The dual role of income in the spread of HIV in Africa.
Every day more people in Africa are infected with HIV despite prevention efforts. These new infections and those already infected are not evenly spread throughout the continent. Substantial variations in HIV prevalence exist within and between countries. Understanding these variations helps understand what is driving the epidemic and this understanding in turn helps in the design of more appropriate interventions to prevent its further spread. This thesis builds on existing work by attempting to develop a more comprehensive theory of what role income plays in the spread of HIV. To this end the Theory of the Dual Role of Income in the Spread of HIV is outlined and explained. It uses the concepts of relative and absolute income, borrowed from elsewhere in the health economics literature, to separate the different effects income has on individual and group risk of HIV infection. The theory hypothesises that, while higher levels of absolute income (income independent of others) offer protection against infection via better access to health care and information, higher relative income (the income of an individual relative to other members of their social or reference group) increases risk of infection either as a result of more sexual partners or higher risk partners. The theory in no way argues that HIV infections are not related to poverty, but rather that the relationship is, somewhat more complicated and non-linear than often suggested. The explanatory power of the theory is examined with the use primarily of two data sets: firstly using data collected from antenatal clinics in two South African provinces linked with census data and secondly with data on a large South African company. While these data are not ideal, the results from the analysis are in line with the expectations based on the theory. The theory and the results of the analysis presented in this thesis support the argument that environments in which decisions are made and actions taken are important in determining risk of HIV infection. This argument suggests that prevention efforts need to do more then provide information.