Page 1 of 1

GARCH (1,1)

Posted: Fri Jun 19, 2015 6:17 am
by RohanDholakia
Hi,

I am currently doing my dissertation on forecasting volatility but I am having some basic trouble with the GARCH(1,1) model.

I am using non-overlapping data for both the implied volatility I have obtained and the realised volatility based on the Wednesday following the 3rd Friday of each Monday until the 3rd Friday of the following month.

My data period is from 1/1/04 to 1/10/14 roughly so I have 130 monthly observations based on there own independent window.

For my GARCH model I am using up to 1/10/11 for my sample and forecasting the last 36 months.
The part I am stuck on is:
Would I be using the daily returns to forecast the daily volatility for the one day ahead and then re-run taking away the first data point and adding a further data point for return? then average these forecasts for each month window I have to find the forecast for each months volatility?

sorry it is a very basic question but I am confused.

Many thanks,
Rohan