Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Friday, August 31, 2012

Is wage rigidity the problem?

Many economists claim that wage rigidity plays an important role in the (mis)allocation of resources in the labor market. Both "Keynesian" (Krugman) and "Monetarist" (Sumner) thinking emphasizes sticky nominal wages. "Labor Search" theories (Hall and Shimer) emphasize real wage stickiness; see here.

I've always been somewhat skeptical of these stories (note: Keynes himself did not emphasize stick wages in the GT--in fact, he argued that wage flexibility makes matters worse!).

My skepticism is fueled by stories like the one I read today: Majority of New Jobs Pay Lower Wages, Study Finds. An excerpt:
The report looked at 366 occupations tracked by the Labor Department and clumped them into three equal groups by wage, with each representing a third of American employment in 2008. The middle third — occupations in fields like construction, manufacturing and information, with median hourly wages of $13.84 to $21.13 — accounted for 60 percent of job losses from the beginning of 2008 to early 2010.  
The job market has turned around since then, but those fields have represented only 22 percent of total job growth. Higher-wage occupations — those with a median wage of $21.14 to $54.55 — represented 19 percent of job losses when employment was falling, and 20 percent of job gains when employment began growing again.  
Lower-wage occupations, with median hourly wages of $7.69 to $13.83, accounted for 21 percent of job losses during the retraction. Since employment started expanding, they have accounted for 58 percent of all job growth.
Seems to me that even if nominal wages appear to be sticky within occupational groups, there is a high degree of de facto flexibility via occupational choices (on both the supply and demand sides).
My gut feeling is that theories that rely on some "wage stickiness hypothesis" are barking up the wrong tree. The assumption of wage stickiness is often supported by appealing to the empirical evidence. But as I explain here, wages that appear to be sticky to an econometrician may not be sticky in any economically meaningful sense. And as the evidence above suggests, there seems to be much more wage flexibility out there than is commonly assumed. But I'm willing to listen to the other side of the story...


  1. Correct me if I am wrong, but weren't the job losses much larger than the job gains? So if this flexibility exists, it is probably a small effect.

    I say "if", because the data you cite don't distinguish between who lost and who gained. We know that a large number of people dropped out of the job market, who were those people?

    If there are relatively more old people who lost a job compared to those who gained a job, you might have an explanation for the flexibility.

    1. Seems like the job losses were concentrated among the young; see here:

    2. Thanks! I had a quick look at

      If you split up the share of employment into young, middle and old or 15-29, 30-49, and everything above, then you would find that over 2007-2011 that:

      I. the young have lost comparatively less than the middle;
      II.the old have gained compared to the others;
      III. from 2009 the young have stabilized.

      I would interpret this as that the job gains must come more from the young than from the middle.

  2. A more or less conventional view here:

  3. I did a comprehensive subsector study a few months back, and found that low-wage professions have *not* had a stronger recovery than middle- and high-wage professions: The slump and recovery appears to be pretty much down across the board when you carefully sub-divide each sector according to average weekly and hourly pay.

    1. JSeydl,

      That's interesting. I wonder what might account for the apparent discrepancy between your study and the one I link up to above? Or is there a discrepancy?

  4. "The [textbook Keynesian] theory literally implies that if prices were fully flexible, many of the worst properties of recessions would be avoided."

    That's because 'Keynesian' theory assumes that when W adjusts, M stays put. Keynes expressed doubt about that, but I think you go too far in attributing to him the view that wage flexibility makes matters worse; he merely said that if a falling wage-rate causes the money-supply to shrink, that makes matters worse.

    Also, textbook models conflate flexible prices with equilibrium prices. But prices can be very flexible indeed without having any great affinity for equilibrium. It's hard to see that the world would be a better place if plumbers' wages were as volatile as FX rates.

    I don't think that many advocates of sticky-price macro would dispute that it's theoretically unsatisfactory. Krugman has pointed that out numerous times. Mostly they are simply invoking Friedman's 'as-if' approach to modelling.

    1. Kevin,

      Although I can't find the precise reference right now, I seem to recall Keynes (GT) making the argument (in a later chapter) that a fall in W would make things worse by contributing to a decline in AD.

      Also, most textbook models do not mention that spot prices are meaningless in relationships. Please see my "sticky price hypothesis, a critique."

    2. Hi, it is the Chapter on wages (XIX). And then, in Ch XXI, when he evokes the cumulative deflationary effect of flexible wages:
      "If, on the contrary, money-wages were to fall without limit whenever there was a tendency for less than full employment, [...] there would be no resting-place below full employment until either the rate of interest was incapable of falling further or wages were zero. In fact we must have some factor, the value of which in terms of money is, if not fixed, at least sticky, to give us any stability of values in a monetary system.

      hope this helps. My take is that David's interpretation is correct.

  5. Stagnant new firm creation has held back employment in this recovery. Moreover, employment in discretionary services -- particularly lower skilled occupations -- has also lagged. Both sectors (they overlap) likely have amongst the least sticky wages in the economy.

    The drop in net new firm creation arguably has a structural explanation: reduced access to capital due to the demise of home equity lending.

    1. But Diego, not many firms complain about a lack of access to capital. And larger firms are hoarding cash. Your explanation for this?

    2. NFIB surveys do downplay the capital access issue. However, entrepreneurs that fail to start up firms are not part of the survey.

      HELOC's and other mortgage financing "technology" freed would-be entrepreneurs from the binding constraint of their accumulated savings. The expected increase in house prices also freed current savings from having to finance future retirement spending. Both effects were arguably important factors in providing capital for start-ups and increasing net new firm employment.

      Cash hoarding and peak NIPA profits are compatible with the above narrative. High margins are a sign of a reduced entry/expansion threat from other firms. The fact that corp. profits recovered their previous peak just three years after the GFC tells us that access to capital is likely an important factor in restricting competition. Note that this high profits/high unemployment condition is far different from the low profits/high unemployment one that "deficient AD" would produce.

    3. However, entrepreneurs that fail to start up firms are not part of the survey.

      Well, I have to say that this is a very good point. I wonder how one might get a handle on how quantitatively important this phenomenon was/is? I mean, there are all sorts of reasons for smaller firms not starting up -- are there surveys out there documenting the reasons?

      Note that this high profits/high unemployment condition is far different from the low profits/high unemployment one that "deficient AD" would produce.

      Doesn't a standard NK model generate a countercyclical markup?

    4. On "failure to launch", I am not aware of any surveys. While many factors do affect the rate of net new firm creation (regulation, innovation, etc.), arguably none of these had a large delta relative to the availability of start-up capital.

      On AD, I had in mind sticky wages, which I assume would produce procyclical mark-ups. Your point on sticky prices is interesting. Since margins tend to trough with recessions, it would seem that countercyclical markups are not supported by the empirical evidence (although I realize factors other than pricing feed into margins). It was really the behavior of margins a few years into the recovery that shows a reduced intensity of competition. That is, comparing the speed of recovery of peak margins in this recession to prior recessions.

  6. Wage rigidity for whom? There's a lot of tax between the wage paid by the employer and the wage received by the employee. Putting that tax burden on anything but labor would allow wages to fall for employers while remaining rigid for employees.

  7. David,
    I share your feelings that there seems to be something wrong with wage rigidities, but do you know of any labor market model that would not imply higher job creation if wages were lower on a fundamental level? Ultimately, if hirings is forward looking expected discounted productivity minus costs minus wages drive expected profits and ultimately vacancy posting. the empirical stories, all of them, do not observe expected productivity (including if you wish future tax burden, liquidity premia, cost of borrowing constraints whatever), so observed lower wages for newly hired workers is irrelevant as long as one can not compare it to the stream of net gains, because what matters is the difference relative to "net production gains" whatever that means lossely speaking, not the level of wages. Hence, how do you interprete the empirical studies you quote. They seem entirely uninformative given that they only observe one part of the profit flows? rigid wages are only a relative concept, no? of course wages are an endogenous object, so rigidity is only losely defined and always relative to a wage setting mechanism and there are too many anyhow, but my question is whether you are aware of a single paper that implies that an exogenously induced decline in wages somehow administered would not lead to an increase in hirings? I guess that is what you have to find to argue that people take the wrong tree? or?

    1. I share your feelings that there seems to be something wrong with wage rigidities, but do you know of any labor market model that would not imply higher job creation if wages were lower on a fundamental level?

      Efficiency wage models?

  8. David,

    What I'm missing in all these stories is a discussion of what we really observe in the data.

    Suppose, for instance, that 99% of all potential jobs have "sticky wages". That is, firms in this group don't hire as they can't lower the offered wage (say, because of internal norms). But 1% can. What would you observe in this world?

    Probably a very high unemployment rate with simultaneous stories of tremendous wage cuts for those lucky enough to actually find a job.

  9. When comparing the FRBSF data:

    to the NELP data:

    to Joe Seydl's data:

    it's important to keep in mind what it is we're looking at. The FRBSF data "examine the subset of workers in the CPS who did not change jobs during the year in which they reported earnings. We consider both salaried workers and hourly wage workers, excluding those receiving the federal or state minimum wage, since employers cannot legally reduce their pay. For salaried workers, we calculate an hourly wage rate by dividing weekly earnings by hours worked. For hourly workers, we use their reported hourly wage. Following the methodology of Card and Hyslop (1996), we compute a nominal wage change for each worker in our sample based on pay as reported early in the survey and a year later."

    The NELP data "examine employment trends in 366 detailed occupations. We formed three equal groups, each representing a third of U.S. employment in 2008: lower-wage occupations with median hourly wages from $7.69 to $13.83; mid-wage occupations with median hourly wages from $13.84 to $21.13; and higher-wage occupations with median hourly wages from $21.14 to $54.55 (all in 2012 dollars)."

    The Joe Seydl data forms five groups by dividing the 80 NAICS three digit level subsectors into five earnings quintiles of 16 subsectors each based on January 2012 earnings. They consequently have differing shares of employment.

    The FRBSF thus shows wage changes for individuals who did not change jobs. The NELP data and the Joe Seydl data shows average wages by occupation or subsector. The FRBSF supports the claim that the wages themselves are sticky. The other two data sets support the idea that aggregate wage changes are affected by composition.

    Furthermore the NELP data suggests job growth this recovery has increased the most in the bottom third. This would seem to support the idea that average real wages are countercyclical, which would match what we would expect to find if wages were sticky. But historically average real wages in the US have been mildly procyclical.

    But if you look at Figure 3 you see that is a simplistic characterization. Really what is happening over the long run is that there is an increase in earnings hetereogenity with the extremes growing about equally and the middle shrinking. If one looks at the past year, instead of the past two years, growth in the extreme thirds appears to be about equal, and this would not be at odds with procyclicality.

    The Joe Seydl data presents a different picture entirely with relative job growth by quintile depending on whether one ranks subsectors by hourly or weekly earnings.

  10. In any case all of this seems to support something I was trying to argue recently.

    We know aggregate real wages are mildly procyclical, but when one accounts for compositional bias, wages are even more procyclical. Solon, Barsky and Parker addressed the issue of compositional bias in estimates of wage cyclicality nearly 20 years ago:

    But it's not hard to tell a story where real wages themselves are countercyclical due to stickiness, and individual earnings are procyclical based on compositional bias. If nominal wages are fixed and prices rise due to an expansion, real wages fall. However as individuals get promoted and others enter the labor force, their earnings rise. So in aggregate real wages may rise.

    It's true that many New Keynesians theorize that the reason why real wages are procyclical is that wages are less rigid than prices, however I think the evidence for that is rather weak.

    "Although I can't find the precise reference right now, I seem to recall Keynes (GT) making the argument (in a later chapter) that a fall in W would make things worse by contributing to a decline in AD."

    I think you're thinking about this:

    "If, on the contrary, money-wages were to fall without limit whenever there was a tendency for less than full employment, the asymmetry would, indeed, disappear. But in that case there would be no resting-place below full employment until either the rate of interest was incapable of falling further or wages were zero."

    This can be found in Chapter 21, Section IV

    Incidentally I came across this paper by Eichengreen last week which seems to support Keynes:

    Was the New Deal Contractionary?
    Gauti B. Eggertsson
    June 2008

    "Can government policies that increase the monopoly power of firms and the militancy of unions increase output? This paper studies this question in a dynamic general equilibrium model with nominal frictions and shows that these policies are expansionary when certain “emergency” conditions apply. These emergency conditions–zero interest rates and deflation–were satisfied during the Great Depression in the United States. Therefore, the New Deal, which facilitated monopolies and union militancy, was expansionary, according to the model. This conclusion is contrary to the one reached by a large previous literature, e.g. Cole and Ohanian (2004), that argues that the New Deal was contractionary. The main reason for this divergence is that the current model incorporates nominal frictions so that inflation expectations play a central role in the analysis. The New Deal has a strong effect on inflation expectations in the model, changing excessive deflation to modest inflation, thereby lowering real interest rates and stimulating spending."

    In short two pieces of conventional wisdom are consequently busted:
    1) Sticky wages are not really a "Keynesian" concept
    2) NIRA may have been good macroeconomics after all

    1. Mark,

      Wow, you have really gone beyond the call of duty here! Thanks for the reference to Keynes.

      Do you have your own blog? If not, and if you'd like to guest post some day, let me know.

    2. David,

      I found your post very thought provoking. I'm only sorry I didn't comment on it sooner (Jackson Hole was absorbing my attention). Your appreciation has made my day.

      I don't have a blog although Scott Sumner once told me I should start one. I may do that someday when my personal life is somewhat more settled.

      I've guest posted before but your site would be a huge step up for me. I'd love to do it.

    3. Mark, please email me when you get the chance.

  11. Hiya Tech Solutions offer Best SEO Services, Please Check out our SEO Services