Do illiquidity and sticky prices go together? If so, why?

Some goods are easy to buy and sell quickly. They are liquid. Other goods are harder to buy and sell quickly. They are illiquid.

Some goods' prices adjust quickly to changes in demand and supply. They have flexible prices. Other goods' prices adjust slowly to changes in demand and supply. They have sticky prices.

1. Line up all the goods in a list, with the most liquid goods at the top of the list and the most illiquid goods at the bottom of the list.

2. Now line up all the goods in a second list, with the goods that have the most flexible prices at the top of the list and the goods that have the stickiest prices at the bottom of the list.

Delete money from both lists, because money is weird.

Do your two lists tend to line up? Are the positions of goods on the two lists positively correlated? Do the same goods tend to appear near the top of both lists and near the bottom of both lists?

If so, why?

I think the two lists are positively correlated. I can think of three possible reasons why they would be positively correlated:

1. Price stickiness causes illiquidity. Suppose a good had a sticky price. That price would often tend to be a long way away from the market-clearing equilibrium price. Nearly half the time the good would be in excess demand, and hard to buy, and nearly the other half of the time it would be in excess supply, and hard to sell.

2. Illiquidity causes price stickiness. ?

3. The same attributes cause both illiquidity and price stickiness. Some goods are homogenous and have a high volume of trade. That makes them easy to buy and sell quickly, so they are liquid. It also makes it very easy to compare prices, and makes a strong incentive for individual buyers and sellers to raise prices when there is excess demand and cut prices when there is excess supply. Other goods are heterogeneous and have a low volume of trade. That makes them illiquid and have sticky prices.

It is much easier to do economics under the assumption that all goods are perfectly liquid and have perfectly flexible prices. Walrasian (or Arrow Debreu) general equilibrium theory makes both those two assumptions. Most macroeconomists find we need to relax those two assumptions to some extent. Most of us assume that some prices are sticky, in terms of money. And most of us assume monetary exchange; which means that trade of non-money goods against other non-money goods (barter) is prohibitively difficult. Which is like saying that goods are very illiquid, except when traded for money, the most liquid of all goods.

Macroeconomists would like to understand illiquidity better and would like to understand price stickiness better. Maybe the two questions are related?

Just a random musing.

Comments

[Andrew]

There is stuff on this, e.g. Briana Chang's JMP.

[JP Koning]

Interesting musing. Are you confining your query solely to goods markets or are you willing to consider asset markets too? I can draw up lists of liquid and illiquid equities, its quite easy. Just measure relative bid-ask spreads. But I'm not sure how to measure stickiness, especially in equity markets, and from there build up a ranking based on stickiness. What formula should I use? If the idea is that equity markets don't exhibit any stickiness whatsoever, then it's hard to draw a link between stickiness and illiquidity, since many equities surely demonstrate illiquidity.

[Nick Rowe]

Andrew: This one? Looks interesting. Tough going for me. My sense however is that it concentrates on understanding liquidity and firesale prices, rather than (what looks like) "sticky prices".

[Nick Rowe]

JP: I definitely wanted to include all goods, including real assets and financial assets too. And labour. And services. Everything, really. Except money, because it's weird, being the medium of account, and exchange.

I'm not sure how exactly to measure price stickiness either. As a first pass, if we are talking financial assets, if we saw a sequence of trades at exactly the same price, then a discrete jump in price, I might say that looked like stickiness. Or maybe imbalances on the buy side or sell side?

[Frances Woolley]

Interesting post. My reaction is that lumpiness and loss aversion matter, as well as low volume and heterogeneity.

E.g. housing. People won't reduce the price of their houses because they hate the idea of incurring a loss, and the fact that houses are large and lumpy means that a reduction in the price of one's home represents a large change in one's net worth. Ditto the supply of one's labour.

Now do these cause both price stickiness and illiquidity as in your case 3? Or do they cause price stickiness, which then has other consequences as in your case 1? That I have no clue. Back to marking.

[JP Koning]

Nick, here is a way to think about stickiness, maybe. In futures markets, there are two major S&P 500 futures contracts. One is traded 24 hour electronically (the emini), the other is traded old fashioned style in the pits. The emini is far more liquid... narrower bid ask spreads and far larger volume. But this wasn't always the case, the pit traded contract used to be the de facto standard.

Each morning when the pit opens the opening price shows a large discontinuous change from the previous night's close, basically catching up to the emini. Given it's relative inflexibility, one could call the pit price the sticky price, or at least stickier than the emini. But once they are both open, the pit contract and the mini move in line with each other and there is no difference in stickiness.

The reason for the relative stickiness is due to the underlying market structure. One market closes for the night, one stays open continuously.

Back in the 1990s, the pit contract was both more sticky and more liquid. Over time, due to the 24 hour nature of the emini contract (ie. it's lack of stickiness) the liquidity migrated to the emini.

So if you're willing to go along with my example and definitions, I'd say that the design of a financial market will determine if prices are more or less sticky. The design with the least stickiness will eventually attract the most liquidity.

[Patrick]

Stuff that's hard to sell (illiquid) is sometimes so because it's hard/expensive to find someone who wants to buy the thing, regardless of price stickiness. eBay and Kijiji are maybe examples of how illiquid markets (I'm thinking of the the market for other peoples used crap) can be made more liquid by making it easier for potential buyers and sellers to find each other. The mountain of plastic kids toys squirrelled away under my deck is testament to (my wife's contribution to) the increased liquidity for said goods that a change in market structure can bring about. :)

[Lord]

Other qualities that affect liquidity are size and expense, large units (lumpiness above) and expensive units being less liquid. Both tend towards smaller markets, lower volumes, more costly errors, more sensitivity to interest rates as financing is involved, and often more uniqueness.

[jh]

Matejka and Sims (here) is a very interesting take on stickiness, using Sims's rational inattention idea.

It's different than the 'type' of stickiness that FW seems to be hinting at. Instead of being merely 'close' to the true value (being reluctant to take a large loss, but not refusing to go down marginally), it argues that only a finite set of prices will be observed (when individuals have an information processing constraint). These both seem to be plausible 'types' of stickiness (to me, anyway), and it may be that they need to be separated.

(And personally, I think labor may be a whole other kettle of fish, at least in terms of the mechanism. There, I think about Truman Bewley's interviews, Rosen-style tournaments, and some of Robert Frank's ideas. But that's just me.)

[fmb]

Re futures: I think an unchanging last with no trades is not really stickiness, stickiness requires trades at substantially the same price despite evolving fundamentals. When the pit is closed, liquidity is nil and there are no trades to be at sticky prices.

But that's not why I wanted to comment.

I think the first part of #3 is about right: homogenous goods with a high volume of trade should generally be both liquid and flexible. However, for heterogeneous goods, I think illiquidity and stickiness are 2 different things:

a) Illiquidity can be related solely to the returns on searching. A well-informed & rational seller of real estate likely faces a tradeoff between price and speed of sale (due to search costs for a buyer that complements the asset well) where an expected speed of sale of months will look best. In a statistical sense, when you face steep contango, you store and sell forward. Such a buyer might adjust continuously to any changes in fundamentals, but the asset will appear (and be) illiquid because search takes time.

Of course, as FW points out, an irrational seller might overestimate the returns to search and add stickiness and exacerbate their illiquidity. While this surely happens, it's not necessary for illiquidity.

b) I think the interesting sources of price stickiness really arise in repeated trade scenarios, such as ongoing employment. No one would say that the coupon on a bond is a sticky price; instead it's a contractual term that as time passes, prevailing interest rate changes might lead buyer or the seller to be better off than expected. Most employment arrangements aren't so contractual, but there are quasi-rents at stake for both parties, and uncertainty for both parties, that leads (I think) to a kind of tentative arrangement to renew continuously at a similar wage, with ambiguous contingencies for surprising developments. Both parties need to be concerned about attempts to misappropriate quasi-rents through hold-up, though, and so employers seek credibility (possibly through not lowering wages), and employees play a game of chicken where they won't accept pay cuts. Ambiguity about whether a pay cut is "fundamentally right" and whether everyone knows that everyone knows this exacerbates the issue.

In a sense, like with a bond, this would not so much be a sticky price as a contract where one side is getting an unexpected portion of the surplus. Pre-specifying all contingencies is too costly, so some contingencies require a threshold before renegotiation occurs. This can become problematic when renegotiation becomes appropriate, but the cost of doing so is high and the need is widespread.

[david]

I can't remember who pointed it out, but I remember a critique of New Keynesianism arguing that NK focused on sticky prices whereas industrial studies actually supported sticky quantities, aka illiquidity, in non-inventory assets (capital, etc.). Of course in principle one can just have both.

I do wonder whether the vague principle of there being an adjustment cost that is what unites the two. Not the menu cost (prices)/inventory maintenance cost (quantities) precisely, but the more general point that when something unexpected happens that requires changing anticipated plans regarding some good, changing plans for some good is more costly than changing plans for other goods. Insofar as the vast majority of economic labour activity involves managing machines and other labour rather than operating machines itself, the act of management is itself in scarce supply.

[Greg Ransom]

Illiquidity causes price stickiness: You are trying to sell new shares in you rumored-shaky bank / financial institution during a financial storm, and you know that the government and/or Fed will bail out/buy your existing shares at some fixed price if worst comes to worse and your financial firm is about to fail.

New shares in a rumored-shaky firm are fairly illiquid because of asymmetrical information problems, market uncertainty, etc., and this illiquidity in conjunction with To Big To Fail produces price stickiness -- existing share holders won't drop the price of new shares to a market clearing level.

[Greg Ransom]

"Macroeconomists would like to understand illiquidity better and would like to understand price stickiness better. Maybe the two questions are related?"

Isn't Wall Street one place to look? E.g. look at how Mortgage Backed Securities and Credit Default Swaps gained and lost in liquidity, or how the price of a controlling stake in various financial firms became less or more sticky, etc.

There is significant heterogeneity between banks & financial institutions and there is low turnover of ownership or control with these institutions.

http://faculty.unlv.edu/msullivan/Sweeney%20-%20Money%20supply%20and%20inflation.pdf

[JP Koning]

"I think an unchanging last with no trades is not really stickiness, stickiness requires trades at substantially the same price despite evolving fundamentals. When the pit is closed, liquidity is nil and there are no trades to be at sticky prices."

I don't disagree, my futures scenario was a bit of a hail mary.

Ok, continuing in the vein of stickiness in financial markets, how about a currency that trades at a pegged rate versus one that freely floats? Obviously the pegged currency is the sticky price, the floating one the non-sticky one. In this case, some sort of market power creates the sticky price. You'll see this in equity markets for short periods of time, no more than an hour or two, when one large trader caps a stock at some price or holds it up.

I've never seen evidence, either statistical or anecdotal, that sticky currencies are more liquid than non-sticky currencies.

[Nick Rowe]

All: maybe I should just add. I saw goods like labour down the bottom of both lists. Illiquid, and sticky prices. At the top of both lists were goods like stocks and bonds that were highly liquid and flexible prices. Illiquidity and sticky prices are conceptually different. We can imagine one without the other. But empirically they do seem (at least at first glance) to be correlated.

Frances: thanks! Loss aversion seems to apply to liquid financial assets too. I read (somewhere) that people don't like to sell shares if the current price is more than the price they paid. But that just affects the position of the supply curve, rather than causing sticky prices and illiquidity. Whereas in housing markets, as you say, loss aversion might cause sticky prices. People put the house on the market, but keep the price at what they paid for it.

JP: I was thinking just of the attributes of the goods themselves. You are right to say that the attributes of the market in which those goods are traded matters too. But if a market closes overnight, that makes the good less liquid, but I wouldn't say it causes price stickiness. There ought to be a gap between last night's closing price and this morning's opening price, if new information comes in overnight.

Patrick: a lot of my thinking about liquidity is based on my experience of buying a used MX6 ("other peoples used crap"???!!!!) on Kijiji. The longer you wait, the better the deal you can get, from both sides of the market. A lot of it is sheer luck (a buyer and seller arriving on the market at nearly the same time in nearly the same location with a car that is nearly the colour and condition the buyer wants most), but there are differences in patience too, on both sides of the market, so you get the best deal if you are patient and wait for a person on the other side of the market who is impatient. This heterogeneity of cars and buyers, plus the thinness of the market, plus the absence of data on transactions prices, makes it very hard to know if a price is too high or too low, and what the trade-off would be to changing your bid or ask price. I suspect that trade-off is rather flat (like in "small menu cost" models), so there is little cost to having a sticky price (or little benefit to adjusting your price in response to new information).

jh: that Matejka and Sims paper looks really promising, if I understand it, but the math is impossibly hard for me. If I understand it right, it does seem to suggest that we would tend to see stickier prices where the information processing costs are higher, which would also correlate with illiquidity? If so, it would seem to explain the sort of correlation I'm talking about here.

fmb: "However, for heterogeneous goods, I think illiquidity and stickiness are 2 different things:"

I agree they are (conceptually) two different things, but empirically I think they tend to go together.

"Such a buyer [of a house or MX6] might adjust continuously to any changes in fundamentals, but the asset will appear (and be) illiquid because search takes time."

They might, but do they? My hunch is they won't.

I agree on your point b. If you buy a new car on installments, the fact that the monthly payments are sticky ex post doesn't mean the ex ante price of the car is sticky in the relevant sense, if it's hard to back out of the deal ex post. But if you can exit the deal, there might be some sort of stickiness due to the fear that if you adjust your price you make yourself vulnerable to gaming by the other side.

david: the way I read it, there are adjustment costs to quantities too. Like hiring costs. Hiring costs mean that employment will respond slowly to increases in AD, even if prices and wages are sticky. Simple NK models ignore those sticky quantities, but I think you need to build them in for higher frequency data.

Greg: Too big to fail is a bit like government price supports for wheat and stuff. Less a sticky price, than the government adding in a perfectly elastic demand curve at the support level.

[Nick Rowe]

JP: there are two types of pegged exchange rates: where it's pegged by law, regardless of supply and demand, which is like a minimum wage law; where it's pegged by the central bank intervening in the market to buy and sell and so creating a perfectly elastic supply/demand curve at the pegged price.

Cuba does the first; most countries do the second. I wouldn't call the second a sticky price, because the price never deviates from the supply/demand equilibrium.

[JP Koning]

"I wouldn't call the second a sticky price, because the price never deviates from the supply/demand equilibrium."

In that case, I can't think of any situation in which equity markets exhibit differing levels of price stickiness. Never seen it... bids and offers for illiquid stocks adjust just as quickly to news as bids and offers for liquid stocks. But I can easily rank equities from illiquid to liquid.

So I'd reject the hypothesis of positive correlation... that the same equities tend to appear near the top of the liquidity/flexible list, and the same appear near the bottom. That's because I don't think it's possible to build a list of equities ranked by stickiness.

[Greg Ransom]

"Greg: Too big to fail is a bit like government price supports for wheat and stuff. Less a sticky price, than the government adding in a perfectly elastic demand curve at the support level."

But TBTF is not a guaranteed price or insurance policy -- it can make a price stick that then collapses.

[Greg Ransom]

Can't we tell similar stories about why labor, especially union labor is sticky? The government is raising is adding costs and restricting the freedom of action upon firms, and is adding a perfectly elastic demand curve for non-employment, which is on the other side of the labor market. There are frequently such stories to be told.

"Greg: Too big to fail is a bit like government price supports for wheat and stuff. Less a sticky price, than the government adding in a perfectly elastic demand curve at the support level."

[Greg Ransom]

The government gives price support for non-employment -- which expands in length as the bust extends in length.

That is part of the sticky wage environment.

[jh]

On Matejka and Sims -- M&S is more about the form of the stickiness than the 'extent' of it. Some of Sims's earlier papers on rational inattention might be more relevant to what you're thinking about. I don't know if you're right about the link with illiquidity. It seems possible, but if more complicated decision problems are also the most important (which seems to be true, at least with housing and labor), then more 'attention' will be allocated to these, which could potentially lead to 'less stickiness'. Plus, the stickiness of market prices will depend on the characteristics of both sides of the market. In one of Matejka's earlier papers (cited in M&S), he shows that a monopolist will choose discrete pricing if the consumers are attention constrained, even if he is not.

[Nick Rowe]

jh: "I don't know if you're right about the link with illiquidity. It seems possible, but if more complicated decision problems are also the most important (which seems to be true, at least with housing and labor), then more 'attention' will be allocated to these, which could potentially lead to 'less stickiness'."

My intuition went the other way. Attention would be more costly in more complicated decision problems, so those prices would be more sticky. If it takes you hours to figure out the new optimal price, you wouldn't want to figure it out every day. If it takes you only seconds to figure out the new optimal price, you would figure it out every day.

[Jon]

Liquidity is valuation risk. If it is easy to know the NPV of an asset, it is easy to sell. What happens usually is that is that the owner of asset has lower valuation risk than the potential buyer; that opens the wedge between the bid and ask.

Price stickiness is also a potential wedge between bid and ask. Thing is that market participants have inclined supply and demand schedules, so the wedge appears in the quantity exchanged.

[jh]

Oh, yeah, I see what you mean. I'm just saying that it might go the other way, because (roughly) complicated is correlated with important. Considering the importance of labor and housing to most people, a great deal of information processing should be allocated on those problems. (At least when they arise exogenously -- I mean, someone need not figure her home value every day, only when a decision is made to move.) Same thing, it seems to me, for buyers of labor and housing. The overall effect in these markets seems ambiguous.

I guess, intuitively, I buy FW's loss aversion explanation better for housing. (And maybe something else entirely for labor.) But generalizing from Matejka, it might be interesting to look at the entire market, how the distribution of attention between people could affect the way market prices are formed.

[Jon]

Hmm I don't think what wrote is quite right. I stand by the first part though, liquidity is from the risk in valuation and it manifests as the bid/ask spread.

[edeast]

Hi, Nick what about Edgeworth conjecture. The less liquid goods have less traders, not as forced to meet the competitive equilibrium. Prices seem sticky,ie the core is wider than the strict competitive equilibrium in less liquid goods.

[Andy Harless]

Shouldn't a good New Keynesian think that the underlying variable is market power? If a seller has market power, then (1) she will set a price at which she would prefer to sell more than purchasers wish to buy, so that, from her point of view, the good will be illiquid, in that she can't sell a marginal unit easily and (2) assuming there are menu costs, she will have no incentive to change her prices when a small shock hits, so prices will be sticky. If a seller has no market power, she will have no choice about what price to charge, and she can render the good perfectly liquid with an infinitessimal discount below the market price, and she will be forced to respond to a shock, no matter how small, by changing her price, so prices will be flexible.

[Lee Kelly]

I can't think of a satisfying explanation of why liquidity and price stickiness correlate, but they sure appear to. Noting such a correlation may also help us better explain recessions.

Stickier prices fall last. Liquid assets tend to have less sticky prices. When there is a shortage of money, liquid assets will tend to fall in value first, and exacerbating the excess demand for money.

[Greg Ransom]

What is most important is when assets _switch_ at the turn of boom-to-bust from being highly liquid to being highly illiquid, or highly sticky and to highly slippery.

MBSs and CDSs were highly liquid and then suddenly were highly illiquid. Ditto all sorts of financial instruments, assets, securities, and shares. And many different assets and instruments were rather sticky, and then suddenly were rather slippery.

Lee writes,

"Stickier prices fall last. Liquid assets tend to have less sticky prices. When there is a shortage of money, liquid assets will tend to fall in value first, and exacerbating the excess demand for money."

[Sergei]

Most goods that are important are flow goods. They are in general liquid but their stickiness depends on stickiness of other flow/stock goods.

Much smaller share of important goods are stock goods. The view on their price stickiness is generally independent from the liquidity view because there is no overlapping dimension in these two measures. Real estate is illiquid on a completely different dimension from its sticky price.

[Lord]

Real estate is interesting. Rents are slow to change and don't generally change that much. Capitalizing its instantaneous value at current interest rates would yield vastly fluctuating prices, but instantaneous value would be next to worthless as a measure. Real estate is bought to be held and someday perhaps sold, so both the value now, value then, and value in between become important. Differing expectations for interest rates, differing windows in time, differing capacities for patience or urgent necessities, all enter, not just for each but for the market as a whole. Time becomes averaged into the cap rate. It is not stickiness as much as differing values for each person and purpose. Both liquidity and time come at a price. What liquidity is desired or required? Who can wait whom out? What is your time horizon? What is the price? What are you willing to pay?

[Becky Hargrove]

I will probably return to this post from time to time, very interesting. One thought about housing: a major part of illiquidity happens because of the loan itself, which has to be maintained as it is created and represents a time intensive investment. Perhaps a solution would be modular housing unit components which the consumer can put together as she has money to do so. This 'lego' house could be created on either rented or personally owned property. Any loss if she needs to pull apart and sell those modular pieces is her own, instead of the bank.

[Jacques René Giguère]

Or the houses made to be easily moved, as they are in New Zealand

http://www.youtube.com/watch?v=ARQPIqU62do

[K]

What Andy says. It's monopolistic competition that permits stickiness. Monopolies also enforce illiquidity. Just try to resell your plane ticket, or phone contract. So both are caused by monopoly power.

[ian mact]

I'm late to the party and I seem to be missing something!

Let's start with a different definition. A market for an asset or good is liquid if it is possible to sell without causing the price to decrease much. A market is illiquid if in order to sell one must accept a significantly lower price. (Buyers take the opposite perspective.)

So price volatility is a sign of illiquidity.

Nick's definition is more or less the Random House Dictionary definition: "the ability or ease with which assets can be converted into cash." But "ease" is very much about price. Because anything can be sold if the price is low enough, and anything can be bought if the price is high enough. Sellers complain that the market is "illiquid" if they can't get a high enough price. It's not that they can't sell, it's that they don't want to take the price on offer.

A corporation may face a liquidity crunch if it can't sell its assets at a high enough price. It's not that the corporation can't sell its assets for any price.

The other point is that it's just not the case that "Some goods are easy to buy and sell quickly" and "Other goods are harder to buy and sell quickly."

In general, most goods are easy to buy quickly and those same goods are hard to sell quickly. (Didn't Nick wonder a few weeks ago why there are more paid salespeople than paid buyers? Selling is hard, even when it's your specialty.) The only exceptions that come to mind are some financial assets. The only examples of assets or goods that are hard to buy quickly yet easier to sell quickly that readily come to mind are rarities or unique items like certain works of art like say the Mona Lisa. Such examples seem irrelevant.

If we think about stock prices, stocks that have fewer market-makers have bigger bid-ask spreads, and more volatile prices. Fewer market-makers is a sign of illiquidity, and the result is price volatility.

[Nick Rowe]

edeast: "The less liquid goods have less traders, not as forced to meet the competitive equilibrium. Prices seem sticky,ie the core is wider than the strict competitive equilibrium in less liquid goods."

Good point. But, as long as we are not outside the core, we could say that those prices may be sticky, but that stickiness doesn't have any allocative consequences. It isn't the same as Keynesian stickiness, where you get suboptimal trading.

Andy: "Shouldn't a good New Keynesian think that the underlying variable is market power?"

Yep. Good point. Market power doesn't itself cause stickiness, in NK models, but it does make the profit function smooth near the top, so that small menu costs can cause stickiness and have big consequences.

Sergei: "Most goods that are important are flow goods. They are in general liquid but their stickiness depends on stickiness of other flow/stock goods."

Dunno. Labour (services) is a very important flow good. But labour is very illiquid.

ian: yep. My definitions of liquidity and stickiness were very loose and inadequate. That was partly laziness, and wanting to keep it short. And partly because the concepts are a bit fuzzy. And I wanted to keep them a bit fuzzy. Because maybe there's a correlation between illiquidity and stickiness if you define them one way, and not if you define them another way, which in itself would be interesting.

[reason]

Aren't 2 and 3 the same. Both relate to uncertainty about the market clearing price (and so the buyer and seller perceptions of price can diverge).

[Daniel Solomon]

This link between "illiquidity" and price stickiness brings to mind the good old Barro/Andolfatto critique:
the markets with the stickiest prices are markets where search frictions favour long term contracts (e.g many
services where customer loyalty and switching costs are important, or markets where many buyers only look at
one competing offer so sellers can get away with sticky prices- see for example the new monetarist model of
sticky prices by Head, Liu and Wright) so that prices don't move but quantities may not move much either. And
when quantities do move it's because there's no price at which it's worth keeping the trading relation under the new bad aggregate conditions. Or prices adjust at the margin to try to preserve those trading relations that are worth keeping. In such a world individual prices can be sticky, but the aggregate price level may behave much more like that in a flexible price model than a Keynesian sticky price model. Which doesn't mean the outcomes are optimal: there are by now many models of credit and expectations driven recessions in flexible price economies. But the rational inattention theory of price stickiness sounds plausible to me. It's strange in macro how you can look at the same real world observations and see multiple equally plausible interpretations with quite different policy implications, with econometric evidence never being decisive enough. Maybe like string theory, except more people care about the conclusions?