Re: [R] alpha_1 + beta_1 >1 in GARCH(1,1)

This would have been a more appropriate question on r-sig-finance. The short answer is: no, you can't trust the coefficients. You don't say how much data you have, but this situation is common when you don't have much data (meaning fewer than 2000 daily observations). The sum of those two param

[R] alpha_1 + beta_1 >1 in GARCH(1,1)

Hi, as i suppose to know in a stationary GARCH(1,1) model the sum of alpha and beta has to be smaller than 1. But if i use the garchfit() function from the package fGarch for my timeseries the sum is bigger than 1. The adf.test tells me a p-value smaller than 0.01 instead. What does this mean for