Problem with dynamic zero restrictions in a VAR

[fructuoso]

Dear Tom,

I would like to ask you a question regarding zero restrictions by using @irfrestrict.
I am trying to make zero a series of IRFs, but my problema is the following: I can zero'ed without any problema until the four period, but if I try to increase the number of zero shocks for more periods, the following message appears after I run the Cvmodel:

Covariance Model-Likelihood - Estimation by BFGS
NO CONVERGENCE IN 4 ITERATIONS
LAST CRITERION WAS 0.0000000
SUBITERATIONS LIMIT EXCEEDED.
ESTIMATION POSSIBLY HAS STALLED OR MACHINE ROUNDOFF IS MAKING FURTHER PROGRESS DIFFICULT
TRY HIGHER SUBITERATIONS LIMIT, TIGHTER CVCRIT, DIFFERENT SETTING FOR EXACTLINE OR ALPHA ON NLPAR
RESTARTING ESTIMATION FROM LAST ESTIMATES OR DIFFERENT INITIAL GUESSES MIGHT ALSO WORK
Observations 98
Log Likelihood -3850.8465
Log Likelihood Unrestricted 246.1813

Variable Coeff Std Error T-Stat Signif
************************************************************************************
1. THETA(1) -0.058242936 1.000000000 -0.05824 0.95355512
2. THETA(2) -0.002041910 1.000000000 -0.00204 0.99837079
3. THETA(3) 0.046331125 1.000000000 0.04633 0.96304633
4. THETA(4) -0.007261677 1.000000000 -0.00726 0.99420607
5. THETA(5) 0.011302774 1.000000000 0.01130 0.99098188
6. THETA(6) -0.099983751 1.000000000 -0.09998 0.92035723
7. THETA(7) 0.054855107 1.000000000 0.05486 0.95625390
8. THETA(8) 0.019920869 1.000000000 0.01992 0.98410650
9. THETA(9) -0.010902027 1.000000000 -0.01090 0.99130161
10. THETA(10) 0.033020830 1.000000000 0.03302 0.97365798
11. THETA(11) 0.000000000 1.000000000 0.00000 1.00000000
12. THETA(12) 0.000000000 1.000000000 0.00000 1.00000000
13. THETA(13) 0.000000000 1.000000000 0.00000 1.00000000
14. THETA(14) 0.564608494 1.000000000 0.56461 0.57234009
15. THETA(15) -0.018195165 1.000000000 -0.01820 0.98548316
16. THETA(16) 0.000000000 1.000000000 0.00000 1.00000000
17. THETA(17) 0.000000000 1.000000000 0.00000 1.00000000
18. THETA(18) 0.000000000 1.000000000 0.00000 1.00000000
19. THETA(19) 0.000000000 1.000000000 0.00000 1.00000000
20. THETA(20) 0.217924404 1.000000000 0.21792 0.82748801


***

I looked for a solution in the fórums, so I included BFGS as pmethod with piters=5.
I am attaching a part of my dataset and of my code.

Thank you very much for your support.

Fructuoso Alberto Borrallo

[TomDoan]

1. What's identifying the other four shocks?
2. You have eight restrictions on a single shock in a five variable model. The most "shape" restrictions you can have on a single shock is N-1 (in this case 4).

The multi-step shocks are subject to the standard Rubio-Ramirez-Waggoner-Zha counting rules, so your model is way short of identification. Compare with the example:

@IRFRestrict