Hi all, I am getting the appropriate output for the following code.
Can someone tell me what the second table is and why there are no confidence intervals etc. given there? Where is the difference with the first table?
Also, I am not getting a P value for the Cragg-Donald and Kleibergen-Paap. Without a P-value, can I assume that the instruments are bad because of the low F-statistic (<10)).
does anyone have an idea what I can improve on the model?
Many Thanks,
Julia
Code:
ivreg2h selfempl (cognability=) financelit age agesq selfafl riskcareer gradgdr gender lnnetinc nrchildren homeowner married [pw= weights_2009_hhsize_inc_all]
Too few excluded instruments: standard IV model not estimable
IV with Generated Instruments only
Instruments created from Z:
financelit age agesq selfafl riskcareer gradgdr gender lnnetinc nrchildren homeowner married
(sum of wgt is 9.0251e+02)
IV (2SLS) estimation
--------------------
Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity
Number of obs = 968
F( 12, 955) = 3.79
Prob > F = 0.0000
Total (centered) SS = 81.86712807 Centered R2 = -0.0385
Total (uncentered) SS = 90.28865901 Uncentered R2 = 0.0583
Residual SS = 85.02096193 Root MSE = .2964
------------------------------------------------------------------------------
| Robust
selfempl | Coefficient std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
cognability | .0882119 .0562729 1.57 0.117 -.022081 .1985048
financelit | -.0082473 .0087339 -0.94 0.345 -.0253654 .0088709
age | -.005932 .0074462 -0.80 0.426 -.0205263 .0086623
agesq | .0000994 .0000853 1.17 0.244 -.0000677 .0002666
selfafl | .0301166 .01181 2.55 0.011 .0069694 .0532639
riskcareer | .0171983 .0043941 3.91 0.000 .008586 .0258106
gradgdr | .0545762 .0244645 2.23 0.026 .0066266 .1025258
gender | -.0104425 .0227119 -0.46 0.646 -.0549571 .034072
lnnetinc | .017764 .028726 0.62 0.536 -.0385379 .0740659
nrchildren | -.0239217 .0081197 -2.95 0.003 -.0398359 -.0080075
homeowner | -.0038161 .0247614 -0.15 0.878 -.0523477 .0447154
married | .0001008 .0257867 0.00 0.997 -.0504402 .0506418
_cons | -.2115411 .2465575 -0.86 0.391 -.6947848 .2717026
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic): 19.419
Chi-sq(11) P-val = 0.0540
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic): 5.083
(Kleibergen-Paap rk Wald F statistic): 2.411
Stock-Yogo weak ID test critical values: 5% maximal IV relative bias 20.90
10% maximal IV relative bias 11.51
20% maximal IV relative bias 6.56
30% maximal IV relative bias 4.80
10% maximal IV size 40.90
15% maximal IV size 22.06
20% maximal IV size 15.56
25% maximal IV size 12.23
Source: Stock-Yogo (2005). Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments): 12.525
Chi-sq(10) P-val = 0.2514
------------------------------------------------------------------------------
Instrumented: cognability
Included instruments: financelit age agesq selfafl riskcareer gradgdr gender
lnnetinc nrchildren homeowner married
Excluded instruments: cognability_financelit_g cognability_age_g
cognability_agesq_g cognability_selfafl_g
cognability_riskcareer_g cognability_gradgdr_g
cognability_gender_g cognability_lnnetinc_g
cognability_nrchildren_g cognability_homeowner_g
cognability_married_g
------------------------------------------------------------------------------
Warning: variables have been centered
Warning: variables have been centered
-----------------------------
Variable | GenInst
-------------+---------------
cognability | .08821
| .0563
financelit | -.008247
| .00873
age | -.005932
| .00745
agesq | .00009944
| 8.5e-05
selfafl | .03012
| .0118
riskcareer | .0172
| .00439
gradgdr | .05458
| .0245
gender | -.01044
| .0227
lnnetinc | .01776
| .0287
nrchildren | -.02392
| .00812
homeowner | -.003816
| .0248
married | .0001008
| .0258
_cons | -.2115
| .247
-------------+---------------
N | 968
rmse | .296
j | 12.5
jdf | 10
jp | .251
-----------------------------
Legend: b/se
.
end of do-file
Also, I am not getting a P value for the Cragg-Donald and Kleibergen-Paap. Without a P-value, can I assume that the instruments are bad because of the low F-statistic (<10)).
does anyone have an idea what I can improve on the model?
Many Thanks,
Julia
Comment