Assessing Changes in U.S. Monetary Policy in a Regime-Switching Rational Expectations Model

We develop a new method for solving forward-looking rational expectations models with regime change and we apply it to the case of switches in monetary regime in U.S. data. Existing solutions to this problem are nonlinear since the parameters of a Markov switching model are functions of the state. We show how to enlarge the state space to obtain an equivalent problem that is linear in parameters. We show that the solution to the equivalent problem is a VAR with state dependent parameters that can be estimated by conventional methods and we present an application to US data