Prices and Information

by Jerry O’Driscoll  

In the recent discussion of Say’s Law, the issue of “sticky” prices came up. The term is the source of much confusion. The opposite of “sticky” prices is not “flexible” prices, but infinitely flexible prices. No matter how flexible a price, short of infinite flexibility, there will be quantity responses. Quantity responses are an inherent part of market economies, do not signal market imperfections, and do not typically trigger income-constrained processes.  

Price setting incurs information costs. How does a producer or  retailer know when to change his prices? Where does that information  come from?  

Normally, price changes are triggered when inventories either  accumulate or decumulate at abnormal rates. If, for instance, a  single retailer faces an increase in demand for his products (say  dresses), he will be able to purchase more at current prices from his  supplier’s inventory. If many retailers face an increase in demand,  then the effects are pushed back to the production stage. Prices will  start rising for all the inputs in the production of dresses and pass  through all the production and distribution stages to retailers, who  then raise his prices on the dresses beings sold to his customers.  Each retailer will tell his customer that he is raising prices because  his costs have risen. The economist understands, however, that the  higher prices reflect higher final demand.  

First quantity changes, then prices change.  

Infinitely flexible prices in goods markets are difficult even to  comprehend. Among other things, there would need to be instantaneous  production. Less than infinitely flexible prices and inventories  buffer against demand changes and avoid costly changes in production  rates.  

In many ways, the current recession is a classic inventory recession (build-up of goods to be worked off). Except the inventory build-up was of residences. Once the recession spread to other goods markets, response was very rapid. Indeed for about a year inventories of most goods were falling as demand fell. That suggests very rapid price responses — even anticipatory price cuts to clear inventory in the face of future drops in demand. That’s about as flexible as prices can get in the real world.  

The problem with “sticky” price models is they prove too much. We would constantly be in depression if they had universal application.  Demand is always shifting, triggering quantity responses and  income-constrained processes. Since the economy is not always in depression, something other than less-than-infinitely flexible prices  must be the culprit.

37 thoughts on “Prices and Information

  1. This is a good criticism of some microeconomic underpinnings of non-Austrian business cycle theories. (Would it be helpful if these theorists spent some time working in businesses?)
    But what about the growing service economy, where firms don’t have inventories?
    While we can point to the role of house inventories in the most recent cycle, what about the tech stock boom-and-bust of 1995 to 2002? Tech firms didn’t have inventories for the most part.
    How to explain price changes in service industries?
    While we can point to the unexpected accumulation/decumulation of inventories as a triggering mechanism for price changes for some firms, is there more to the story?

  2. Thanks. Good points and good questions. I’d say the inventory of service firms is chiefly its labor force (labor services). Labor services are perishable, so the analysis probably needs to be adjusted. Like seats on an airplane, there is an incentive to sell the inventory every day.

  3. I love this post, and I really hope to see more like it on this blog. Why don’t economists consider the relative rate of flow between different links in supply and demand? Seems pretty obvious that if the overall rate of flow in demand for a particular good (e.g., houses) were much slower than the rate of supply at a given moment then there is a bubble forming.

    I do have one (somewhat nitpicky) point though:

    Prices will start rising for all the inputs in the production of dresses and pass through all the production and distribution stages to retailers, who then raise his prices on the dresses beings sold to his customers. Each retailer will tell his customer that he is raising prices because his costs have risen. The economist understands, however, that the higher prices reflect higher final demand.

    In this story, either the retailers are colluding to raise prices in order to prevent an inventory shortage or demand is highly inelastic for each and every retailer of the good in question. Why? Because if there are relatively good substitutes or no collusion then somebody will undersell to gain market share until the market clears.

    Whether increased costs can be passed on to customers is of great concern to investors. CEOs love to brag about new technology and products but most never bother to explain how or why the costs of these new things will be passed on to the customer rather than borne by the investors.

    One thing that I’m somewhat obsessed with is the fact that whatever the answer we give to these questions (is demand elastic? how big a shortage?), that answer has time-dependence. The same answer probably won’t be valid at different times because of the way suppliers and customers react to price changes.

  4. I see that you were careful to say “tell the customer” rather than simply raise prices. But the same argument about elasticity applies to the suppliers passing costs onto the retailers too.

  5. Prices that are less than perfectly flexible in the context of an excess demand for money lead to recession.

    If there were a shortage of money, and prices were (implausibly) “infinitely” flexible, then the real quantity of money would adjust to the demand, and the real volume of expenditures would be unaffected.

    I don’t believe that sticky prices and wages cause recession whenever there is an increase in saving. While the market process by which saving results in expanded investment might lead to an excess demand for money, it doesn’t have to, and it is the excess demand for money that is the problem.

    Because we live in a world of sticky prices, an excess demand for money results in lower output and employment.

    By the way, falling inventories as demand falls don’t necessarily mean rapidly falling prices. Why can’t production drop faster than demand?

    Also, there are actually measures of new orders and final sales. If rapidly dropping prices are causing inventories to fall, it would be due to higher real purchases.

    I regularly get emails about all sorts of measures of the real volume of sales. It is all way lower than a year ago.

    If the problem is now a shift of resources from the porduction of houses to other capital goods, why aren’t prices, profits, production, and employment all expanding in those industries?

    Why isn’t the news–well, I have a bunch of applications from carpenters, but I need machinists. Where are all the machinists?

    The problem is that those who were funding the houses though money market funds investmd in commercial paper invested in mortgage backed securities are now holding T-bills and FDIC insured bank deposits. The interest rate on bank reserves is higher than the T-bill yield, and so, there is an excess demand for base money, despite the historically huge increases in the quantity of base money. Nominal expenditure is falling. And, if prices and wages fell enough, then real expenditure wouldn’t have fallen because of these financial changes. But prices and wages didn’t fall enough. Real expenditure did fall, and output and employment has fallen nearly across the board.

    The economy is not expanding in those sectors that had been starved for resources because of malinvestment into overproduced single family homes.

    Anyway, I think the Fed needs to stop paying interet on reserves and increase base money more than it has. Nominal income needs to return to its past growth path. Given that environment, and stable growth in nominal expenditure, the market can adjust the allocation of resources between single family homes, consumer goods in general, and other capital goods just fine.

  6. Lots of ideas and I will just respond to a couple. First, I think the term sticky prices should be abandoned. There are just prices. Infinitely flexible prices would be prohibitively costly. Armen Alchian did the most to drive that point home. Whatever the putative costs of less-than-infinitely flexible prices, they are less than the costs of infinitely flexible prices. Second, economists have understood for centuries that an increase in money demand MAY cause problems. What are the likely causes of increasing money demand? Under what system can changes in money demand be best accommodated? That is where the focus should be.

  7. “Infinitely flexible prices in goods markets are difficult even to comprehend. Among other things, there would need to be instantaneous production.”

    I think that prices that would be infinitely flexible are prices that would never change:

    If prices adjust instantly, then the future would be always correctly anticipated by the entrepreneurs. Hence, if the future is always know, then the present prices already reflect all changes that will happen in the future (like the change in demand for dresses in the example, we would have in the present all the factors correctly allocated to supply the future increase in demand).

    Well, extending this logic further, all transactions for all possible future commodities would be made in the present. Why await to buy something that you already know you will need in the future and whose price never changes? This is pretty much the arrow-debreu notion equilibrium. So I think that if we are not in a full equilibrium we doesn’t have perfectly flexible prices.

    Even with costs to change prices, with unchanging prices there is no cost. For example, if demand for money increases in the future, prices all already reflect the future increase in demand in the present, hence, no recession. In this case the menu costs explanation of inflexible prices is not very rigorous, even if it explain why supermarkets doesn’t change the price of tomatoes every day. But, in this case (full equilibrium) there is no real uncertainty, so money doesn’t exists anyway.

  8. The version of Says law that is correct describes a long-run trend.

    However, as Bill says recessions are short-run occurrences. I expect we all agree that there will be recovery in the long-term no matter what the money supply is. The important thing for business cycle theory though is that some amounts of money will produce more growth in the long-term than others. That is why Say’s law isn’t really very useful in these discussions. I used to think it was, but I’ve since changed my mind.

  9. I’m sorry my argument has been lost on Current. Say’s Law is relevant in the short run, too, because it correctly directs our attention to the microeconomic forces at work even in a recession. The effectiveness of discretionary monetary policy depends, at least in part, on expectations. I took that as the gravamen of Rafael Guthmann’s post. It is also the core truth of ratex. Mises called it Lincoln’s Law.

  10. “Say’s Law is relevant in the short run, too, because it correctly directs our attention to the microeconomic forces at work even in a recession.”

    Fair enough, I see your point.

  11. If we assume Say’s Law is correct, must we also assume that individual preferences are stable over time (or at least that they do not vary over the period associated with less-than-infinitely flexible price equilibriation)?

  12. By “stable over time” I didn’t mean constant in time, although I could have been clearer. I mean deterministic in the sense of Becker and Stigler.

    Suppose my preferences are, in part, a function of the preferences of those I am in communication with through market price signals. Now even if it is true that everybody behaves in accordance with their rational expectations, there might be sticky prices when there is less-than-infinitely-flexible price equilibration. The time-scale of preference formation may be faster than the time-scale of equilibriation.

    Say’s Law is true so long as the change in preferences is not so rapid that price cannot reach equilibrium before another shift in preferences.

  13. Okay. I wish I could agree. I simply don’t have enough info about preferences to know whether I can or cannot. It seems possible to me at least that what look like deterministic preferences are distributions that are not stationary in time. I haven’t see any measurements of preferences to disprove that possibility. If so, then I think Say’s Law cannot always be valid.

  14. Imagine a rope. Each strand of the rope is only short, but the overall rope is long. Each strand represents a preference, the distance along the rope is time.

    In order for the problem Michael F. Martin mentions to occur there would have to be a sudden change of very many preferences. This is not likely unless we all become Zombies and hence prefer brains to anything else.

  15. 😀 No. That sudden and massive changes in preferences are possible.

    If each individual’s preferences are coupled to other individuals’ preferences through market price signals, and market price signals are convex (i.e., increase slower the higher they rise), then the preferences of everybody in the market can spontaneously and rapidly synchronize. That’s the Kuramoto Model.

    Zombies aside, I think there is some empirical evidence for this. I think Keynes had something like this in mind when he talked about how markets tend to move ith “animal spirits.”

  16. None of this has anything to do with whether preferences are stable or deterministic. You are trying to understand a theory (Say’s Law) in terms of an incompatible theory. One may be correct and the other incorrect, but that cannot be determined by inspecting one in light of the other. More fundamentally, there is a clash here on what one does in economics: inspect the mathematical properties of abstract models, or understand human behavior. And that is all that I have to say on this issue.

  17. Michael F. Martin: “If each individual’s preferences are coupled to other individuals’ preferences through market price signals”

    I don’t think you understand what economists mean by “Preferences”. Suppose that person A wants X because they can sell it to person B. That is not a preference on the part of person A.

    Michael F. Martin: “and market price signals are convex (i.e., increase slower the higher they rise), then the preferences of everybody in the market can spontaneously and rapidly synchronize. That’s the Kuramoto Model.”

    I see from google that the Kuramoto model has something to do with synchronizing oscillators. Why do you think this has anything to do with this topic? Your explanation above doesn’t explain any sort of clear connection.

    Michael F. Martin: “Zombies aside, I think there is some empirical evidence for this.”

    Really? What is it then?

    I think it’s fairly silly. Since the current recession began I haven’t changed my preferences for anything. I’ve seen a few of my friends take up knitting and home-brewing, that’s about it.

    Michael F. Martin: “I think Keynes had something like this in mind when he talked about how markets tend to move ith ‘animal spirits.'”

    Keynes was talking about investors.

  18. I do understand what economists mean by preferences: an ordinal ranking of goods that is the argument to an individual’s utility function. But you’re right that I don’t have exactly the same thing in mind. When economists talk about preferences, they’re talking about revealed preferences, wherein the revelation is through transactions. So long as the principle of revealed preferences is the only way to measure preferences, economists have to be agnostic about what happens between transactions. What I have in mind is a model of preferences based on observation of the frequency distribution of consumption or production for a group. If the frequency distributions are stationary in time, then you get the same predictions you would from the principle of revealed preferences.

    Also, I shouldn’t have said “market prices are convex.” What I meant is “utility functions are convex” — a consequence of the principle of diminishing marginal utility.

    Here’s how Kuramoto applies. Assume each market participant has an average period between acts of buying or selling so that the market as a whole has some frequency distribution. Now assume that these periods between buys and sells are dependent in part on the absolute price. Some people buy often even at high price. Others buy never at high price, etc. Finally, assume that everybody is watching the price at all times so that buying and selling decisions are all-to-all coupled through the market price signal. Kuramoto model says under these conditions buying and selling may become synced.


    and papers cited therein

    There’s no difference between buying and selling securities and buying and selling apples as far as this theory is concerned. (Except for measured differences in the average period between buying and selling.)

  19. But forget about zombies and the Kuramoto Model, which are only interesting because they don’t necessarily require conscious agreement. This defense of Say’s Law is unconvincing for a much more obvious reason:

    Quantity responses are an inherent part of market economies, do not signal market imperfections, and do not typically trigger income-constrained processes.

    Oh, yeah? SO it’s not a market imperfection when the buyers or sellers collude to raise prices against potential counter-parties? Anybody with experience in actual markets knows that people collude to raise or lower prices, and that Say’s Law to the extent that it predicts that competition will eventually break up the conspiracy, is irrelevant as a practical matter.

  20. I don’t understand the motivation for Michael Martin’s gratuitous attack on something I wrote earlier in an entirely different context. Say’s Law has nothing to do with the issue of collusion. An “income-constrained process” is a term of art familar to macroeconomists. In my experience, “Anybody…knows” is a construction employed by people without an argument or evidence. It is beneath contempt.

  21. Bill Stepp: “Tech firms didn’t have inventories for the most part.”

    I don’t think this is quite right: hiring numerous Java programmers certainly builds up an “inventory” of a sort.

  22. I’ll concede that I did not understand “income-constrained process” to preclude collusion. But with that big a caveat, my skepticism about Say’s Law does not decrease! I am sorry for coming off flip.

  23. M. Martin wrote: “I can certainly agree that the choice among two inconsistent theories should be based on data, not aesthetics.”

    Yes, so in 1590, when the astronomical data were on the side of Ptolemy/Brahe, but the aesthetics were on the side of Copernicus, you would have wisely sides with Ptolemy/Brahe, hey?

  24. “I appreciate the humor, Gene.”

    But I’m making a quite serious point. Kepler and Galileo went with Copernicus based on aesthetics, not data. Were they wrong to do so? And similar incidents have been frequent in the history of science. Einstein had won over Newton based on aesthetics well before the famous “eclipse test” involving the perihelion of Mercury, and much evidence contradicting Einstein was (wisely, I say) ignored during the teens and twenties.

  25. “To say that the choice between two inconsistent theories should be made on the basis of data is not to deny that aesthetics should play a role in theory.”

    OK, but you did very explicitly deny that the choice should ever be made on aesthetics over data. Yet that is just what Galileo and Kepler did. (Yes, the two of them provided good data for heliocentrism, but that was only after and because they had already long before committed to heliocentrism based on aesthetics, and then went looking for the data to prove it correct.)

    So, I reiterate: Do you think they were wrong to do so?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s