EFFICIENT ETHICAL PRINCIPLES FOR MAKING FATAL CHOICES.

• Author: Viscusi, W. Kip

INTRODUCTION

The monetization of reduced risks to life and health is the linchpin of evaluation procedures to ascertain the economic desirability of interventions affecting public health. (1) The function of these values and the degree of their personalization to the specific context are different than for damages in tort cases. Compensation in personal injury cases is ex post and is targeted to address what has been lost as a result of the injury, usually including the present value of the financial harm. (2) To provide compensation for the specific harms, the payment amounts are tailored to the particular circumstances of the case, such as the family's financial loss. (3) Public valuations of health risks focus on reducing the risks to broad population groups, where these risks are valued ex ante. (4) The valuation amounts are broadly based and generally reflect average valuations for the general population rather than being tailored to correspond to the demographic profile of the target population. (5) Bioethicists' discussions of medical rationing in the context of the COVID-19 pandemic have taken a more targeted approach, incorporating distinctions across different patient groups. However, this tailoring of the benefit assessments is not linked to people's own preferences and has not been guided by any economic efficiency principles. Instead, recent bioethicist discussions of the proper response to the COVID-19 pandemic and related medical rationing often incorporate highly problematic ethical judgments as well as a myopic conceptualization of medical resource allocation decisions.

The dominant benefit component in policy evaluations of health risks consists of the value that people attach to reducing mortality risks, or the value of a statistical life. (6) Part I presents a brief summary of this benefit-assessment approach as well as related measures, such as those dealing with very short life extensions. Government agencies generally use populationwide estimates of the VSL to monetize mortality reduction benefits. (7) Although agencies usually rely on average values, that is not because there is no heterogeneity in the VSL. The discussion in Part II indicates that the economic literature has documented several sources of heterogeneity, with the most prominent being differences by age and by income. Part III explores the ethical aspects of different valuation approaches. All of those measures grounded in economic efficiency are related to private values as reflected in the willingness to pay to reduce risks by those who are exposed to the risk. As discussed in Part IV, the literature on medical resource allocation decisions often displays little or no understanding of these underlying economic principles for resource allocation. This failure has received particular prominence with respect to allocating access to ventilators during the COVID-19 pandemic. Proposals to impose arbitrary age cutoffs on access to medical care are both illegal and divorced from economic efficiency concerns. The concluding Part V recommends that the conceptualizations of medical resource decisions be reframed to focus on decisions that are more foresighted in nature, reflecting the prospective willingness to pay of those who will ultimately be bearing the risks.

1. ECONOMIC VALUATION CONCEPTS

   The pertinent matter of interest for assessing policy benefits is society's willingness to pay for the benefit. (8) In the case of health risks, it is the value of the incremental change in the risk that is most relevant, as that is the risk-reduction benefit that is being provided. Schelling first analyzed the economic importance of this tradeoff rate between mortality risk and money, which is now most widely known as the value of a statistical life ("VSL"). (9) The VSL represents the willingness to pay for safety-per-unit risk. For example, suppose that you are willing to pay an extra $900 for a safer car that reduces your fatality risk by 1/10,000. Then your VSL is given by the willingness-to-pay amount divided by the reduction in the risk, or $900/(1/10,000), which equals $9 million. Viewed in another way, if a group of 10,000 people each faced an annual fatality risk of 1/10,000 and each of them would be willing to pay $900 to eliminate that risk, it would be feasible to raise $900 x 10,000 people, or $9 million, to eliminate this group risk of one expected death.

   Although Schelling introduced the concept now known as the VSL, he did not indicate what the source of these numbers should be, was skeptical of whether they could be estimated empirically, and did not discuss how they should be used for policy purposes. (10) My introduction of the use of the VSL in government policies utilized my labor-market estimates of the VSL derived from wage-risk tradeoffs, which incorporated two principles. (11) First, the valuations that are most pertinent are those of the individuals being affected by the risks, not the expert assessments of government officials or academicians. In effect, the VSL incorporates the personal valuations revealed in this implicit market for risk and, as such, respects individual preferences. Second, the best estimates of the VSL are those based on people's actual risk-taking decisions in market contexts. (12) Fortunately, in the United States there is detailed information on the fatality risks of different occupations that can be matched to employment information to estimate statistically the wage premium for fatality risks, making it feasible to calculate the VSL. (13) The alternative approach is to use interview methods designed to elicit the valuation of hypothetical risks. (14) Although these stated preference estimates are sometimes useful when data limitations are encountered, they do not provide as meaningful measures of the VSL. (15)

   My mean estimate of the VSL after adjusting for publication selection effects is $11 million in 2019 USD. (16) The mean estimate without such adjustments is about $2 million higher. (17) Current estimates applied by government agencies are in a similar range after making appropriate adjustments for inflation. The U.S. Department of Transportation recommends a value of $9.6 million in 2016 USD, (18) the U.S. Environmental Protection Agency recommends a value of $7.9 million in 2008 USD, (19) and the U.S. Department of Health and Human Services recommends a value of $9.6 million in 2014 USD. (20) The VSL estimates that have been used in regulatory impact analyses by agencies have increased over time, but in recent years have been in the $9 million to $11 million range across federal agencies. (21)

   In situations in which the life extension that results from a policy is extremely short, it is also feasible to use an analogous concept, the value of a statistical life year ("VSLY"). (22) This measure imputes the value of an additional expected year of life based on the VSL estimate and the remaining life expectancy of the person at risk. The current estimate of the VSLY in U.S. Department of Health and Human Services guidance in current dollars is just over $500,000. (23) Valuation of a new cancer treatment that would extend life by one expected year consequently would have a benefit value of about $500,000 rather than the full VSL.

   Following a similar approach to that of the VSL, there are also estimates of how much people value risks of cancer and other health impacts. (24) As in the case of the VSL, the objective is to base the monetization of the risk reduction on how much those who are exposed to the risk value changes in the risk level.

   Although the VSL approach has received its greatest prominence in policy valuation situations, plaintiffs have also proposed its use in wrongful death cases. Basing compensatory damages on financial losses addresses the monetary impacts on survivors but does not address the loss of life for the deceased. Particularly in states where the loss of enjoyment of life is an element of damages, there have been attempts to introduce the VSL as a measure of the decedent's loss. (25) There have also been other proposals advocating more general application of the VSL in wrongful death cases, such as that by Posner and Sunstein as well as by Polinsky and Shavell, each of which favor making the VSL a routine component of damages in wrongful death cases. (26)

   Efforts to introduce the VSL as a compensatory damages component have had only limited success. Most courts have generally rejected attempts to introduce the VSL estimates as a guideline for jury determination of damages, though there may be some situations in which discussion of the methodology may be permitted. (27) From an economic standpoint, compensation based on the VSL provides excessive levels of insurance. In products liability situations, the purchaser of the product in effect will be paying for this excessive insurance since the expected costs of the compensation will be embedded in the product price. This level of compensation is more than the person would choose to provide after death. Given the economic underpinnings of the VSL, its use in setting damages would be more suitable for setting the total level of damages in situations in which establishing incentives for deterrence is of paramount concern, which usually are when punitive damages are warranted. (28)

2. HETEROGENEITY OF THE VSL AND POLICY IMPACTS

   The VSL is not a natural constant and does not have a uniform value across the population. That attitudes toward health risks may vary is no more surprising than the presence of differences in other types of preferences. The two most prominent personal characteristics that have been analyzed with respect to the heterogeneity of the VSL are age and income. (29) Age is consequential because it plays a critical role in determining the remaining life expectancy that is at risk, which is the commodity that is being valued in the VSL. Income levels also have a critical role since...