Versión de HTML Básico
Analíti a
k
9
Revista de Análisis Estadístico
Journal of Statistical Analysis
Analítika, Revista de análisis estadístico, (2015), Vol. 9
Impact of Fertility on Female Labor Supply
17
Table 7:
Exclusion restriction: Women aged 21-35 with two children
All women
Expenditure
Controls
sd Treated sd p-value
Per-Child Clothing and shoes expenditure
35.57 (35.663) 36.63 (35.798) [0.408]
Per-Child clothing expenditure
25.68 (27.476) 26.29 (26.986) [0.534]
Per-Child shoes expenditure
10.41 (10.287) 10.84 (10.943) [0.268]
Per-Child spending on clothing children 3 to 12 years
8.26 (9.887)
8.29 (9.432) [0.930]
Per-Child spending on shoes children 3 to 12 years
4.06 (4.006)
4.08 (3.976) [0.908]
N
1648
1506
Note: National Household Survey of Income and Expenditure 2011-2012 INEC. Treated group is samesex
equals 1 and control group is same sex equals 0.
5 Exclusion restriction
One concern in relation to the instrument applied in this paper is raised by Rosenzweig
and Wolpin (2000), who say that the same sex instrument can affect labor supply through
economies of scale, and thereby reducing the cost of childcare. To evaluate this, I used data
from the Ecuadorian survey of income and expenses and I found that expenses that may
involve some form of economies of scale do not have difference between households with two
children of the same sex compared to households with two children of different sexes (Table
7).
6 Conclusions
The OLS estimates indicate that women with more than two children are 8 and 9 percentage
points less likely to work than women with two children for the sample of all women and
married women, respectively.
To estimate the causal effect of fertility on female labor supply I use sex composition of
first two children as instrumental variable. The first stage shows that families with two boys
or two girls are 3.6 and 3.8 percentage points more likely to have a third child than families
with one boy and one girl, for the sample of all and married women respectively.
The 2SLS results show a causal impact 8 and 9 percentage points of decrease of female
labor supply by having a third child in all and married women respectively.
Two aspects are important here: first, the results apply for the “compliers”. This means
that the result is local (Local Average Treatment Effect-LATE), that is, for women who
changed their fertility decision due to the instrument. Second, these results refer for moving
a negative and significant impact (at 10% of confidence level) of about 7 percentage points on the outcome
“paid work” and don’t present impact on the “work” for both samples (Annex 1).
11