Adaptive voting: an empirical analysis of participation and choice

Dynamic models of learning and adaptation have provided realistic predictions in terms of voting behavior. This study aims at contributing to their empirical verification by investigating voting behavior in terms of participation as well as choice. We test through panel data methods an outcome-based learning mechanism based on the following assumptions: (a) people expect that the party they do not support will be unable to bring economic improvements; (b) they receive a feedback whose impact depends on the consistency between their last voting behavior and personal economic improvements (or worsening) from the last election; (c) they tend to discard choices associated to an inconsistent feedback. Results show that feedbacks of this sort affect persistence of voting behavior, interpreted as participation and voting choice. Age and trade union affiliation reinforce this adaptive behavior. The analysis also investigates the intensity of the learning feedback, differentiating between a strong inconsistent feedback, which leads to a vote switch in favor of the opponent party, and a weak inconsistent feedback, which induces just abstention rather than a vote switch.