Reporting on pay gaps

Last week saw a bit of press about a Ministry of Women's Affairs paper on salary gaps, noted here. The paper is now available.

Sometimes, reading these sorts of things, I ask what I'd do if I were asked to grade the paper as an honours project here in the Department. In this case, I'd be heading over to the student's supervisor's office and asking what went wrong: did the supervisor assign a reasonable project but then unforeseen data limitations make a hash of it; did the student only start working on it the week before it was due; did the supervisor just not provide enough guidance for a weaker student?

In fairness, the paper notes that it's just "a preliminary step" in "understanding the differences in pay between male and female workers with tertiary education". So I probably ought to be viewing this as more akin to the first presentation given by an honours student a few weeks into the project. But we'd never think about putting out press releases about a student's preliminary work.

One obvious problem with the study, anticipated in the prior post, is that gender imbalances across disciplines combined with different average salaries in different disciplines can quickly lead to findings of pay gaps even if all workers in the same area are paid differently. Economics is lumped in with languages and library studies. Any guesses as to gender and salary differences in those fields?

A second problem is that the paper uses only money salary and doesn't include nonpecuniary benefits. This is of course harder. But more flexible work arrangements are usually more valued by women than men and come at a price in salary. Total compensation, as judged by the worker herself, may well be better, but we're only measuring the salary part.

Third, the paper notes high drop-off in earners in the sample. They think this is mostly due to young folks going overseas to get some experience. And a lot of young Kiwis do this. But more men than women: that's why we have more young women around than men (potential immigrants to NZ take note). If relatively more men drop out to go overseas while relatively more women drop out for childbearing, that will also explain some of the gap: finding a permanent replacement for a young worker going overseas (who hasn't yet acquired a lot of firm-specific knowledge) is less costly for a firm than dealing with temporary replacements and holding the position open for up to a year. The authors think this unlikely to be what's going on as average age of childbearing is 28, but it isn't the average that matters; rather, it's the proportion of women in the ages studied that do exit the workforce for childbearing. I haven't the numbers at hand, but even if the average is 28, you could easily have a large fraction (less than half, of course) leaving the workforce at age 24 or 25. You've gotta take the integral under the curve to the left of the cut-off age for the study.

I'd say that Ministry of Women's Affairs was a bit silly to promote so preliminary a piece of work. I'd say that, except that the media here seems completely incapable of critically evaluating a produced number. They get the press release, turn it into a headline, then ask people how the number makes them feel. Expecting more of the TV news just sets you up for disappointment. The Ministry would reasonably have expected this; getting folks more riled up about pay gaps is a good way to help their department push for more funds relative to other departments. I'd view this more as a salvo in a war of public opinion to draw greater funds to the department than as any kind of serious attempt at research. 'Cause if I were to grade it as research, well...it ain't there yet.