Print
E-MailAntoinette Schoar is the Stewart C. Myers-Horn Family Professor of Finance and Entrepreneurship at the MIT Sloan School of Management; Kelvin Yeung is a PhD student at the Cornell University Samuel Curtis Johnson Graduate School of Management; and Luo Zuo is Associate Professor at the Cornell University Samuel Curtis Johnson Graduate School of Management. This post is based on their recent paper.
In the paper The Effect of Managers on Systematic Risk, we ask whether top manager-specific differences account for part of the unexplained variation in traditional asset pricing models. A key principle in asset pricing theory is that investors are compensated for bearing systematic risk, but not idiosyncratic risk. Drawing on this insight, empirical asset pricing models decompose stock return variability into systematic and idiosyncratic components. The systematic risk of a stock is determined by its beta, which measures the sensitivity of the stock’s return to common risk factors such as the market factor. A large literature investigates the determinants of beta, but a general conclusion from these studies is that a large amount of variation in systematic risk cannot be explained by firm-, industry-, or market-level variables.
Tracking the movement of top managers across firms, we document the importance of manager-specific fixed effects in explaining heterogeneity in firm exposures to systematic risk. We show that these differences in systematic risk are partially explained by managers’ corporate strategies, such as their preferences for internal growth and financial conservatism. We follow the approach of Bertrand and Schoar (2003) to document that such person-specific styles explain a significant amount of variation in firms’ capital structures, investment decisions, and organizational structures. The notion that CEOs differ in their styles is reinforced by Bennedsen, Pérez-González, and Wolfenzon (2020) who exploit hospitalizations to examine variation in firms’ exposures to their CEOs. Similarly, a vibrant literature suggests that managers’ personal traits play a role in shaping their management approach. Our results suggest that managerial style explains a substantial fraction of the variation in both idiosyncratic risk and systematic risk.
Starting from a single-factor market model, we decompose stock return variability into systematic and idiosyncratic components. Total risk (TVOL) is the standard deviation of a firm’s daily stock returns within its fiscal year. Our measure of systematic risk (β_MKT) is the slope coefficient on the excess return of the market portfolio and our measure of idiosyncratic risk (IVOL) is the standard deviation of the residuals. Although we focus on the single-index model, our results generalize to other empirical asset pricing models such as the Fama-French Six-Factor Model.
In our base model, we regress each measure of risk on firm fixed effects and year fixed effects. Then, we add manager fixed effects to our base model and examine whether the manager fixed effects have incremental explanatory power. Intuitively, we test whether systematic risk is correlated across at least two firms when the same manager is present, controlling for time-invariant firm characteristics (firm fixed effects) and year-specific cross-sectional effects (year fixed effects).
Our results indicate that managerial style is an important determinant of systematic risk. We observe a 7.16% increase in adjusted R2 when we add manager fixed effects to the model with β_MKT as the dependent variable, which translates to a 16.67% increase relative to the base model. For comparison, we observe a 4.43% increase in adjusted R2 when we use IVOL as the dependent variable and we observe a 4.56% increase in adjusted R2 when we use TVOL as the dependent variable. Adjusted R2 increases by 7.19% and 7.36% relative to the base model for IVOL and TVOL, respectively. Furthermore, the frequency of significant manager fixed effects is far greater than would be expected under the null hypothesis that managerial style is not a determinant of systematic risk: 49.26% of the manager fixed effects are significant at the 10% level, 43.18% of the manager fixed effects are significant at the 5% level, and 35.01% of the manager fixed effects are significant at the 1% level. In terms of economic magnitude, hiring a manager at the 25th percentile leads to a 0.201 decrease in β_MKT and hiring a manager at the 75th percentile leads to a 0.161 increase in β_MKT.
To understand the channels through which top managers affect systematic risk, we analyze whether specific firm-level decisions that managers undertake translate into differential loadings on systematic risk. We first examine whether manager fixed effects on the real side of the firm, such as capital structure decisions and other firm policies, explain manager fixed effects on beta, β_MKT. We conduct factor analysis which shows that these manager fixed effects vary along three dimensions: internal growth, financial conservatism, and external growth. Manager fixed effects on β_MKT are positively related to managers’ preferences for internal growth and negatively related to managers’ preferences for financial conservatism, but do not vary systematically with measures of external growth.
To analyze the importance of managerial strategies in explaining manager fixed effects on β_MKT, we also rerun our regressions, but directly control for time-varying firm characteristics. This specification directly absorbs changes on the real side of the firm that top managers might be undertaking. When we control for time-varying firm characteristics, adjusted R2 increases by 6.86% (compared to 7.16% in the benchmark specification). Our results indicate that time-varying firm characteristics partially explain manager fixed effects on β_MKT, However, manager fixed effects have significant explanatory power after controlling for time-varying firm characteristics. The above results suggest that managers affect their firm’s loading on systematic risk via the corporate strategies they adopt, but a large amount of the variation in manager fixed effects on β_MKT remains unexplained by these observable dimensions.
In a next step, we analyze whether observable manager characteristics explain manager fixed effects with respect to systematic risk. We find that manager fixed effects on β_MKT are related to managers’ early-career experiences. On average, the signed effect on β_MKT is 0.240 smaller for managers who originally entered the labor market during recessions. These results are in line with the findings in Schoar and Zuo (2017) that managers who enter the labor market during recessions adopt more conservative corporate strategies, such as lower SG&A or reduced leverage. We do not find evidence that other characteristics like age or gender are related to manager fixed effects on β_MKT.
To shed light on the settings in which managerial style matters more for systematic risk, we analyze whether certain firm and market conditions moderate the effect of managers on beta. We find that manager fixed effects on β_MKT are more pronounced in smaller firms compared to larger firms, which is consistent with the notion that managers have more discretion over firm outcomes in smaller firms.
The complete paper is available here.
