Jim,

Your last post does a really impressive job of explaining both the theory and results of my Journal of Law and Economics paper (19 J.L. Econ. & Org. 517) with Eric Helland. Thanks!

Lawyers do find our proposition that restricting contingency fees will cause lawyers to turn towards hourly compensation difficult to accept. Part of the problem, however, is that Helland and I should have spelled out more clearly that by "hourly compensation" we simply mean non-contingent compensation. Hourly compensation doesn't have to be literal payment by the hour, instead it can be cost-sharing. In legal cases of all kinds there are costs (deposition costs, discovery costs, witness cost etc.) Lawyers paid on contingent fee sometimes bear these costs but they need not do so and there is quite a bit of flexibility in how these costs are assigned. (This is another reason, by the way, why the apparent inflexibility of the nominal contingency fee near 1/3rd can actually mask fairly flexible total payments.) Is it so hard to believe that when contingent fees are restricted there will be greater cost sharing? I don't think so.

Now let's turn to your alternative theory. Once again, I am impressed. Your comments are the most sophisticated that we have received on this paper. I found especially useful how you laid out the order of pleading, discovery, judgement, this all makes a lot of sense. Indeed, I agree that the theory you outline can explain the same results that we presented in our paper.

Thus, we have two theories that can explain the same set of results. What do we do now? I am tempted to say that at this stage a lawyer rests but an economist gets to work! The two theories have the same implication for drops but what else do they predict? Can we find some other testable proposition that the theories disagree upon?

Here is one answer. Our theory says that when contingent fees are restricted lawyers take on bum cases that they would not have taken on under contingent fees (these cases are then dropped as plaintiffs figure out that they are wasting their money). Our theory, therefore, implies that expected awards should decrease when contingent fees are restricted. More bum cases under c. fees thus lower expected awards (lawyer income doesn't necessarily fall since they get more hourly compensation but there are more bum cases so expected awards will be lower.)

Your theory says that when c. fees are restricted lawyers take on the same initial set of cases but then they drop weak cases that they would not have dropped under c. fees. Thus, your theory implies that expected awards should increase when contingent fees are restricted. That may surprise you but I think the reasoning is right. Since it's the weak cases which are being dropped the cases that remain must have higher expected value.

Ok, so what do we find? I went back to our data from Florida but instead of looking at drops I ran the same regression (with all the control variables) on awards. I had not previously run this regression or if I had I did not remember the results.

The regression says that after limits are put into place expected awards decrease by about 30-50 thousand dollars. The decrease is statistically significant.

Well, it's 2:30 in the morning so take the results with a grain of salt but as for me I think I can now sleep peacefully.