Spatial theory would predict voters to react to parties’ position switches on the left-right axis, with ‘partisan sorting’: either follow the party with adjusting their own left-right position, or vote for a different party in the succeeding election. However, this ‘partisan sorting’ does not take place as citizens do not react on a significant level to changes in parties’ policy positions as represented in election campaigns (Adams et al., 2011; 2012). The research paper aims at solving this puzzle. I argue that pure spatial modeling cannot account for voters’ reactions to policy change because voting is not a purely rational activity. On one hand, voters are often not best informed about policy change, on the other hand, voters’ emotional attachments to parties have a strong effect on their loyalty to the party. Using comparative data on individual voting behaviour from the Comparative Study of Electoral Systems (CSES), Modules 2-4, and the Comparative Manifesto Project (CMP), I run a 2-level multinomial model which takes account of voters’ reactions to party movements on the left-right axis by comparing voters before their current vote decision. I show that parties’ position change does not necessarily lead voters to abandon the party they previously voted for, but that for vote switching it is most important that voters know party positions precisely. For parties’ position change to have a significant effect on vote switching it is necessary that voters are willing and able to notice this change.
