Overconfidence and Information Aggregation

This paper explores the macroeconomic implications of overconfidence in noisy-information models. Relative overconfidence—where agents underestimate the precision of others’ signals—reduces welfare by intensifying the use of public information. Absolute overconfidence—where agents overestimate the precision of their private signals—can improve aggregate welfare in static settings by mitigating inefficiencies stemming from an overreliance on public signals. Welfare gains can be as large as 24% of the welfare in a rational benchmark. I generalize the result to a dynamic framework, showing that absolute overconfidence can partially counteract inefficiencies arising from strategic complementarities, improving long-run welfare.