A model for policy interest rates

Abstract This paper introduces a model that addresses the key worldwide features of modern monetary policy making: discreteness interest rates both in magnitude and timing, preponderance status quo decisions, inertia regime switching. We capture them by developing new dynamic ordered-choice with switching among three latent regimes (easing, neutral tightening). The simulations an application to federal funds rate target demonstrate ignoring these leads biased estimates, worse in- out-of-sample predictions, qualitatively different inference. Using all Federal Open Market Committee’s (FOMC) decisions made at scheduled unscheduled meetings as sample observations, we Reserve’s response real-time data available right before each meeting. model, fitted for Greenspan’s tenure, detects oscillating switches regimes, identifies types correctly predicts out 90% next 111 FOMC on rate, clearly outperforms linear models (including Taylor rule), conventional ordered probit other discrete-choice from literature.