Question

Why do researchers prefer GARCH(1,1) models to pure ARCH(q) models?

Answer 1

The GARCH model i.e. (generalized autoregressive conditional heteroskedasticity) has three parameters only, which allows to influence the conditional variance by an infinite number of squared roots. This unique characteristic enables GARCH to be better than ARCH model. So, GARCH is better for modelling time series data when the data exhibits heteroskedasticity and volatility clustering.

But in some situations, some aspects of the GARCH model can be modified or improved to detect more precisely. For example, a standard model sometimes fails to capture leverage effects which can be observed in the financial time series. i.e. the good and the bad information’s has the same effect and it won’t get detected.

Hence , we can tell that the GARCH model is slightly better than the arch model.