7 - Microfoundations of Incumbency Bias

Summary

Chapter 7 employs original survey experiments to test the theory’s microfoundations. The experimental settings display wide variation in incumbency bias and the designs balance the tradeoff between abstraction and control. The results are consistent with bounded accountability: citizens process information about fiscal shocks in a rationally. In Brazil, when the hypothetical nature of the scenario deprives them of prior information about candidates, citizens only respond to information about a fiscal windfall when it is effectively deployed in their district. In Argentina, where the scenario is real and citizens thus hold prior views about incumbents, citizens react according to the predictions of rational updating – that is, improving low evaluations when they learn that incumbents have high responsibility and downgrading evaluations after being told that incumbents have access to external resources. The Brazil experiment also provides evidence consistent with a key assumption of bounded accountability: when given the opportunity, citizens substitute exogenously driven performance for more informative shortcuts – such as party labels and programmatic differences.