Table 6

Estimated coefficients of the ordinary least squares (OLS), instrumental variable (IV) and fixed effects (FE) regressions of relative income price (RIP) of cigarettes in Bangladesh, 2009–2015

	OLS			IV			FE
	Coefficient	Robust SE	P values	Coefficient	Robust SE	P values	Coefficient	SE	P values
Tax	0.43	0.03	0.00	0.56	0.09	0.00	0.34	0.02	0.00
Age	−0.03	0.01	0.00	−0.03	0.01	0.00	–	–	–
Female	−1.37	0.91	0.14	−1.15	0.92	0.21	–	–	–
Years of education
1–8	−0.39	0.38	0.30	−0.16	0.41	0.70	−0.81	0.58	0.16
9 or more	−2.26	0.50	0.00	−1.96	0.54	0.00	−1.35	0.76	0.08
Occupation
Self-employed in non-farm agriculture	−1.10	0.84	0.19	−1.55	0.86	0.07	−0.04	0.91	0.97
Self-employed in non-agricultural activity	−0.77	0.38	0.04	−1.55	0.40	0.00	−0.50	0.51	0.33
Farm wage labourer	1.54	0.50	0.00	2.11	0.53	0.00	0.80	0.61	0.19
Non-farm agricultural wage labourer	0.13	0.74	0.86	−0.79	0.70	0.26	0.44	0.94	0.64
Non-agricultural wage labourer	1.43	0.49	0.00	0.74	0.48	0.12	1.31	0.54	0.02
Professional	−1.73	0.98	0.08	−1.33	1.14	0.24	−0.93	1.20	0.44
Managerial, administrative or clerking	−1.26	0.68	0.06	−2.54	0.76	0.00	−0.19	0.76	0.81
Student	2.54	1.37	0.06	2.37	1.56	0.13	2.66	1.12	0.02
Unemployed	1.27	0.68	0.06	0.43	0.73	0.56	0.66	0.70	0.35
Housewife/housekeeper/household manager	1.37	0.98	0.16	0.96	1.06	0.37	0.99	1.10	0.37
Others	0.41	0.43	0.34	−0.08	0.43	0.86	0.00	0.48	0.99
Socioeconomic status
Moderate	−0.59	0.37	0.11	0.03	0.39	0.95	–	–	–
High	−2.67	0.36	0.00	−2.09	0.38	0.00	–	–	–
Rural area of residence	2.39	2.10	0.26	2.76	0.37	0.00	–	–	–
RIP
Cigarette brands
Medium	−0.99	0.53	0.06	−3.25	1.75	0.06	0.17	0.47	0.72
High	−2.31	1.29	0.07	−9.41	4.24	0.03	3.42	0.88	0.00
Premium	−10.05	2.34	0.00	−24.50	7.94	0.00	1.50	1.51	0.32
Trend
Wave 2	−1.29	0.34	0.00	−1.14	0.35	0.00	−1.52	0.27	0.00
Wave 3	−2.48	0.42	0.00	−3.39	0.84	0.00	−2.88	0.35	0.00
Wave 4	−7.11	0.46	0.00	−8.90	1.29	0.00	−7.22	0.39	0.00
Constant	12.12	2.69	0.00	7.97	0.77	0.00	7.69	0.70	0.00
Number of observations	5882			5882			5882
R²	0.54			0.47			0.35

Reference categories include male, persons with 0 year of education, owner/tenant farmers, low socioeconomic status, urban area of residence, low cigarette brands and Wave 1 (2009).
The SEs of OLS estimates are adjusted for autocorrelation of error terms of multiple observations on the same individual. The village effects in OLS regression are suppressed for the brevity of presentation.
The village variable is used as an instrument for tax in the IV estimation. The test statistics for exogeneity of the tax variable using the Durbin-Wu-Hausman test is statistically insignificant (robust regression F(1, 2552)=2.18, p=0.1403) suggesting exogeneity. The adjusted R² from the first-stage regression of tax is 0.9021, suggesting strong predictability of variations in tax using village-level variation.
The age variable drops out of the FE regression due to perfect collinearity with the wave variables. Gender, socioeconomic status and area of residence variables are time invariant and hence drop out of the FE regression as well.