Relative Risks and the Market for Sex: Teenagers, Sugar Daddies and HIV in Kenya

This study reports on an information campaign that provided Kenyan teenagers in randomly selected schools with the information that HIV prevalence was much higher among adult men and their partners than among teenage boys. This led to a 65% decrease in the incidence of teenage pregnancies by adult partners in the treatment group relative to the comparison group and suggests a large reduction in the incidence of unprotected cross-generational sex. The information campaign did not increase pregnancies among teenage couples. These results suggest that the behavioural choices of teenagers are responsive to information on the relative risks of different varieties of a risky activity. The author argues that policies that focus only on the complete elimination of a risky activity and do not address smaller risk reduction strategies, such as avoiding sex with a particular age group, may be ignoring a margin on which they can have substantial impact.

In sub-Saharan Africa the prevalence of HIV is considerably higher among young women than among young men. In Kenya and Zambia, for example, prevalence in the 15-19 year old age group has been found to be at least five times higher among girls. Recent studies suggest that this discrepancy is due, in part, to the high incidence of consensual, unsafe, cross-generational sexual relationships – that is, unprotected sex between teenage girls and adult men five or more years their senior. Men involved in these relationships, often called "sugar daddies", are more likely to be infected with HIV than teenage boys since they have been sexually active for longer. In addition, as older men are financially more powerful than young men, adult men are able to negotiate for unprotected sex. Thus, compared to relationships with teenage boys, cross-generational relationships pose a higher risk of HIV infection for teenage girls.

By looking at the results of a randomised experimental information campaign conducted in 328 schools in Kenya, the author argues that providing information on HIV prevalence disaggregated by gender and age group - highlighting the higher risk of sex with older men - could help young women reduce their risk. An information campaign conducted by a non-governmental organisation (NGO) in 71 randomly selected schools provided teenagers with information on the prevalence of HIV disaggregated by age and gender group. Teenagers in the remaining 257 schools did not receive this information and served as the comparison group. The randomised design ensured that there would be no systematic difference in the prior information held by the students across the treatment and comparison groups before the start of the information campaign. This allowed the research to identify the impact of the additional information by comparing behaviours and outcomes across groups 6 to 12 months after the intervention.

The research found that the information campaign led to a 65% decrease in the incidence of pregnancies by adult partners among teenage girls in the treatment group relative to the comparison, suggesting a large reduction in the incidence of unsafe, cross-generational sex. The information campaign also led to an increase in self-reported sexual activity among teenagers with partners from within their cohort, suggesting a substitution effect. However, there was no increase in pregnancies among teenage couples, since condom use also increased. Data to measure the exact impact on the incidence of HIV infections is not currently available, but this report suggests a long-term follow-up using biomarkers of sexually transmitted infections could be conducted in the future.

According to the author, the findings of this paper have important implications. Public health interventions often focus their efforts on the extensive margin of a risky behaviour: they aim at the complete elimination of the behaviour and urge complete abstention from the activity. Accordingly, they rarely provide information on the relative riskiness of different varieties of a risky activity - information that would enable people to reduce the intensity of their exposure to risk while remaining active. The author states that the amount of information that a prevention campaign should provide in order to maximise its health impact depends on the relative size of two elasticities: the elasticity between high-and low-risk varieties of an activity and the elasticity between the low-risk variety and no activity at all. The empirical evidence presented in this paper suggests that, in the case of sexual behaviour, the former is larger than the latter. This result suggests that HIV education campaigns may achieve a wider health impact if they include both risk reduction and risk avoidance information. The author suggests that further work to assess the overall welfare implications of providing information on relative risks will be important in order to allow governments, as well as health and education specialists, to design the most effective information campaigns.