2
$\begingroup$

Lets say I have a GARCH(1,1) model,

First, I model the conditional MEAN,

$$Y_t=\delta+\beta Y_{t-1}+\varepsilon_t$$

NextI gather the residuals $\varepsilon_t$ and model the conditional variance,

$$h_t=\omega + \alpha_i\varepsilon_{t-1}^2+\beta_ih_{t-1}$$

I need to get the standard residuals,

my attempt at this is,

$$U_t=\varepsilon_t/\sqrt{h_{t}}$$

Is this correct?

I am struggling to find this in any textbooks or papers so a reference would be appreciated if possible!

$\endgroup$
2
  • $\begingroup$ Thanks Richard, yes! $\endgroup$ Commented Jun 10, 2019 at 12:06
  • $\begingroup$ Indeed, when we talk about standardization, we are talking about "uniforming" the standard deviations. $\sqrt{r_t}$ is as good of an approximation for it as we can get (unless the process actually follows an exact GARCH, which never happens) $\endgroup$
    – David
    Commented Jun 10, 2019 at 14:16

1 Answer 1

1
$\begingroup$

This is correct.

References:

$\endgroup$
2
  • $\begingroup$ Thanks very much for your help! $\endgroup$ Commented Jun 10, 2019 at 12:07
  • 1
    $\begingroup$ @FrancisOrigi, You are welcome! $\endgroup$ Commented Jun 10, 2019 at 12:07

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.