Measuring the Effects of Monetary Policy: A DSGE-DFM Approach

I propose new method of measuring the effect of monetary policy on a large number of macroeconomic series by combining dynamic stochastic general equilibrium (DSGE) with a dynamic factor model (DFM) in a data-rich environment including a broad range of useful information for both central banks and the private sector. This method can resolve two problems inherent in the FAVAR approach proposed by Bernanke, Boivin and Eliasz (2005), yet can enjoy its benefits as well. Other positive aspects of this method are to split off components of model concepts (or common factors) and measurement errors from all observable variables, and to identify structural shocks, including monetary policy shocks, from the point of view of DSGE based on a microeconomic foundation with rational expectations. This new method calculates historical decomposition, which provides useful information for policy analysis for researchers and policymakers, as well as theoretically reasonable impulse response functions and variance decompositions for all observable variables of large panel data. Through this method I will provide empirical illustrations of 55 Japanese macroeconomic series during the late 1980s and the 1990s using a Smets-Wouter (2003, 2007) type medium size DSGE model.