I am doing some research on MS GARCH, specifically the model proposed by Haas et al., however I am really stuck on the moment conditions derived in the appendix of the paper.

  • Right before equation ($34$) I am unsure why do we have to make use of the iid random variables. Also I am not quite following the reasoning how equations $34$ are derived, and thus how we can write the conditional expectation as equation ($35$). I know that we can relate $$ \mathbb{E}[\boldsymbol{\sigma}_n \mid \mathcal{F}_{n-\tau-1},\boldsymbol{\pi}_{n-\tau-1}]=(\boldsymbol{1}_k' \otimes \boldsymbol{I}_k)\boldsymbol{M}^\tau \boldsymbol{\Pi}_1+(\boldsymbol{1}_k' \otimes \boldsymbol{I}_k)\sum\limits_{i=0}^{\tau-1}\boldsymbol{M}^i\boldsymbol{A}_{n-1-i}^1$$ to $$\mathbb{E}(\boldsymbol{\sigma_{n+1}} \mid \Delta_{n-1}=j, \boldsymbol{\sigma}_n)=\boldsymbol{\alpha_0}+\sum\limits_{i=1}^k \boldsymbol{M}_{ji}\boldsymbol{\sigma_n}$$ however the conditional in these 2 instances is not the same.
  • Can someone direct me from where exactly was this equation extracted as I can't seem to recognize it from Karlin and Taylor. $$\pi_{y-\tau-i}=\pi_\infty + \delta^{\tau-1-i}(\pi_{t-\tau}-\pi_\infty)$$
  • How are we generalizing for k>2 using perron's formula?

Thank-you in advance!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.