We used a two-part model for the estimation of the price elasticity of participation and consumption of cigarettes by the duration of the smoking habit and a continuous-time split-population model for the estimation of prevalence and duration of smoking onset and smoking addiction, allowing for covariates in the participation component of the model. Results: We computed the total price elasticity of consumption of cigarettes by quartiles of addiction and found that for the people located in the lowest quartile of addiction the total price elasticity is around -0.51; while for those located in the highest quartile of addiction this figure is only -0.19. Then, a 10% increase in cigarette prices, via taxes, reduces the consumption of those in the early stages of the addiction by 5% and for those with a longer history of addiction by only 1.9%. Estimating the continuous-time split-population model we found that, at the mean starting age of 15 years, an increase of 10% in real cigarette prices is expected to delay smoking onset by almost two and a half years. On the other hand, the same policy is less effective to reduce the duration of the habit because there is no meaningful relationship between the duration of the smoking habit and the real price of cigarettes.The policy of raising cigarette excise taxes, to increment prices, seems to be more effective to delay smoking onset. On the other hand, the same policy is less effective to reduce the duration of the habit. A policy recommendation that emerges from this evidence is that for people with a developed addiction a combination of increasing taxes and other public health policies, like cessation therapies, could prove more effective.
Keywords: cigarette prices and taxes; quitting smoking; smoking onset; split-population model; survival analysis.