ing does seem to be strongly linked to the level of adult drinking, and the level of adult drinking—at the very least—probably places clear upper bounds on the effectiveness of any given set of policies to control underage drinking. This evidence highlights the challenge, to which we referred in Chapter 1, of trying to suppress underage drinking in a culture in which drinking is normative behavior.

The relationship between underage drinking and adult drinking is relevant to our charge for a second reason. Since the level of adult drinking might be an important determinant of underage drinking, it is at least logically possible that the most “cost-effective strategy to reduce underage drinking” includes policies that produce their main effects not on underage drinking, but rather on the overall level of drinking in the population. The question to be faced, then, is whether to construe our mandate (to propose a cost-effective strategy to reduce and prevent underage drinking) as including: a review of all policy instruments that could produce an effect on underage drinking, including those that are not directed specifically at underage drinking, such as taxes and other general policies affecting price and availability, or a review only of policy instruments that are specific to underage drinking, such as the enforcement of laws prohibiting underage drinking, or the development of special media campaigns targeted only on underage drinking, or the strict regulation of venues in which underage drinking is most likely to occur.

The committee decided that it would focus on policies specifically aimed at underage drinking, but that it would not close its eyes, categorically, to policies that affect all drinking. Instead, we have carefully reviewed the evidence regarding the effects of general alcohol policies on underage drinking and have included in our proposed strategy one of these general components (raising excise taxes) because a substantial increase can be expected to have a robust impact on underage drinking and can also strengthen the nation’s capacity to implement a strategy aiming to reduce underage drinking.

Some people have argued that recent declines in underage drinking negate the need for significant new interventions. As noted in Chapter 2, the prevalence of alcohol use in the past 30 days among high school seniors has declined from a high of about 72 percent in 1979 to about 49 percent in 2002; similarly, the prevalence of heavy drinking within the past 2 weeks has declined from a high of 41 percent in 1981 to about 29 percent in 2002. The proportion of youth fatalities involving alcohol-involved underage drivers has also declined, from 55.8 percent in 1982 to 30.1 percent in 2000, although there has been little change in recent years (National Highway