Cost-Benefit, Oster, and AIDS
Picking up the theme Ben mentioned earlier, economist Emily Oster suggests (see our prior blog post here) that exports helped determine the rate of diffusion of AIDS in Uganda. As prices of Ugandan coffee exports declined, men had less pocket money which made it harder for them to have more sexual partners. This view actually could make it possible for Helen Epstein’s arguments on concurrency to co-exist with Oster’s work, as the Center for Global Development argues. They suggest that there could be a problem of people having multiple sexual partners and also true that export price fluctuations might alter how many partners they do have. Here’s a video link to Oster’s presentation at TED (Technology, Entertainment, Design).
It’s interesting to think about Oster’s thesis, and I’m not sure it sits that well with Epstein’s recent book. If Oster is right, then it may be harder to replicate Be Faithful programs around the continent through conscious efforts. It may be harder for people to be faithful if commodity prices increase and give people more income to be able to afford to support more sexual partners.
However, I’m not sure if Oster’s thesis is internally consistent. If people have more money, then, according to Oster, a cost-benefit ratio over the long-term would suggest they have more to live for, so they shouldn’t wreck it by being risky. On the other hand, she also finds that a large proportion of the variance in AIDS rates is explained by variations in export revenue. With higher export revenue, AIDS prevalance goes up. What gives? Am I wrong or are these two conjectures at odds? I believe there may be a correlation there with declining export revenue and prevalance in Uganda, but is the causal connection between walk around money and how many girlfriends men could support right?
Postscript: Oster’s old paper is the one that made the claim that richer people should be more risk averse because they have more to live for. Her newer paper makes the claim that higher export earnings are correlated with higher prevalence levels. It seems to be driven by volume and the time truckers are away from home rather than individual earnings power. More exports means more time demanded of truckers to be away from their families. However, that must also mean more money in their pockets which should mean that, as any forward thinking rational animal, truckers should put a sock on it. Sounds like they don’t which sits uneasily with Oster’s previous work, as far as I can tell.
Postscript 2: I had a good interchange with Oster about this issue, and she sought to reconcile it by talking about long-term vs. short-term behavior change. She wrote that the subject of her first paper was how making people richer over their whole lives could motivate behavior change. It wouldn’t happen overnight, as it relied on shifts in people’s expectations about the future, and she noted that the effects were small.
Whereas the first paper relied on long-term behavior change, she suggests the export story in her new paper is really short-term, as trend data has been explicitly removed. With more exports she says, there are more truck drivers. They have more sex partners, driving epidemic rates up. Over the long-term, high exports could make the country richer, leading people to protect themselves, but in the short-run, the issue is one of more exports–>more truckers–>more risky sex–>more HIV.
The more I read her paper, and from her e-mail, it’s really not about income, but it’s about export volume and the presence of more truckers on the road. She disputed the notion that export volumes in the short run would necessarily mean more income. But, unless truckers have more pocket money, how are they able to enter into these sexual relationships?
I think the mechanism influencing the number of sexual partners needs more elaboration. Do truckers’ change their number of sexual partners in response to education/information, income, or time away from home. It seems like Oster’s thesis comes down to the third explanation. Guys are away from home, and no matter how much money they have in their pocket, they will engage in risky sex. The fact that Uganda experienced a a decline in prevalence wasn’t because the truckers got information or even that they had less money, but there were simply fewer of them on the road. This may be a hasty read of her argument, but I still think it needs some tweaking/refinement.
Filed under: Epidemiology, Politics and Policy
Still more research would be required before an association can be established between AIDS and the Economic situation.