Parenthood decisions and the transformation of the family structure

(eng) The study of demographic subjects among economists has flourished after the work of the Nobel Prize winner Gary Becker. This thesis contributes to this literature by looking at aspects of parenthood decisions that have already been studied by demographers and sociologists but much less by economists. The first two chapters study childlessness and its determinants. The last chapter studies how childcare is related to the education of parents. The first chapter is the first study in the economic literature on fertility to consider childlessness on its own. I show that childlessness is not a mere extension of a fertility decision problem. Over the long run, completed fertility and childlessness can have a positive relationship for some cohorts of women. This chapter provides an explanation for this counterintuitive fact using an endogenous fertility model where individuals have different preferences for children. The main mechanism considered goes through the inter-generational evolution of preferences. I show that a reduction in the gender wage gap, or an increase in the fixed cost of becoming a parent, have a negative effect on both fertility and childlessness. As the wage gap closes, fertility decreases because women earn more and this increases their opportunity cost of childrearing time. An increase in the cost of parenthood decreases fertility due to an income effect. As for childlessness, its reduction is due to a composition effect: the decrease in fertility makes small families to shrink more than larger families and this reduces childlessness because children from smaller families have a larger probability to be childless. This study does not consider either involuntary childlessness nor the effect of marriage on both types of childlessness. My second chapter includes these two important issues. In "DINKs, DEWKs and Co., Marriage, Fertility and Childlessness in the United States", co-authored with Thomas Baudin and my thesis supervisor, David de la Croix, we show how family patterns have been shaped by the rise in education and wage inequality, and by the shrinking gender wage gap in the United States. In particular, we explain that the U-shaped relationship between childlessness and education, reflects a predominance of women remaining childless for sub-fecundity reasons among the uneducated and, among the highly educated, a prevailing number of women remaining childless because the opportunity cost of motherhood in terms of foregone income is too high. Our theoretical framework allows quantifying the importance of voluntary and involuntary childlessness. Using the model to understand the changes that occurred over the period 1960-1990 we have learned that an increase in the education of men leads to a decrease in both involuntary and voluntary childlessness and an increase in the marriage rate of educated people. We have also shown that closing the gender wage gap is a powerful tool for limiting the proportion of involuntary childlessness. We also explore how marriage affects childlessness decisions. For uneducated women, marriage is a way to pool income and insure against involuntary childlessness. For highly educated women, husbands provide part of the childrearing duties, diminishing the opportunity cost of motherhood. The last chapter of my thesis builds on the idea that sharing childcare duties between partners provides an incentive to be married. This chapter compares the respective merits of two theoretical frameworks in explaining households' choices on childcare: a collective decision process and a semi-cooperative one. In particular, the facts to be explained are: (i) that the amount of time spent providing childcare increases with the education of both members of a couple, and (ii) that the gap between childcare provided by women relative to that provided by men decreases as both members of a couple are more educated. The collective framework used is benchmark in this literature (see the several works by Chiappori and his co-authors). The semi-cooperative decision process assumes that households make decisions in two steps. First, they cooperatively choose the amount of labor supplied by each member and then each partner chooses childcare individually. This timing relies on the assumption that partners can commit on what their labor supply will be (as members have to sign contracts at the beginning of adult life) but cannot commit on how much childrearing time each of them will provide. The collective approach does not allow to replicate the fact that the amount of childcare provided by the husband increases with the education of its female partner. The reason is that, with the existence of a wage gap between genders, efficiency implies that the woman will be more likely to specialize in childcare and, if her education is larger than that of her partner, she will both work and provide childcare. The semi-cooperative setup allows for a good matching with the data. The assumption about the timing of the choices implies that households have to make predictions about future decisions on childcare. This generates indeterminacy about the final equilibrium. To choose among the possible equilibria, I compare three selection criteria, the first selects the equilibrium that brings the highest utility to the husband, the second is the symmetric case for the wife, and the third chooses the final equilibrium in a random way, where each possible outcome has a probability to appear equal to the inverse of the number of possibilities. The random choice selection criteria is the one that provides the best fit with the data. The reason is that, among the possible equilibria, are situations where both partners work but only one specializes on the educational part of childcare provided to children. In this case, childcare increases with both partners education because the increase in one's education increases the opportunity cost of not providing childcare (the production of child quality positively depends on the education of parents) and also increases with the education of the partner as a highly educated partner decreases the amount of work done by the one supplying childcare. The change in the composition of the possible equilibria, with respect to the education of both partners, allows to explaining the facts.