Structural vector autoregressions with smooth transition in variances

Abstract

In structural vector autoregressive analysis identifying the shocks of interest via heteroskedasticity has become a standard tool. Unfortunately, the approaches currently used for modeling heteroskedasticity all have drawbacks. For instance, assuming known dates for variance changes is often unrealistic while more flexible models based on GARCH or Markov switching residuals are difficult to handle from a statistical and computational point of view. Therefore we propose a model based on a smooth change in variance that is flexible as well as relatively easy to estimate and illustrate its use by analysis of the interaction between monetary policy and the stock market based on a five-dimensional system of U.S. variables. For the benchmark setup it is found that previously used conventional identification schemes in this context are rejected by the data if heteroskedasticity is allowed for. We also illustrate the implications of using different transition variables and varying the sample period.