue ownership has less added theoretical value to spatial proximity than previous literature suggests.

Keywords:

• Issue ownership
• political parties
• policy position
• median voter

Acknowledgements

I would like to thank reviewers and the editors of this journal for valuable comments on earlier versions of the manuscript together with colleagues in the Political Parties Research Group at the Department of Political Science, Aarhus University, participants at presentations at the Department of Government, University of Vienna and the Department of Government, University of Gothenburg in 2018, and participants at the annual meetings of the American Political Science Association and Danish Political Science Association in 2017.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 In line with Petrocik’s original conception of issue ownership, this study is concerned with issue ownership as a party’s reputation among the electorate for being able to competently handle problems on an issue rather than merely voters’ ‘spontaneous association’ between a party and an issue (Walgrave et al. 2012; see also Stubager 2018).

2 The countries are Sweden, Norway, Denmark, Finland, France, Germany, Austria, the United Kingdom, the United States, Canada, Australia and New Zealand.

3 I used the issue content codebook from the Comparative Agendas Project (Baumgartner et al. 2011).

4 To reflect the perception of welfare in the United States, this was coded as social security.

5 The log of emphasis, to be more precise, following Lowe et al. (2011).

6 Lelkes measures bimodality as bimodality = $\frac{m^2+1}{n+3\frac{{(n-1)}^2}{\left(n-2\right)\left(n-3\right)}},$ where m is the skewness of the distribution, and n is the kurtosis score of the distribution.

7 The coding strategy adopted by Dolezal et al. (2013) is admirable but unrealistic to carry out across a large number of countries. The measurement strategy by Meyer and Wagner (2017) also uses CMP data but is constrained to apply only to overarching issues such as the economy, with many sub-issues from which to calculate a party’s position and emphasis on that issue. This is not the case in this analysis, which focuses on many specific issues like immigration and education.

8 Right-wing parties are liberal, conservative, Christian democratic, agrarian, and far-right parties, and left-wing parties are green parties, communists, socialists and social democrats (using the party families in the Comparative Manifesto Project).

9 The EU is unsurprisingly slightly odd, and there appears to be no connection between the position of the median voter, party positions, and issue ownership (Green-Pedersen 2012).

10 If the welfare issue is excluded or Germany is excluded, the levels of statistical significance for the party proximity variable are p < 0.13 and p < 0.15, respectively, in Tables A6 and A7. If the issues of welfare or the environment are excluded or Sweden is excluded, the levels of statistical significance for the interaction term between party issue emphasis and party position proximity are p < 0.15 and p < 0.12, respectively, in Tables A8 and A9. Hence, in a few instances, the reduction in the number of observations when excluding an issue or a country in the analysis generates p-values that border conventional levels of statistical significance.

11 The interaction between party proximity and bimodality is only statistically significant in parts of the bimodality interval (see Figure 6) and the interaction term in the second column in Table 4 therefore does not reach conventional levels of statistical significance. This means that the evaluation of the robustness of the result across issues and countries is mainly concerned with the extent to which the size and direction of the coefficient changes. This does not appear to be the case.

12 With this alternative measure of party proximity, the level of statistical significance for the interaction term is p < 0.11 and it remains positive.

Additional information

Notes on contributors

Henrik Bech Seeberg

Henrik Bech Seeberg is Associate Professor in the Department of Political Science, Aarhus University. His research focuses on party competition, voting behaviour and political agenda setting, and has been published in the European Journal of Political Research, Journal of Public Administration Research and Theory, Electoral Studies and Party Politics, among others. [[email protected]]