I try to fit a GARCH(1,1) to the log-returns of the german stock index (dax) data from 2014-06-03 till 2016-01-01 using ugarchfit from the ugarch. I found that there is a difference to the coefs calculated by using garch from tseries.
Why there is a difference and which one is the more precise?
Here is my code:
library("tseries")
library("rugarch")
# GARCH(1,1)-Specification using rugarch-package
specgarch0 <- ugarchspec(variance.model=list(model="sGARCH", garchOrder=c(1,1)), mean.model=list(armaOrder=c(0,0)), distribution="norm")
dax_mod0<- ugarchfit(data=dax_ret, spec=specgarch0)
show(dax_mod0)
coef(dax_mod0)
# GARCH(1,1)-Specification using garch(...) by tseries-Package
dmod<-garch(dax_ret)
summary(dmod)
dmod$coef
> dmod$coef
a0 a1 b1
9.763928e-06 1.001589e-01 8.496407e-01
> coef(dax_mod0)
mu omega alpha1 beta1
6.105980e-04 1.022263e-05 1.036797e-01 8.436064e-01
