Quick Thought on Jeffrey Keefe’s Research on Public Employees

by Reihan Salam

Ezra Klein cites new research from Jeffrey Keefe on the compensation of state and local government employees:

The data (pdf) come from Rutgers’s Jeffrey Keefe, and he also ran “a separate calculation that controls for full-time status, education level, years of experience, age, gender, race, employer organizational size, industry, and hours worked,” which found that “public employees are compensated 2-7% less than equivalent private sector employees.”

Which makes sense: You never hear public employees say that they went into government for the money. But to make the more counterintuitive point, this is a fairly counterproductive conversation. We want really good regulators watching Wall Street. We want talented people teaching our children. We want our trash picked up by workers who want to keep their job and we want the DMV staffed by individuals who aren’t incompetent. That isn’t to say we should pay every public worker millions and millions of dollars, but very few go into government work because they’re of a charitable bent. It’s a job, like any other, and it attracts talent only by paying it well.

This is an interesting and worthwhile take. After reading Keefe’s report, however, it’s not clear that paying for talent is the governing paradigm for public sector compensation:

On average, state and local public-sector workers are more highly educated than the private-sector workforce; 54% of full-time state and local public sector workers hold at least a four year college degree compared to 35% of full-time private-sector workers. State and local governments pay college-educated labor on average 25% less than private employers. The earnings differential is greatest for professional employees, lawyers, and doctors. On the other hand, the public sector appears to set a floor on compensation. The compensation of workers with a high school education is higher for state or local government employees, when compared to similarly educated workers in the private sector. [Emphasis added.]

Of course, education is not always a reliable proxy for the quality of the workforce. But do we know for sure that workers with a high school education in the public sector are more industrious, punctual, and hard-working than comparable workers in the private sector? This seems like a question worth asking.

Interestingly, Keefe also notes the following:

A standard earnings equation produced a surprising result: full-time state and local employees are undercom- pensated by 6.3%. Full-time public employees, however, work fewer hours, particularly employees with bachelor’s, master’s, and professional degrees. A re-estimated total compensation equation controlling for work hours of full-time employees demonstrates that there is still a significant public-sector penalty of 3.7% in total compensation between full-time state and local employees and private-sector employees. At closer examination, the penalty disappears for local government employees, but remains for state workers who in 2009 had a 7.5% compensation penalty. [Emphasis added.]

So if the penalty disappears for local government employees, is it fair to use Keefe’s analysis as guide to public sector workers being undercompensated relative to private sector workers? It looks as though full-time state employees with more than a high school education might be undercompensated. That is a far narrower claim. One gets the impression that Keefe’s analysis is being overinterpreted.

Moreover, research by Andrew G. Biggs and Jason Richwine on compensation for federal employees suggests that there might be a missing dimension in Keefe’s analysis:

Federal workers are right to protest simplistic public/private pay comparisons. USA Today recently reported that average federal pay is double that of the private sector. But federal workers have greater education and work experience, for which they deserve higher pay.

But not that much higher. Even after controlling for a large set of variables related to individual productivity, federal workers still appear overpaid. Our analysis finds that federal employees received salaries 12 percent higher than private sector workers with the same age, education, experience, race, gender, and other personal characteristics.

This conclusion is consistent with academic evidence spanning three decades. Combined with higher benefits, average federal compensation may be $14,000 above what a similar private sector worker would receive.

Yet claims continue that federal employees are underpaid. Colleen Kelley, NTEU’s president, recently cited the 22 percent pay gap figure and called for increased federal pay. An OPM official referenced the same alleged pay gap in a press conference, and director John Berry released a statement claiming federal workers “are making a significant sacrifice in pay.”

How does the 22 percent underpayment figure coexist with the academic evidence that federal workers are overpaid? Because the government asks a different question than do economists. The government compares jobs, while economists compare employees.

Each year a bureaucratic entity called the President’s Pay Agent compares public and private sector wages in various occupations and at different levels of responsibility. The answer is that, on average, federal positions pay 22 percent less.

But this does not mean federal employees earn less than they could in the private sector. The reason is that the federal government promotes employees to higher positions than they would hold in the private sector.

I should re-emphasize that Biggs and Richwine are talking about federal employees here. But one wonders if we see a parallel phenomenon at the state level that wouldn’t be captured by Keefe’s controls.

Because it is difficult to compare public sector and private work in all instances, Keefe writes the following:

Ideally, we would compare workers performing similar work in the public sector with the private sector, but this is not always possible. There are too many critical occupations in the public sector, for example, police, fire, and corrections, without appropriate private-sector analogs. Even private and public teaching is significantly different. Public schools accept all students, while private schools are sometimes highly selective and may exclude or remove any poor performers, special needs, or disruptive students. Consequently, comparing workers of similar “human capital” or personal productive characteristics and labor market skills is considered the best alternative, and well accepted by labor economists. Analyses based on personal characteristics comparisons capture most of the important and salient attributes observed in the comparable work studies. 

Prior research reveals that education level is the single most important earnings predictor. Education helps create work-relevant skills. People invest heavily in their own and their children’s education, by buying homes in communities with good schools and by paying or taking on debt to attend schools, colleges, and universities. Empirically, education is followed by experience in advancing earnings. People learn by doing and by working in a variety of job tasks as they advance through occupational levels. Most occupations reward experience, since experience is associated with more competent and complex performance, arising from on-the-job learning. [Emphasis added.]

Could it be that security of tenure in the public sector leads to a sorting process in which more risk-averse high human capital workers gravitate towards public sector work? That is, security of tenure might itself a form of compensation that should be factored in. And remember, Keefe writes:

Full-time public employees, however, work fewer hours, particularly employees with bachelor’s, master’s, and professional degrees.

To some extent, this phenomenon reflects the norms of the workplace. It could be that higher human capital workers who’ve sorted into public employment find long hours distasteful. Many argue that long hours in professional services reflect a desire to minimize the frictional loss that’s involved in adding more members to a team devoted to a project. That is, long hours are at least in part about keeping a small team “in-the-loop” and working in concert, an organizational strategy designed to increase productivity. Higher productivity is then rewarded with higher compensation. As the organizational capital literature suggests, pay-for-performance and worker autonomy have been a key drivers of rising productivity in professional services and other domains of knowledge work. Public sector employment tends to be defined by fairly rigid pay scales and a lack of autonomy. 

Given that we’re now talking about a very narrow group — college-educated state government employees — Keefe’s analysis looks decidedly un-sweeping in its conclusions. And even when we look at college-educated state government employees, the only group Keefe seems to be saying is undercompensated relative to their private sector counterparts, it is easy to identify confounding variables.

To return to Ezra’s point:

But to make the more counterintuitive point, this is a fairly counterproductive conversation. We want really good regulators watching Wall Street. We want talented people teaching our children. We want our trash picked up by workers who want to keep their job and we want the DMV staffed by individuals who aren’t incompetent. That isn’t to say we should pay every public worker millions and millions of dollars, but very few go into government work because they’re of a charitable bent. It’s a job, like any other, and it attracts talent only by paying it well.

We do want good regulators watching Wall Street. This is a federal job, and it’s a job where I’d be happy to see somewhat higher compensation. We do want talented people teaching our children, but teachers are generally local government employees, and Keefe’s analysis states that “the penalty disappears for local government employees.” Moreover, talented teachers often value autonomy over cash compensation, thus suggesting that structural reform could improve the talent pool more cost-effectively than simply increasing compensation in a system that devalues the contributions of “superstar teachers.” 

Moreover, Keefe’s analysis suggests that less-skilled workers in the public sector — a category that might include sanitation workers — are compensated more generously than their private sector counterparts. And Indiana’s Bureau of Motor Vehicles is widely considered one of the nation’s best despite the fact that Governor Mitch Daniels has reversed an executive order that granted state employees collective bargaining rights. This isn’t dispositive, of course, but it does point us in a different direction.

The Agenda

NRO’s domestic-policy blog, by Reihan Salam.