Versión de HTML Básico
Assessing the Effect of Conditional Cash Transfers in Children Chronic Stunting: The Human Development Bonus in Ecuador
107
13
Analiti a, Revista de análisis estadístico, Vol. 13 (1), 2017
Table 7:
FRDD estimates (parametric)
VARIABLES
(1)
(2)
(3)
D
-1,951 -1,928
-2,147
(1,503) (1,499)
(1,410)
X
0,00553 0,346
0,228
(0,0425) (1,275)
(1,299)
X2
-0,00601 -0,00411
(0,0225)
(0,0230)
age (months)
-8,26e-05
(0,000166)
sex (1: male)
-0,225*
(0,121)
ethnicity (1: indigenous)
0,113
(0,299)
mother’s height
0,0496***
(0,0113)
mother’s education level
-0,340
(0,260)
Constant
-0,574 -5,385
-9,994
-1,772 (18,13)
(18,15)
Observations
662
662
662
Robust standard errors in parentheses
*** p
<
0,01, ** p
<
0,05, * p
<
0,1
Dep var: HAZ (Y)
D instrumented by Z
Note:
Coefficients of D in (1), (2) and (3) are estimates of the effect of BDH recep-
tion on stunting z-scores. None of these programme effect estimates are statistically
significant.
5.3 Non parametric estimates
For this approach, I took advantage of the tools developed Calonico et al. (2014) and Calonico
et al. (2016a) which allows to compare a) conventional estimates with conventional variance
estimators, b) conventional but bias-corrected and c) bias-corrected robust non-parametric
estimators, for different bandwidths and polynomial fits of the forcing variable for both, the
effect estimates and the confidence intervals.
The logic for the especifications selection was similar to the parametric section, therefore
the same three were used, though the difference (other than the method itself) is that the
25