Dynare forums

Hi,

I would like to write a DSGE model with endogenous growth. For this sake I first wanted to "play around" with the model by Kung (JFE, 2014) and tried to replicate it. I stationarized the model and derived the SS. Using the same parameter values as in the paper I get that the BK conditions are not satisfied. I'm probably doing something wrong, but I double(triple)-checked the model equations and I cannot find the mistake. Could anyone maybe take a look at it?

Thanks a lot in advance.

The attached model does not run due to various problems. Are you sure you uploaded the correct version?

Sorry for the incovenience. I edited the previous post. Now it should be the correct version.

Thanks a lot for your help.

I can only recommend checking the timing of the equations. In particular, look at the Euler equations if you e.g. missed a (+1) somewhere, e.g. in the inner derivatives.

The BK conditions are now satisfied (I had written the Taylor Rule with Output growth instead of deviations from steady state).
I still haven't managed to replicate the code as there's quite some information missing from the paper. Did anyone manage to replicate it by any chance?

Another question concerns the steady state. I have one equation that in steady state isn't completely satisfied apparently but the residual is quite small (-3.3379e-06) and the code runs despite the error. Is that a problem? Moreover I don't get why I have this error, given that the equation is just a simple definition of an auxiliary variable.

Thanks in advance for your help

Epstein-Zin preferences introduce big nonlinearities into the model that sometimes lead to numerical problems. Your U1sigma is about 21 in steady state. It enters as

Code: Select all

exp(U1sigma)

which is huge and can result in overflow. Your are lucky in that the residual is still within Dynare's accepted tolerance. Thus, there is not too much to worry. But best practice in these cases is to introduce a normalizing constant into the utility function that makes U equal to 1 in steady state. That avoids most problems of this type.

Thanks a lot for the quick reply