The Impact Of Overestimation On US Trading Patterns And Diversification Of Portfolio Holdings

Abstract

This paper examines whether US households that overestimate their financial knowledge trade more actively and hold less diversified portfolios. Using microdata from the 2022 Survey of Consumer Finances, a nationally representative sample of 4,595 US households, an overconfidence gap is constructed by subtracting the z-scored financial literacy score (FINLIT) from the z-scored self-rated financial knowledge score (KNOWL). OLS regression models are estimated separately across five imputed datasets and combined using Rubin's rules. Results consistently support the first hypothesis: a one-unit increase in the overconfidence gap is associated with a 12.3% increase in trading activity among brokerage-account holders, robust to demographic, wealth, and risk controls. The second hypothesis is not supported; more overconfident households hold significantly more (1.52 unique stocks) rather than fewer distinct stocks. This finding is interpreted as reflecting a higher willingness to participate in individual equity markets rather than true risk-reducing portfolio diversification. These findings extend prior literature by testing overconfidence within a representative population sample rather than single-brokerage customer data, and suggest that policy aimed at improving financial self-calibration, not just financial knowledge, may reduce the welfare costs of overconfident investing.

Read full paper (PDF) →