Hi, I'm trying to do a gym regression on a Taylor rule with smoothing parameter, but I cannot get the matrix to converge.
This is my code
gmm (ESTER-((1-{b1})*({b2}+HICP*{b3}+GDPGAP*{b4})+ l1.ESTER*{b1})), igmm instruments(L(1/4).HICP L(1/4).GDPGAP L(1/4).ESTER L(1/4).CMP L(1/4).M03) wmatrix(hac bartlett opt) variables(HICP GDPGAP ESTER) vce(hac bartlett opt)
and this is what I get:
Step 1
Iteration 0: GMM criterion Q(b) = 4.5181651
Iteration 1: GMM criterion Q(b) = .01089994
Iteration 2: GMM criterion Q(b) = .00591039
Iteration 3: GMM criterion Q(b) = .00591039
....
Step 300
Iteration 0: GMM criterion Q(b) = .07218202
Iteration 1: GMM criterion Q(b) = .07218201
note: maximum number of GMM iterations reached.
GMM estimation
Number of parameters = 4
Number of moments = 21
Initial weight matrix: Unadjusted Number of obs = 196
GMM weight matrix: HAC Bartlett 20
(lags chosen by Newey–West)
------------------------------------------------------------------------------
| HAC
| Coefficient std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
/b1 | .9699711 .004906 197.71 0.000 .9603556 .9795866
/b2 | -3.656506 .5887776 -6.21 0.000 -4.810489 -2.502523
/b3 | 2.419334 .3795486 6.37 0.000 1.675433 3.163236
/b4 | .0747057 .1326703 0.56 0.573 -.1853232 .3347347
------------------------------------------------------------------------------
HAC standard errors based on Bartlett kernel with 20 lags.
Lags chosen by Newey–West method.
Instruments for equation 1: L.HICP L2.HICP L3.HICP L4.HICP L.GDPGAP L2.GDPGAP L3.GDPGAP L4.GDPGAP L.ESTER
L2.ESTER L3.ESTER L4.ESTER L.CMP L2.CMP L3.CMP L4.CMP L.M03 L2.M03 L3.M03 L4.M03 _cons
Warning: Convergence not achieved.
I don't think this happens because my moments are more than the parameters, because I've seen others do the regression and get a result with the same structure. what can I do?
This is my code
gmm (ESTER-((1-{b1})*({b2}+HICP*{b3}+GDPGAP*{b4})+ l1.ESTER*{b1})), igmm instruments(L(1/4).HICP L(1/4).GDPGAP L(1/4).ESTER L(1/4).CMP L(1/4).M03) wmatrix(hac bartlett opt) variables(HICP GDPGAP ESTER) vce(hac bartlett opt)
and this is what I get:
Step 1
Iteration 0: GMM criterion Q(b) = 4.5181651
Iteration 1: GMM criterion Q(b) = .01089994
Iteration 2: GMM criterion Q(b) = .00591039
Iteration 3: GMM criterion Q(b) = .00591039
....
Step 300
Iteration 0: GMM criterion Q(b) = .07218202
Iteration 1: GMM criterion Q(b) = .07218201
note: maximum number of GMM iterations reached.
GMM estimation
Number of parameters = 4
Number of moments = 21
Initial weight matrix: Unadjusted Number of obs = 196
GMM weight matrix: HAC Bartlett 20
(lags chosen by Newey–West)
------------------------------------------------------------------------------
| HAC
| Coefficient std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
/b1 | .9699711 .004906 197.71 0.000 .9603556 .9795866
/b2 | -3.656506 .5887776 -6.21 0.000 -4.810489 -2.502523
/b3 | 2.419334 .3795486 6.37 0.000 1.675433 3.163236
/b4 | .0747057 .1326703 0.56 0.573 -.1853232 .3347347
------------------------------------------------------------------------------
HAC standard errors based on Bartlett kernel with 20 lags.
Lags chosen by Newey–West method.
Instruments for equation 1: L.HICP L2.HICP L3.HICP L4.HICP L.GDPGAP L2.GDPGAP L3.GDPGAP L4.GDPGAP L.ESTER
L2.ESTER L3.ESTER L4.ESTER L.CMP L2.CMP L3.CMP L4.CMP L.M03 L2.M03 L3.M03 L4.M03 _cons
Warning: Convergence not achieved.
I don't think this happens because my moments are more than the parameters, because I've seen others do the regression and get a result with the same structure. what can I do?
