A revolution in economics? It’s just getting started…

Abstract

Abstract We have each experienced thrills and pain while supporting the mission of the Abdul Latif Jameel Poverty Action Lab, which facilitated many of the experiments described in the 2019 Nobel Prize citation. J-PAL in many ways seeks to fulfill what Angrist and Pischke called the “Credibility Revolution in Empirical Economics.” Even though (or perhaps because) we have conducted many RCTs, we share many of the concerns that critics have highlighted: high cost, long time lags, and limits to generalizability. Yet, we are quite optimistic that the impact and reach of experimental work in economics and policy will only grow. We see two complementary developments which will make RCTs cheaper, faster, larger, and ultimately substantially more insightful. First, a new research literature seeks to improve the design of experiments, and what we can learn from them, through improved methodologies, meta-analyses, and improved understanding of heterogeneity. Second, the rise of administrative data rapidly opens new frontiers of investigation, in particular the possibility of ‘closed-loop’ data environments, in which interventions can be delivered and evaluated digitally, often on very large samples, and often iteratively.