In this article we attempt to evaluate the impact of various information shocks on cigarette consumption. In contrast to the existing studies, we do not impose any break points a priori.We use recently developed techniques for sample splitting in the data on US cigarette demand, and find that using a time index as a transition variable can cluster cigarette demand into four distinct regimes. In the past four decades a myopic-to-rational switch in cigarette consumption behaviour was observed in response to various anti-smoking information events. In particular, we find strong evidence in support of the rational addiction model in the 1980s and 1990s. Moreover, in line with the theoretical prediction, our estimate of the long-run elasticity is twice as large as the short-run elasticity. Our empirical framework should improve policy makers' understanding of the dynamics of cigarette consumption in response to various anti-smoking campaigns, and stress the role of sound policy making in improving smoking control measures.
All Science Journal Classification (ASJC) codes
- Economics and Econometrics