同问

在拟合GARCH模型后,Q统计量始终显示存在高阶自回归(P值小)

不管怎样调整均值方程中的ARMA,均有上述问题,请教高手怎么样才能解决呢

ARMA is used to model the first moment and ARCH type model second moment. they are a bit different things.

不管怎样调整均值方程中的ARMA,均有上述问题 this should be natural there is genuine ARCH effect.

AS a matter of fact, people just demean the log returns and directly model the demeaned log returns with ARCH model without condsidering ARMA terms in the mean equation.

This is sensible since, in practice, for stock price, random walk model predicts better than ARMA model.