Economics, Entitlements

Public-Private Pay Divide: Part II

In a recent article for The American and an earlier blog post, I discussed a study for the Center for State and Local Government Excellence which argues that state and local government employees are underpaid relative to private-sector workers. My comments on the paper, by economists Keith Bender and John Heywood of the University of Wisconsin-Milwaukee, looked beyond pay to the relative generosity of pension and retirement health benefits, arguing that these could make up the difference.

Here, however, I want to look at the basic results for salaries alone. Using data from the Current Population Survey, Bender and Heywood concluded that state and local workers, on average, received salaries around 11 percent below those of otherwise-similar private-sector employees. In work with Jason Richwine of the Heritage Foundation, we’ve produced similar results.

But there’s one factor that is hard to account for: the role of labor unions. A much higher percentage of public-sector employees than private-sector employees are unionized. Around 40 percent of state employees and 51 percent of local employees are covered by union contracts, versus only around 9 percent of private-sector workers. The chart below is recreated from the Bender-Heywood study, showing how different personal attributes tend to increase or decrease average earnings. As in the private sector, males, whites, educated workers, and older workers tend to earn more than others. These coefficients won’t perfectly predict individuals’ wages, as there are factors that can’t be measured or accounted for, but equations like this can account for much of the variation in wages between different employees.


The coefficients of two variables are of interest here: “state/local employee” and “covered by union contract.” All other things equal, a public-sector worker receives a salary 11 percent below a private-sector worker. That’s the paper’s main takeaway result: that state and local employees are underpaid.

But the unionization variable shows that individuals who are covered by a union contract tend to earn around 16.2 percent more than those who are not. Here’s where things get tricky, since it’s not clear what’s driving state/local pay: the fact that the government is the employer or the fact that much of the workforce is unionized. Since the two tend to go together, it’s not easy statistically to disentangle them.

We can say two things, however: First, the pay gap for union workers is larger in the private sector than in state and local government, so the population-wide results in the Bender-Heywood paper are driven in part by factors that have nothing to do with public-sector pay.

Second, if unionization predictably leads to higher pay, then a government that permits its employees to unionize is implicitly giving its workers higher wages. So in a sense, unionization isn’t different than other conditions of employment that may be more favorable to public-sector work, such a higher benefits, greater job security, and so on.

A reader of the Bender-Heywood report shouldn’t conclude that public-sector workers are like the typical American worker. Rather, they’re more like the typical unionized worker, which is a little bit different.

Comments are closed.