Econ 522 – Lecture 10 (Oct 72008)

Before we start, a quick example, not directly related to the course, on how economists look at the world differently than everyone else.

There’s a well-publicized story recently of a 90-year-old woman in Akron, Ohio, Addie Polk, who became sort of a symbol of the current mortgage crisis when she shot herself while a bank was trying to evict her after foreclosing on her home.

Obviously, a tragic story – that someone at that point in her life was being thrown out of her home, and that she was desperate enough to try to kill herself. (She shot herself twice in the upper body, was hospitalized, and lived.)

Friday, I saw an article online that Fannie Mae had decided to give her the house – to forgive the remaining balance on the mortgage, and allow her to keep it.

To most people, a sweet end to a sad story – a big company “doing the right thing” to address a sad situation.

I saw the article as a link from an economics blog I read. It was under the heading “No, no, no!” And the comment after a quote from the article was:

“In other words, the taxpayers are now subsidizing self-injury.”

(In fact, this serves as a kind of example of the incentive problem of bailouts in general. After the fact – once the situation has already occurred – giving her the house may seem quite reasonable, just like rescuing a failing institution (or several) may seem necessary. But in terms of incentives; if people anticipate that they’ll be “rescued” if their business starts to fail, it creates an incentive to take unreasonable risks. Just like if people expect they can keep their house by shooting themselves, it may create an incentive for a behavior we don’t really want to encourage. We’ll have to see whether there are any more foreclosure-related suicide attempts.)

Last Thursday, we asked the question of what promises the law should enforce, and introduced contract law as the attempt to answer that question.

  • We talked about one early attempt to answer the question, the bargain theory of contracts, and some of the problems with it.
  • Under the bargain theory, promises are legally enforceable if they were given as part of a bargain; there are three elements that must be present, offer, acceptance, and consideration
  • We showed an example of an agency game, where my inability to commit to a future action (in that case, returning your investment) led to a breakdown in cooperation…
  • …we said that the first purpose of contract law is to enable cooperation, by turning games with noncooperative solutions into games with cooperative solutions…
  • …and we argued that efficiency generally requires a promise to be enforceable if both the promisor and the promisee wanted it to be enforceablewhen it was made
  • A question I got after class…
  • We saw an example of how asymmetric information can inhibit trade…
  • (the example of you being unable to buy my used car, because I “know too much” about its condition and have no incentive to tell you what’s wrong with it)
  • …and claimed that the second purpose of contract law is to encourage the efficient disclosure of information
  • We discussed the fact that efficiency sometimes requires breaching a contract…
  • …and said that the third purpose of contract law is to secure optimal commitment to performing…
  • …and argued that setting the promisor’s liability equal to the promisee’s benefit – expectation damages – accomplishes this goal
  • We introduced the idea of reliance, that is, investments made by the promisee to increase their benefit from the promise…
  • …and said that the fourth purpose of contract law is to secure the optimal level of reliance…
  • …and then we ran out of time.

Today, I want to go back over the example of efficient breach, since I think I went through that a bit fast… do an example of reliance… then move on to default rules and mandatory rules

We begin with an example of efficient breach. Suppose that I build airplanes, and you contract to buy one from me. You value the airplane at $500,000. We agree on a price of $350,000. It will simplify the example if we assume you paid me up front; so let’s assume this contract was money-for-a-promise: you already paid up front and I promised to deliver a plane. (This doesn’t really matter much, it just makes all the numbers positive.)

The rule for efficient breach is:

If [ Promisor’s Cost ] > [ Promisee’s Benefit ]  Efficient to breach

If [ Promisor’s Cost ] < [ Promisee’s Benefit ]  Efficient to perform

Since the promisee’s benefit is known to be $500,000, it is efficient to perform whenever the cost of building the airplane is below $500,000, and efficient to breach whenever the cost is above $500,000.

Since the promisor only looks at his own private cost and benefit when deciding whether to breach or perform,

If [ Promisor’s Cost ] > [ Liability ]  Promisor will breach

If [ Promisor’s Cost ] < [ Liability ]  Promisor will perform

In the case of perfect expectation damages, the promisor’s liability would be the amount of benefit the promisee would have received, which is $500,000; this leads to the promisor performing whenever the cost of building the airplane is less than $500,000, which is exactly what efficiency would require. Setting the promisor’s liability at any other level would lead to some instances of either inefficient breach (if liability were too low) or inefficient performance (if liability were too high).

(Why? First, suppose liability were $350,000 – I can get out of building you the plane by returning your money. If the cost of building it rises to $400,000, though, now I’ll refuse to build the plane; but it would be efficient for me to build it.

So what, you might ask? If transaction costs are low, we can just negotiate again and agree that you’ll pay me $450,000 for the plane. But if I can get out of a promise just by returning your money, it might be tempting for me to do this too often, just to try to raise the price. Suppose the true cost stays the same; but I know you agreed to pay $350,000, so you must value the plane more highly than that. So I go to you with a made-up story about one of my workers threatening to quit unless he gets a raise, I tell you my costs went up, and you can only have the plane if you’re willing to pay me $400,000. If it’s too easy to get out of a contract, agreements become meaningless. And if transaction costs are high – say, you’re mad because I breached the first contract, and don’t want to deal with me anymore – then you don’t get your plane, even though it would be efficient for me to build it.

On the other hand, suppose it was too hard to get out of a contract. Suppose that contract law specified that if I breach a contract, I owe you $10,000,000. Now my costs actually go up, and it would cost me $1,000,000 to build you a plane. I go to you and say, “look, I know I promised you a plane, but it would cost me $1,000,000 to build, can we just agree to undo the deal?” And you say, “sure, just give me the $1,000,000.” And I say, “Why? You were only going to get $500,000 of benefit out of the plane.” And you say, “Yeah, so what? If you don’t give me a plane, you owe me a lot more than that. Give me $1,000,000, or else.”

What’s the problem with that? Well, in a static world, nothing – me having to pay you $1,000,000 to get out of my promise is annoying to me, but it’s not inefficient. But now go back to when we were originally agreeing to the contract. If I know there’s a small chance my costs will go way up, and if contracts are strong enough that you could do this to me, then maybe I don’t want to take the risk of making the promise in the first place. So now even though it might be efficient for me to agree to build you a plane, I’d be afraid to make that promise.

On the other hand, if the cost to me of breaching the contract – my liability – is exactly equal to your benefit, there is no problem. I’ll still build the plane exactly when it’s efficient for me to build the plane – whether or not transaction costs are low enough that we could renegotiate. We’ll see later that expectation damages doesn’t perfectly solve the problem of efficient signing – being willing to sign the right contract in the first place – but at least it doesn’t make this problem too severe.)

So that’s efficient breach. Next, reliance. You’ll recall that reliance is any investment the promisee makes that increases the value of performance. So you contract to buy my painting, and go buy a frame for it; or you contract to buy an airplane from me, and you start building a hangar.

Since reliance increases the value of the promise to you, it increases my liability for breach under the concept of expectation damages as we’ve defined them. (That is, if I break the promise, I’m responsible for making you as well off as you would have been if I had kept my word; so if you’ve built a hangar, now I have to reimburse you for the value of the plane with a hangar, rather than without a hangar.) So reliance increases my losses under breach. But you don’t take that into account when deciding how much to invest in reliance, so there is no guarantee that the level of reliance will be efficient.

We’ll use the same example – you contract to buy a plane from me. You value the plane at $500,000, and agree to pay $350,000 for it. Let’s assume that this time, the bargain is promise-for-a-promise – you agree to pay on delivery. And there are expectation damages.

Now you have the option of building yourself a hangar. Building a hangar costs $75,000, and increases the value of owning a plane from $500,000 to $600,000.

Suppose that it’s most likely that building the plane will cost me $250,000; but that there’s some probability p that it will instead cost $1,000,000. Clearly, if it costs $1,000,000, I won’t build it; I’ll just breach the contract and accept that I have to pay you damages.

Let’s look at what happens in each case.

First, suppose the cost of the plane is $250,000, so I build it. Our payoffs (in thousands):

If you relied (built the hangar): you get 600 – 75– 350 = 175

I get 350– 250= 100

If you didn’t rely (build)you get 500– 350= 150

I get the same 350 – 250 = 100

Now look at the case where the cost of sheet metal went through the roof and I choose to breach. Assuming I owe perfect expectation damages as we’ve defined them – that is, enough to make you as well off as if I’d performed…

If you relied (built the hangar): Your surplus would have been 600 – 350 = 250 from the plane, so I owe you 250 in damages

And you paid 75 to build the hangar

So you end up with payoff of 175

I get –250 (since I have to pay you 250 in damages)

If you didn’t rely (build)Your surplus would have been 500 – 350 = 150, so I owe you 150 in damages, which is your payoff

I get –150 after paying you damages

So whether or not I perform, you get 175if you relied, 150if you didn’t. So clearly, reliance makes you better off.

But then the question is, is reliance efficient? That depends on how likely I am to breach. If you rely, our combined expected payoffs are

(1–p) (175 + 100) + p (175 – 250) = 275 (1–p) – 75 p = 275 – 350 p

If you didn’t rely, our combined expected payoffs are

(1-p) (150 + 100) + p (150 – 150) = 250 – 250 p

So the total social gain from you building the hangar is

(275 – 350 p) – (250 – 250 p) = 25 – 100 p

So it turns out that when p < ¼, reliance is efficient – it increases our combined payoffs. When p > ¼, reliance is inefficient – it decreases our combined payoffs.

This is indicative of a more general idea: when the probability of breach is low, more reliance tends to be efficient; when the probability of breach is high, less reliance tends to be efficient.

But if my damages cover your benefit whether or not it’s efficient, then you don’t care about the risk of breach – you end up just as well off whether or not I breach. So you’ll clearly choose the higher level of reliance, whether it’s efficient or not. This will sometimes lead to overreliance – more reliance than is efficient.

So how do we fix this? Cooter and Ulen adjust their definition of expectation damages in the following way:

Perfect expectation damages restore the promisee to the level of well-being he would have had, had the promise been kept, and had he relied the optimal amount.

(This is why they attach the word “perfect” to expectation damages)

Thus, the promisee is rewarded for efficient reliance – this increases his payoff from performance of the promise, and also increases his payoff from breach, since it increases the amount of damages he receives. But the promisee is not rewarded for excessive reliance – overreliance – since damages are limited to the benefit he would have received given the optimal level of reliance.

It’s a nice idea, but it seems like it would be very hard in general for a court to determine after the fact what the optimal level of reliance was. (It might also be hard for the promisee to know this, since he may not know the probability of breach.)

What is actually done in practice? One important legal doctrine is that liability is generally limited to a level of reliance that is foreseeable.

  • Reliance is foreseeable if the promisor could reasonably expect the promisee to rely that much under the circumstances
  • Reliance is unforeseeable if it would not be reasonably expected
  • American and British law tend to define overreliance as unforeseeable, and therefore noncompensable.

An example given in the book is a telegraph company failing to transmit a stockbroker’s message, resulting in millions of dollars in losses. The telegraph company could not reasonably expect the stockbroker to rely that heavily on one message, and so would not be liable for the extent of the losses.

Another example: the rich uncle’s nephew, when he was promised a trip around the world, went out and bought “a white silk suit for the tropics and matching diamond belt buckle”. After the uncle refuses to pay for the trip, the nephew sells the suit and belt buckle at a loss, and sues his uncle for the difference. The court might find the silk suit foreseeable reliance, but the diamond belt buckle unforeseeable, and only award him the loss on the suit. (The book points out that “in American law, gift promises are usually enforceable to the extent of reasonable reliance.”)

Reliance is part of the issue in the famous case of Hadley v Baxendale, a precedent-setting English case decided in the 1850s. I give a link to the actual court decision on the syllabus. Here is Cooter and Ulen’s summary of the case:

Hadley owned a gristmill; the main shaft of the mill broke; and Hadley hired a shipping firm where Baxendale worked to transport the shaft for repair. The damaged shaft was the only one in Hadley’s mill, which remained closed awaiting return of the repaired shaft.

The shaft was supposed to be delivered in one day; Baxendale decided to ship it by boat instead of by train, and as a result, it arrived a week late. Hadley sued for the profits he lost during that extra week in which the mill was shut down. Quoting again:

The shipper assumed that Hadley, like most millers, kept a spare shaft. The shipper contended that Hadley did not inform him of the special urgency in getting the shaft repaired. The shipper prevailed in court on the damages issue, and the case subsequently stands for the principle that recovery for breach of contract is limited to foreseeable damages.

The ruling was that the lost profits were not foreseeable – the court specifically listed several circumstances in which a broken crankshaft would not force a mill to shut down – and that Baxendale was only liable for damages he could reasonably have foreseen. However, this isn’t just a question of reliance; part of the issue is that Hadley knew about the urgency of getting the crankshaft fixed quickly, but did not tell Baxendale. We’ll come back to this question of information shortly.

default rules

If transaction costs are 0, then the two sides to a contract could spell out exactly what should occur in every possible contingency – what happens if the cost of sheet metal rises, what happens if my uncle wants my painting, what happens if a shipment is delayed, and so on. This would make contract law much simpler – courts could simply enforce the letter of the contract, since nothing was left unclear.

However, in reality, some circumstances are impossible to foresee; and even if they weren’t, the cost and complexity of writing a contract to deal with every possibility would make perfect contracts unworkable.

Risks or circumstances that aren’t specifically addressed in a contract are called gaps; default rules are rules that the court applies to fill in these gaps.

Gaps can be inadvertent or deliberate. Our contract to sell you my painting might not have addressed my uncle wanting the painting because I didn’t know he was coming to visit, or because I never would have imagined he would be so excited about it. On the other hand, we could have imagined that it was at least possible for the price of raw materials for building an airplane to go up significantly; however, we might have felt it was such a remote risk that it was not worth the time and effort to build it into the contract.

Cooter and Ulen point out the decision to leave a gap or fill it (specifically address a particular contingency) is the difference between the need to allocate a loss after it has occurred (ex post) versus the need to allocate a risk before it becomes a loss (ex ante). In the first case, allocating the risk, the cost of adding it to the contract is definitely incurred; in the second case, allocating a loss that has occurred, the cost of allocating the loss is only incurred when the loss occurs. Thus, it is often rational to leave gaps when the risk is very remote. (On the other hand, it is usually cheaper to allocate a risk ex ante than a loss ex post.)