Page 1 of 1

noncontinuous function definition

PostPosted: Wed Aug 12, 2015 2:57 pm
by leonard007
Hello ,
During running my codes, I got a basic problem I don't know how to proceed! one of my endogenous state variables is shadow price of the outside equity of the bank (mue). In the steady state, the return of deposit(R) and the return of outside equity(Re)( because of Euler equation) are equal so mue tends to zero in the steady state. There is another function (x) which is an increasing function of mue. Function x is not a continuous function and when mue is zero, it would be a fix value (=-epsilon/k) [while in the continuous form when mue tends to zero, x tends to infinity, and this makes problem for dynare]. Now it is my question: How Can I solve this problem in dynare and say dynare that x is a noncontinuous function. in initial values I put x=-epsilon/k , but in the model if I put the continuous function of x there is an error, and if i put the fix function my model is not correct!

Re: noncontinuous function definition

PostPosted: Thu Aug 13, 2015 1:27 pm
by jpfeifer
Dynare's perturbation techniques cannot easily handle non-differentiabilities like the one you describe. Depending on what you are doing, Iacoviello/Guerrieri's Occbin might be a way out.