Auction Markets and Optimally Sticky Prices

by Joseph T. Salerno

Keynesian macroeconomists, old and new, have long criticized their classical and contemporary opponents for ignoring reality and treating the market economy as a giant auction in which prices are “perfectly flexible,” responding instantly to changes in supply and demand.  This charge is wrong on two counts.  First, all markets for outputs and inputs function precisely like auctions; and, second, auctions are not characterized by perfectly flexible prices but by an optimal degree of stickiness in prices that is determined by the market itself.

In this post I will deal with second point, because it has been generally neglected in responding to the Keynesians.  To illustrate this point I will use the example of a one day on-site auction of 49 unsold condominiums at a 73-unit complex that recently took place in Auburn, Alabama and which I also happened to attend. (This auction is described in detail by Doug French.)  The condo complex was designed as a “game-day center” to cater to out-of-town Auburn alumni and other fans who crowd the town and double its population during the seven or eight autumn weekends when the Auburn University football team plays its home games.   Several of these complexes were started near the peak of the housing bubble and, like the one in question, were not ready for sale until after the bubble began to rapidly deflate in 2007.

As French describes, 12 of the 49 suites on the auction block were to be sold “absolute,” that is, at the price offered by the highest bidder, regardless of its height.  The remaining 37 units were to be auctioned with a predetermined “reserve” price set by the seller.  In the case of the latter, if the highest bid price for a particular unit did not exceed the reserve price, then the condo would be withdrawn from sale.

The condo units to be sold absolute were auctioned off first.  The two deluxe two-bedroom suites originally advertised at $389,000 were sold for $260,000 and $245,000 respectively.  The four two-bedroom models, initially priced at $237,000, sold for $125,000 to $135,000, while the four deluxe one-bedroom suites all garnered high bids of $105,000, a deep discount from their original $279,000 asking price.  Smaller one bedroom units sold at even steeper discounts, with the $210,000 and $188,000 units fetching only $70,000 and the $63,000 respectively at the auction.

When all condos for absolute sale had been sold, the auctioneer initiated the bidding for condos the developer had designated for sale with reserve.  The first unit was a two-bedroom deluxe condo and the bidding stopped at a price of $200,000, well below the lowest price ($245,000) that  a similar unit had sold for in the absolute segment of the auction.  What happened next was very odd:  the auctioneer refused the bid price and abruptly declared the auction closed so as not to further “embarrass the property.”  All remaining 37 units were withdrawn from sale. In an instant, prices had become absolutely rigid, leaving 37 condo units unsold.  The reason for this development is clear.  The owner was testing the market with the preliminary auction of the units for absolute sale, while possibly also raising needed cash to satisfy the minimum repayment demands of the banks that had extended him construction loans.  To attain these purposes prices needed to be, and indeed were, perfectly flexible. However, after experiencing the state of the market with the initial sales and the highest bid received and refused on the first unit for sale with reserve, the auction company quickly ascertained that the developer’s reserve prices for the respective models would not be met and the auction was halted.  These reserve prices were established because the developer of the complex was speculating that higher prices could be realized by withholding the remaining units from current sale and offering them in a future market.  This aim was served by absolutely rigid prices.  (Actually since the auction reserve prices were not announced, it may have been the case that the prices of the various models were not perfectly rigid at a given reserve price, but that for each of the several models there was a scale of increasing minimum prices at which progressively more units would have been sold until the stock was exhausted.)

In any case, our first conclusion is that auction prices are as sticky or as flexible as they need to be to accomplish the goals of the sellers.  That is, in technical jargon, the supply curve may be: vertical (perfectly flexible prices); upward-sloping to the right beginning at a reservation price that exceeds zero (semi-flexible prices); or L-shaped with a horizontal segment up to the total stock available at which point the curve becomes vertical or upward sloping (absolutely rigid or sticky at the reserve price).  The point to be emphasized is that the subjective purposes, expectations and plans of sellers directly determine the degree of flexibility of prices and the corresponding shapes of the supply curves.  (Ultimately, of course, it is the present and anticipated demands of buyers which determine the sellers’ plans and thus condition the shape of the supply curve)

The second lesson to be derived from this auction market example is that the market clears and all gains from exchange are exhausted whether prices are perfectly flexible or absolutely rigid.  The 37 condo units of the 49 units offered for sale that remained unsold at the auction were not unwanted surplus from the point of view of the developer, who actively pulled them off the auction block.  At the end of the day all 49 units, whether sold or unsold were allocated by the market to their most “most capable possessors,” to use Böhm-Bawerk’s felicitous term.   Thus sticky prices on an unhampered do not reveal market failure, unwanted surpluses and frustrated buyers and sellers—far from it, they reveal the sensitive adaptation of the pricing process to entrepreneurial discovery and planning.

In a subsequent post I shall argue that nominal price rigidities based on objective factors like “menu costs” are pure illusion.  Like all “selling costs” the costs associated with changing prices associated with different pricing technologies previously chosen by entrepreneurs are sunk costs and do not affect pricing decisions on current stock.

25 thoughts on “Auction Markets and Optimally Sticky Prices

  1. Joe, I can’t see what this has to do with the sticky prices argument. Of course, Keynesians realize that people are deciding not to take a lower price. But the point, it seems to me, is that Keynesians think they are mistaken in doing so. If, a year from now, these condo units upon which bids of $200,000 are refused sell for $100,000, then the Keynesian will say, “See, the price was unduly sticky — the condo was kept out of use for a year, only to sell at an even lower price than was earlier refused.”

  2. I actually studied the auction process when I was working for a developer. We were upside down on a condo project and the lender was putting pressure on us. We talked to the three biggest auctioneers in the country. The whole process is gamed by the auctioneers who understand what humans do when they are bottom fishing. The absolute sales are the teasers to get people in the door. It’s an interesting process. What the developer is looking at is to recover enough money for the construction lender; anything left over goes to the investors or mezz guys. If he can’t hit at least the lender’s per unit release prices for his loan, then he is facing a deficiency judgment for the difference. The only reason that the developer would take them off the block is if the lender is willing to work with him. Otherwise, the lender wants his money.

    So while this doesn’t have much to do with the stickiness factor, it shows that the market isn’t sticky. Now, the interesting thing is what the lenders have at risk. If they get cheap bailout money, they are less flexible with the borrower-developer because they can roll the dice a bit with the extra backing and push the developer to foreclosure. The Feds are breathing down their necks. If they took no bailout money, perhaps they would be more willing to work with the developer.

    I am seeing this happen now. Our big local bank that took the bailout money is pressured to liquidate its loans. The little local bank that didn’t take the bailout money is working with its borrowers.

  3. The whole business of sticky price and wages is very odd.

    According to the normal definitions a sticky wage is a *wage*. That is it is a fixed price for an hour of labour. The number of hours concerned must be variable. So we have a ratio money/hour with one part fixed.

    If however a worker is paid a salary they do not have a “sticky wage”. A salary defines both the number of hours worked and the payment received. Both parts of the money/hour ratio are fixed, so the compensation is done per month.

    This leads to a very bizarre conclusion. If other prices are not sticky we have the following situation. If we all had flexible wages then according to New Keynesian theory there would not be recessions. Also, if we were all salary-men then according to New Keynesian theory there would also be no recessions.

    Scott Sumner first pointed this out to me.

  4. Gene,

    New Keynesians believe that sticky prices result from nominal rigidities” caused by menu costs, long term contracts,monopolistic competition, etc. These are objective factors that prevent the sellers from reducing prices in the short run. As a result, Keynesisan argue, prices get “stuck” at a level above the equilibrium level resulting in quantity adjustments instead, i.e., surpluses and unwanted build up of inventories. This is why leading intermediate macroeconomics textbooks, such as Mankiw and Abel, Bernanke and Croushore depict the Aggregare Supply Curve as perfectly horizontal in the short run, so that a change in AD demand elicits only a quantity response with no change in prices. A constant price level, of course, is also the foundational assumption of the IS-LM model that is used in all macro texts to explain the Great Depression.

    So, Gene, I believe you are quite wrong in claiming “Of course, Keynesians realize that people are deciding not to take a lower price. But the point, it seems to me, is that Keynesians think they are mistaken in doing so.” Keynesians are not referring to individuals who are speculating incorrectly in setting reservation prices, but rather believe that external non-subjective factors such as menu costs, market structure and so on make it unprofitable to immediately lower prices to levels that will clear markets and return the economy to general equilibrium.

    I suggest a quick look at Mankiw’s Macroeconomics, if you are interested in learning what Keynesians think on these matters.


  5. Excellent post, Joe.

    Current, it seems like you have several kind of farcical assumptions (wage/hour ratio remaining fixed for salaries; actually there would just be a maximum ratio likely). In any case, to a New Keynesian, I think the main point is that although recessions would occur, they would be smaller with flexible wages.

  6. Interesting post.

    One possible quibble would be with the statement that “it is the present and anticipated demand of buyers which determine the sellers’ plans and conditions the shape of the supply curve”. If the supply curve is a function of the demand curve, we have only one curve, and no supply-and-demand market. The other interpretation is that all suppliers have some monopoly power (in which case there is no supply curve).

  7. Pierre,

    You are quite right, and I apologize for the ambiguity. In saying that the anticipated demand curve “conditions the shape of the supply curve,” I did not mean “functionally dertermines the supply curve.” At any point in time the location of the supply curve in P-Q space is determined by the past production decisions of entrepreneurs. Absent any speculative components, the supply curve for any product will be perfectly vertical at the quantity of stock currently in existence. In Austrian-Wicksteedian causal-realist price theory, an upward slope of the supply curve or a reverse L-shape, as the case may be, reflects entrepreneurial expectations of the future development of the product price. In this sense alone, as mediated through entrepreneurial price appraisements, are anticipated demand curves an influence on the shape of the current supply curve.


  8. Admiral: “it seems like you have several kind of farcical assumptions (wage/hour ratio remaining fixed for salaries; actually there would just be a maximum ratio likely).”

    Yes, a salary is the sort of maximum ratio you mention.

    My point though is that at least some New Keynesian analysis works by these sort of strange assumptions. Salaried workers are excluded from the analysis. Except in that they may represent a sticky price. Since salary renegotiations occur and people change jobs, and become unemployed, etc.

    Admiral :”In any case, to a New Keynesian, I think the main point is that although recessions would occur, they would be smaller with flexible wages.”

    Yes. That was why I mentioned that other prices would have to be sufficiently flexible. The other criteria that could cause a recession in New Keynesian theory would be an external shock.

  9. I thought “the money illusion” (the concept, not Sumner’s blog!) was all about workers being mistaken regarding real vs nominal wages.

  10. I don’t get it. What’s the point — that the price of apartments rise when they are expected to appreciate in value? Because that’s all that happened, it seems to me. There were a awful lot of words though.

  11. Lee Kelly is right. What is being described here is an expectations failure. I don’t see how this is a refutation of of any New Keynesian proposition. Actually, more recent New Keynesian research focuses on how incomplete information can result in price stickiness. The agenda of New Keynesian is to show how rational actions by individuals can lead to macroeconomic distortions. These actions might be optimal for the individuals themselves but far from optimal for the economy as a whole. This post seems to assume that the price stickiness of the auction has no external effects.

  12. TGGP, the “Money Illusion” isn’t quite the same thing as wage stickiness or price stickiness.

  13. Current, correct me if I’m wrong, but doesn’t “the money illusion” refer to how workers react differently to real vs nominal paycuts? If that’s not the case, what is it?

  14. I applaud Joe Salerno for his post. I have in the past argued we should drop the term “sticky prices,” because it presumes there are “non-sticky prices.” But the latter exist only in economists’ models, not in actual markets. Also, modern macroeconomists are “Keynesian” only in an ironic sense. As Leijonhufvud demonstrated 40 years ago, sticky wages were not central to Keynes in the GT.

  15. “I have in the past argued we should drop the term ‘sticky prices,’ because it presumes there are ‘non-sticky prices.’ But the latter exist only in economists’ models, not in actual markets.”

    There is nothing “optimally sticky” about the apartment price. The seller is just speculating that he can get a higher price in the future and witholding sale. A price which is able to respone to seller expectations is not “sticky”, even if it may *seem* that way to some presumptuous Keynesian.

    The problem with Keynesians is that they seem to think that exchange is a good thing in and of itself. Even though the seller evidently valued the apartment more than the potential buyers, willing to forgo $200,000 for it, this cost is not registered by accountants as “spending”. Since an transaction that would have contributed to “GDP” has not occurred, then Keynesians consider the high price to be a problem.

  16. One point should perhaps be clarified here. The sale of an asset does not contribute to GDP, any more than any other sale. GDP measures production, not exchange nor asset values.

  17. Pierre,

    GDP is equal to “total spending on all final goods and services.” Home purchases are counted as “investment,” and so I assume that an apartment purchase in Auburn would count. It seems to me that GDS (or gross domestic spending) would be a more accurate initialism.

    In any case, a mistaken exchange (i.e. when buyer and seller both receive something valued less than what was given up) will presumably register on GDP, even though both parties would have been better off if the exchange had never taken place.

    It seems, then, if we can somehow trick people into making mistaken exchanges, then we can increase GDS — increase the velocity of money. And is this not exactly inflation via monetary expansion actually does? People accept money in exchange for goods and services which they later discover has less purchasing power than expected. There is still a cost involved when an exchange doesn’t occur (i.e. what could have been had in exchange), but this “expenditure” is not detected by accountants and so is missing from GDS.

    When economic performance is gauged by the ups and downs of GDS (with the ups being good), a bias then emerges toward seeing all non-exchanges as being a problem — a failure of the market.

  18. Lee,

    I don’t want to defend the usefulness of GDP. I have often attacked the concept when it is used to measure welfare — not to mention the statistical problems of measuring it. And you are right to raise the problem of what happens to the normative significance of GDP when relative prices are distorted.

    My only point is that GDP should be understood for what it is. It is an estimate of total production using as weights the prices that represent the marginal value of each goods or service. This total production is equal to total incomes or, alternatively, to total spending on all final goods and services, as you rightfully say. If an appartment (condominium) is produced this year and is sold $100,000 this year, it is counted as $100,000 in this year’s GDP. If it is not sold, it is counted as an increase of $100,000 in inventories (goods to be sold later). [It may actually be considered capital, I would have to check, but at any rate, inventories are just short-term capital.]

    Now, assume that this appartment is later sold again in three years’ time for $120,000 (without having been renovated in any way). This will not register at all in GDP, as far as this appartment is concerned. What will register in this t+3 GDP (and in previous years’ GDP, if previous savings were involved) is the $120,000 in goods and services the buyer has produced and transferred to the former appartment owner.

    All these values are, of course, based on the assumption that every individual knows better what is good for him, so if somebody purchases $120,000 of gadgets (the former owner of the appartment, for example), they are worth for him at least that amount. If somebody doesn’t purchased the gadgets, we only know they are not worth at least $120,000 to him. In an analogous but distinguishable way, the only things that get produced and enter GDP (one time) are those which require resources that are not worth more in producing something else for somebody else.




    I should perhaps have mentioned that GDP is the production of final goods and services FOR THE PURPOSE OF EXCHANGE ON MARKETS (GENERALLY FOR MONEY). The meals cooked by the housewife are not counted in GDP. But this is a rather well-kown characteristic and weakness of the GDP concept.

    Note, however, that this does not mean that if the family now decides to eat out, GDP increases. For the purchase of the restaurant meals will have to be made in exchange for a part of what the family produces and adds to GDP. Except if the bread earner decides to produce more in order to be able to take his family to the restaurant, he will have to reduce his consumption of something else. In a very real sense, “real” GDP is not changed; only the constellation of goods produced is.

    On the other hand, if the housewife starts working for a restaurant instead of cooking her family’s meals, then of course GDP is increased.

  20. Lee Kelly makes my point while missing it. Theer is no infinite price flexibility (whatever that would be) in markets. Additionally, I did not use the term “optimally” and would never do so. It is again imposing criteria of models on markets, instead of using models to understand markets.

  21. Pierre,

    You are absolutely right. I had forgotten entirely when writing those previous posts. Thank you correcting me, especially since I now remember about half a dozen other objections I have to GDP (as a measure of economic performance).


    If I make your point while missing it, then I think we must agree after all. By the way, the word “optimal” was used by Joe in the original post; I didn’t mean to put words in your mouth.

  22. TGGP: “correct me if I’m wrong, but doesn’t “the money illusion” refer to how workers react differently to real vs nominal paycuts? If that’s not the case, what is it?”

    The idea of “money illusion” is that agents plan in terms of money. They don’t revise their plans fully or accurately when a change in the value of money occurs. Wages are a special case of this.

    Sticky prices and wages are something slightly different though. Money illusion is sometimes suggested as a reason for sticky prices, or prices being more sticky than usual. Menu costs are another reason given. However, there is no direct link between money illusion and sticky prices. As others have said in this thread, there are in reality no non-sticky prices.

    The reason for the stickiness, as Joe mentions isn’t money illusion.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s