CMStatistics 2018: Start Registration
View Submission - CFE
A1636
Title: Estimation of permanent and temporary shocks in a factor model framework Authors:  Carlos Montes-Galdon - European Central Bank (Germany) [presenting]
Abstract: An efficient algorithm is presented to recover permanent and temporary shocks from a factor model. There are several contributions. First, the algorithm makes use of a new way to introduce sign restrictions in a reduced form model to identify structural shocks. Moreover, via a state space model with restrictions on the persistence of the shocks, we can disentangle the impact of, for example, permanent productivity shocks or temporary ones. Also on the technical side, it is well known that in factor models the estimates of the loadings have very slow convergence properties. This problem is also present (and worsened) when we introduce sign restrictions. We propose an extension of very recent econometric techniques to overcome this problem. The final algorithm is extremely efficient and reliable. Then, based on restrictions that come from a DSGE model, looking at both changes in the steady state and the cyclical components, we estimate a factor model under the presence of both permanent and temporary developments. One of the main findings is that the great recession was mostly driven by permanent structural shocks (mainly investment specific shocks) which could explain the slow recovery.