How to set a lower bound at version 4.4.3?

Dear all.

we’re proceeding our tasks using SIR DSGE model.
But, the given dynare codes is only operated with dynare version 4.4.3( especially for sim1.m)

However, our future tasks require us to set a lower bound to some endogenous variables.

So, we want to ask how to set a lower bound at dynare version 4.4.3.
if there’s no solution, then we want to know the method with which the given dynare codes can be operated in the recent version of dynare that can use lmmcp.

Sincerely, Thanks.

Î am not sure I understand. See https://github.com/JohannesPfeifer/DSGE_mod/tree/master/Stock_SIR_2020

Thank you for replying!!

I’m really sorry for the poor explanation.
Actually, we are replicating Eichenbaum, Rebelo, & Tranbandt(2020) paper.
They opened their dynare code.

I really want to attach the benchmark.mod file down.
but, I’m not sure that I can attach it without their approval.

The mod file is here!
ert_model.mod (17.2 KB)

So, I write down their homepage address.
https://faculty.wcas.northwestern.edu/~yona/research.html

one of their paper named [Epidemics in the Neoclassical and New-Keynesian Models],
they announced that their dynare codes should be operated only with dynare version 4.4.3.

However, we have some trouble with restricting some endogenous variables in a specific range.

For instance, if there exists an endogenous variable m_t, we want to limit the range of m_t as “m_t>=0”.

However, we could not find any method to set a lower bound to the endogenous variable in dynare version 4.4.3.

If it is not possible in dynare version 4.4.3, then, we would like to run the modified code with recent dynare version.

However, sim1 were not run properly with the recent dynare version.
How can we modify the last part of the benchmark.mod to be properly operated in recent dynare version?

I hope you see the benchmark.mod(specific location in the zip file: \ERTcodes_neo_nk\competitive\ert_model.mod).

Sincerely, Thanks.

Which file exactly do you want to run in a new Dynare version?

Thank you for replying!

We want to run this mode file.
ert_model.mod (17.2 KB)

You need to replace the calls to sim1 by perfect_foresight_solver. See ert_model.mod (17.3 KB)

We really thank to your devotion!
now, we can run everything we want.
Thank you professor!

We are sorry to say that we have one more question.

we want to regulate some endogenous variables to be always over the steady-state level.

To be specific,

//Production 
y=pbreve*A*k(-1)^(1-alfa)*n^alfa;

//Marginal csot
mc=1/(A*alfa^alfa*(1-alfa)^(1-alfa))*w^alfa*rk^(1-alfa);

//Cost mininizing inputs
w=mc*alfa*A*n^(alfa-1)*k(-1)^(1-alfa);

//Law of motion capital
k=x+(1-delta)*k(-1);

//Aggregate resources(1)
y=x+m+g_ss;

//ms, susceptible
ms=cso+ksis;

//ksi, susceptible
ksis=cns-Tr;

//mi, infected
mi=coi+ksii;

//ksi, infected
ksii=cni-Tr;

//mr, recovered
mr=cor+ksir;

//ksi, recovered
ksir=cnr-Tr;

In the above code, we’d like to make ksis, ksii, ksir be always over the steady state level, 348.7205.

Is there any possible way for that?

Sincerely, Thanks.

That very much depends on the model. Usually, endogenous objects are jointly determined at time t and cannot be easily restricted unless it is part of the model setup itself. An example is the zero lower bound on interest rates.