The health financing system in the Democratic Republic of Congo (DRC) presents an extreme example of low government investment, high dependency on user fees and poor harmonization across donors. Within this context, performance-based financing mechanisms are being implemented by various donors in the expectation that they will improve health worker motivation and service delivery performance. Drawing on qualitative and quantitative data at different levels of the health system, this study focuses on one such programme in Katanga Province, which combines paying for performance (P4P) with a reduction in fees to users. Despite adding considerably to facility resources (providing the majority of the resources in the case study facilities), there was no evidence of benefits in terms of any of the service inputs, processes or outputs measured. The findings suggest that the positive effects on health worker motivation cannot be taken as a given, particularly, when staff are often expected to increase their workload to achieve the performance objectives and when another source of income, the income from user fees, may be reduced due to a fall in the prices of services. Moreover, in a context where health workers were already almost entirely dependent on users for their remuneration before the donor programme was introduced, the incentive effects of a performance contract may be muted. In addition, other income sources have particular value for staff, it seems-even though salaries and government allowances were low, and frequently delayed, health workers were highly dissatisfied at not receiving them. Salaries were seen as a more assured and long-term source of funding and an important recognition of their role as agents of the state. The authors conclude that while there may be a role for P4P in fragile contexts such as the DRC, to be effective it needs to be rooted in wider financing and human resource policy reforms.

Rwanda demonstrates how a process of community consultation and participation is able to identify and rank community members according to 'social poverty', drawing on the Ubudehe tradition which is considered a strength of Rwanda's social fabric. However, with the Ubudehe categorisation now the basis for determining eligibility to a range of social benefits, the process has come under some strain. This article highlights two issues related to targeting and graduation: (1) the difficulty in identifying the poor/non-poor and ranking the population using community participatory techniques; and (2) the sensitivity of eligibility criteria and graduation thresholds to different targeting modalities. Our primary interest is to establish whether improvements for identifying the poor and non-poor can be made without undermining community ownership and what these improvements would look like. This will be useful for policymakers in Rwanda as the new five-year development strategy places importance on graduating households out of extreme poverty.

No abstract