I was just have trouble interpreting and understanding the asymptotic distribution of the MLE of the parameters for and ARMA(1,1) model (and an ARMA(p,q) model in general).
It has been given that
$$\left(\begin{matrix} \hat\phi \\ \hat\theta \end{matrix}\right) \Rightarrow AN\left(\left(\begin{matrix} \phi \\ \theta \end{matrix}\right), \frac{\sigma^{2}}{n} \Gamma^{-1}\right)$$
Where
$$\Gamma^{-1}=\frac{1-\phi\theta}{(\phi-\theta)^{2}}\left(\begin{matrix} (1-\phi^{2})(1-\phi\theta) & -(1-\theta^{2})(1-\phi^{2}) \\ -(1-\theta^{2})(1-\phi^{2}) & (1-\theta^{2})(1-\phi\theta) \end{matrix}\right)$$
I can see that the means of the MLEs are $\phi$ and $\theta$. But what are the variances and how are they obtained? How would this generalise to an ARMA(p,q) model.
Any advice would be appreciated. Cheers