Panel A :analysis of variance | F-state | R-sq |
Cigarettes | 2.32*** | 0.43 |
Chewed tobacco | 7.71*** | 0.72 |
Panel B: price elasticity | Cigarettes | Chewed |
Cigarettes | −1.069*** (0.115) | −0.027 (0.208) |
Chewed tobacco | −0.103* (0.054) | −0.551*** (0.154) |
***p<0.01, *p<0.1. Bootstrapped SEs in parentheses. The within-cluster regressions of unit values and budget share estimate the coefficients of log of total household expenditure for cigarette to be 0.037 (SE=0.003) and 0.077 (SE=0.036) for the two equations, respectively. For chewed tobacco, these are found to be –0.016 (SE=0.022) and –0.058 (SE=0.012). The overall tables for within-cluster regressions for all the models are available upon request from the authors.
Authors’ calculations from PSLM 2015–2016 (HIICS) data.
HIICS, Household Integrated Income and Consumption Survey.