David Andolfatto has written a very good post criticising sticky-price macroeconomics. This post really needed to be written. It explains why a minority of macroeconomists don't like the mainstream assumption of sticky prices. After you have read it you should understand why those crazy freshwater macroeconomists aren't quite so crazy after all.
I agree with so much that he says. But in the end, I take a different path. I stick with my assumption of sticky prices, even though I hate it. Here's why I hate it; but why I stick with it nevertheless.
(I'm going to start out using "price-stickiness" and "wage-stickiness" more or less interchangeably. Wage is just a price, the price of labour, and both wages and prices seem to be sticky. The distinction matters later in the post.)
I dislike the assumption of sticky prices for much the same reasons that David does.
One of the jobs that economists are supposed to be doing, and have been doing for the last couple of centuries, is to explain prices. To assume prices are fixed is not to explain them; it's to refuse to explain them. We aren't doing our job. It is only a little better to assume prices are "sticky", which means imperfectly flexible, so they adjust only slowly from one equilibrium to another. We don't explain why they are sticky; we just assume it. For example, the Calvo model of the Phillips curve, that underlies most New Keynesian macroeconomic models, simply assumes that firms face a fixed probability d per period of being "allowed" to change price. Why isn't d=100%? Will d change if monetary policy changes? Who knows.
It get's worse.
What's so wrong with just assuming sticky prices, even if it would be nice if we could explain them? Even if it's our job to explain prices? Economists make lots of assumptions. We (usually) assume fixed preferences and technology too.
What's so wrong about assuming sticky prices is that it's possibly internally inconsistent with our assumption about how quantities are determined. In the micro part of ECON 1000 we adopt the standard Marshallian "scissors" of demand and supply curves, where buyers and sellers try to buy or sell however much or little they want, at the going price. If something like rent controls or minimum wage laws prevents the price adjusting to the market equilibrium, we assume that the "short side" of the market determines the quantity exchanged: Q=min{Qd,Qs}. And if the government stays out of the market, price adjusts so that Qd=Qs, so that Q=min{Qd,Qs}=Qd=Qs. Everything is internally consistent in micro. The theory of price determination and quantity determination both come from the same model.
In macro, we assume that prices are inherently sticky, for some reason. We don't know what that reason is, but we recognise there must be something wrong with the Marshallian model of markets, and if we understood what it was, we would be able to explain sticky prices. And yet at the same time we continue to use the Marshallian model to explain quantity determination given sticky prices. Q=min{Qd,Qs}, but P=Pbar.
Robert Barro was the first economist to point out this internal contradiction between the then-prevalent theory of sticky wages (risk-averse workers buy insurance against wage fluctuations from risk-neutral firms) and the theory of employment determination given sticky wages. (It's paradoxical that his 1977 paper could be both so influential and yet seemingly forgotten today.) If firms and workers set wages in a long-term contract, they would want to set employment too in a long term contract. And if they did so, employment would be determined very differently than it would be in a Marshallian spot-auction market with an exogenously fixed wage. Q would not equal min{Qd,Qs}. Instead, employment would be set to equalise the value marginal product of labour with the value of the marginal disutility of labour. Employment would be set exactly as if wages were perfectly flexible (except for a minor wealth effect on labour supply due to the insurance). Sticky wages were a facade.
Now that's just one theory of sticky prices (or wages) that was shown to be inconsistent with the Marshallian theory of quantity determination given sticky prices. David points to a second.
Each worker and each job is in some ways unique. Getting the right worker in the right job is a non-trivial matching process, that takes resources. In equilibrium there must be quasi-rents to the relationship between a particular firm and worker sufficient on average to cover the cost of those resources needed to find the right match. Investment in firm-specific human capital while on the job can further increase those quasi-rents. Unlike the equilibrium in a Marshallian competitive spot auction market, where a tiny rise in an individual worker's wage, relative to the market wage, would make no employer want to hire him, and a tiny decrease in one firm's wages, relative to the market wage, would make no worker want to work there, those quasi-rents create a zone of indeterminacy within which an individual worker's or firm's wage can fluctuate without leading either to want to break the relationship. This zone of indeterminacy both allows for sticky wages, and yet, at the same time, is inconsistent with the Marshallian theory of quantity determination given exogenously sticky wages.
But in the search-theoretic or matching model, unlike the models that Barro was criticising, wage stickiness can still have an effect on employment, even though it will be different from the Marshallian model plus assumed sticky wages. If wages are stuck "too high" for example, existing firm and worker matches may continue. But unemployed workers will devote more resources to finding a match, and firms with a vacancy will devote fewer resources, because the match will be worth more to the worker and less to the firm. (I learned this from Roger Farmer). This inefficient mix of resources to finding a match will mean fewer new matches are made. It doesn't cause layoffs, but it does reduce new hires.
But yes, it's still very different from employment determination in a Marshallian spot auction market with an exogenously fixed wage.
So, if I hate the assumption so much, why am I still a sticky price macroeconomist?
First, because when I go to the supermarket, what I see and what I do seems to fit my macroeconomic model pretty well. I buy whatever I feel like buying on the day, at the price they set, and that price doesn't change very often. And when it does change, it usually changes for a temporary sale, then goes right back to exactly where it was before the sale (true, this does contradict the Calvo model, but it also means that prices actually change less frequently than reported, once we set aside these temporary sales).
Now it's a puzzle that quantity is nearly always demand-determined, and almost never supply-determined. Q=Qd seems to work almost as well as the more general Q=min{Qd,Qs}. But that puzzle is quickly resolved when we recognise that monopolistic competition is the rule, and perfect competition the exception. If prices were sticky up and down in a perfectly competitive market, we would see excess demand (quantities supply-determined) as often as excess supply (quantities demand-determined). Since monopolistically competitive firms will have price above Marginal Cost in equilibrium, the price would have to be stuck a long way below the equilibrium before any firm would want to ration customers. And the best thing about New Keynesian macroeconomics is that it replaced perfect with imperfect competition, so its assumption that quantity is demand-determined became more internally consistent.
I don't know why my supermarket has sticky prices. There must be some "friction", or "long term contract", or something that makes it depart from the Marshallian spot auction market. But whatever it is that causes sticky prices, it's something on the supply-side. My demand curve still slopes continuously down. If they raise prices a little, I buy a little less, and shop elsewhere a little more. And it's my demand curve that determines quantities, at the prices they set, not their supply-side. Whatever it is that's weird, it's something weird on their supply side; it only matters in determining the price. That price, plus my perfectly normal Marshallian demand curve, are a sufficient statistic for quantity determination.
Maybe supermarkets see themselves as having long-term obligations to their customers. But the customers don't see themselves as having long term obligations to their supermarkets. They buy whatever quantity they want, from whichever supermarket they want. At the prices the supermarkets set. It sure looks like quantity is determined by quantity demanded on a standard Marshallian demand curve. Just like standard sticky-price macro. And that "supermarket" model looks fairly typical for most newly-produced consumer goods. There may be rare exceptions. Maybe I sometimes stick with a trusted mechanic, or other service provider, and buy more from them when they need the work, just to maintain the relationship, and because I know I may need them to help me at some future time when they don't really want the extra work. But those exceptions are I think rare.
What about the labour market? The labour market really does seem different from the supermarket model. Buyers of groceries don't feel many qualms about switching supermarkets to get a slightly better deal. But buyers of labour certainly would be made to feel qualms (or a lot worse) if they fired one worker and replaced him with another worker who offered a slightly better deal. In fact, it's the seller of labour who will feel few qualms about switching to another job that offered a slightly better deal. (And this asymmetry in who can easily quit their relationship is surely what places more bargaining power in the hands of the worker than the firm, and explains why the labour market typically has excess supply of labour rather than excess demand, because workers in a tight labour market can credibly threaten to switch to another firm if they don't agree to a pay raise, but firms can't credibly threaten to switch to another worker if they don't agree to a pay cut.)
The simple Marshallian model of sticky prices and Q=Qd may work well for the supermarket, but looks a lot less plausible for the labour market. Might that invalidate sticky-price macroeconomics? Well, it might indeed create problems for sticky wage macroeconomics, but sticky price macroeconomics may have an escape. That's because the demand for labour is (usually) a derived demand. It's derived from the demand for the goods that the labour produces. And if the price of those goods is sticky, and quantity of those goods is demand-determined, then whether wages and other input prices are sticky or not doesn't really matter.
This point has been known since Patinkin in the 1960's, and formalised by Barro and Grossman in 1971. If firms are sales-constrained, because prices are sticky and output is demand-determined, then the effective demand curve for inputs is perfectly inelastic, for any input price below the Marginal Revenue Product of inputs. Regardless of how cheap inputs are, and how productive they are, there is no point in a firm hiring any more inputs if the firm can't sell the extra output those extra inputs could produce. Sure, if wages fell relative to rental prices of capital and land, a firm might want to substitute and hire more labour and less capital and land. But if all firms did this that would just leave capital and land idle, instead of labour. If output is demand-determined with sticky prices, then that determines total employment of inputs, even if input prices all fell to zero.
That's why (or that should be why) New Keynesian macroeconomists focus on sticky prices, not on sticky wages. Given sticky output prices, sticky wages and other input prices are irrelevant. Whether the labour market does or does not conform to the Marshallian model doesn't matter. Output determines employment. Employment is determined by whatever is needed, given the technology, to produce the amount of output that firms can sell, which is determined by the amount of output that supermarket customers are willing to buy at the sticky prices.
Search models of the labour market might be very useful in understanding wage-determination, and the natural rate of unemployment. But they tell us nothing about cyclical fluctuations in employment, if those fluctuations are caused by fluctuations in the demand for output given sticky output prices.
Now there's an implicit assumption in sticky-price macroeconomics that many sticky price macroeconomists themselves may not be aware of. And David may not be aware of it either. There is a market (or set of markets) in which output goods are bought and sold, and a market (or set of markets) in which labour (or the different types of labour and other inputs) are bought and sold. They are bought and sold for money, the medium of exchange. We know that this must be a monetary exchange economy, because there is no market in which labour can be exchanged directly for output. If such a market did exist, then firms that cannot sell extra output, and workers who cannot sell extra labour, would certainly want to participate in that market. Even if nominal prices were too high, and nominal wages too high, so there was excess supply in both output and labour markets, if real wages were at the right level (because nominal prices and wages were too high in the same proportion), unemployed workers would do direct barter deals with firms at that real wage, and there could not be any recession due to deficient aggregate demand.
A Marshallian economy has many markets, one for each of the non-money goods, where that good is exchanged for money. A Walrasian economy has one big centralised market, where all goods are exchanged simultaneously for all other goods. Sticky-price macroeconomics is Marshallian, in that sense, and definitely not Walrasian. By ignoring the possibility of a market in which labour is exchanged directly for output, sticky-price macroeconomists are implicitly assuming a monetary exchange economy. Barter is ruled out. Recessions are inherently a monetary exchange phenomenon in these models. Recessions are caused by a shortage of money – an excess demand for the medium of exchange. (Though once the recession begins, we cannot speak of one excess demand for money, since there are as many excess demands for money as there are markets.)
A model of recessions that says they are caused by an excess demand for the medium of exchange seems right to me. I think I see more resort to home production, barter, and private monies during a recession, and I do see more incentive for people to do so. We see proxies for excess supply generally increase: sellers need more effort to sell; buyers need less effort to buy. The prices that do seem more flexible, because we see them rise and fall daily, tend to fall. It looks right.
[And the fact of monetary exchange means that sticky prices cannot just be a facade. Start in equilibrium, then halve the nominal money supply, holding all nominal prices (and wages etc.) fixed. Is it possible, as Barro 1977 asserts, that buyers and sellers in long-term relationships could honour their implicit contracts and continue to exploit all potentially Pareto-efficient trades, so that there is no effect on quantities traded? Even though all relative prices are unchanged (by assumption), so that a barter economy, or Walrasian market economy, could carry on as before, the real stock of money is now halved. That is a real effect, and will interfere with people's ability to conduct transactions. And taking those extra trading frictions into account, the real volume of efficient trades will be lower. Sticky nominal prices, if they are the prices that determine how much money changes hands when people take delivery of goods, cannot be a facade.]
Now David would want to model those trading frictions that motivate monetary exchange explicitly. And he is right to want to do so (someone needs to do this). But paradoxically, the very anonymity of buyers or sellers that usually underlies explicit models of monetary exchange is what David complains about in the Marshallian/Walrasian markets of sticky-price macroeconomics. I buy what I want at the supermarket, because I am anonymous; and they make me pay cash, because I am anonymous.
Yep, the assumption of sticky prices is a very bad assumption. I hope to God we can do better than this someday. And I applaud all those economists who are trying to do better than this. But right now I can't do better than this, and I don't think you can either. But someday you will, or someone will.
Worthwhile Canadian Initiative 
Your narrative objectifies the store. Surely the store observes the quantity exchanged and adjusts it’s prices too, or things are different in Canada.
In my experience, processed foods have constant prices and fresh foods have highly variable prices. The reason should be clear, processed foods take commodity inputs and warehouse well.
Jon: presumably the store does adjust prices in response to changes in demand and costs, but does so infrequently. Remember my post at the end of May? I came back from a 4 day canoe trip to find the value of my loonie had changed a lot in the forex market and the stock market, but not at the supermarket.
The price of fresh food does vary much more. But usually between fixed price points (e.g. 99c/lb=$2.18 per kg). So I don’t know if it’s really varying in the way that theory says it should.
Crude oil is like processed food: a commodity that warehouses well. But its price varies continuously. Gold too. Why not cans of sardines?
I’m lost in the 3 paragraphs 2 off from the end.
Just to be clear,
“Sticky nominal prices, if they are the prices that determine how much money changes hands when people take delivery of goods, cannot be a facade”
follows from the previous sentence, on the real effect of peoples ability for transactions,
meaning:
The sticky prices are sellers taking into account the looming inefficiencies of lower money supply.
And how does this compare to the monopolistic competition sticky prices, of the grocery store.
” But whatever it is that causes sticky prices, it’s something on the supply-side “
I don’t think you are giving discounts to “list price” (couponing, sales, and non-price promotions) their due. Firms typically discuss these as “giving up 2 points of pricing because the environment was promotional”. Further, I agree with Jon–the “outer ring” of a supermarket (fresh items) has quite variable prices, and the range you observe may just be a function of the limited variation in input costs.
Truly sticky-price supermarket items that are ultimately a sign of imperfect competition. These firms will hold nominal prices and let output vary, but only because their input cost inflation expectations are tame. In higher-inflation countries, this is not the case, and bar-coding allows for fairly constant changes in supermarket prices. This renders the concept of “sticky prices” tautological. In other words, “prices are sticky because input costs are sticky; and input costs are sticky because prices are sticky.”
Does one explain sticky up, and the other sticky down?
@Jon
See this paper for some work on that topic: http://www.nber.org/papers/w13829
The gist is that nominal rigidities do not necessarily take the form of sticky prices, but may take the form of sticky “reference prices” that adjust infrequently and prices regularly return to.
edeast: sorry, but you lost me on your “you lost me” comment. Which bit don’t you understand?
The sentence: “And the fact of monetary exchange means that sticky prices cannot just be a facade.” states the thesis of a new argument put forward in the paragraph which it begins. It’s not a summary of the argument above. You could even delete that paragraph and the rest of the post would still cohere.
Does that help clarify?
David: imperfect competition does not explain sticky prices (though it might make sticky prices easier to explain, since the loss in profits of failing to adjust prices are lower under imperfect competition). And even if MC curves were flat, and elasticities of demand constant, all you get are sticky real (relative( prices at the individual firm level, not sticky nominal prices at the aggregate level.
jsalvati: good find! That was exactly what I had in mind when I was talking about “temporary sales” in the post, and “fixed price points” in my response to Jon. Is that Eichenbaum et al paper the seminal paper on that point? (I think it is, but my brain is mush — especially in the heat here this last few days, 34C!).
“I don’t know why my supermarket has sticky prices.”
You could ask them? But I guess that wouldn’t be scientific.
If you don’t know why your supermarket has sticky prices, you could ask them.
I’m amazed at the tendency for academic economists to invent puzzles where none exist.
here’s how I read it,
The first part of the blog concerns, sticky prices, and the monopolistic competition,
You explain grocery store sticky prices are marshallian, then you describe how the labor market doesn’t matter, because “Given sticky output prices, sticky wages and other input prices are irrelevant.”
I wouldn’t mind an explanation of your monopolistic competition, maybe now with graphs 😉 (A stronger or simpler proof of sticky output prices.)
The confusion I’m having is with the second part of the essay where you describe the hidden assumption of sticky prices. I read those paragraphs as building a case.
Paragraph 1, describes whether a marshallian or walrusian,
Paragraph 2, describes the assumption,
“…,sticky-price macroeconomists are implicitly assuming a monetary exchange economy” ie marshallian.
Paragraph 3, is your defense, on your intuition, that recessions are caused by “excess demand for the medium”
Paragrah 4. The assumption becomes fact. You assert with the beginning sentence and the final sentence that sticky prices are not a façade. And the middle sentences are I take it your attempt, at guessing why. If you halv the money supply, frictions are introduced. Ineffeciencies etc.
So I took that as meaning, the sticky prices value came from the frictions of the monetary recession. And I was trying to compare this to your previous output sticky prices. Now you want to delete this paragraph but I find it the most interesting.
Darren: It’s been tried. Robert Hall (? IIRC) did it 30 years ago. It’s worth doing, but it doesn’t really answer the question. It’s a bit like asking giraffes why they have long necks. Even if giraffes could speak our language, it wouldn’t really distinguish the Darwinian from the Lamarkian(sp?) theories.
“Sticky prices” is really a system property, not merely an individual firm property. (You can build models where each individual price is sticky but the average price level is perfectly flexible, Caplin and Spulber?). Plus the individual firm might know only that “it works”, not why it works. Same with giraffes.
edeast: thanks for explaining. Damn! I didn’t write that post clearly enough. The logical structure of the essay isn’t transparent.
Paragraph 4 (“And the fact of monetary exchange means that sticky prices cannot just be a facade…..”) is really just an aside. I’m throwing an extra stone at Barro’s 1977 argument, showing that it can’t work even if you do follow all his assumptions.
I’m going to stick brackets around that whole paragraph. An ugly fix, but better than nothing.
Your asides and blogging style, allow non-economists to get some traction. My fault for not understading Barro, but I do have my preferred, mental model, that I’m trying to build a case for so responsibility lies with readers as well.
This discussion reminds me of fascinating explanation I heard from Max Keiser about the role of market specialists in the stock market. This was done in the context of explaining high-frequency trading and the role of virtual market specialists in the Flash Crash last May.
Max Keiser does a radio internet show out of London& Paris. A former stock trader in the 80’s, in the 90’s he invented the Hollywood Virtual Exchange, and the virtual market specialist module is the basis for many of the high frequency trading programs used by the big Wall Street banks today.
In the stock market, bids and offers come in randomly, and in the short run very rarely match. So in order to have a stable market and smooth price changes, “market specialists” are specialized traders that hold inventories of stock, and can see in-coming orders, and they use their inventories to offset short-term gaps in supply and demand. So prices fluctuate not so much because of the immediate supply and demand conditions, but more from the of the expectations of the market specialists about the longer-term trend in market conditions.
So, to bring this to your example of the grocery store, the store acts as a market specialist, keeping inventories of groceries, and they have a superior knowledge of market conditions as compared to most sellers and buyers in the food market. Instead of adjusting prices instantaneously based on instantaneous fluctuations in suppply and demand, the retailer will use adjustements in their inventories to match immediate demand and supply, and will only change prices slowly according to their expectations of the more general trend of supply and demand conditions.
In short, if you see a market with stable or smoothly changing prices, that’s probably because their are “market specialists” managing holdong inventories and using them to manage short-term fluctuations in supply and demand.
Max Keiser’s radio and TV shows can be watched or listened to via his website at maxkeiser.com
I don’t believe prices are sticky. Gray’s Papaya has a recession special (2 hot dogs and a drink $4.45 instead of 5.45).
Nick, In your comment you referred to models where individual prices are fixed and the overall price level is flexible. I seem to recall reading the following about price stickiness:
1. If individual prices fully adjust whenever the current price is more than X% out of equilibrium, then individual price stickiness doesn’t create aggregate price stickiness.
2. If individual prices fully adjust after X period of time, then individual price stickiness does create aggregate price stickiness.
Does that sound right? I suppose that’s one reason I have always put more weight on wage stickiness, which fits the second category.
I was surprised that at the end of David’s paper he indicated that his real target wasn’t just sticky price models, but any model where nominal shocks have real effects. He seems perplexed that some economists would believe in models that imply that the 50% fall in US NGDP between 1929 and 1933 might have actually had some sort of causal role in the Great Depression. Alternatively, he seems to think that if whatever had caused the Great Depression had not happened, then a 50% fall in NGDP would not have created even a minor recession.
I’ve devoted too much of my life to economic history to find those sorts of models plausible.
Nick, no disagreement that commodity (spot) prices are volatile. My point is that a quantity variation in a given processed food is unlike to move the commodity markets because they are much deeper than a single consumer good, and that its very easy to hedge against price fluctuations.
I think your case about markets routinely differ from the model is unsubstantiated. The big sources of sticky prices are labor contracts, good-will, and rented capital.
jsalvati,
I’ve reviewed that paper. My thoughts: the paper is really much more effective at damaging the claim of monopolistic competition in the retail grocery market. This comports with my knowledge of the business that its based on razor-thin margins.
The authors claim that the retailers appear to set their prices based on cost but attenuate cost fluctuations as well. Again, this comports well with my understanding of the business: there are fixed costs associated with the price changes. Given some budget constraint, the retailer will strategically change prices, but its not optimal to change all prices, all the time.
The retailers appear to pass-thru costs because their production function is irrelevant for determining prices. Its the aggregate production function that matters. i.e., variations in consumer demand and producer supply change the equilibrium prices of the costs. If one store receives rush orders, the store usually reacts by withdrawing from the sale of that item because the store is a price taker. Other stores can and will supply at the old passthru cost.
I use to think that the stickiness i relative -the result of relative slow change in prices. There are assets, that are very quicly in adjust their prices, and there are good markets, very more slowly, and labor market, very very slow, indeed. So, these different velocity of adjustment make trouble in money market, because the velocity of circulation fall, and money disappear from some market.
If the velocity to adjust were the same in all markets, I suppose there will be no problem -the quantity of money need not to be adjusted to maintain the level of activity: Perfect Say´s law working!. So I think that the different velocity of price adjustment is the problem.
So, I agree with your argument, it explain very well the stickiness, but I think that the real problem is the big difference in changes in prices.
I think you need to think of employment as a long term contract, not a spot price.
Both the employee and the firm view employment as a long term relationship, much like a marriage, that will last through significant market fluctuations. It is to the advantage of both to ignore short run fluctuations.
But a spot price is just a momentary relationship that does not last longer than a very short time and may not be valid after that. At another time it may take a very different price to clear a spot market. But that does not apply to a long term contract. Even long term contracts to supply and purchase a good typically include terms to change the price if conditions change. For example, firms may buy and sell a volume of a good at a certain price, but if volume fall below or rises above that amount the price may also rise or fall for the higher(lower) amount.
I think I get it now.
Is it true that anonymity of buyers or sellers underlies explicit models? because David’s whole post seemed to be about relationships.
edeast: I’m not an expert on modern explicit models of monetary exchange. But from what I understand, a common and easy way to explain why people use money is to assume decentralised markets plus anonymity. Anonymity means you can’t use some sort of credit system where you get the goods now and promise to provide goods in return at some future date or some other location. They don’t know who you are, so can’t check whether you keep your promise.
spencer: I agree that employment looks a lot like that, with long term contracts, either explicit or implicit. That was what Barro’s 1977 paper was about. But the output market usually doesn’t look like that. And even in the labour market, firms typically reserve the right to layoff workers in the event of a decline in demand for their product. The question is whether as many workers are laid off than they would be if wages were flexible. (That’s distinct from the question of whether employment on average over the business cycle would be higher if wages were more flexible.
Luis: I disagree. The fact that some prices are more flexible than others causes changes in relative prices and the composition of output and excess supplies. But even if all prices were equally rigid, in the face of an exogenous increased demand or decreased supply of money would still cause excess supplies of all goods, and a recession.
Scott: That’s about right. The Caplin Spulber(?) model, IIRC, had an s,S pricing rule. When your price gets 1-s% below the equilibrium adjust it to S-1% above the equilibrium. This was in a model where the equilibrium price was always rising over time, but at faster or slower rates. And where s and S were independent of the expected equilibrium path, or how much time had elapsed. Carleton’s cafeteria changes prices once per year. That’s more like the Taylor model than Calvo.
I really dislike the Calvo assumption of purely random times at which to change price. We know it’s wrong, and it rules out inflation inertia (which seems to exist). But we do it because it’s mathematically tractable, since all firms have the same chance of changing prices each period, so we can use representative firm analysis.
Funnily enough “history”, in the form of the 1982 recession, was influential in my thinking too. I wonder why that is? Is it because we get a richer reading of events by living through them, or by studying them as a historian, than by looking at aggregate data on P and Y etc.? And this helps us identify shocks to money demand and supply better? (Like Friedman and Schwartz’s Monetary History of the US?).
Jon: supermarkets can face downward-sloping demand curves, and still have zero margins of P over ATC in equilibrium. Are you sure those “razor thin margins” are margins of P over MC, or over ATC? I can’t believe that individual supermarkets face horizontal demand curves. The cost of travelling to the next nearest supermarket is not trivial. Plus, they are all different in other ways, so not every customer will switch 100% for a 5% increase in relative price.
Nick: I tend to agree with David that the big Macro events aren’t explicable on the basis of price stickiness. This is recognized by the SPs themselves insofar as they rely on the liquidity trap, which of course can’t be fixed by flexible prices. But why is there so much resistance to the multiple real equilibria interpretation of macroeconomic experience? – whether Farmer, or Diamond’s search externality model, or any number of others. The sticky price stuff can explain what goes on this side of Leijohufvud’s corridor. On the other side I think you need the idea of multiple Pareto-ranked real equilibria.
kevin: I think there are two questions:
1. Is current Y determined by the AD curve plus sticky prices? In other words, are we off the LRAS curve in the short run. All sticky-price (or sticky wage) macroeconomists must answer “Yes” to this question, by definition.
2. Is the AD curve downward-sloping, or vertical? In other words, would increased price flexibility get us back to the LRAS curve? If you believe in liquidity traps, for example, you believe the AD is vertical (at least under some circumstances), and so if prices were perfectly flexible we would now face explosive deflation, but that would not get us back to the LRAS curve. A sticky-price macroeconomist could answer yes or no to this question.
I think that sticky-price macroeconomists, plus their critics, tend to conflate these two questions. I say “yes” to 1, and sometimes refuse to answer 2, on the grounds that it’s a hypothetical! Since prices are sticky, we need to work with that, and what would happen if they were perfectly flexible (whether we would return immediately to full employment, or just have explosive deflation) is not a relevant policy question.
Kevin: continued: and the multiple equilibria people are either saying:
3. There’s not one LRAS curve, but several, so that if AD falls by a big enough amount, we might get onto a lower LRAS curve, with lower Y, and yet zero downward pressure on prices, and actually upward pressure on prices if we shift AD right a little bit.
4. There’s not one AD curve, but several. If a big temporary shock shifts it left, we can’t easily shift it right again. There’s hysterisis in AD. But if AD falls, there will be permanent deflationary pressure (unlike 3 above).
Nick: to take the most extreme example the typical markup over wholesale at walmart is one cent. Those are tight margins, and this has been transformational in that segment of the retail market.
The situation among low-cost grocers just isn’t comparable to an upmarket retailer like H&M or niche chains such as Whole Foods.
Look at the price history of this property in the Seattle area. Sticky prices? Perhaps. The mechanism in this case is a slow-step reverse auction combined with seller patience / stubbornness / optimism / perception that product is semi-unique or specially valuable?
http://www.redfin.com/WA/Seattle/7500-S-Taft-St-98178/home/177690
Jon: I just find it hard to imagine how Walmart can cover its costs of labour, inventory-keeping, etc., at that sort of margin. Even if they do pay low wages and have very high volume and excellent inventory control. But even if Walmart is perfectly competitive, their suppliers might not be, and retailers on average might not be. (And does Walmart not have a lot of prices at $4.99 etc., which is hard for me to reconcile with PC and price flexibility).
Indy: Good example. Apparent price stickiness in the housing market has always interested me. There are no long term implicit contracts there between sellers and potential buyers. Both the “jumpiness” of listed prices (which may not necessarily mean jumpiness of actual transactions prices). Plus the fact that we always seem to have large inventories when prices are falling, and small inventories when prices are rising. That certainly makes it look like slow adjustment of prices towards equilibrium prices, with variations in excess supply or demand being the result. Typically, it’s harder to buy and easier to sell in a rising market, and the converse in a falling market. If prices adjusted instantly to changes in demand or supply, even if houses and buyers are heterogeneous (so it’s a search and matching market), we wouldn’t have the concepts of “buyers’ market” and “sellers’ market”, and they wouldn’t correlate with falling and rising prices.
The housing market always makes me think that there might be something to Lucas 1972 after all. Nobody is forcing sellers and buyers to have sticky prices. It’s just that their expectations of market equilibrium do seem to take a long time to adjust.
Nick: a fair point about wholesale costs, but a fairer point to me would be that unit costs at walmart are frequently fractions of a dollar, so that could still be tens of a percent.
And its true that walmart pays below the general wholesale because they are very good at browbeating suppliers and finding lower-cost suppliers.
Still, you have to make a contrast with other retailers. For instance, Whole Foods had been earning bottom-line profits around twenty-percent of revenue which suggest phenomenal margins.
I don’t shop at walmart, so I don’t know if they offer a lot of goods at 4.99, do they? What wouldn’t surprise me is that they still practice price (search) discrimination techniques. Some goods being priced close to cost (or below), and others priced too high to advantage the one-stop shop format that they have.
They used to have a compaign where they explicitly advertised that they didn’t have 99c prices.
Okay, so net income is about 3.5% of revenue and operating income is about 6% of revenue. That doesn’t really tell you how far apart marginal revenue and marginal cost are though.
I’m reading a paper that might fit into mankiw’s information stickyness. It uses rational inattention to explain why prices are flexible to “idiosyncratic” shocks but slow to nominal.
Rather it’s against Mankiw, is 125 cititations good in economics? I am partial to information theory, but I’m not sure how well received Christopher Sims work is, in econ.
Here is asecond paper based on Sims work. It covers discrete price changes, which you mention in your next post.
Nick:
I have been out of town until now. Just read your reply to my post; very well done. (It’s a pleasure to read such thoughtful and learned responses of this sort — the general discussion has been great too).
Your point on anonymity and money was a good one. True, the demand for money (in the modern monetary literature) is motivated by some degree of anonymity (otherwise, credit is possible). I will have to think about how one might square monetary trade (anonymity) with the relationship trade I stressed.
Let me summarize my main points. These are as follows. Nominal price “rigidities” appear to be an empirical fact. But do not confuse fact with theory; there are potentially many different ways to interpret price stickiness. This is potentially important because different interpretations can lead to very different policy conclusions. To simply assume that prices are sticky and then arrive at a very strong policy conclusion (a la Paul Krugman) seems wrong to me (and potentially dangerous).
Nick, you appear to be sympathetic to this. And you defend the sticky price hypothesis in an admirable way. I am fine with this. I would only add that policy conclusions based on this assumption should not be held with religious conviction (I am not accusing you of this…you know who I am talking about).
I do have one question for you. You say:
“A model of recessions that says they are caused by an excess demand for the medium of exchange seems right to me. I think I see more resort to home production, barter, and private monies during a recession, and I do see more incentive for people to do so.”
What, in your view, causes an “excess demand for the medium of exchange?” Do you believe that the recent recession what caused by a sudden excess demand for US government money and treasuries? Do you have something deeper than that in mind?
Thanks David! I only just saw your comment. Nearly missed it altogether.
“What, in your view, causes an “excess demand for the medium of exchange?” Do you believe that the recent recession what caused by a sudden excess demand for US government money and treasuries? Do you have something deeper than that in mind?”
Good question. I wish I had a better answer. If I did have a better answer, my theory would be a lot more testable. This is all I’ve got:
Lots of things might. (Good cop out answer!)
Sometimes, like in the 1982 recession, a negative shock to money supply looks like the reason. The Bank of Canada wanted to reduce inflation, and finally tightened monetary policy enough to do so. Trouble is, even here we can’t expect to be able to clearly identify a supply shock from a demand shock, since lower expected inflation would increase the demand for money.
Sometimes, it’s a fall in the natural rate of interest. (1991 recession?)
Sometimes, like in the recent recession, it’s caused by a rush to liquidity, when some previously liquid assets become a lot less liquid, and people and FIs might have to unwind positions very quickly. The medium of exchange is the most liquid of all assets. The US dollar, as the world’s reserve currency, is money to all the other monies. If you want to exchange Loonies for Aussie dollars, you have to go through the US dollar. Just as if you want to exchange apples for bananas, you have to go through money.
That’s not really much deeper than “a sudden excess demand for US government money and treasuries”, but it’s a little bit deeper.