The Effect of Maternal Healthcare on the Probability of Child Survival in Azerbaijan
Table 1
The effect of delivery in healthcare facility on probability of child survival.
2SLS model
Bivariate probit model
Coef.
Std. Err.
Coef.
Std. Err.
First equation: instrumented variable is delivery in health care facility; instrumental variables are wealth and birth order
Age 20 or younger
−0.104
0.035
0.003
−0.357
0.116
0.002
Age 36 or older
0.094
0.053
0.076
0.344
0.207
0.098
Low birthweight
0.061
0.045
0.171
0.219
0.179
0.222
Higher education
0.077
0.022
0.000
0.660
0.199
0.001
Wealth
0.198
0.032
0.000
0.891
0.144
0.000
Birth order
−0.073
0.013
0.000
−0.247
0.040
0.000
Constant
0.854
0.038
0.000
1.049
0.135
0.000
Main equation: outcome variable is probability of child survival
Delivery in healthcare facility
0.151
0.063
0.016
0.923
0.347
0.008
Age 20 or younger
0.012
0.016
0.451
0.080
0.154
0.601
Age 36 or older
−0.017
0.031
0.584
−0.136
0.247
0.582
Low birthweight
−0.020
0.019
0.288
−0.174
0.167
0.297
Higher education
−0.032
0.022
0.144
−0.164
0.199
0.410
Constant
0.843
0.051
0.000
0.969
0.319
0.002
Number of observations
2285
(5, 311)
1.31
Prob >
0.000
Number of observations
2285
Log pseudo-likelihood
−1450000000
Wald (11)
126.52
Prob >
0.000
Test of endogeneity
Durbin-Wu-Hausman test and (P value)
10.49 (0.001)
-statistic and P value
5.647 (0.017)
(Rho)
−0.424
Wald test and P value
4.15 (0.041)
Tests for overidentifying restrictions
Hansen statistic and P value
0.407 (0.686)
Tests for the adequacy of instruments
Kleibergen-Paap LM statistic and P value
42.93 (0.000)
Effects of treatment
ATE (average effect of treatment)
0.161
ATT (average effect of treatment to the treated)
0.184
Notes: (1) dependent variable in the first stage is healthcarefacility delivery = 1; otherwise = 0. Dependent variable in the second stage is child survival = 1; otherwise = 0. (2) , , and . (3) Results adjusted to heteroskedasticity and clustering. (4) Data are rounded up Source: 2006 Azerbaijan Demographic and Health Survey [17].