Skip to content
All library documents

Adding Exogenous Variables to a GARCH Variance Equation

Article Quant Q&A · Author: Shikamekika SK
Read the original

Summary

The document asks whether MATLAB’s `garchset` and `garchfit` interfaces allow dummy or explanatory variables in the conditional variance equation, analogous to regressors in the mean equation. It seeks a way to specify such a model, rather than only adding regressors to the conditional mean.

The answer reports that, to the respondent’s knowledge, MATLAB had no public implementation for this feature and points to Ox’s G@rch and R’s rugarch packages as alternatives. It provides no example specification, package version, or comparison of capabilities, so the software guidance may be time-sensitive and should be checked against current documentation. The exchange explains the modeling distinction between including predictors in the mean and including them in the variance, but does not describe estimation or assess the statistical validity of any particular variance model.

Key ideas

  • The question distinguishes regressors in the conditional mean from predictors in the GARCH variance equation.
  • The respondent reports no known public MATLAB implementation for variance regressors.
  • Ox’s G@rch and R’s rugarch are named as possible package alternatives.
  • The exchange gives no implementation details, version checks, or model evaluation.

Tags

VolatilityStatisticsMachine learning
Full text
# Does Matlab support exogenous variables in GARCH models?


# Does Matlab support exogenous variables in GARCH models?












Is it possible to introduce dummy variables or explanatory variables in the GARCH variance equation (`garchset` and `garchfit`) in Matlab?

This is done in the mean (ARMAX) equation through the input `Regress` in `garchset`, but I have not seen how it can be done for the variance equation as well. So how can I find a closed form solution to do it? I will appreciate any advice on the question.

## Answer by Malick (score 1)

https://quant.stackexchange.com/a/24704

To the best of my knowledge there is no public implementation in matlab. However both R and Ox have some packages devoted to this end :

-Ox - see G@rch package

-R - see rugarch package

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.