Vertical equity consequences of very high cigarette tax increases

"Traditionally, cigarette excise taxes have been seen as regressive, due to both the higher prevalence of smoking among lower income groups and the regressivity of any sales or excise tax. One challenge to this view says that "cigarette tax increases may not be regressive," because poorer individuals are more elastic, and therefore may cut back sufficiently to make the share of income spent on cigarette taxes by the rich increase by more than that spent by the poor. We test this challenge empirically. First, we estimate how the sensitivity of cigarette consumption to price varies with income, using a two-part model and pooled cross-sections from the CPS, merging the tobacco use supplements with the February/March CPS from 1993-2002. Then, we predict the regressivity of large cigarette tax increases using the traditional tax expenditure-based definition of progressivity and traditional welfare measures. We focus on the progressivity of changes in these measures. We find that the price elasticity of smoking participation is -.14 for the lowest income tercile, -.05 for the middle income, and -.21 for the high income. We find that the price sensitivity of conditional consumption, cigarettes smoked by smokers, shows no robust pattern with income and is frequently insignificant. Thus, our results challenge the conventional view that price sensitivity falls monotonically with income. Our predictions of the equity consequences of tax increases show that using all traditional measures of progressivity, whether based on tax expenditures or welfare, cigarette tax increases are not close to progressive"--National Bureau of Economic Research web site.