Skip to content
All library documents

Specifying EViews GJR-GARCH Models for Volatility Spillovers

Article Quant Q&A · Author: Lilly-Ann
Read the original

Summary

The question asks how to specify a multivariate GJR or threshold GARCH model in EViews to test whether volatility in two stock markets spills over into a third. The response explains that EViews estimates a GJR-GARCH specification through its GARCH/TARCH option when a threshold order is set. It also notes that EViews calls this option TARCH even though its variance equation follows the GJR form, rather than the original TARCH formulation based on conditional standard deviation.

The answer points to the model dialog’s distribution selector for choosing a different error distribution. It does not provide the actual variance-regressor setup needed to encode cross-market spillovers, nor does it demonstrate estimation or testing. The guidance is therefore a brief clarification of model naming and software options, not a complete recipe for fitting or validating a multivariate spillover model.

Key ideas

  • In EViews, selecting GARCH/TARCH and setting a threshold order produces a GJR-GARCH specification.
  • EViews’ TARCH label refers to a variance equation matching GJR-GARCH rather than the original standard-deviation formulation.
  • The ARCH-GARCH dialog allows the user to select a different error distribution.
  • The answer does not explain how to specify cross-market variance regressors or test spillovers.

Tags

EquitiesVolatilityStatistics
Full text
# How to estimate a multivariate GJR or TARCH model in Eviews?


# How to estimate a multivariate GJR or TARCH model in Eviews?












How do I specify the GARCH/TARCH equation in Eviews 6 in the variance regressors frame, if I want to find out whether there are volatilty spillovers from stock markets A and B to stock market C?

P.S. I know I have to set threshold order to 1 ;)

## Answer by Aarthi (score 2)

https://quant.stackexchange.com/a/1474

Would this work?

> EViews actually estimates the GJR-GARCH model when you select the GARCH/TARCH option and specify a threshold order. Original TARCH model works on conditional standard deviation. However, as you can verify it from the user's guide, EViews' TARCH model uses the same specification as GJR model does. Given that, in order to employ a different error distribution, all you have to do is select from the drop down menu in the ARCH-GARCH models dialog box.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.