Arnold Kling  

Price Discrimination Explains...Macro?

PRINT
My First Look at Strictly C... Harry & Teddy and High Inflati...

In chapter 10 of Macroeconomic Patterns and Stories, Ed Leamer writes,


in recessions, the most price sensitive customers drop by the wayside, and what remains are the wealthy who don't care much about price. Facing this kind of customer base, firms have little incentive to cut price.

He is trying to tell a story to explain the old macroeconomics conundrum. Why do recessions manifest themselves in terms of quantities rather than prices? Why don't prices adjust to clear markets?

When I wrote that Price Discrimination Explains Everything, it did not occur to me that it explains macro. In that post, I suggested that stores encounter shoppers on Black Friday who are more price-sensitive than customers on other days, so they lower they offer sale prices on Black Friday. Here, Leamer is suggesting that stores have their least price-sensitive customers during recessions, so they do not lower their prices. I worry that this is a just-so story. In recessions, aren't non-rich customers likely to become more price-sensitive?

If you don't like that story, Leamer has some others to offer.


Most of us love our homes, and have a hard time dealing with the idea that the market value is less than we think. To adjust to a reduced market price can be tough emotionally.

The result if that in a housing downturn, the volume of transactions plummets. With flexible prices, you would expect to see the same volume--just at lower prices. Leamer offers a similar argument for why people will not accept wage reductions.

A different Leamer story is that employment is closer to a relationship than to a market. When you are making hiring and firing decisions, small differences in the wage rate do not mean much. If you really don't want to go out to eat with a guy, how much difference would it make if he offered to split the check 60-40 instead of 50-50? By the same token, if you really do not want to hire someone, how much difference would it make if that person would take a 10 percent lower wage?

In a way, this is an "animal spirits" theory of employment. Keynes argued that investment was not very sensitive to the interest rate, because the "state of long-run expectations" of the entrepreneur was more important. Similarly, if labor is capital, then the state of long-run expectations determines your willingness to hire, and hiring is not very sensitive to the wage rate.


Comments and Sharing





COMMENTS (10 to date)
Doc Merlin writes:

Hrm, if money is neutral-ish, then it really depends entirely on what the Fed is targeting. If they are targeting prices then quantities will adjust, if they are targeting quantities (RGDP) then prices will adjust.

agnostic writes:

I'm just winging this, haven't thought it out much. But is there an analogy to a host that gets sudden information that he's infected with food poisoning, and keeping such food that he just at at bay for a long time? It doesn't matter if you lower the price of the food -- he perceives it as a likely poison, and won't take it in. Only after a long time passes does he regain his trust of this food (perhaps never, though).

Or look at a society that's taking in immigrants and learns that they're bringing in an infectious epidemic disease -- like the native Americans upon the arrival of Europeans and Africans. They would try to expel them and keep them out. It wouldn't matter if the would-be immigrants offered to lower the price to the host society of their joining -- the host society sees them as too great a risk, and puts them under quarantine (in the literal sense), or doesn't let them in ever again.

These are both reductions in the quantity of something taken in by the host, and are blind to the "price" of taking in a unit. For things that could really undo us, we don't care about tiny variations in the price of taking them in, but rather focus on the very large magnitude of damage that could be done if something should go wrong Black Swan-style.

I think we have this kind of (over?) reaction, i.e. keeping down the quantity we let in, when we realize we've let too much bad stuff in for awhile, no matter what the domain is -- food poisoning, epidemics from abroad, a spy or traitor who got close to you when you were trying to enlarge your social circle as much as possible, etc.

agnostic writes:

To make the connection to housing downturns explicit, the pool of potential home-buyers are the host. When a supposedly perfectly safe asset like a house goes into a meltdown, the host views the pool of houses for sale like a crop of apples that news reports say have been stricken by some infectious blight.

They are wary of taking in a house on the market, no matter if the sellers docked the prices low enough to clear the market. Knowing this to be the mindset of potential buyers, the sellers don't bother to reduce their prices in the first place and sell to those who are more risk-taking -- those who would buy an apple from a rotten crop, i.e. someone with unusually robust health and excellent health coverage. Or in this case, someone who's incredibly wealthy and protected in case this one house doesn't perform so well.

PrometheeFeu writes:

I think that view of employment has something to it in some cases. If you are making widgets, most likely your workforce is basically homogeneous. Your workers follow a process to make the widget and you know how long it takes of their time to make X widgets. From there and knowing the sales price of widget and the cost of your other inputs you can make a simple calculation as to how much you should pay them and offer that. But what if you hire someone to invent the next widget? Even if you can calculate the future discounted value of that invention and make that the team budget, how can you calculate the contribution of an engineer of varying skill? Basically you can't. So most likely you set the budget for the post and from the pool of candidates who are within the budget you take the best. And you won't go for the second best even if he offers to take a 10% pay cut, because you can't calculate the difference in value between first and second best. The two markets should behave very differently.

agnostic writes:

OK, I took a nap and let my mind wander, and I solved the "why quantities instead of prices" problem.

I already showed why potential home-buyers are unwilling to buy at just about any price -- they see houses as an infected crop, and should the house they buy turn out to be a Black Swan, the magnitude of the blow-up is not forecastable. Buying a certain home could wipe the buyer out for life. Because the magnitude of the blow-up is not calculable, the seller cannot lower the price of the home in order to "compensate for expected damages" -- since there is no typical magnitude for such a blow-up. So, prices don't adjust and only quantities do.

The challenge is to apply this to firms who could hire a bunch more un/under-employed workers. Unlike a single apple from a rotten crop, a single immigrant from an infected population, or a single house from a toxic class, how can a single new worker -- whose value is a lot less than you previously thought -- threaten to undo the firm?

But when we're in a recession, we're not talking about adding a single worker who walks in off the street as during business as usual. We're talking about a firm hiring a large pool of under-employed workers in order to bring unemployment way down. Let's just say the decision is over whether or not to add 1000 new workers within a year.

The worry is that, having received news that a lot of workers out there are worth a lot less to companies than the companies had thought, this group of 1000 might turn out to be a rotten crop for whatever reason. What's the big deal, you might ask -- couldn't you just get rid of them?

Assume even a libertarian utopia: they sign binding contracts that they won't join a union, waive their right to sue if they're cut loose, and just to make this a real gimme, that they'll pay back whatever costs the firm incurred in hiring and keeping them on for the brief period before they were let go.

Cont. below...

agnostic writes:

*Even then*, your firm's bottom line will still take a huge Black Swan hit, i.e. one whose magnitude cannot be estimated. Why?

Because of convexity in the damage to your reputation done by firing X workers. If X is 1-5, you basically suffer no damage -- a single worker is liquidated every week or month, and no one thinks worse of your firm for it.

But if you discover that these 1000 new workers are a rotten crop, you have to liquidate all of them at once. Dumping 1000 workers at once makes investors, analysts, consumers of news, consumers of your product, etc., all panic like it's a five-alarm fire. Or at any rate, the damage done to your reputation is a lot more than 1000 times the damage done by shedding 1-5 workers. The huge damage to your reputation / trust / confidence will kill your stock price. And again the sky is the limit for this damage, so the workers cannot estimate the "expected damages" that they'll stake should they turn out to be rotten.

There could also be non-linearities in the transaction costs of liquidating 1000 vs. 1-5 workers at a single stroke. But I don't see how right away -- in this libertarian utopia, you're just dumping them out on the street, not trying to sell them to another buyer all at once. The big factor seems to be the non-linear, convex damage done to your reputation when Bloomberg News announces that your small business just cut loose X workers.

So because the group of 1000 workers can't estimate the magnitude of damage done to the potential hirer should the group turn out to be rotten, the hirer will refrain from hiring them at any price. Thus only quantities of labor are affected, not the price.

agnostic writes:

Last, obviously the real world brings even more non-linear, convex damages to firing X workers. If you fire 1-5 at once, they may not feel they have strength in numbers to get back at you (through sabotage, class action lawsuit, calling union friends, etc.). But if you fire 1000 at once, they may feel emboldened enough to take you to the cleaners. This damage will be more than 1000 times the damage done by firing only 1-5, and it has no upper-bound or predictability to it (especially if we're talking damages awarded in a lawsuit).

Think about the damages done by punishment-minded consumers who see you downsizing X workers. If you fire 1-5 at once, they don't feel very vindictive -- they probably won't punish you at all. But if you fire 1000 at once, now the damages are a lot more than 1000 times the damages done by firing a handful (which was probably 0 to begin with).

"You can't throw that many people out on the street all at once. It's like your some dictator going on a massive, impulsive purge."

So they organize a boycott to punish you. Again there is no typical level of damages done by a boycott, where the sky's the limit -- maybe it'll be a tiny hit to your bottom line, and maybe it'll wipe you out because people have been galvanized to see you as though you trafficked in human flesh.

These other real-world non-linear, convex damages from firing X workers just make the basic problem of investors panicking about your reputation all the more severe.

Hyena writes:

Wages are a prestige issue. If a company announces that all employees will take the same percentage deduction, people are much more amenable to it because they know their coworkers will not rise in status against them.

I mean, you did take into account the fact that you are usually forbidden from discussing your wages, right?

azmyth writes:

Nick Rowe wrote a related article you might find interesting (link here).

mattmc writes:

"Similarly, if labor is capital, then the state of long-run expectations determines your willingness to hire, and hiring is not very sensitive to the wage rate."

Except that we do decide to hire people where it is cheaper to hire them. (Or all macro questions naturally scoped to a country's borders?)

Comments for this entry have been closed
Return to top