Monday, September 28, 2009
Last Spring, I participated in a conference at Wake Forest on the Restatement (Third) of Torts. As is usually the case when a group of law professors get together, a number of panelists spoke about whether this or that Restatement provision is economically efficient. This reminded me of a question that I have pondered since law school but never researched—what empirical evidence is there that the existence of common-law tort decisions in fact deters future tortious conduct? I posed this question to a couple of the law and economists attending the conference. After thinking about it for a while, they knew of no such evidence, and one admitted that he was a little embarrassed never even to have considered the question.
Now that I have surveyed the existing literature, I am not surprised by the answers I received—and the professor should not have felt embarrassed. Very little study of the question exists. Law and economics has dominated discussion of tort law for nearly four decades. Virtually every piece of L&E literature relies on the assumption that tort rules and decisions are at least capable of deterring future conduct. Moreover, as Professor Hylton suggested in his blog, judges too find the proposition “noncontroversial, even obvious.” And yet the basic premise is grossly under-tested.
A number of scholars have attempted to answer the question with regard to specific areas of tort law—the effects of no-fault auto insurance rules, workers’ compensation schemes, and medical malpractice reforms have served as the most common subjects. There have also been at least three efforts—by Gary Schwartz; Don Dewees, David Duff, & Michael Trebilcock; and Michael Smith—to aggregate evidence of deterrence to get a look at the bigger picture. Results are mixed. No study has found that tort law serves as a comprehensive deterrent as Landes, Posner, and Shavell, for example, have proposed. Some have found limited evidence that tort acts as a weak deterrent with respect to certain behaviors. Still others have found no evidence of deterrence and even, in a few cases, a negative association—that certain tort rules actually lead to an increase in injuries.
In one respect, it is not surprising that there exists so little study of tort law’s deterrent effects. Empirical investigation of the question poses significant challenges. Obtaining data measuring individuals’ risky behavior is difficult; the number of variables affecting behavior levels is potentially overwhelming; and the rules governing common-law tort liability do not generally vary much among American jurisdictions (providing limited opportunities for comparison).
In light of the challenges facing real-world empirical study of deterrence, Albert Yoon and I are attempting a behavioral-science approach to the question. As a first bite at the apple, this Fall we recruited a number of fellow torts profs to offer to incoming 1L students a series of four surveys designed to elicit evidence of whether the general threat of tort liability deters individuals’ risky conduct. Each survey was comprised of three parts. The first part contained nine vignettes, instructing subjects to imagine themselves facing decisions to engage in risky behaviors under the backdrop of a particular legal regime. The second part consisted of a previously-validated, thirty-question measure of subjects’ inherent risk proclivities (an attempt to segregate subjects’ inherent risk proclivities from any found deterrent effect). The third part collected demographic information.
Each student took only one survey. Each survey was identical, except for the legal regime applicable to the vignettes. Survey #1 told subjects that they could not be sued and would not be held liable for any injury they might cause in doing the conduct described, but that there was a law against the conduct that imposed a criminal fine. Survey #2 informed subjects that there was no criminal law barring the behavior, but that they might be sued and might have to pay for any resulting injury. Survey #3 contained a prompt to the effect that neither criminal nor tort law barred the activity. And Survey #4 contained no legal prompt at all, leaving in place subjects’ background understanding of relevant law. At the end of each vignette, subjects were asked to rate the likelihood that they would engage in the behavior described.
Thus far, we have gathered over 700 survey responses, and we are just beginning to analyze the data. Our hypothesis tracks conventional wisdom with regard to deterrence. Our guess was that Survey #3 (no crim or tort liability) would produce responses indicating the highest likelihood that subjects would engage in the potentially tortious behavior. Survey #4 (no prompt) seemed likely to produce a lower likelihood because although subjects might presume the possibility of liability, they were not prompted to actively consider it. Survey #2 (tort, no crim) seemed likely to have some deterrent effect, and Survey #1 (crim, no tort) might well prove to be the strongest deterrent.
But I wonder whether this hypothesis will bear out. What if our behavioral-science investigation produces evidence that tort law does not deter individuals from engaging in risky conduct? Would that be so surprising? It would perhaps be jarring in light of 40 years of efficiency-based literature, but would it render tort law superfluous? It is fairly clear, as an historical matter, that tort law did not arise as a means of deterring future conduct, but as a means of resolving individual disputes over who should pay for allegedly wrongful acts that resulted in injury. This notion is not quaint even in today’s complex, industrialized world. Nor would such evidence render efficiency irrelevant in tort law—perhaps tort law deters commercial actors, if not individuals.
Of course, Albert and my study has significant limitations. Our pilot only involved law students, a group that is hardly representative of the general population in terms of race, gender, age, education, or socio-economic background. Moreover, the study only tested subjects’ stated willingness to engage in the described behaviors—and behavioral-science studies have shown a frequent disconnect between how people say they will act and how they in fact do act. Finally, subjects were asked to consider their actions a priori—but tortious behavior frequently consists instead of snap judgments or less-than-conscious errors.
Even with these limitations, I think that the study will prove to be a valuable first salvo into an area in considerable need of empirical investigation. If any of you are interested in assigning the surveys to your torts students, the study provides a nice conversation-starter with regard to the purposes of tort law. You can reach me at firstname.lastname@example.org if you are interested.
Dorothy Salmon Professor of Law