Commodity price shocks and misattribution of responsibility for the economy: observational and experimental evidence

Abstract

This paper examines correction of misattribution bias in the context of retrospective voting. We first report observational evidence from Brazil and Ecuador that is compatible with the existence of misattribution bias (i.e. attribution of economic outcomes caused by exogenous factors to presidents' competence). (Hemisphere 2018)We then conduct survey-experiments in both countries, in which we assess whether the provision of information about relative economic performance or about the occurrence of commodity shocks in uence respondents' assessment of presidents' competence. Results reveal that information has a small yet very consistent effect, across experimental designs and countries, which is nonetheless restricted to sophisticated voters|those that understand the appropriate normative rules of inference. For those that do not know these rules, new information is irrelevant. These results imply that misattribution bias is not driven solely by lack of information, but also by cognitive limitations to process the information that is available.