I am setting up a SOE model with a bank, deposits, loans (domestic and foreign) and bank capital. But I would like to make the rate on deposits (and foreign loans) a monetary policy rate and shock it.
However, adding such rate as an AR(1) implies that I have an extra equation in the dynare code.
I would like to know what step to take further. Should I remove one of the budget constraints or a first order condition? I am not sure about the validity of any of the options.
Thank you very much in advance
Dear Fernando, you do not need an extra-equation: just impose that the deposit rate is equal to the policy rate, which in turn is determined by the Taylor rule. Then you can add a shock to the Taylor rule and you are done. I think most of papers with a banking sector just assume that the deposit nominal rate is the policy rate.
Hello, Valerio, thanks for your response.
What I wonder is what to write in the code. For instance, if R is the deposits rate, adding Rp as new variable, then adding the equation “Rp = R”; plus including the actual Taylor rule eq “Rp = …;” yields the message that I have an extra equation.
Please correct me if I get you wrong. So you suggest to substitute R as Rp anywhere it appears in the code and then add the Taylor rule equation of “Rp = …;” I imagine this has errors messages because the approximations with respect to R when solving. I am solving the model-code in levels.
Yes this is my suggestion, but I think you will get the same error message because you will still have one equation more, right?. Send me the code, I’ll try to help you.
Thanks, Valerio, I will contact you by email or PM. Best
Here I post the MOD file, thanks for the help.modeloshared.mod (6.1 KB)
That does not make sense. If
then R is already uniquely determined. You cannot add another equation to again determine
R. Either it will be the same and is redundant or it will contradict you model. Usually, the way to go is to have the Taylor rule determining
R and replacing every occurrence of
do you have nominal rigidities in the model?
Thanks for help as well.
If I follow correctly, in my code, I want to make R_H as the mon policy (Rp).
Hence, (not doing anything else) I add Rp as variable in the var section.
I substitute the R_H for Rp in the whole model section.
Add the taylor rule equation for Rp in the model area.
Include the initial value of Rp in the corresponding section.
Doing these changes, none of the algorithms is able to find the steady state values. That’s why I am a bit confused about this procedure.
I have no nominal rigidities in the model so far.
Ok that’s the point. In your model a Taylor rule makes little sense, because your model is real. If you want, you can add 2 variables (nominal rate and inflation) and two equations (Taylor rule and Fisher equation to define the nominal rate), but these two equations would be completely disconnected from the rest of the model: in other words, in your model the classical dicothomy holds.
Thanks again for the help.
I did try your suggestions, it wasn´t finding the steady state values. Now it does find them, I just need to work around the eigenvalues and forwards looking variables.
What I did was to set the initial values of the variables with most of the decimals I could. I guess providing a steady state file would help in this regard.
If Dynare found a steady state then using an analytical steady state does not help unless there are multiple steady states and Dynare found the wrong one.
Thanks, Johannes, I think I did solve the code issue.