SOCIAL COSTS OF ALCOHOL AND DRUG ABUSE

Drinking, smoking, and the use of psychotropic drugs have a variety of consequences for those who partake of them, for their families and associates, and for society at large. A number of these consequences are negative. Smokers die young from heart or lung disease, drinkers get into traffic accidents and fights, drug injectors spread the HIV virus. In the context of public policymaking, where priorities must be set for the use of scarce resources, it seems important to have a measure of the overall magnitude of the social burden engendered by such consequences. One familiar approach is to express the magnitude of the problem in terms of the number of people who die each year. When we learn that there are 107,400 deaths per year in the United States from Alcohol abuse (Harwood et al., 1998) and perhaps four times that number from Tobacco use, we know that the stakes are very high in devising sound policies for controlling drinking and smoking. Such statistics, compelling as they are, tell only part of the story. In addition to causing early death, substance abuse makes for a variety of consequences that reduce the quality of life, both for users and other people.

To capture this broad array of consequences in a single number, analysts have estimated various measures of social cost. The estimates are important because they figure in the political process by which federal funds are allocated to the National Institutes of Health and to other agencies that play a role in combating substance abuse. The most prominent estimates of social costs for substance abuse have utilized a conceptual apparatus developed by a task force of the U.S. Public Health Service chaired by Dorothy Rice (Hodgson & Meiners, 1979). In 1994, the International Symposium on the Economic and Social Costs of Substance Abuse issued guidelines recommending the use of this cost-of-illness method in an attempt to establish a common foundation and enhance the comparability of cost studies conducted in different countries (ICAP, 1999).

Although prominent in policy debate, the cost-of-illness (COI) method has been faulted for its emphasis on production as the measure of social welfare. Economists favor a quite different approach that measures social welfare from the perspective of the consumer. The economists' preferred accounting framework is referred to in this article as the "external social-cost" approach.

THE TWO FRAMEWORKS APPLIED TO SUBSTANCE ABUSE

A coherent assessment of the social costs of substance abuse requires an accounting framework that specifies criteria for judging which of the myriad effects are properly deemed to be of public concern. For example, in the case of drinking, on any one drinking occasion there may be unwanted, harmful consequences: social embarrassment, loss of reputation or affection, failure to discharge some responsibility at work or home, physical injury from an accident, victimization by a mugger or rapist, and nausea or hangover. Chronic heavy drinking may result in still other consequences, including rejection by family and friends, loss of a job or of an opportunity for promotion, progressive deterioration in physical health, and an early death. In order to capture these and other negative consequences in a single number, the list of consequences must be reviewed to determine which should be considered in establishing priorities for substance abuse policy. The consequences deemed relevant must then be quantified, translated into a standard unit of account (dollars), and summed.

The Cost-of-Illness Framework.

The COI approach is concerned with measuring the loss or diversion of productive resources resulting from an illness or activity. In the case of alcohol abuse, human capital resources are lost and the gross national product reduced by the morbidity and early death suffered by some drinkers, whether as a result of injuries sustained in alcohol-related traffic accidents or violent crime or as a result of organ damage and other diseases stemming from chronic heavy drinking. The loss to society in these cases is equal to the loss of the marginal product of the victims' labor, valued at the market wage. Unpaid work at home, including housework and child care, is included in the computation, with values being assigned according to how much households pay for such services when they are performed by paid help.

The COI approach also takes account of the diversion of resources from other productive uses necessitated by alcohol abuse. Thus the costs of medical care for alcohol-related illness, treatment for Alcoholism, and research on prevention and treatment are incorporated in the social-cost estimate. Similarly, the value of law-enforcement and justice resources devoted to alcohol-related crimes are included, as are the costs of replacing property damaged in traffic crashes and fires caused by drinking.

Several prominent estimates of the total costs of alcohol abuse for the United States have utilized the COI framework (Berry & Boland, 1977; Harwood et al., 1984 & 1998). In 1998, Harwood et al. published the most complete COI study to date. Using figures from 1992, the most recent year for which complete data were available, they found that the economic costs to society of alcohol abuse totaled $148 billion, broken down as follows:

About three quarters ($107 billion) of the total cost in this tabulation is the value of labor Productivity lost as the result of illness, injury, or early death. The human capital lost as a result of alcohol-related mortality was computed for all those who died in 1992 from causes in which intoxication or chronic heavy drinking played a role. These include traffic fatalities and deaths from liver cirrhosis, among other causes. The lost human capital was valued by estimating how much the deceased would have earned if they had lived and worked until retirement age.

The human capital lost as a result of morbidity was calculated by estimating the reduction in the productivity of the labor force resulting from alcohol dependence or abuse. Harwood et al. combined two sets of estimates to arrive at this number: first, the percentage of the labor force in 1992 that was or had ever been subject to a diagnosis of alcohol dependence or abuse; and second, an estimate of the loss in earnings associated with such a diagnosis.

Critique.

Estimates of this sort have been challenged for two reasons. The first challenge is to the statistical methods used to generate the estimates of morbidity, mortality, and lost earnings (Cook, 1991). The second challenge is more fundamental, for it concerns the basic principles that inform the COI accounting framework.

The COI procedure estimates the cost of morbidity and mortality in terms of lost productivity, but this emphasis on production as the measure of social welfare seems misplaced. A more liberal perspective, favored by economists among others, shifts the emphasis to consumption and interprets the task of measuring social welfare in terms of aggregating individual preferences. Consumers are the best judges of their own welfare, and if sometimes they make choices that fail to maximize their productivity, that should not in itself be regarded as problematic. In this view, the choices that people make concerning how hard to work and when to retire are of little public concern. The same goes for choices that place one's own health and safety at risk. Thus in economics there is a strong presumption in favor of consumer sovereignty, the principle that the individual consumer is in the best position to define what is best for him or her, and that social welfare is enhanced by free choice within certain limits. A negative consequence is deemed to be of public concern only when the actions of one individual impinge negatively on the welfare of others. The basic distinction, then, is between internal and external consequences of individual decisions, where the latter impose an involuntary cost on other people.

In the case of alcohol abuse, the internal costs include those suffered by drinkers and are foreseeable as a natural consequence of their choices. A small example explains the reasoning here. Suppose a woman decides to drink heavily tonight despite knowing that she may be tired and unproductive tomorrow. By making this decision, she is indicating that for her the pleasure of partying outweighs the "morning-after" costs. If no one else is harmed by this decision, the external costs are zero. If she were to drive after drinking, however, the accounting would change. She would be risking serious injury to herself and to others on the highway. Her injury would have external costs to the extent that a third party (group insurance or Medicaid) paid her medical expenses. The risk that she might injure other people while driving is also a negative externality, to be valued at the expected loss to them. That cost, incidentally, is not limited to their lost earnings, but also includes their pain and suffering and the suffering of those who care about them.

In sum, the most fundamental challenge to the COI framework relates to its presumption that social welfare is synonymous with national product. Economists argue instead that the preferences of individuals are the proper measure of their well-being and that social welfare is the sum total of individual welfare. Some of the major costs in the COI framework, especially lost earnings, are less important in the external social-cost view, whereas a number of costs that are ignored in COI become important when the focus is on external costs.

The External Social-Cost Framework.

In a study at the Rand Corporation, economists applied the ESC framework to alcohol abuse and other poor health habits (Manning et al., 1989, 1991). Their estimate for alcohol abuse amounted to about $30 billion in 1985, less than half the COI estimate presented above for the same year. The accounting procedures used to generate this estimate of the ESC can be briefly summarized:

1. Earnings . Heavy drinkers might earn less than they otherwise would have during their careers and might have their careers cut short by poor health and early death. Although the most obvious effect was a reduced standard of living, which was properly considered a private cost, a number of programs created a collective interest in the productivity of each individual. For example, those who died young saved their fellow citizens the expense of years of pension payments and medical costs. Those who retired early (perhaps because of poor health) imposed financial costs on others in the sense that their contributions to the Social Security system were reduced. Thus these collective financing arrangements had the effect of creating both external costs and benefits in relation to heavy drinking. The net effect, according to Manning et al. (1991), was negative, and equaled about 22 percent of the total external cost.
2. Traffic Fatalities . Heien (1996) reported that about 3,765 of the 13,984 people who died in alcohol-related traffic accidents in 1993 were "innocent," in the sense that they had not been drinking at the time. Their lives had value not because their work increased the size of the national product, but because they enjoyed life. People are willing to pay to reduce the risk of a fatal accident, and the social cost of these innocent deaths is in principle equal to the total amount the public would be willing to pay to eliminate the threat of being killed by a drunk driver. Manning et al. (1991) employed this willingness-to-pay approach and found that nearly half of the social cost of alcohol abuse stemmed from traffic fatalities.
3. Other Costs . The remaining $7.2 billion in Manning et al.'s (1991) social cost estimate stemmed primarily from the burden of alcohol-related cases on the criminal justice system, and the share of collision insurance costs accounted for by the property damage caused by drunk drivers.

It appears that in several respects these estimates are incomplete. The costs of alcohol-related injuries to innocent victims are far higher than indicated by Manning et al., since they omitted the financial and personal costs of nonfatal injuries in traffic accidents (Miller & Blincoe, 1993), and also the costs of both fatal and nonfatal injuries from violent crimes perpetrated by drunks. In addition, recent research has suggested that moderate alcohol consumption carries measurable health benefits, which must also be figured into any equation attempting to assess social costs (ICAP, 1999).

An even more interesting controversy has arisen over the basic perspective that informs these external social-cost estimates. Some critics reject outright the liberal doctrine that individual preferences are to be accorded primacy in the definition of social welfare and social cost. They postulate a collective interest that can somehow be defined without reference to the choices made by individuals (Beauchamp, 1980). The COI approach reflects one such definition. Other critics accept the liberal doctrine but argue about its application. A particularly difficult set of philosophical and practical issues arise in setting the boundary between internal and external costs in the context of the family. Manning et al. (1991) view the family as a unit and accept the presumption that each member of the family will internalize the concerns of the others and act accordingly. Harwood et al. found that, in 1992, abusers and their households bore $66.8 billion of the total cost of alcohol abuse. If the father is a heavy drinker or smoker, it is not because he is unaware or unconcerned about the consequences for his wife and children of his drinking or smoking, but because his enjoyment of these activities in some sense outweighs the costs to them. That presumption may seem particularly problematic in the case where the mother's substance abuse causes her baby to be born defective.

COSTS OF SMOKING AND DRUG ABUSE

Manning et al. (1989) provided an estimate of the social costs of smoking that utilized the same general approach as their estimate of drinking costs. They found that over their lifetime smokers experienced higher medical costs than they would have if they had never smoked, amounting to an average of $0.38 per pack. Since these costs were for the most part paid by insurance, government programs, or other collective sources, they included them in the external social-cost estimate. Other important external costs were the reduced contributions to the Social Security system and related programs ($0.65 per pack) resulting from the early termination of the average smoker's career, and the increased cost to group life insurance programs resulting from the reduced life expectancy of smokers ($0.11 per pack). Interestingly, these external costs were much less than the external benefits conferred by smoking. Because smokers died young, the pension payments were much less than they would have been otherwise ($1.82 per pack), and the likelihood that they would be housed in a collectively financed nursing home was also substantially reduced ($0.26 per pack). The result was that each pack of cigarettes smoked conferred a net social benefit amounting to $0.91.

The calculations used to arrive at these figures are quite complex. Cigarettes smoked in different years may have variant health effects. Tar content in cigarettes, for example, has decreased three to four percent since World War II. It is generally believed that cigarettes containing lower amounts of tar cause fewer health problems. Since over the course of a smoking career the social costs generally precede the benefits, the net benefit to society was reduced if future costs and benefits were discounted (standard practice in accounting). The appropriate discount rate to be applied to these calculations is a matter of some dispute. It turned out that with a discount rate of five percent, the lifetime present value of the external effects of smoking amounted to a net external cost of $0.15. Manning et al. point out that smokers more than pay this cost in the form of the state and federal excise taxes imposed on tobacco. The external effects in this calculation are all financial; they stem from private and government programs that have the effect of forcing us to pay for each other's medical care, retirement, and other benefits. Smoking, however, also causes external effects directly, since smoke pollutes the air we all breathe. The value of clean air for non-smokers could in principle be estimated and added to the total external cost. Manning et al. chose not to do so, in part because they believe that the bulk of the costs of secondhand smoke is borne by those in the same household as smokers. However, in 1995, taking into account the consequences of second-hand smoke, Viscusi brought the estimate of the net social benefit of smoking down to a more modest, but still beneficial, $0.07 per pack, assessing the costs of second-hand smoke at $0.25 per pack.

Applying the external social-cost framework to smoking and other harmfully addictive activities raises another issue. The vast majority of smokers begin their habit as adolescents, so the obvious question is whether people at that age are making well-informed decisions that take proper account of the lifetime consequences (Goodin, 1989). Adolescents tend to be as well informed about the health risks of smoking as adults, and both groups, if anything, exaggerate these risks (Viscusi, 1992). However well informed they are, most people who acquire a smoking habit nevertheless end up wishing they could quit.

In considering the social costs of illicit drug use, the illegal status of these drugs makes an enormous difference (Kleiman, 1992). The consequences of criminalizing transactions in these drugs include the bloody wars between rival drug-dealing organizations, crime by addicts seeking funds for their next fix, and the spread of disease through use of unclean needles, as well as the billions of dollars spent in law-enforcement efforts. Harwood et al. estimated that, in 1992, drug abuse problems incurred a social cost of $97.7 billion.

CONCLUSION

In conclusion, the effort to produce estimates of the social costs of drinking, smoking, and drug abuse is motivated by an interest in establishing a scientific basis for setting priorities in government programs. This effort has produced some useful results and a good deal of controversy surrounding the issue of what is to be counted and how. The task of estimating the social costs of substance abuse requires an accounting framework, and the choice of a framework is not a technical, scientific issue but rather a matter of political philosophy. This is surely one area where the numbers do not speak for themselves.

(See also: Accidents and Injuries ; Complications ; Economic Costs of Alcohol Abuse and Alcohol Dependence ; Productivity: Effects of Drugs and Alcohol on )

BIBLIOGRAPHY

Beauchamp, D. E. (1980). Beyond alcoholism: alcohol and public health policy. Philadelphia: Temple University Press.

Berry, R. E., & Boland, J. P. (1977). The economic cost of alcohol abuse. New York: Free Press.

Cook, P. J. (1991). The social costs of drinking. In Expert meeting on the negative social consequences of alcohol abuse. Oslo: Norwegian Ministry of Health and Social Affairs.

Goodin, R. E. (1989). No smoking: The ethical issues. Chicago: The University of Chicago Press.

Harwood, H. J., et al. (1998). The economic costs of alcohol and drug abuse in the United States: 1992. Rockville, MD: The National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism.

Heien, D. M. (1996). Are higher alcohol taxes justified? The Cato Journal, 15 (2-3).

Hodgson, T., & Meiners, M. (1979). Guidelines for cost-of-illness studies in the public health service (Task Force on Cost-of-Illness Studies). Bethesda, MD: Public Health Service.

International Center for Alcohol Policies. (1999). Estimating costs associated with alcohol abuse: Towards a patterns approach. ICAP Reports, 7.

Kleiman, M. A. R. (1992). Against excess: drug policy for results. New York: Basic Books.

Manning, W. G., et al. (1991). The costs of poor health habits. Cambridge, MA: Harvard University Press.

Manning, W. G., et al. (1989). The taxes of sin: Do smokers and drinkers pay their way? Journal of the American Medical Association, 261, 1604-1609.

Miller, T. R., & Blincoe, L. J. (1993). Incidence and cost of alcohol-involved crashes. Unpublished manuscript.

Rice, D. P., et al. (1990). The economic costs of alcohol and drug abuse and mental illness: 1985 (Report submitted to the Office of Financing and Coverage Policy of the Alcohol, Drug Abuse, and Mental Health Administration, U.S. Department of Health and Human Services). San Francisco: University of California, Institute for Health and Aging.

Rice, D. P. (1999). Economic costs of substance abuse, 1995. Proceedings of the Association of American Physicians.

Viscusi, W. K. (1995). Cigarette taxation and the social consequences of smoking. Tax Policy and the Economy, 9. Cambridge, MA: National Bureau of Economic Research.

Viscusi, W. K. (1992). Smoking: Making the risky decision. New York: Oxford University Press.

Philip J. Cook

Revised by Sarah Knox