algebra at the beginning of conditional_forecast

In the algebraic description of conditional_forecast in section 4.20 of the manual, lagged uncontrolled endogenous variables are missing

changed milestone to %6.x

added documentation label

@MichelJuillard I don't really understand this section and I can't remember if I wrote this part. I will rewrite it.

assigned to @stepan-a

I think that the algebra should be something like that:

  \begin{itemize}
  \item<+-> Solution function
    \[
      y_t = A y_{t+1} + Bu_t
    \]
  \item<+-> Decompose in controlled and uncontrolled variables
    \[
      \begin{bmatrix} y_{u,t}\\y_{c,t}\end{bmatrix}
      =
      \begin{bmatrix}
        A_{uu} & A_{uc}\\
        A_{cu} & A_{cc}
      \end{bmatrix}
      \begin{bmatrix} y_{u,t-1}\\y_{c,t-1}\end{bmatrix}
      +
      \begin{bmatrix}
        B_{uu} & B_{uc}\\
        B_{cu} & B_{cc}
      \end{bmatrix}
      \begin{bmatrix} u_{u,t}\\u_{c,t}\end{bmatrix}
    \]
  \item Solution
    \begin{align*}
      u_{c,t} = B_{cc}^{-1}\left[y_{c,t} - A_{cu}y_{u,t-1} -
      A_{cc}y_{c,t-1} - B_{cu}u_{u,t}\right]\\
      y_{u,t} = A_{uu}y_{u,t-1} + A_{uc}y_{c,t-1} + B_{uu}u_{u,t} +
      B_{uc}u_{c,t}
    \end{align*}
  \end{itemize}
``

Yes. I am not sure that we need to define explicitly

u_{c,t}

, but we should definitely provide a correct partition of the system.

I wrote that part. The mistake in the current description is that it must be

T(𝑐𝑜𝑛𝑡𝑟_𝑣𝑎𝑟𝑠,:)𝑦_{𝑡−1}

on the right. But a full description may be good. Note that it needs to be ported to mcforecast3.m as well

closed with commit c9825c80

mentioned in commit fb3dbb75

mentioned in commit c9825c80