How to Pay “Cash-on-Delivery” for HIV Infections Averted: Two Measurement Approaches and Ten Payout Functions

In contrast to current donor policy, which funds a recipient country’s national AIDS control program, this paper proposes a measurement strategy to enable a donor to reward a recipient country’s success at HIV prevention, irrespective of the inputs, activities, or who gets the credit. In accordance with the “cash-on-delivery” model of foreign assistance, the objective is not to replace traditional input- or activity-oriented aid, but to complement it by enhancing the motivation for local actors and their partners (including the traditional bilateral and multilateral funding agencies and their agents) to achieve measurable reductions in the rate of new HIV infections. This paper proposes two approaches to measuring the number of HIV infections averted between a baseline survey and a follow-up survey and explores the properties of ten alternative “payout functions” which would link measured epidemic changes to the size of the reward to be paid. All measurement approaches include the possibility of statistical error and thus a risk of rewarding the country too little or too much. This risk depends on the initial rate of infection and on HIV prevention success and can be reduced by either increasing the survey sample size or increasing the interval between surveys. By negotiating in advance the choice of one of these measurement approaches and one of a menu of payout functions, the donor and recipient agree on the recipient’s incentive structure with respect to the magnitude and precision of the estimated reduction in the rate of new infection.