"A process of deliberation, in which policymakers exchange information prior to formal voting procedures, precedes almost  very collective decision. Yet, beyond scarce evidence coming from field and laboratory experiments, few studies have analyzed the role played by sequential deliberation in policy-relevant decision-making bodies. To fill this gap, I estimate an empirical model of policy-making that incorporates social learning via deliberation. In the model, committee members speak in sequence, allowing them to weight their own information and biases against recommendations made by others. The empirical model is structurally estimated using historical transcripts from the Federal Open Market Committee (FOMC), which is the body in charge of implementing monetary policy in the United States. I find the process of deliberation significantly changes members’ behavior: a typical FOMC member would modify her policy recommendation in 36% of the meetings after listening to previous speakers, with respect to the scenario where members exclusively follow their private information. Counterfactual simulations show modest gains of modifying the order of speech on the quality of the committee’s policy choices. Incorporating sequential learning explains the pattern of individual recommendations and collective choices extremely well and improves the fit over behavioral models that ignore deliberation."