If the volatility forecast reached the long term Vol value at a certain step (in case of my data sample, @ step 15), would it be meaningless to forecast beyond that step (i.e., 16 + weeks)?
You are correct, the forecast will revert to the mean (or in this case long-run volatility). In few cases, the volatility forecast kind-of oscillates (looks like a sine/cosine wave) in the beginning before it converges to the log-run volatility.
To be on the safe side, I would forecast to all horizons of interest.
コメント
You are correct, the forecast will revert to the mean (or in this case long-run volatility). In few cases, the volatility forecast kind-of oscillates (looks like a sine/cosine wave) in the beginning before it converges to the log-run volatility.
To be on the safe side, I would forecast to all horizons of interest.
サインインしてコメントを残してください。