Page 1 of 1

running error with New Keynesian Model

PostPosted: Thu Aug 19, 2010 8:46 pm
by marvinsyrene
I am estiamting a pretty standard new Keynesian monetary model with US data, every time when I specify three observables:y pai i, there is a error message:
??? Error using ==> chol
Matrix must be positive definite.

Error in ==> metropolis_hastings_initialization at 52
d = chol(vv);

everything works fine with two observables: y pai.
this is data for output, pai and i.
output is detrended log, multiplied by 400.
pai is the log difference of CPI, multiplied,
i is the fed fund rate

Re: running error with New Keynesian Model

PostPosted: Fri Aug 20, 2010 6:43 am
by jpfeifer
Hi, use mode_compute=6. There are a lot of posts dealing with this problem. See also the Wiki http://www.dynare.org/DynareWiki/MonteCarloOptimization