Table 4.
Results of the budget share regression (equation 3)
Variables | (2005) | (2009) |
w | W | |
lnx | −0.056*** | −0.065*** |
(0.017) | (0.023) | |
Size | 0.002 | 0.039 |
(0.031) | (0.043) | |
Adults | 0.008 | 0.092 |
(0.072) | (0.103) | |
Males | 0.013 | 0.010 |
(0.059) | (0.068) | |
Education | −0.001 | −0.012 |
(0.020) | (0.025) | |
Age | 0.028 | −0.077 |
(0.044) | (0.072) | |
Gender | −0.038 | −0.108* |
(0.037) | (0.056) | |
Work | 0.037 | 0.058 |
(0.037) | (0.039) | |
Year | ||
Constant | 0.533*** | 0.963*** |
(0.193) | (0.292) | |
No. of households | 233 | 147 |
R2 | 0.866 | 0.909 |
Results of the regression of the cigarette budget share (w) on the log of household expenditure (lnx) and other household characteristics (see equation 3). Household size (Size), education of household head (Education) and age of household head (Age) are in natural logarithms. Adults refers to the proportion of adults in a household and adults are defined as aged 18 years or older. Males is the proportion of males in a household. Gender is a dummy variable which takes on the value of 1 if the household head is male and zero if they are female. Work is a dummy variable which takes on the value of 1 if the household head is employed and zero otherwise. SEs are in parentheses. *P<0.1, **P<0.05, ***P<0.01. Cluster fixed effects are suppressed for space reasons but are jointly statistically significant at the 1% level for the 2005 sample and at 10% for the 2009 sample.