Schooling and household welfare: The case of Sri Lanka from 1990 to 2006

• Published date: 01 May 2018
• Date: 01 May 2018
• Author: Rozana Himaz,Harsha Aturupana
• DOI: http://doi.org/10.1111/rode.12355

REGULAR ARTICLE

Schooling and household welfare: The case of Sri

Lanka from 1990 to 2006

Rozana Himaz

1,2

|

Harsha Aturupana

3

1

Department of Accounting, Finance and

Economics, Oxford Brookes University,

Headington Rd, Oxford, OX3 0BP, UK

2

Centre for the Study of African

Economies (CSAE), Economics

Department Manor Road, University of

Oxford, Oxford, OX1 3UQ, UK

3

World Bank, Washington, D.C

Correspondence

Rozana Himaz, Department of

Accounting, Finance and Economics,

Oxford Brookes University, Headington

Rd, Oxford OX3 0BP, UK; and Centre

for the Study of African Economies

(CSAE), Economics Department Manor

Road, University of Oxford, Oxford OX1

3UQ, UK.

E-mail: rhimaz@brookes.ac.uk

Abstract

This paper looks at the effect schooling has had on

household welfare in Sri Lanka during the 1990–2006

period, on average and across the welfare distribution.

We account for the endogeneity of schooling using quan-

tile instrumental variable estimation as developed in

Chernozhukov, Fern

andez-Val, and Kowalski (2015). We

use pooled data from four cross-sectional Household

Income Expenditure Surveys. The results show that an

extra year of schooling on the part of the most educated

adult member in the household can increase welfare

(proxied by real per capita consumption expenditure) by

3.8 percent on average. However, the effect varies con-

siderably across the welfare distribution: At the lower

end, around the 20th and 25th quantiles, an extra year of

education increases welfare by 6 and 5 percent, respec-

tively, while at the median it is around 3.5 percent. At

the higher, 90th quantile it is much less, at 1 percent.

Thus the marginal effect of schooling on welfare is sig-

nificant and positive at all levels of the welfare distribu-

tion, but highest at the lower and middle quartiles. This

result is different from findings in the literature that tend

to show larger effects at higher quantiles, when endo-

geneity is uncorrected.

DOI: 10.1111/rode.12355

592

|

©2017 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/rode Rev Dev Econ. 2018;22:592–609.

1

|

INTRODUCTION

Sri Lanka is a developing country with deep and widespread schooling that has been well recog-

nized in the literature. In the 1980s, for instance, it was recognized as being an “outlier”among

developing countries with levels of literacy that were much higher than can be expected at the

given low income levels and high gender parity in education access and attainment (Isenman,

1980; Sen, 1981; Osmani, 1994). More than 30 years (and a three-decade-long civil war) later, in

the 2000s, Sri Lanka can boast of adult literacy levels even higher than before as stocks of those

who completed secondary education have doubled and women surpass men in terms of educat ion

attainment (World Bank 2005).

Although many studies have looked at private returns to education at the individual level in Sri

Lanka (Aturupane, 1993; Gunawardane, Abeyrathna, Ellagala, Rajakaruna, & Rajendran, 2008; De

Silva, 2009; Himaz, 2010, Himaz & Aturupane, 2016), few studies look at how much an addi-

tional year of schooling contributes toward household economic welfare levels, proxied by con-

sumption. This is an important aspect to look at, because even though an extra year of education

may lead to higher incremental returns at the individual level –on average around 5 percent,

according to Himaz and Aturupane (2016) –this does not mean that incremental household wel-

fare with respect to an extra year of education on the part of the most educated adult in the house-

hold (or some other proxy for the level of household education) is 5 percent. This is because of a

key problem that arises due to the accuracy of individual wages or earnings based estimations of

returns to education in developing countries: in Sri Lanka, as in many developing countries, non-

wage earners are a considerable share of the labor force, engaged in self-employment activities or

working for a family farm or venture, earning an income for the household rather than the individ-

ual.

1

Since it is difficult to map an individual’s education attainment to their contribution to house-

hold income, surveys often assign household earnings as those attributable to the household head.

Thus even if earnings are used instead of wages, the estimation of individual returns to education

becomes problematic. In any case, individual returns estimation says little about household welfare

and education or the effect education may have on welfare in rich versus poor households.

Several papers in the relevant literature, such as Glewwe (1991) and Teal (2004), look directly

at the impact of education on household welfare given prior asset accumulation and household

characteristics, enabling an interpretation that suggests not simply a correlation but causality.

Glewwe finds, for instance, that education has a positive effect on welfare for those employed in

the public sector in C^

ote d’Ivoire. Teal finds that an extra year of education increased household

welfare by 1.9–2.9 percent in Ghana in the 1990s. Following an approach similar to Glewwe

(1991) and Teal (2004), Rolleston (2011) finds not only that schooling plays an imp ortant role in

household welfare in Ghana for the 1991–2006 period but also that higher levels of schooling have

larger and increasing benefits compared to lower levels of schooling. This type of result –that

education significantly impacts household welfare (measured using real per capita consumption or

income) –is evident in several other more recent studies as well, such as Haddad and Maluccio

(2003), Bellemare (2012), Alem and S€

oderbom (2012), and Arouri, Nguyen, and Youssef (2015).

A few papers also look at the impact of education on household consumption between poorer and

richer households. For example, Nguyen, Albrecht, Vroma, and Westbrook (2007, p. 477) use

quantile regression (QR) techniques to find, inter alia, that upper quantiles see larger increases in

the marginal impact of education on household welfare in the South of Vietnam, although compa-

rable patterns are not evident in the North. In the North the impact is stable across the quantiles,

apart from the very top. In a similar study for 12 Arab countries, Hassine (2015) finds that the

marginal effect of education on household spending is generally larger at higher quantiles. Himaz

HIMAZ AND ATURUPANA

|

593