Arnold Kling on Government Wages

Arnold Kling writes a post responding to my post on government wages and titles it not “Seeing Like a Labor Econometrician” but Seeing Like a Central Planner:

…This is part of an argument the government workers are not underpaid. As a central planner, you are convinced that you can measure a worker’s value by looking at characteristics such as education level.

Two points I would make.

1. The government can never know the value of a government worker. We do not know the value of the output that they produce. The socialist calculation problem is very real.

2. We might be able to make a guess about the opportunity cost of a government worker. That is, what does the worker forego in order to work for the government? To make this guess, we would want to look at the willingness of people to take government jobs and the willingness of people to leave government jobs. If the government has a hard time filling positions, then those positions are likely to be underpaid relative to opportunity cost. If the government has a hard time retaining workers, then those workers are likely to be underpaid relative to opportunity cost.

It seems to me that no matter how many studies one does of pay relative to education or other characteristics, they will only be convincing to dedicated believers in central planning. If you held a gun to my head and made me a central planner, I suppose that I would try to allocate labor by measuring characteristics of workers and aligning those characteristics to jobs….

Three back:

1.  Several people have mentioned a weird thing about being uncomfortable controlling for education, as if I’m making stuff up on the fly. If you don’t know this you’ll just have to believe me, but controlling for education while comparing wages is necessary. If they didn’t do that, you’d have to ignore and throw out the study. I simply can’t imagine the idea of not controlling for years of education and/or an education level when comparing wage levels, either between industries, genders, locations, and/or employers (public versus private).

2.  I also don’t fully understand what Kling is arguing against. State and local government workers are underpaid, particularly at the high end of education. I understand the arguments about how to properly control for long-term benefits, but on the first approximation a college-level worker makes 25% less in total compensation and a lawyer or other professional worker 37% less.

And yes, the government does hire in a market where it has to compete with private employers. If they paid too little and demanded too much workers would go elsewhere. Now if workers were overpaid relative to private market workers, we’d have a discussion. But they are paid too little here at the state and local level. I think Kling’s argument is that government workers are (or perhaps government work is) lazy and unaccountable to productivity. Fine. But that goes to why people would take the wage cut when they have high human capital. I’d like to think a sense of civic purpose too, but not necessarily.  (Separately, at the Federal level, many of the higher paying jobs are ridiculously competitive, often pulling from the most elite colleges.)

Markets exist in all kinds of labor markets where the “marginal product” of a unit of labor doesn’t make a ton of sense or isn’t very clear, from the non-profit where I work to places like social networking, national defense, and radio production.  Yet there are competitive markets for the labor in each of these.

At the “high school” through “associate” degree level there is a compensation premium (not wages, which are lower across the board), but that is more the result of declining wages and benefits in the private market and not a runaway government spending spree. Yes, you could probably hire cops for $8/hr and no benefits to squeeze government workers down to private level wages.  But is a failure of the current private market for benefits, not the public one.

3.  Fun thought of the day that I’m not certain is in anyway correct, but let’s double down: We do have a reason why government wages might be lower – the government is a monopsony purchaser of labor in many fields, particularly those that monopolize violence, and particularly those that can confer certain types of prestige.   Also it makes a monopsony market for labor in many places and locations too.  (FYI: Monopsony is to buyers what a monopoly is to sellers; from our point of view firms purchase labor.) For those who want to be in criminal justice and/or administer to the peace, for those that want to join the military, for those who want to be regulators, work in the bureaucracy, work in or for the courts, etc. there is often only one game in town: the government. As such, it can squeeze wages downward.   I don’t know if this is why certain types of high human-capital jobs get paid less, but it is a thought.

This entry was posted in Uncategorized. Bookmark the permalink.

7 Responses to Arnold Kling on Government Wages

  1. sraffa says:

    Sigh. Do we always have to resort to red-baiting? Kling doesn’t reference all the other studies that are screaming gov’t workers are overpaid! That use the same “socialist” calculation. Is there any other better way. If there isn’t a non-”socialist” way to calculate whether gov’t workers are overpaid, then the justification for a massive cut to government workers is pure speculation. If we can calculate the wage differential, then govt workers aren’t overpaid. Engaging in these kind of ad hominems shows that the right is losing the debate.

    Who are these workers that are overpaid? Firefighters? Soldiers? Police? Teachers? the VA? It’s easy when you talk in terms of abstract, instead of thinking about your neighbors.

    Let’s talk something me and Kling should both know well- Universities. As George Mason is a public University, Arnold Kling is a public employee. Is he underpaid? How could we even tell? It’s well known that private universities, even of similar quality, pay better. No socialist calculation, just my take from inside that market.

  2. “The government can never know the value of a government worker. We do not know the value of the output that they produce. The socialist calculation problem is very real.”

    Arnold, get used to the fact that we live in a world of probabilities (and assumptions, if even the most basic, like what I see is actually there). You basically can’t know anything 100% for sure. You can’t know the (utility) value of a private sector worker precisely for sure either. First, there are, of course, very substantial externalities (especially positional/context/prestige), asymmetric information, zero marginal cost of ideas/information, inability/impracticality to patent, etc., etc. On top of this, if you’ve ever had a cost accounting class you know how hard it is to accurately discern the costs and benefits of an individual worker to the firm’s bottom line.

    But, as with government workers too, with some smart analysis, and very mild assumptions, you can come up with high probability confidence intervals that aren’t that wide.

    I mean think about it objectively if you’re willing to think about something that goes against your libertarian dogma. Many people in government control millions, billions. You don’t want to pay for a skilled competent person to do that efficiently and well. Some of these people control more money than CEOs paid, without exaggeration, over 1,000 times as much.

    I’ll leave you with a quote I hope you’ll think about, from celebrated growth economist and perennial Nobel shortlister, Paul Romer of Stanford:

    As just one example, recall that the increasing returns to scale that is implied by nonrivalry leads to the failure of Adam Smith’s famous invisible hand result. The institutions of complete property rights and perfect competition that work so well in a world consisting solely of rival goods no longer deliver the optimal allocation of resources in a world containing ideas.

    – Forthcoming American Economic Journal paper, page 8, at:

    Of course, you may not care if there’s tremendous loss, suffering, and decreased growth in wealth, science and medicine, if you avoid losing even one smidgeon of economic freedom that hardly anyone will even notice, but luckily, few voters are extremist libertarians.

  3. Arnold Kling says:

    1. A central planner and a labor market econometrician see the labor market the same way. That is, both think that wages should be a deterministic function of measurable characteristics.

    2. You impute to me all sorts of opinions that I did not express. That is not constructive.

    3. You say that if government as an employer paid too little and demanded too much, then workers would go elsewhere. Workers aren’t going elsewhere. Therefore, ….

    4. To commenters: I am not an employee of George Mason University.

    • “You say that if government as an employer paid too little and demanded too much, then workers would go elsewhere. Workers aren’t going elsewhere. Therefore, ….”

      It’s a matter of quality. You’ll still get workers, but of lower quality.

      If the pay of a CEO went from $1 billion per year to $20,000 (and no perks, or indirect monetary gain), you would still be able to fill that job, just with someone a lot less skilled.

      As many people in government control millions, or even billions of dollars, and vital public safety, it makes sense to pay to get skilled competent quality people for these jobs.

  4. David Wiczer says:

    It’s a bit silly to control for worker quality by only using education. At least include age. Is the age profile different? And the composition?
    Anyway, you already know that wages in the private sector, residual of any observable quality controls have very high variance, both within a year across workers and within an employee’s life. You’re not doing science by comparing wages of education groups because that’s not a useful partition when residual wages are so important.

    It’s perfectly reasonable that government “stability” (from yearly fluctuations and from uncertainty when I match with an employer) compensates for low pay. Either workers are indifferent between higher compensation or different types (by risk tolerance) are attracted.

  5. sraffa says:

    Apologies to Arnold- seems he is a former George Mason University employee

  6. JCD says:

    I would love to see Mike and/or Arnold address the issue that this study uses average compensation (rather than median compensation, or multiple quartile or quintile compensation values).

    The average compensation values used in this study are based on average salary values (provided in a similar table in the study). A comparison of average income is incredibly misleading, as the average income is way above median income rates.

    The study lists average earnings in the private sector for a professional degree at $152k and for a bachelor’s degree at $71k, while in the public sector it is only $88k and $48k respectively. I am sure that someone more informed than I could provide even more recent numbers, but a quick google search turned up 2006 census data that revealed that median personal income for full time employed persons 25+ (who will usually be earning more than someone younger) with a professional degree is $100k, and with a bachelor’s degree is $50k.

    The study completely ignores the fact that average income in the private sector is skewed by extremely high income individuals that drive the mean up. This is a factor that simply is not present in the public sector.

    Basically, I would posit that the top 1% of the private sector makes so much more than the top 1% of the public sector so as to overall make the average private sector income be higher than the average public sector income, but that this has absolutely no bearing on whether most public sector employees are under compensated.

    I also note that this theory would suggest that this is only a factor for workers having a bachelor’s degree and above, and this is exactly where the huge disparities in average income are seen in the study.

    I take no position on whether most public sector employees are, in fact, over or under compensated; instead, I merely think that using mean income poses problems.

    Overall, more granularity is needed. There are some public sector employees that are overworked and underpaid, while there are other public sector employees that are, unquestionably, overpaid. Similarly, there are some private sector employees that are underpaid, while there are some private sector employees who are overpaid. Unfortunately, because this study uses misleading mean incomes as a starting point, I cannot see how it is anything but misleading as a starting point for discourse.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s