Interventions available to policymakers aim either to lower lifetime consumption of alcohol or to limit the number of cases of periodic intoxication. This is in accord with the classification discussed earlier (see Section II). Policy interventions related to prolonged consumption (group I) can be grouped into ex ante interventions designed to lower the probability of contracting cirrhosis or cancer, primarily by lowering per capita consumption, and ex post treatments to extend the number of years of life and to reduce pain and suffering by providing health services, such as chemotherapy or palliative care.
Given that the majority of the direct effects related to group 11 problems are accident and motor vehicle related, policies oriented toward a reduction in the number of cases of periodic intoxication aim to lower the morbidity and mortality with drinking and driving accidents. Because the number of offenses committed by an individual is inversely related to the cost of each offense, many of the policy interventions attempt to reduce drunk driving ex post by imposing mandatory penalties (pecuniary and non-pecuniary) for conviction of driving under the influence, open container laws, and minimum legal drinking ages for alcoholic beverages. Ex ante efforts to limit periods of intoxication include regulations on advertising, excise taxes, and information and education campaigns to inform consumers of the potential hazards of excessive drinking. It is in this area that the most promise for reducing alcohol related problems lies.
It is reasonable to assume that the full costs to society of alcohol consumption in the form of negative externalities are not fully reflected in a free market price. Given that one response of government to an externality is the imposition of a tax that causes producers of the externality (alcohol abusers) to internalize the additional social cost (or benefit) of their actions (Pigou, 1947), it would seem that setting prices so that private costs and social costs are in alignment would be an excellent government policy. Moreover, existing econometric evidence suggests that the demand for beer, wine and spirits is price inelastic-between -0.11 for beer and -0.61 for spirits-and the inverse elasticity maxim of public finance would therefore suggest higher proportional taxes on these commodities'. However, price elasticities should not be considered in isolation from other influences of consumption. The evidence of income elasticities greater than 2.0 for spirits and as high as 1.9 for wine suggests that any effect on consumption of an excise tax could be rapidly eroded with income growth. Further, even if alcohol users face all the relevant costs of their actions consumption could still be excessive if users underestimate the potential hazard due to imperfect information.
From a public health perspective changing tax rates is also a good way to influence consumer behavior. In addition to reducing consumption, taxation of alcohol production and consumption also generates government revenues through taxes on industry and excise taxes. Heavy taxation, on the other hand, may stimulate illicit production and black market sales as witnessed by the period of prohibition in the United States, and more importantly, reduce the disposal income of heavy drinkers and their families as a larger and larger share of disposable income is spent on alcohol.
The effectiveness of these interventions can evaluated by following the ex ante/ex post framework developed in the previous section. The effectiveness of ex post treatments associated with cancer of the liver and esophagus and cirrhosis of the liver is approximated using data from Barnum and Greenberg on the change in survival rates for cancer in the United States.' From 1950 to 1980, the relative survival rates for esophageal and liver cancer show little improvement. Due to the level of technology needed to treat cirrhosis, it is likely that cirrhosis survival rates in developing countries are also low, and that the years of life gained (YLG) through curative treatments is, therefore, likely small as well.
The effectiveness of non-curative interventions can be estimated by globally measuring the number of deaths averted by various policies. Although outcomes are measured in terms of the number of deaths averted, the overall objective of these interventions is to lower the consumption of alcoholic beverages. The efficacy by which lower consumption reduces the number of alcohol-related problems is highlighted in the historical examples of Prohibition in the U.S. and wine rationing in France between 1942 and 1948. In the latter case, upon the enactment of wine rationing, the cirrhosis mortality rate in Paris declined from 35/100,000 in 1941 to a low of 6/100,000 in 1945 and 1946, quickly returning to its forma level following the repeal of wine rationing in 1948 (ferris, 1967, p. 2077)
To the extent that reductions in consumption translate into reductions in alcohol-related problems, one of the most promising non-curative interventions is taxation. Evidence indicates that adjusting the tax rates to cover the social and economic costs of alcohol abuse is an effective measure in reducing the burden on society by raising the price of alcohol and consequently lowering demand. Yet, surprisingly, there are very few studies which examine the manner in which increased prices lower alcohol consumption. Cook and Tauchen examined cirrhosis mortality data and consumption by chronic heavy drinkers and found that alcohol taxation can be used as an effective policy lever. In particular, they found that a one dollar increase in the liquor excise tax per gallon proof reduces the liver cirrhosis mortality rate by 5.4 percent in the short tam and perhaps twice that in the longer term. Using 1986 as the base, this translates into between 1,400 and 2,600 deaths averted. In a related study, Charles Phelps, 1988, estimated the optimal tax rates for beer in the U.S. He also concluded that beer taxation reduces both beer consumption by youths and the auto fatalities they cause; fatalities drop by 33 percent in response to a 20 percent tax and 40 percent in response to a 30 percent tar'. Given the weak tax administration in many developing countries, eagerness to raise taxes on alcohol should be tempered by the recognition that there are many other ways to lower alcohol consumption.
A 1991 study by Saffer and Grossman estimated the impact of a range of policy options on the number of motor vehicle fatalities in the U.S. The results indicate that the most effective policies are increased alcohol taxes and mandatory administrative license actions. Through various policy simulations, they found that increasing the beer tax to a value equivalent to the real 1951 value would have averted 5,174 deaths (32% of these deaths would be among the 18-20 year old age group); while mandatory license sanctions would have reduced fatalities by 9 percent or 4,202 deaths (19% in the 18-20 year old range). The next most effective policies were a minimum legal drinking age and relatively large mandatory fines, which reduced fatalities by about 5 to 6 percent. Laws such as seat belt requirements, mandatory jail sentences, and open container laws had a very small deterrent effect on drunk driving.
Information and education campaigns and restrictions on advertising also change behavior-though there is very little empirical evidence. The evidence that does exist (Saffer, 1989) shows that countries which have adopted bans on beer and wine advertising have about 23 percent lower alcohol consumption and traffic fatalities than countries with no bans.
Cost-effectiveness, While a rigorous cost-effectiveness analysis is beyond the scope of this paper, the cost-effectiveness of the alternative policy options should be considered. Although the evidence presented must be interpreted the large differences that exist in the impact pa dollar spent on alternative interventions points to the appropriate government response. Analyzing the cost per YLG estimates presented by Barnum and Greenberg, it is clear that the high cost treatments associated with esophageal and liver cancer, as well as cirrhosis, are relatively cost ineffective. For example, the cost per YLG from treatment for liver cancer is over 100 times that for cancers of the breast, mouth, and cervix. Furthermore, in a developing country setting, the marginal cost of such a treatment is surely much higher, as resources directed to high cost treatments divert valuable resources from basic activities, such as immunizations and the provision of safe water, which have large positive externalities.
On the other hand, policy changes which either alter the consumption of alcoholic beverages, e.g. increased excise taxes or advertising, or alter the behavior of intoxicated consumers, such as a one year license sanction and a $500 fine for driving under the influence, have a low cost pa death averted. In fact, in many cases the only cost associated with the policy change is the political cost plus a small increase in administrative or personnel costs. The fact that a large share of the deaths averted by such policies come from the 18-24 age group, primarily due to motor vehicle fatalities, indicates that the cost pa discounted year of life gained is probably even lower expected. Because it appears the cirrhotic process can be slowed or accelerated in response to changes in the rate of alcohol consumption, policies which decrease pa capita consumption are likely to be highly cost-effective'.
Final considerations in designing government policy, Designing a government policy to diminish the impact of alcohol-related problems, using many of the policy instruments discussed earlier, is a complex, multi-dimensional problem. It is critical to remember that reductions in alcohol consumption will affect the economy in a variety of ways. To name a few: government revenues will fall (excise taxes on alcohol account for anywhere from 5 to 15 percent of government revenues), government will, therefore, need to encourage an increase in the production of other consumer goods to meet the increasing demand for goods which are substitutes for alcohol. Furthermore, reductions in alcohol-related deaths may have implications on the social security or insurance markets. The following box examines the impact of alcohol policy in the forma Soviet Union to highlight a few of the issues and potential problems related to the design of alcohol policy.
[Ukrainian] [English] [Russian]