Itâs generally difficult to find real-world evidence that decreasing bias and hiring more diverse teams directly cause better performance. But because the sex of children is random â providing natural experimental and control groups â and because it affected the likelihood of hiring female partners, the study was able to compare the effect of diversity on performance. The researchers were Paul Gompers, a professor at Harvard Business School, and Sophie Wang, a Harvard economics Ph.D. student.
The probability of a venture firmâs succeeding, as measured by start-ups going public or being acquired for more than the amount invested, was 3 percent higher in firms in which investing partners had more daughters. The rate of return for the funds also increased 3 percent. The researchers controlled for factors like firmsâ size, age and demographic makeup.
âItâs a kind of pure test, to have a separate source of variation and find that diversity does good,â said Larry Summers, a Harvard economist and former Treasury secretary. He called it âa clever and important study that seems applicable to a variety of professions.â
Venture capital, even more than other industries, tends to be homogeneous â mostly white or Asian-American men with similar education and work backgrounds. The investors control billions of dollars and work in small teams that hire infrequently, so even small biases in hiring have industrywide effects. Only about 10 percent of new hires in the industry are women, and three-quarters of firms have never had a female senior investor, according to the data.
The researchers ran a calculation comparing firms, and found that on average, the presence of one more daughter instead of a son for each senior partner increased the probability of hiring a female senior investor over a male by 24 percent. In firms with more daughters, 12 percent of new hires were women, compared with 9 percent in firms with more sons.
The researchers used the Dow Jones VentureSource database to track venture-backed start-ups and their investors from 1990 to 2016. They created a data set of venture capitalistsâ children by using publicly available information, like alumni directories, and by sending personal email queries. The data set of children is not representative because they could not find data for every investor, but they included 1,403 of 3,329 people.
The researchers wanted to know whether working only with people of similar backgrounds is a positive thing, because it could make communication easier and more efficient, or detrimental, because it could lead to harmful groupthink. The latter appears to be true, Mr. Gompers said.
âIf you have five white guys who all went to Harvard Business School and worked at Google, itâs likely when they see an investment or problem or decision, they frame it in the same way,â he said. âHaving a diverse set of people really means youâre much less likely to have blind spots.â
Another reason for the results, the researchers said, could be that the pool of qualified women is so untapped that female hires tend to be of exceptionally high quality, which could lead to higher returns. Women also tend to have more diverse networks, so they have access to a bigger range of deals.
The benefits to the firms in the study seemed to come from a genuine decrease in gender bias from having daughters. Other research has shown that artificial increases in diversity, like quotas requiring that a certain share of executives are female, have little effect on businesses.
All executives and policy makers canât magically be given daughters. Still, the research demonstrates the value of including a diverse set of people in the room.
âIt becomes really hard to deny we probably would see very large effects of other experiences, too, like serving in the military, growing up in a disadvantaged neighborhood, going on a mission, being a first-time college graduate,â Ms. Sen said. âItâs a nice illustration of the potential benefits of lived experience.â