1. IntroductionGovernment regulation of the market for tobacco products can be justified in a number of ways. Smoking is associated with market failures such as negative externalities and imperfect information among market participants, and these failures provide one rationale for government intervention. Another is the huge health care costs associated with the health consequences of smoking. The cost of medical treatment for smokers and second-hand smokers inflates health insurance premiums for everyone, regardless of smoking participation; in addition, many of these expenses are paid from public funds.Youth is of particular interest for public policy makers and economists who study smoking behavior. The evidence from recent economic studies indicates that adolescents are significantly more responsive than adults to changes in cigarette prices (U.S. Department of Health and Human Services [USDHHS] 1994). In addition, the vast majority of smokers complete their initiation prior to their 21st birthday (Gilpin et al. 1994). Therefore, focusing preventive efforts on youth seems to be the most effective way to achieve a long-term reduction in smoking prevalence. Public policy makers are also concerned with youth-specific smoking externalities. Almost all first use of cigarettes occurs during the high school years (Kessler 1995). At that age, consumers typically underestimate the health consequences of smoking and the risk of nicotine addiction (Kessler 1995; Johnston, O'Malley, and Bachman 2001), thus underestimating the price of smoking to them.The annual prevalence of cigarette smoking in the United States has been declining since the 1970s, stabilizing in the 1990s with approximately 62 million smokers in 1996, which represented 23.2% of the U.S. population (USDHHS 1996). Even though this figure is not high relative to smoking in other countries (the world average smoking prevalence in 1997 was 29%; World Health Organization, 1997), the declining trend in cigarette consumption has substantially slowed in the 1990s.It is particularly troubling that the slight decrease in smoking prevalence among adults in the 1990s was accompanied by an increase in smoking participation among youth and young adults. The evidence of this trend was detected in several nationally representative surveys. For example, the 1997 Youth Risk Behavior Survey (YRBS) reported an increase in average smoking prevalence among high school students from 27.5% in 1991 to 36.4% in 1997. According to the Centers for Disease Control and Prevention, the number of 12th grade high school students who started smoking as a daily habit jumped from 708,000 in 1988 to 1,200,000 in 1996, an increase of 73% (CDC 1995, 1996, 1997, 1998).However, statistics at the end of the 1990s suggest a respite from this trend in increased youth smoking prevalence. The YRBS reported a decline in smoking participation for 9th graders between 1997 and 1999 (by 17%), but very little change for 10th and 11th graders (decline by 2%), and a slight increase in smoking prevalence among 12th graders (by 8%). In 2000, the Monitoring the Future Surveys (MTFS) reported a decrease in smoking participation among all three surveyed high school ages (Johnston, O'Malley, and Bachman 2001).There is an economic explanation for this trend in smoking prevalence among youth. Even though the federal cigarette excise tax was raised twice in the beginning of the 1990s, real prices of cigarettes fell in the subsequent period. Between 1993 and 1996, the real price of a pack of cigarettes, adjusted for inflation, dropped by 10% (Tobacco Institute 1997). Real cigarette prices began to rise again toward the end of the 1990s. Between 1997 and 1999, the real price of a pack of cigarettes adjusted for inflation increased by 48% (Orzechowski and Walker 2000). This sudden change was partly triggered by a new financial liability of tobacco companies toward 46 states under the Master Settlement Agreement (November 1998) amounting to a $206 billion financial burden for the industry over the following 25 years. …