Trying to replicate a model. Help improving ?

[btamler]

Hi, I'm a student trying to learn more about DSGE Models. I'm trying to replicate a model (a very hard for me). Although I can run the model my IRF's are much different from the paper's model. The paper which I'm trying to replicate is:

"Capital Requirement and Financial Frictions in Banking: Macroeconomic Implications" from Ali Dib (Bank of Canada).

The paper can be found here:
http://www.bankofcanada.ca/wp-content/u ... p10-26.pdf

But I'm also attaching the file. The paper estimates a DSGE model with a financial accelerator à la BGG (Bernanke, Gertler and Gilchrist 1999 Seminal Paper).

The problem that I'm having is that my IRF's are much different from the paper. My IRF's are really unstable. I'm sending my .mod file and also a .pdf file which contains all the equations of the non-linear system, the steady-state and the log-linearized system. I guess that my errors (and I'm certain that I've commited a lot of them) are in the log-linearized system, which I made by hand, and in the steady-state parameters that are not included in the paper. Most of them I had to calculate from the non-linear equations.

Could some one help me improving ? I know that I've made a lot of errors...

I've also found some problems with the BK conditions. The thing is: when the equations are in the order of the .mod file attached the model runs. But, let's say that I put the same equations in a different order, for instance in the order as the .pdf file (just organizing them in a differente way), the file doesn't run since BK conditions are not satisfied. I really don't understand why this happens. The "check" command gives me different information with the order of the equations altered. With the .mod file attached I have "9 eingenvalues larger than 1 for 9 forward-looking variables". But, when I run the model as the .pdf file I encounter "13 eingenvalues larger than 1 for 9 forward-looking variables". Why the different order makes this error ? Sorry for the question but I really didn't found this in the manual!

Thank you so much for the help!