Determinants of party policy preferences: evidence from party manifestos in Belgium

Our study aimed at contributing to the understanding of party policy preferences in Belgium, i.e. how far can we explain the attention that political parties allocate to different policy issues in their electoral manifestos. One of the added values of our study was located in our research design. We reversed the perspective that is used in most of previous works and we used party policy preferences – via the analysis of party manifestos – as a dependent rather than an independent variable. We were not interested in what could be explained by party preferences but in what could actually explain them. In addition, we observed that, in the few works that considered party policy preferences as a dependent variable, party competition was rarely mobilised as an explanatory variable. Our study aimed at contributing to the understanding of the content of party manifestos by confronting the hypothesis of party competition with other possible explanations.<p><p>With the help of new data about party preferences in Belgian party manifestos collected in the framework of the CAP project, we quantitatively analysed the content of all party manifestos between 1977 and 2007 and coded them according policy domains and issues (up to about 250 policy issues and coding categories). Based on the selection of four specific policy issues (environment, decentralisation, migration and morality issues) and on regression analyses (panel data), we hypothesised that party preferences on theses policy issues is best explained by party competition variables.<p><p>Our findings confirmed that party policy preferences are not static but rather that they evolve over time. Party preferences are different over time and space and we aimed at providing clues about what could explain these differences. Based on the literature, a large set of potential explanatory variables has been mobilised in order to explain these differences. But most of these independent variables have no or few impact on party preferences, such as the fractionalisation of the party system or ‘real-world’ indicators. Contrary to previous findings, changes in party manifestos are not explained by the fact that the party grows in size and gets older or by the fact that it wins or loses the elections. Similarly, we observed that party strategies – including party name change and the creation of electoral alliances – had no impact of the content of part manifestos. Even if our bivariate analyses indicated the importance of phenomena related to the government formation and participation, we found out that this effect disappears in multi-variate analyses.<p><p>Still, the introduction of our party competition variables – based on the niche party’s size, electoral fortunes and government participation – provided ambiguous results, depending on the policy issue at stake. Party competition contributes to the understanding of party policy preferences on environment and migration. Nonetheless, our models do not demonstrate an impact of party competition on preferences concerning decentralisation and morality. When controlling for party families, we observe that party competition has a significant impact on party preferences, meaning that political parties react to the electoral strength of a niche party by paying more attention to the niche party’s issue in their manifesto. Finally, the observed impact of party competition on policy preferences concerns certain parties only and the other parties display preferences that appear independent from the existing patterns of party competition.Doctorat en Sciences politiques et sociales