Reproductive Health and Labour Outcomes

Abstract

This thesis formalizes the interaction between income and fertility outcomes using a fertility model taking account of mediation. Despite the diversity of studies concerned with the determinants of fertility, one consistent nding is that the timing and the number of children born to a couple are ultimately determined by the interaction between a couple's fecundity and their contraceptive behaviour. For example, fecund individuals may or may not be sexually active, may or may not be using contraception, which would in uence their degree of risk for pregnancy. In this context, fecundity is a necessary but insu cient cause of fertility. The two variables, fecundity and contraception, are mediator variables in the causal path between income and fertility (Baron and Kenny, 1986). It follows that, understanding the interaction between biological and behavioural determinants of fertility holds the key to understanding fertility trends in any society. With our understanding of causal-mediation in place, this thesis extends the health investment model to develop a model of the sequential nature of the human reproductive process. Because reproductive health is commonly understood as the capability to procreate and the freedom to decide if, when and how often to do so, poor reproductive health outcomes have long been acknowledged as a main cause of economic hardship for women and their children, accounting for 18% of the total global burden of disease and representing 32% of the burden among women of reproductive age. Thus, the presented theoretical model of the demand for reproductive health services is a derived demand for reproductive health (with adjustment costs). Our theory suggests that the relationship between family planning services, their marginal costs, the stock of reproductive health capital and the woman's earnings volatility is essentially nonlinear. The theoretical predictions arising from the model are empirically supported from individual level South African data. Because a number of the health costs are associated with `mistakes', in the sense that childbearing might be ill-timed, because the reproduction process cannot be completely controlled, at least yet, the previous model is extended to account for uncertainty in the process. In other words, an assumption that contraception gives women perfect control over their fertility outcomes is considered to be too simplistic. Therefore, the focus remains on the e ectiveness of contraceptive use in controlling fertility, but in a setting of uncertainty. The model developed ts nicely into the analysis of the sequential nature of the human reproductive process, as it makes use of the continuation or option value of contraception, which is assumed to depend on potential future labor market prospects, and the uncertainty inherent in the reproduction process. The model naturally leads to a mixed hitting time (MHT) empricial model; a woman becomes pregnant when the underlying stochastic reproductive health process rst crosses a threshold in zero. We nd empirical support for our theoretical analysis.