### INTRODUCTION

### MATERIALS AND METHODS

### Cigarette demand by different income groups

### Models

##### (1)

*Z*is a dichotomous variable which detects whether or not y is observed,

_{i}*y*being observed only when

_{i}*Z*= 1,

_{i}*W*denotes observable features including the overlapping variables

_{i}*X*, and

_{i}*α*denotes the vectors of parameters to be estimated. In the linear model,

*y*denotes the dependent variable,

_{i}*X*denotes the observable features of the independent variables,

_{i}*β*is the parameters to be estimated and

*μ*is a normally distributed error term with a mean of zero and a standard deviation σ to be estimated [16].

_{i}### Model specification

*η*is the overall elasticity,

_{t}*η*is the participation elasticity, and

_{p}*η*is the conditional elasticity.

_{c}*Abridged SimSmoke*model represents that a larger proportion of decreased smoking prevalence would be attributable to tobacco taxation rather than other tobacco policies included in MPOWER package. It has been modeled that Increasing cigarette excise taxes to 75.0% of the retail price alone would reduce smoking prevalence by 27.3% within 5 years [20].

##### (4)

*Smoke*Dummy variable that takes a value of 1 if the household spent any money on cigarettes;

*Income*Total monthly household income of employed family members from self-employment and salary by Iranian rials (in natural logarithm);

*Divorce rate*A province-level variable representing total divorce rate of province where household residing in;

*Unemployment rate*A province-level variable representing total unemployment rate of province where household residing in;

*share of members aged 15 to 18*Proportion of household members at age between 15 and 18 to household size;

*share of jobless member*Proportion of unemployed household members aged between15 and 64 to household size;

*share of primary educated*A variable showing proportion of household with at least one member who had a primary education;

*share of university educated*share of household with at least one member who hold a university degree;

*Age*Age of the household head; Sex Dummy variable that takes a value of 1 if the head of household is male

##### (5)

*β*1 and price elasticity from

*β*2.