This study develops a theoretical model to analyze the asset substitution problem over insurance decisions. We find that agency conflict is related to a firm’s risk level and capital structure. In particular, at the optimal leverage, agency conflict occurs only when the risk level is relatively high, which explains why insurance covenants are typically for significant pure risks. Moreover, when the risk level is specified, agency conflict over insurance decisions occurs within a specific leverage range. This is consistent with the findings of some research on the asset substitution problem over speculative risk choices. In addition, we consider premium loadings and conclude that full hedging is not a firm’s optimal risk management strategy, contributing to the literature on optimal hedging decisions with transaction frictions. Our framework with premium loadings can also explain many insurance phenomena, such as risk retention for small losses and subsidies for catastrophe insurance.

How prior trust moderates investor responses to restatements is unknown. We examine how societal trust affects the changes in institutional investors’ shareholdings around a restatement. We consider two competing hypotheses based on the erosion of trust and confirmatory bias. We find the change in institutional investors’ shareholdings around a restatement is more negative for investors from high trust areas compared to low trust areas, consistent with an erosion of trust where high trust institutional investors view the restatement as a violation of trust. Further analyses show that our findings vary with the regulatory or economic environment, type of institution, and type of restatement. Our results are also robust to different tests that address endogeneity and use alternative societal trust measures. Overall, we contribute to the literature by examining the role of societal trust in a dynamic setting where investors’ trust-based beliefs about the credibility of accounting information are not realized. Data Availability: GSS Sensitive Data Files are not available from the authors. Persons interested in obtaining these data should contact the GSS at [email protected]. Other data are available from the public sources cited in the text.