% Function: solab
% Purpose: Solves for the recursive representation of the stable solution to a system
% of linear difference equations.
% Inputs: Two square matrices a and b and a natural number nk
% a and b are the coefficient matrices of the difference equation
% a*x(t+1) = b*x(t)
% where x(t) is arranged so that the state variables come first, and
% nk is the number of state variables.
% Outputs: the decision rule f and the law of motion p. If we write
% x(t) = [k(t);u(t)] where k(t) contains precisely the state variables, then
% u(t) = f*k(t) and
% k(t+1) = p*k(t).
%
% Calls: reorder
function [f,p] = solab(a,b,nk);
[s,t,q,z] = qz(a,b); % upper triangular factorization of the matrix pencil b-za
[s,t,q,z] = reorder(s,t,q,z); % reordering of generalized eigenvalues in ascending order
z21 = z(nk+1:end,1:nk);
z11 = z(1:nk,1:nk);
if rank(z11)abs(s(nk,nk)) | abs(t(nk+1,nk+1))