I read Michael Luo and Michael Cooper’s report on rival studies concerning the putative compensation gap between public and private sector workers with great interest. One friend of mine suggested that Luo and Cooper found no indication that public workers are overpaid. I found this odd, as I was not under the impression that Luo and Cooper were trying very hard, not least because they relied on data that left out benefits, public workers, and other pieces of information that they explicitly say would be “crucial for a fuller comparison.”
As I’ve tried to explain in this space, the problem with public sector compensation tends to be its structure, i.e., the manner in which compensation is determined. This shouldn’t be too hard a concept to understand, and I suspect, in my more cynical moments, that some of the folks writing on these issues are deliberately obfuscating the important issues. Given that the terms “overpaid” and “underpaid” have normative content, it seems pretty clear that at least some of our friends will never agree that any public workers are “overpaid.” Many people believe, for example, that all employers should be obligated to pay a “living wage,” regardless of the market-clearing wage. I tend to think that while public wage subsidies might be appropriate, there is a danger that a price floor for labor will in some cases exacerbate unemployment. It should come as no surprise that partisans of living wage regulations believe that public wages are a political issue rather than a management issue, i.e., a question of how to deliver the highest-quality services at the lowest cost to the taxpayer.
The Luo and Cooper article on the 25th relies on crude comparisons of workers by education levels:
“Because the public sector is much more likely to be highly educated, we would fully expect them to earn more on average because of that, just like we would expect somebody with a master’s degree to earn more than somebody with a high school education,” said Keith A. Bender, an economics professor at the University of Wisconsin-Milwaukee, who has studied compensation in the public and private sectors.
Would we expect somebody with a master’s degree in social work to earn more than somebody with a bachelor’s degree yet who has experience that can’t be captured in a raw measure of years of educational attainment in a formal setting? Moreover, the public sector, and in particular public schools, often rewards the attainment of graduate degrees with a step increase in compensation, a practice that is far less common among private firms.
Keith A. Bender’s remarks reminded me of the work of Thomas Lemieux, an economist who has closely studied the interaction between post-secondary education and wage inequality. Lemieux, drawing on the work of Mincer and others, notes that within-group wage dispersion is higher for more educated groups.
Since the early 1980s, the U.S. workforce has grown more educated and experienced, two factors strongly associated with more within-group wage dispersion. Most of the growth in residual wage dispersion is due to the increasing weight put on groups of workers with more dispersed wages, as opposed to growth in wage dispersion within these narrowly defined groups of workers.
While we might generally expect that workers with master’s degrees will earn more than workers with bachelor’s degrees, this tells us little about the pattern of wage dispersion within the class of workers with master’s degrees, and indeed within the class of workers with master’s degrees in a particular discipline.
One of the reasons I’ve been so dispirited by the coverage of the pay gap is that I’ve seen so much coverage of the EPI study and so little coverage of the response by Andrew Biggs of AEI. In an article published on the 26th, Luo and Cooper briefly cite Biggs:
A recent study, for example, by Keith A. Bender and John S. Heywood, economics professors at the University of Wisconsin-Milwaukee, for the Center for State and Local Government Excellence and the National Institute on Retirement Security, found that state workers earn 11 percent less than their private sector counterparts, even including benefits, after other factors like education and age are accounted for.
But the economists also controlled for unionization, meaning they used statistical methods to remove it as a factor in their comparisons. That significantly affected their findings, as noted in a critique by Andrew Biggs of the American Enterprise Institute for Public Policy Research.
Unfortunately, they don’t reference Biggs’s more recent work on the subject. And again, Luo and Cooper continue citing a study that ignores benefit levels. Rather than shedding light on the compensation gap, Luo and Cooper have written articles that are maddeningly repetitive and unclear.