Not all complaints can be true at the same time

How sticky are real wages anyway?  The mainstream view is that real wages are almost completely acyclical.  But a new paper by Basu and House challenges this, here is part of their conclusion:

…there are indications that the allocative wage — the wage that governs hours worked and that firms internalize when making production and pricing decisions — may not equal the contemporaneous remitted wage.  In particular, firms and workers may well have an implicit understanding that the remitted wage will be a smoothed version of the expected allocative wage.  By estimating the expected present value of wage payments, one can construct a “user cost of labor,” which should measure the underlying allocative wage.

You can think of the employer as adjusting the “career ladder” of the worker while the contemporaneous wage remains more or less constant.  So in good times that “true” real wage goes up, and in bad times it goes down.

I would say the verdict on this idea remains out, but I found this a stimulating piece to read and it also offers an excellent survey of the literature.

Here’s the catch: on the internet I’ve read dozens — no, hundreds of times — that real wages haven’t gone up more because the Fed chokes off real wage hikes every time the economy nears recovery.  You will notice that this claim is simply flat out wrong if the mainstream view of real wage acyclicality is correct.  Somehow that never seems to come up.  (Yet it seems this ability of the Fed to stifle real wage hikes feels like it is true.)

Now, the Basu and House paper, if it is correct, actually creates an avenue through which you could start (partially) viewing the Fed as a real wage villain.  Maybe.  If some other auxiliary hypotheses were to kick in, which is not guaranteed.

But then here is the thing: you could no longer believe in traditional doctrines of wage stickiness.  And you would have to change many of your views on macroeconomics, and indeed on the efficacy of labor markets more generally.

People, repeat after me: Not all complaints can be true at the same time.  I know it is hard to live with that reality, but maybe it is worth a try.  And if you don’t like it, you can complain about that too.


I have asked this question before: when did psychology (or is it voodoo) start masquerading as economics? I understand why economists wish to be soothsayers since humans are obsessed with (afraid of) the future, but psychology is the last refuge of the clueless.

It's easy to find people making just the opposite complaint: "economists are clueless because they ignore human behavior and assume everyone is perfectly rational" etc.

Sounds like animal spirits to me.

Actually, you might find a correlate of similar data in the willingness of corporations to support executive education, sponsorship of MBA tuition, seminars, trips, etc. instead of pay because those features can be turned on and off based on the business cycle, and are of benefit to the employee.

Clearly, repositioning employees within an organization can be a model for pay increases or if no repositioning, pay stagnation.

Adding more rungs to the ladder or implicitly expanding duties to give a pay raise only goes so far: The trouble is that the world can support only so many chiefs, and it still needs Indians. (Sorry, no offense intended to Native Americans, or Bengalis). This model can only go so far because not all employees receive this implicit wage or benefit increase.

So it seems that macro is suffering from a schizophrenia between "fundamental analysis" and "technical analysis" to borrow terms from the securities business. Fundamental analysis in the macro sense looks at the quality of inputs (human capital, natural resources, innovation) and the structure of the markets (political, tax, regulatory, trade, legal) and attempts to derive conclusions and prescriptions from that exercise. Technical analysis in the macro sense looks at historical data and massages it in an infinite variety of ways to attempt to predict the future through modeling and data mining. It's possible that both have their place but it should be clear that the issues that can be addressed by fundamental analysis are vastly more important and less subject to data and model selection and bias. The basic inputs to a successful economy are not that complex and they have been known for a long time: stable families, strong cultural value of education, rule of law, reasonable taxes and regulatory policy, reasonably open trade, sound fiscal and monetary policy. If you fail on many of those fronts talking about how sticky wages are or what the Fed is doing is just a fairly uninteresting ex post exercise.

I think this paper is more narrow than that: it's saying that real wages are sticky. TC is saying that if that's true, then the traditional theories of long-run money neutrality may be wrong, and the Fed is a 'real villain'

I personally think, based on the Ben S. Bernanke 2002 FAVAR econometrics paper, which found Fed policy affects a range of economic variables only 3.2% to 13.2% (out of 100% of the change), that Fed policy is trivial. Statistically significant, as the Bernanke paper said, but not much better than random noise. 3.2% change? Nothing. It means 96.8% of the change has nothing to do with Fed policy. Money is endogenous, whose value changes as the economy expands and contracts, and is largely neutral (has no real effect except in nominal terms).

Isn't that statement meriting a 'duh!'? Economists should go back to pouring over stock exchange data. There is lots of it and an abundance of theories yet to be elucidated.

How the rest of the economy works in reality is far more complicated. And humbling. And liable to drive any confidence that a prescriptive measure will do anything near what you expect.

There are industries or sectors where quasi piece work wage schemes exist, where people get laid off seasonally. Or project based. But you can't have an indebted consumer and piecework, nor an expensive entitlement regime where the contributions are sporadic. As well in spite of the valiant efforts over a half dozen generations of business managers, people have value larger than the current short term output.

A story I heard. Calgary has been hit hard by the drop in commodity prices, oil companies of all sizes getting rid of people. But one very large residential construction firm is maintaining their production by using all kinds of financing schemes. They do a few hundred houses a year, mostly piecework labor arrangements, but it they don't have work for these people they will go somewhere else. And then when things turn around, and they will, it will take a couple of years to build up their production capacity. If it doesn't turn around, then we see a second wave layout/bankruptcy cycle that takes years to climb out of.

I know in my industry that it you lay your people off in the slow times they aren't there when there is work to do.

I'm not at all sure the instutional strucutre of financial markets is consistent with the general price model in economics. It seems to be a somewhat different institutional strucutre -- for instances, instantaineoius S & D may appear to be monotonic and correctly sloped but comparing the date to something like Cambells Tomato soup will, I suspectt profuce some rather different conclusions.

I guess the idea is to endorse any old theory so long as it justifies one's policy preferences. Who's gonna follow up on all my Tweets to make sure I actually articulated a consistent view of reality? And anyway, if I myself don't invest in a consistent world view, then I might actually believe that I can hold all these beliefs simultaneously.


"All psychological defenses have a common structure: that two legitimate but contradictory beliefs are held simultaneously, one consciously, one unconsciously, alternating variously. That way all possibilities are covered. Change is neutralized."

"But then here is the thing: you could no longer believe in traditional doctrines of wage stickiness."

Ack. This seems like a technicality. The traditional doctrine of wage stickiness is people hate getting pay cuts, so wages end up sticking. The revised version suggests that employers know this but sometimes want to cut wages anyway so they try to disguise the pay cut by keeping it from showing up in the bottom line number on the check. That is pretty close to the traditional doctrine isn't it?

Or is the objection that wages aren't really sticky so wage stickiness doesn't cause unemployment? That would only be true if employers have infinitely effective ways to hide the wage cuts, which seem implausible.

Given that a lot of the equalibrium produced in the labor markets seems to pass through the lens of corporate wage policy making don't we want to considered labor markets in a context a bit more like that of health insurance where there is a rather large (perhaps quasi-market itself) that is mediating the consumer/workers preferences and the external market position?

I'm not familiar with the literature on sticky real wages, but nominal wages seem to be sticky to the downside more than the upside.
And if Fed policy attempts to keep unemployment from getting "too low" because that would spark wage inflation, you could claim that it is restricting growth in real wages. In a tight labor mkt, labor's share of income should be higher and it is very reasonable to expect wage growth to outpace CPI growth. Rightly or wrongly, the Fed often acts to prevent a tight labor mkt.
Also, to some degree the fact that the Fed is following a certain rule should cloud the data and make it harder to infer what would happen if the Fed were following a different rule.

"You can think of the employer as adjusting the “career ladder” of the worker while the contemporaneous wage remains more or less constant. So in good times that “true” real wage goes up, and in bad times it goes down."

I'm hardly an expert, but how does adjusting the career ladder disprove sticky wages. How does it result in wages going down?

I would agree that employers would seem more likely to give lateral "promotions" or promotions without real pay raises during bad times. But not getting real (or even nominal) raises doesn't contradict the main view of sticky wages, as far as I know.

I thought the conventional view was nominal wages are sticky, at least to the downside, thus inflation helps while deflation hurts.

I know there are theories of real wage stickiness. But the old fashioned Irving Fisher and J. M. Keynes view is that the NOMINAL wage was sticky, especially in regards to absolute declines. (They may be slow to adjust upward also.)

The theory was to explain why deflationary shocks caused unemployment instead of just nominal wage declines.

As for real wage stickiness, never reason from a price change. (H/T to Scott.) The labor market can be hit by demand or supply shocks. A backward shock to supply would push real wages up and employment down, a backward shock to demand would push real wages down and employment down. If cycles are a mixture of demand and supply shocks, averaging over all cycles, we could get consistent declines in employment but conflicting results for real wages.

Not being an economist, intuitively I would have thought wages had a sticky nominal floor, but real wage growth would vary with the relative bargaining power of firms versus labor. This is a function of the unemployment rate, which is cyclical.

Comments for this post are closed