How Costly is Diversity?

Introducing affirmative action causes a substantial increase in the number of female competitors, and this supply effect reduces the cost of requiring equal representation of women.

Despite decades of striving for gender equality, there are still large differences between men and women in the labor market. Women are more likely to hold clerical or nurturing jobs while men are more visible in manufacturing. Across fields, men are disproportionately found in professional and managerial occupations. Even among graduates of top tier business schools, female MBAs are more likely to work in the non-profit sector, to work part time, or to drop out of the work force entirely. One theory suggests that women are underrepresented in many high-profile jobs, and across entire professions, because of the way they respond to competition: men are eager to compete, while women often shy away from competitive environments.

In How Costly is Diversity? Affirmative Action in Light of Gender Differences in Competitiveness (NBER Working Paper No. 13923), authors Muriel Niederle, Carmit Segal, and Lise Vesterlund devise a series of experiments to investigate how affirmative action might affect participants' willingness to compete, and at what cost. Specifically, they observe 42 men and 42 women at Harvard who competed in a timed tournament involving adding series of 5 two-digit numbers. The tournament rules under affirmative action required that out of two winners, at least one must be a woman. Niederle and her co-authors find that when women are guaranteed equal representation among winners, as in this case, more women and fewer men enter competitions - and the response is even larger than one might predict given the changes in the odds of winning. The response is explained by the affirmative action competition being more gender specific: for example, to win the competition a woman only needs to outperform the other women.

Interestingly, both beliefs about relative performance and the willingness to compete will change in more gender-specific competitions.

The researchers then ask how costly it is to insure that women be represented equally among those who win competitions - in other words, what is the cost of affirmative action? In particular, how much lower will the performance threshold be for women? How many better-performing men will have to be passed by to hire a woman? To what extent will reverse discrimination arise?

The authors find that introducing affirmative action causes a substantial increase in the number of female competitors, and this supply effect reduces the cost of requiring equal representation of women. The costs of affirmative action depend on how much lower the minimum performance threshold must be to secure gender parity, compared to performance of a group in which gender is not taken into account. If there were no change in male-female entry, then equal representation of women would result in lower minimum performance for women. However, the change in tournament entry implies that women are better represented among the set of entrants, and in particular that more high performing women are in the applicant pool. Thus, it becomes much less costly to achieve equal representation, and the resulting minimum performance threshold changes very little if at all under affirmative action.

-- Lester Picker