I am stuck at BK conditions not being satisfied despite anything I do! Any help would be much appreciated.

As a guide to the attached files - I was initially working on a log-linearised model but quickly realised that I needed to explicitly calculate Steady state of my model even if some of the values didn’t appear explicitly in the model. So I ‘non-linearised’ the model by putting in definitions of STEADY_STATE, and then linked it up with standard SS computation files. I am calibrating a couple of parameters to make way for great ratios that I am imposing on the model.

The first three files pertain to those. It solves SS fine. But as you will find there is a rank mismatch.

So in order to check if there was something weird happening with all this transformation, I ‘rewrote’ the log-linearised model in the fourth file (Final_Haul_24_llss.mod), using the relevant SS values of variables from the non-linear set up, that feature in the log-linearised equations. They are part of the parameters block in this model.

And BK condition still remains unsatisfied.

I have tried playing around with parameter values; and the ‘predetermined_variables’ command for my predetermined variables (capital, domestic and foreign bonds according to my understanding)…but nothing seems to be working.

As always in these cases, start with a smaller working model. Also, check that everything you want to determine endogenously can be determined endogenously, see [An infinity of steady states with Taylor rules)

Thanks a lot for your reply. I went through the link that you shared - but unfortunately I don’t think I have been able to make too much sense out of it (also with the mod file missing in that thread!). Can I probe a little further on your comment ‘everything you want to determine endogenously can be determined endogenously’. If I have as many endogenous variables as equations and that my steady state works out fine; how can this be a problem? And if so, how should I go about checking if any of the endogenous variables cannot be determined?

My model is borrowed heavily from the basic Medina and Soto (07) model - which from what I understand is quite standard in literature. Do you have any suggestions, from a bird’s eye view of my code, if there is any equation here that immediately jumps out to you as odd?

What I was saying is the following: Sometimes people try to e.g. compute the steady state nominal interest rate and the steady state inflation rate in the model endogenously. But both are often not uniquely determined. For every inflation rate, the Fisher equation gives you a corresponding nominal interest rate. Thus, using the steady_state-operator to endogenously determine the two steady states for these variables does not work. You need to tell the model the steady state of one of these variables for the other to be determined.