By Ted Frank
I attended yesterday's AEI forum debating the Texas study of Professors Black, Silver, Hyman, and Sage. Based on their study, the authors made a broad statement in a NY Times op-ed: "The medical malpractice system has many problems, but a crisis in claims, payouts and jury verdicts is not among them. Thus, the federal 'solution' that Mr. Bush proposes is both overbroad and directed at the wrong problem." At AEI, however, Professors Black and Hyman made the following two concessions (which I paraphrase from my notes):
- Over the long run, malpractice insurance prices reflect claim outcomes.
- Caps on non-economic damages should reduce insurance rates.
Don't expect ATLA to acknowledge these straightforward economic principles in its above-the-fold front-page link to its press release on the study. But once everyone agrees on this central common ground, then, to paraphrase Rabbi Hillel, that's the whole tamale and the rest is just commentary.
After all, it doesn't appear to be the case that the doctors complaining about a "crisis" are concerned only about the first derivative of claims costs. The claim is also the subjective one that insurance costs are "too high" relative to the benefits of the current system of malpractice liability. One can agree or disagree with that value proposition but, once one acknowledges that insurance premiums are a function of claim costs, and further acknowledges that non-economic damage caps would reduce premiums, the authors' op-ed becomes a non sequitur. Perhaps claims costs aren't rising in such a way to cause insurance costs to rise. But proving that premise does not demonstrate that there isn't a "crisis", and doesn't demonstrate flaws in the administration proposal. Ironically, it's the op-ed that is "overbroad and directed at the wrong problem."
And it's far from clear that the premise--that there is no problem of rising risk to insurers from a change in the malpractice environment--is true. As Professor Klick noted at the conference, insurance premiums reflect not just the expected incurred cost, but a risk premium to compensate for the variance of that expectation. And the Texas study did not look at variance. Let's take a look at the Texas Department of Insurance data set used by the Texas study. What does the closed cost data look like in 2000, the year that insurance costs started to rise in earnest?
Extseq #32300902 was a $65,000,000 verdict issued by a jury on December 13, 1999. �The case appears to have been subject to a hi-lo settlement, so it wouldn't have much impact on the closed claims paid data that the Texas study relies upon -- but the fact of such a large verdict has to have had some impact on insurers' pricing. �In addition, there was a verdict of over $33 million that closed in 2000 after settling for about $11 million. �
Let's compare this to the world in 1990, the pre-2000 year in the TDI data set with the highest mean verdict. That year, the most aggressive outlier Texas jury awarded $17 million, a bit less than $23 million in 2000 dollars.
In 1990, there were about seven closed claims with verdicts above $1 million. In 2000, there were about sixteen closed claims with verdicts above $1.32 million (about $1 million in 1990 dollars).
The tail is about twice as tall at the $1 million mark in 2000 than in 1990. There are two outlier verdicts in the 2000 TDI closed claims data that average nearly twice as much as the single largest verdict in the 1990 TDI closed claims data, even after inflation. Yet the Texas study makes the blanket statement that juries weren't any more aggressive over time. Just from the TDI data, this seems to be false.
But wait, there's more. On November 3, 2000, a Dallas jury awarded $268 million to the family of a 15-year-old cerebral palsy patient who died of a medication overdose. This verdict, twelve times the size of the largest 1990 verdict in real dollars, is not in the TDI closed claims data at all, and was thus excluded from the study. The authors acknowledge the incompleteness of their data set, which omits certain self-insured entities that need not report to TDI. I don't think this critique is ambiguous, but let me make clear that I'm not accusing the authors of doing anything unethical by failing to include this data point. But they do err in overestimating the power of the conclusions that can be drawn from the data they do have. Their regressions don't account for increased variance or for the fact that the top three verdicts in twelve months averaged a record $100 million, nearly an order of magnitude higher than those of a decade ago, but the actuaries calculating insurance rates in Texas certainly did. This evidence is certainly evidence is anecdotal, but these anecdotes are large enough pebbles to distort the entire market; in the world of insurance, it's the outliers that create the need for reserves.
"Wait!" the reader protests. Doctors don't get eight or nine digits worth of insurance. Shouldn't insurance companies be indifferent between a $10 million verdict and a quarter-billion-dollar verdict?
No. First, the ratchet of increasing jury verdicts creates settlement leverage for malpractice plaintiffs by increasing the range of plausible results. There's a lottery-ticket effect: even a small chance of winning tens of millions of dollars has significant value. Caps target precisely this unmoored lottery-ticket effect.
Second, insurers aren't just at risk from plaintiffs; they're at legal risk to their insured. If a plaintiff offers a settlement at the policy limits, the insurer refuses, and then the jury awards more, the insurer is potentially liable to the insured for alleged wrongful refusal to settle--what Texas law calls the Stowers duty. E.g., American Physicians Ins. Exch. v. Garcia, 876 S.W.2d 842 (Tex. 1994). This also creates pressure to settle at policy limits to the extent the plaintiff has a chance to overshoot these limits by a great deal.
The Texas study's data set ends in 2002, which means it has very little data for the period of time after the landscape changed. It will be interesting to see what 2003 and 2004 data show. There appears to be an increase in the length of time between opening and closing a claim between 1999 and 2002, though once the data is appropriately weighted, the difference may turn out not to be significant (or even to exist). On the other hand, a significant difference would tend to show that the spate of large 2000 verdicts encouraged medical-malpractice plaintiffs to press their claims harder, thus likely increasing settlement costs and defense costs. If the length of the open-claims tail changed over time, it would require adjustments to the closed-claim data that may change conclusions.