I estimate a model of policymaking that incorporates social learning via deliberation. In the model, committee members speak in sequence, balancing their private information and biases against recommendations made by others. I quantify the extent of social learning in the Federal Open Market Committee. I find that the process of deliberation significantly changes individual behavior and aggregate monetary policy: the average committee member changes her recommendation in approximately 36 percent of the meetings after listening to previous speakers. Counterfactual simulations show that the observed deliberation order within the FOMC has been effective at reducing the probability of mistakes.
Download Paper (Last Version: September, 2018
Download Appendix