The fascinating connections between cigarette addiction and educational attainment

Listen to this article

When it comes to cigarettes, public policy analysts would like to know the answers to two fundamental questions. First, does addiction to cigarettes attenuate educational attainment? Second, does education reduce smoking? Thought-provoking new research by Rong Hai and James Heckman provides credible answers to these two questions. The authors begin by pointing out that the relationship between cigarette smoking and education is prima facie apparent in survey data — young people who begin smoking earlier obtain fewer years of schooling, and more highly educated adults smoke at lower rates. However, the key problem is showing a causal connection between these two things. This is because of what economists call selection bias (shared underlying factors may drive both choices) and dynamic selection (past choices may reshape future preferences). To address these challenges, the authors estimate a structural life-cycle model embedded in a rational addiction framework.

This model tracks young men from age 15 to 64, during which time they make dynamic decisions about smoking, school enrollment, part-time work while in school, consumption, and savings, all subject to borrowing constraints. A key feature of the model is the “addiction stock,” which accumulates with smoking and depreciates over time. This stock affects current utility from smoking and, critically, shrinks a person’s subjective discount factor — making addicted individuals more short-sighted and therefore less inclined to invest in education, which yields returns over the long run.

Education, in turn, reduces the flow utility of smoking, increases the discount factor, raises future earnings, and relaxes the budget constraints confronting young men. Three channels thus link smoking negatively to education: addiction raises myopia, smoking costs divert financial resources from tuition, and agents anticipating that education will reduce their smoking enjoyment may rationally invest less in schooling. Education affects smoking through analogous channels in reverse. Individual heterogeneity is modeled through cognitive and non-cognitive abilities as well as parental education. These endowments shape preferences for schooling and smoking, the discount factor, and earnings capacity.

The authors use data from the National Longitudinal Survey of Youth 1997 (NLSY97), covering 1,605 males from age 15 through their early 30s, merged with state-year cigarette price and excise tax data from the CDC. The econometric analysis undertaken shows that the model fits the data well.

Consider first the effect of education on smoking. Assigning each individual one additional year of schooling at age 20 reduces the smoking rate at that age by 4 percent. By age 30, the cumulative effect grows to an 8 percent reduction. The treatment effect is strongest among individuals with medium cognitive and non-cognitive abilities, and non-cognitive skills — reflecting traits like self-control and discipline — are the primary driver. A one standard deviation increase in non-cognitive ability raises the likelihood of quitting smoking by 14 percent.

Next, consider the effect of smoking on education. In a counterfactual where smoking is eliminated as a choice, the college attendance rate among benchmark ever-smokers rises by 6 percent. Across the full population, college attendance rises by 3 percent. Interestingly, while non-cognitive ability is more instrumental in avoiding harmful behavior when smoking is possible, cognitive ability becomes the dominant driver of college attendance in a world without smoking.

Finally, consider the use of excise taxation as policy. A more practical policy — a 40 percent additional excise tax on cigarettes, with revenues redistributed as tuition subsidies — achieves the same increase in college attendance among would-be smokers as eliminating smoking entirely. It also reduces the age-30 smoking rate from 27 percent to 24 percent. The combined tax-and-subsidy design leverages both the deterrent effect of higher cigarette prices and the incentive effect of reduced tuition costs.

Importantly, this research establishes that youth cigarette addiction and educational attainment are causally and reciprocally linked, with effects that compound over the life-cycle through a variety of factors such as the discount factor, personal utilities, and budget constraints. Policy impacts are highly heterogeneous across cognitive and non-cognitive ability levels, suggesting targeted interventions may be most effective. The authors rightly point out that their research framework is applicable to other addictive behaviors — such as marijuana and e-cigarettes — whose long-run economic consequences are not yet observable in data, thereby making this analysis a useful template for future research and the design of public policy.

Batabyal is a Distinguished Professor and the Interim Director of the Golisano Institute for Sustainability in the Rochester Institute of Technology but these views are his own.