ARMA-GARCH estimation

[Christian Deguzman]

We're going to estimate an ARMAX-GARCH model using the log series of PSEi market returns to assess the impact of natural disasters on market returns and volatility. We are trying to modify the model from Nguyen and Chaiechi’s (2021) “The Effects of Natural Disasters on Stock Market Return and Volatility in Hong Kong”

To account for the occurrence and location of the disaster, we are going to use regional dummy binary variables as the exogenous variables in the mean equation of the ARMAX-GARCH model. We are looking for ways to estimate the ARMAX and GARCH components simultaneously using Stata. However we only found out in Stata forums and other reference materials how to employ this using the "rugarch" package in R.

Do you have any tips on how to run this model in Stata?

Once again, we truly appreciate your time and input.Thank you!

[Dario Maimone Ansaldo Patti]

I am not sure I picked up your question correctly, but if so you want to estimate the volatility of a series specifying an arms structure for the error term in the mean equation and assuming always in the mean equation that your returns depends on a regressor x. If so why don't you try:

arch returns x, arch(1) garch(1) ar(1) ma(1)