Proximity or Directional Model of Voting for the Turkish Voter?

Abstract

Voting behavior is a very complex type of political behavior. Therefore, understanding why voters vote for a particular political party or a candidate requires developing complex models. In 1957, Anthony Downs, who built his model on Hottelings' and Smithies' models, argued that political parties’ and candidates’ ideological and issue positions can be expressed on a one-dimensional space. On one hand, it was highly reductionist to argue that political ideas on a particular issue can be expressed this way, on the other, it was highly practical from analytical point of view. Locating parties, candidates and voters on a one-dimensional space according to their ideological or issue positions was then a revolutionaly idea and helped comparing party, candidate and voter ideological and issue positions within and across countries. These models, which were called spatial models of party competition were further developed over time and helped understanding voting behavior. Currently, spatial models of party competition have two major competing models linking voter ideological positions with party ideological positions. Simply, while the proximity model proposes that voters vote for the parties or candidates that hold ideological positions in the political space that are closest to their own, the directional model suggests that the voters vote for the parties or candidates that are on their side of the two-dimensional political spectrum and more extreme than their own while being within the acceptability region. This research aims to test the applicability of these two voting models for the Turkish voter. Türkiye constitutes an interesting case study with its long-term PR electoral system as it was suggested in the extant literature that proximity model is a more appropriate tool to explain voting behavior in Proportional (PR) systems. Thus, we hypothesize that in Türkiye, where a PR electoral system is in effect for parliamentary elections, voter electoral preferences are better explained by the proximity model than the directional model. Our research analyzes Comparative Study of Electoral Systems (CSES) data for voters of the four major political parties in Türkiye, the Justice and Development Party (JDP), the Republican People's Party (RPP), the National Action Party (NAP), and the People's Democratic Party (PDP). A series of Multiple Linear Regression Analyses were conducted to reveal associations between the dependent and the independent variables. Voter embracement, as expressed as like-dislike of each political party for each voter, is seperately used as the dependent variable for each analysis. Issue distance and issue scalar product were used as key independent variables representing the formulas for the proximity and the directional models, respectively. Additionally, education, age, gender and income were recruited as classical control variables. Comparing explanatory powers of the statistical models showed that, contrary to the findings of MacDonald and his colleagues, the proximity model of voting is a more appropriate tool than the directional model to explain voting behavior in Türkiye. From a macro-political perspective, this finding supports Westholm's (1997) argument that the PR provides a more appropriate tool to explain voting behavior in PR systems. Yet, it should be noted that further multi-country comperative analyses required for certain results.