A Quantitative Theory of HIV Diffusion

This paper develops a model of HIV diffusion among heterosexual individuals who choose (i) whether to engage in sex, i.e. the source of potential infection, (ii) the nature of their relationship as measured by the endogenous rate of partnership destruction, (iii) the presence of extra-marital partners (i.e., concurrency) and (iv) whether to have a child or not. We study the general equilibrium of this economy where the distribution of the population is endogenously determined in stationary and nonstationary environments. We use a calibrated version of the model to analyze the quantitative implications of a wide range of reproductive health policies that target AIDS treatment and HIV prevention. Our preliminary results look promising. First, the theory is able to generate high and low aggregate HIV prevalence scenarios associated with mature epidemics. Second, we find that changes in several reproductive health policy parameters generate large aggregate and distributional changes in the long-run patterns of HIV infection and prevalence. For example, free universal access to antiretrovirals generates an increase in aggregate HIV prevalence from 9.7% to 34.4%; increasing the stability of sexual partnerships by 2 years substantially reduces HIV prevalence to almost zero; reductions in the level of infectiousness (condom use, gels, etc.) per sexual act succeed in lowering the aggregate HIV prevalence to zero; reducing mother-to-child transmissions to zero implies a reduction of 33% of the aggregate HIV prevalence; reductions in the efficiency of the matching technology reduce HIV prevalence by a factor of .8. Distributional effects across gender, marital status and type of sexual partner (which includes the possibility of serodiscordant couples) are also analyzed.