Fitting GARCH Models with Exogenous Variables in Python

Summary

The document concerns fitting a time series using exogenous predictors while allowing its volatility to follow a GARCH process. The question notes that the author could not find Python support for multiple predictors and wonders whether a package they found was limited to one. A response points readers to the ARCH package, stating that it now includes the capability to fit models with exogenous variables alongside volatility specifications such as ARCH, GARCH, TARCH, EGARCH, and EWMA/RiskMetrics.

This is a brief pointer to software availability rather than a modeling guide. It gives no example, model equations, estimation details, comparison of implementations, or empirical results. It therefore does not establish how to specify multiple predictors, choose a volatility model, or assess fit; those details would need to be checked in the package documentation and relevant statistical references.

Key ideas

• The question is how to include exogenous predictors while modeling conditional volatility with GARCH.
• The response says the ARCH package supports this capability.
• The document lists several volatility model families but does not explain their specification or use.
• The software pointer provides no fitting example or evidence about model performance.

Tags

# How to fit exogenous + GARCH Model In Python?


# How to fit exogenous + GARCH Model In Python?












I am studying a textbook of statistics / econometrics, using Python for my computational needs. I have encountered GARCH models and my understanding is that this is a commonly used model. In an exercise, I need to fit a time series to some exogenous variables, and allow for GARCH effects. I looked but found no package in Python to do it. I found this but I think it only supports 1 exogenous variable - I have a bunch of them. This surprises me because I thought this would be something that some quants do every other day... Have I been looking in the wrong places?

Thank you very much

## Answer by P RAY (score 2)

https://quant.stackexchange.com/a/60305

It is an old thread. Just pointing out that capability is available in ARCH package now for the benefit of future readers. https://pypi.org/project/arch/

Volatility models ARCH GARCH TARCH EGARCH EWMA/RiskMetrics

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.