Sex, AIDS & Exports in Africa

CGD held a fascinating lunchtime discussion earlier this week with "scientist morphed into journalist" Helen Epstein, about her new book, The Invisible Cure: Africa, the West, and the Fight Against AIDS. During her talk, Helen argued persuasively that many international policymakers lack an adequate understanding of the important role of long-term concurrent sexual relationships that have been a primary driver behind the epidemic's rapid spread in the worst-affected African countries (discussed further here). She presents evidence to suggest that the remarkable declines in HIV prevalence in Uganda can be attributed to unique community-based characteristics of Uganda’s AIDS prevention campaign that helped average adults to recognize the risk in concurrent partnerships and to reduce the numbers of such partners. She concludes that other countries should shift the mix of their prevention interventions to assure that people understand the risk of unprotected sex with long-term concurrent partners.At first glance, Helen’s argument that concurrency is the most important determinant of the spread of HIV in much of Africa seems to directly contradict the work of economist Emily Oster (see, for example, her recent presentation at the TED conference). Emily's prior research has included looking at increased transmission rates due to co-infection with other STIs as a primary driver of Africa's AIDS epidemic, combined with behavioral differences resulting from lower discounted life expectancy; now, she has further identified export patterns as a key determinant of changes in infection rates over time. In Uganda, for example, she finds that much of the fluctuation in new HIV infections can actually be attributed to national economic fluctuations as measured by changes in Ugandan exports. (Incidentally, Emily also gets enormous credit for the hard work of generating new and quite plausible estimates of HIV infection, both incidence and prevalence, by inferring from existing sibling-based mortality data; using this method, she finds that UNAIDS might have been substantially overestimating prevalence rates throughout the 1990s.)In my view the two theories do not contradict, but rather complement each other. First, since Emily is arguing that exports only explain between 30% and 70% of the decline in the prevalence rate in Uganda, that leaves between 70% and 30% to be explained by other factors - including Uganda's HIV prevention program. Second, a decline in the number of concurrent partners could be driven by both a decline in men’s pocket money and by a community-based HIV prevention program. A recent Washington Post article suggests that a reduction of economic resources among men can reduce the demand for concurrent partners. With reduced exports, there is reduced income and reduced travel among men who might have several concurrent partners. In the language of demographers, reduced numbers of concurrent partners might be a "proximate" determinant of HIV incidence, while exports are a "distal" determinant. An HIV prevention program that enabled men and women to better understand the risks with multiple concurrent partners may well have been synergistic with reduced exports and the consequently reduced incomes in reducing the number of new infections in Uganda.Third, and most intriguing is that Emily shows a strong estimated impact of exports on new HIV cases in several African countries, not just in Uganda (see Figure 8 as well as Figure 2 of her online working paper). The graphs alone are convincing for most of the countries, but the argument also holds up when she does sophisticated econometrics, using prices as instruments for exports. Unfortunately, data isn't available on concurrent partnerships or on the characteristics of HIV/AIDS prevention campaigns in these other countries that would enable us to follow the mechanism by which export fluctuations affect HIV incidence and to understand how the characteristics of HIV/AIDS prevention programs affected the outcomes. In my view the most important contribution of Helen's work is to signal that sex is risky not only with commercial sex workers and casual partners, but also with regular long-term partners who might have other partners. She is right to urge that the riskiness of concurrent long-term partners be communicated in HIV prevention programs. Some people who hear these messages will reduce the number of their partners, even in booming economies. However, as long as increases in men's income translate into increased demand for sexual partners, and as long as their increased incomes and mobility make them more attractive and available to women, access to barrier prevention methods like male and female condoms will continue to be essential elements of government-supported HIV prevention programs.Update: CGD non-resident fellow Bill Easterly shares his own thoughts on The Invisible Cure in the New York Review of Books.