ivreg2h warning variables have been centered and invalid name r(198);

[Julia Sophie]

Hello all,

I am trying a probit IV regression with weights. My command is as follows

ivreg2h self_empl (cogn_ability=f20s1) finance_literacy age age_squared sa_fl graduation_gdr gender mth_net_hh_inc risktolerance number_chi
> ldren homeowner married, first

Standard IV Results
 

First-stage regressions
-----------------------


First-stage regression of cogn_ability:

Statistics consistent for homoskedasticity only
Number of obs =                    969
----------------------------------------------------------------------------------
    cogn_ability | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-----------------+----------------------------------------------------------------
           f20s1 |   .0737916   .0237149     3.11   0.002     .0272523    .1203308
finance_literacy |   .1428539   .0137185    10.41   0.000     .1159321    .1697757
             age |   .0274964   .0262839     1.05   0.296    -.0240844    .0790771
    age_squared2 |  -.0004098   .0002887    -1.42   0.156    -.0009764    .0001567
           sa_fl |   .0711833   .0337132     2.11   0.035     .0050228    .1373438
  graduation_gdr |   -.138416    .076785    -1.80   0.072    -.2891026    .0122705
          gender |   .1592251   .0667075     2.39   0.017      .028315    .2901351
  mth_net_hh_inc |   .0000456   .0000233     1.96   0.051    -1.73e-07    .0000913
   risktolerance |   .0070372   .0122796     0.57   0.567    -.0170609    .0311352
 number_children |   .0597235   .0268744     2.22   0.026     .0069838    .1124631
       homeowner |   .1829295    .069381     2.64   0.009     .0467727    .3190862
         married |  -.0750075    .075888    -0.99   0.323    -.2239338    .0739187
           _cons |  -.9842482   .5917053    -1.66   0.097    -2.145439     .176943
----------------------------------------------------------------------------------
F test of excluded instruments:
  F(  1,   956) =     9.68
  Prob > F      =   0.0019
Sanderson-Windmeijer multivariate F test of excluded instruments:
  F(  1,   956) =     9.68
  Prob > F      =   0.0019



Summary results for first-stage regressions
-------------------------------------------

                                           (Underid)            (Weak id)
Variable     | F(  1,   956)  P-val | SW Chi-sq(  1) P-val | SW F(  1,   956)
cogn_ability |       9.68    0.0019 |        9.81   0.0017 |        9.68

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 Sanderson-Windmeijer F statistic.

Underidentification test
Ho: matrix of reduced form coefficients has rank=K1-1 (underidentified)
Ha: matrix has rank=K1 (identified)
Anderson canon. corr. LM statistic       Chi-sq(1)=9.72     P-val=0.0018

Weak identification test
Ho: equation is weakly identified
Cragg-Donald Wald F statistic                                       9.68

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.

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(1,956)=      12.83     P-val=0.0004
Anderson-Rubin Wald test           Chi-sq(1)=     13.01     P-val=0.0003
Stock-Wright LM S statistic        Chi-sq(1)=     12.84     P-val=0.0003

Number of observations               N  =        969
Number of regressors                 K  =         13
Number of endogenous regressors      K1 =          1
Number of instruments                L  =         13
Number of excluded instruments       L1 =          1

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics consistent for homoskedasticity only

                                                      Number of obs =      969
                                                      F( 12,   956) =     2.94
                                                      Prob > F      =   0.0005
Total (centered) SS     =   80.8255934                Centered R2   =  -1.3017
Total (uncentered) SS   =           89                Uncentered R2 =  -1.0903
Residual SS             =  186.0381302                Root MSE      =    .4382

----------------------------------------------------------------------------------
       self_empl | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-----------------+----------------------------------------------------------------
    cogn_ability |   .3222591   .1416957     2.27   0.023     .0445407    .5999775
finance_literacy |  -.0436765   .0223474    -1.95   0.051    -.0874766    .0001237
             age |  -.0141219   .0117364    -1.20   0.229    -.0371249     .008881
    age_squared2 |    .000222    .000134     1.66   0.097    -.0000406    .0004846
           sa_fl |   .0060097   .0180043     0.33   0.739     -.029278    .0412974
  graduation_gdr |   .0769905   .0347261     2.22   0.027     .0089286    .1450523
          gender |  -.0481954   .0365726    -1.32   0.188    -.1198764    .0234856
  mth_net_hh_inc |  -1.05e-07   .0000134    -0.01   0.994    -.0000263    .0000261
   risktolerance |   .0125772   .0055883     2.25   0.024     .0016243    .0235301
 number_children |  -.0389959   .0140751    -2.77   0.006    -.0665827   -.0114091
       homeowner |   -.054245    .040002    -1.36   0.175    -.1326476    .0241575
         married |   .0219434   .0359337     0.61   0.541    -.0484855    .0923722
           _cons |   .1665422   .2708018     0.61   0.539    -.3642196    .6973041
----------------------------------------------------------------------------------
Underidentification test (Anderson canon. corr. LM statistic):           9.715
                                                   Chi-sq(1) P-val =    0.0018
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):                9.682
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.
------------------------------------------------------------------------------
Sargan statistic (overidentification test of all instruments):           0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         cogn_ability
Included instruments: finance_literacy age age_squared2 sa_fl graduation_gdr
                      gender mth_net_hh_inc risktolerance number_children
                      homeowner married
Excluded instruments: f20s1
------------------------------------------------------------------------------

Warning: variables have been centered
_z_cogn_ability_finance_literacy_eps invalid name
r(198);

Can somebody explain why I get the warning: variables have been centered
_z_cogn_ability_finance_literacy_eps invalid name
r(198);?

How can I solve this problem?

[Andrew Musau]

ivreg2h is from SSC (FAQ Advice #12).

Can somebody explain why I get the warning: variables have been centered
_z_cogn_ability_finance_literacy_eps invalid name
r(198);?

These are two different issues. On the centering, the command is simply informing you that centering has been done on the regression sample. This (subtracting a constant from every value of a variable) affects the intercept, but not the slopes.

_z_cogn_ability_finance_literacy_eps invalid name

The authors of ivreg2h define a variable as _z_+ endogenous variable name+ "_" + regressor name+ "_"+ eps in their internal routines but are not too careful to consider Stata variable name length limits. So as in your case, you encounter errors if you have long variable names to begin with and the combined variable name is greater than 32 characters.

help limits

You can resolve this by renaming your long variable names to something shorter, e.g.,

rename (finance_literacy graduation_gdr number_children ///
mth_net_hh_inc risktolerance) (fin_lit grad_gdr nkids nethhinc risktol)

ivreg2h self_empl (cogn_ability=f20s1) fin_lit age age_squared sa_fl ///
grad_gdr gender nethhinc risktol nkids homeowner married, first