Factor Augmented Autoregressive Distributed Lag Models

This paper proposes a factor augmented autoregressive distributed lag (FADL) framework for analyzing the dynamic effects of common and idiosyncratic shocks. We first estimate the common shocks from a large panel of data with a strong factor structure. Impulse responses are then obtained from an autoregression, augmented with a distributed lag of the estimated common shocks. The approach has three distinctive features. First, identification restrictions, especially those based on recursive or block recursive ordering, are very easy to impose. Second, the dynamic response to the common shocks can be constructed for variables not necessarily in the panel. Third, the restrictions imposed by the factor model can be tested. The relation to other identification schemes used in the FAVAR literature is discussed. The methodology is used to study the effects of monetary policy and news shocks. JEL Classification: C32, E17