Suppose a policymaker is interested in the impact of an existing social program. Impact estimates using observational data suffer potential bias, while unbiased experimental estimates are often limited to other contexts. This creates a practical trade-off between internal and external validity for evidence-based policymaking. We explore this trade-off empirically for several common policies analyzed in development economics, including microcredit, migration, and education interventions. Based on mean-squared error, non-experimental evidence within context outperforms experimental evidence from another context. This advantage declines, but may not reverse, with experimental replication. We offer four reasons these findings are of general relevance to policy evaluation.

In 2016, the Liberian government delegated management of 93 randomly selected public schools to private providers. Providers received US$50 per pupil, on top of US$50 per pupil annual expenditure in control schools. After one academic year, students in outsourced schools scored 0.18 σ higher in English and mathematics. We do not find heterogeneity in learning gains or enrollment by student characteristics, but there is significant heterogeneity across providers. While outsourcing appears to be a cost-effective way to use new resources to improve test scores, some providers engaged in unforeseen and potentially harmful behavior, complicating any assessment of welfare gains. (JEL H41, I21, I28, O15)