Research

I develop a measure of managerial attention reallocation using earnings-call transcripts from 2007 to 2023. Using BERTopic, I compare the distribution of topics in each firm's management presentation with its prior communication pattern. Unusual attention reallocation predicts negative abnormal returns, elevated crash risk over horizons of up to two years, persistently lower long-run returns, greater analyst forecast dispersion, and higher option-implied uncertainty. Attention reallocation also predicts subsequent adverse firm developments: a one-standard-deviation increase is associated with an 8.6% increase in bad-news 8-K filings over the following 90 days, as well as a greater incidence of ESG incidents. Markets and analysts do not fully incorporate the signal at the time of the call, suggesting that changes in managerial attention allocation contain information about future firm outcomes that is only partially reflected in prices.

We develop a novel machine-learning technique analyzing earnings call transcripts to measure whether a manager's focus is on shareholders or external stakeholders. We find that managers focusing on shareholders generate significantly higher stock returns in subsequent periods. However, these higher returns do not come to the detriment of other stakeholders; we find that shareholder-focused firms experience far fewer adverse environmental and social incidents, and those that do occur are less severe. Using natural disasters as an instrument for managerial focus and exploiting exogenous shifts in firm orientation induced by CEO transitions, we find that a shareholder focus causally reduces adverse E&S outcomes across all horizons examined. We conclude that a shareholder-focused strategy creates maximum value for both shareholders and stakeholders.

In a sample of ballot items for which proxy advisor recommendations oppose management, we find that the likelihood of a mutual fund actively voting (deviating from proxy advisor recommendations) is significantly higher when the fund has larger monetary incentives tied to the portfolio holding. To address endogenous portfolio selection issues, we find the same result in a sample of passive index funds for which portfolio allocations are exogenous. Firms that are more commonly held by less incentivized funds have lower CEO delta and lower CEO turnover-performance sensitivity.

We construct a measure of corporate tax attention from the text of quarterly earnings calls for U.S. public firms over 2007 to 2023 using an unsupervised topic model. The measure captures within-firm deviations in the attention corporate executives pay to tax related issues from the firm's own historical same-quarter baseline. We find that elevated tax attention predicts lower cash effective tax rates over horizons of up to three years and generates positive announcement returns. A triple difference-in-differences design exploiting the Tax Cuts and Jobs Act of 2017 shows that firms with high pre-reform tax attention and high effective tax rates captured the largest post-reform reductions in effective tax rates. Analysts revise earnings forecasts upward and realized earnings subsequently exceed their revised forecasts, consistent with partial incorporation of a signal that is available but costly to extract.

We employ two novel datasets to evaluate whether the Paycheck Protection Program (PPP) met short- term and longer-term goals. One dataset matches PPP with small business members of the National Federation of Business (NFIB) that report their financial constraints and their employees’ economic constraints. We find more alleviation of both constraints for PPP recipients, consistent with short-term goals. We employ data on post- versus pre-crisis county employment, wages, and new businesses to test the longer-term goal of improved community crisis recoveries from the short-term constraint relief. We find to the contrary – worse longer-term recoveries in counties with more short-term constraint relief.