This study evaluates the impact of six different types of tort reforms on the frequency, size and number of total annual settlements in medical malpractice cases between 1991 and 1998. Previous studies have failed to correctly identify the effective dates of reforms, to account for the retroactive applicability of striking down reforms, or used highly selected samples of jury verdicts or litigated cases. I employ a new legal data set of tort reforms, which carefully evaluates effective dates as well as when certain laws were overturned. Medical malpractice data comes from the National Practitioner Data Bank, which contains more than 100,000 malpractice settlement payments in the study time frame. The data represent the universe of cases in which doctors paid a positive settlement. Thus, the present study has significant advantages over previous work for being the first study to systematically and adequately explore the impact of tort reform on settlements (in contrast to judgments). Of the six tort reforms examined, two reforms (caps on pain-and-suffering damages and limitations on joint and several liability) reduced the number of annual payments, and two reforms (caps on pain-and-suffering damages and the periodic payment reform) reduced average awards. Caps on non-economic damages had an effect on total annual payments, although the statistical significance of that effect was weak. The joint effect of enacting all six reforms was statistically significant for reducing the number of cases but not the state level average award or total payments.

This report offers a framework for assessing the effects of tort reforms. It attempts to provide a coherent structure for systematically thinking about how research can contribute to the policy debate over tort reform. It identifies four basic policy issues critical to assessing the effects of tort reforms on the tort system: (1) how soon we can expect to see effects of reforms; (2) whether reforms have affected the outcomes of disputes; (3) who won, who lost, and how much; and (4) whether reforms have affected economic behavior. The author points out that the kinds of data needed to assess the effects of reform are generally not available, and suggests that three types of new data collection systems need to be considered: (1) systematic efforts to obtain data from insurers and self-insured defendants on the aggregate outcomes of liability claims; (2) special surveys of claimants, the bar, and insurers to obtain the detailed individual claim information needed to identify the winners and losers in the reformed system; and (3) systems for collecting information both on the other factors that affect the behavior of participants in the tort system, and on economic outcomes and injuries.

Tort reforms enacted in response to asserted crises date back to the 1970s and have emphasized the highly visible areas of punitive damages, medical malpractice, and products liability. Little evidence exists that reform of punitive damages affected the ratio between punitive and compensatory damages. This is consistent with the absence of evidence that punitive damages were ever out of control and in need of reform. Evidence of the effect of tort reform in the medical malpractice field is mixed. Caps on non-economic damages have reduced costs, thereby likely decreasing pressure on hospitals to improve care. Consistent evidence of effects on physician behavior and physician supply has not emerged. Tort reform has rarely sought to address the well-established problem of widespread harm caused by poor quality care. Products liability plaintiffs have had decreasing success over time. While one cannot rule out specific statutory reforms as achieving more favorable results for defendants, the national scope of plaintiffs' declining success supports an explanation based on the social construction of knowledge by well-funded industry groups.

A theoretical and empirical analysis of the disposition of malpractice claims compares the actual outcomes with the legal standard of payment equal to damages, if, and only if, negligence occurred. An economic model of the settlement process assumes that the litigants attempt to maximize wealth, subject to the legal standards of liability and damage, and the costs of litigating. The model predicts that awards in settlements out of court will reflect the expected verdict, the probability of the plaintiff's winning, and the costs of going to court. The probability of the plaintiff's winning in settlement will also reflect the probability of winning in court, the size of the expected verdict, and the costs of going to court. Evidence from three malpractice claims surveys is consistent with the legal standard but departs in ways predicted by the model. Characteristics of the injury, the plaintiff, and the defense influence the outcome.

This article is a Reply to the critique by Black, Hyman, and Silver (BHS) of our 2007 Journal of Legal Studies article, “An Empirical Assessment of Early Offer Reform for Medical Malpractice.” The early offer reform gives insurers the option of making an early offer that will expedite payment of claimants' economic losses and reasonable attorney fees. Using data on closed medical malpractice claims from the Texas Department of Insurance (TDI), our 2007 article estimates the financial impact of this proposal by comparing the expected payments to claimants under the early offer reform to the payments under current tort rules. A central component of our calculation of expected payments is unique information on insurers' reserves associated with the claim; actual payments are reported in the TDI data for all litigated and settled cases involving payments of at least $10,000. The BHS article misrepresents fundamental aspects of our empirical analysis, including the following. BHS set out to correct our purported “false assumption” that all claims have a 1.0 probability of success, which is a problem that arises because BHS omit the probability of claimant success from the formula that is presented in our paper. BHS's error is compounded as their discussion of our paper fails to recognize that our use of reserve amounts in the analysis incorporates the insurers' estimates of the likelihood of claimant success, Indeed, they neither acknowledge our use of the insurer reserve data, nor do they use the insurer reserve information in their paper. BHS claim incorrectly that our analysis does not discount deferred payments whereas in fact it does. Our early offer analysis uses data for both litigated and settled claims, avoiding the selection bias and measurement error problems associated with BHS's extrapolation from the 2% of paid claims that are litigated to the universe of all settled and litigated claims. In addition to these and other errors in their characterization of our empirical analysis, the BHS article reflects a misunderstanding of the operation of the early offer reform, which leads them to erroneous statements regarding how the parties would behave if the early offer reform were implemented. We also provide a brief critique of the BHS two-sided version of the early offer proposal, which would not be workable and would not reduce litigation costs significantly.