Versión de HTML Básico
Assessing the Effect of Conditional Cash Transfers in Children Chronic Stunting: The Human Development Bonus in Ecuador
95
13
Analiti a, Revista de análisis estadístico, Vol. 13 (1), 2017
number of households), which represents 4’734.481 individuals (29,68% from total popula-
tion). Within the BDH recipients’ subset 672.714 households benefit from the mother’s type
(15,5% from total number of households), which represents 3’368.833 individuals (21,12%
from total population) (INEC, 2014). When comparing the BDH beneficiaries’ coverage
from the LSMS to the official registries published by the MCDS, the percentages that the
families represent from the total population
3
are either equal or very similar. Specifically,
for all type of beneficiaries’ the registries estimate shows a 6,99% coverage while the LSMS
a 7,04%, and for the mother’s type a 2,8% compared to a 4,2%. These percentages are also
consistent between datasets when analysing by sex and ethnicity.
For the index replica, the complete set of surveyed individuals is employed to reflect
the living conditions of the Ecuadorian population. Though, the analysis faces some other
criteria which reduces sample’s power. First, I kept only those who declared to be mother
type recipient’s households and non-beneficiaries. Afterwards, I selected only those surveyed
from April 2014, since the RSII only started to be applied from that date. Lastly, the
largest reduction was made by keeping only households with children under 5 years. The
final sample is composed by 6.317 kids and 1.356 households, though, if we only consider the
valid cases the sample is reduced by 143 individuals. Table 2 shows t-test for the stunting
prevalence and the HAZ by BDH declared reception. The statistics evidence a significant
difference in both measures when analysing by group.
Even though, the LSMS allows to group individuals into nuclear families, the construction
of this unit was not clean (e.g. not all declared a family head), also, the LSMS focuses
its design on individuals grouped into households. Because of this, I decided to do the
analysis on the household level. The following subsection details the steps undertaken for
the methodological design of the research including descriptive statistics and intermediate
results.
4.2 Methodological approach and implementation
4.2.1 RSII index replica
The eligibility score is a proxy means testing index which is expected to be similar to the
consumption poverty, but with a multidimensional approach based on Bourguignon and
Chakravarty (2003). This type of index is the most commonly used in Latin American
programmes. In Ecuador, as mentioned, the index changed depending on the living condi-
tions of the population and was estimated through NLPCA. Compared to the traditional
PCA, this method does not rely on the assumptions of linear relationships between numer-
ical variables nor the multivariate normality of the data. Additionally, Optimal Scaling by
3
The ratios where obtained by dividing the number of families receiving BDH to the total population,
since MCDS makes public only the number of household but not the individuals they represent. For the
registries ratio, I used the total population estimated for 2014 in the bases the 2010 population census; while
for the LSMS I used the total weighted sample population.
13