Wednesday, September 4, 2013

Why the (Stiglerian version of the) naive Coase theorem?

David Henderson has a nice tribute to Coase in the Wall Street Journal. His treatment of the more naïve versions of Coase is very clear and I feel like it makes the point I was making the other day - that the Stigler version isn't really wrong (it's actually a corollary of the version Coase was interested in), it's just not the interesting case, and certainly not the realistic case. David writes:
"George Stigler, who won the 1982 Nobel Prize in economics, thought the exciting part of Coase's insight was what happened when transactions costs were zero. Stigler labeled that insight the "Coase Theorem." If transactions costs were zero, no government intervention was needed.

Deirdre McCloskey, an economist at the University of Illinois at Chicago Circle, thought the no-transactions-cost insight was trivial. The interesting part, according to Ms. McCloskey, is what happens when reality intrudes in the form of positive transactions costs. Then it matters how courts assign liability. If, in the above example, a court gave a rancher the right but the rancher valued the steer at less than the damages to the farmer's crop, transactions costs could prevent the efficient solution from emerging.

Coase himself rejected both the Stigler view (that zero transactions costs are the important case) and the Pigou view, both of which he derisively called "blackboard economics."
But I find the last part, where Pigou comes back into the discussion, a little odd. That made me turn over why Stigler might have presented the version of Coase that he did. One possible reason, it seems to me, is that, the high transaction cost world where liability matters and bargaining may be problematic is to a large extent a Pigovian world, and Stigler obviously knew that Coase was trying to add to - if not completely overthrow Pigou.

When you get Pigou on externalities the goal is to internalize the externalities. The very definition of the problem offers the first solution to internalizing the externality: rearrange property rights. The whole problem is that I don't have rights associated with the third party harm I'm experiencing, so clearly if I had rights the situation would be different. Pigou adds that there are public policies that can approximate this solution in circumstances where assigning property rights isn't feasible.

What Coase really adds that's interesting is the possibility of bargaining and the conditions for the feasibility of bargaining (reasonably low transaction costs). The high transaction cost world (let's just say infinite transaction cost world for the sake of argument) is the Pigou world. We have property rights solutions (or, what's equivalent, liability) and some public policy solutions. The real expansion of our vision from Coase had to do with the world where transaction costs are sufficiently low that private arrangements are possible.

Although it's not really what Coase thought was most important, it doesn't seem all that surprising that Stigler chose to focus on this side of the discussion. Although that was not Coase's interest that's where a lot of the new insights came in (and as transaction costs increased of course we only gradually scale back the real world viability of those insights). Coase also has implications for trying to influence transaction costs themselves, of course.

This still leaves open the point I was hitting on a lot yesterday about normative assessments. We may have strong ethical reactions about who has standing (no one invokes Coase, for example, to argue that a rape victim should consider paying off her rapist before the law gets involved - the transaction costs are low, after all, so a bargain could be struck).

Why is this still so contentions?

There are three broad answers to externalities I mentioned above (I am assuming that most economists reject quantity restrictions because of insurmountable knowledge problems):

(1.) reassignment of property rights,
(2.) taxes/subsidies, and
(3.) bargaining between parties

A couple factors help us determine which is best. First, of course, is how easy it is to reassign property rights. Transaction costs also matter. Whether or not we consider the problem legitimately reciprocal matters. Public choice issues matter (you can think of these as political transaction costs).

I think of (and I think most economists think) that in 1920 when Pigou published his book we were working with answers (1.) and (2.). That is the Pigovian world. Coase adds (3.) so the full picture we have now has all three. I get the impression from wading through posts and comments that there are many libertarians out there who think that the Pigovian world is just number 2 and that we didn't have numbers 1 or 3 until Coase.

If you actually think that, of course, it's much more understandable why you might have big fights between Stigler and McCloskey.

I think McCloskey is certainly right in addressing what Coase thought. I think Stigler has a point in his description of how the landscape of economic science changed after Coase.

1 comment:

  1. An addendum - I disagree with McCloskey that the zero transaction cost case is "trivial" too. It's less interesting precisely because it's not realistic or descriptive of the real world, but it's still a non-trivial bit of economics to get your head around. It certainly wasn't commonly discussed before Coase, which seems to me to be a good indication that it was not "trivial".

    But of course I agree with her on what the most interesting point is and where a lot of our focus should be.

    The zero transaction cost world is excellent for pedagogical purposes - to get the logic across of another solution to the externality problem.


All anonymous comments will be deleted. Consistent pseudonyms are fine.