Descriptive statistics of variables used in the first-stage regression
| Observations | Mean | SD | Min | Max | |
| Consumers | |||||
| Unit value, cigarettes | 5775 | 0.993 | 0.380 | 0.312 | 2.472 |
| Budget share, cigarettes | 5775 | 0.046 | 0.037 | 0.000 | 0.496 |
| Total expenditure | 5775 | 756.128 | 389.698 | 63.988 | 2900.703 |
| Household size | 5775 | 3.497 | 1.831 | 1.000 | 56.000 |
| Male ratio* | 5775 | 0.505 | 0.245 | 0.000 | 1.000 |
| Adult ratio* | 5775 | 0.881 | 0.189 | 0.250 | 1.000 |
| Mean of years of education | 5775 | 10.676 | 1.766 | 0.000 | 20.000 |
| Maximum education | 5775 | 14.132 | 2.546 | 0.000 | 20.000 |
| Household type* | |||||
| Unemployed | 5775 | 0.035 | 0.185 | 0.000 | 1.000 |
| Pensioners | 5775 | 0.192 | 0.394 | 0.000 | 1.000 |
| Employed | 5775 | 0.772 | 0.419 | 0.000 | 1.000 |
| Whole sample | |||||
| Unit value, cigarettes (ln) | 5775 | 0.993 | 0.380 | 0.312 | 2.472 |
| Budget share, cigarettes | 12 503 | 0.021 | 0.034 | 0.000 | 0.496 |
| Total expenditure (ln) | 12 503 | 689.281 | 387.071 | 33.687 | 2630.04 |
| Household size (ln) | 12 503 | 3.206 | 1.792 | 1.000 | 56.000 |
| Male ratio* | 12 503 | 0.464 | 0.271 | 0.000 | 1.000 |
| Adult ratio* | 12 503 | 0.889 | 0.188 | 0.200 | 1.000 |
| Mean of years of education | 12 503 | 10.062 | 2.884 | 0.000 | 20.000 |
| Maximum education | 12 503 | 11.924 | 3.159 | 0.000 | 20.000 |
| Household type* | |||||
| Unemployed | 12 503 | 0.034 | 0.181 | 0.000 | 1.000 |
| Pensioners | 12 503 | 0.250 | 0.433 | 0.000 | 1.000 |
| Employed | 12 503 | 0.716 | 0.451 | 0.000 | 1.000 |
Author’s calculation based on the Household Budget Survey data.
T statistics difference between variables means: budget share t=−94.733121; total expenditure t=−17.645353; household size t=−16.984248; male ratio t=−16.344199; adult ratio t=4.2696627; mean education t=−5.2837905; max education=−10.555047; employed t=−0.79802037; pensioners t=14.127229 and unemployed t=−13.193474.
*Adult ratio represents percentage of adults in household, including members aged 15 or more. Male ratio represents percentage of males in household. We controlled for the household economic activity by splitting households into three groups: (1) unemployed (each member unemployed), (2) pensioners (no employees, only pensioners), (3) employed (at least one member employed- person in paid employment).