Age at marriage, social norms, and female education in Nepal
| Author | Gabriel Montes‐Rojas,Saqib Jafarey,Ram Mainali |
| Date | 01 August 2020 |
| DOI | http://doi.org/10.1111/rode.12692 |
| Published date | 01 August 2020 |
878
|
wileyonlinelibrary.com/journal/rode Rev Dev Econ. 2020;24:878–909.
© 2020 John Wiley & Sons Ltd
1
|
INTRODUCTION
The millennium development goals sponsored by the United Nations (in particular, Sustainable
Development Goal Five, SDG5, United Nations General Assembly, 2015) call on member countries
to promote gender equality, an important dimension of which is related to education. While most
countries have succeeded in narrowing the educational gender gap at primary and lower secondary
levels, it continues to affect higher levels, particularly in South Asia. In this paper, we study the impact
of age at marriage on female education, instrumenting the former via variations in cultural influences
on female marriage age across ethnic groups in Nepal.
Demographers and ethnographers recognize that the age at which a typical man or woman will
marry is influenced by prevailing social norms.1 Neugarten, Moore, and Lowe (1965) and Billari,
Received: 14 January 2016
|
Revised: 22 May 2020
|
Accepted: 25 May 2020
DOI: 10.1111/rode.12692
REGULAR ARTICLE
Age at marriage, social norms, and female education
in Nepal
SaqibJafarey1
|
RamMainali2
|
GabrielMontes-Rojas3
1Department of Economics, City University
of London, London, UK
2Economic Policy Analysis Division,
Ministry of Finance, Kathmandu, Nepal
3Instituto Interdisciplinario de Economía
Política de Buenos Aires (IIEP-BAIRES-
UBA), Facultad de Ciencias Económicas,
Universidad de Buenos Aires, Buenos
Aires, Argentina
Correspondence
Gabriel Montes-Rojas, Instituto
Interdisciplinario de Economía Política
de Buenos Aires (IIEP-BAIRES-UBA),
Facultad de Ciencias Económicas,
Universidad de Buenos Aires, Av. Córdoba
2122 2do piso, C1120AAQ, Ciudad
Autónoma de Buenos Aires, Argentina.
Email: gabriel.montes@fce.uba.ar
Abstract
We study the impact of age at marriage on female educa-
tion. We hypothesize that in cultures where women marry
young, parents discount the pecuniary benefits of educating
girls; the earlier the anticipated age at marriage the greater
this discount. We empirically test this effect using house-
hold data from Nepal. We control for potential endogene-
ity of age at marriage by exploiting variations in cultural
norms regarding dowry and differences in the average age
of female marriage among ethnicities and regions as in-
strumental variables. The econometric results support the
hypothesis that female education is negatively affected by
cultural norms that favor early marriage.
KEYWORDS
education, gender, marriage, Nepal
JEL CLASSIFICATION
I20; J12; J16
|
879
JAFAREY Et Al.
Prskawetz, and Furnkranz (2003) argue that social norms directly determine the acceptable age range
for individuals to enter marriage. The main hypothesis of this paper is that the earlier the culturally
determined age at which a woman is expected to marry, the less education she will receive, even in the
years prior to her marriage. Our paper empirically tests this effect using data from a Nepalese survey
that identified married women by the age at which they married, their educational attainment, and
their ethnicity, along with other characteristics of their marital and parental households.
Nepal is a relevant country for studying this issue. It has a high incidence of child marriage (de-
fined by the United Nations Children's Fund [UNICEF] as marriage before reaching age 18). As
Table1 shows, 40% of women in the age group 20–24years were married before age 18, a rate that is
second only to Bangladesh in South Asia. At the same time, Nepal’s gender gap in education has been
practically eliminated at primary level, greatly narrowed at lower secondary level but remains fairly
high at the upper secondary level, exceeding that of Bangladesh, India, and Pakistan. This raises the
question as to whether the phenomenon of early marriage and the customs driving it are undermining
female education beyond the basic levels.2
Nepal has considerable variation in age at first marriage across ethnic groups and regions, a feature
that we exploit in implementing an instrumental variable (IV) strategy. Our first instrument indicates
membership to the Maithili-speaking community. Maithili is one of Nepal’s three main languages;
its speakers traditionally belong to a region called Mithila that straddles the Indo-Nepalese border,
stretching from northern Bihar in India to Janakpur in southeastern Nepal. Maithili speakers inhabit
both sides of the border.
Although Nepali Maithili speakers are not counted as a distinct ethnic group within Nepal, mainly
because they are themselves divided into castes and religions, they share cultural traits with their
neighbors in Bihar. One of these is the tradition of dowry. According to widespread popular percep-
tion, dowry-giving is more widely practiced by the Maithili community than by other communities of
Nepal. Although less so in the case of Nepal, academic studies on India have identified Bihar, along
with neighboring Uttar Pradesh, as two states in which dowry is implemented more strictly than in
other parts of India (Ashraf, 1997; Dyson & Moore, 1983; Srinivasan & Lee, 2004). The Maithili cus-
tom of Tilak Pratha, which links the value of the dowry to the grooms’ economic status, encourages
parents to marry their daughters young.
TABLE 1 Comparisons of female education and marriage age across South Asia
Country
School completion ratesa Female marriage
Primary Lower secondary Upper secondary Average age at first
marriagea
Percentage
married on or
before
Male Female Male Female Male Female 15a 18a
Afghanistan 67.2 40.2 49.3 25.6 32.3 14.4 21.2 8.8 34.8
Bangladesh 76.3 89.1 59.2 70.5 31.5 27.3 19.2 22.4 58.6
India 91.5 91.3 82.4 79.3 46.5 39.7 20.7 6.8 27.3
Nepal 83.9 82.3 71.4 67.9 51.9 38.5 20.7 7 39.5
Pakistan 64 55.4 54.7 44.6 23.6 23.3 23.1 2.8 21
ahttps://data.unicef.org (Demographic and Health Survey). Year of data: Afghanistan 2015; India, Nepal 2016; Bangladesh 2019
(education), 2014 (marriage age); Pakistan 2018 (education), 2013 (marriage age).
bhttps://datab ank.world bank.org (Gender Statistics Database). Year of data: Afghanistan, Bangladesh, Pakistan 2013; India 2011;
Nepal 2014.
880
|
JAFAREY Et Al.
The impact of dowry traditions on the age of marriage has been a matter of some discussion in
the empirical literature, albeit in contexts outside Nepal. Ashraf, Bau, Nunn, and Voena (2020) study
the impact of school construction policies in Indonesia and Zambia and how the policy interacts with
marriage customs. They find that girls’ education increased only as a result of this policy in places
where it is common for the brides’ family to receive a payment when their daughter gets married (i.e.,
families invest in their daughters’ education only if they can monetize it). Corno, Hildebrandt, and
Voena (2020) look at the impact of weather shocks on child marriage in Africa and India and find that
the sign of the impacts depends on whether the prevalent marriage payments are dowry or bride price.
In Section 3.2, we justify the use of this instrument in greater detail, and in Section 2.2, we discuss
other identification strategies implemented to study similar effects. However, we acknowledge the
possibility of two sources of bias arising from this variable: first, there may be an inherent cultural bias
in the Maithili community that disfavors girls’ schooling on its own account; second, the practice of
dowry might influence girls’ education because of income effects that might be especially significant
for poor families.
We attempt to control for these potential biases by (1) including mother’s education as a control
for household attitudes to female education and (2) rerunning our IV regressions using only the upper
two quartiles of households in terms of wealth distribution. Our results remain robust to both checks.
Nonetheless, we go on to employ a second instrument, namely, the average age at marriage within a
respondent’s peer group, which we define as the ethno regional group into which she was born and
conduct the aforementioned robustness checks on this instrument as well. Our results continue to hold.
While the IV estimates have large confidence intervals, so that we cannot say that they are signifi-
cantly different from the ordinary least squares (OLS) estimates, they are positive and significant in
the majority of cases. We can infer that the effect is between 0.2 and 0.5years of schooling per 1-year
increase in age-at-marriage.
The rest of the paper is organized as follows. Section 2 compares female education and age at
marriage across the countries of South Asia and reviews the literature on gender gaps in education,
female age at marriage, and the interaction of the two. Section 3 explains the econometric model and
discusses the identification assumptions. Section 4 describes the data and presents summary statistics.
Section 5 reports the empirical results. Section 6 concludes.
2
|
FEMALE EDUCATION AND EARLY MARRIAGE IN
SOUTH ASIA: STATISTICS AND LITERATURE
In this section, we shall first provide some background details, based on recent statistics and informa-
tion taken from secondary sources, concerning marriage age and school completion rates in Nepal and
other South Asian countries. We shall then review the broader literature on girls’ schooling, age at
marriage, and their mutual relationship.
2.1
|
The South Asian background
We acknowledge that since the survey from which our data are taken, Nepal has made steady progress
in both female educational attainment and the incidence of child marriage. Yet in Nepal and through-
out South Asia, policymakers and academics remain concerned about both issues and believe that
further progress needs to be made.
To continue reading
Request your trial