Mother’s Educational Attainment: Does a Mother’s Age Group Moderate the Heterogeneity Effect of Fertility and Child Health Outcome?

Reader Impact Factor Score
[Total: 3 Average: 5]

Published on International Journal of Social, Politics & Humanities
Publication Date: May 28, 2019

Nubonyin Hilda Fokong
Department of Economics, The University of Bamenda

Journal Full Text PDF: Mother’s Educational Attainment in Cameroon: Does a Mother’s Age Group Moderate the Heterogeneity Effect of Fertility and Child Health Outcome?.

This paper examines the role played by mother’s education in expanding children’s nutritional capabilities in Cameroon, by focusing on whether mother’s age group moderated the association between maternal fertility and child height for age z score. Data on 11732 women aged 15-49 years drawn from 2001 Cameroon Demographic and Health Survey were used. Instrumental Variable-Cum-Heckman Selectivity Models were used for analysis. The results reveal that maternal fertility correlates negatively and significantly with child health outcome, more so for mothers less than 30 years and greater than 40 years. These results have implications for addressing maternal fertility, maternal educational attainment and child health concerns in the on-going process of growth and poverty reduction in terms of improving access to family planning, medical and educational services.

Keywords: Mother’s educational attainment, mother’s age group, heterogeneity effect, fertility, child health outcome, Cameroon.

1. Introduction
The education of women not only brings benefits to themselves but also to the next generation. Studies have suggested that women who have more equality in education, employment and decision-making tend to have lower fertility and their children have lower risk of malnutrition (Baye and Sitan, 2016; Fambon and Baye, 2017). Higher women’s education is also associated with fertility via the age of women’s at first birth (Tambi, 2014). In Cameroon, the age cohorts 10-16 and 17-19 contribute for 2.82 per cent and 11.415 per cent of deliveries between 2003 and 2005 respectively (Tebeu, Kemfang, Sandjong, Kongnyuy, Halle and Doh, 2010). This age-specific fertility rates indicate that, there is a problem of Cameroonian women giving birth to children at a very early age (Tebeu et al., 2010; Walker, 2012). Based on the above evident one is lured to find out whether the diverse maternal fertility pattern implications of the age cohorts (10-16 and 17-19) in Cameroon explains the problem of child health outcomes that are more likely to be poor.
In 2004 specifically, 5.3 per cent babies in Cameroon weighed less than 2500g or 2.5Kg (National Institute of Statistics, 2004). According to WHO, UNICEF, and the World Bank (2015) the rate of children below the age of five stunting in Africa is 31.6 per cent which is better than the 33 per cent in Cameroon in 2011. Related to this view is the observation that, the fertility rate in Cameroon is 4.8 per cent live births per woman (International Monetary Fund, 2018) with 7 per cent of women having a body mass index (BMI) less than 18.5 and 29 per cent over 25 (Nanfosso and Zamo-Akono, 2010). However, the highest BMI (25.5) is being observed in Yaoundé and Douala (Nanfosso and Zamo-Akono, 2010). The above statistics points to the fact that, linkages between women health status or fertility and child health conditions in Cameroon are not better. Thus, to provide a lasting solution to the problem the government of Cameroon adopted the sustainable development health goals (United Nations, 2015). Notwithstanding government efforts, preliminary research explains that low maternal and child weights, early marriages, low level of schooling, low maternal labour force participation, unwanted pregnancies, mother’s age at first birth, inadequate use of contraceptives among other fertility factors continue to affect child health in Cameroon (Baye and Sitan, 2016; Tambi, 2014).
It is worthy to note that, the relationship between women’s fertility and child health outcome has been a major research focus in theories of human capital and fertility decline (Becker, 1962; 1965 and 1981). The improved education of women is hypothesized to increase their health outcome, economic opportunities, leading to improvement in child health in part through a higher prevalence of contraceptive use and thus, declining fertility. Even though the topic has been intensively studied, just how women’s fertility affects aspect of child health remains a matter of debate for several reasons. The first reason is the wide variation in the conceptualizations and frameworks to measure child health. The second reason is the inconclusive linkages that have been established between women’s fertility and child health. It is on this precedence that the objectives of this paper are: to determine the factors influencing child health outcome in Cameroon, to evaluate the effects of maternal fertility on child health in Cameroon and to examine the effects of maternal fertility on child health by mother’s age group in Cameroon.

2. Literature Review
From a theoretical perspective, the health status of individuals has been explained by the human capital theory (Becker, 2007; Grossman, 2000; Becker, 1981). Indeed Becker (1981) analyzed the demand behaviour for child health services by a mother using a microeconomics fertility model in which the child is embedded in the utility function. Thus the model assumes that households maximize utility over market goods, child health (quantity and quality of children), and leisure. The utility function is assumed to have all the standard properties of neoclassical production function (quasi-concavity, continuity and differentiability). The model assumes that the household faces three constraints: a time constraint, a budget constraint, and a child health production function.
It is worthy to note that, Becker’s (1981) model had limitations in developing countries. For example the model is too static as it does not allow for changes in preferences during lifetimes (De Bruijn, 1999). However, for Cameroon, empirical findings indicate that education and health constitute key components of household economic welfare because they directly and indirectly affect household utility and production functions (Tchombe, 1994; Epo and Baye, 2011). This implies that Becker’s model has been strongly utilized in developing countries like Cameroon, since the framework of the model reveals that maternal education affects child health in part through changing maternal health behaviour.
From an empirical perspective, Delemarre-Van DeWaal (1993) investigated the influence of environmental factors on pubertal development in Europe, North America and Australia. He used regression analysis. Interestingly, the findings show that primiparous mothers either older than 38 years or younger than 20 years of age have an increased risk of giving birth to shorter babies. The study concluded that postnal environmental factors such as maternal age, birth order, nutrition, chronic diseases, socio economic status, body size, season/climate, urbanization as well as industrialization had an effect on children’s height for age z score. However, the study was mostly descriptive in nature. In another study, Pitt and Sigle (1997) investigated the effects of seasonality and weather shocks on the joint determination of fertility and child survival. They used 1986 demographic and health survey of Senegal merged with monthly rainfall data. They used a bivariate random effects probit model. The results show that rainfall correlates positively with fertility while fertility correlates negatively with schooling. Interestingly, the findings reveal that infants born in colder months were shorter than those born in hot months meaning the cost of raising children increases. They further argued that increase temperature increase birth length.
Desai and Alva (1998) examined the causal relationship between maternal education and child health. They used data from the first round of demographic and health surveys for 22 developing countries. They analyzed the data using ordinary least square, logit regression and fixed effects linear regression estimation techniques. Findings from the study showed that the effect of maternal education on height-for-age was significant only in 5 of the 22 countries studied. This corroborates the findings from an exploratory study that, children from ‘carer-mothers in Australia’ who were employed were healthier than those whose mothers had more education (Ewald and Boughton, 2002). However, both studies were constrained by the size of the available population and this led to incomplete information being collected for some sample.
Using DHS data from Sub-Saharan Africa, Giroux (2008) examined the relationship between maternal fertility and child stunting in Burkina Faso, Ghana, Madagascar, Tanzania, Zimbabwe, and Zambia. He used the logistic regression model. Giroux (2008) results reveal that stunting declines as fertility levels decline however, other countries (such as Cameroon and Burkina Faso), have experienced an increase in stunting during fertility decline. The results also indicate that children from families with more children in the household under age 15 were less likely to be stunted. The findings further reveal that mothers’ education is positively correlated with child stunting in all countries included in the analysis, with the exception of Zimbabwe. Child age is positively associated with stunting, whereas family socio economic status reduces the changes in stunting in all models. However, two major limitations with Giroux (2008) study were that of the sample of children only including those who live with their biological mother and neglecting fostered children. In addition, only children of women age 15-49 years were included in the survey.
In a further twist, Mwabu (2008) evaluated the child health production using a structural model of birth weight. He used data from 10,000 households and the instrumental variable approach. Mwabu (2008) results show that maternal education is negatively associated with the probability of first birth. His findings further depict that babies born to immunized mothers were 769 grams heavier than babies of mothers who had not received tetanus vaccination. The findings rely on untested assumptions and unavailable data on some indicators which acted as a limitation to the study.
Miller and Rodgers (2009) estimated the effects of mother’s education and sex of the child on child health outcomes. Data from Cambodia’s 2005 Demographic and Health Survey were used. Miller and Rodgers (2009) bivariate and multivariate regression results reveal that mother’s education was strongly inversely associated with risks of both small birth size and stunting (height for age z score) but, not with wasting (weight for height z score). Based on the full multivariate results, children born to women with at least secondary education were roughly one-third less likely than those with uneducated mothers to be stunted.
A similar pattern of results is reported by Güneş (2013) who used the most recent Turkish Demographic and Health Survey (TDHS-2008) and investigated the causal relationship between mother’s primary school completion and infant health at birth. Güneş (2013) analyzed the data using instrumental variable estimation technique. His findings indicate that, mother’s primary school completion improves infant health. The results also show that effects of maternal education on child health depend on province characteristics and the sex of the child. The results also suggest that, education, region, urban/rural residence as well as pregnancy experience and ethnicity are important factors of infant health and health care utilization respectively. Makoka (2013) used DHS datasets (from Malawi, Tanzania and Zimbabwe), bivariate analyses and Pearson Chi-square test of independence to test the association of maternal education and child nutrition. Makoka (2013) results indicate that child nutritional status significantly decreases with increased levels of mother’s education. The results further show no clear pattern between the mother’s body mass and child stunting in the three countries. While mothers who were overweight (BMI ≥25.00) were less likely to have stunted children, the results were only significant for Malawi and Tanzania. He argued that wealth is the main socioeconomic factor that seems to explain whether a child is stunted or not.
Rutstein and Winter (2014) assessed the effects of spacing between birth, maternal age at the child’s birth and the child’s birth order on child survival and nutritional status. They used data from 45 DHS surveys between 2006 and 2012.They used bivariate and multivariate logistic regression. Their results indicate that that optimal age of mothers to conceive children is between 25 and 39 years while the optimal month of birth interval is 36 to 59 months. They argued that birth interval of 24 to 35 months for these mothers are associated with 20% greater risk of having stunted children. Children whose mothers were 40 years or over at the time of their birth have 23% lower likelihood of being stunted. With regards to birth order the findings reveal that increasing birth order is associated with child being stunted because of poor nutrition outcomes, possibly through sibling competition and the mother’s availability for child care and decrease household socio economic status. Increasing mother’s age at first birth reduces a child’s risk of being stunted. The study stresses that, the more fertility behavioural risk factors a child faces the greater are the chances of dying and being malnourished.
In Cameroon, Tambi (2014) used the control function approach and DHS data (NIS, 2011) to investigate the implications of mother’s age at first birth on child health at birth and birth weight. The result reveals that mother’s age at first birth is negatively correlating with birth weight. The findings further show that the three principal variables that commonly influence child health at birth are: mother works in the agricultural sector, household residence and the presence of father in the house. These variables are all negatively correlated with birth weight. Using data from the Uganda Demographic and Health Surveys (DHS), Malaria Indicator Surveys (MIS), and AIDS Indicator Surveys (AIS), Keats (2016) examined the role of women’s education on fertility and child health in Uganda. He used the instrumental variable approach. The findings of this study reveal that women with more schooling get married at a later age, they delay fertility and as such reduce their overall fertility, they increase early child health investments and above all they have less chronically malnourished children.
Iftikhar, Bari, Bano and Masood (2017) studied the impact of maternal education, employment, and family size on nutritional status of children. They collected data from 340 children (170 cases and 170 controls) with age range of six months to five years and 340 mothers. The data collected by Iftikhar et al. (2017) were analysed using univariate regression technique and logistic regression analysis. The findings established that maternal education had definite and significant effect on nutritional status of children.
Consistent with earlier claims by Baye and Sitan (2016) that, maternal fertility exhibited a negative and significant effect on child health, recent research by Fambon and Baye (2017) in Cameroon concluded that household consumption expenditure is positively and significantly associated with child nutritional status. Notably, the aforementioned studies in Cameroon have made use of the control function approach and instrumental variable approach respectively. Unfortunately, these works used different types of anthropometric indicator(s) for child health outcome, different datasets and different sets of instrumental variables. Interestingly, the present study differs from these previous studies in terms of the anthropometric indicator for child health outcome, the DHS dataset and sets of instrumental variables.