Reducing dimensions in large time-varying parameter VAR models

Event Date: 9 December 2015

Time: 4.15pm

Location:  Strathclyde Business School, Cathedral Wing (CW) 404b

Speaker: Rodney Strachan, University of Queensland 

Abstract: This paper proposes a new approach to estimating high dimensional time varying parameter vector autoregressive models (TVP-VARs). Such models are rarely used with more than 4-5 variables. However recent work has shown the advantages of modelling VARs with large numbers of variables and interest has naturally increased in modelling large dimensional TVP-VARs.

We propose a specification that uses strong and perfect correlations in a factor-like structure to estimate a TVP-VAR for 15 variables. We show clear empirical evidence in favour of our model and improvements in estimates of impulse responses.

Published: 8 December 2015