Purpose The paper investigates how investor overconfidence affects stock price crash risk. Design/methodology/approach Following Adebambo and Yan (2018), we use mutual fund data from Thomson Financial, CRSP Survivorship Bias Free Mutual Fund Database and Morningstar Direct to construct our investor overconfidence proxy. We then conduct our analysis using the regression method in the US market for the sample period between 1988 and 2018. Findings We find that managers, to respond to unrealistic expectations from overconfident investors, are more likely to withhold bad news and overinvest, which increases stock price crash risk. Furthermore, firms with overconfident investors are more likely to have breaks in a string of consecutive earnings increases and engage in earnings management. Employing the Regulation SHO Pilot Program (Russell Index Reconstitution) as exogenous shocks to valuation (the participation of active investors), we find that the relation between investor overconfidence and stock price crash risk is more pronounced among overvalued firms (firms with a higher share of active investors). Finally, we show that strong corporate governance can discipline managers against catering for overconfident investors. Originality/value While existing literature has focused on how investor overconfidence increases stock price crash risk via excessive trading activities, we show a very different mechanism through firm's disclosure response to investor overconfidence.