Hello,
I am trying to see whether financial development (measured as liquid liabilities divided to GDP) has an effect on the convergence between countries. For that purpose, I have generated an interaction term "linitial_LL".
I am using Cross-country IV Regressions to address the problem of endogeneity that comes from the initial value of growth of real GDP per capita. The results below come with two tests: Hasen J test and Underidentification test. I know that my Hansen test shows that the instruments used are not correlated with residuals. Nonetheless, I am not entirely sure about the interpretation of Underidentification test and whether it shows that I can accept my results. Can somebody help me with the interpretation?
ivreg2h gr linitial log_infl log_trade log_govsize log_school (linitial_LL Lliab Lliab2=) , gmm2s robust
Too few excluded instruments: standard IV model not estimable
IV with Generated Instruments only
Instruments created from Z:
linitial log_infl log_trade log_govsize log_school
2-Step GMM estimation
Estimates efficient for arbitrary heteroskedasticity
Statistics robust to heteroskedasticity
Number of obs = 37
F( 8, 28) = 25.52
Prob > F = 0.0000
Total (centered) SS = 44.63952131 Centered R2 = 0.6601
Total (uncentered) SS = 161.3884031 Uncentered R2 = 0.9060
Residual SS = 15.17405041 Root MSE = .6404
Robust
gr Coef. Std. Err. z P>z [95% Conf. Interval]
linitial_LL .2557908 .0572746 4.47 0.000 .1435346 .368047
Lliab -3.93395 .8873929 -4.43 0.000 -5.673208 -2.194692
Lliab2 .4522403 .1071857 4.22 0.000 .2421603 .6623204
linitial -1.084343 .2176225 -4.98 0.000 -1.510875 -.6578104
log_infl -.2873005 .056014 -5.13 0.000 -.3970859 -.1775151
log_trade .6776841 .0998588 6.79 0.000 .4819644 .8734038
log_govsize -1.101553 .3356589 -3.28 0.001 -1.759432 -.4436736
log_school 2.787074 .3847577 7.24 0.000 2.032963 3.541185
_cons 10.33173 2.546661 4.06 0.000 5.340368 15.3231
Underidentification test (Kleibergen-Paap rk LM statistic): 14.632
Chi-sq(13) P-val = 0.3309
Weak identification test (Cragg-Donald Wald F statistic): 3.029
(Kleibergen-Paap rk Wald F statistic): 10.108
Stock-Yogo weak ID test critical values: 5% maximal IV relative bias 18.73
10% maximal IV relative bias 10.33
20% maximal IV relative bias 5.94
30% maximal IV relative bias 4.37
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): 8.213
Chi-sq(12) P-val = 0.7683
Instrumented: linitial_LL Lliab Lliab2
Included instruments: linitial log_infl log_trade log_govsize log_school
Excluded instruments: linitial_LL_linitial_g linitial_LL_log_infl_g
linitial_LL_log_trade_g linitial_LL_log_govsize_g
linitial_LL_log_school_g Lliab_linitial_g Lliab_log_infl_g
Lliab_log_trade_g Lliab_log_govsize_g Lliab_log_school_g
Lliab2_linitial_g Lliab2_log_infl_g Lliab2_log_trade_g
Lliab2_log_govsize_g Lliab2_log_school_g
I am trying to see whether financial development (measured as liquid liabilities divided to GDP) has an effect on the convergence between countries. For that purpose, I have generated an interaction term "linitial_LL".
I am using Cross-country IV Regressions to address the problem of endogeneity that comes from the initial value of growth of real GDP per capita. The results below come with two tests: Hasen J test and Underidentification test. I know that my Hansen test shows that the instruments used are not correlated with residuals. Nonetheless, I am not entirely sure about the interpretation of Underidentification test and whether it shows that I can accept my results. Can somebody help me with the interpretation?
ivreg2h gr linitial log_infl log_trade log_govsize log_school (linitial_LL Lliab Lliab2=) , gmm2s robust
Too few excluded instruments: standard IV model not estimable
IV with Generated Instruments only
Instruments created from Z:
linitial log_infl log_trade log_govsize log_school
2-Step GMM estimation
Estimates efficient for arbitrary heteroskedasticity
Statistics robust to heteroskedasticity
Number of obs = 37
F( 8, 28) = 25.52
Prob > F = 0.0000
Total (centered) SS = 44.63952131 Centered R2 = 0.6601
Total (uncentered) SS = 161.3884031 Uncentered R2 = 0.9060
Residual SS = 15.17405041 Root MSE = .6404
Robust
gr Coef. Std. Err. z P>z [95% Conf. Interval]
linitial_LL .2557908 .0572746 4.47 0.000 .1435346 .368047
Lliab -3.93395 .8873929 -4.43 0.000 -5.673208 -2.194692
Lliab2 .4522403 .1071857 4.22 0.000 .2421603 .6623204
linitial -1.084343 .2176225 -4.98 0.000 -1.510875 -.6578104
log_infl -.2873005 .056014 -5.13 0.000 -.3970859 -.1775151
log_trade .6776841 .0998588 6.79 0.000 .4819644 .8734038
log_govsize -1.101553 .3356589 -3.28 0.001 -1.759432 -.4436736
log_school 2.787074 .3847577 7.24 0.000 2.032963 3.541185
_cons 10.33173 2.546661 4.06 0.000 5.340368 15.3231
Underidentification test (Kleibergen-Paap rk LM statistic): 14.632
Chi-sq(13) P-val = 0.3309
Weak identification test (Cragg-Donald Wald F statistic): 3.029
(Kleibergen-Paap rk Wald F statistic): 10.108
Stock-Yogo weak ID test critical values: 5% maximal IV relative bias 18.73
10% maximal IV relative bias 10.33
20% maximal IV relative bias 5.94
30% maximal IV relative bias 4.37
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): 8.213
Chi-sq(12) P-val = 0.7683
Instrumented: linitial_LL Lliab Lliab2
Included instruments: linitial log_infl log_trade log_govsize log_school
Excluded instruments: linitial_LL_linitial_g linitial_LL_log_infl_g
linitial_LL_log_trade_g linitial_LL_log_govsize_g
linitial_LL_log_school_g Lliab_linitial_g Lliab_log_infl_g
Lliab_log_trade_g Lliab_log_govsize_g Lliab_log_school_g
Lliab2_linitial_g Lliab2_log_infl_g Lliab2_log_trade_g
Lliab2_log_govsize_g Lliab2_log_school_g