Pay gap reporting

The NBR's Dita De Boni asked me for comment on the push for companies to report their gender pay gaps. Her story's now up here ($).

As always, space constraints mean she could only use a bit of what I'd sent through (all used fairly). Here's my full comment, for those who were interested:

Pay equity reporting mandates are fraught where worker compensation is more than just pay. Roles vary considerably, and people also vary in how much they value pay as compared to flexibility in hours or location of work or being on-call.

If a worker’s total compensation bundle isn’t just pay, but also include whether they’re able to work flexible hours, or whether they’re expected to be on-call at odd hours, or whether they’re regularly expected to put in overtime or exempt from having to do so, then measuring only one aspect of pay and reporting on it could provide a rather inaccurate picture.

The best evidence we have on gender wage disparities, at least out of the United States, suggests that a lot of existing differences in pay are explained by those kinds of differences in overall job conditions. Claudia Goldin’s work in particular shows a substantial pay penalty for workers putting in fewer than 40 hours per week, regardless of gender. But if there are gender differences in willingness to work longer hours, that would show up as a gap in measured pay. Is that the kind of gap that policy or companies should care about? If so, what should they do? If there are real advantages for the company in some workers being willing to work longer hours, forcing that to pare back could hurt overall productivity.

And what evidence we have out of the US is consistent with that worry. A recent NBER Working Paper showed that Danish legislation requiring publication of gender-disaggregated wage statistics did reduce the measured wage gap – by slowing wage growth among male employees. The reduction in pay growth for male employees was matched by a decline in firm productivity, so overall firm profitability was not much hit (pay less for less work) – but is that really a good outcome?

Disclosure legislation could easily have a lot of perverse consequences. For example, consider a firm making staff diversity a priority in new hiring. If new hiring tends to be at more junior levels, because there is usually greater churn at junior levels, then the measured ethnic or gender pay gap would increase because of that round of new hiring. Conversely, if the firm wanted to reduce its measured pay gap, it [could] hire junior staff from groups that the company pays more on average, to help bring down that average – potentially at the expense of diversity in overall staffing.

The exercise is inherently fraught because compensation is complex.