Abstract
There is minimal evidence causally estimating the relationship between tobacco tax policy, population health and earnings. This article uses state tobacco taxes as an instrument for life expectancy to estimate the effect on earnings per capita in a panel of almost 3000 counties in the United States over the period 1970 to 2000. In the first stage of the model, we find that a one dollar increase in state tobacco taxes significantly increases life expectancy between 1.9% and 2.1%. Despite showing that tobacco tax is correlated with higher life expectancy and that, theoretically, improvements in health from reductions in smoking should lead to increased earnings, we find an insignificant impact of the induced gains in life expectancy from tobacco policy on county level earnings per capita over the period 1970 to 2000. The lack of significance is explained through local average treatment effects and the smaller economic impact of certain policies that impact health later in life. These results provide further evidence that the most effective interventions for improving income later in life are policy interventions with a direct impact on the health of younger cohorts rather than older cohorts.
