Hi

I am new in dynare I would like to ask some questions regarding calculation steady state values.

I am trying to replicate a one paper from Pierre Agenor called “optimal fiscal policy under commodity fiscal policy”. the log-linearized and steady state equations are used to solve this model.

There are 39 non-linear steady state equations. To use steady state values so as to solve log-linearized equations (such as I*Iss=IG*IGss+IPss*IP is loglinearized equation and Iss is coming from steady state solutions), I am using fsolve.

Firstly, would like to ask Can I introduce multiplicative constants for some of 39 steady state equations to solve these 39 nonlinear equations?

For example, I have DGss=(PRss^(1/pr))*Yss equation for goverment debt and Fss=(chi*TOss)/phiF for sovereign fund. Can I introduce DGss=0,1*Yss and Fss=0,02*Yss multiplicative constants instead of previous two equations .

Secondly, will it be wrong to use real data which I will have from data source such imf instead of using steady state values which I obtain from solving 39 nonlinear steady state equations by using fsolve .

If you help me I will be very happy

Thanks in advance

Sincerely yours

Cumhur

Sorry, but I don’t understand what you are trying to do. If you need a numerical solver for the steady states, you should work with the nonlinear first order conditions, not the linearized ones.

It would be wrong to use values from the data, because the steady state is the solution to the static model equations and they must be consistent with the dynamic model.

Dear Professor firstly really thanks for your attention and this nice forum. To find steady state i am working nonlinear first order equation as in picture which I attached here. I solve these nonlinear first order equation in matlab.

As at the end of picture, writer of this paper stated that some multiplicative constant are added to in one or two equation. For example, I have DGss=(PRss^(1/pr))Yss equation for goverment debt and . Writer introduced DGss=0,1Yss multiplicative constant instead of previous equation. I would like to ask whether this way is wrong? If you answer me Professor I will be very happy and thankful dear professor [Screenshot_2018-05-29-23-45-43|281x500]

That sounds as if the authors was calibrating the model by imposing some steady state ratios. That is common and correct.

Thanks for your attention .

Sincerely yours

Hi @Cumhur125, have you been able to replicate Agenor’s model. If so, I would be happy if you share your solution here.

Thank you for your time and consideration.