[A,B,C,D,F,G,H,J,List,Nf] = system(M)
M [ model ] - Model object whose system matrices will be returned.A, B, C, D, F, G, H ,J [ numeric ] - Matrices describing the unsolved system, see Description.
List [ cell ] - Lists of measurement variables, transition variables includint their auxiliary lags and leads, and shocks as they appear in the rows and columns of the system matrices.
Nf [ numeric ] - Number of non-predetermined (forward-looking) transition variables (multiplied by the first Nf columns of matrices A and B).
'linear=' [ @auto | true | false ] - Compute the model using a linear approach, i.e. differentiating around zero and not the currently assigned steady state.
'select=' [ true | false ] - Automatically detect which equations need to be re-differentiated based on parameter changes from the last time the system matrices were calculated.
'sparse=' [ true | false ] - Return matrices A, B, D, F, G, and J as sparse matrices; can be set to true only in models with one parameterization.
The system before the model is solved has the following form:
A E[xf;xb] + B [xf(-1);xb(-1)] + C + D e = 0
F y + G xb + H + J e = 0
where E is a conditional expectations operator, xf is a vector of non-predetermined (forward-looking) transition variables, xb is a vector of predetermined (backward-looking) transition variables, y is a vector of measurement variables, and e is a vector of transition and measurement shocks.