Krugman Phones One In

I rarely post in the middle of the day, but this seems to call for an immediate response:

Paul Krugman, feisty as ever, scoffs at the claim that public-sector employees are overcompensated. True, salaries are 13% higher in the public sector. But, says, Krugman, you’ve got to correct for the fact that public employees are (on average) better educated. After the correction, those public servants earn 4% less than the rest of us.

Well, Krugman is certainly right that you can’t take the raw data at face value. But, at least if you’re trying to be honest, you don’t get to pick and choose what you correct for either. Sure, let’s correct for education levels. Let’s also correct for the fact that public sector employees work fewer hours per week. And for differences in pension plans, and job security, and working conditions.

How can we ever be sure we’ve counted everything important? We can’t, as long as we do it Krugman’s way. So let’s do something sensible instead. Let’s look at quit rates. Quit rates in the public sector are about one third what they are elsewhere. In other words, government employees sure do seem to like holding on to their jobs. More than just about anyone else, in fact. Doesn’t that tell us everything we need to know about who’s overcompensated?

Share/Save

41 Responses to “Krugman Phones One In”


  1. 1 1 Ax

    Low risk people take less risky actions and settle into low-risk routines.
    High risk people take riskier action and seek out high-risk/reward ventures.

    I’m not sure sure how this plays out in terms of salary, but you can’t just take quit rates at face value.

  2. 2 2 GV

    Why are quit rates something unique. People who like stable jobs probably go work for the state. And guess what, they don’t like to rock the boat, so they stay.

    He does have a point when he says, no ambitious kid is dying to work for the govt. Clearly, not the most sought after job. I find that more revealing than quit rates.

  3. 3 3 Steve Landsburg

    Ax and GV: You both have a good point, at least in principle. But surely the vast majority of private sector workers are in routine salaried jobs, not high-flying entrepreneurs. So I’m at least initially skeptical that their risk preferences are going to be all that much different from the public employees’.

    I’m open to being proved wrong about this. But in any event, I am very sure that quit rates come a lot closer to being the right measure than a correction for education that ignores every other relevant variable.

  4. 4 4 nobody.really

    Given Landsburg’s propensity to look to markets as the arbiter of so many things, I’m curious to learn more about Landburg’s theories about the labor market.

    Let’s assume that public employees are “overpaid.” Wouldn’t we then expect that ever more people, and better qualified people, would respond to this price incentive and apply for these highly-compensated jobs? Wouldn’t we expect that government employers would then hire the best of this applicant pool, and that this process would squeeze out the arbitrage advantage?

  5. 5 5 Steve Landsburg

    nobody.really: Wouldn’t we expect that government employers would then hire the best of this applicant pool, and that this process would squeeze out the arbitrage advantage?

    This would require some theory of what government employers are maximizing.

  6. 6 6 EricK

    I don’t know how the proportions work out in the US, but if, for example, the vast majority of teachers are employed by the state then quit rates would be quite low simply because if they want to stay in their profession there would be very few openings outside of the public sector.

  7. 7 7 Steve Landsburg

    EricK: if they want to stay in their profession there would be very few openings outside of the public sector.

    This is another way of saying that they really really like their jobs (compared to the alternatives), which counts as a form of compensation.

  8. 8 8 RJ

    Krugman at least backs up his claims with data from John Schmitt’s research. You, however, are just continuing your anti-Krugman Krusade without any credible research to suggest contrary. And, to be honest, it’s a real turn-off from taking your blogging seriously anymore.

    Also, according to Schmitt, the higher you’re paid as a public sector worker the higher your wage penalty is as a result of wage premiums. From what I remember, the majority of these state and local public sector employees are teachers (K-12). Their burnout rate is vastly greater than the HS dropout rate; nearly half quit within the first five years of employment.

    http://www.timesdaily.com/article/20090222/ARTICLES/902220325?Title=Teacher-dropout-rate-higher-than-students-

    Also, I don’t buy your 1/3rd quit rates claim, even though I am willing to accept that it might overall be lower than within the private sector. I want a link to a legitimate source.

  9. 9 9 nobody.really

    Conversely, perhaps Landsburg has succeeded in persuading the world that there’s something dishonorable about receiving a paycheck from taxpayers. Consequently public employees really, really hate their work. And, as Adam Smith observed, the rate of pay varies with the honorableness or dishonorableness of the work; consequently we must pay public employees more to endure the opprobrium!

  10. 10 10 Steve Landsburg

    RJ: I got the quit rates from BLS data. What additional “research” do you think is missing?

    And how exactly does getting your numbers from John Schmitt mitigate the fact that you’re looking at the wrong numbers?

  11. 11 11 RJ

    Then ‘link’ it.

    I got my data that says you’re wrong from NBER, I swear I did!

    See how that works?

    Or for the matter, do reveal something that would discredit Schmitt’s research and/or support your claim that quit rates are a more accurate measure than a control over age, gender, and education. I believe race and geographic distribution were accounted for as well.

  12. 12 12 Steve Landsburg

    RJ: How did Schmitt account for education? By counting a BA in electrical engineering as equivalent to a BA in education?

    Aside from the fact that it’s nearly impossible to get these controls right, I said in the post that controlling for worker characteristics is not the same as controlling for working conditions, hours, etc. Or did you miss that part?

    I am on my way out the door right now; will add a link to BLS data later. But these numbers are not hard to find.

    Edited to add: Here are the most recent quit rate data: http://www.bls.gov/news.release/jolts.t04.htm . You can find data for previous years at the same site.

  13. 13 13 nobody.really

    Also, according to Schmitt, the higher you’re paid as a public sector worker the higher your wage penalty is as a result of wage premiums.

    I have to wonder if there isn’t something to this. That is, perhaps public sector clerical staff is better paid relative to the private sector, but public health officials and chemistry professors are less well paid relative to their private sector cousins?

    But again, testing this theory requires consideration of all relevant factors of employment – where “relevance” is determined by the employee at the margin. Presumably those public health officials and chemistry profs accepted their positions because they liked the combination of pecuniary and non pecuniary attributes — just as their private-sector cousins did in choosing their private-sector pursuits.

    While we can try to model this, we generally rely on markets to send the appropriate signals.

    Wouldn’t we expect that government employers would then hire the best of this applicant pool, and that this process would squeeze out the arbitrage advantage?

    This would require some theory of what government employers are maximizing.

    Fine. So, to test the idea that public employees are overpaid, we can look to an (imperfect) labor market or we can look to (imperfect) models of a perfect labor market. Choose your poison.

  14. 14 14 floccina

    I think that the fact that something like 12% of mail carriers have college degrees is a sign that mail carriers are over paid. How to adjust for that?

    IMO a job is overpaid if you could get qualified people to do it for less. That is all. Gov. jobs seem to highly sought after so I would assume that they are over paid.

  15. 15 15 EricK

    EricK: if they want to stay in their profession there would be very few openings outside of the public sector.

    Steve:This is another way of saying that they really really like their jobs (compared to the alternatives), which counts as a form of compensation.

    Having to change profession can be considered a cost, and I don’t see how not being overly motivated to suffer that cost can be viewed as being a form of compensation. For instance, if you don’t see people leaving profession A to work in profession B or vice versa, you can’t deduce that both professions are overpaid! Rather this would suggest that there are barriers of one sort or another inhibiting the move.

    When you look at the quit rates in the private sector, do you see most people going to work in a similar role for another company or do you see them change profession altogether? If the latter, then you may have a point, but if it is the former, then all you can say is that public sector teachers like to continue teaching rather than retrain.

  16. 16 16 PaulRoscelli

    1/3? Are you kidding me? You are being far to generous to Krugman’s point. I work at a public community college. Years ago I was asked to work on the dismissal process (process? Hint right there it takes a bit longer to see bad apples leave). I asked how many people, in the prior 25 some odd years had left the college. I got answers that varied from puzzled looks to numbers in the single digits!! I had recently moved from a CPA firm where quit rates were 25% a year–that’s 100% in 4 years and these guys were telling me about three guys named Bob!

    This guy, and not Krugman, should have gotten the Nobel

  17. 17 17 Seth

    Nice ‘true measure’. This reminds me of the ‘true measure’ of default rates in lending. When lending institutions were accused of racial discrimination, someone asked a good question, “shouldn’t that show up in the default rates?” In other words, if lending institutions held certain races to higher credit standards and other races to lower standards, shouldn’t the former have lower default rates? I don’t believe the data panned out.

  18. 18 18 floccina

    EricK your argument would carry weight if teacher pay had been higher and had recently dropped but this is not the case. Also they could quite and go to a different school district.

  19. 19 19 RJ

    How did Schmitt account for education? By counting a BA in electrical engineering as equivalent to a BA in education?

    What was measured, interestingly enough, is that more people within the public sector employment field has disproportionately more degrees as well as advanced degrees (50% vs. 30% for four year and 25% vs. 10% for advanced degree.)

    Aside from the fact that it’s nearly impossible to get these controls right, I said in the post that controlling for worker characteristics is not the same as controlling for working conditions, hours, etc. Or did you miss that part?

    Oh, I got that part. I also read you emphatically stating…

    So let’s do something sensible instead. Let’s look at quit rates. Quit rates in the public sector are about one third what they are elsewhere…Doesn’t that tell us everything we need to know about who’s overcompensated?

    And…

    But in any event, I am very sure that quit rates come a lot closer to being the right measure than a correction for education that ignores every other relevant variable.

    Again, education wasn’t the sole variable. Nor have you yet provided any legitimate reason or scholarly research as to why quit rates are a sign of overcompensation. That could easily suggest greater labour mobility between different firms of a given industry versus another. You see this a lot in the restaurant industry, for example. According to the data you provided, if we’re following your logic, Accomidation and Food Services has a significantly higher quit rate than Construction…are those in the Construction field therefore overcompensated based on the nature of their work?

    Here, I’ll charitably help you establish your argument better…

    http://www.bls.gov/news.release/ebs2.nr0.htm

    Seriously though, my point to you is to establish that your anti-Krugman Krusade has become an obnoxious obsession and you actually stray away from adding an alternative side to the debate.

    I am on my way out the door right now; will add a link to BLS data later. But these numbers are not hard to find.

  20. 20 20 Josh

    RJ, Construction was overcompensated during the boom.

  21. 21 21 RJ

    Josh,

    During the boom? Well it so happens that Steve’s stats from the BLS are from May 2009 to May 2010. Some ‘boom’ they’re having then and now.

  22. 22 22 Steve Landsburg

    RJ: Well it so happens that Steve’s stats from the BLS are from May 2009 to May 2010.

    It so happens that the stats I linked to are from May 2009 to May 2010. It also so happens that the analogous stats from the several years previous (all on the same site) show pretty much exactly the same thing.

    (This does not, of course, dispose of your point about different labor mobility in different industries.)

  23. 23 23 Neil

    If government workers are overpaid, and I do not discount the possibility, it is a relatively minor problem. I can only speak from my experience, but many times I have been impressed with the quality and dedication of the much maligned government worker. However, government managers are another thing. There are news items tonight about people being paid millions to manage small towns. WTF?

  24. 24 24 Cos

    I’m getting the impression that you love bashing most things Krugman says without bothering to spend any real thought on the matter, and that’s how you end up with supremely facile posts like this one.

    I’ll leave it an exercise for you to think up a few other reasons why quit rates might differ – something I know you could do if you weren’t blinded by ideological bias. At the very least you should conclude that quit rates on their own tell you nothing useful here.

    You might the move on to reading comprehension, and notice that Krugman never claims that he’s *proving* that public sector employees are underpaid or overpaid relative to private sector employees – he merely claims that 1) the evidence that they’re overpaid is inconclusive, and 2) that regardless of which way the discrepancy goes, it seems not to be as big a difference as to merit the claims being made about it.

    You might then properly feel a bit silly for responding to his “look, this evidence can be read other ways, it’s not conclusive” with your rebuttal “hey, Krugman, this evidence in incomplete and inconclusive!” – you’re saying the same thing on point #1 (while you completely ignore point #2).

  25. 25 25 Brian Golden

    My ex-wife is a teacher. She works in a very wealthy school district and her and her co-workers would constantly complain that they are underpaid.

    I would say to her, “do you remember when you applied for your current job? Do you remember you were 1 of 600 applicants? If you’re so underpaid, why are there 599 other people who want your job?”

    She didn’t appreciate my logic.

    (Just to give you an idea, the high school Home-Ec teacher was making 120K/year).

  26. 26 26 Josh

    Cos said “You might the move on to reading comprehension, and notice that Krugman never claims that he’s *proving* that public sector employees are underpaid or overpaid relative to private sector employees”

    Krugman said “But as Schmitt shows, that’s an apples and oranges comparison: state and local workers are much better educated and somewhat older than private-sector workers, and once you correct for that the comparison actually seems to go the other way.”

    Aren’t we playing with words a bit, Cos? You claim Krugman never said he’s “proving” anything, but in the sentence above he states that it seems like his view is accurate based upon the fact that state and local workers are much better educated and somewhat older than private-sector workers. He’s making a case, no?

  27. 27 27 Henry

    This is another way of saying that they really really like their jobs (compared to the alternatives), which counts as a form of compensation.

    That doesn’t mean that others would like their jobs relative to the alternatives. A blind beggar child might not want to quit for anything, that doesn’t mean that blind beggar children are overcompensated and that we should be seeing queues of people waiting to blind themselves in order to take up this lucrative job.

  28. 28 28 John Dewey

    Some of these comments cause me to wonder if the commentors are not public employees trying to defend their excessive pay and benefits.

    The more that hard information is revealed about public sector pay and benefits, the more outraged the taxpayers become. Decades ago we used to argue that public sector employees were paid less because they enjoyed greater job security than private sector employees. Not anymore.

    Experiences we have all had with public sector emnployees continue to confirm their arrogance. They know how hard it is for them to be fired, and it usually shows in their demeanors. Has anyone ever had a good experience at his state department of motor vehicles or state driver’s license bureau?

    Why do we put up with their arrogance? I haven’t figured that out, but my guess is that we know they have better access to people with guns who will find ways to lock us up.

  29. 29 29 Dallas

    The extreme security of government jobs or any “no employment risk” job combined with higher pay and benefits creates a situation where the boss can be a real ass and treat the staff poorly, without them quiting. That is why we have the term “going postal”. Being in a low paying private sector business area where everyone could quit and get a higher paying job forces a manager or owners to behave.

    With technical employees for regulatory agencies, I find it curious that most of them do not stay technically current and never seem to broaden their scientific understanding of the world via self learning and study. Could it be that only politically sensitive people who follow the organizational agenda get promoted and upgrading your actual skills and knowledge doesn’t have any rewards. Top that off with being a technical staffer whom only deal with sycophants and supplicants who need permits and would never tell him he is full of BS and his self opinion of his ability will remain high as his ignorance increases. Give even a bright Ph.D. with two decades of non-thinking bureaucracy and he will become a fool who thinks he knows what he is doing.

    As an employer, I hired people based upon their knowledge and ability, not their degrees. Government HR seems to assume that if you have the desired education and paper work you actually know what your doing and that assumption is not true. In 35 years in my own technical business, I never hired an ex-government employee that was worth a dam and I was always glad to see them go.

    Normalizing for “education” doesn’t address the knowledge and breadth differences found between government bureaucrats with their narrow thinking and what is required in the private sector. If we could somehow include normalization on actual knowledge and skill, the government employees would be even more overpaid.

  30. 30 30 Ken B

    I am struck at how many miss the point. Krugman picked one factor — good looking people are paid more, and postal workers are really hot so once you correct for that they aren’t overpaid, etc. SL pointed out some flaws in this logic, and proposed that if you want to look at just one number there are some that do a beeter job of measuring overall satisfaction.

    This should alert Steven to one of the main problems to teaching economics: economics is political. Need to pick a more abstract and neutral example to make your point or you run into partisanship over everything.

  31. 31 31 Sam Grove

    Maybe the overall problem isn’t how much government workers are paid, but how many government workers are being paid.

    How many government jobs are make-work?

  32. 32 32 John Dewey

    “Government HR seems to assume that if you have the desired education and paper work you actually know what your doing and that assumption is not true.”

    Not sure if this is still true, but I think government HR was for decades more concerned about correcting perceived discrimination than about finding effective employees. HR folks I have known more often have prided themselves as the “conscience” of an organization than as a key player in increasing organization effectiveness.

    As a Vietnam-era veteran, I would have received preferential treatment when I applied for a government job. So also would have an African American, a Native American, a Hispanic-surnamed American, and many more members of “disadvantaged” groups.

    Could a hiring manager discriminate in any way and have a chance of hiring a potentially effective employee who happened to be a non-veteran, non-disabled white man? I don’t think so, unless the hiring manager decided to insist that a college degree be required. Forty years ago college graduates were disproportionately white males. Of course, that situation didn’t last long. Once the supposedly disadvantaged groups realized the value of the college degree, everyone who wanted a government job managed to earn degrees of some sort.

  33. 33 33 DividedLine

    Krugman didn’t pick one factor, he referenced an article that pointed to two different economic studies which attempted to correct for two factors relative to total compensation: education and experience. Follow the links and you can read the two studies. Cohn, whom Krugman cited, picked the more conservative of the two. Here is the Cohn article with the links to the studies:

    http://www.tnr.com/blog/jonathan-cohn/76884/why-your-fireman-has-better-pension-you

    I always assumed education and experience were worth something, but maybe that’s not right (Economics speaks of the signaling value of a degree). If quit rates are a better metric than education and experience for determining pay, are there studies that we can evaluate? What does quit rate theory indicate is the optimal pay rate for government employees? I agree satisfaction is worth something. How much? Should human satisfaction be a required consideration for all economic studies? Have military personnel been factored in or out of government quit rates (the quit rate figures supplied were national)? etc.

    What Krugman did add that was new to the discussion is that state government employees account for 28% of state budgets. Cutting their pay by 10% would reduce state budgets by ~3%. He writes: “A few percent either way in workers’ compensation would not make a big difference to state and local spending. This is a phony issue.” The math supports his point.

  34. 34 34 John Dewey

    “If quit rates are a better metric than education and experience for determining pay”

    Didn’t Steve write that quit rates are a metric for determining overcompensation? That’s different than determining pay levels.

  35. 35 35 John Dewey

    “I always assumed education and experience were worth something, but maybe that’s not right”

    Maybe it’s not that simple. It really depends on the content of the education and the quality of the experience, doesn’t it?

    A bachelor’s degree in cultural studies or English literature is not going to mean much to most employers. A degree in electrical engineering or nursing or accounting will definitely increase one’s earnings potential.

    Work ten years on a drilling rig and you’ve learned some skills a number of oil field services companies will pay for. Work ten years at a window receiving traffic fines and your skill set has not appreciated much at all.

  36. 36 36 Russell Nelson

    Cos, you’re defending Krugman by saying that he phoned this one in (by posting mere speculation). Well, read the title of this posting: Krugman Phones One In.

    Given that Krugman has enough undeserved respect to have gotten a Nobel Prize THIS YEAR, it’s fair to point out when he takes himself down a notch.

  37. 37 37 Russell

    One other thing that most surveys don’t take into account is over-education in the gov’t sector. I am a pharmacist, and work for a gov’t agency and separately for a private company.

    My degree is necessary (based on licensing laws) for my private job. However, my gov’t job could be performed adequately by someone with a high school diploma. But the gov’t insists on hiring medical professionals for positions such as mine. It’s true that my compensation is approximately equal at both jobs, but the gov’t job could be done for probably 30-40% of what they pay me.

    Gov’t hiring of medical professionals where they are not truly needed also distorts the market for those positions, thus increasing what the private sector must pay for these positions (which also leads to increases in gov’t salaries of these people).

  38. 38 38 Bruce True

    I would think that you could argue (some above may already) that government agencies would pay for workers more educated than optimal to do the work, for the prestige. The argument above that a premium on the job gives the boss more power over the works makes sense also. I think you could explain some of this based on a read of Alchian and Allen’s textbook.

    I don’t know enough about any of the data above to comment on a number of substantial points made above.

  39. 39 39 Jeffrey

    Another factor is that the government employers are essentially never allowed to fire people or lower their pay once it’s raised, and consequently can be very slow to raise the pay of those who deserve it. Suppose a $40,000/yr private sector employee does something absolutely amazing and his salary shoots up to $80,000/yr, but he’ll lose it the moment he fails to maintain his performance. Meanwhile, suppose his public sector cousin does the same thing and gets a grade promotion to $50,000/yr that can never be taken away. (I can certainly imagine these numbers being tweaked so that I’d be indifferent as to which promotion I would prefer.)

    Due to the means in which they are paid for their performance, the government employee sticks around to collect the rest of the bonus that he earned in the past, while the private sector guy has no compelling reason to either quit or stick keep the same job for a long time.

    If the pay system is the reason government employees don’t quit, then this points to a problem of poor incentives that encourage people to keep working for the government for far too long, rather than to a problem of overpaid government workers.

  40. 40 40 van nestor

    I wonder what the quit rates are for tenured college professors

  41. 41 41 nobody.really

    Let’s assume that public employees are “overpaid.” Wouldn’t we then expect that ever more people, and better qualified people, would respond to this price incentive and apply for these highly-compensated jobs? Wouldn’t we expect that government employers would then hire the best of this applicant pool, and that this process would squeeze out the arbitrage advantage?

    This would require some theory of what government employers are maximizing.

    Of course, they’re maximizing the intelligence and qualifications of the public workforce. After all, how do you suppose government is able to retain all those smart, qualified people who watch over everything?

    Oh, wait a minute….

  1. 1 Krugman Phones One In « Daniel Joseph Smith
  2. 2 Some Links
  3. 3 Tweets that mention Krugman Phones One In at Steven Landsburg | The Big Questions: Tackling the Problems of Philosophy with Ideas from Mathematics, Economics, and Physics -- Topsy.com
  4. 4 Public Sector Gets Even Fatter | Daily Libertarian
  5. 5 Interesting fact but doesn’t settle the matter « Belligerati
  6. 6 The American Spectator : AmSpecBlog : How to Tell if Your Public Employees Are Overpaid
  7. 7 KRUGMAN WATCH › Picked apart… twice!!
Comments are currently closed.