egarch

[h-godarzi]

hi
i am working on egarch and tgarch model.as a decision rule when gamma is negative and significant in egarch model the model detect the leverage effect.in tgarch when gamma is posetive and significant the model is adequatly detect the leverage effect.i was wondering if i could ask you to describe why in both model it should be negative and posetive respectively?
thanks
hogo

[MartinFalch]

Try inserting a negative value of the previous epsilon - the coefficient that is multiplied with this should have the appropriate value, in order for it to generate an INCREASE in the current conditional volatility. This is due to Black finding a so called "leverage effect" in financial time series (a stylized fact, so to speak), meaning that negative returns tend to increase volatility in subsequent periods more than positive returns do. Hope that answers your question :)

[h-godarzi]

dear sir
please describe more.
thanking you
hogo