Hi,I'm trying to estimate the AGDCC-model of Cappiello et al (2006) for a multi-asset portfolio. My results (the agdccparamters) seems quite strange and I wonder if I'm missing some constraints on the parameters. In their paper Cappiello et al mention just a constraint concerning the positive definiteness of the covariance matrix. I'm not sure if this is enough (not questioning their paper, just wondering if I'm overlooking sth.!) since their model is in spirit of an GJR-GARCH (and GJR imposes some constraints on the GARCH parameters). I already tried GJR-constraints (extended to the multivariate case), but results are still strange. Any comments would be helpful!Best regards,Martin

Did you backtest a general GJR-GARCH framework? Do you get good results for impulse response and forecasts?-ADCC(L,S,U)model: G = g1/2,A = a1/2,B = b1/2 you're also right about the positive definitiveness of the var-covar matrix