It seems to be the de-facto method; and I see how we get it from log-normal assumption.

However volatility scaling seems to be way more sensitive to $\tau$ than mean scaling -- as in two ~ 2 times (be it using simple or log returns) in my sample.

And it indeed it is: \begin{equation} \frac{d}{d\tau}\big(\sigma_{1}(\tau)\big) > \frac{d}{d\tau}\big(\mu_{1}(\tau) \big)\hspace{2.5 mm} \mid \hspace{2.5 mm} \tau > 0 \end{equation}

But is this good for practical use?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.