Let $a_t=\sigma_t\epsilon_t$ and let $a_t$ be modeled e.g. by an ARCH(1) model, that is $a_t^2=\alpha_0+\alpha_1 a_{t-1}^2$.

Are $\sigma_t$ and $\epsilon_t$ independent?


1 Answer 1

  • $\sigma_t$ can be expressed as a deterministic function of an infinite amount of lags of $\varepsilon_t$ (by recursively substituting for lags of $a_{t-1}^2$ in the equation $a_t=\alpha_0+\alpha_1 a_{t-1}^2$).
  • $\varepsilon_t$ are i.i.d., thus independent.
  • A function of of some variables that are independent of another variable will be independent of that variable.

Hence, $\sigma_t$ and $\varepsilon_t$ are independent.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.