The law and economics of products liability.

• Author: Hylton, Keith N.
• Position: III. Products Liability Law: Economic Analysis 1. Consumption Effects c. Adverse Selection, Moral Hazard, and Correlated Claims iii. Correlated and Cascading Claims through V. Conclusion, with footnotes, p. 2485-2514

iii. Correlated and Cascading Claims

Insurance is difficult to offer when claims are positively correlated. (112) In the standard life or health insurance setting we typically do not see positively correlated claims. The likelihood of one insurance customer dying from a heart attack does not influence the likelihood that another customer will die from a similar event. By holding a large portfolio of potential claims, the insurer can virtually eliminate uncertainty connected to the expected payment of claims within a fixed time period.

When claims are positively correlated, however, the insurer cannot virtually eliminate uncertainty with respect to expected payouts by holding a large portfolio. This is a phenomenon we observe in the case of natural disasters, such as earthquakes and hurricanes. If one person claims, we are likely to see hundreds of others claiming.

Insurance is not impossible in the context of correlated claims. We observe earthquake and hurricane insurance. However, because of special difficulties in these markets, we often see state mandates requiring that the insurance be offered and sometimes state subsidies to encourage customers to purchase the insurance. (113)

The claims experiences connected to various products differ. With respect to most products on the market, we are unlikely to see positively correlated claims. However, with respect to some products, we should anticipate positive correlation. For example, products that affect health and the environment are likely to generate positively correlated claims. The asbestos cases and the silicone implant cases provide examples. Asbestos claims have flooded the courts over the past thirty years. (114) Once established as presenting a health risk, asbestos generated an avalanche of products liability claims. Similarly, silicone implants generated an avalanche of claims. (115) However, the scientific evidence of causation has never been established with respect to silicone implants. (116) The silicone claims appear to have been strategic efforts by plaintiffs to gain compensation for illnesses unconnected to silicone implants.

I prefer to think of the silicone implant example as one of cascading claims rather than correlated claims. I use the term cascade here to refer to a process in which a legal theory is adopted successively by actors who have rational bases to do so. In the silicone context, the scientific causation evidence suggests that the avalanche of claims was the result of a concerted effort to create a pool of funds for the purpose of compensating claimants and lawyers. The heart of the process is an information cascade rather than a natural disaster generating positively correlated injury claims. (117) In the information cascade process, one lawyer or group of lawyers proposes a theory of injury causation, which is passed on to other lawyers and widely advertised to the public. Lawyers have incentives to join the cascade, even if they privately doubt the underlying causal theory. The reason is that as more claims join the cascade, the greater the probability that the target will be forced to settle rather than fight all of the claims in court. Thus, cascading claims generate a self-fulfilling prophecy.

The correlated and cascading claims scenarios imply that the output reduction effect of strict products liability can be substantially larger than anticipated. Like adverse selection, this is a case in which the output reduction effect can reach such a level that the product cannot be offered. This is perhaps not a troubling result in the case of some products, such as asbestos, where the harms are generally well understood in terms of causation principles. (118) It is a troubling result, however, in the cases where plaintiffs' causal theories are invalid or unsupported by science.

2. Design Choice Effects

   The discussion in Part III.B.2.b of the design choice implications of strict products liability assumed, as did the initial presentation of consumption effects, that there are no litigation costs, and that courts are perfectly accurate as well as predictable. If courts are not accurate and if litigation costs are substantial, there is no guarantee that strict products liability will optimally regulate the design choice decision. The liability system may "over-internalize" or "under-internalize" injury costs to the producer, depending on whether damages typically exceed the cost of bringing a claim to court. In other words, products liability may tax the producer by an amount that exceeds the expected value of the injury risks it creates, or by an amount that is less than the value of the risks it creates. Since products liability damages are likely to exceed the cost of bringing a claim to court, and given the active class action practice in products liability, (119) the more plausible result is over-internalization of risk costs to the producer.

   In addition, the insurance market inefficiencies discussed earlier could easily lead to inefficient regulatory outcomes. If adverse selection occurs in the market for the risky product, leading to excessive purchases by high-risk consumers, the liability system will internalize to the producer costs that exceed the expected risk costs based on the entire population of consumers. As a consequence, the costs shifted to producers as a result of products liability may be significantly larger than the real risk costs generated by the products.

   On the assumption that liability costs to producers over-internalize the risk-related costs, products liability law will not optimally regulate the design choice. Since the risk cost internalized to the producer is likely to exceed the injury risk to the consumer, producers will be excessively reluctant to adopt net-positive utility designs that entail additional risk to consumers.

   Now consider innovation, under the assumption that liability over-internalizes risk costs to producers. I noted earlier, in the analysis under ideal conditions (no litigation costs, uncertainty, etc.), that strict products liability failed to provide optimal innovation incentives, but that the liability rule did not alter the outcome that would be observed in a full-information market. With litigation costs taken into account, strict products liability exacerbates the divergence between the private and social incentive to innovate. Relative to a market in which consumers are fully informed, products liability excessively discourages innovation.

   4. Effects of Legal Tests

   The preceding discussion examined the effects of liability under a very simple rule of strict or absolute liability. Of course, the law has not generally adopted such a test--the closest the law comes to it is the strict liability rule that applies to manufacturing defects. The general tests adopted in the law are the consumer expectations test and the risk-utility test. In this Part I will examine the incentive effects of these tests, taking advantage of the arguments developed in the foregoing analysis of the strict producer liability rule.

1. Consumer Expectations Test

   1. Consumption and Design Incentives

      Although the preceding analysis examined a simple version of strict liability, the analysis explains the effects of the consumer expectations test, the first test for liability proposed in the Restatement (Second) of Torts. (120) Under the consumer expectations test, the manufacturer is strictly liable for defective conditions of which the consumer is unaware, and is not liable when the consumer is fully aware of the defective conditions. (121) The consumer expectations test is an especially refined form of strict producer liability that imposes liability when consumers are uninformed and provides immunity from liability when consumers are informed.

      Based on the preceding analysis, and all of its simplifying assumptions, the consumer expectations test is an optimal rule for regulating consumption decisions. The rule internalizes the costs of injury risk to manufacturers when consumers are not aware of them, and this leads to optimal production decisions. On the other hand, when consumers are fully aware of the risk costs, the consumer expectations test leaves them to suffer the costs of their decisions, which leads to optimal consumption choices. As the preceding analysis indicates, optimal consumption patterns result from the consumer expectations test.

      As for product design decisions, the preceding analysis implies, again, that the consumer expectations test provides an optimal regulatory framework, in the sense that the test leads the producer to choose the risky design when and only when the incremental utility from the risky design is greater than its incremental risk. If the incremental risk exceeds the incremental utility, and consumers are uninformed, the liability imposed under the consumer expectations test would induce the manufacturer to choose the safe design instead of the risky design. If the incremental risk is less than the incremental utility, and consumers are uninformed, the liability imposed under the consumer expectations test would not alter the manufacturer's preference for the risky design. If consumers are informed, the consumer expectations test permits the market to regulate design choice, which is preferable in that case.

      The consumer expectations test's only potential flaw becomes apparent when we consider innovation. Although the test appears to lead to optimal choices between two existing product designs, it replicates the market's incentives when it comes to innovation of new product designs that improve consumer welfare.

   2. Real-World Complications: Consumer Expectations Test

      As commentators have noted, there are practical problems in implementing a consumer expectations test. (122) Some consumers may be fully aware of the risks while other consumers are not. Whose awareness level should the consumer expectations test incorporate? My analysis to this point...