Cowen and Smith on monopoly and stimulus

In a recent Bloomberg debate, there was an interesting exchange.  First, Tyler Cowen:

But Noah, I have a question for you. You’ve written several columns about how the American economy is becoming more monopolistic. If true (and it is not exactly my view), that implies output could be much higher with current resources, even at full employment. A boost in demand could spur firms to produce more, rather than restricting output so much. So are you now a fan of these Trumpian deficits? They may not be your preferred form of deficit spending, but do you see them still as a net positive?

Then Noah Smith:

As you say, monopoly power could potentially increase the case for stimulus in bad times.

Actually, that’s not what Tyler said, nor is what Tyler said true.  (Now everyone will be annoyed at me.)

One of the fundamental principles of modern macro is that demand-side stimulus cannot solve real problems.  It can overcome problems such as high unemployment caused by sticky wages and prices combined with inadequate spending, but that’s all it can do.  Inefficiencies associated with monopoly are a real problem, and cannot be solved by printing money.  There are actually a number of issues here that need to be disentangled, some of which are quite subtle.

1.   Monopoly is a microeconomic problem, not a macroeconomic problem.  Thus it’s quite possible to have low unemployment rates and high levels of monopolization.  Indeed, I’d argue that’s true in America right now.  Employment in the monopolistic sector is indeed lower than we’d like, but the result of this is not unemployment, it’s workers being employed in the less efficient competitive sector of the economy.  This is important, because the mechanism by which demand stimulus creates growth is by encouraging more employment (not moving workers between sectors).  But we are already at full employment.

2.  Suppose I’m wrong, and monopolization causes the natural rate of unemployment to be higher than otherwise. Say America’s natural unemployment rate rises to French levels, due to monopolization.  Is Tyler correct in that case?  No, demand stimulus is still not called for even if monopolization causes the natural rate of unemployment to be higher than otherwise.  That’s because when you are at the natural rate, demand stimulus basically tricks workers and firms into producing more output than they’d like, by pushing up nominal spending in the face of sticky wages and prices.

So doesn’t that make us better off?  In the short run yes, but only at the cost of being worse off in the long run.  When prices are sticky, demand stimulus can reduce a monopoly’s real price, which is its price relative to NGDP.  But once the monopoly catches on to the higher NGDP, it will raise the real price again.  That might not sound so bad, but it leads to cyclical instability.  Ditto for wage stickiness.  Demand stimulus will give monopolies an incentive to hire more workers, as long as nominal wages are sticky.  That will indeed make the economy more efficient for a short time (this may have been Tyler’s intuition), but at a cost of future instability.

This is why we have independent central banks.  Because our economy is riddled with inefficient policies such as minimum wage laws and taxes on labor, our natural rate of output is suboptimal.  Demand stimulus tricks us into producing more, and we move closer to the optimal position for the economy.  But it’s not sustainable. It’s a sugar rush.  Minimum wages eventually get increased with NGDP, and workers renegotiate contracts.  In the short run, the stimulus really does make us better off as a country (with or without monopoly), but it overheats the economy and leads to a painful recession in future years.  Once mainstream monetarist and New Keynesian economists understood this problem, they decided the best we could do was to keep the economy close to the natural rate of unemployment, and then advocated setting up independent central banks that would be immune from pressure by a corrupt future president that might have a big ego and a short attention span.  (Hmmm . . . . )

Unless I’m mistaken, there is nothing particularly controversial about this post.  Think about a standard NK model, which produces an optimal policy of 2% inflation.  How would the existence of monopoly change the optimal policy?  Make it more expansionary?  But what does that even mean?  In the standard model, money is neutral in the long run.  Going to 3% inflation doesn’t have any long run benefit.  I suppose you could advocate steadily rising inflation, ending up in hyperinflation, but that won’t work if there are any welfare costs of inflation.  In fact, the optimal policy under a NK model is no more expansionary with monopoly than without.

Now I suppose there might be models where the optimal policy is more countercyclical if there is monopoly, and this seems to be what Noah is hinting at.  But that doesn’t help Tyler’s argument, as in that case policy should actually be more contractionary when unemployment is 4.1%.  And I’m not even sure that claim is true; do NK models imply that more weight should be put on output, and lesson inflation, when the economy is more monopolistic? Are there any models that do so?

PS.  Tyler might argue that the monopoly argument was not his view, just the implication of Keynesian models with which he does not agree.  But I’m saying even that’s not true.  The argument he makes is not even an implication of any sound Keynesian model that I’m aware of.

PPS.  I have a new post criticizing proposals to “experiment” with a hot economy, over at Econlog.




21 Responses to “Cowen and Smith on monopoly and stimulus”

  1. Gravatar of Student Student
    16. February 2018 at 11:35


    1.) that we have the tax cuts and budget deficits no matter what at this point,
    2.) this isn’t a permanent thing,
    3.) we do have high levels of monopolization;

    Then wouldn’t it make sense to simply roll with the state of things as they actually are and just to a higher inflation target?

  2. Gravatar of Student Student
    16. February 2018 at 11:36

    *move to a higher…

  3. Gravatar of Mike T Mike T
    16. February 2018 at 11:37

    Good post. I agree stimulus cannot solve real problems. This point about monopolization from Smith or Cowen is not clear at all. I suppose the only argument that seems sensible is that on average central banks and government may have to change size of counter-cyclical required to fix sticky price/information distortions. But I’m not sure this is even clear.

    Cochrane’s recent paper in the JME takes the baseline NK model. And the less flexible prices become (up to perfectly flexible)the more effective stimulus becomes, and recessions become deflation is initially worse to a shock to the real rate. But there is a giant discontinuity at the point of perfectly flexible prices. Moreover these results drastically change by equilibrium path specification.

    My point is not to support this paradox in the NK model. But simply to highlight that its not clear at all from the NK model that these permanent interventions are justified to correct for the increased monopolization. I think the only outcome thats justified is a change in the path of r relative to r* for longer to hit the SAME ngdp target.


  4. Gravatar of Tyler Cowen Tyler Cowen
    16. February 2018 at 11:40

    A lot of the current shock is fiscal and thus has a significant real component, for one thing. Also, in models where nominal and real stickiness interact it is easy enough to get the desired result from nominal shocks too. Circa 2018, my first point is more important, but either way I think your main claim is wrong!

  5. Gravatar of Mike T Mike T
    16. February 2018 at 11:44

    typo “…the more effective stimulus becomes, and deflation in recessions is worse as prices become MORE flexible.”

  6. Gravatar of ssumner ssumner
    16. February 2018 at 12:24

    Student, No, that would be procyclical and hence very destabilizing. If we are going to move to a higher inflation target, the transition should occur during the next recession.

  7. Gravatar of ssumner ssumner
    16. February 2018 at 12:25

    Tyler, See my new post.

  8. Gravatar of Mike T Mike T
    16. February 2018 at 13:02

    Wow I should do a better job proofing. Cochrane directly “This increase in multipliers presents an interesting policy paradox. Microeconomic efforts to reduce price stickiness make the depression worse, according to Fig. 2. But such efforts make multipliers larger, increasing the effectiveness of fiscal or broken-window stimulus.”

  9. Gravatar of Rajat Rajat
    16. February 2018 at 17:37

    Slightly off-topic, and maybe you’ve covered this elsewhere, but why do you think the US has high levels of monopolization?

    I don’t think you simply mean high levels of concentration – which I find easier to accept – because you go on to say that employment in the monopolistic sector is lower than optimal. That would only be the case if the sectors in question were exercising market power. It doesn’t seem that tech firms or retailers are withholding supply to raise prices and profits, although you’ve said before that such firms make up a relatively small share of the economy. So which sectors did you have in mind? Regulated sectors such as healthcare, higher education, and perhaps banking? Others?

  10. Gravatar of Jerry Brown Jerry Brown
    17. February 2018 at 06:34

    When Tyler says in his question that because of monopolies, output could be much higher with current resources even at full employment, is that because monopolies tend to use their current resources inefficiently? And if there were more demand then that would spur them to use the same resources more productively? That would work if it were the case that monopolies are just bad managers of resources, right? But why would monopolies not have an incentive to use resources most productively right now? Maybe a lack of competition allows them to pass on the higher than average costs to the customer.

  11. Gravatar of ssumner ssumner
    17. February 2018 at 15:34

    Rajat, I wouldn’t say levels of monopolization are all that high, but I do think the IT industry has caused them to be higher than before, due to patents and network effects.

    Jerry, No, he’s saying monopolies produce too little output.

  12. Gravatar of Jerry Brown Jerry Brown
    17. February 2018 at 20:05

    Yeah but why? It doesn’t seem to make any sense otherwise. How can production go up if you are already at full employment without increasing productivity? I think Tyler has the argument wrong.

  13. Gravatar of dtoh dtoh
    17. February 2018 at 20:07


    “He’s saying monopolies produce too little output.”

    I have a hard time following this. I can see how monopolies cause a mis-allocation of production between industries, but I have a hard time seeing a significant mechanism whereby this would translate into too little aggregate output. And if the problem is not too little aggregate output, I don’t see how stimulus would impact the problem.

    And even if monopolies cause some form of sub-optimization, I don’t see how this ranks very high on the list of economic problems that need to be solved.

  14. Gravatar of Scott Sumner Scott Sumner
    18. February 2018 at 14:56

    Jerry, I suppose there are a couple possible mechanisms; I’m not sure which one he’s suggesting.

    dtoh, I can even see how it would slightly reduce overall output by moving production away from the sectors where it is most useful. But even then I fail to see how it is a problem that demand stimulus could address. It seems like an efficiency problem.

  15. Gravatar of dtoh dtoh
    18. February 2018 at 19:22

    I see how that reduces utility, but I don’t really see how it reduces aggregate output. I don’t even see it as an efficiency problem….. you can produce the wrong things very efficiently.

  16. Gravatar of Jerry Brown Jerry Brown
    19. February 2018 at 07:16

    Thanks. In case it wasn’t clear, I think you are correct when you say “Tyler might argue that the monopoly argument was not his view, just the implication of Keynesian models with which he does not agree. But I’m saying even that’s not true. The argument he makes is not even an implication of any sound Keynesian model that I’m aware of.”

  17. Gravatar of ssumner ssumner
    19. February 2018 at 18:00

    dtoh, Output is supposed to measure utility. If it reduces utility then it reduces RGDP, properly measured. Suppose we produce fewer iPhones and more wheat. The physical output might be the same but real GDP will fall, as iPhones count more in real GDP than bushels of wheat.

  18. Gravatar of dtoh dtoh
    19. February 2018 at 20:32

    Not really following. If we produce $10 of iphone or $10 wheat. Isn’t impact on RGDP the same?

  19. Gravatar of ssumner ssumner
    20. February 2018 at 11:03

    dtoh, Not if workers are more productive at making iPhones than wheat. GDP is not measured in pounds or tons, it’s based on consumer willingness to pay. If monopolies cause too little production of the highest valued stuff, then GDP falls.

  20. Gravatar of dtoh dtoh
    20. February 2018 at 17:05

    Not sure that follows without making some specific (and probably not necessarily valid) assumptions such as productivity is higher in the monopoly industry than in other industries and that the slope for marginal productivity of labor is very steep in non-monopoly industries.

    Basically, I think the analysis boils down to reduced output in the monopoly industry leaves extra labor and capital which gets deployed in other industries and produces other goods at a market clearing price.

  21. Gravatar of ssumner ssumner
    21. February 2018 at 10:15

    dtoh, The fact that productivity is higher in the monopoly industry is the basic assumption of the textbook monopoly model that Tyler relies on. If productivity is not higher at the margin, then monopoly is not a social problem in the first place. There would be no deadweight loss from monopoly.

Leave a Reply