This paper develops a theory of optimal fertility behavior under mortality shocks. In an OLG model, young adults determine their optimal fertility, labor supply and life-cycle consumption with both exogenous child and adult mortality risks. We show that a rise in adult mortality exerts an ambiguous effect on both net and total fertility in a general equilibrium framework, while child mortality shocks unambiguously lead to a rise in total fertility, leaving net fertility unchanged. We complement our theory with an empirical analysis using a sample of 39 Sub-Saharan African (SSA) countries over the 1980-2004 period, examining the overall effects of the child and adult mortality channels on both total and net fertility. We find child mortality to exert a robust, positive impact on total fertility but no impact on net fertility, whereas a rise in adult mortality is found to negatively influence both total and net fertility. Given the particular demographic profile of the HIV/AIDS epidemic (killing essentially young, active adults), we then conclude in favor of an unambiguous negative effect of the HIV/AIDS epidemic on net fertility in SSA.

International audienceThis paper introduces variable markups in a horizontal-differentiation growth model by considering a larger class of preferences that nests the classic “CES” specification usually present in the workhorse love-for-variety models. Our first result is to obtain a generalized characterization of the Euler condition for this broader class of utility functions: in our model, the Euler rule features a supplementary term aiming at compensating the consumer for variations in the preference for variety along the consumption level. We are then also able to demonstrate that in our generalized framework, the economy’s balanced growth path displays both endogenous markups and a strictly positive growth rate of the number of available varieties (being the engine of growth). Finally, we show that under endogenous markups, the economy’s growth rate and firms’ market power can display a negative correlation, as opposed to the standard result obtained in the CES framework

No abstract