To measure the gender bias, I examine stock returns around CEO announcement dates using a modified event-study methodology and ExecuComp, CompuStat, and CRSP data from 1992 to 2016. Essentially, the behavior of stock returns around (female) CEO announcement dates should reveal a gender bias if CEO gender affects investors’ perceptions of the risks to which the firm—and, thus, its earnings—are exposed. This is because investors price stocks based on risk: riskier stocks require lower prices and, thus, higher returns.
This research question is important, because a gender bias in financial markets would imply an inefficient allocation of capital, which necessarily constrains economic performance and social welfare more generally. Existing event studies are inconclusive as to whether or not such a bias exists. In part, this is because earlier studies necessarily relied on inherently small samples of female CEOs (N = 70 or fewer), leading to inconclusive results. However, recently, the number of female CEOs has increased. My dataset, which extends to 2016 and has over 90 female CEO observations, and the modified event study methodology I employ afford my results statistical power that earlier studies could not achieve. Preliminarily, I find gender biases exist—investors perceive female CEOs as relatively risk averse—and these biases are time varying.