(You may , but need not, leave any question blank and receive 20% of the points for that question)

I. Explain briefly, in your own words, what it means in this course to say that a change is an economic improvement--i.e. leads to increased efficiency. Give two examples of ways in which a change might be an economic improvement and yet be regarded by many people as undesirable. (10 points)

It means that if you measure gains and losses to those affected in dollars by willingness to pay and add them all together, the sum is positive.

1. Many people do not accept revealed preference as corresponding to value. They might consider undesirable a change that gave a heroin addict a shot of heroin that he was willing to pay $100 for, at the cost of not producing other goods that other people were willing to pay only $80 for.

2. Many people do not accept the legitimacy of doing interpersonal comparisons in dollars. They might consider undesirable a change that benefitted a wealthy man by $10 and hurt a poor man by $9.

[Answers that depend on the fact that many people care more about their own welfare than that of other people are also legitimate. Note that you were supposed to give two examples of "ways," not two examples of one way, so two examples that both depend on the same problem--for example individual's caring more about their own welfare--do not get full credit.]

II. I summed up Coase's analysis as consisting of three parts: Nothing works. Everything works. It all depends (on transaction costs). Pick one of the three and explain that part of the argument. (10 points)

1: Nothing Works. This is Coase's critique of Pigou: An externality is typically a cost that results from decisions by both parties--A's decision to play loud music plus his roommate B's decision to try to sleep or study at the same time. We do not, in the general case, know which party can eliminate the cost more easily--and the best solution might involve adjustment by both parties--A can play loud music except during exam week. So a legal rule of the sort proposed by Pigou, which makes one party liable (to the other or to the state) for the cost, will only give the right result if the best solution happens to consist of only that party adjusting.

2. Everything Works (aka the Coase Theorem): In a world of zero transaction costs, any initial assignment of rights will lead to an efficient outcome. The reason is that if the outcome were inefficient, there would be some bargain to change it that would benefit all concerned--and in a world of zero transaction costs, such a bargain will get made.

3. It All Depends (on transaction costs): Getting from some initial assignments of rights (for example, the factory may pollute the lake) to some final outcomes (for example, the factory stops polluting) requires a transaction (the resort owners pay the factory to stop polluting). Such transactions may be costly--in the example, if there are many resort owners they face a public good problem in organizing to pay the factory to stop polluting. So in choosing an initial assignment, one should consider not only how likely it is to lead directly to the efficient outcome (in the example, if the efficient outcome were for the factory to pollute and the resorts to adjust) but also how hard it will be to get to other outcomes (no pollution, for example) if they turn out to be the efficient ones, and how likely that is.

III. Answer both parts A and B: (10 points)

A. What is the difference between protecting my ownership of something with a property rule, and protecting it with a liability rule?

Under a property rule someone who wants to use someone else's property must first get the owner's permission. If he doesn't, he suffers a penalty designed to get him to do so--which might be much more than the actual cost he imposes on the owner. If you take my car for the weekend without my permission and get caught you get punished severely--even if I didn't actually plan to use the car that weekend.

Under a liability rule someone who wants to use someone else's property does so, and is then liable to the owner for the cost the user inflicts on the owner by using the property. Under a liability rule, if you take my car (or, more plausibly, run into it) you are liable to me for the cost--the loss to me from not having my car available (or the repair bill if you ran into it).

B. What are the reasons for choosing one or the other?

A liability rule makes more sense the higher the transaction costs of allocating something by voluntary transactions. For instance, it would be impractical to get everyone else in Santa Clara County to agree to let me drive, even though I impose some small risk on them by doing so, before I left my driveway. A liability rule makes less sense the more difficult it is for a court to measure the cost that someone's use of something imposes on the owner.

IV. Answer either part A or part B: (10 points)

A. Expain briefly why property rights are sometimes undesirable--why it is sometimes more efficient to treat things as a commons. Illustrate your explanation with examples from the case of primitive peoples--reasons why it sometimes is and sometimes is not desirable for them to treat land, or other things, as private property.

In order to have property rights, one must define, enforce, and assign them; in order to have property used by someone other than the owner, one requires transactions. All of these activities may be costly, and if their costs are greater than the benefits of having property we are better off with a commons.

For example, it would be very costly, when hunting large animals, to have to stop each time you reached a property boundary and go get the owner's permission; that is a transaction cost, and a reason why primitive peoples who are using their land to hunt large animals across typically do not have property rights in the land--although they may have property rights in the animal (first to chase the animal owns it), and in land at other times of the year. Such societies do not have property rights in live wild animals that are not being chased, because it would be too costly to define and defend them--to keep track of what animal belonged to whom. And the benefit would be small, since there isn't much primitive people can do that affects the availability of wild animals--not much elasticity on the supply side of the problem.

B. Under current law, copyrights are given easily and protect for a long time; patents are much harder to get and protect for a much shorter term. Does this difference make sense? Why or why not?

Yes it makes sense. Copyrights are easier and cheaper to define and enforce, both because it is usually obvious if one work is copied from another and because independent invention is unlikely, so a system of copyrights does not force writers to spend lots of time and energy making sure that someone else has not written their book first. Patents are harder to define and enforce, because ideas don't have sharp boundaries; it may not be clear whether a new invention infringes on an existing patent, which leads both to unintentional infringement and litigation costs. So patents have larger social costs than copyrights, which is a reason to be less willing to grant them.

Furthermore, since independent creation of a writing is unlikely, someone who writes a book and copyrights it does not make other writers worse off by significantly reducing their opportunities, so his revenue from his copyright does not overstate the social value of what he has produced, so it does not give him an inefficiently strong incentive to write the book. Independent invention is much more likely, so part of the revenue from a patent may represent, not net social value, but a transfer from others who would have invented the idea themselves--and that becomes more true the longer the patent lasts. In the extreme case of a patent race, the winner gets all of the benefit of the invention even though the loser would have invented it only a little later--giving both parties an inefficiently high incentive to invent as fast as possible.

One further argument is that modern technology makes copying of writings very cheap, so the lack of copyright protection would make it very much harder for the original author to make money from his writing. This is somewhat less true of patents.

V. Apple is contracting with Motorola for the provision of special computer chips for the next generation of Apple Computers. Among the possible problems they anticipate are: (10 points)

1. Designing the chips might take longer and cost more than Motorola currently expects.

2. Apple might change its plans while Motorola is designing the chips, requiring expensive modifications in the design.

The contract can specify which party bears the risk in each case--whether, for example, in case 1, Motorola is required to stick to the previously agreed price, or may revise the contract to pass on the additional costs to Apple.

A. You are an Apple attorney helping to draft the contract. Which risks should you assign to Motorola? Which risks should you assign to Apple? Why?

You should assign risk 1 to Motorola and risk 2 to Apple. In each case you are assigning the risk to the party that best knows it, thus reducing the problem of adverse selection, and to the party that is in the best position to control the risk--and will if that party bears the resulting cost.

B. Your law partner asks you what you are doing. You reply that you are trying to draft an efficient contract. He asks "Why? Shouldn't we just draw up whatever terms for dealing with such risks are most favorable to our client?" How do you answer him?

By drafting a more efficient contract you increase the combined gain to both parties from agreeing to it and thus the amount that your client can collect in exchange for its agreement. Terms that inefficiently benefit your client are likely to cost him more elsewhere in the agreement--for example in the price--than they are worth to him.