Multilevel determinants of fertility behaviour in Eswatini

The study re-examines the levels, trends and determinants of fertility in Eswatini. The overall objective of the study is to understand fertility behaviour of Swati women. Fertility behaviour is explained by actual (current and lifetime) fertility and proximate determinants of fertility measures. The following variables were considered as regression dependent variables: i) lifetime fertility (children ever born), ii) current use of contraception. The study used data from four population censuses (1976, 1986, 1997 and 2007) and 2007 Eswatini Demographic and Health Survey (DHS). The study employed direct and indirect estimation methods (P/F ratio methods-Brass, Trussell, Arriaga, Feeney; Relational Gompertz model and regression based methods-Rele’s, Palmore’s, Gunasekaran and Palmore methods), to estimate the level and trends of fertility in Eswatini. The impact of the proximate determinants of fertility on the level of fertility was assessed using the Bongaarts model and its modifications. Bivariate and multivariate (or multilevel) Poisson and logistic regressions analyses were used to examine the correlates of fertility and contraceptive use in Eswatini. The study found that between 1966 and 1976, fertility remained constant and thereafter fertility declined from 7.0 in 1976 to 4.6 children per woman in 2007. Contraceptive use among sexually active women is 46.1% of which 1.9% used traditional methods and 44.2% used modern contraceptives. Among the individual-level factors influencing fertility were child mortality, women’s empowerment, ideal number of children, employment, marriage, female education, wealth, age at first sex and age at first birth. At community-level, rural-urban residence played a significant role in shaping fertility in Eswatini. On the one hand, fertility was positively associated with child mortality, women’s empowerment, ideal number of children, employment, marriage and rural residence. On the other hand, fertility was negatively associated with female education, wealth, age at first sex, age at first birth, but not associated with mass media exposure and contraceptive use. The analyses of the proximate determinants of fertility revealed that contraception had the greatest impact on fertility reduction, then postpartum infecundability, sexual activity and induced abortion. In terms of contraceptive use, the study indicated that the following individual-level variables are significantly associated with contraceptive use in Eswatini: occupation, wealth, education, exposure to mass media, women’s empowerment, marital/union status, age, parity and age at first birth. The study recommends, among other things, improving female education, delaying age at first birth and sexual debut, and reducing poverty, child mortality and desired number of children as measures that can be adopted to accelerate fertility decline in the country. In addition, the study recommends more studies on aspects such as women’s empowerment, employment, marriage patterns, and rural-urban divide disadvantages which could provide additional insights into fertility behaviour. Efficient and effective fertility reducing programmes and interventions should be directed towards individual-level differences as well as considering the urban-rural community contexts in Eswatini.
A thesis submitted to the Faculty of Humanities, University of the Witwatersrand, Johannesburg, in fulfillment of the requirements for the award of the degree of Doctor of Philosophy in Demography and Population Studies Johannesburg, 2018
Chemhaka, Garikayi Bernard. (2019). Multilevel determinants of fertility behaviour in Eswatini. University of the Witwatersrand.