<i>Schumpeter Lecture 2025</i> : The New Causal Macroeconomics of Surveys and Experiments

Abstract We discuss how randomized control trials (RCTs) can be used to study the causal effects of expectations on the decisions of households, firms, and other economic agents. Specifically, information provision in RCTs can create exogenous variation in the beliefs of survey participants and thus can address a key identification challenge that has plagued research efforts focused on understanding the role of expectations in shaping economic decisions. When linked to either external information on their actions or subsequent survey waves that measure their ex-post decisions, RCTs can provide clear causal identification of the passthrough from expectations to decisions. We review recent evidence using this strategy and discuss potential challenges associated with this approach.