A microeconometric analysis of smoking in the UK health and lifestyle survey

This paper provides an overview of microeconometric evidence on various aspects of smoking using data from the UK Health and Lifestyle Survey (HALS). In doing so it illustrates how the choice of an appropriate statistical model is influenced by the survey design and the definition and measurement of variables. The HALS is a rich source of information on cigarette smoking which provides data on all aspects of the “life-cycle” of a smoker. Results are reported for parametric survival analysis of the age of starting and the number of years of smoking; univariate and bivariate probit models suggest that addiction, proxied by previous peak consumption, does not affect the desire to quit but does lead to a significant reduction in the chance of succeeding. There is clear socio-economic gradient in success in quitting, but not in the desire to quit. Those with other smokers in the household are less likely top have quit. Results for the duration analysis suggest that the determinants of the number of years smoked are broadly similar for those for quitting. The results for the Box-Cox double hurdle model are consistent with a “fixed cost” model of addiction. Addiction reduces the probability of quitting and raises the number of cigarettes smoked. However, the effects of other variables work in opposing directions for participation and consumption. The implication is that, after controlling for addiction, the more someone smokes the more likely they are to try and quit, as the potential benefits of quitting are greater.