Recursive GARCH Forecasting with Rolling Re-estimation in R

Summary

The document addresses how to produce recursive one-step-ahead GARCH forecasts when model coefficients are re-estimated at each step. It identifies an R package function designed to automate rolling estimation and forecasting, so each forecast can use parameters fitted with the data available at that point.

The answer points readers to an example and package documentation for implementation details. It does not compare software options, explain how to specify or diagnose a GARCH model, or provide forecast results. The material is therefore a concise pointer to a workflow rather than a complete guide to volatility modeling. Users still need to choose a suitable model, estimation window, and evaluation procedure for their data.

Key ideas

• Recursive one-step-ahead GARCH forecasts can re-estimate coefficients at each forecast step.
• The R package rugarch provides an automated rolling estimation and forecasting function.
• The answer directs users to package examples and its vignette for implementation guidance.
• It does not offer model-selection advice or evidence comparing forecast accuracy.

Tags

# Most GUI user friendly Time series Econometrics software for modelling and Forecasting GARCH models


# Most GUI user friendly Time series Econometrics software for modelling and Forecasting GARCH models












I think the question is simple enough. I have been using Eviews, but it is unable to do recursive one step ahead forecasting directly and requires me to use coding, which I'm not very good at.

I need to know which is the best software which will allow me to do recursive one step ahead forecasts, whereby coefficients are re-estimated at each step, and these are used for forecasts.

Additionally if anyone can also answer the following question linked, it would be highly appreciated.

Correct procedure for modelling GARCH for forecasting volatility of stock Index returns

## Answer by Richard Hardy (score 0, accepted)

https://quant.stackexchange.com/a/33454

I will answer the part of your question on

> ...software which will allow me to do recursive one step ahead forecasts, whereby coefficients are re-estimated at each step, and these are used for forecasts.

Automated recursive GARCH estimation and forecasting is available in R package `rugarch` via function ugarchroll (you will find an example in the link, too). See also the package vignette for details and examples.

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.