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
9
the theoretical model. Section 3 presents the data, summary statistics and the construction
of the instrumental variable. Section 4 presents the main impacts of fertility on labor supply.
Section 5 presents exclusion restrictions related to the instruments used in this study. Finally,
section 6 presents the conclusions.
2 Identification Strategy and Theoretical Model
2.1 Identification Strategy
As Rosenzweig and Wolpin (2000) indicate, impact of fertility on the participation of married
women in the labor force was investigated first in the United States. Taking into account that
fertility and participation on labor force are endogenous, some identification strategies were
the use of twins on the first birth (Rosenzweig and Wolpin (1980); Bronars and Grogger
(1994)) and the Angrist and Evans’ natural instrument (1998) of the sex of the first two
births, specifically sex-sameness. Some recent strategies involve infertility shocks (Aguero
and Marks (2008)), exploit the multiple births in higher parities (Delpiano (2012)) or research
about a particular sample (Cristia (2008)).
This paper employs the Angrist and Evans’ instrument which “exploits the parental
preferences for a mixed sibling-sex composition that can be thought as randomly assigned.
Then, a dummy variable that indicates whether the sex of the second child matches the sex
of the first one provides a plausible instrument for further childbearing among women with
at least two children” (Angrist and Evans (1998):451).
This instrument must meet two conditions to be valid:
•
It must be relevant. It must be found that parents whom their first two children have
the same sex are more likely to have an additional child compared to those who their
first two children have different sex (with the sex-sameness instrument).
•
I must accomplish the exclusion restriction. It must be confirmed that to have a
children of the same sex in the first two births does not directly affect the subsequent
labor supply of either parent except through its effect on having an additional birth.
This seems plausible since parents cannot control or manipulate their children’s sex.
With this in mind, the following regression models is proposed:
Y
i
=
α
+
w δ
+
βx
i
+
i
(1)
Where:
Y
i
is a measure of labor supply,
w
is a covariates vector: age of women, age at first birth,
indicators for the sex of first and second child, dummies by indigenous and urban status.
3