Women’s Empowerment and Child Malnutrition: The Case of Mozambique

• Date: 01 June 2019
• DOI: http://doi.org/10.1111/saje.12223
• Author: Joseph Deutsch,Jacques Silber
• Published date: 01 June 2019

South African Journal of Economics Vol. 87:2 June 2019

doi : 10.1111/ saje .122 23

139

© 2019 UNU-WIDER. South A frican Journa l of Economics published by John W iley & Sons Ltd on behal f of

Economic Societ y of South Africa.

WOMEN’S EMPOWERMENT AND CHILD MALNUTRITION:

THE CASE OF MOZA MBIQUE

JOSEPH DEUTSCH† A ND JACQUES SILBER† ‡§,*

Abstract

Using data from the 2011 Demographic and Health Survey in Mozambique this paper checks

whether women empowerment has an impact on the nutritional status of children. We evaluate

the degree of empowerment of women via multidimensional approaches, making a distinction

between five domains: decision making, use of violence by husband/partner, attitude of the woman

towards this use of violence, available information, material resources. Each domain includes

several questions reflecting different aspects of empowerment. For each domain of empowerment,

three different methods of aggregation are used: correspondence analysis, the so-called Alkire

and Foster methodology and the “fuzzy sets” approach. The impact of women empowerment on

the nutritional status of children is analyzed via the MIMIC approach. No clear-cut conclusion

concerning the possible impact of women’s empowerment on the nutritional status of children

could be drawn. But, ceteris paribus, the material wealth of the household, the educational level

of the mother and her BMI are positively correlated with the nutritional status of children which

is also higher when the child is female. Finally, there are important differences in the nutritional

status of children between the various regions of Mozambique and this nutritional status is in most

regions lower in rural areas.

JEL Classification: D13, I15, J16, O15, O55

Keywords: Alkire an d Foster approach, correspondence analysis, de mographic and health Survey,

“ fuzzy sets” approach, health, MIMIC approach, Mozambi que, women empowerment

1. INTRODUCTION

One of the most robust features of economic development is that in countries with higher

income, lifespans are longer, infant mortalit y lower and illness reduced throughout the life

* Corresponding author. Depart ment of Economics, Bar-Ilan Universit y, Ramat Gan, Israel.

E-mail: jsilber_2000@yahoo.com

† Department of Economics, Ba r-Ilan University

‡ LISER

§ Centro Camilo Da gum, Tuscan Interuniversity C entre, Advanced Statistics for Equitable and

Sustain able Development

A first version of this paper was prepared for a workshop organized by WIDER in Maputo,

November 27–28 2017. A second version appeared as WIDER Working Paper 2017/211. The

authors thank t wo anonymous referees for the very useful comments they made on the second

version of this paper.

This is an ope n access arti cle under the terms of t he Creative Commons A ttribution-NonComm ercial-Share Alike Licen se, which

permits use a nd distribution in a ny medium, provided t he original work is pr operly cited, the use i s

non-commercia l and the content is of fered under identica l terms.

South African Journal

of Economics

140 South African Journal of Economics Vol. 87:2 June 2019

© 2019 UNU-WIDER. South A frican Journa l of Economics published by John W iley & Sons Ltd on behal f of

Economic Societ y of South Africa.

course (Deaton, 2013). Similar conclusions seem to hold within countries. As argued by

Wilkinson (2000), death rates are two to three times higher among people at the bottom

than at the top of the social hierarchy. Wilkinson adds that such differences may be larger

in some countries than in others but they are found wherever data a re available, and it does

not matter whether people are ranked by education, income, occupation, or living area.

One of the main determinants of illness turns out to be malnutrition. Malnutrition,

severe or otherwise, is estimated to be a contributing factor in over 50% of child deaths

(Prudhon et al., 2006). Müller and Krawinkel (2005) also argue that ma lnutrition “is

globally the most important risk factor for illness and death, with hundreds of millions

of pregnant women and young children particularly affected.” Malnutrition can refer

to a deficiency in macronutrients (protein, carbohydrates and fat) causing thus protein–

energy malnutrition, but there exists also malnutrition in micronutrients (electrolytes,

minerals and vitamins).

Among children, protein–energy ma lnutrition occurs when a child suffers of under-

weight (weight falling below two standard deviations under the normal weight for age),

stunting (height falling below two standard deviations under the normal height for age)

and/or wasting (weight for height falling below two standard deviations under the nor-

mal weight for height). While wasting is a sign of recent weight loss, stunting is the

consequence of chronic weight loss. These anthropometric measures a re hence indicators

of malnutrition.

Smith and Haddad (2000), in a cross-country analysis covering the 1970–1995 pe-

riod, stressed four underlying determinants of child malnutrition: health environment,

women’s education, women’s relative status (measured via the ratio of female to male life

expectancy at birth) and per capita food availability.

While education may be considered as an element of women empowerment, the latter

includes other aspects. Among these other features of women empowerment are the ex-

tent of a woman’s decision making within t he household, the absence of violence experi-

enced by the woman as well as her attitude towards this kind of violence, the amount of

material resources that the household enjoys, and the extent of the information available

to the woman.

The purpose of this paper is precisely to look at the link that may exist between these

other components of women empowerment and the health status of children, measured

via two anthropometric indicators, the height for age percentile, a variable referring to

stunting and the weight for age percentile, a variable referring to wasting. Since the

nutritional status of a child is not really observed and is in fact a latent variable, we use

a MIMIC model (see Section 3 and Appendix B, for more details on this technique)

where the two anthropometric indicators that have just been mentioned are considered

as observed measures of t he latent variable representing the nutritional status of the child.

The paper contributes in different ways to the literature that examined the link be-

tween women empowerment and children’s nutritional status. First, as mentioned pre-

viously, it takes into account several domains of women empowerment and looks at

their specific relationship with the two anthropometric indicators of child malnutrition

that we use. Second, for each of the five domains of women empowerment taken into

account, we construct an aggregate variable derived from the answers to the specific

questions asked for each domain, and implement several techniques of aggregation of

variables (see Section 3 and Appendix B for more details on these techniques). Finally,

141South African Journal of Economics Vol. 87:2 June 2019

© 2019 UNU-WIDER. South A frican Journa l of Economics published by John W iley & Sons Ltd on behal f of

Economic Societ y of South Africa.

we examine the relationship bet ween women empowerment and child malnutrition indi-

cators in Mozambique, a country where this issue ha s not yet been examined. Moreover,

Mozambique is a particularly interesting case study because child mortality fell from

about 250 deaths per 1000 live births in the late 1980s to 71.3 deaths per 1000 live births

in 2016, a drastic reduction in child mortality indeed. This decrease in mortality was

certainly related to the end of the civil war that was fought from 1977 to 1992.

The paper is organised as follows. In Section 2, we review the literature examining

the relationship between women empowerment and children’s health-related outcomes.

We first mention studies where changes in women empowerment related to individual

choices had been neutralised thanks to some natural experiment. But we also review

studies where such a “neutralization” was not possible so that the link between women

empowerment and children’s health status cannot be given a causal interpretation. In

Section 3, we describe the data sources and ex plain succinctly, more details being given in

Appendix B, the various techniques we use d to aggregate the variables that were available

for each domain of women empowerment. In Section 4, we describe the main aspects

of the MIMIC model while Section 5 presents the results of our empirical investigation,

based on the MIMIC model. Section 6 briefly concludes.

The main findings of our investigation are as follows. When looking at the whole

sample, no clear-cut conclusion could be drawn concerning the possible impact of wom-

en’s empowerment on the nutritional status of children. However, when limiting the

analysis to male-headed households, the ex tent of a woman’s decision making as well as a

negative attitude towards her husband’s or partner’s violence, had often a positive impact

on the (latent) nutritional staus of her children. The material wealth of the household,

the educational level of the mother and her BMI (body mass index) were, ceteris pa ribus,

positively correlated with the nutritional status of children. The latter was also generally

higher when the child was female. Finally, there were important differences in the nutri-

tional status of children bet ween the various regions of Mozambique and this nutritional

status was in most regions lower in rural areas.

2. ON WOMEN EMPOWER MENT AND CHILDREN ’S HEALTH-RELATED

OUTCOME S

The relationship between women empowerment and family outcomes, such as child

health, was discussed, quite at length, by Duflo (2012). Concerning the fact that the

health of children is more correlated with the education of mothers than with that of

fathers, Duflo (2012) stresses that such a difference raises two important issues. On the

one hand a woman’s education is likely to be correlated with “unobserved dimensions

of her ability, family or community background” so that if these factors have a direct

impact on the health of children, the correlation between women’s education and chil-

dren’s health is not a causal relationship. Moreover, as emphasised by Duflo (2012), when

comparing the coeff icients of husbands’ and wives’ education, one should not forget that

there is probably a correlation between the education of wives and unobserved cha racter-

istics of their husbands. First more educated women “may be able to marry men who care

more about their children”(Duflo, 2012). Second if a husband is “progressive enough to

allow his wife to seek employment, then this same progressive attitude may make him

treat his children better” (Duflo, 2012).