Does Involving Local NGOs Improve Outcomes?

In an effort to enhance the delivery of public services in areas where resources are scarce, governments and development agencies often seek to involve existing local nongovernmental organizations (NGOs) in implementing programs. The development literature has a limited understanding of the challenges of involving local nonstate actors, however, and many questions remain about when, where, and how to do so. Discussions often highlight the potential benefits from leveraging local capacity, including expanding the reach of basic services and improving quality. But collaboration between state and nonstate actors also has potential pitfalls, including mistrustful relationships between governments and their nonstate counterparts. So, what might distinguish a successful government-NGO collaboration from a troubled one? A recent paper by Das, Friedman, and Kandpal takes advantage of an experimental evaluation of a malaria control intervention in India to examine whether involving local NGOs in program implementation has a significant influence on the individual health outcomes targeted by the intervention. The paper finds that involving local NGOs in program implementation can indeed lead to positive outcomes, though success likely depends on the quality of the NGOs. And in certain underresourced areas there may be little choice in the selection of qualified local NGOs, thus diminishing the likelihood of success. The malaria control intervention studied in the paper was conducted simultaneously by three NGOs in two endemic districts in the state of Orissa. It turns out that the effect of NGO involvement on patterns in the use of mosquito nets and the seeking of medical care for fever—the main targeted outcomes of interest—differs significantly between the two districts. Using either simple regression decomposition or propensity weighting, the paper examines three potential causes for the differences in outcomes: differences between the two districts in the characteristics of the population, differences in the characteristics of health workers, and differences in the characteristics of the NGOs. The analysis shows that the differences in observed outcomes are due mainly to the first and third factors as well as to their possible interaction. These findings raise important questions about the ability of the health system to benefit from limited nonstate capacity in underresourced areas. They also raise questions about the extent to which evidence taken from interventions implemented by heterogeneous actors can be generalized. The paper thus contributes to the small but important literature on the external validity of small-scale policy evaluations, especially those conducted in the nonstate sector. Contextual knowledge and implementation capacity are surely key determinants of the effectiveness of development interventions, and there are many potential nonstate actors with varying degrees of capability. Without a better understanding of the context and of the capabilities of the study participants, as well as standardized measures of implementation capacity, it will be difficult to generalize findings from any relatively small-scale evaluation. More narrowly, on the direct question of leveraging preexisting local NGOs in the delivery of public services, underresourced areas such as those studied by the paper often have underresourced local nonstate capacity. Involving nonstate actors therefore will not necessarily result in better outcomes. This lack of local nonstate capacity in many underserved and underresourced areas speaks to the complex interactions between poverty and local characteristics. In highly impoverished areas a combination of locally targeted investments and antipoverty policies may be necessary to improve public service delivery and, consequently, targeted welfare outcomes.