1
$\begingroup$

I'd like to see if the oil price trend (as measured by WTI or Brent indexes) can help to predict the volatility of a stock.

I am using a GARCH approach to model volatility of the FTSE.

Now, if I want to know if the oil price volatility has an impact on FTSE's volatilty, can I use the returns on oil price as a variance regressor in my GARCH model?

$\endgroup$
  • 1
    $\begingroup$ You would use squared returns, right? Because raw returns could produce negative values (fitted or predicted) of conditional variance, which you do not want. $\endgroup$ – Richard Hardy Sep 20 '16 at 16:37
  • $\begingroup$ James, your first paragraph does not quite fit with the last two paragraphs. Could you clarify your question? $\endgroup$ – Richard Hardy Sep 20 '16 at 18:57
  • $\begingroup$ @RichardHardy As far as you first comment is concerned, I thought that the positive and negative shocks have the same impact on conditional volatility in a linear GARCH model. Doesn't this mean that there is no need to square the returns? Anyway, what I want to see is wheter the oil price can help to predict the conditional volatility of a stock. Actually I don't know if I should use oil price, its returns, or its [(un)conditional] volatility. Could you help me? $\endgroup$ – James Sep 20 '16 at 19:50
  • $\begingroup$ There is a fundamental problem in having negative fitted/predicted volatility, and therefore you have to make the regressors in the cond. variance equation nonnegative. Squaring the returns is one way of doing so. Taking absolute values or making other transformations could be considered as well. In general, your question seems to be less about statistics and more about suject matter (finance). I am not sure if this is the right place to look for expertise in finance. $\endgroup$ – Richard Hardy Sep 20 '16 at 19:58
  • 1
    $\begingroup$ I don't really do personal consulting, so it's better if you formulate your problem as a question that can be answered here. (Your problem might or might not be a good fit for Cross Validated.) $\endgroup$ – Richard Hardy Sep 25 '16 at 16:57

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.