Research

Abstract: Prior studies focus on how investors' sustainability preferences incentivize firms to reallocate resources from dirty to clean physical capital. However, the impact of investors' preferences on capital allocation depends critically on whether clean capital and dirty capital are substitutable. I develop a novel empirical strategy showing that dirty capital and clean capital are highly complementary. Theoretically, I explore firms' investment decisions, assuming that investors dislike carbon emissions through both risk and nonpecuniary utility channels. Given the current level of complementarity, investors' preferences have a limited impact on investment decisions, underscoring the need for technological innovation to address this production friction. 

Abstract: Realized default rates and losses are too low relative to high corporate credit spreads in the data. I explain this "credit spread puzzle" through a dynamic capital structure model with long-term bonds and disaster risk. In contrast to models with one-period bonds, long-term bonds serve as hedge investments under disaster risk concerns, generating different economic mechanisms. I explore the effects at firm level. Disaster risk affects corporate credit spreads through default risk, risk premium, and corporate capital structure. In disaster states, default risk dominates other channels due to firms' high likelihood of default. In normal times, the disaster risk premium induces lower optimal capital levels, which leads to higher leverage and credit spreads. When exploring the effect of real and financing frictions in the model, I find that both lead firms to act conservatively and reduce their leverage, giving rise to lower credit spreads. However, I also find that, after a disaster, financially constrained firms lose more equity value, and their credit spreads sharply increase in the model. 

Abstract: Investors can influence firms only through what outsiders can measure. We present a simple model in which sustainable demand creates a pricing wedge, but the firm chooses both a costly internal social action and the precision of disclosure about it. When disclosure is too noisy or too costly, the wedge does not translate into real action. We test this prediction using confidential U.S. Census administrative wage records to measure firm-year employee gender pay gaps. With rich fixed effects, socially responsible ownership is unrelated to pay gaps. The relation turns more negative when boards include female directors, and these firms talk more about gender equality in earnings calls. Our results suggest that information asymmetry limits investors’ ability to change hard-to-observe internal policies.

Findings: We estimate a structural model to demonstrate that the heterogeneity in investors’ relative risk aversion significantly explains the time variation in the aggregate equity premium empirically.