The value of statistical life is a number that informs analyses of the benefits and costs of regulations. But the number that the US Environmental Protection Agency currently uses needs attention: it hasn’t been updated since 1997.
The value of a statistical life (VSL) is used by government agencies in evaluations of prospective regulatory policy to calculate the benefits derived from reduced mortality. The process of regulatory benefit-cost analysis, which was initiated in the 1980s, must compare the benefits of any prospective regulation with the costs of imposing that regulation. This comparison is called a regulatory impact analysis and is performed with the economic methods that are outlined by the US Office of Management and Budget in the agency’s “Circular A4” guidance (which currently is being revised). Reductions in mortality risk contribute most of the benefits in regulatory impact analyses for policies that aim to protect health and increase public safety, such as regulations that aim to reduce road accidents, air pollution, or the occurrence of foodborne illness.
The VSL often has been misinterpreted as the value that the government assigns to a person’s life in a given regulatory impact analysis. But “it’s not the value of saving a life; it’s the value of reducing your risk of death by a little bit,” as Resources for the Future (RFF) Senior Fellow Alan Krupnick explains on an episode of Resources Radio. A “little bit” usually is something on the order of 1 in 10,000; for example, if 10,000 people experience a 1/10,000 reduction in the risk of dying from exposure to pollution in a given year, then the sum of those reductions in risk equates to one statistical life.
History of the Value of a Statistical Life
As a short history lesson, the US government’s monetization of reductions in mortality risk originates in Cold War–era economic analysis of bombing strategies by the US Air Force. Initial methods primarily relied on using an individual’s earnings to determine the value of an avoided death, which is known as the “human capital” approach, though the use of these methods led to the obvious issue of excluding people outside the labor force and disregarding people’s preferences for reducing their mortality risk.
The idea of valuing small reductions in mortality risk as an alternative to the human capital approach was not presented until the economist Thomas Schelling published his essay “The Life You Save May be Your Own.” In the essay, Schelling describes how to value consumer preferences for reductions in mortality risk and coins the term “value of statistical life.” Schelling’s idea spurred further research, and studies that calculate the VSL increased throughout the 1970s.
The VSL had its federal debut after Executive Order 12291, which established the practice of regulatory benefit-cost analysis. Economist W. Kip Viscusi used the VSL to settle a dispute over the calculation of policy benefits between the US Occupational Health and Safety Administration and the US Office of Management and Budget. The Reagan administration accepted the VSL, which relevant agencies then adopted for use in their own cost-benefit analyses, as spearheaded by the US Environmental Protection Agency (EPA) and Department of Transportation.
The VSL that EPA currently uses is of interest, because it is used by academics and other federal agencies, governments, and institutions around the world. This VSL is derived from studies that use stated preference approaches (in which researchers ask people hypothetical questions about their willingness to pay for reductions in mortality risk) and revealed preference approaches (in which researchers analyze data that reveal actual choices that people have made that are related to the avoidance and acceptance of mortality risk). EPA’s VSL was established in a retrospective analysis of the Clean Air Act in 1997 and draws from a review of 22 studies in Viscusi’s 1992 book Fatal Tradeoffs.
While, over time, EPA has adjusted the official VSL value of $7.4 million ($2006) for inflation and increases in income when applying it, the value has not been updated with new research since 1997. Our recent RFF working paper provides a review of major research on the VSL published in the United States since 1997. In this new paper, we outline the advances in this research and propose criteria for suitable revealed and stated preference studies that EPA could use to reexamine the studies that underlie the VSL.
How Research Informs the Value of a Statistical Life
All of the revealed preference studies that underlie EPA’s current VSL apply what economists call the “hedonic wage” methodology: many of tthese studies examine trade-offs between wages and the risks of on-the-job mortality. In the labor market, higher risks demand higher wages; these studies “reveal” worker preferences when trading off wages and risk by indicating the additional wages that workers will require before accepting an additional level of risk on the job.
Hedonic wage studies are useful, because they utilize real-world data on the wage-risk trade-off—but these studies do have drawbacks. A hedonic wage VSL can be estimated only for workers, meaning that people who are not in the labor force or who have retired are left out of the sample. This exclusion matters. Research suggests that EPA’s regulations greatly affect health outcomes for people who are 65 and older. Furthermore, the risks that workers face on the job, such as transportation accidents, are not subject to EPA regulation. Another issue is that many older hedonic wage studies exclude female workers. Nevertheless, since EPA determined its VSL in 1997, hedonic wage studies have benefited from better data that are available through the Census of Fatal Occupational Injuries. This census includes data on both female and male workers and has adopted comprehensive categories for measuring risk by occupation and industry.
Hedonic wage studies are only one type of revealed preference study. The VSL also can be calculated from purchases of homes, medical care, and cars, as we highlight in our working paper. In the case of home purchases, environmental disamenities (so-called “bads”) that are associated with a property, such as proximity to an EPA Superfund site or a highway, are tied to health risks that affect the home’s price relative to similar homes that are located elsewhere. These studies are less abundant, as controlling for all the other “goods” and “bads” that impact home price can be difficult.
Stated preference studies avoid the issue of restricted samples but do not capture real-world trade-offs between money and risk, like revealed preference studies do. In stated preference studies, a respondent is asked how much they would be willing to pay for small reductions in risk. These surveys are carefully crafted and rely on a respondent’s clear understanding of their baseline risk and the specific risk they are paying to reduce; for example, the risk of developing bladder cancer due to exposure to lead in drinking water. The answers, as well as debriefing questions and personal characteristics of respondents, are used to compute an average VSL for the sample of respondents. Contemporary stated preference studies are more sophisticated and credible than the stated preference studies that were available when EPA estimated the VSL in 1997, given the greater emphasis on risk communication and debriefing questions in newer studies.
Updating the Value of a Statistical Life
Beyond a quantitative revision of the VSL, the time also may be right to rebrand the VSL itself. The term “value of a statistical life” has led to confusion: it is sometimes interpreted as the value of a named individual’s life, rather than the sum of the values that people place on small reductions of risk. Although alternative names for the VSL, such as the “value of mortality risk reductions,” have been proposed, a new name has yet to replace the original.
Criticism also has stemmed from controversy in the mid-2000s over the “senior death discount.” EPA experienced backlash when it suggested using a lower VSL for people older than 65 and cited the frequently observed inverted-U relationship between age and the VSL. This revision of the VSL by EPA was construed in the popular media as the government devaluing the lives of senior citizens, which led EPA to declare that it would not distinguish VSL by age. However, using multiple VSLs potentially could be useful if, for instance, reliable estimates are available that differentiate by mortality risk.
EPA’s National Center for Environmental Economics has produced in-depth white papers on the VSL and potential avenues for a re-estimation in 2010 and 2016 and has received subsequent responses from the Science Advisory Board (a feedback process that is governed by the Federal Advisory Committee Act). The correspondence between the two bodies has yet to yield a consensus, which suggests that EPA could benefit from an approach that circumvents the formal process; for example, the US Department of Health and Human Services revised its own VSL in 2016 by working with experts on a structured review.
As we conclude in our working paper, “Whatever EPA’s process, reexamination of the VSL is imperative. Newer, better designed, and analytically rigorous US studies are available in both the revealed and stated preference literatures. Past attempts to revise the VSL have been derailed by reviewers’ observation that new studies are not perfect. The best thus becomes the enemy of the good. The current VSL has been in place for 30 years. It is time to reexamine it.”