The stock market appears not to discriminate against women CEOs in the form of under-estimating the stock of their firms. However, there are so few female CEOs that the jury’s out for now.

In Diagnosing Discrimination: Stock Returns and CEO Gender, Justin Wolfers uses the stock market to determine if discrimination is the principle reason for the disproportionately small number of female CEOs. Financial markets provide a constant stream of the market’s perception of firms’ values. Therefore, if female-headed corporations are under-estimated due to discrimination, then these firms would outperform expectations and yield additional returns. From S&P 1500 firms from 1992 to 2004, Wolfers finds no systematic difference in returns to the holding of female-run firms versus those run by men. He concludes, however, that results expose the weak statistical power of his test, rather than the inference that financial markets under-estimate female managers. Essentially, the results are out until more females become CEOs of large corporations.

Women make up half the population, yet between 1992 and 2004, only 1.3% of CEO-years were worked by women. This phenomenon is called a “glass ceiling” for females striving to become the heads of corporations. However, the number of female CEOs has been creeping up through time, from a low of 4 in January 1992, to 34 in December 1994. Three sets of explanations have been offered to account for the lower proportion of women CEOs:

1. There may be unobserved differences in productivity or preferences that are correlated with gender (i.e. There is a legitimate reason why females are underrepresented.)
2. Discrimination: Animosity by co-workers or customers may decrease the firm’s marginal revenue product if it promotes women, or the employer may be willing to accept lower profits in order to avoid promoting women
3. The ability of women is systematically mis-assessed. If consistent and systemic under-assessment of female-head firms occurs then the undervaluing of these firms will continue.

The logic behind Wolfer’s tests is simple. Equity markets traders constantly assemble information regarding companies and are influenced at least partly on their assessments of the ability of the CEO. If expectations about the ability of women are systematically biased and discriminatory, then continually betting on strong female performance would yield excess returns because these firms would consistently beat the lower predictions by biased analysts.

Wolfers uses Execucomp data from 1992 of 2004 to test whether the market detects the discrimination of female CEOs. His data identified a total of 64 female CEO’s and 4175 male CEOs. On the surface, he finds that samples of male- and female-headed firms are surprisingly similar in terms of size, book-to-market value, price-earnings ratios, firm-level betas and even industry compositions. While there are some differences—men tend to head slightly larger and more cyclical firms while women are over-represented in retail trade and information sectors—he stresses that these differences should not be exaggerated given the very small sample of female CEOs.

Wolfers uses the three approaches to measure differential returns between men and women-run firms:

1. portfolio returns- These are returns to holding zero investment portfolios that follow a strategy of buying female-headed firms and shorting male-headed firms in portfolios which weigh positions proportionately to market capitalization two months prior.
2. Fama-Macbeth regressions- Fama-Macbeth approach allows for the control of many firm characteristics like whether the firm is traded on the Nasdaq, whether it is a member of the S&P500, its book-to-market ratio and dividend yield, its stock price, the natural logs of market capitalization and trading volume
3. matching estimates- For these the author did a head to head comparison of female-headed and male-headed firms that he had matched. The matched firms were in the same narrow industry and had a similar market capitalization.

Can you profit by buying female companies? Does the stock market discriminate against women CEOs? With all three tests, the difference in returns between female- and male-headed firms is not statistically discernible from zero. These results are consistent with the hypothesis that markets do not systematically under-estimate female-headed firms.

Wolfers also realizes that the true ability of a CEO is revealed during the first few years of their tenure Any mis-assessments, such as those caused by discrimination, should be reversed during that period. Therefore, one might expect such the effect of discrimination to be most strongly present in the first few years of a CEO’s tenure. He tests this hypothesis and also finds no statistically significant evidence of this, either.

Wolfers concludes that the hypothesis that the stock market reflects the discrimination of female CEOs cannot be rejected because the standard errors are too large. There may be substantial bias against female CEOs, but more data is needed to confirm this. The best conclusion from this analysis is that differences in long-term returns to holding stock in female- and male-headed companies cannot yet falsify the null hypothesis that investors under-, or over-estimate the ability of female CEOs. He determines that in his current data analysis, differences in annual returns of 5% may not be detectable. He calculates that in order to get the standard error down to 1%, we would need around ten times this sample size.