Economics' gender problem

For a long time, I assumed that the underrepresentation of women in economics was a legacy problem. There used to be prohibitions against (married) women in the workplace. Older colleagues treated their female assistants differently and derided their female co-workers. I thought that these attitudes would disappear one retirement at a time, and the genders would eventually balance. How wrong I was.

The IDEAS/RePEc data on female representation by cohort shows a steady rise from 6% in 1971 to 22% in 1998. It has been stuck around that number for the last 25 years.

I read Alice Wu’s paper, but Erin Hengel’s really caught my attention. The ill-treatment of two female colleagues and fellow unionists reinforced the message but did not really drive it home. I had looked at the data for Energy Economics and did not find a gender gap. How wrong I was.

Erin quickly noted that I had not properly accounted for composition. So I gave her access to the entire electronic archive of the journal. (I also offered the paper archive of earlier years but she wisely declined.) We convinced 32 other journals to share summary data for their submissions. The — now proper — analysis shows that female-authored papers are delayed by one or two months.

The paper also shows that I was wrong a third and fourth time. My prior was that there were a few outright misogynists in the pool of referees. Removing them would solve the problem. Instead, there is pervasive, low-level statistical discrimination. The working hypothesis is that inexperienced referees have seen so few female-authored papers in journals and at conferences that they treat such papers with extra suspicion. There is no easy fix to this.

My other prior was that more female editors would solve the problem. The evidence from the 32 journals is thin — only 32 journals after all, and fewer with a changing gender balance in the editorial board — but does not offer support for this hypothesis either.

The paper with Diane, Olga, and Erin only shows the delay at one journal. It also finds a slightly higher rejection rate for female authors. If, as is common, a paper gets rejected by a few journals before it is published, then the cumulative delay easily becomes six months — long enough to materially affect tenure decisions and promotions.

We have a real problem so, that will not heal itself over time. Female colleagues do need extra protection, when submitting papers and when applying for promotion or tenure. Statistical discrimination is a vicious circle — unfamiliarity breeds suspicion breeds unfamiliarity. Vicious circles become virtuous once broken. So that’s what we need to do.