I was fortunate enough to spend part of last Summer at the Duke Center for the History of Political Economy Summer Institute studying the history of economic thought. For one class session, Steve Medema lectured on the Coase theorem. Medema is probably the foremost historian of law and economics in general, and of Ronald Coase in particular (see work here). The Coase theorem* (which I’ve written about a bit before) states that in a world with absolutely no transaction costs, the assignment of property rights does not affect the efficiency of an outcome. That is, in a world where bargaining is completely costless and instantaneous, and where all actors have complete and perfect information, it doesn’t matter who initially has a property rights claim, the most efficient outcome will obtain. The classic example is quoted by David Frum in a recent blog post:

A famous Coase example is the following. Consider a railroad that passes through wheat fields. The passing trains let off sparks which can burn the wheat. If the legal rights are on the side of the farmers, then they could require the trains to buy and install spark catchers to eliminate these fires. However, if that is expensive (i.e. more than the value of the burned wheat), the train owners may just pay the farmers for the damage done to the crops. If the legal rights are with the trains, the farmers may just put up with burned crops or (if that is expensive) they could pay the trains to put on spark catchers. Either way, the socially efficient outcome (install spark catchers or burn crops) is what happens and the legal rights just determine who has to pay.

So, to reiterate: in a world without transaction costs, the assignment of property rights does not affect the efficiency of the outcome (but it does affect who actually pays, and thus the inequality/fairness of the outcome).

Pretty great, right? Here’s the thing: the world is chock full of transaction costs. The point of the Coase theorem, following Medema’s lead, is that because transaction costs do exist, the assignment of property rights does matter. Coase’s research program calls for in-depth case studies to determine whether or not particular institutional arrangements (allocations of property rights) are efficient, and when the are not, to recommend changes. The Coase theorem functions as the rallying cry or starting point: it lays out the conditions under which property rights don’t matter to show how absurd those conditions are.

Now let’s get back to David Frum and the example of charity auctions. Frum argues:

The local humanities council approaches Ayers/Dohrn: we’d like to auction off a dinner cooked by the two of you.

Or perhaps it’s the Global Poverty Project that asks Rupert Murdoch to auction off a lunch in his company.
…
Why don’t the targets here just say: “How much do you expect to raise from this dinner/lunch/donation?” Almost invariably, the amount raised is less than the target would pay not to have do what’s being asked. Why don’t Ayers/Murdoch/the ski chalet donate the money themselves to avoid doing what’s asked?

Here’s a fun game: spot all the places where transaction costs might get in the way of the “Coasian bargain” that Frum proposes to solve the charity auction problem. To start with, the targets of the donation request probably don’t know how much their donation will raise. The charity probably doesn’t as well. Figuring that out would take time and resources – especially if we take any indirect publicity effects into account (e.g. a high-profile donation-in-kind might increase bids on other products).

The point is not that Frum’s solution is a bad one, and I am sure some high profile donors do respond with money when asked for favors**. Rather, I just want to point out that the Coase Theorem doesn’t solve the problem because, as always, we live in a world full of transaction costs. The Coase Theorem should be a starting point for institutional analysis, not a rhetorical stick used to bludgeon bits of the world that seem irrational to scholars and critics who believe just a bit too strongly in the perfection of actually existing markets.

* Note that Coase himself did not refer to it as a the Coase Theorem, or a theorem of any sort. I think that coinage dates to a textbook by Stigler in the 1970s.
** Though, as a sociologist and not an economist, I will point out that there is a lot of research on how cash vs. in-kind gifts sometimes work very differently (see Viviana Zelizer’s work on the meaning of money, among many others). Also, in some ways, Frum’s use of the Coase Theorem here is a bit weird, as property rights claims are noticeably absent in the situation of requesting donations. But I’ll follow Frum and focus on the possibility of efficient side bargains, and ignore the context switch.