105

CHAPTER 15

WHAT IS THE VALUE OF HUMAN LIFE?

Reasons and Methods for Placing Finite Values on Unidentified Human Lives

Imagine if we placed an infinite value on human life. It would then be rational for us to protect it at all costs. Every health-care expenditure that would save a life would be infinitely more valuable than expenditures on, say, entertainment, so we wouldn’t start spending on movies, sports, or amusement parks until every beneficial health-care measure had been adopted. We would spend no money on nonessential goods, such as holiday gifts, furniture, and artwork, until all research with a positive probability of yielding cures for deadly diseases had been performed. Products would become far more expensive because product and workplace safety standards would be set high enough to eliminate risks to life. For instance, no amount would be too much to spend on safety measures in automobiles because the extra money it takes to build cars stronger than tanks would be well worth the priceless lives saved. Activities such as smoking, drinking alcohol, eating red meat, and crossing streets are preventable causes of death, threatening the loss of infinite value and trumping any finite value for the benefits of those activities. We exchange risks of death for money whenever we decline options for side-impact airbags and antilock brakes on cars and whenever we accept money for increasing our risk of death by leaving the safety of our home to go to work. Trade-offs between lives and dollars imply a finite value of life, but is it necessary to specify that value? And, if so, who’s got the number?

106

As much as anyone hates to place values on human lives, there are several important reasons why this is done. When making policy decisions, legislators must decide where to draw the line when they require manufacturers to spend money to comply with regulations that will save lives. Fire safety standards for aircraft cabins that save lives for $148,000 each and requirements that children’s sleepwear be nonflammable at an added cost of $1.2 million per life saved may be no-brainers. How about asbestos regulations that cost $163 million per life saved? Should we stick with legal limits on occupational exposure to formaldehyde that costs $127 billion per life saved? We could always spend more money to avoid increasingly remote chances of death, but, as a practical matter, we must draw the line somewhere.

When contemplating unidentified lives that we can predict will be lost on the basis of statistics, we want to identify the value of a statistical life. For example, we know that every year in the United States there are about 22 traffic fatalities for every 100,000 licensed drivers. We don’t know who the victims will be, but everyone who drives faces a risk, and with 200 million licensed drivers we have reasonable certainty that about 44,000 unidentified lives will be lost each year. As we look into the future and make decisions about appropriate speed limits, automobile safety standards, and drunk driving policies, we are discussing not specific lives but statistical lives. A nationwide speed limit of 10 miles per hour would save many statistical lives but at a tremendous cost. The fact that we allow cars to go 70 mph on many highways indicates that we have accepted a trade-off between the statistical lives that will be lost in accidents at such speeds and the benefits of getting places quickly. Naturally, the value we place on a particular life, especially our own, is in another ballpark.

In a wrongful-death lawsuit, one person’s death is attributable to the willful or negligent act of another. Jurors must decide how much the victim’s family should receive from the defendant. If the goal is to elicit efficient decisions about dangerous behavior and safety precautions that might affect lives, then a deterrence value for a life equal to the full value of a statistical life is needed. If decision makers know they must pay the full value of a statistical life in the event that their actions kill someone, they will not take those actions unless the benefits exceed the expected loss. Automobile manufacturers, for example, are often faced with the dilemma of whether to spend more money on safety features that will save lives in accidents. In the 1970s, the Ford Motor Company estimated that 180 deaths and 180 serious burn injuries would be prevented if it installed an $11 part to avoid fuel fires on each new Pinto (a line of small cars). With the expectation of being required to pay $200,000 per lost life and $67,000 for each serious injury, Ford calculated that the benefit of installing the parts would be $49.5 million, whereas the cost would be $137 million.¹ On that basis, the company made the unfortunate and inefficient decision to forgo the safety measure. If Ford had anticipated court-imposed penalties similar to the actual value of a statistical life, the company would have installed the part.

¹ See www.fordpinto.com/blowup.htm. These estimates of the payment per lost life and per serious injury are from the National Highway Safety Administration.

In some wrongful-death cases the goal is to compensate the family of the victim for the financial burden of losing a wage earner, in which case an award equal to the value of lost earnings will suffice. When we purchase life insurance, it is our own task to decide how much money our families should receive if we die. This value—the insurance value of a life—is generally less than the deterrence value of a life. The insurance value is really a bequest to others that the victim will not enjoy directly, and when deciding how much to leave for others, people typically choose an amount that is less than the full value of their lost earnings. W. Kip Viscusi and William Evans estimate that the insurance value of a life is about 85 percent of the value of lost earnings.²

² “Utility Functions That Depend on Health Status: Estimates and Economic Implications,” American Economic Review (1990), 80:3, 353–374.

107

The dilemma of how to value lives does not lend itself well to surveys. The task of assigning a value to life on the spot is daunting, and responses tend to be at one extreme or the other. In wrongful-death cases the basis is often the value of lost earnings over the expected lifetime of the victim, sometimes with adjustments including family members’ losses of love, affection, and companionship from the victim (termed consortium in legalese). After the September 11, 2001, terrorist attacks, Congress set up the September 11 Victim Compensation Fund for the families of people who had been killed or injured. To compensate for nonmonetary damages, including pain and suffering, lost consortium, and emotional distress, the families of victims received $250,000 plus $100,000 for the spouse (if any) and for each dependent of the victim. Each family also received an award to compensate for the victim’s lost earnings less the amount of any “collateral sources of compensation,” such as life insurance, Social Security benefits, and pension funds. On average, the total Victim Compensation Fund award per family was $2,082,035.³

³ See www.usdoj.gov/final_report.pdf.

Although they are popular in lawsuits, values of life based on earnings involve a critical flaw: We are not what we earn. The value of retired people, homemakers, and chronically unemployed people is clearly not 0, even though their earnings are 0. Lost-earnings values provide guidance for the compensation of families, but they are much smaller than the values society places on statistical lives. To make a better value-of-life estimate, we need to know the number of dollars people would accept in exchange for a certain statistical probability of death. Suppose you would accept $800 in exchange for taking a 1 in 10,000 risk of death (roughly the annual fatality risk for U.S. manufacturing workers). You are effectively valuing 1/10,000 of your life at $800, indicating an implicit statistical value of life of $8 million (10,000 × $800). The following sections describe assessments of the true value people place on statistical lives, calculated on the basis of real-world trade-offs between money and risks rather than on surveys or lost earnings.

During the past 30 years, economists have favored estimates of the value of human life that are based on trade-offs made in labor markets. Other things being equal, jobs that entail greater risk of death pay more than safer jobs. Compared to jobs with similar prerequisites, for example, miners, firefighters, contractors in war zones, and construction workers receive wage premiums in exchange for their higher risks of death. More subtle variations in risks and wages appear throughout the workforce. Using econometric analysis, economists can input data on hundreds of thousands of workers and have computers adjust for differences in worker age, experience, education, and other determinants of wages to calculate the trade-off between wages and risk of death apart from other influences. Extrapolating as we did with the $800 figure in the preceding example, the amount the typical worker is willing to accept for a small risk of death can be used to estimate the value of a statistical life. In similar ways, 27 labor market–based studies determined an average value for a statistical life of $8.76 million. These estimates can be adjusted to fit individuals who are, for example, younger or older than the average worker. Discomfort with suggestions that some people are worth more than others makes applying the average figure to everyone an attractive alternative.

108

Orley Ashenfelter and Michael Greenstone used the choice of speed limits to assess the implicit value of a statistical life.⁴ They examined the choices of 21 states to raise their speed limits from 55 mph to 65 mph when the federal government permitted that change in 1987. The increase in the speed limit increased highway fatality rates by 35 percent and saved about 125,000 hours of driving time per life lost. By valuing time at the average hourly wage rate, Ashenfelter and Greenstone estimated that the speed limit increase saved $1.89 million worth of time per life lost, suggesting that amount as the upper bound for the implicit value of statistical lives in those states. (States that did not raise the speed limit indicated a higher value for life.) Other value-of-life studies have considered dollar-risk trade-offs involving property values and the mortality effects of air pollution, expenditures on smoke detectors, and expenditures on automobiles with varying accident risks. The results were all in the range of $1 million to $6 million.

⁴ “Using Mandated Speed Limits to Measure the Value of a Statistical Life,” Journal of Political Economy (2004), 112:S1, S226–S267.

Although we can save lives with regulations, if regulations go too far in requiring spending on certain safety measures, they actually cause more deaths than they prevent. The problem is that expenditures on compliance with regulations divert money away from other expenditures, some of which influence our risk of death. Each dollar spent to comply with regulations can’t be spent on organic food, a health club membership, or a car with side airbags. Of course, many of the dollars spent on regulations would otherwise go toward goods and services that have no bearing on the length of one’s life, but health and safety items are among those normal goods that consumers purchase less of when incomes fall.

Consider the Food and Drug Administration’s ban on a cattle-fattening hormone called diethylstilbestrol (DES), which has been linked to cancer. The Office of Management and Budget estimates that the ban on DES in cattle feed costs $185 million per premature death avoided. This cost, which is reflected in the incomes of workers in the cattle industry and is shared by beef consumers via higher prices, competes with expenditures on everything else, including safer vehicles, healthy food, medical care, exercise equipment, and so on. Randall Lutter, John F. Morrall III, and W. Kip Viscusi examined how the demand for lifesaving and life-endangering goods is influenced by income levels and estimated that each expenditure of $17 million on regulations diverts enough money away from life-affecting goods to cause the loss of a statistical life.⁵ Thus, regulations that save lives at a cost of more than $17 million each may also cause the loss of more lives than they save. In fact, each life saved by the DES ban corresponds with the loss of about 11 other lives.

⁵ “The Cost-per-Life-Saved Cutoff for Safety-Enhancing Regulations,” Economic Inquiry (1999), 37:4, 599–608.

109

A popular 1970s television show had the telling title The Six Million Dollar Man. Who knew that each of us is worth approximately that amount? It’s a touchy subject, given our respect for humanity, but we must place values on human lives for the purposes of drawing the line on regulations, compensating the families of victims, and establishing appropriate incentives for those whose decisions affect lives. Even if we chose to devote the entire value of the U.S. gross domestic product to avoiding fatal accidents, we would be able to spend less than $100 million per fatality. In our everyday decisions we express even more modest valuations of statistical lives by accepting risks that earn or save measurable amounts of money. The alternative would be to save lives at all costs and to live in a world with many lives and few freedoms or luxuries. Before you let anyone convince you that it’s cold to place a dollar value on human lives, ask them if they’ve ever taken money that could have been spent on health care and used it to buy a television set instead, or if they’ve ever crossed the street to receive a paycheck.