When I was a young economics graduate student at UCLA in the early 1970s during the heyday of that wonderful department, convinced, like most of the other UCLA grad students, that I was in the best graduate program in the country (and therefore the world), where the deepest, most creative, economic theorists in the world, under the relaxed and amiable leadership of Armen Alchian — whose failure to win a Nobel Prize is really the failure of the Nobel selection committee — would eventually succeed in unifying economic theory by reformulating macroeconomics on correctly specified microfoundations, I naively entertained for a while the idea that I would write a Ph.D. dissertation on some aspect of the problem of price rigidity. However, I never could do more than formulate some general observations about the theoretical role of price rigidity in macroeconomic models and compose a superficial survey of the empirical literature on price rigidity, starting with the work of Gardiner Means in the 1930s on administered prices up to a volume, The Behavior of Industrial Prices, by Stigler and Kindahl that had just appeared. After many months of frustrating and fruitless efforts to come up with a hook on which to base a dissertation, I came to the painful conclusion that my theoretical ambition exceeded my intellectual resources and I would have to look for a dissertation topic that I could get my arms around.
I then went to the other extreme, choosing a straightforward empirical topic, a comparison of car insurance premiums in states with different regulatory regimes, eventually finding that, as suggested by the regulatory capture theory, stricter regulation was associated with higher, not lower, premiums. But in choosing a dissertation topic in which I had little emotional or intellectual investment, I paid an unforeseen cost of another sort; my lack of passion for my dissertation made me a poor candidate in the job market. So the only job offer I got was from an undistinguished upper-midwestern economics department.
These not so pleasant reminiscences were triggered by my response a couple of days ago to the comments of Nick Rowe, Scott Sumner, Anon, and Wonks Anonymous on my posting “Krugman on Keynes and the Moderns.” In that post, I had chided Krugman for dismissing without any, much less adequate, explanation Robert Barro’s assertion that Keynes’s theory of high unemployment could be reduced to the following: wages and prices are stuck at a level too high to allow full employment; the problem could therefore be solved by monetary expansion raising equilibrium wages and prices, thereby obviating the reduction in wages and prices that “price stickiness” had been blocking. In response, Krugman merely harrumphs, and complains that Barro doesn’t get it.
In his comment, Nick distinguished between the cause for the decline in AD and why, once AD has declined, it is translated into a reduction in output and employment rather than a pure reduction in prices and wages with unchanged output and employment. According to Nick’s interpretation, the Krugman/Barro dispute is seemingly reduced to a verbal dispute about the meaning of the word “cause.” But, presumably, Krugman wants to say that a reduction in AD, must, for reasons deeper than mere “price stickiness,” have output and employment effects, not just price effects.
Then Scott weighed in with the following comment.
If wage and price stickiness aren’t needed for the Keynesian model, do the model without them. I don’t see how it can be done. I’ve never seen a Keynesian model with complete wage and price flexibility. If NGDP falls 99%, and so do wages and prices, how is there unemployment?
I responded to Nick and Scott as well as to further comments by Anon and Wonks Anonymous, but my comments may have been too terse to have been comprehensible. At any rate, there was no response to my comment, so I don’t know if readers came away nodding or shaking their heads.
At the risk of boring even those who have made it this far, let me try to make a couple of crucial, but rarely noted, distinctions about terms like “price stickiness,” “price inflexibility,” or “price rigidity.” On the one hand, “rigid prices” can mean that, even though supply and demand have shifted in a way that implies that the price should change, the price doesn’t change. On this interpretation, “price stickiness” is a kind of (perhaps externally imposed) market failure. The virtue of the price system, as Hayek taught us, is that it transmits information about alternative uses for and supplies of resources, leading to efficient resource allocation with no need for central direction, purely through voluntary responses to price signals.
On this interpretation, “sticky” prices constitute a fundamental failure; prices and wages, either because of direct government intervention or monopoly privileges, do not adjust in the “normal” or “proper” way to changes in demand and supply; the price mechanism doesn’t function as it is supposed to. Well, then, the argument goes, if the price mechanism is malfunctioning, because prices are “stuck” at levels too high for markets to clear, of course output and employment will contract. If that is all there is to what Keynes was saying, what’s the big deal? Everybody knew that.
I interpret this to have been Barro’s point about which Krugman complained. And, Scott Sumner seems to agree with Barro that Keynesian economics is nothing other than the economics of sticky prices. Nick Rowe, however, if I understand him correctly, at least wants to leave open the possibility that Keynesian economics is about more than just the assumption that the price mechanism is malfunctioning.
Here is where my UCLA training, and my seemingly fruitless efforts nearly 40 years ago, may give me (but certainly not just me) some added insight into the problem. Let me ask the following question. When aggregate demand drops, would we really expect workers immediately to take wage cuts and businesses immediately to reduce prices, the decline in aggregate demand being entirely reflected in instantaneously falling prices and wages, with no reduction in output and employment? I doubt it. At the moment aggregate demand falls, how many people are even aware of what has just happened? It’s not easy to distinguish between a general decline in demand and a decline limited to just your own product. If you are a worker told by your employer that you are being laid off because his sales are down, would it be more rational for you to ask how big a wage cut would allow you to hang on to your job, or to assume that some other employer would be willing to pay you a wage close to what you had been earning. (And if there was none, why was your old employer paying you a much higher wage than anyone else was?)
This is the search rationale for unemployment developed at UCLA in the early 1960s and discussed in the classic introductory text University Economics by Alchian and Allen, later developed by others like Peter Diamond into a theory for which Diamond, not undeservingly, did win a Nobel Prize. Axel Leijonhufvud, who came to UCLA while turning his own Ph. D. dissertation into a wonderful book On Keynesian Economic and the Economics of Keynes, used the search-theoretic explanation of unemployment to suggest an interpretation of Keynes that didn’t rely on wage and price rigidity in the first sense.
Another UCLA luminary, Earl Thompson, restated the same basic point more elegantly in a Hicksian temporary equilibrium framework. In temporary equilibrium, as understood by Hicks, individual supply and demand decisions depend on the possibly incorrect and conflicting price expectations of each transactor. Only in full general equilibrium are all price expectations correct, but in temporary equilibrium prices do adjust to clear markets despite incorrect price expectations. Suppose price expectations are too high, as they are after a decline in aggregate demand, then the quantities offered for sale by transactors will be less than would have been offered if expected prices were lower. Given that expected future prices are too high, the price mechanism is working as well as it can. Prices are not sticky; no price adjustment would induce a mutually beneficial transaction given the price expectations held by the transactors. But those incorrect price expectations, nevertheless, cause a cumulative contraction of output, with shrinking aggregate demand. Even though price expectations may be revised downward, the fall in aggregate demand may prevent restoration of full equilibrium with correct price expectations.
That is how you can have declining real output and employment even with fully “flexible” prices in the only sense of the term that I can conceive of. If someone wants to call this Keynesian or involuntary unemployment, it is fine with me. But I don’t think that the formal apparatus of what is commonly understood to be Keynesian economics is at all necessary to understand the mechanism. And I’m not so sure that the Keynesian model really helps us understand the nature of the dynamic at work in this situation.
Those were the days.
Uneasy Money 
Great Post. Clear (and nostalgic). Nick Rowe today tried something along does lines, but this time I don´t think he brought his point across very clearly (at least to me).
http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/07/the-peanut-theory-of-recessions.html
LikeLike
As Steve Horwitz likes to point out, this is just the flip side of the Mises-Hayek theory of a false expansion in the case of money supply exceeding money demand. And as there, so here, a nice illustration of the effects of heterogeneous (non-perfectly substitutible) capital that must be re-priced through a process of Hayekian discovery.
LikeLike
Nick Rowe has a further related post here: http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/07/why-we-are-almost-always-less-wealthy-than-we-think-we-are.html
The underlying imperfection in your example is that there is no prediction market for AD (or equivalent). Because if there were such a market, speculation would quickly produce more accurate expectations (assuming that the market was sufficiently liquid, due to e.g. noise trading or hedging activity).
This is actually a big problem for ABCT as well, which posits that entrepreneurs can be systematically mistaken about future interest rates as a result of central banks’ policy actions. But liquid markets for future rates do exist, and their predictions are not obviously inaccurate.
LikeLike
I don’t have a problem with that, but then I’d still call that sticky prices. Which suggests we are just differing on semantics.
Do you see a link to the Lucas monetary misperceptions model?
LikeLike
Contracts are sticky.
LikeLike
Rummaging around my econ class days, I recall this: That many employees have skills particular to their employer. After proving yourself on the job, and learning the ins and outs of your particular employer, you are more valuable than an equally talented guy off the street. Training up someone new would be expensive, and even risky. After several months, the new guy might prove out not be as good as you. Then your employer would have to start searching anew, while suffering reduced output or lost sales etc.
So the employer has a stake in your continued employment. Cutting wages to every wrinkle in the economy is not done. So downward stickiness is a problem.
Employees also know they are more valuable to their employer than a new employer. This reality I think keeps people in jobs longer, at steadier wages, than might be thought otherwise.
LikeLike
Marcus, Thanks again for your kind comments and the reference to Nick’s post, which I haven’t yet had a chance to look at. Perhaps I will have a further comment after I look at it.
Eric, Thanks for pointing out the connection. I met Steve many years ago and we have been in touch on and off over the years, but not recently. But just to be slightly argumentative, there is nothing, repeat nothing, in the theoretical account of Hayekian discovery to guarantee that discovery process will arrive at, or even approach, an equilibrium price vector that brings the system back towards full employment. In his magnificent essay “Economics and Knowledge” Hayek stated explicitly that the only basis for assuming that there is any tendency toward intertemporal equilibrium is past experience that this has been the case. There is no a priori necessity for individual expectations of future prices to be correct or even in agreement, and there is no market mechanism for bringing about the agreement of individual expectations. Ludwig Lachmann was the one Austrian who was explicit in drawing out the logical implications of this fact, which most Austrians seem to ignore.
Anon, Thanks for the additional reference. I agree that the absence of a prediction market for AD is a critical incompleteness in the system. I am not totally convinced that adding such a market is a sufficient condition for completing the system. On the latter point, I am not trying to be argumentative, just acknowledging that I haven’t thought it through to my satisfaction. I agree also with your comment about ABCT, a criticism, by the way which is implicit in Hawtrey’s discussion of the relation between short and long-term interest rates in A Century of Bank Rate.
Scott, I didn’t think that you would have a problem with it, but I wasn’t totally sure, so I am glad to have your response. And I agree that the issue is semantic, but it is not just semantic, because you want to use the same word to describe two distinct phenomena. Yes, I do see a link to Lucas. I think Lucas posed the problem in the wrong way and came up with a not very useful way of analyzing the problem which helped pave the way for the disastrous RBC theory, which even he is not willing to buy into.
Carl, Agreed. I don’t deny that there may be prices that are sticky. I wanted to distinguish between different kinds of stickiness. Actually market prices can sticky even without contracts. Dennis Carlton succeeded in making an important theoretical and empirical contribution about sticky prices in his doctoral dissertation at Chicago. In his work, he distinguishes between the sticky prices that businesses set for their “regular” customers who in return for paying a price that on average is higher than the “spot” price get priority in being supplied in times of “shortage” when the “spot” price rises sharply but the “sticky” price is maintained at the customary level.
LikeLike
As Hicks tells us in many places, the Hicks work was inspired by Hayek ….
LikeLike
I agree there can be latency/stickiness in wages and pricing aside from contracts. But why argue about it? And there are contracts such as purchase agreements that stipulate or guarantee nominal prices. But I was talking about loans. They’re kindof a big deal. Households on average live something like 5-15 years in their future with respect to 1/3 of their gross income. Nominal shocks are a Big Deal to these people. End of argument, no? Regardless of whether Amazon updates their prices every 10 seconds, or minimum wage is abolished, there are still enough mortgages, car loans, etc. to make nominal shocks real.
LikeLike
Benjamin, That is all correct and provides an explanation for why wages don’t adjust to every change in demand. Workers who have been working for an employer have developed specific capital which makes other workers imperfectly substitutable for them and vice versa. But the question still arises why wages aren’t cut instead of laying workers off in a recession.
Greg, I totally agree that Hicks was very much influenced by Hayek, and I never suggested otherwise. However, in The Pure Theory of Capital, Hayek takes issue with Hicks’s formulation of the concept of temporary equilibrium even though it is a direct application of Hayek’s concept of intertemporal equilibrium. I never understood why Hayek disliked Hicksian temporary equilibrium. I once asked him about that and I didn’t follow his answer which, since I asked him about it cold, was probably not a question that he was prepared to answer off the cuff. I think that he may have said that he needed to go back and rethink what he said about temporary equilibrium, but it is at least 25 years since we talked and my own recollection is fading.
Carl, That’s an interesting question to ask a blogger. What else are we here for? At any rate, I don’t think the fixed mortgage payment is a good example, because the payment is already fixed so that no change in the allocation of resources depends on the fixed mortgage payments. They are a pure transfer of wealth from one party to another. Changes in the value of money affect the size of the transfer not the allocation of resources. Of course, default is possible, but that is not what we normally think of when we talk about sticky prices affecting output and employment.
LikeLike
This is how I see it…. Keynesians have two arguments which they combine. Firstly there is the old argument about price stickiness affecting real output, which is quite correct. Then there is the argument about Say’s law. As I understand it the essence of the argument about Say’s law is that demand may shift towards assets that are not reproducible in the short-run. Demand can shift towards money, land, or towards safe bonds and share in the ford motor company, these can’t be reproduced in the short run even if ford cars can be (that’s the point of the “musical chairs” analogy). Meeting the demand for money doesn’t provide all of the solution here to a Keynesian.
What the Krugman is saying here is “look we have two arguments, not just wage stickiness”. They do, but what he misses here is that the second argument doesn’t work without the first. If prices are perfectly flexible then they can adjust to any set of prices of assets, and produce full-employment. Notice this the same can’t be said for real output. If all investors decide that non-reproducible input resources are much more valuable in the long-term than they have previously thought then they won’t allow as much to be used for output now.
anon,
All the rates of short-run interest are visible to the investor. But, in ABCT that’s only a small part of the problem. The larger part is that an unexpected monetary injection causes business profits to rise. Businesses can’t tell if extra profits are caused by their own entrepreneurship or by monetary influences. It’s a weak complaint against ABCT.
LikeLike
>the payment is already fixed so that no change in the allocation of resources depends on the fixed mortgage payments
Eh? I thought the whole point is that if prices, wages, and loan payments were indexed to inflation, inflation would have no ill effects. When some prices adjust sooner than others, resource allocation gets screwed up. Precisely because most loan payments are fixed, they are sufficient to mess up resource allocation.
LikeLike
Current, Well, as I said, “price stickiness” covers more than one phenomenon. The tendency for output to fall in response to a less than perfectly anticipated decline in aggregate demand would operate even if there were no legal or contractual barriers to price adjustment to clear markets continuously. And if there is nothing to prevent markets from clearing continuously, it is not clear what makes prices sticky.
Your point about Say’s Law is interesting, but I don’t see why a shift demand toward goods in fixed supply would do anything other than drive up the prices of those specific goods. Why would that result in a cumulative decline in output? In the case of an increase in the demand for money, the only reason that a decline in output results is because the only way that increase in demand can be accommodated is for the prices of all other goods to fall. And even an increase in the demand for money has that result only if it involves an increase in the demand for a stock of money in fixed supply. Obviously competitive banks would be perfectly capable of responding to an increase in the demand to hold money by correspondingly increasing the amount of money they make available. You are right that in ABCT the increase in business profits resulting from monetary expansion may affect incentives to invest as well as the reduction in the interest rate. But I think that any fair reading of the Austrian literature would show that it is the rate of interest not the windfall from increased prices that is the key variable in the Austrian story.
Carl, I agree that in principle, at least, full indexation might or might not eliminate all adverse effects from inflation. But the indexation of interest payments already agreed to is supposed to avoid distribution effects. If I sell you my car on credit tomorrow on credit at a fixed interest rate, and then the price level doubles the week after, the inflation has no effect on the allocation of the car, it simply transfer some of my wealth to you.
LikeLike
I think they simply had a very different explanatory strategy.
The IE construct plays a different explanatory role for Hayek than Temp Equil. plays for Hicks.
“Hayek takes issue with Hicks’s formulation of the concept of temporary equilibrium even though it is a direct application of Hayek’s concept of intertemporal equilibrium. I never understood why Hayek disliked Hicksian temporary equilibrium.”
Too bad there isn’t a good book explaining how Hayek at the LSE & in Britain completely changed economics in all sorts of surprising ways.
LikeLike
Greg, You may be right about the different explanatory role, but that still doesn’t help me understand the nature of Hayek’s objection to temporary equilibrium.
LikeLike
“Well, as I said, “price stickiness” covers more than one phenomenon. The tendency for output to fall in response to a less than perfectly anticipated decline in aggregate demand would operate even if there were no legal or contractual barriers to price adjustment to clear markets continuously. And if there is nothing to prevent markets from clearing continuously, it is not clear what makes prices sticky.”
I agree with what you say here, but I’m not really sure what you meaning by “price stickiness”. Legal and contractural barriers can prevent price adjustment. So though can simple business issues. I’m sure you know all these: the cost of changing prices, that information must be gathered and expectations formulated in order to do it and the chains of prices that must change. I’ve always considered the later to be part of the price stickiness argument in general. Are you thinking about those things as something else?
“Your point about Say’s Law is interesting, but I don’t see why a shift demand toward goods in fixed supply would do anything other than drive up the prices of those specific goods. Why would that result in a cumulative decline in output?”
I didn’t mean a cumulative decline in output. The question here is: does the profit from turning non-reproducible goods into reproducible goods remain the same?
Suppose Robinson Crusoe produces roast pork by killing wild pigs and eating them. If Crusoe believes that the wild pigs are becoming rarer he may eat less of them, reducing his output even though his capability of producing output hasn’t fallen.
(Of course Crusoe could start farming pigs, for this a problem to occur there must be no way of using investment to solve it problem. We must assume that we’ve reached what could be called maximum roundaboutness for available technology).
In even rotation the problem here is an interest rate fall that can’t be accomodated by an increase in capital intensity/roundaboutness.
Please note, I don’t think this is actually a major problem in practice. The impression I get is that Keynesians think it is though. Take Krugman’s “debt overhang” problem in “Keynes and the Moderns”. If debtors must transfer more wealth to creditors this isn’t an issue if both are in a similar position. I think what he’s worried about is that the creditors won’t spend what they receive on components of output.
“And even an increase in the demand for money has that result only if it involves an increase in the demand for a stock of money in fixed supply.”
I agree with you there and I think that monetary equilibrium is the real problem in business cycles.
“But I think that any fair reading of the Austrian literature would show that it is the rate of interest not the windfall from increased prices that is the key variable in the Austrian story.”
Well, it’s one interpretation. Both Hayek and Mises though described a process where an injection of money supply beyond the demand for it causes profits to rise across industries. Hayek talks about this a lot in “Profits, Interest and Investment”.
LikeLike
Current, My point was exactly that there are a number of different sources of price stickiness and that we should try harder to specify which source of price stickiness we have in mind when we make some statement to the effect that price stickiness is responsible for some property of a macromodel. The source of price stickiness that I was focusing on in my posting was incorrect (overly optimistic) expectations of future prices. That source of stickiness allows nominal changes to produce real effects even though there is no exogenous restriction on and no cost of price change.
Sorry, but I am not following your argument on Say’s Law, perhaps you can clarify (or simplify) it for me.
You know your Hayek. It’s been a while since I looked at Profits, Interest and Investment, and now I do recall that he tried to restate his theory in a different way. In his later discussions of inflation, he liked to emphasize that one of the problems caused by inflation was that some businesses that were unprofitable and would have shut down are kept in business by inflation thereby preventing the reallocation of resources to higher valued uses. But I think that it is not the mechanism that is usually highlighted in expositions of the Austrian business cycle theory.
LikeLike
“Current, My point was exactly that there are a number of different sources of price stickiness and that we should try harder to specify which source of price stickiness …. That source of stickiness allows nominal changes to produce real effects even though there is no exogenous restriction on and no cost of price change.”
I see now, I agree with that. I’ve argued the same thing with the Austrian supporters of 100% reserve banking who insist that there would be no price stickiness without government interference.
“Sorry, but I am not following your argument on Say’s Law, perhaps you can clarify (or simplify) it for me. ”
This is difficult because I’m trying to explain something I don’t quite believe myself. As far as I can see Keynesians are saying that sometimes circumstances arise where it isn’t profitable for businesses to take input and make it into output. This is what discussions about a low rate of “marginal efficiency of capital” are about. That was the point of my simple example with Robinson Crusoe. He changes his mind about the value of the inputs he has at hand, and because of that he produces less output. Keynesians are worried about things like this. They think that even if we have monetary equilbrium there may be the situation where expectations of the future are so bad that expected future rates of return remain low, which means that rather than spend income on new investments it will be spent bidding up the price of existing investments. I’m not really sure that this is a practical issue (or that anything can be done about it if it is).
My point earlier was that if prices were perfectly flexible in the sense that monetary disequilibrium can’t arise then this problem theoretically still remains.
“But I think that it is not the mechanism that is usually highlighted in expositions of the Austrian business cycle theory.”
It’s in both Hayek and Mises. But you’re right that explanations of ABCT often fail to emphasis it.
LikeLike
> If I sell you my car on credit tomorrow on credit at a fixed interest rate, and then the price level doubles the week after, the inflation has no effect on the allocation of the car, it simply transfer some of my wealth to you.
Simply? Such an unplanned additional transfer has major effects on the allocation of resources – you can buy less and I more, but for no good reason in the trades and choices we made.
In real life, I bought a house last month. The interest rate isn’t indexed to anything, and if at any time over the next 30 years inflation strays much from the 2% both I and my bank assumed, I guarantee it will effect the behavior of both me and my bank.
LikeLike
Current, OK I think that we have finally reached a meeting of the minds. On Say’s Law, you could also get a similar result if people decides that they want to work less. Output would go down, but the result could not be attributed to any failure of the price mechanism to coordinate the plans of households and firms. I think the same would be true in the case that you posit. Total output and income might go down, but that would simply reflect a change (an unfortunate one to be sure) in the underlying initial conditions.
Carl, The distinction that I am making is, I think, a fairly standard one in economics, namely between the resource allocation effects of a change and the distributional effects. If the preferences of the parties between whom wealth is being redistributed have identical preferences, then the final equilibrium would be identical to the initial equilibrium except for the shift in consumption between the parties. That is a pure redistribution effect. Since, in general, one would not expect the parties to the redistribution to have identical preferences, there would be a change in the final equilibrium but it would only differ from the original equilibrium to the extent of the difference in preferences. And if the initial equilibrium was pareto-optimal, so would the final equilibrium. The case we started with was a case in which stickiness in pricing causes a failure to attain any equilibrium. In the case we are now dealing with, we go straight to a new equilibrium, it’s just different from the first owing to a shift in wealth distribution. Sorry for this giving you such a long-winded response.
LikeLike
That’s right, that’s what I mean.
LikeLike
@David
> The distinction that I am making is, I think, a fairly standard one in economics, namely between the resource allocation effects of a change and the distributional effects. If the preferences of the parties between whom wealth is being redistributed have identical preferences, then the final equilibrium would be identical to the initial equilibrium except for the shift in consumption between the parties. That is a pure redistribution effect.[snip]
I must be confused (entirely possible – I’m a computer programmer by trade). The distinction seems valid, and perhaps the assumption that preferences of the two parties are identical is too. But it seems to me there is a factor of 10 (the money multiplier) operating on this wealth transfer — ten units of wealth are lost to lenders for every one gained by savers. It seems to me that might produce substantial real effects without any need to invoke things like menu costs (which seem to me rather contrived.. and must have plummeted over the last ten years, at the very least).
LikeLike