jmaih
commented
1 year ago Hi Jose,
This is a big fat nothing burger :)
Just index the parameters with time and RISE will take care of the rest.
RISE solves the generic problem Ef(x{t+1},x{t},x{t-1},e{t},theta{t},theta{t+1})=0
where x{t} is the vector of endogenous variables, e{t} the vector of shocks, theta{t} the vector of parameters.
Cheers,
J.
On Wed, Sep 21, 2022 at 6:13 PM Jose Andrés Ortega @.***> wrote:
Dear Mr. Maih,
I am currently working with a model which has just one regime-dependent parameter (say $\lambda(s_t)$), but this parameter appears both depending on $st$ and on $s{t+1}$ in the equations of the model.
I wanted to ask you if there is a recommended way to deal with future regimes appearing in the model. My current approach is to define a compund regime: \begin{equation} s_t^ = \begin{cases} 1, \quad \text{if } s_{t+1} = 1, st=1 \ 1, \quad \text{if } s{t+1} = 2, st=1 \ 1, \quad \text{if } s{t+1} = 1, st=2 \ 1, \quad \text{if } s{t+1} = 2, st=2 \end{cases} \end{equation} and transition matrix given by: \begin{equation} p{11} & p{12} & 0 & 0 \ 0 & 0 & p{21} & p{22} \ p{11} & p{12} & 0 & 0 \ 0 & 0 & p{21} & p{22} \end{equation*} where $p{ij}$ are the original transition probabilities of $s_t$. Is this the right way or would you recommend using a different approach?
Thank yoy very much in advance, Jose
— Reply to this email directly, view it on GitHub https://github.com/jmaih/RISE_toolbox/issues/173, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATKBT5MF6P766KHFWJP5HLV7MX4BANCNFSM6AAAAAAQSHMPKU . You are receiving this because you are subscribed to this thread.Message ID: @.***>
jortegab
Dear Mr. Maih,
I am currently working with a model which has just one regime-dependent parameter - say $\lambda(s_t)$ -, but this parameter appears both depending on $st$ and on $\mathbb{E} s{t+1}$ in the equations of the model.
I wanted to ask you if there is a recommended way to deal with future regimes appearing in the model. My current approach is to define a compund regime:
$$ \begin{equation} st^* = \begin{cases} 1, \quad \text{if } \mathbb{E} s{t+1} = 1, st=1 \ 2, \quad \text{if } \mathbb{E} s{t+1} = 2, st=1 \ 3, \quad \text{if } \mathbb{E} s{t+1} = 1, st=2 \ 4, \quad \text{if } \mathbb{E} s{t+1} = 2, s_t=2 \end{cases} \end{equation} $$
and transition matrix given by:
$$ \begin{equation} \begin{bmatrix} p{11} & p{12} & 0 & 0 \ 0 & 0 & p{21} & p{22} \ p{11} & p{12} & 0 & 0 \ 0 & 0 & p{21} & p{22} \end{bmatrix} \end{equation} $$
where $p_{ij}$ are the original transition probabilities of $s_t$. Is this the right way or would you recommend using a different approach?
Thank yoy very much in advance, Jose