Dear all,
Can I interpret the following output as the instruments are not valid by looking at the Sanderson-Windmeijer multivariate F test.
How can I proceed if I cannot find a better instrument?
Thank you,
Nazlı
xi: xtivreg2 `t' (`z'=instrument_p) i.year i.ccode , fe first cluster (ccode)
outreg2 using CIFCIILII.xls, ctitle(`y', K-INTENSİVE in CII) keep( `z' ) e(N N_g F r2_a r2u r2c widstat j idstat idp cdf rkf arf arfp archi2 archi2p sstat sstatp)
panel variable: id (unbalanced)
time variable: year, 2006 to 2015, but with gaps
delta: 1 unit
i.year _Iyear_2006-2015 (naturally coded; _Iyear_2006 omitted)
i.ccode _Iccode_1-81 (naturally coded; _Iccode_1 omitted)
Warning - singleton groups detected. 3352 observation(s) not used.
Warning - collinearities detected
Vars dropped: _Iccode_2
FIXED EFFECTS ESTIMATION
------------------------
Number of groups = 2127 Obs per group: min = 2
avg = 3.1
max = 10
Warning - collinearities detected
Vars dropped: _Iccode_2 _
First-stage regressions
-----------------------
FIXED EFFECTS ESTIMATION
------------------------
Number of groups = 2127 Obs per group: min = 2
avg = 3.1
max = 10
First-stage regression of im_den_city:
Statistics robust to heteroskedasticity and clustering on ccode
Number of obs = 6511
Number of clusters (ccode) = 77
------------------------------------------------------------------------------
| Robust
im_den_city | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
instrument_p | 1.647439 1.957921 0.84 0.400 -2.191079 5.485957
_Iyear_2007 | -.0215135 .0236983 -0.91 0.364 -.0679742 .0249473
_Iyear_2008 | -.0230585 .0272326 -0.85 0.397 -.0764481 .0303312
_Iyear_2009 | -.059271 .0584188 -1.01 0.310 -.1738014 .0552595
_Iyear_2010 | .055957 .1071557 0.52 0.602 -.1541224 .2660365
_Iyear_2011 | -.0242156 .059115 -0.41 0.682 -.140111 .0916798
_Iyear_2012 | .0617291 .1225663 0.50 0.615 -.1785628 .3020211
_Iyear_2013 | .3920116 .1453562 2.70 0.007 .1070398 .6769834
_Iyear_2014 | .8483978 .2831119 3.00 0.003 .293355 1.40344
_Iyear_2015 | 1.747242 .4438663 3.94 0.000 .8770397 2.617445
_Iccode_2 | 0 (omitted)
_Iccode_81 | 0 (omitted)
------------------------------------------------------------------------------
F test of excluded instruments:
F( 1, 76) = 0.71
Prob > F = 0.4028
Sanderson-Windmeijer multivariate F test of excluded instruments:
F( 1, 76) = 0.71
Prob > F = 0.4028
Summary results for first-stage regressions
-------------------------------------------
(Underid) (Weak id)
Variable | F( 1, 76) P-val | SW Chi-sq( 1) P-val | SW F( 1, 76)
im_den_city | 0.71 0.4028 | 0.72 0.3967 | 0.71
NB: first-stage test statistics cluster-robust
Stock-Yogo weak ID F test critical values for single endogenous regressor:
10% maximal IV size 16.38
15% maximal IV size 8.96
20% maximal IV size 6.66
25% maximal IV size 5.53
Source: Stock-Yogo (2005). Reproduced by permission.
NB: Critical values are for i.i.d. errors only.
Underidentification test
Ho: matrix of reduced form coefficients has rank=K1-1 (underidentified)
Ha: matrix has rank=K1 (identified)
Kleibergen-Paap rk LM statistic Chi-sq(1)=0.67 P-val=0.4143
Weak identification test
Ho: equation is weakly identified
Cragg-Donald Wald F statistic 128.68
Kleibergen-Paap Wald rk F statistic 0.71
Stock-Yogo weak ID test critical values for K1=1 and L1=1:
10% maximal IV size 16.38
15% maximal IV size 8.96
20% maximal IV size 6.66
25% maximal IV size 5.53
Source: Stock-Yogo (2005). Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
Weak-instrument-robust inference
Tests of joint significance of endogenous regressors B1 in main equation
Ho: B1=0 and orthogonality conditions are valid
Anderson-Rubin Wald test F(1,76)= 7.22 P-val=0.0089
Anderson-Rubin Wald test Chi-sq(1)= 7.32 P-val=0.0068
Stock-Wright LM S statistic Chi-sq(1)= 6.24 P-val=0.0125
NB: Underidentification, weak identification and weak-identification-robust
test statistics cluster-robust
Number of clusters N_clust = 77
Number of observations N = 6511
Number of regressors K = 10
Number of endogenous regressors K1 = 1
Number of instruments L = 10
Number of excluded instruments L1 = 1
IV (2SLS) estimation
--------------------
Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on ccode
Number of clusters (ccode) = 77 Number of obs = 6511
F( 10, 76) = 2.95
Prob > F = 0.0035
Total (centered) SS = 207489.9952 Centered R2 = -0.0955
Total (uncentered) SS = 207489.9952 Uncentered R2 = -0.0955
Residual SS = 227309.2738 Root MSE = 7.201
------------------------------------------------------------------------------
| Robust
reldsales_c | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
im_den_city | -1.823125 2.236623 -0.82 0.415 -6.206826 2.560575
_Iyear_2007 | -.1671247 .3489624 -0.48 0.632 -.8510783 .516829
_Iyear_2008 | -.3170098 .4293094 -0.74 0.460 -1.158441 .5244212
_Iyear_2009 | -.4779769 .5637315 -0.85 0.397 -1.58287 .6269166
_Iyear_2010 | -1.195521 .5330187 -2.24 0.025 -2.240219 -.1508239
_Iyear_2011 | -1.827432 .6027818 -3.03 0.002 -3.008863 -.6460016
_Iyear_2012 | -2.739736 .7133271 -3.84 0.000 -4.137832 -1.341641
_Iyear_2013 | -2.736951 1.133316 -2.41 0.016 -4.95821 -.515692
_Iyear_2014 | -.9507677 2.409706 -0.39 0.693 -5.673705 3.772169
_Iyear_2015 | 1.109389 4.590752 0.24 0.809 -7.888319 10.1071
_Iccode_2 | 0 (omitted)
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic): 0.666
Chi-sq(1) P-val = 0.4143
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic): 128.676
(Kleibergen-Paap rk Wald F statistic): 0.708
Stock-Yogo weak ID test critical values: 10% maximal IV size 16.38
15% maximal IV size 8.96
20% maximal IV size 6.66
25% maximal IV size 5.53
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): 0.000
(equation exactly identified)
------------------------------------------------------------------------------
Instrumented: im_den_city
Included instruments: _Iyear_2007 _Iyear_2008 _Iyear_2009 _Iyear_2010
_Iyear_2011 _Iyear_2012 _Iyear_2013 _Iyear_2014
_Iyear_2015
Excluded instruments: instrument_p
Dropped collinear: _Iccode_2 _Iccode_3 _Iccode_4 _Iccode_5 _Iccode_6
_Iccode_7 _Iccode_8 _Iccode_9 _Iccode_10 _Iccode_11
_Iccode_12 _Iccode_13 _Iccode_14 _Iccode_15 _Iccode_16
_Iccode_17 _Iccode_18 _Iccode_19 _Iccode_20 _Iccode_21
_Iccode_22 _Iccode_23 _Iccode_24 _Iccode_25 _Iccode_26
_Iccode_27 _Iccode_28 _Iccode_29 _Iccode_30 _Iccode_31
_Iccode_32 _Iccode_33 _Iccode_34 _Iccode_35 _Iccode_36
_Iccode_37 _Iccode_38 _Iccode_39 _Iccode_40 _Iccode_41
_Iccode_42 _Iccode_43 _Iccode_44 _Iccode_45 _Iccode_46
_Iccode_47 _Iccode_48 _Iccode_49 _Iccode_50 _Iccode_51
_Iccode_52 _Iccode_53 _Iccode_54 _Iccode_55 _Iccode_56
_Iccode_57 _Iccode_58 _Iccode_59 _Iccode_60 _Iccode_61
_Iccode_62 _Iccode_63 _Iccode_64 _Iccode_65 _Iccode_66
_Iccode_67 _Iccode_68 _Iccode_69 _Iccode_70 _Iccode_71
_Iccode_72 _Iccode_73 _Iccode_74 _Iccode_75 _Iccode_76
_Iccode_77 _Iccode_78 _Iccode_79 _Iccode_80 _Iccode_81
------------------------------------------------------------------------------
Related Posts with xtivreg2 interpretation
Create Risk Prediction Variable from 2 Different Heckman Correction ModelsHi everyone, I'm wondering if there's a way to create a risk prediction variable from the results o…
floatplot updated on SSCThanks as always to Kit Baum, the files for floatplot on SSC have been updated. Somehow or other th…
Multilevel modeling a binary outcome taking account a complex survey design (with psu, strata, and weights) using svy:melogitProgramme: Stata SE 16.1 Desired Analysis: Multi-level modelling for complex survey design (with psu…
How 'mixed' Command Identifies Level-1 and Level-2 Variables?My confusion is depicted in the picture below: Array …
Should I use "c." or "i." prefix or something else?Dear all, I have a panel dataset of rated firms and want to study the effect of busyness cycles on …
Subscribe to:
Post Comments (Atom)
0 Response to xtivreg2 interpretation
Post a Comment