Hello,
I am trying to use Occbin available on Dynare versions 5.0 and higher, but run into perplexing regime histories.
The constraint that I would like to implement is a collateral constraint of the following type. V_t\geq exp(\zeta_t)(P_{kt}K_t + P_{bt}B_t) where \mu_t is the lagrangian multiplier associated with this constraint and \zeta_t is an exogenous variable meant to capture a shock to the constraint.
I define a variable \Delta_t = V_t - exp(\zeta_t)(P_{kt}K_t + P_{bt}B_t).
There are two regimes, one where the constraint is slack (\mu_t=0, \Delta_t>0) and another where the constraint binds (\mu_t>0, \Delta_t=0).
I write down the regime-specific equations as follows.
[name=‘constraint’,bind = ‘CC’]
Del =0;
[name=‘constraint’,relax = ‘CC’]
mu=0;
I write the occbin_constraint block as follows.
occbin_constraints;
name ‘CC’; bind mu>0; relax Del>0;
end;
When I solve the model and simulate it using “surprise” shocks, I end up getting sometimes negative values of \mu_t or \Delta_t.
Have I incorrectly specified the occbin_constraint block?
I have also tried the following setup:
occbin_constraints;
name ‘CC’; bind Del<=0; relax mu<=0;
end;
but the solution will not converge now.