Tort Reform and Tradeoffs [warning: long post]

As I was reading Dwight Meredith’s latest interesting discussion of tort reform, lead poisoning and lead-based paint came to mind. I was trying to picture the reaction that courts in the 1960s would have had to hypothetical plaintiffs alleging that, “the paint in our house is causing our childrens’ learning disabilities and seizures.” That certainly sounds frivolous, but of course the allegation is true. The point is that frivolity is a lot easier to discern with hindsight. Proscribing, ex-ante, seemingly silly claims puts a lot of potentially valid claims at risk.

There are two categories of errors in scientific hypothesis testing: a false negative (rejecting a true hypothesis), and a false positive (failing to reject a false hypothesis). These are called, respectively, Type I and Type II errors (there’s a third, more widely committed error, the “Type III error”, which is forgetting which error is Type I and which is Type II).

Limiting the scope of pursuable lawsuits will undoubtedly decrease the number of false positives (less unmeritorious lawsuits will be won), but it will come at the expense of more false negatives (lawsuits that should be filed and won because the claims are valid will not be filed and won). As statisticians well know, there’s generally a tradeoff: as the odds of a false negative decrease (the “significance level” of the test increases), the odds of a false positive (the “β” of the test; 1-β is the “power” of a test) increase, and vice-versa. Assuming sound statistical techniques are being employed, the only way to simultaneously decrease the probability of both types of errors is to add more observations.

Here’s an extreme example that illustrates the tradeoff in the context of lawsuits: if we want the rate of successful frivolous lawsuits to be zero (which would presumably dramatically reduce the filing of such lawsuits), we could adopt this simple rule: “reject all lawsuits”. Voila! Zero probability of a false positive. However, this entails a very high risk of false negatives–the probability of a false negative will equal the proportion of lawsuits that are deserving (because they will all be rejected). Conversely, a rule like “accept all lawsuits” will eliminate the risk of a false negative but lead to a high rate of false positives. As you move away from one type of error, you move toward the other.

Continuing with my statistics/lawsuits analogy, the analog of improving statistical techniques is to improve the judicial system. Replacing stupid judges with smart ones, or bad rules of evidence with better ones, will decrease the chances of both types of errors. But beyond that the only way to improve on both dimensions is to gather more data. All other reforms involve a tradeoff, which brings me to my point.

The debate about eliminating frivolous lawsuits is rarely discussed in a meaningful way. The costs of the two types of errors need to be reckoned. So an honest statement by a tort reform adherent would be something along these lines:

I believe that the benefits to society of reducing the number of successful-yet-invalid lawsuits would outweigh the costs of the necessarily concomitant increase in the number of unsuccessful-yet-valid lawsuits. Therefore we should raise the bar and make it more difficult to file lawsuits.

Then we could have an honest discussion about the relative costs and appropriate rate of tradeoff between the two types of errors.(*) But outside of academic journals, meaning in the press and in politics, I don’t hear the debate formulated in this fashion. Instead, it’s always

Frivolous lawsuits are costly. Let’s make it harder to file lawsuits, thereby reducing the number of frivolous lawsuits, and we’ll all be better off.

Note that mention of the hidden cost, more false negatives, is omitted. Some might recommend that we only make it harder to file frivolous lawsuits while not making it harder to file meritorious ones–and where possible that’s a good idea–but generally, it cannot be known ex-ante which are which.

Because across the board caps on damages lower the returns to all lawsuits, not just the returns to bad ones, the administration’s various proposed tort reform measures to cap damages on awards are a perfect example of a policy proposal that considers only the beneficial effects of less false positives while ignoring completely the social costs of more false negatives.

AB

(*) You’ve probably heard a phrase like “Better that ten guilty persons escape than that one innocent suffer.” This 10:1 ratio is called Blackstone’s Ratio, after its author, 19th century English legal scholar William Blackstone. In this calculus, false positives (convicting the innocent) are no less than ten times worse than false negatives (freeing the guilty). We could make jury instructions such that juries are more likely to convict, reducing false negatives, but that would necessarily increase the false positives. So much of the debate over legal reform is really a debate over the right “n:1” ratio, or the correct rate of tradeoff between false positives and false negatives. It’s a subjective issue, and the answer may well vary according to context (e.g., death penalty cases have a higher “n” than shoplifting cases). But remember that most judicial reforms involve altering the ratio in one direction or another, not keeping one type of error constant while reducing the odds of the other–even though proposals are all too often cast in terms of the latter.

Alexander Volokh (brother of Volokh Conspirator Eugene), has an interesting essay on the web, “n Guilty Men”, that gives a detailed discussion of the history of thought about the right “n”. To wit, he leads with a recounting of Abraham’s efforts on behalf of Sodom (Genesis 18:23-32):

And Abraham drew near and said, Wilt thou also destroy the righteous [of Sodom] with the wicked? Peradventure there be fifty righteous within the city: wilt thou also destroy and not spare the place for the fifty righteous that are therein? … And the Lord said, If I find in Sodom fifty righteous within the city, then I will spare all the place for their sakes.

…And he [Abraham] said, Oh let not the Lord be angry, and I will speak yet but this once: Peradventure ten shall be found there. And he said, I will not destroy it for ten’s sake.

I was unable to find an estimate of the population of Sodom at that time, but let’s conservatively hypothesize that it was 1000 (in Sodom and surrounding towns). Based on this, God’s “n” would be 99, at least in the context of imposing death. That is, 10 false positives (the righteous being destroyed) was too high a price for avoiding 990 false negatives (the wicked going unpunished). But it would be worth enduring 9 or less false positives in order to avoid 990 false negatives. On the other hand, in the Biblical telling the 3.5 righteous people of Sodom (it’s unclear whether Lot’s wife counts as righteous or not, given her pillar of salt ending) were spared (though they lost their home and posessions).