Decomposition of cohort, age and time effects in Iranian households’ cigarette consumption

The purpose of this paper is to examine cigarette consumption behavior of younger cohorts in the urban and rural areas of Iran. The authors use Iran’s annual Household Income and Expenditures Surveys (HIES) database over 2007–2013 for the statistical analysis. In order to control for a large number of households with zero expenditure on cigarette consumption, the authors have used the double-hurdle modeling approach for counting the outcomes of interest. The authors have also limited the sample to cases in which the head of household is between the ages of 21 and 45 and all children are younger than 18.

In this study, the authors have conducted a multivariate econometric analysis to identify the impact of age and birth year cohort on the demand for cigarettes among Iranian households. The authors have used the HIES data for multiple years in the analysis. The ideal data set for the analysis is a panel data that include information on cigarette consumption of various age cohorts over a long period of time. Since no suitable panel data are available, the authors have constructed a multi-year cohort data by extracting cohort data from the annual HIES data set. Due to the unique properties of cigarette consumption, the authors have used the double-hurdle econometric model with appropriate diagnostics.

After controlling for price and demographic factors, which affect the demand for cigarettes, the authors find that the younger cohorts in rural areas, who smoke, tend to consume fewer cigarettes than the older ones; however, the opposite is true among urban households. The probability of being a non-smoker is larger for younger cohorts in both rural and urban areas. Among smokers, the authors observe an inverse U-shape relation between age and quantity of cigarettes consumed per day. The trend is positive up to age 45, but diminishes for older smokers because of health concerns.

In comparison to previous studies of tobacco consumption in Iran, the authors have used a more comprehensive household income and expenditure survey data set with a large number of observations. Furthermore, the authors have applied an econometric method (the double-hurdle model), which is suitable for the analysis of the determinants of demand for cigarettes when a subset of households report no cigarette consumption.