Middlebury economics professors Caitlin Myers and Jessica Holmes


What sparked the idea for a study linking testosterone to wages?

Jessica: Believe it or not, the seeds were sown during a sideline conversation at a Middlebury College women’s soccer game. I made an anecdotal observation to biology professor Jeremy Ward that females with twin brothers seemed to be more aggressive athletes than those without. (One of the more forceful players on the team had a twin brother.) Jeremy then shared the little-known fact (to us economists anyway!) that in both litter-bearing animals and humans, female fetuses are exposed to testosterone shocks from their male littermates, and these testosterone boosts are linked to greater risk-taking and assertiveness later in life. This was a “Eureka!” moment for Caitlin and me because we immediately saw the potential to contribute to an ongoing nature vs. nurture debate about why women tend to earn less than men. 

Can you explain more about the gender wage gap and how it relates to your research? 

Caitlin: The current gender wage gap is about 77 cents on the male dollar. Economists can explain about half of the gap by observable factors like part-time status, occupational choice and labor market experience. The remaining half is considered “unexplained,” but the usual suspect is labor market discrimination. Recently, social scientists have started to consider a third potential explanation: that perhaps it has something to do with differences between men and women in preferences and personality.  We’re interested in whether any such differences might be biological.  Is it possible that women are not wired to “Lean In,” and that this helps to explain the wage gap? 

Can you explain your basic premise and methodology? 

Jessica: Essentially, we use differences in prenatal exposure to testosterone between opposite sex twins, single sex twins and singletons to examine “nature’s” contribution to the gender wage gap. Twins provide us with a unique natural experiment (one that allows us to avoid randomly injecting a sample of pregnant women with testosterone to see how things turn out for their children!) The empirical test is simple. We know that females with a twin brother experience a greater testosterone “shock” in utero than those with a twin sister. We also know that prenatal testosterone wires the brain for more typically masculine preferences and behaviors including risk-taking and aggression. So we use econometric techniques to test whether twins who receive a prenatal testosterone “boost” earn higher wages later in life. 

Has anyone else ever taken this approach?

Caitlin: We are building on a large literature from biology, neuroscience and psychology. To our knowledge, we are the first to use this approach to look at economic outcomes and also the first to use closely-spaced singletons to control for potential socialization effects. Another group of economic researchers recently used a similar strategy to look at investment portfolio selection and financial risk-taking. 

It seems like an impossible task to find a large enough data sample that tracks twins from infancy through professional life. How did you go about that?


Were you surprised by any of your findings?

: We have the typical impartial mien of empiricists; it’s hard to surprise us! What we find particularly interesting about the results, though, is that they suggest that testosterone exposure benefits men in the labor market, but does not help — and may even hurt — women. Our results are consistent with mounting research that suggests that men and women are evaluated differently in the workplace. Men who are assertive and competitive are regarded as strong leaders and as such are more highly compensated and more likely to be promoted. Women who behave similarly, that is women who are ambitious, assertive and competitive, are perceived as competent but less socially skilled… and they are less likely to be promoted than less aggressive women. So a testosterone boost in utero boosts wages for men but has no effect (and may actually reduce) wages for women.

What’s next? Did this study suggest other avenues that you plan to pursue?

Jessica: We are investigating the possibility of using a large database of Swedish twins to test whether prenatal exposure to testosterone affects other outcomes such as cognitive test scores and occupational choice.