A question about the policy function

I want to take the Policy function out and do something about it.
However, I find that the state var of the policy function stored in oo_.dr.ghx does not correspond to the state var of the policy function output from the console. In other words, the policy function stored in oo_.dr.ghx has less state var than the one printed on the console.What’s going on here?
Anybody know? Thank you for your attention

That cannot be. ghx and ghu must have the same number of rows as there are rows in the display of policy rules at first order.

I think I may have found the problem.

The coefficients of the missing variables are close to zero, so Dynare omitted them when presenting the policy function…

Also, if I want to get the two matrices used to solve the policy function(B&K method or Schur method), do Dynare store them? Where can I find them?

Thank you very much for your kindly help.

  1. You stated the opposite. Indeed, the display will suppress all states with coefficients smaller than options_.dr_display_tol, while the matrices will contain all entries.
  2. The matrices going to the Schur decomposition can be found in dyn_first_order_solver.m in
    [ss, tt, w, sdim, dr.eigval, info1] = mjdgges(E, D, DynareOptions.qz_criterium, DynareOptions.qz_zero_threshold);

Oh, I see. Thank you for your kind help.