« Another rant about the Globe and Mail’s coverage of economics | Main | How has the crisis changed the teaching of economics? »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Why is that darn money multiplier so unstable?

westslope: I don't see the relation between the money multiplier and L-data.

Nick, Phelps has a book "Structural Slumps" that is all about what you're talking about here. Basically the natural unemployment rate varies with the liquidity of the labour market.

Maybe the BLS' Job Opening and Labor Turnover Survey (JOLTS)?

"There's something more going on than what is captured in the P-data and Q-data."

I'd agree with your points about liquidity's importance. So would Carl Menger, who had the same beefs you did some 128 years ago. I can't resist quoting:

"But the fact that different goods cannot be exchanged for each other with equal facility was given only scant attention until now. Yet the obvious differences in the marketability of commodities is a phenomenon of such far-reaching practical importance, the success of the economic activity of producers and merchants depending to a very great extent on a correct understanding of the influences here operative, that science cannot, in the long run, avoid an exact investigation of its nature and causes." Principles of Economics, 1871.

As for data, in stock markets we evaluate liquidity through bid-ask spreads, the size of bids and offers, the number of average trades, whether a stock is in the index or not (that adds significantly to liquidity), if there's a course-issuer bid, if stocks are closely held by insiders or not etc.

But you're asking about liquidity data in labour markets, small businesses, etc. That's a tough one. Buildups in inventories might - among other things - indicate a lack of liquidity, especially unplanned buildups. Duration of unemployment (CANSIM 2820047) might work as an indicator of labour's saleability to prospective buyers. Is there an analog for this; duration of time spent trying to fill a position?

Could you interpret how others might interpret the vacancy data differently? That would have been my suggestion. It seems to me that a Beveridge Curve type relation measures this concept decently well. That introduces the question of whether unemployment is "L-data". I think it's safe to assume the unemployed are looking for jobs, so in that sense it is, but it really oughta be combined with people currently in a job and people out of the labor force who are also looking for jobs.

Some measure of vertical integration might potentially be useful. We know from Smith and Stigler that illiquid markets lead to more in-house production. That's just off the top of my head - would that make sense?

In the housing market I was thinking of inventory and sales which may be Q data, selling time Qi/(Qs/period) which is probably L data, though inventory may change with perceived difficulty of selling or desire of selling so Qi may change for more than one reason.

Adam: I think that's closely related, but not quite the same.

Mazda MX6's are rare and therefore less liquid than the more common Honda Civics. They have a higher "natural rate of unemployment" that Civics. The Beveridge curve for MX6's is further away from the origin. But that doesn't mean there's a buyers' market in one and a sellers' market in the other. MX6's are *both* harder to buy *and* harder to sell.

In the spring, it's more of a seller's market for sports cars, and in late fall, when sellers without a garage get desperate to sell, it's more of a buyer's market.

It's coming to me clearer now. There's L-data, which is about the ease of buying and selling in general. I expect that what I am especially interested in, from the business cycle perspective, is a particular subset of L-data, which is more like the relative ease of buying vs selling. Something like the ratio of ease of buying to ease of selling would be some sort of measure of whether there's a buyers' or sellers' market.

Stephen: I'm not sure. Vacancies, job turnover, duration of unemployment, all seem indicative to me of whether there's a buyers' or sellers' market in labour. But it's not always clear cut. High turnover might mean there's a sellers' market, so workers quit because they know it will be easy to find another job. Or it might mean they get laid off quickly. An increase in duration of unemployment might mean that it's become a buyers' market, and harder to find a job. Or it might be because EI has got more generous (or something), so workers are choosing to take longer before accepting a job. I wish my head were clearer on this.

JP: Yep, and Menger's view of money was directly based on this way of viewing the world. I wonder how much we have advanced?

Do people ever talk about buyers' markets or sellers' markets in financial assets? We don't hear those concepts used in normal times, and to me that ties in with the idea that financial assets have very flexible prices, so if a buyers' market ever arose, the price would immediately fall to eliminate the excess supply. But in a financial crisis, when trade dries up, we do see something that looks like a buyer's market?

(Notice by the way that I avoided talking about excess demand and excess supply in the post? Because those are theoretical constructs, and not always applicable, and not directly observable. But a first step towards modeling a buyers' market would be to hold P fixed, and let demand fall or supply increase.)

Daniel: We can at least imagine a world in which, when unemployment is at the natural rate, there are no vacancies, or lots of vacancies. It might be very costly to have a vacancy ("we need a worker right now, or the whole factory shuts down"), or very cheap to have a vacancy ("we need a worker sometime, but no rush, and it's not crucial if we never find one").

Fire engines are a good example. They have an extremely high natural rate of unemployment. Most fire engines are just sitting there, nearly all the time, waiting for a job. And there are very few vacancies for fire engines (fires that need a fire engine), and they get filled within minutes, in a well-run municipality. The natural rate is right up one end of the Beveridge curve.

I think it's safe to assume that nearly all the unemployed are looking for jobs. Stats Canada asks them if they are actively looking, and (with minor exceptions) doesn't count them as unemployed unless they answer "yes". But an increase in the unemployment rate doesn't necessarily mean the labour market has shifted to a buyers' market. It might be structural, or it might be that workers are less desperate to accept a job quickly.

Unemployment can also be viewed as vertical integration and in-house production. The question is whether the unemployed are growing veggies to feed themselves, rather than working in the market and buying veggies with their wages, because the labour market has become a buyer's market, or because there's a negative technology shock.

Underlying this whole post, of course, is my view that the business cycle is some sort of failure of the market mechanism, as opposed to the RBC view. And a deficient demand failure, rather than a recalculation failure. A deficient demand (monetary) theory of the business cycle says that Q and P fall in a recession because most markets have become buyers' markets.

Lord: that's about how I would think about it too.

You know, it's weird. Other people talk about buyers' markets or sellers' markets, but economists rarely do.

"Do people ever talk about buyers' markets or sellers' markets in financial assets?"

One can say that a stock or a market is well-bid (or well-offered). Roughly more bids than offers (or vice versa).

"You know, it's weird. Other people talk about buyers' markets or sellers' markets, but economists rarely do."

I think we do, except we don't use the same phrase. We talk about sticky prices, which is basically the same thing.

Consider a negative demand shock which should lower the equilibrium price. But if prices are sticky, then the market price is above the equilibrium price, leading to Qs > Qd. It's really easy to find a seller if you want to buy the good at the market price, but really difficult to find a buyer at the market price if you want to sell.

Laypeople and economists are talking about the same thing. The two groups just use different languages.

JP: that makes sense. So if a market is "well bid", it's easier to sell a lot than to buy a lot at close to the equilibrium price.

Mike: The way I see it is that excess supply is a *theory* of what causes a buyer's market, rather than an alternative term to describe a buyers' market.

Ahh.. okay. I had to read that 3 times, but I've got the distinction now. :)

Nick: Right. It also applies to bond auctions and such. Although in industry speak, you try to buy/sell near the last price, not the equilibrium price.

Nick:"Do people ever talk about buyers' markets or sellers' markets in financial assets?"

All the time, but usually it's nonsense, ie there is no sensible sense in which there is a bid/ask mismatch. But we do sometimes see markets which are not clearing all the interest due to short selling constraints. Eg my desire to sell a stock is not reflected in its price because I can't borrow it or because the the cost of failing is too high and is asymetric between the longs and shorts. We see this sometimes in bankrupt companies where the stock continues to trade well above $0 despite every market professional knowing that there is 0 chance of any recovery and forward prices being much lower (in backwardation). So the price is formed by a consensus of the bulls and the recently bullish. Like a collapsing housing market this is a good example of a buyers market - bid the market up for size and you will get hammered by sellers who were regretting not having sold at that price previously. ABS/CDO markets similarly failed to clear partly because of short selling constraints on individual securities and because the longs willfully suppressed acknowledgment of their losses. Then there's market corners - especially effective when combined with the short squeeze. VW in '07 became the single most valuable auto maker on the planet for a brief while due to a stealthy corner by Porsche. The price was obviously mad but only the longs could sell it which wasn't enough to balance the desparate short covering. The solution to finding an efficient equilibrium in all these cases is to modify market mechanisms to permit (operationally) risk free shorting.

So there are definitely (sucker) buyer's markets defined by the price being inefficiently high. But what about seller's markets? Not sure. Going to think about it.

And what about non-clearing labour markets? By analogy, if we could short labour we could sell it down quickly to it's clearing price. But in this case the microeconomic solution obviously neglects the role of labour in demand and would bring us away from equilibrium.

Mr. Rowe,

My primary education is in law, not economics (although where I come from all lawyers have to learn basic micro and macro), so I may be completely wrong with this, but if I'm understanding your post correctly, it might help to think of buyers' and sellers' markets (and therefore liquidity) in terms of risk/reward distribution and factors other than price and quantity, at least in some markets.

Relative advantage could be expressed through price, of course, but also through contract terms: When someone takes advantage of a monopoly or high transaction costs, they saddle the counterparty with risks and costs that they themselves would bear in a theoretical optimum. Alternately, to facilitate sales, a company might "sweeten the deal" with extensive guarantees, additional perks, frequent flier miles, or whatever. Such terms are (potentially) not costless, but aren't directly reflected in price. Why wouldn't the same principle hold for buyers' and sellers' markets in general? If prices are sticky, economic actors might be tempted to change these attendant factors instead. In a buyers' market, a company might, for example, offer extensive guarantees for its products at price x. In a sellers' market, conversely, a company might decide to keep price at x but reduce the duration or scope of its guarantees. Or, it might try to saddle the customer with risk of defects (if the legal system permits it). This would probably only hold when a product is not entirely homogenous, though.

Generally, I would attempt to explain it like this: On average, you have a certain frequency of transactions during a time period, and these transactions have certain legal characteristics beyond price. During a sellers' market, the frequency increases, implying higher demand and prices, but that also gives a seller opportunity to negotiate better contract terms (which can reduce frequency, but reduce costs). During a buyers' market, companies selling products at a certain price might increase the frequency of sales in a time period (and sales in general) by offering more attractive contract terms, which can raise their costs, and in that regard has an anologous effect to lower price. That way, you could define a buyers' and sellers' market with deviations from average contract terms (as far as the law permits).

Anyway, I might be completely off with this. Or, it might not be useful on a macro-level, since contract terms are specific to particular markets. But in case this is of some use, I thought it best to post this.

Another angle: If we assume imperfect competition then it makes sense for a rational actor to set price (reducing quantity) to a level where the Courtnot point lies. But, unless demand is completely inelastic, some people will cease to buy a product because of the substitution effect: At that price level other products become more attractive.

Now consider imperfect information. If a company making durable goods, such as TVs, offers a five-year guarantee, that increases its costs, but is advantageous to the customer - in effect, it makes the product better, or, alternately, reduces the asking price: A rational customer who knows that a TV has a 5% chance per year of breaking down and that repairs on average cost x would factor that into his demand if he would have to bear the cost. Now assume imperfect information: A company knows that a TV has an x chance of breaking down every year and that repairs on average cost y. On the other hand, a customer probably does not know the exact chance of a TV breaking down or the average cost of repairs - or he might not care ("I want that awesome TV now!"), so he might underestimate the potential cost he has to bear, in effect overpaying for the TV.

To a company, therefore, it might make sense - to a point - to worsen contract terms in a sellers' market instead of raising the price: It reduces their costs, increasing their profit (which is what raising prices would do), while the customers might underestimate the costs they're being saddled with, paying more than they would if they had all the relevant information, with a similar effect to misreading the price. That could allow a company to dodge a part of the substitution effect, knowingly or not. Therefore, it might make sense to look for sellers' markets by looking for a worsening of contract terms.

T: You are partly off=track, but not completely.

If demand falls, or supply increases, there's an excess supply (a buyer's market) at existing prices. In response to that excess supply, sellers might cut price, or they might, as you suggest, sweeten the deal by changing other non-price attributes. We can think of those sweeteners as either a quality improvement, or as a hidden price cut. But a better way to think of it is that the whole market upward-sloping curve of price against quality shifts down. So you pay a lower price for the same quality, or get a better quality for the same price. Same thing. And Statistics Canada, in its CPI data, for example, does at least try to adjust the prices they report for quality changes (though it probably fails, because it is hard to do this).

So, my first response to you would be to say "No, that's just a different type of price cut".

But I think my first response would leave something out. Why is it that firms offer sweeteners rather than cut price? And might those sweeteners be at least correlated with the sort of L-data I'm talking about? That's my second response: you might be on to something.

Third response: What's the difference between a cut in prices and a line-up of sellers? Answer: a cut in prices redistributes income from seller to buyer; a line-up is a waste of resources. In other words, a price cut is a non-wasteful way of eliminating an excess supply. A big line-up of unsuccessful sellers, each trying to be first in line to sell, is a purely wasteful way of "eliminating" the excess supply. And many of the sweeteners you are talking about may be partially wasteful ways of eliminating the excess supply, because the sweetener is worth less to the buyer than it costs the seller. So your sweeteners may be halfway between P-data and L-data.

I have an off topic question:

I have been interested in macro micro-foundations for a little while, particularly the "monetary equilibrium" approach. Macroeconomists in general don't seem to talk about micro-foundations very much. Are there other kinds of micro-foundations? Why don't macroeconomists talk in terms of micro-foundations more?

jsalvati: quick and dirty answer: you don't talk about something when everybody is doing it. You only talk about things that are new, and controversial. There's a lot more microfoundations of macro today than 40 years ago, so there's less talk about it.

Nick: “But we really need to distinguish the ease of buying from the ease of selling. They aren't the same.”

and

“I expect that what I am especially interested in, from the business cycle perspective, is a particular subset of L-data, which is more like the relative ease of buying vs selling. Something like the ratio of ease of buying to ease of selling would be some sort of measure of whether there's a buyers' or sellers' market.”

This may be pedantic, maybe not. Yes, the ease of buying and selling aren't the same. But that's because we are talking about 2 different assets. The ease of buying (say a house) is really just the ease of selling cash (or some other medium of exchange that has been earmarked for the purpose of buying a house).

A bad buyer's market in houses is really just a bad seller's market in cash. You say that it's a buyers' market for labour nowadays, but I'd call it a good seller's market for cash (and all the other mediums of exchange used to attract workers).

There are no buyer's and seller's markets. Everyone is a seller, and each of these sellers’ markets is either good or bad. Or everyone is a buyer, and each of these buyers’ markets is either good or bad.

In a recession, a number of items become easier to sell (like cash, low priced necessities, survival gear) while other items become harder to sell (like cars, luxuries).

The key ratio you are interested in is the ease of selling some good x to the ease of selling cash (or some other medium of exchange) that been earmarked for purchasing x. You can measure the former easily by watching how long it takes for someone to sell x. ie. How long has your house been on the market? But it would be very difficult to watch how long it takes for someone to sell the cash they have specifically earmarked for x. How long has your cash been on the market? An individual's store of cash and other mediums of exchange are homogenous and can be earmarked for so many different purchases.

K: I don't think that's totally true. Talking about a buyers and sellers market makes a good deal of sense for market making activities. The role the market makers is to adjust prices until the market is no longer a buyer or sellers market.


I think the asymmetry arises from lopsided scope for price discrimination in markets with pockets of imperfect competition. In a buyers' market, the distribution of sellers' reservation prices is skewed, with some being very low. Most transactions still happen at the "typical" price, but buyers that hold out for a deal might discover a seller with a low reservation price and do very well. In the housing bubble, there were stories of amazing deals for sellers and very few for buyers, while in the current environment this is reversed.

So, measuring the skew of the distribution of (appropriately normalized) transaction prices could be a source of L-Data. Negative skew indicates a buyers' market.

Observing changes in negotiating behavior (i.e. this model predicts more effort to identify high-surplus counterparties and trade with them at favorable prices) may also work.

I think this might arise due to varying risk aversion. Right now the risk (both real and perceived) for some house sellers and some job seekers could be very high, while the risk for the other side is lower and less varied. In the housing bubble, buyers that were "short" housing were the ones with the (sometimes) very high risk aversion (perhaps irrationally responding to missing out on recent rapid appreciation).

Nick,
Would the Fed's Senior Loan Officer Survey and the lending surveys of other central banks count as L-data? How about the the ISM questions on suplier delivery times and order backlogs?

JP: That's not pedantic; it's crucial. Your pedantic point underlies my whole perspective on business cycles. We live in a monetary exchange economy. When we say it's a buyers' market for goods, that means it's a seller's market for money. It's an excess demand for money that is at the root of an excess supply of goods. General gluts are always and everywhere a monetary phenomenon.

(But I wanted to leave my monetarist leanings out of this post).

fmb: neat idea. I don't know it it's right, or how generally it's right, but it sounds like it might be.

Gregor: Good examples. The loan officer survey (the BoC has something similar) asks about the ease of obtaining credit, I think. Order backlogs are like queues, and queues are great L-data. Remember the stories of workers queueing for hours to apply for a job (in the pre-internet days)? That was strong evidence of excess supply of labour, buyers' market.

jsalvati:  There is no such thing as a "buyer's" market due to market makers not yet having noticed that supply is outstripping demand at the current price.  Financial markets, except to extremely sophisticated high-frequency trading systems (these are actually market making systems),  are immune to the kind of inefficiency you are proposing.  If you think market makers keep prices at non-equilibrium levels long enough for you to make a trade, then arbitrage them.  They wont be in business for long. 

fmb:  I think you have it backwards.  In the US over the past 2 years, it's the buyers who have gotten stuffed.  If the market had cleared in '07-'08 prices would have collapsed then to the equilibrium level.  Instead it's taken till now and for all I know, it could keep going for awhile.  So anyone who bought in the "buyer's market" has watched their investment dwindle.  If they had checked out housing forward markets, like the OTC RPX market, they would have seen that they were buying into a market which rational investors expect to decline (in backwardation).  Buyer's market = price *above* equilibrium = sellers are getting the "amazing" deal. 

Nick:  I think forwards/futures, in markets where they exist, would be an excellent indication of buyer's vs seller's markets where prices are sticky.  Backwardation, adjusted for carry costs/convenience yields, would be an indication of a buyer's market; i.e. easy to buy because price above equilibrium.  Contango an indication of a seller's market. We need unemployment futures.

K: Good point. The "amazing" deal is relative, in the context of other prices at the same time, but it does often seem that a "buyer's" market is also characterized by prices above equilibrium. It might be the case that every buyer gets a *bad* deal, just some are much less bad than others.

The same hypothesized skew in transaction prices could arise: a price above equilibrium is much harder to push up but might occasionally move down a lot as a careful buyer successfully negotiates a trade close to equilibrium with a desparate counterparty.

K: "Buyer's market = price *above* equilibrium = sellers are getting the "amazing" deal."

Yes, good point. If it's easy to buy something, maybe you are overpaying.

Nick, Excellent post. But I have a thought about unemployment. Couldn't you argue that employment is Q data but unemployment is L data? After all, unemployment (at least in the US), isn't really people who are unemployed, it's people who are unemployed and want to be employed.

Thanks Scott! Employment is clearly Q-data. I tend to interpret unemployment as L-data. Unemployment is like a line-up for jobs, and the higher is unemployment, the harder it is to get jobs. Trouble is, we can't say for certain if an increased unemployment rate means it's harder to get a job or if people are just pickier or less desperate to accept a job.

fmb: now that I have gotten past my point about who is getting a good deal, I am able to understand your point about skew; I think it's a good one. And I think it matches perfectly the psychology of a market that is slowly trending towards equilibrium.

This thread has clarified my thoughts about sticky prices. There may be a million technical explanations of the microstructure of any particular sticky price, but fundamentally, there is no reason to expect prices to behave like martingales where there is no arbitrage mechanism to bring them to equilibrium instantly. At the very least, this requires the ability to short sell. But if we can't have that, the next best thing is a futures market that can point us to equilibrium and help us get there quickly by eg allowing buyers to purchase forward at a discount and thereby dropping out of the spot market.

Nick: The slope of an unemployment futures curve would be an excellent indication of to whether unemployment is caused by a drop in supply or demand.

Correction: it's a wage future contract that's needed; not an unemployment future.

K: "Backwardation, adjusted for carry costs/convenience yields, would be an indication of a buyer's market"

I wasn't sure what you meant by this. You seem to be talking about a market already in backwardation. If the influence of the convenience yield is somehow removed from the equation, and the contract is still in backwardation after this adjustment, then you've got a buyers market?

JP Koning: Yes, that's what I meant. Ie, adjusted for rational explanations (rates, carry, convenience, etc) the curve is still not flat. So it must be sticky. And if it's sloping down, the price is above it's equilibrium value, which implies a buyer's market.

I'm having problems getting my head around this because it merges monetary theory with futures theory, and they're both already tough to understand. Keynes followed this route; he was into both thinking about money and futures speculation. Maybe there are some hints there.

Hang on with the future market guys. Remember, there's nothing to say that the price of (say) fresh strawberries will follow a martingale. There's no reason to expect the futures price to equal the current price, on average, for every month of the year. Some stuff isn't costlessly storable.

Nick:"Some stuff isn't costlessly storable"

yup. That's why I said "adjusted for carry costs/convenience yields". You'd have to estimate those somehow. For houses we have a pretty good idea from rents, mortgage rates etc. RPX swap prices were indicating 25% annual drop in housing prices at points over the last 3 years, way more than can be accounted for by convenience yield. The only explanation is that prices are sticky and that forwards are pointing the way towards equilibrium. The only agents who were capable of "arbitraging" the market were buyers who could have dropped out of the spot market and bought forward. In the case of macro variables like unemployment I'll admit it gets more tricky. I'm going to have to give it some more thought. Sorry.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad