Paper short abstract:

Health care funding has increased considerably in recent years. This paper tests six hypotheses on possible determinants to allocate health assistance to a country. Health performance matters but bilateral relations, institutional quality, donor’s preferences and other donors’ behavior matter too.

Paper long abstract:

This paper analyzes the targeting of development assistance for health across countries in a multivariate regression framework, based on data from 22 bilateral donors to 160 recipients between 1990 and 2007. The results show that health performance is considered when allocating health assistance. HIV prevalence significantly increases aid receipts, while under-five mortality and maternal mortality do not influence allocation decisions for health by the average donor. Yet, health performance is not the only aspect that matters. Bilateral relations influence the allocation process. In fact, economic links as trade relations, cultural proximity as the same dominant language or religion, or the common colonial past influence these decisions significantly. Yet, political ties or geographic proximity are insignificant factors for the decision-making process. In the discussion about aid effectiveness, a major issue is the sensitivity of donors as regards potential problems related to a weak institutional environment such as corruption. The results show that more corrupt countries do receive less health assistance. The level of effort put into national health by the recipient country is insignificant. Neither the level of public health expenditures nor the immunization coverage has any statistical effect. The available evidence suggests that the preference of the donor for health on the national political agenda increases the provision of health assistance. The average donor does not decide independently but acts strategically. Particularly the donor with a large budget is inclined to complement allocations made by the US and by multilateral donors. However, the effect is not substantial.