Risk and Rationality : Effects of contextual risk and cognitive dissonance on (sexual) incentives

Sammanfattning: Paper [I] theoretically analyzes how the level and uncertainty of future prospects affect incentives to abstain from sexual risk taking in the presence of HIV. The results suggest that, for individuals with limited access to HIV treatment, uncertainty of future health may be an important factor driving unsafe sex practices and support the empirical finding of a weak link between sexual behavior, HIV prevalence, and HIV knowledge in poor countries; therefore suggesting that AIDS policy needs to be calibrated in order to fit within different social contexts.Paper [II] empirically tests the link between uncertainty of future prospects and sexual risk taking in a group of young adults in Cape Town, South Africa. The findings indicate that expected income and health and future uncertainty are significant determinants of current patterns of sexual risk taking. However, the empirical results only provide limited support to a link between expected health and sexual risk taking.Paper [III] theoretically analyzes effects of affect and defensive denial on incentives to engage in sexual risk taking related to HIV. The results of the theoretical analysis suggest that the effect of rationalization of personal risk depends on the risk of being HIV positive. Although rationalization causes excessive risk taking behavior for individuals with a relatively low lifetime risk, it may prevent fatalism among individuals whose lifetime risk of HIV is perceived as overwhelming.Paper [IV] theoretically analyzes the role of identity conflict for the evolution of female labor supply over time. The results suggest the fear of becoming an outsider in society may have prevented a complete transition of women from housewives to breadwinners. In addition, our analysis shows that not recognizing that the weights attached to different social identities are endogenous may imply that the long-run effects on labor supply of a higher wage may be underestimated.