i am writing my thesis, where I am estimating different GARCH models with STATA 9.1

i already did the estimation, for different assets log.returns, but i am not sure, if I used the proper STATA command for the models.

sometimes STATA is not able to estimate the parameters, i think because of some stability problems or something... is that normal?

maybe you can check?

all the estimations were done with 1000 data points.

1. i did the estimation for the option pricing model of DUAN. here is the process (logreturn), that needed to be estimated:

View attachment 355

I estimated the model for m=1, n=1, so the parameters that needed to be estimated are lambda, alpha0, alpha1 and beta1.

i used the STATA command:

"arch logreturn, arch(1/1) garch(1/1) archm archmexp(X^0.5)"

2. i did the estimation for the option pricing model of HESTON/NANDI. here is the process (logreturn), that needed to be estimated:

View attachment 356

I estimated the model for m=1, n=1, so the parameters that needed to be estimated are lambda, alpha0, alpha1, beta1 and tau.

i used the STATA command:

"arch logreturn, narch(1/1) garch(1/1) archm archmexp(X^0.5)"

3. i did the estimation for an Power ARCH option pricing model . here is the process (logreturn), that needed to be estimated:

View attachment 357

I estimated the model for n=1, so the parameters that needed to be estimated are lambda, alpha0, alpha1 and delta.

i used the STATA command:

"arch logreturn, parch(1/1) archm archmexp(X^0.5)"

I hope, that information is clear enough, and that you can help me

thanks

ps. sorry for my english, but i am not from the US