Nonzero residuals of static equations

Steady state
1

Hi professor
I have two equations:
1- b=d+kb
2- jb=rbb-rdd.
In steady state model block i use these two equations as below:
1- d=ln(exp(b)-exp(kb));
2- jb=ln(exp(rb)*exp(b)-exp(d)*exp(d));

But in residuals of static equations part, the residuals of two equations mentioned earlier are nonzero.
Would you please help me?
Best regards.
AAA.mod (16.5 KB)

2

Your equations above are impossible to read, because some multiplication signs must be missing.

It is usually better to work with auxiliary equations to get a log-linearization. See Question about understanding irfs in dynare

3

Excuse me
the two equations are:
1- b=d+kb;
2- jb=brb-drd
thanks for your tips, but if you see the steady state model block i find the steady state value of all variables by mentioned method and residuals of all get zero instead of 2 mentioned equations.
would you please check it again?
Best regards

4

Hi professor
as you said in the topics, I performed an exp() substituion and the right form of above equations are:
1- exp(b)=exp(d)+exp(kb);
2-exp(jb)=exp(b)*exp(rb)-exp(d)*exp(rd);
would you now check the model again?

5

In

d=log(exp(b)-exp(kb));

you are getting a negative number inside of the bracket.

6

Excuse me professor
How do you understand that the exparation (ln(exp(b)-exp(kb)) become negative?
Because i experience this problem before and dynare give an error that “resid: The initial values for the steady state of the following variables are complex”.
But dynare do not give me this error now?
How i can fix the error?

7

You can execute the parameter definitions and the steady state model equations using F9 in Matlab. You will then see the problem. Tip: use % instead of // to add comments. Otherwise Matlab will complain if you execute lines with appended comments.

8

Hi Prof. jpfeifer, I have similar doubts.

If my model has, e.g. 20 equations, by resid MATLAB/dynare only reports the first 2 equations nonzero,

does it mean (i) I only need to check these 2 equations,
or (ii) I have to check all 20 equations
to remove these 2 nonzeros??

9

This means that all equation may be wrong, but for debugging, the place to start are the two equations with non-zero residuals.

1 Like
10

Furthermore, one follow-up question:
This post is in “Steady state” category, does it mean
(i) I only need to check my steady-state settings/file,
or (ii) the whole equation system and my model settings may have flaws??

11

Option ii). Your entered equations and the computed steady state values are inconsistent. There must be a mistake somewhere, but it is unclear whether it is the model equations or the steady state computation that has the mistake.

1 Like