1. Financial Incentives & HIV prevention in Sub-Saharan Africa: Evidence from Three Experimental Interventions Berk Özler & Damien de Walque Development Research Group (DECRG) The World Bank November 10, 2009

2. PREMISE AIDS Treatment: Important, but not the solution to slow the HIV/AIDS epidemic. Estimation that $1 in prevention averted $43 in treatment in Thailand (Over et al. 2007) Vaccine technology: Probably still many years in future. Existing prevention efforts: Low effectiveness in most African settings. Novel approaches needed! Conditional cash transfers (CCT) have been promising in other domains … could they be adapted to help slow the AIDS epidemic?

3. The SIEF Cluster tests 3 different ways to apply the CCT concept for HIV prevention The program in Malawi is a traditional CCT program for schooling, but evaluates the impact of the program on the sexual behavior and HIV risk of its young female beneficiaries. In Tanzania, the intervention offers cash transfers conditional on periodic negative tests for curable STIs. In Lesotho, lottery tickets are offered conditional on periodic negative tests for curable STIs. 3

4. Common features Application of the CCT concept Rigorous impact evaluations with baseline and follow-ups and random allocation of the intervention All three evaluations are collecting biomarker data on HIV and other STIs in addition to self-reported data on sexual behavior. The 3 evaluations are at different stages. One year impacts will be presented for Malawi, while we will present the evaluation design for Tanzania and Lesotho. 4

5. Rewarding STI Prevention and Control in Tanzania (RESPECT Project) Implementing institutions Ifakara Health Institute World Bank Development Economics Research Group UC-Berkeley Funding from the World Bank (Research Committee and Spanish Strategic Impact Evaluation Fund) and Hewlett Foundation/Population Research Bureau.

6. OVERVIEW What: “Proof of concept” evaluation of randomized CCT to incentivize reduction in risky sex. Why: Goal is to decrease HIV incidence, with potential subsequent long-run health and economic benefits. How: Condition cash incentives on periodic negative STI tests for young adults aged 18-30. Where: Ifakara Demographic Surveillance Site in rural Tanzania.

7. CCTs to Reduce Risky Sexual Behavior? Conditionality: Increase “price” of risky sex, if positive STI test causes loss of CCT. Time discounting: Bring rewards of risk reduction closer to present, rather than just avoiding AIDS many years in future. Income effects: Potential direct impact on sexual behavior, could facilitate behavioral change, especially for women.

8. SPECIFIC OBJECTIVES Evaluate the impact of the combined CCT/counseling intervention during the intervention period (immediate and short-term effects) on STI incidence overall and by specific subgroups Evaluate the economic outcomes of the reward. Examine the long-term effects of the intervention – and its withdrawal – with final round of STI testing and surveying in the same population 12-months after the intervention has ended Compare the impact of the CCT intervention in the high-value cash transfer arm to that in the low-value cash transfer arm

9. CONDITIONAL CASH TRANSFER Conditionality: Treatment group  Testing negative for a set of curable STIs (Chlamydia, gonorrhea, trichomonas, mycoplasma genitalium, syphilis).  Although we are testing for them, we do not condition on HIV and HSV-2 because they are not curable. Amount:  USD 20 every four months (high-value)  USD 10 every four months (low-value) We provide free STI treatment if positive, and we allow returning under CCT once they are negative again

10. TANZANIA: STUDY DESIGN The study is a Randomised control trial Participants 3000 Counseling and life-skills training, free STI treatment Intervention group (Conditional cash) N=1,500 High-value N=750 Control group N= 1,500 Low-value N=750 Sample drawn from 10 villages, randomization at the individual level.

11. The impact of short-term financial incentives on sexual behavior and HIV incidence among youth in Lesotho Partnership between the World Bank and PSI Lesotho Same idea as in Tanzania, but instead of getting a cash transfer for sure if they test negative for a set of curable STIs, study participants will get a lottery ticket for a larger prize (500 rands/65 $ or 1000 rands/130$) Lottery system might be easier and cheaper to scale- up Lottery should be especially attractive to risk takers. 11

12. LESOTHO: STUDY DESIGN The study is a Randomised control trial Participants 5000 Pre and post test counseling, free STI treatment Intervention group (Lottery tickets) N=3,000 High-value lottery N=1,500 Control group N= 2,000 Low-value lottery N=1,500 Sample drawn from villages where PSI is conducting mobile VCT clinics, randomization at the individual level