New shock Rank isn't verified

Stochastic simulations
1

Hey guys. I want to introduce a new shock to bank discount factor(nu), so that it can make a difference in the leverages of two banks.

To be specific, I add vnu in varexo, add random_nu in variable.
and a new equation:

log(nu)=(1-rho_nu)*log(nuss)+rho_nu*log(nu(-1))+vnu;

then it shows that the Rank ISN’T verified. It confuses me a bit because I follow exactly the way the original author does.
untitled_nu.mod (7.8 KB)
consolechatnu.m (3.5 KB)

And here is the original model which can run well but the leverages in two banks are exactly the same.
untitled_chat3.mod (7.5 KB)
consolechat3.m (3.4 KB)
My purpose is to make a difference of the leverages. Their lines should have a rough symmetry and complement each other, but I don’t know how to set it.

Thank you for your assistance.
Wish you all the best !

2

You have to rethink your approach. The original equation

nu=1-chi+chi*beta*lambda(+1)/lambda*nu(+1)*((rb(+1)-r/pi(+1))*lev+r/pi(+1));

describes a forward-looking difference equation. You now made the left-hand side follow an exogenous backward-looking AR process. The resulting second order difference equation for nu seems to be explosive.

1 Like
3

Thanks a lot Professor Pfeifer. I am going to forget about that shock and concentrate on an additional household department or a controlled experiment with limit to nu.