Reply to Jonathan Gruber and Michael Frakes

We address Gruber and Frakes's criticisms that (1) the state excise tax on a package of cigarettes is a more exogenous measure of the cost of cigarettes than the state-specific price of a package of cigarettes and (2) that it is preferable to control for the effects of unmeasured variables that vary over time with the use of time dummies instead of with a quadratic time trend. We also point out that our specification of the cost effect differs from theirs because we allow it to be nonlinear while they force it to be linear.