Should I perform a ARCH & GARCH analysis ?

[Seeking_Knowledge]

Hello everybody

I'm new to this wonderful program and would like to learn more, by performing a logical list of tests for a problem statement.

1) For example in this relation, let's say I want to test whether the price development of copper has an effect on a certain stock.

1.1) Problem statement, does the price volatility of copper have an effect on stock A ?

I've been reading about ARCH, and various GARCH tests + extensions, however it seems to me that these are forecasting tools, and not effect testing tools.

I hope someone experienced in these matters, could perhaps enlighten me a bit.

Looking forward to some answers :-)

[BobJ]

GARCH and ARCH models are designed to capture certain features of asset price volatility. So, if I understand your question correctly, you may use a multivariate GARCH model to examine whether copper price volatility and the volatility of stock price you're interested affect each other (this is called a volatility spillover effect). Note that the multivariate GARCH is about second moments (volatilities/variances). It would be a different story if your interest is to examine, say, the effect of copper price changes (not volatility) on the volatility of the stock you're interested interested in. You can simply do that by adding the change in copper prices as an explanatory variable in the mean or variance equation (or both) of a GARCH model (this can be easily done in Eviews).

Hope this helps.

Bob