Natalia Nollenberger, Núria Rodríguez-Planas, and Almudena Sevilla have a paper on this topic (click on the AEAweb pdf), presented at this year’s AEA meetings, the core result is this:
This paper investigates the effect of gender-related culture on the math gender gap by analysing math test scores of second-generation immigrants, who are all exposed to a common set of host country laws and institutions. We find that immigrant girls whose parents come from more gender-equal countries perform better (relative to similar boys) than immigrant girls whose parents come from less gender-equal countries, suggesting an important role of cultural beliefs on the role of women in society on the math gender gap. The transmission of cultural beliefs accounts for at least two thirds of the overall contribution of gender-related factors [emphasis added by TC].
I believe we will learn more yet when women stop improving, relative to men, at chess. But so far that has not yet happened.