Education and generalized trust: a test of two mechanisms

Abstract Generalized trust supports social cooperation and institutional performance. Formal education is often assumed to foster trust, but competing theories make opposite predictions: the social-intelligence view holds that education sharpens belief accuracy about others’ pro-social behavior, while the selection/exposure view expects a systematic optimistic bias. We test these mechanisms using an original survey in which 800 respondents in Spain estimated the return rates of “lost wallets” across four countries, based on results from previous cross-national field experiments. Respondents misjudged others’ honesty (overestimating returns when no money was involved and underestimating them when money was present) and predicted a decline in honesty as the amount of money in the wallet increased, while actual return rates rose across those conditions. Higher education does not correct these errors or produce consistent optimism. The findings thus challenge both explanations and suggest that the education–trust link operates through other, yet-to-be-identified, institutional or normative pathways.