Revision of Departmental Guidance on Treatment of the Value of Life and Injuries
This
document provides background information to accompany the updated Department of
Transportation guidance memorandum on Treatment of Value of Life and Injuries
in Preparing Economic Evaluations.[1]
Background:
A defined value of life has been pursued by governmental agencies here and abroad for over thirty years. It is regarded as an essential element of cost-benefit analysis to guide public policy in the areas of regulation and investment in health and safety. Some measure is necessary to ensure prudent management of public and private resources. Although saving an identifiable life is often regarded as a moral imperative on which no monetary value can be placed, prevention of every possible accidental death would be intolerably costly in terms of both money and the quality of life. The term "value of a statistical life" (VSL) is widely used to emphasize that value is placed, not on a particular life, but on safety measures that reduce the statistically expected number of accidental fatalities by one.
In
the past, VSL was commonly based on a concept of human capital, or prospective
earning power. This approach is
unsatisfactory from a theoretical perspective, because economics derives the
values of goods in market transactions from the rates at which individuals are
subjectively willing to exchange them for money or other goods. Correspondingly, willingness to spend more
for increased safety than the statistically expected improvement in future
earnings should not be treated as irrational or ignored by a representative
government. Therefore, agencies
regulating or investing in safety rely on the values they believe citizens
themselves would assign to the desired benefits if they were purchasing the
protection directly, a standard termed "willingness to pay" (WTP).
On
January 8, 1993, DOT published a guidance memorandum, "Treatment of Value
of Life and Injuries in Preparing Economic Evaluations," recommending a value of $2.5 million as the benefit of averting an
accidental fatality, to be periodically updated by the GDP implicit price
deflator.[2] Since 1996, $2.7 million has been prescribed as the VSL in
departmental regulatory and investment analyses. This value will now be increased to $3.0 million.
Although
economic evidence of WTP provides a necessary point of reference, the value of
life is identified as much with concepts of justice as with empirical
measurement. To say that it is worth
spending so much, but no more, to guard against a risk is to invoke a broad
social consensus about the job that government should be doing. A practical and acceptable standard must
aggregate cases that may, in fact, give rise to very different WTP in private
markets. Age, health, income, specific
type of risk, and degree of personal risk aversion may all be expected to
influence empirical estimates. Most DOT
actions involving safety affect groups that are diverse or unspecified in age
and income, but safety provisions for infants are sometimes distinct, and air
travelers tend to have measurably higher incomes than auto passengers. For the purpose of specifying a quantitative
VSL standard, we have embraced a policy that no distinctions of value should be
made among the lives we are charged with protecting. In conformity with OMB guidance, the benefits of saving lives in
the future, as well as the costs, are discounted at a rate of 7 percent.
Sources:
Three
principal methods have been employed to estimate WTP for reducing risks to
life. Perhaps the most common is
estimation of wage-risk tradeoffs, the wage premiums that must be paid to
induce workers to accept riskier employment.
A second category of revealed-preference studies (using data on
decisions observed in real markets) includes various forms of consumer
behavior, such as willingness to accept cost or inconvenience in exchange for
safety improvements from smoke detectors, automobile seat belts, bicycle
helmets, etc. Third are contingent
valuation (CV) or stated-preference techniques that rely on verbal responses to
carefully structured hypothetical questions.
These techniques are still being refined, and a number of prescriptions
have been developed to guard against biases and inconsistencies in the
responses.
In addition, there are a number of secondary studies that synthesize the results of primary research. These attempt to reach more broadly based conclusions regarding "best practice". Some simply collate studies, identifying those that are more convincing or deriving a simple mean of their estimate. Others are "meta-analyses," applying statistical techniques to the results of independent studies. This approach is used to estimate the parameters accounting for differing results in studies separated by years and national boundaries and to quantify the impact of variables that may not have been estimated directly.
For those who wish to investigate the sources of VSL estimates in greater depth, the documents cited in the footnotes below contain references to the original research that has contributed to analyses.
The
1993 departmental guidance memorandum was based primarily on a survey of the
research literature by Ted R. Miller, which yielded a likely VSL of $2.2
million in 1988 dollars.[3] Additional information was obtained from a study prepared by W.
Kip Viscusi, who found most estimates to be clustered in the range of $3
million to $7 million .[4] The body of research surveyed was essentially identical to that
reviewed by Miller. While Miller
excluded 18 studies he considered methodologically invalid (out of 65) and attempted
to correct for biases in an additional 15, Viscusi made no such adjustments and
did not recommend a single value. The
VSL of $5 million used by FDA, CPSC, and, initially, by EPA is based
substantially on Viscusi's work.
EPA's
Guidelines for Preparing Economic Analyses are posted on the Web.[5] In a currently circulating draft, EPA presents the conclusion of
its consultants that the currently recommended VSL of $6.4 million be increased
to $7.9 million, the mean estimate of 16 selected studies (out of 60 reviewed).[6] The distribution of the original 60 estimates is markedly skewed,
however, with most falling in the $0-3 million and $3-6 million ranges. Therefore, it is possible that the mean is
disproportionately influenced by a small number of studies that yielded high
values.
In
2000, Miller published a meta-analysis drawing on 68 studies he considered
sound, including the original 47.[7] Miller's best estimate for VSL in the United States is $3.670
million in 1995 dollars, with a range between $3.3 million and $4.5
million. The deflator used was the wage
and salary component of the US employment cost index, which grew more rapidly
than the GDP implicit price deflator adopted by DOT. Thus, different assumptions about the growth of VSL from the date
of earlier studies would produce significant changes in present recommended
values.
A
meta-analysis employing VSL estimates derived from labor market data has
recently been conducted by Janusz R. Mrozek and Laura O. Taylor.[8] This study controlled for job characteristics that had been
omitted by earlier research and suggested a VSL range of approximately $1.5 to
$2.5 million in 1998 dollars, with a recommended value of $2.0 million..
[3] Miller, T. R. (1990): "The Plausible Range for the Value of Life – Red Herrings among the Mackerel". Journal of Forensic Economics, 3, 17-40.
[4] Viscusi, W. Kip (1993): "The Value of Risks to Life and Health". Journal of Economic Literature, 31, 1912-46.
[6] DeMocker, James (June 7, 2001): "Benefits and Costs of the Clean Air Act 1990 - 2020:Draft Analytical Plan For EPA's Second Prospective Analysis". Office of Policy Analysis and Review, U.S. Environmental Protection Agency. Democker.Jim@epamail.epa.gov 202-564-1673
[7] Miller, T. R. (2000): "Variations between Countries in Values of Statistical Life". Journal of Transport Economics and Policy, 34, 169-188.
[8] Mrozek, Janusz R. and Laura O. Taylor (Aug. 29, 2001): "What Determines the Value of Life? A Meta-Analysis". http://www.gsu.edu/~ecolot/research.html.