public health burden of tobacco use, especially when history so clearly documents the capacity of the tobacco industry to mount effective countermeasures against and to neutralize potentially effective innovations in tobacco control.

The proposed tobacco policy development office might sensibly be located in the Food and Drug Administration (FDA) after Congress confers on FDA the necessary authority to regulate tobacco products.

One of the policy office’s first tasks should be to foster improvements in tobacco policy simulation models. The current generation of tobacco policy models has been valuable to the committee, and the committee is confident that the projections provided to the committee represent the best that can be produced with the currently available tools. However, the implications of tobacco policy decision making are of sufficient magnitude to warrant a greater investment to help policy analysts advance the state of the art. The current tobacco policy models could benefit from the following improvements.

Smoking initiation is influenced by peers, particularly peers’ rates of smoking. More generally, individuals’ decisions to start, stop, or restart smoking may be influenced by the smoking or nonsmoking of others. These feedbacks from current prevalence to various flow rates include personal interaction effects (e.g., when smoking teens encourage their friends to smoke), societal-level effects (e.g., if smoking is rare, it is more likely to be shunned, which might reduce relapse rates), and market-level effects (e.g., if there are fewer smokers, and hence less demand for tobacco, the market equilibrium price for cigarettes might be affected, which in turn can affect smoking rates).