ITHACA, N.Y. -- Boosting taxes on cigarettes will have a far less dramatic effect on rates of teen-age smoking than politicians are hoping, a new Cornell University study finds. In fact, say the researchers, higher taxes will have "a statistically insignificant impact" on whether young people decide to start smoking.
The Cornell researchers say that a 20-cent a pack tax increase would reduce the number of new teen-age smokers in grades 8-12 by less than one-half a percentage point. They estimate that a $1.50 a pack tax boost would reduce the number by about two percentage points. Some researchers have estimated this tax increase would reduce the number by as much as half.
The study was carried out by Don Kenkel and Alan Mathios, both economists and associate professors in policy analysis and management in Cornell's College of Human Ecology, and Phil Decicca, a graduate student in their department.
Their findings are particularly relevant in light of the bill approved by the U.S. Senate Commerce Committee last week that would raise cigarette prices by $1.10 a pack over the next five years with hopes of cutting teen-age smoking by 40 percent over the next 10 years. Also, President Clinton has said he favors price hikes by as much as $1.50 a pack in order to cut youth smoking by half over the next 10 years. Although smoking has dropped to 25 percent of U.S. adults from 40 percent since 1964, teen-age smoking has been rising since 1990.
"This reliance on higher prices as a way to discourage youth smoking has widespread support, yet our study indicates that higher prices don't work to deter youth smoking," says Mathios.
"There is good evidence that higher taxes reduce total tobacco sales and adult smoking," says Kenkel, a health economist with strong interests in public-policy issues. "Higher taxes reduce adult smoking by encouraging current smokers to cut down or quit." However, he says, the latest study shows that "higher taxes will have a statistically insignificant impact on youth decisions to start smoking."
The Cornell economists used what they call statistical predictive modeling techniques to make a computer analysis of data from the National Education Longitudinal Survey, a 1988 study of about 12,000 students in grades
8, 10 and 12. "This allowed us to study the impact of taxes on smoking behavior during exactly the time in adolescence when most smokers start their habit," says Mathios.
The researchers' findings were at variance with previous studies that have found that price hikes do in fact curb teen-age smoking. However, says Mathios, most of these studies attempted to analyze how youth smoking rates were affected by differences in cigarette taxes from state to state. "The problem with this cross-sectional approach, however, is that each state has different antismoking sentiments which could be influencing the onset of smoking among youth," says Kenkel.
"For example, the tobacco-producing states have the lowest tax rates. We also find that youth smoking is higher in states that have smokers' rights laws, again probably because there is less stigma attached to smoking in these states," Kenkel says. "Our data allowed us to directly examine the impact of changes in tax rates on youth smoking behavior and our results indicate this impact is small or nonexistent."
The Cornell researchers found that when they looked at the original 1988 national survey data they found that higher taxes were indeed linked to lower teen-age smoking rates. But when they also analyzed data from the survey's two biennial followups, and when they eliminated 8th graders from the sample -- 95 percent of whom had not started to smoke in 1988 --the effect that price increases seemed to have on teen smoking disappeared.
The researchers presented their findings at the American Economics Association annual meeting in Chicago in early January. The research was supported by the Bronfenbrenner Life Course Center at Cornell and the National Cancer Institute.
The above story is based on materials provided by Cornell University. Note: Materials may be edited for content and length.
Cite This Page: