Policy-making in developing countries: from prediction to planning

The claim that randomized controlled trials (RCTs) can improve policy-making in developing countries seemed to be severely challenged by the widespread belief that RCTs lack external validity. This belief led to the articulation of strategies to alleviate this problem which developed under the assumption that policy-making in developing countries can be best understood as a problem of prediction. The causal structures that characterize developing contexts, however, render this framing inadequate. I propose to understand policy-making in developing contexts as a planning process. In this new framework, the relevance of several kinds of evidence is highlighted and the role of causal effects refocused. Interestingly, also the concern for the external validity of causal effects seems now less forceful than the critics of RCTs suggest.