RANK CONDITION NOT VERIFIED

[St_aueb]

I have a simple 3-period OLG model. I am doing some tax/spending reforms by altering each time the residually determined fiscal policy instrument in the government budget constraint. I can find the steady-state, but it the rank condition message appears in some cases, depending on the adjusting instrument. I attach two mod files

In OLG_log.mod :

There are 12 eigenvalue(s) larger than 1 in modulus
for 12 forward-looking variable(s)

The rank condition ISN'T verified!

But everything its OK in OLG_log_sGs_adjusts.mod. I think it's a timing problem but I cannot find it. Please, any ideas. I m going crazy here!!!!!

Thank you!

[jpfeifer]

I don't think this is a timing issue. It's not the Blanchard-Kahn conditions that are violated, but the rank condition. There may be a good economic reason why for some instruments, there is no unique determinate solution. Have you tried thinking about the economic intuition behind the different results for different instruments? Is there a pattern for which instruments it works?

[St_aueb]

Thank you very much Johannes for your quick reply

I have many instruments (4 spending instruments, 4 tax instruments) so it is difficult to find a pattern for which instruments it works. It works fine for 5-6 instruments. It gives error for some others. For the main instruments that I am particularly interested in, it works. I was just checking the rest of them for robustness. So I conclude that I have to think more intuitively as you propose. I ran model diagnostics and there is no error so I guess timing is not the problem as you said.

Best
Stelios
Athens University of Economics and Business

[St_aueb]

I don't think this is a timing issue. It's not the Blanchard-Kahn conditions that are violated, but the rank condition. There may be a good economic reason why for some instruments, there is no unique determinate solution. Have you tried thinking about the economic intuition behind the different results for different instruments? Is there a pattern for which instruments it works?

I noticed that when I remove debt from my model, the problem disappears. This hods for every instrument

[jpfeifer]

Maybe this is a problem with Ricardian equivalence then. Debt and lump sum transfers are identical.