Wednesday, June 12, 2019

Litigation Versus Disclosure Versus Regulation as Means of Inducing Cost-Internalization: A Comment on the Insys Bankruptcy

by Michael C. Dorf

As then-Professor Guido Calabresi's pathbreaking work explained, the tort system can be a kind of substitute for or complement to regulation. As applied to manufacturers, tort damages serve not only to compensate victims but also to induce cost internalization. In this blog post, I shall use the news of the bankruptcy filing by Insys Therapeutics--maker of fentanyl--as an occasion to discuss some of the advantages and disadvantages of the tort system relative to other mechanisms for promoting cost internalization. I'll discuss taxes, regulation, markets, and disclosure.

We have grown accustomed to thinking of internalization as a process by which customers come to pay a price that reflects not only the cost of producing a good but also the cost its use imposes on the larger society. For example, a Pigovian tax on fossil fuels means that the price of a gallon of gasoline reflects not just what it costs to get the gas to the pump but also the social cost of pollution. The very term "internalization" implies that absent the intervention (such as a tax), the product creates negative externalities, i.e., harms inflicted on others.

However, internalization can also be encouraged relative to hidden costs to the consumer/purchaser him or herself. Suppose that installation of some safety mechanism will substantially reduce the likelihood of a certain kind of fatal-to-drivers accident, while increasing the cost of cars. If consumers are unaware of the reduction of the fatal risk, then they might prefer the cheaper models without the safety mechanism. Depending on the numbers, that could be an irrational preference. How might the legal system respond to this schematic example and others like it? I have already mentioned taxation as one approach. Let's consider a few more possibilities.

Rely on the market. If cars with the safety mechanism are sufficiently safer than cars without the mechanism, then car manufacturers might compete on safety. Even if there is no tort liability for failure to install the safety mechanism, it could become standard equipment if car company A markets its cars as safer than those of cars sold by competitors. In this scenario we would still need government to police false advertising (to prevent car company B from misleading consumers by saying that its cars without the mechanism are as safe as A's cars), but if the safety mechanism is cost-justified, consumers will prefer it. Or consumers will sort themselves out based on their preferences for risk, cost, and other factors. Safety-conscious consumers will buy Volvos; cost-conscious consumers will buy used small cars; and thrill seekers will buy high-performance sports cars.

In principle, the market-based approach would work, but we do not rely on it alone for two main reasons. First, we think that consumers are unlikely to have sufficient information--even after the marketing efforts of Volvo and like companies--to make fully informed choices. Second, we have paternalistic grounds for thinking consumers will make poor choices even if they have full information. The same sorts of considerations that led state legislatures to mandate seatbelt use led Congress to mandate that new cars be equipped with seatbelts and other safety equipment.

Disclosure. To the extent that consumers simply lack information, government-mandated disclosure can help. Consider a schematic comparison. Suppose that a widget sold without a shield carries a sufficient risk of injury to the average user that, once one does the math, it imposes a statistical cost of $10 per unit on users. Government could mandate a $10 tax per unit sold, with the proceeds going into a fund to compensate people who suffer the injury associated with the unshielded widget. Or government could mandate that all widgets be sold with shields. If adding a shield increases the cost of widgets by $10 or less, mandating shields makes sense, but if it costs more than $10 then it "over-protects" and is not a cost-justified intervention.

An alternative to both taxation and command-and-control regulation is mandated disclosure. Let's say you're thinking about buying a widget for the going price of $100. Perhaps you could buy one with a shield from the Volvo of widget makers, but that adds a cost of $20, which makes the widget not worth your money (because you have an average risk tolerance profile). If you have full information, you will only buy the $100 unshielded widget if its value to you is at least $110, because you need to be compensated for the additional cost of the risk. Raising the effective price in this way will lower demand, in principle by the same amount as the $10 tax would.

By contrast with the tax, however, relying on purchasers to make their own calculations means that there will be no public fund to distribute aid to the people for whom the risk materializes. In theory that's also not a problem, because the consumer who only purchases the $100 widget if it has a value of $110 to him sets aside those $10 either to self-insure or to purchase a widget-induced-accident insurance policy that mimics the public fund under the tax scenario.

If the idea of consumers purchasing widget insurance or making mental price adjustments based on risk strikes you as far-fetched, then you will conclude that disclosure alone has the same drawbacks as relying purely on the market. Disclosure won't really solve the problem of consumer ignorance, at least for reasonably complex products like cars or medications.

Think about the disclosures of the risks associated with medications. Potential side effects are listed in advertisements without any quantification, much less with reference to the underlying studies that form the basis for the listing of the potential side effects. One strongly suspects that a typical potential consumer suffering from the condition that the medication treats simply relies on a doctor's recommendation that the drug's benefits outweigh its risks, all things considered. And insofar as prescription medication insurance covers much of the cost of a medication, price does little to inform or reflect consumer preferences.

Tort Liability. Much of the scholarly literature on the tort system as a mechanism for inducing cost internalization concerns the choice between a negligence rule (under which manufacturers are only liable for the foreseeable harm that could have been avoided by taking cost-justified measures) and a strict liability rule (under which they are liable even if the measures needed to avoid the harm were not cost-justified). Let's set that debate aside. Under either a negligence rule or a strict-liability rule, the goal is a product whose price reflects its true costs: production plus internalized harm. Consumers (who do not have exotic risk preferences) do not need to make calculations of risk. If they think they'll get at least $110 of value from a widget priced at $110, they buy it.

The fact that the tort system bypasses consumers' lack of information and potentially bad decision making seems like an advantage over the unregulated market or disclosure for the reasons discussed above. This seems especially true when we come (finally!) to the news story that inspires this blog post: liability for harm caused by opioids. After all, consumers who become addicted to a drug can hardly be expected to make rational choices about whether to continue to use the drug.

Nonetheless, the tort system has at least two very substantial problems. One is very high administrative costs. To be sure, tort reform advocates often exaggerate the system's costs while understating its benefits. But even fair-minded assessments would be high. Most tort plaintiffs hire contingency-fee-based lawyers who take a cut of about a third (or up to 40% post-trial) of any award. That's a higher percentage than in health insurance, where administrative costs are routinely decried as very high. It does not follow that the tort system should be abolished or reformed along the lines proposed by tort reformers, but one cannot simply gainsay the high administrative costs.

To be sure, high administrative costs are less of a problem for cost-internalization than for compensation of victims. If a producer pays $10 million in damages for harm caused by its products, price will reflect that $10 million, even though only $6 million to $6.67 million will find its way into the pockets of injured parties. The result will be proper cost internalization, equivalent to a consumer paying $10 for the purchase of a widget insurance policy but only getting a policy worth $6 to $6.67. That's a problem but not a cost internalization problem.

A second substantial problem with the tort system does relate to cost internalization:  uncertainty. Here's how it works in theory: A manufacturer of a product is in the best position to calculate its costs and benefits. After making cost-justified safety improvements, the manufacturer calculates how much in tort-compensable injuries it will need to pay (in principle this is zero in a negligence-rule regime) and raises the price accordingly. Or the manufacturer's insurance carrier makes these calculations in order to determine the manufacturer's premium, which in turn is passed on to consumers in the price.

In practice, however, neither firms nor their insurers can make anything like precise calculations of potential exposure for non-recurring harms. Actuaries using population-wide data can do a very good job of predicting the number of automobile fatalities in one year based on data from the previous year. But the marketers of a complex new product have a much harder time predicting and pricing the harm it will cause. That's a problem for all of the mechanisms I've discussed thus far, but it will be compounded in the tort system, where liability standards interact with juries and multiple jurisdictions.

The Insys bankruptcy filing is a cautionary tale. It comes less than a week after Insys agreed to pay $225 million to the federal government to settle criminal and civil actions. States and individuals harmed by Insys's marketing of fentanyl might still be able to recover something from Insys and/or its insurance, but given the magnitude of the harm, it seems unlikely that everyone will recover everything to which they are legally entitled. Thus, the Insys case--like litigation involving asbestos a generation ago--shows another risk of relying on the tort system: undercapitalization and underinsurance.

Now it might be objected that the likely limits of liability of opioid makers, while tragic, do not directly bear on the topic I've chosen to address today. After all, the opioid makers stand accused of what amount to intentional torts: deliberately marketing addictive drugs. That's true, but the behavior that gives rise to mass liability will usually include a mix of inadvertence and intentional misconduct--such as failure to disclose or act on risks known by the firm but not regulators or the general public.

Accordingly, we see that there is no close-to-perfect approach. That probably explains why the legal system deploys a combination of command-and-control regulation, Pigovian and other taxes, disclosure requirements, market incentives, and tort liability.

No comments: