Incentives and HIV/AIDS

I'm now in Cape Town, South Africa, finishing off some research with my colleague Brendan Maughan-Brown and starting some fresh work, as well. It's definitely great to be back here, and away from the dissertation and financial crisis (kind of - the South African market went tumbling after the failure of the bailout, and my e-trade account looks destitute), if at least for a few days.

More so than usual, incentives and behavior have been front and center in my mind. Not surprising, as these topics figure centrally in a paper Brendan and I just presented and submitted for publication (see here for an earlier post on the topic). Some background: HIV/AIDS is a big problem in South Africa, with perhaps as much as 20% of the working age population afflicted by the disease. South African public policy towards the disease involves two main components: government provided antiretrovirals and disability grants. The latter is the focus of our work.

Disability grants are very large transfers given to poor individuals who are too sick to work. Once these individuals become healthy again, the grant is revoked. However, the problem in South Africa is that high rates of unemployment (up to 40% in some areas) make it likely that many individuals will not be able to find jobs post losing their grants. A fair amount of scholars and policy-wonks worry that this may create a perverse incentive for individuals to forego or modulate their antiretroviral treatment in order to remain eligible for grants.

We decided to investigate this issue by using some interesting data on individuals with HIV/AIDS taking treatment who were followed during the period 2004-2007. Our main finding is that, while losing a disability grant is associated with drops in individual and household income, people do not get sicker and continue to adhere to treatment. We find that households appear to use other strategies to cope with the loss of a grant, including transferring young children (dependents) to other households. In all of our models, we try to account for a host of factors that might bias our results (individuals on disability grants might be worse off because they need to meet a means test, or better off because they are good at pro-actively getting what they want or obtaining social transfers; the causality between the outcomes and disability grant receipt might run the other way, etc).

I'm not too surprised by these findings. It seems a bit much to think that someone who has been AIDS-sick would want to go through that again under any circumstances. In the words of a University of Cape Town researcher who attended the seminar Brendan and I gave: "being poor is better than being dead."

Another interesting thing we found was that individuals were receiving disability grants long after their eligibility should have been up (people on highly active antiretrovirals get better within six to nine months). This, along some other interesting evidence, leads us to believe that people are really good at gaming the social security system. So perhaps this is another reason why people aren't responding to the perverse incentive: there are other ways to survive (that confer far less risk and disutility), including taking advantage of a broken system.

Drop me a line if you have any questions about our work: I'm really excited about this project and the several extensions we have planned, and would certainly love to discuss.