Arnold Kling  

Self-Discrediting

PRINT
We Need a Tea Party... A Rationally Optimistic Price?...

This study comes from the Economic Policy Institute.


The study analyzes workers with similar human capital. It controls for education, experience, hours of work, organizational size, gender, race, ethnicity and disability and finds that, compared to workers in the private sector, state government employees are undercompensated by 7.55% and local government employees are undercompensated by 1.84%. The study also finds that the benefits that state and local government workers receive do not offset the lower wages they are paid.

In that case, laying off public sector workers is doing them a favor. They should enjoy the higher remuneration available elsewhere.

Thanks to Mark Thoma for the pointer. If some projects are self-recommending, then this sort of project is self-discrediting.


Comments and Sharing


CATEGORIES: Economic Methods



COMMENTS (14 to date)
Stefan writes:

And laying off private sector workers is doing them a favor since they are paid less than public sector workers, in your view?

It strikes me that you need more than a 7.55% pay differential to off set the costs of getting fired (for most workers).

John writes:

Well played. There are just some claims that can be disputed by the plain evidence (e.g., if the U.S. is such a terrible proto-fascist place, as some claim, why is there such robust immigration?).

I'm sick of hearing that public employees are undercompensated. The reason? Everyone I know wants a government job, but they're incredibly hard to get. In a lot of cases (state and local especially), you need to know someone on the inside. If the demand is high for a government jobs, you'd expect wages to drop. But you don't see that, in part because of nepotism/patronage, in part because of unionization, and in part because the government has no profit motive to force its hiring practices to be cost-effective.

tjames writes:

May I nominate at least 2 howlers I have so far read that are additional strong indicators that this work should be dis-credited?

#1
"Prior research reveals that education level is the single most important earnings predictor. Education helps create work-relevant skills."

I believe the authors have missed the on-going debate about the relevance of college education vs. job skills training. Some disciplines exist to transfer a large body of relevant knowledge to their practitioners (e.g. medicine, law, engineering, accounting), some explicitly do not, but instead aim for an educated adult. Whether or not the degrees received transfer value work related skills or not is easily seen in the private sector by compensation differentials. This is suppressed in the public sector, where often a bachelor's degree is considered a uniform good, regardless of the institution granting it or the discipline.

#2
"State and local public employees are substantially more educated than their private-sector counterparts."

They might be more credentialed than private sector counterparts. This does not make them more educated, nor is it any indicator of the acquisition of skills that would be valued in the private market. One gets the impression that the authors equate human capital with credentials. I notice the self-employed (I assume this means many small business owners, since they employ themselves) are conspicuously left out, despite the fact that many have self-generated enough human capital to be judged worthy of compensation by the market.

What a sloppy piece of work.

Richard A. writes:

If public sector workers are underpaid relative to private sector workers, then public sector workers must have a higher turnover rate then private sector workers. I don't think this is so.


Floccina writes:

You cannot go by human capital level because higher compensation attracts higher human capital. If you pay DMV workers $200K/year and base hiring on school credentials you will have a many PHDs working at the DMV. This would be double bad because they would be overpaid and would be keeping more potentially productive people unproductive. (I wonder if even $200K PHDs would still seem become like zombies after a few years on the job. )

Dave writes:

Landsburg already handled this idea eloquently here.

mark writes:

As far as I can tell, the study author (a state employee by the way) is simply comparing the average compensation as among all public sector workers with the same metric among all private sector workers, which I infer means that the private sector benchmark is skewed by outliers - CEOs, bankers, athletes, high paid lawyers and doctors. If we give every public sector worker average priced seats at sporting events for free, we would not think it was justified because the average price of the seats occupied by other attendees was greater.

8 writes:

If you know how state government works, employees get raises for earning higher degrees. I once asked the members of the school board and the principal whether they had ever seen a study that shows teachers with a Master's degree are more effective in the classroom. They had no idea, it's just he way the pay scale works. Same for cops, etc.

Government workers are probably overpaid by 30% or more based on their contribution to the economy, which in many cases is a negative number. Firing them would be a net gain.

Peter Twieg writes:

I can understand a lot of the controls, but... organizational size? Even if there's a link between organizational size and wages, why would that be a relevant control?

Bob Knaus writes:

I'm assisting in an org review for a large state agency providing centralized computer services to the other state agencies. Public sector undercompensation is a common complaint in the IT field. We found that 25% of agency staff are eligible to retire this year... but only 2.5% of them will.

Better opportunties in the private sector? I guess not.

eccdogg writes:

As others have mentioned turnover rate is the only thing you need to look at.

Last I looked only 2% of public employees voluntarily left their job in a given year compared to something like 17% in the private sector.

That tells you all you need to know. When a public employee gets a govt job there are not a lot of higher bidders out there.

two things writes:

(Nominal) hours worked is not a good proxy for productivity.

Mark Brady writes:

How do professors and instructors at state universities and colleges compare with their counterparts at private institutions?

Reality Check writes:

Arnold, you're getting kind of Krugman in all of this (i.e., it's becoming a tirade).

As the first responder, above, pointed out, it's hard to see you applying the logic you endorse for the private sector, so why do so for the public sector?

And even if your logic was something you were willing to apply equally, it wouldn't do anything to support your claim that the study was self-refuting, as it stands or falls on its own merits.

Finally, the idea (put forth by others in the comments as well as by Landsburg) that you only need to compare quit rates is equally batty. Landsburg tells us that we can't control for all possible variables when comparing wages so we should look at quit rates without ever telling us why controlling for outside variables is unnecessary here. If outside variables affect wages, how could they not effect quit rates?

As a federal employee (I work for BLS in the wage program that is allegedly used to set federal pay), I've come to believe that we are, likely, overpaid, in many (if not most) areas, so I'm sympathetic to your thesis...

...but it's starting to look personal.

Comments for this entry have been closed
Return to top