State tobacco taxation, education and smoking: Controlling for the effectsof omitted variables

There is a potential bias in cross-sectional estimates of the effects of ci garette prices on cigarette consumption. States with the strongest antismok ing sentiment will likely have the highest cigarette taxes, which result in the highest prices. Some of the lower consumption of cigarettes ill high-t ax stairs will result from such sentiments, rather than from higher taxes, so the estimated effect of cigarette taxes on consumption will be overstate d. This study corrects for such bias, employing panel data for U.S. states from 1960 to 1990. We find that controlling for this bias reduces the estim ated consumer response to cigarette price change by 40-50 percent.