The Behavioral Dynamics of Youth Smoking

While individual smoking behavior persists over time, it is unknown whether this repeated behavior is due to addiction or individual propensities to smoke. To address this issue, we develop a dynamic empirical model of smoking decisions which explicitly accounts for the impact of previous smoking behavior and allows for unobserved individual heterogeneity. The model is estimated using longitudinal data on a representative sample of teens from all 50 United States from 1988 to 1992. We find that current smokers are both more likely to continue smoking and are less price sensitive than current non-smokers. For example, smoking in 8th grade (as opposed to not smoking) increased the probability of smoking two years later three fold, while smoking participation rates are double four years later. The estimated price sensitivities of previous non-smokers and previous smokers are -0.32 and 0.08, respectively. This suggests that a cigarette price increase will have a larger aggregate effect in the long run than in the short run as more individuals accumulate in the price-sensitive non-smoking group. In total, a dollar increase in cigarette prices reduces (age 18) smoking participation predictions by four percentage points more when unobserved individual heterogeneity and behavior modification associated with previous price changes are taken into account than when they are ignored.