The realized volatility in the forecast dialog


I have an EGARCH Model and I would like for forecast the local and term-structure volatility. In the forecast dialog, I have the option to provide a realized volatility for most recent dates. Should I leave it blank (default), use EWMA, historical volatility, etc. ?

Which one's more appropriate for the financial historical data I used?


1 comment
  • The realized volatility serves as the initial conditions for the forecast. If the EWMA happens to be closer from the GARCH/EGARCH long-run volatility, it will converges faster.

    The question falls not on the convergence speed, but on your belief that EWMA or the GARCH realized volatility are representative to market, as it drives the forecast. This is why there are numerous methods to compute realized volatility (e.g. Yang-Zand, EWMA, HI-LO, moving standard deviation, implied volatility, etc.), as it affects your forecast.

    I suggest, you start with "blanks" (i.e. EGARCH vols), then try with EWMA and compare the two. Bear in mind EWMA is a constrained form of GARCH itself, but it happens to be used widely in industry to measure volatility

    Comment actions Permalink

Please sign in to leave a comment.

Didn't find what you were looking for?

New post