Pseudo Panel Data Estimation Technique and Rational Addiction Model: An Analysis of Tobacco, Alcohol and Coffee Demands

In this paper, we generalize the rational addiction model to include three addictive goods: cigarettes, alcohol and coffee. We use a pseudo-panel data approach which has many advantages compared to aggregate and panel data. While cigarette and coffee demands fit well with the rational addiction model, alcohol demand does not. This result might be due to possible inventory effects. Our results suggest that although cigarettes and alcohol reinforce each other in consumption, consumers substitute them when there are permanent changes in relative prices. In the semi-reduced system, the cross-price elasticity of coffee demand with respect to cigarette price is positive and significant. Long-run cross-price elasticities derived from the semi-reduced system and the Morishima elasticities show that when relative prices increase, consumers substitute addictive goods with other addictive goods. This is likely due to compensation and income effects. When there is a permanent increase in relative prices, addicts cut the consumption of a harmful addictive substance, and substitute it with another addictive substance to compensate for the resulting stress. Moreover, when the consumption of an addictive substance decreases after a price increase, relative consumption of other substances increase due to the positive income effect