Smoking cessation: a model of planned vs. actual behavior for time-inconsistent consumers

We offer a simple model of intertemporal choice to characterize how planned versus actual behaviors evolve for time-inconsistent smokers. Our results suggest that smokers' participation and cessation decisions are governed by the interplay between three effects. The cessation effect leads smokers to advance their plans to quit smoking, whereas the procrastination effect leads them to consecutively revise their planned quitting age upwards. Consequently, the duration of smoking is effectively governed by which one of these two effects is dominant. Finally, for certain consumer segments, a threshold effect causes an "all or nothing" type of extreme smoking behavior based on certain critical values of present-biased preferences. Our results provide some preliminary evidence that both marketing efforts by tobacco firms and public policy initiatives can have a significant influence on smoking behavior. In particular, we find that reductions in the age at which individuals start smoking may not only vastly extend their duration of smoking, but also convert potential "never smokers" into lifetime smokers. Finally, we estimate a hazard model using survey data from over 800 smokers to provide evidence in support of our theoretical model.