Traders works on the floor of the New York Stock Exchange.

In his 1957 book, The Economics of Discrimination, the late Nobel Prize-winning economist Gary Becker argued that well-functioning markets work inexorably to eliminate racial, and other forms of, discrimination.

By Becker’s reasoning, if an employer, for example, had an ingrained prejudice against workers from a particular group, then that employer would be at a competitive disadvantage relative to employers who did not have the same prejudice. If, say, an employer believes—incorrectly—that workers from one group are inherently less productive, then he or she will offer members of that group lower wages. Meanwhile, his or her competitors who do not hold the same prejudice will offer members of the same group higher wages that are commensurate with their actual level of productivity.

Oversimplifying somewhat, Becker argued that in the long run, nondiscriminating employers will attract more productive workers and will grow, while those employers who discriminate will see their businesses wither and eventually die. This kind of theoretical argument has led many economists over the years to doubt the existence and especially the persistence of labor-market discrimination based on race, gender, religion, and other characteristics. A new paper by Mark Egan of the University of Minnesota, Gregor Matvos of the University of Chicago, and Amit Seru of Stanford University, however, joins a large body of more empirically oriented research that strongly suggests that discrimination remains an important feature of contemporary labor markets.

Egan, Matvos, and Seru have constructed a unique dataset containing information on all of the 1.2 million registered financial advisors operating in the United States between 2005 and 2015. Their data include information on each advisor’s gender, employment history, assets under management, professional licenses held, and industry exams passed, as well as records of any formally reported customer disputes and related disciplinary actions. Misconduct, as measured by the authors of the paper, included churning accounts, breaches of fiduciary duty, fraud, negligence, risky investments, and other forms of financial wrongdoing. Each year, about 0.7 percent of male advisors and about 0.3 percent of female advisors were involved in these forms of misconduct.

The three researchers use the new dataset to ask whether firms treat male and female financial advisors differently after they have been accused of misconduct. The data point solidly in the direction of employer discrimination against female financial advisors. Relative to male advisors, female advisors were less likely to engage in any kind of reported financial misconduct, were less likely to be repeat offenders, and when women did engage in misconduct, the infractions were less costly to the firms where they worked.

Yet women who engaged in misconduct were, on average, subject to much harsher punishments. After engaging in misconduct, female advisors were 20 percent more likely than male advisors to lose their jobs and were 30 percent less likely to be hired subsequently by another financial firm. Women who did manage to land a new job as a financial adviser took significantly longer than their male counterparts to find their new job.

Egan, Matvos, and Seru painstakingly eliminate a host of alternative explanations for these results. Maybe the female advisors were less productive than their male counterparts? The results hold even after controlling for three reasonable proxies for advisors’ productivity: assets brought to the firm, value of assets under management, and an independent, proprietary quality assessment available for a large subset of their sample. Maybe the women have less experience or more career interruptions? Again, the results hold even after controlling for on-the-job experience. Maybe the female advisors have fewer formal qualifications or certifications? No, not the case.

Their analysis leads the researchers to conclude that there are “large and pervasive differences in the treatment of male and female advisers” and the “financial advisory industry is willing to give male advisers a second chance, while female advisers are likely to be cast from the industry.”