Research

I demonstrate that institutional equity investors influence corporate debt financing through their dual roles as informed traders and shareholders. These roles operate through two distinct channels: an information transmission channel, where institutional equity investors transmit information to debt investors, and a benchmarking channel, where investors’ benchmarking mandates alter the discount factors used to value firm cash flows. Empirically, I provide causal evidence for both channels by exploiting a mutual fund disclosure reform and using equity benchmarking intensity as an instrument. Theoretically, I develop a dynamic general equilibrium model that incorporates both channels and endogenizes firms’ financing choices and investors’ shareholdings. I show that the information channel accounts for over 20% of the observed relationship between institutional ownership and corporate debt issuance, with the associated information costs amounting to roughly 5% of equity value. Furthermore, an increase in benchmarking intensity comparable to Russell 2000 inclusion amplifies the relationship by 9%. These results highlight the importance of cross-market investor linkages and firm responses to investor characteristics in understanding financing outcomes.

Machine learning algorithms are known to outperform human analysts in predicting corporate earnings, leading to their rapid adoption. However, we show that leading methods (gradient boosting, neural nets, ChatGPT) systematically exhibit over- and under-reaction biases similar to human forecasts. The behavioral bias is primarily due to biases in the training data and we show that it cannot be eliminated without compromising accuracy. Analysts with machine learning training overreact much less than do traditional analysts. Our findings suggest that AI tools reduce but do not eliminate behavioral biases in financial markets. We provide a model showing that there is a tradeoff between predictive power and rational behavior.