Education is well recognized as one of the more potent ways that hastens social
mobility. Its importance in overall development of a country is also well recognized. This
is clear from the sustained and widespread attention it has received in the development
literature. One does need to dig deep to realize that what underlies education progress
or retrogress is the decision of households to invest in the education of their children. It
is, therefore, always important to contribute to the understanding of this process. This is
the ultimate object of this paper.
Relative to countries with about the same level of development, the Philippines is known
for high school attendance at all levels. Even with its relatively low per capita income, it
has achieved attendance rates that approximate those found in high-income countries
that led analysts to consider the performance of the Philippines in this area an outlier
(see for instance Berhman, 1990, Behrman and Schneider, 1994). This advantage,
however, is fast eroding in recent years. For instance, UNESCO data show that Thailand
has surpassed the Philippines in attendance rates at the secondary and tertiary levels
since the late 1990s2
. But what is even more alarming, as this paper will later show, is
that this erosion is faster among larger and also poorer families. The segment of society
that needs most higher education investment to hasten poverty alleviation is in fact
investing lower than those who need it less.
The paper formulates and estimates a model of the determinants of the proportion of
school-age children attending school considering the endogeneity of the number of
children and using an instrument for it. As far as the author knows, this is the first paper
that has taken into account the endogeneity of the number of children in the school
attendance equation using Philippine data. The quantity-quality literature spawned by
the seminal treatment in Becker and Lewis (1973) clearly argues for the endogeneity of
the number of children in education equations. Under this framework, OLS estimates of
the education equation will be biased and inconsistent. Instrumental variables estimation
is needed to generate consistent estimates.
The paper is divided as follows. The next section presents a brief review of the previous
literatures. Following that is the presentation of the methodology, instrument and data
used. The results are provided next. The final section summarizes and identifies some
implications for policy.