Structural Factor-Augmented VAR (SFAVAR) and the Effects of Monetary Policy

Factor-augmented VARs (FAVARs) have combined standard VARs with factor analysis to exploit large data sets in the study of monetary policy. FAVARs enjoy a number of advantages over VARs: they allow a better identification of the monetary policy shock; they can avoid the use of a single variable to proxy theoretical constructs, such as the output gap; they allow researchers to compute impulse responses for hundreds of variables. Their shortcoming, however, is that the factors are not identified and, therefore, lack any economic interpretation. This paper seeks to provide an interpretation to the factors. We propose a novel Structural Factor-Augmented VAR (SFAVAR) model, where the factors have a clear meaning: 'Real Activity' factor, 'Price Pressures' factor, 'Financial Market' factor, 'Credit Conditions' factor, 'Expectations' factor, etc. The paper employs a Bayesian approach to extract the factors and jointly estimate the model. This framework is then suited to study the effects on a wide range of macroeconomic variables of monetary policy and non-policy shocks.