Economics for Accountants (Part 1)

I gave a short talk a couple of weeks ago to some of the MAcc courses here at the U of U. The Masters of Accounting is a professional degree that prepares folks for a career in auditing and public accounting. It's one step on the path to a CPA. (And by the way, demand for CPAs remains strong. The MAcc is a great degree, and our School of Accounting does a super job with it.)

I was talking about the economy generally and the complex chain of events that led us into the deep recession we're now in. But I only had an hour and twenty minutes. So I didn't get to do everything I wanted to do.

One of the key ideas for understanding various aspects of the crisis is called "The Market for Lemons." This is one of the very most important ideas in economics. George Akerlof won the 2001 Nobel Prize for it. Put it in the same "really important ideas" bin as Hayek's notion that prices convey information.

The Market for Lemons doesn't just help us understand the crisis --- it actually explains why financial accounting exists in the first place. So it's an important idea for accountants (and everyone else, for that matter) to get a grasp on.

I was hoping to explain it to the MAcc students, but didn't have time. I'll explain it here instead, and hope that they'll come read. It's a big idea with lots of moving parts, so it will probably take me a few posts to get everything out.

So here's the one sentence version: "Asymmetric information can lead to market failure."

And here an example showing what that means.

Let's consider the market for used cars. We have two buyers and two sellers. One of the sellers has a really nice car to sell --- great shape and plus it runs good. One of the sellers has a junker: dents all over and maybe it needs a transmission.

The seller of the "plum" car is willing to sell as long as she gets at least $6000 in exchange for the car. The lemon seller is willing to sell as long as she gets at least $2000.

There are buyers out there as well. Buyers like plums, of course, but they're willing to buy a lemon if the price is right. Buyers are willing to pay up to $7000 for a plum, and up to $3000 for a lemon.

Because the most the buyer is willing to pay is bigger than the least the seller is willing to accept, there are gains from trade. That is, one buyer can offer the lemon seller, say, $2500 for the lemon. The buyer thinks "I'd be willing to pay $3000, so if I get it for $2500, that's giving me $500 worth of surplus." The seller things "I'd be willing to sell for $2000, so if I sell it for $2500, I get $500 of surplus." Both buyer and seller are made better off through this trade. Same for the plum; a price of $6500 makes both better off.

Note what happened here: In each case, the car ended up with the person who valued it the most. The lemon buyer values the car at $3000, while the seller values it at $2000. The buyer ends up buying it. The plum buyer values the car at $7000, while the seller values it at $6000. The buyer ends up buying it. This market worked, in the sense that the cars are put to their highest-value use.

Now I'm going to make one small change to this market interaction, and it's going to mess things up pretty badly.

I'm going to assume that the buyers cannot determine whether a car is a plum or a lemon and that sellers know whether their cars are plums or lemons. To make this more concrete, have a look at this article about post-Hurricane-Katrina flood cars. Turns out that it's pretty easy to take a flood-damaged car and make it run. The problem with flood damage is that it shows up over time. This is exactly the sort of problem that a seller might know about, but where it'd be hard for a buyer to detect prior to purchase. You could imagine the same sort of issue surrounding a faulty transmission, or balky radiator.

Just to throw some jargon at you, these assumptions mean there's asymmetric information. The seller knows something, and the buyer doesn't. I'm also going to assume that the buyer knows he doesn't know, and the seller knows the buyer doesn't know, and the buyer knows the seller knows the buyer doesn't know, and so on. That is, the asymmetry of information is common knowledge.

How does this change things?

To see the problems, we're going to talk through a buyer's thought process about what offer to make to a seller. Recall that this problem was pretty easy in the "symmetric information" case above. The buyer above makes a high offer if the car is a plum, and a low offer if the car is a lemon.

In the "asymmetric information" case, things are more complex. Suppose the buyer sees a car. He doesn't know whether it's a plum or a lemon, but figures there's half a chance of either. What should he offer?

One possible offer is $6500.

If the buyer offers $6500 and the car is a plum, the seller will probably accept the offer. In this case, the buyer ends up happy, because he's just bought a car that he values at $7000. As above, he ends up with $500 in surplus.

If the buyer offers $6500 and the car is a lemon, an unscrupulous seller will.... accept the offer (and secretly scream in joy) In this case the buyer has just purchased a car that he values at $3000 but has paid $6500 for it. This is $3500 in "negative" surplus for the buyer.

Let's summarize: If the buyer offers $6500, he gets a car for sure. Half the time the car is a plum (in which case the buyer gets $500 in surplus) and half the time the car is a lemon (in which case the buyer gets -$3500 in surplus). On average, the buyer gets

(1/2) * 500 + (1/2) * (-3500) = -1500

in surplus. The buyer is better off making no offer than offering $6500!

What about other offers? Maybe the buyer should offer $2500?

If the buyer offers $2500 and the car is a plum, the seller will not accept the offer. No trade happens, and the buyer gets $0 in surplus.

If the buyer offers $2500 and the car is a lemon, the seller will accept. This yield $500 in surplus for the buyer.

Again to summarize: If the buyer offers $2500, he gets a car only if it happens to be a lemon. Half the time the car is a plum (in which case the buyer gets $0 in surplus) and half the time the car is a lemon (in which case the buyer gets $500 in surplus). On average, the buyer gets

(1/2) * 0 + (1/2) * (1500) = 250

in surplus. The buyer is better off when offering $2500 compared to making no offer.

Notice what's going on here. The buyer is afraid to make a high offer, because high offers attract both lemon sellers and plum sellers. Because the buyer won't make a high offer, the plum car doesn't get sold, even though the buyers value the car more than the plum seller. This means there are unrealized gains from trade --- the market has failed to achieve all possible gains from trade.

We call this example "The Market for Lemons" because that market works. It's specifically the market for plums that fails.

Asymmetric information is a big problem in lots of markets. It can cause trouble in labor markets, financial markets, insurance markets, health care markets,... you name it. And while the example I've written down is quite specific, the conclusion of market failure isn't driven by any of the specific assumptions. We could, for example, allow the seller to make offers, let the buyer make multiple offers, assume the seller doesn't know with certainty what kind of car he has (but still has a better idea than the buyer). None of this changes the basic conclusion that asymmetric information can lead to market failure.

In Part 2 of this post, I'll explain why the Market for Lemons explains why financial accounting exists. In Part 3, I'll write about how this idea can explain other "institutions" we observe in the world. Then in Part 4, I'll get back to showing how this idea is useful for explaining various parts of the financial crisis.