Infant and Child Mortality Differentials By Women’s Empowerment: New Insights from the Egyptian Demographic and Health Survey

Guglielmo Maggi, Student
Niccolò Certi, Student
Curzio Checcucci, Student

Short Abstract

This paper focuses on the relationship between women’s empowerment and infant – first year of life – and child – between one and four years of life – mortality. This topic is not new but still represents a crucial and timely issue, especially for the developing countries. Here we explore such relationship by applying event-history techniques to data from the Egyptian Demographic and Health Survey (2014). We proxy women’s empowerment level by relying on three components: i) women’s level of education; ii) their decision-making autonomy; ii) and their attitude towards domestic violence. Our analysis reveals at least three key findings. First, women’s empowerment clearly mitigates both the risk infant and child mortality, and the effect is more pronounced for the child mortality. Second, higher education urban areas facilitates the mitigation effects of women’s decision making autonomy and attitudes towards violence on mortality risks. Third, women’s empowerment variables are not losing importance among more recent mothers’ cohorts. In all, the positive relationship between women’s empowerment and child health care represents an indication for politics towards the need of strengthening gender and social equality.

Extended Abstract



This paper focuses on therelationship between women’s empowerment and infant and child mortality. Thistopic is not new but still represents a crucial and timely issue, especially forthe developing countries. Here we explore such relationship using fresh datafor Egypt, collected in 2014 by the Demographic and Health Survey Program.

            Instudies of this kind, the first and the main complication is the definitionitself of empowerment and then the construction of a specific index that canmeasure it, this constitute the background for ouranalysis and the selection of the empowerment’s variables.Following the recommendations from the literature (Csazar’s 2005 definition of empowermentwith the Max Weber’s definition of power),in this article we proxy women’s empowerment level byusing three components: i) women’s level of education; ii) theirdecision-making autonomy; iii) their attitude towards domestic violence.

According to theDemographic and Health Surveys of 1995, 2005 and 2014, mortality is higher forinfants than for children (62% vs 19,2% in 1995, 33,2% vs 8,1% in 2005 and 22%vs 5% in 2014); mortality shows a negative trend through the years (itdecreased of the 33% in the last 15 years and it is 9 times lower in 2014respect to 1960); in rural areas mortality is higher (almost 30% higher thanurban areas); education, age of the mother at the child birth, sex and weightof the child are important differential factors for infant and child mortality.

Specifically,we address the following research questions: Does women’s empowerment affectchildren mortality? If yes, is the effect direct or indirect (mediated bywomen’s area of residence and higher education)? Are there differences orsimilarities between the determinants of infant (during first year of life) andchild (between one and four years old) mortality?

Data & Methods

For our analysis, we used the data ofthe 2014 Egyptian Demographic and Health Surveys (EDHS), in which theyinterviewed 19.757 Egyptian women, mothers of 59.266 children. The reportincludes almost one thousand questions (for 1.082 variables in the data-set) ondifferent topics: basic information on the respondent (date of birth; residence;job at the interview; education; religion; and so forth) and their partners aswell as on fertility preferences, contraceptive methods, attitudes towarddomestic violence and methods of making decision inside the household.

From the statisticalpoint of view, we use event history analysis, so to follow a life courseperspective. For infant mortality, we follow each child from the birth date tothe death or the first birthday, whichever comes first. Exposure time wascalculated, in this case, in days. For child mortality, we follow each childfrom the first birthday to the death or the fourth birthday, whichever comesfirst. In this second case, we calculated the time in months.

            In both, infant and childmortality, we used the same variables:

For the specification of the basictime factor, we fitted several specifications (log-normal, cox, Weibull, andexponential) and then verified the best fitting using AIC and BIC criteria. Themodel we found the best was the piecewise constant exponential model.


First Results

For both the infant and the child mortality,model output confirms the results obtained in a preliminary analysis conductedthrough Kaplan-Meier survival curves. Indeed, in both cases, higher levels ofwomen’s empowerment are associated with lower levels of infant and childmortality: women with more decision autonomy and higher education level have a lowerrisk of experiencing the loss of their children. In a similar vein, theviolence justification index shows a negative correlation with the childrenmortality.

Demographicfactors – child’s birth order and mother’s age – seem to play the decisive rolefor infant mortality. Differently, for child mortality, socio-economic factorsand, especially women’s emancipation, show greater risk ratios, whilebiological factors lose their importance.

In childmortality, unlike infant, the decision-making autonomy is not significant: thisis due to the effect of the level of education that acts as a confoundingfactor because it simultaneously affects the decision index and the childmortality. By not including the education level in the model, in fact, thedecision-making index appears to be significant. We can therefore consider thelevel of education as an antecedent variable: mothers who make more decisions arethose who are better educated and higher education lowers the risk of childmortality.

By estimating interactions effects between mothers’birth cohorts and our three indexes of women’s empowerment (results not shownhere), we find that women’s education, decision-making autonomy, and attitudestowards domestic violence are not losing explanatory power in predicting infantand child mortality. The positive relationship between women’s empowerment and childhealth care, still important among recent cohorts of Egyptian mothers,represents an indication for politics towards the need of strengthening genderand social equality.

Presented in Session 1186: Mortality and Longevity