We match administrative panel data on portfolio choices with survey measures of financial literacy. We observe that, controlling for portfolio risk, most literate households experience 0.4% higher annual returns than least literate households. We then show that more literate households display...