Scott Sumner  

Don't blame the workers

PRINT
Everything's Amazing and I'm H... There is no sticky wage puzzle...

A popular Keynesian theory of recessions is that nominal wages are sticky, and hence reductions in aggregate demand lead to high unemployment. They often go on to advocate more government spending to put people back to work. Free market economists often get frustrated by this argument. Sometimes they blame unions. Sometimes minimum wage laws. Often there's a suggestion that uneducated workers suffer from some sort of "money illusion" and hence are too dense to understand that they need to accept lower wages to save their jobs during a recession.

I'm a free market economist, and yet I don't think much of this tendency to blame workers. Indeed I see not one but three fallacies involved here:

1. It's not just labor markets: Fluctuations in nominal GDP growth cause two big problems, labor market instability and credit market instability. We all know about nominal wage stickiness, but it turns out that debt contracts exhibit even more nominal stickiness than wages. Suppose GE issues a corporate bond, which is purchased by institutions such as pension funds and insurance companies. What can we say about this transaction?

First, it's a relatively free market. There are no labor unions involved, nor minimum wage laws. There are no uneducated people; the GE finance people and the institutional bond buyers have all taken economics in college. And yet the bonds are almost never indexed to NGDP, they are nominal bonds. This practice leads to financial crises when NGDP growth is highly unstable.

2. It's also companies: It takes two sides to negotiate wage contracts. And wages are not just sticky in the downward direction. In the late 1960s, NGDP growth accelerated to a very high rate. If wages were perfectly flexible they would have risen even faster than they did. Instead wages were sticky due to long term wage contracts, and the faster NGDP growth led to extremely low unemployment by the end of the 1960s, about 3.5%. I doubt it was the workers who objected to breaking the contracts and offering even higher wages.

3. It's an externality problem: A system of complete wage flexibility offers huge external benefits, by making the business cycle milder. But that is mostly an external benefit; it's very rare for workers at individual firms to be able to save their jobs with wage cuts.

Consider a sudden 8% decline in NGDP growth per capita, relative to trend. One example occurred between mid-2008 and mid-2009, when the growth rate plunged from the normal 4% to negative 4%. Now suppose you are a group of nurses. Would you be anxious to accept an 8% wage cut relative to your previous expectation? I don't see why, nurses aren't likely to lose jobs in a recession, nor are lots of other people, such as tenured teachers. But if they don't accept wage cuts then just imagine how big the wage cuts must be in other sectors to reduce the aggregate nominal wage rate by 8% below trend. Even worse, if you are a worker in a windshield wiper plant, good luck saving your job with wage cuts when the automakers you supply stop making cars.

The macroeconomy is a very complex system. Macro problems cannot be understood by thinking in terms of workers at a "representative firm," because very few workers are at such firms. This means there are huge externalities to wage (and price) flexibility, which makes it even less likely that society will reach an equilibrium with sufficient wage flexibility.

Engineers don't bemoan the existence of gravity when they design airplanes. Nor do they blame gravity for air crashes. Free market economists should not blame workers for money illusion, they should design monetary regimes where the existence of money illusion by workers, firms, lenders and borrowers does not lead to business cycles and financial crises. That way the Keynesians won't be able to ask for more government spending.


Comments and Sharing






COMMENTS (17 to date)
mico writes:

I agree with your post. Let me ask you two questions:

1. Can you imagine a free market monetary regime that gives good answers to these questions? Making central banks better is very practical but central banks are fundamentally non-free market institutions.

2. Does manipulating interest rates/the money supply in the way you suggest have any negative externalities that mist be weighed against the benefits?

Scott Sumner writes:

Mico, I actually don't favor manipulating interest rates. I favor letting the market determine the level of interest rates, and the money supply. As far as negative externalities, if the Fed target is stable growth in expected future NGDP, then I would not foresee much in the way of problems. But again, interest rates should be set by the market, not the Fed.

andy weintraub writes:

I've always favored the argument put forth, I think, by George Stigler. That is, that it takes time for information about a downturn to makes its way into the minds of people who've been laid off. And during that time, they are reluctant to accept a lower wage. Meanwhile, firms find it easier to lay off a few workers rather than lower the wages of all.

It's been a while since I read Stigler on that topic, but I think I've got it right.

Kenneth Duda writes:

> Free market economists should not
> blame workers for money illusion, they
> should design monetary regimes
> where the existence of money illusion by
> workers, firms, lenders and borrowers
> does not lead to business cycles
> and financial crises.

This is it exactly. Stabilizing NGDP isn't "meddling with the free market". It's monetary policy that makes it easier for free-market participants to do their jobs --- they can count on at least nominal stability. We can negotiate contracts, credit agreements, supply agreements, employment agreements without worrying about the possibility of economy-wide swings in prices or spending.

Kenneth Duda
Menlo Park, CA

ThomasH writes:

"They often go on to advocate more government spending to put people back to work."

This is generally correct because when there is a recession, monetary authorities generally drive down the rate at which governments can borrow which leads to more projects having positive NPV. If projects also use inputs that are unemployed, then there will be a difference between the price and the marginal cost of using that inputs that would lead an optimizing government to use more of it.

Unfortunately, in the most recent recession many governments have moved in the opposite direction, reducing investment. This results in monetary authorities having to use stronger doses of unconventional instruments.

ThomasH writes:

"That way the Keynesians won't be able to ask for more government spending."

Why should only "Keynesians" advocate additional investment in the circumstance you describe, the 2007-08 downturn in NGDP growth? It seems to be the standard neoclassical response of an optimizing government.

Perhaps you meant that if monetary authorities were more sure footed and could insure that departures of NGDP from trend were extremely short, then the optimal fluctuations in government investment would be quite small.

ThomasH writes:

"I favor letting the market determine the level of interest rates, and the money supply."

I can see that varying interest rates and the money supply are not the only ways that monetary authorities can keep NGDP on target, but isn't it likely that when NGDP is below target whatever instrument they use (buying NDDP futures?) will result in lower interest rates?

Scott Sumner writes:

Andy, That may be part of it, but I think money illusion is also a factor. I'll do a post soon on wage stickiness.

Thanks Ken.

Thomas, You've made that argument for fiscal stimulus many, many times before, including recently. And I've responded many, many times before, including recently. Do you have a response to my response?

Regarding low rates, in general a more expansionary monetary policy is usually associated with higher interest rates than a more contractionary monetary policy, although not always. A good example is 1965-81, when a shift to a highly expansionary monetary policy led to much higher interest rates.

In late 2007 and early 2008 a tight money policy led to much lower rates.

mico writes:

"Mico, I actually don't favor manipulating interest rates. I favor letting the market determine the level of interest rates, and the money supply. As far as negative externalities, if the Fed target is stable growth in expected future NGDP, then I would not foresee much in the way of problems. But again, interest rates should be set by the market, not the Fed."

What I am asking is, imprecisely, how could market monetarism arise without a Fed, in a system in which money is entirely privatised?

Does a true free market system have to accept severe business cycles? If so, that seems like a strong failure of a true free market - much stronger that I've seen in any other sphere.

Daniel writes:

I "love" it how you said free-market economists when everybody knows you really meant "internet austrians".

Why this need to mince words ?

It's not like they're open-minded people who can be persuaded by evidence. Nor are they taken seriously by anyone else.

BC writes:

Mico, if the Fed causes severe demand-side recessions by artificially keeping money too tight and causes severe demand-side inflations by keeping money artificially loose, then I don't see how that represents a failure of the marketplace.

An interesting question, though, is whether financial innovation can help alleviate sensitivity to Fed mistakes. For example, I wonder whether employees that are compensated with company stock tend to face fewer layoffs than those compensated with straight cash wages. Company stock is not the medium of account, so an employees' real wage can fluctuate even if their nominal wage in shares of stock is fixed.

Scott Sumner writes:

Mico, I'm not sure what would happen with a fully privatized monetary regime. Perhaps it would target NGDP, perhaps not. I simply don't know. If such a regime had occurred 150 years ago, I'd guess it would have stabilized the nominal price of gold or silver, which is not a MM target. Today? I'm not quite sure what it would look like.

Benoit Essiambre writes:

This is a great point. No one wants to be the first to adjust their wages downwards to save jobs, but even when some are willing (maybe because the risk of ending up unemployed looks very high to them), it won't help much since to be able to save their job without lowering their wages excessively, not only does their own has to go down, but their employer's suppliers' wages have to go down as well as the suppliers' suppliers'.

In a market suffering from the effect of generalized too low prices, the whole upstream production chain may have to drop wages in a synchronized manner for the burden to be spread broadly enough to have a realistic chance at saving a business.

I don't think we can expect this to happen very smoothly without monetary mechanisms in place.

Otherwise, the rational action for an individual unable to save his or her job, might be to extract the most out of it while employed.

tl;dr Above market, business destroying wages can be a Nash equilibrium.

BC writes:

Scott, wouldn't private NGDP money look something like shares in a mutual fund that invests in securities whose USD returns are correlated to NGDP_USD (NGDP measured in USD)? NGDP measured in shares equals NGDP_USD/NAV, where NAV is Fund net asset value (USD/share). So, if the Fund is long NGDP futures (if they exist), then NGDP_shares will be very stable. If NGDP futures don't exist and the correlation between Fund returns and NGDP is imperfect, then the Fund can adjust NAV through share splits and reverse splits.

With NGDP_shares stable, "all" that is left is to get people to agree to denominate their wages in shares instead of USD. Employers could invest in the Fund and entice employees to accept pay in the form of Fund shares (with no automatic adjustment for splits) in exchange for either guaranteed employment terms (no layoffs) or guaranteed severance pay. Employers should like share-denominated wages because share NAV falls during periods of falling NGDP. For employees, there should be some amount of job security/severance that convinces them to accept the fluctuating wage risk.

I have no idea whether a Fund like this can be implemented using bitcoin technology such that Fund management and administration can be distributed and not require a trusted authority.

Unrelated question: Does the existence of unemployment insurance make workers more willing to accept unemployment risk and, hence, less willing to accept wages that fluctuate with NGDP? Understand your point about externality, but employees do accept some amount of incentive pay (stock, commissions, etc.) that is related to firm revenue that presumably reduces unemployment risk. Perhaps, in the absence of UI, incentive pay would increase.

Scott Sumner writes:

Benoit, That's right.

BC, I don't think workers want to be paid in anything other than the medium of account. Our best plan is to insure the medium of account has a stable value.

I have no idea if a private producer of money would tie the medium of account to NGDP.

I do think that UI makes wages a bit stickier, but only by a small amount.

Michael Byrnes writes:

Excellent post and excellent comment from Benoit Essiambre.

There are some people who seem to think that a fully flexible monetary system would be optimal - I don't buy that that is the case. Surely there are situations where more flexibility in prices would be beneficial, but I don't think that's enough to prove the general case.

Isn't the money illusion sort of a means of coping with limited and imperfect information? In an unstable monetary environment, workers and firms don't know exactly what wage change would be "neutral" in real terms. I imagine that they will have some big picture sense that the economy is entering a recession, but not the cut in their wage that would promote stability. In this situation, workers will, as now, always have a reason to want higher nominal and real wages and firms will always have a reason to want the opposite, at least to the extent they can cut real wages without hurting the firm. (But one can also imagine cases where a struggling form would avoid cutting wages for signialing reasons.)

Nominal income stability makes it easier, not harder, for individuals and firms to make these decisions.

Andrew_FL writes:
Free market economists [...] should design

I could hardly think of anything less free market than presuming to design how it works.

Comments for this entry have been closed
Return to top