Determinants of Smoking Cessation: A Panel Data Analysis Using HRS

Abstract

Using longitudinal data of HRS from, this paper analyzes older smokers’ participation decision and cigarettes consumption decisions using two-part and hurdle models with correlated random effects. This study extends the existing literature by incorporating both cigarette prices and health shocks in one unified framework. Our estimates indicate that raising cigarette price plays an effective role in both reducing participation and amount of consumption of cigarettes. Simultaneously, health shocks like diagnosed heart diseases and cancers are found to raise the probability to quit smoking substantially, which verifies the curative aspect of quitting behavior in the elderly. However, if the smoker does not quit even after a health shock, the shock shows a weak effect in reducing smoking intensity. By a decomposition analysis, we show that whereas both health shocks and cigarette prices explain the reduction in smoking prevalence, increases in cigarette prices are main reasons for the drop in smoking intensity over the period of 1992-2010.

Keywords: Smoking cessation, Health shocks, Hurdle model, Correlated random effects, Selfselection

JEL Classification: I12, I18, C23, C25