The Long-run Impact of Tobacco Control Policies on Smoking Initiation

Tobacco control policies such as cigarette excise taxes, smoke-free air laws, and youth access laws gained popularity in the 1990s and were expanded further in the 2000s. Studies show that tobacco control policies affect the smoking behavior of adults, but adolescents are generally unresponsive. There is also a high correlation in the smoking behavior of parents and their children. If this correlation represents a causal relationship, then we would expect policy-driven changes in parent smoking behavior to also affect the smoking behavior of their children. In this paper, we consider policy changes that occur during childhood (when an individual is age 7 and younger), which have no direct effect on their own smoking behavior, but potentially affect the behavior of parents, older siblings, or other older role models. We consider individual state-level tobacco control policies and a novel tobbaco control policy index that weights each policy by its relative importance in predicting the smoking prevalence rate in each state. We estimate a discrete-time hazard model of the effect of childhood tobacco control policies on the age of smoking initiation, while controlling for the direct effects of these same types of policies at later ages. Individual-level data on smoking histories come from the Panel Study of Income Dynamics and National Longitudinal Survey of Youth 1979 Children and Young Adults. We find that a one standard deviation increase in the childhood tobacco control policy index decreases the proportional hazard of initiation by between 11 and 15 percent, while a similar increase in the contemporaneous tobacco control policy decreases the proportional hazard of initiation between 14 and 16 percent. When we relax the proportional hazards assumption we find evidence that these decreases are driven by early initiators (age 10-12) delaying to slightly older ages (age 13-14) and some late initiators being detered entirely. Our analysis of individual tobacco control policies reveals that increases in cigarette taxes during childhood have a large affect on smoking initiation but increases in taxes at later ages have no effect. The strong indirect effect of tobacco control policies suggests that previous studies focusing solely on immediate effects dramatically underestimate the total impact of these policies due to the spillover effect on the rising generation.