Missing Standard Errors and Wald Statistics after running logistic regression

Jonathan Lau

Join Date: Mar 2018
Posts: 19

Missing Standard Errors and Wald Statistics after running logistic regression

21 Aug 2018, 05:19

Dear Statalisters

I'm running a logistic regression using a panel dataset of 70 countries from 1970-2010. When I regress Debt Crisis dummy variable using a 1-year lag of independent variables, I'm not sure why Stata displays the regression output without any standard errors & Wald Statistics.

Code:

. logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.CA L.Inflation L.NSavings L.D
> omesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovEx
> penditures ///
> L.GDPGrowth L.REER_overvaluation L.EquityPriceIndex,vce(r)  

Logistic regression                             Number of obs     =        257
                                                Wald chi2(0)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood =          0               Pseudo R2         =     1.0000

--------------------------------------------------------------------------------
               |               Robust
External_SDe~m | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------+----------------------------------------------------------------
YrAvgCred_GD~p |
           L1. |   .0002343          .        .       .            .           .
               |
            CA |
           L1. |   1.23e-28          .        .       .            .           .
               |
     Inflation |
           L1. |   18740.06          .        .       .            .           .
               |
      NSavings |
           L1. |   5.31e+13          .        .       .            .           .
               |
DomesticCredit |
           L1. |   18.01216          .        .       .            .           .
               |
 ExportsGrowth |
           L1. |    1933401          .        .       .            .           .
               |
           FDI |
           L1. |   1.80e+33          .        .       .            .           .
               |
  ln_TReserves |
           L1. |   7.75e+54          .        .       .            .           .
               |
ForexReserve~P |
           L1. |          0          .        .       .            .           .
               |
   M2_Reserves |
           L1. |   156.7724          .        .       .            .           .
               |
GlobalInterest |
           L1. |   .0000851          .        .       .            .           .
               |
GovExpenditu~s |
           L1. |   3.05e+34          .        .       .            .           .
               |
     GDPGrowth |
           L1. |   1.16e-17          .        .       .            .           .
               |
REER_overval~n |
           L1. |   .0038201          .        .       .            .           .
               |
EquityPriceI~x |
           L1. |   8.01e-11          .        .       .            .           .
               |
         _cons |          0          .        .       .            .           .
--------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 251 failures and 6 successes completely determined.

Something must have gone wrong in my regression. When I re-run the regression by excluding lags of either YrAvgCred_GDP_Gap, CA,GovExpenditures,REER_overvaluation or EquityPriceIndex, the standard errors & Wald Statistics are reported by Stata again.

Regression output after dropping Cred_GDP_Gap

Code:

. logistic External_SDebt_Dum L.CA L.Inflation L.NSavings L.DomesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovExpenditures ///
> L.GDPGrowth L.REER_overvaluation L.EquityPriceIndex,vce(r)  

Logistic regression                             Number of obs     =        274
                                                Wald chi2(14)     =      17.18
                                                Prob > chi2       =     0.2466
Log pseudolikelihood = -6.9420221               Pseudo R2         =     0.7595

------------------------------------------------------------------------------------
                   |               Robust
External_SDebt_Dum | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
                CA |
               L1. |   .5651487   .2223709    -1.45   0.147     .2613592    1.222046
                   |
         Inflation |
               L1. |   1.412208   .3121428     1.56   0.118     .9157061    2.177917
                   |
          NSavings |
               L1. |   .7176149    .250526    -0.95   0.342     .3620172    1.422505
                   |
    DomesticCredit |
               L1. |   1.025021   .0503808     0.50   0.615     .9308837    1.128678
                   |
     ExportsGrowth |
               L1. |   1.078473   .0918819     0.89   0.375     .9126198    1.274466
                   |
               FDI |
               L1. |   29.57978   47.96341     2.09   0.037     1.232473    709.9249
                   |
      ln_TReserves |
               L1. |   .2320873   .2026167    -1.67   0.094     .0419311    1.284596
                   |
 ForexReserves_GDP |
               L1. |   1.6e-283   4.8e-281    -2.24   0.025            0    2.33e-36
                   |
       M2_Reserves |
               L1. |   .9246002   .0923316    -0.79   0.432     .7602424    1.124491
                   |
    GlobalInterest |
               L1. |   .8622933   .1011469    -1.26   0.207     .6851871    1.085178
                   |
   GovExpenditures |
               L1. |   4.650817   3.297107     2.17   0.030     1.159008     18.6626
                   |
         GDPGrowth |
               L1. |   .5540131   .2484676    -1.32   0.188      .230021    1.334359
                   |
REER_overvaluation |
               L1. |   .8556241   .0768249    -1.74   0.082      .717555     1.02026
                   |
  EquityPriceIndex |
               L1. |   .4660356   .2082362    -1.71   0.088     .1941249    1.118812
                   |
             _cons |   8.15e+10   1.80e+12     1.14   0.254     1.39e-08    4.78e+29
------------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 217 failures and 0 successes completely determined.

Regression output after dropping CA

Code:

. logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.Inflation L.NSavings L.DomesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovExpenditures ///
> L.GDPGrowth L.REER_overvaluation L.EquityPriceIndex,vce(r)  

Logistic regression                             Number of obs     =        257
                                                Wald chi2(14)     =      64.12
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -6.3079508               Pseudo R2         =     0.7785

------------------------------------------------------------------------------------
                   |               Robust
External_SDebt_Dum | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
 YrAvgCred_GDP_Gap |
               L1. |   .7870885   .1072563    -1.76   0.079      .602602    1.028056
                   |
         Inflation |
               L1. |   1.152744    .507368     0.32   0.747     .4865029    2.731368
                   |
          NSavings |
               L1. |   1.302975   .2867231     1.20   0.229     .8465002    2.005604
                   |
    DomesticCredit |
               L1. |   1.075679   .1008831     0.78   0.437     .8950605    1.292746
                   |
     ExportsGrowth |
               L1. |   1.312463   .1720621     2.07   0.038     1.015069    1.696987
                   |
               FDI |
               L1. |   190.8707   773.2515     1.30   0.195     .0679781      535932
                   |
      ln_TReserves |
               L1. |   68.54516   286.2422     1.01   0.311     .0191156    245790.2
                   |
 ForexReserves_GDP |
               L1. |          0          0    -1.24   0.215            0    3.1e+206
                   |
       M2_Reserves |
               L1. |   1.076799   .2161754     0.37   0.712     .7265237    1.595952
                   |
    GlobalInterest |
               L1. |   .7806613   .1115465    -1.73   0.083     .5899794    1.032972
                   |
   GovExpenditures |
               L1. |   7.090031   9.740722     1.43   0.154     .4799708    104.7325
                   |
         GDPGrowth |
               L1. |   .4554118   .1539089    -2.33   0.020     .2348206    .8832273
                   |
REER_overvaluation |
               L1. |    .944055   .1139817    -0.48   0.633     .7451203    1.196102
                   |
  EquityPriceIndex |
               L1. |   .2636178    .264602    -1.33   0.184     .0368634    1.885185
                   |
             _cons |   6.55e-45   6.64e-43    -1.00   0.316     2.9e-131    1.49e+42
------------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 216 failures and 0 successes completely determined.

Regression output after dropping GovExpenditures

Code:

. logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.CA L.Inflation L.NSavings L.DomesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest ///
> L.GDPGrowth L.REER_overvaluation L.EquityPriceIndex,vce(r)  

Logistic regression                             Number of obs     =        257
                                                Wald chi2(14)     =      52.72
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -8.4354062               Pseudo R2         =     0.7037

------------------------------------------------------------------------------------
                   |               Robust
External_SDebt_Dum | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
 YrAvgCred_GDP_Gap |
               L1. |   .9274178   .0968898    -0.72   0.471     .7556977    1.138158
                   |
                CA |
               L1. |   .9745585   .1980865    -0.13   0.899     .6543268    1.451514
                   |
         Inflation |
               L1. |   .7013162   .1218975    -2.04   0.041     .4988437    .9859692
                   |
          NSavings |
               L1. |   1.182645   .4861614     0.41   0.683     .5283756    2.647074
                   |
    DomesticCredit |
               L1. |   1.027624   .0403721     0.69   0.488      .951466    1.109878
                   |
     ExportsGrowth |
               L1. |   1.095546   .1518915     0.66   0.510     .8348651    1.437623
                   |
               FDI |
               L1. |   1.788765   .7420622     1.40   0.161     .7932964    4.033396
                   |
      ln_TReserves |
               L1. |   .7121404   1.009195    -0.24   0.811     .0442897     11.4506
                   |
 ForexReserves_GDP |
               L1. |   1.23e-73   1.41e-71    -1.46   0.143     2.7e-171    5.68e+24
                   |
       M2_Reserves |
               L1. |   1.064476   .1242488     0.54   0.592     .8468004    1.338106
                   |
    GlobalInterest |
               L1. |   .9758268   .0499951    -0.48   0.633     .8825974    1.078904
                   |
         GDPGrowth |
               L1. |   .5412614   .2043194    -1.63   0.104     .2582789    1.134292
                   |
REER_overvaluation |
               L1. |   .8495199   .0411329    -3.37   0.001      .772608    .9340882
                   |
  EquityPriceIndex |
               L1. |   .7348835   .1826574    -1.24   0.215     .4514919    1.196154
                   |
             _cons |    1319761   3.92e+07     0.47   0.636     6.44e-20    2.71e+31
------------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 150 failures and 0 successes completely determined.

Regression output after dropping REER_overvaluation

Code:

 logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.CA L.Inflation L.NSavings L.DomesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovExpenditures ///
> L.GDPGrowth L.EquityPriceIndex,vce(r)  

Logistic regression                             Number of obs     =        333
                                                Wald chi2(14)     =      77.30
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -17.447942               Pseudo R2         =     0.5783

------------------------------------------------------------------------------------
                   |               Robust
External_SDebt_Dum | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
 YrAvgCred_GDP_Gap |
               L1. |   1.006213   .0371285     0.17   0.867     .9360116     1.08168
                   |
                CA |
               L1. |   1.424173   .2531854     1.99   0.047     1.005168    2.017841
                   |
         Inflation |
               L1. |   1.016997   .0462047     0.37   0.711      .930352    1.111711
                   |
          NSavings |
               L1. |   .9579689   .0682905    -0.60   0.547     .8330518    1.101617
                   |
    DomesticCredit |
               L1. |   .9961455   .0183246    -0.21   0.834     .9608697    1.032716
                   |
     ExportsGrowth |
               L1. |   .9133633   .0310205    -2.67   0.008     .8545437    .9762315
                   |
               FDI |
               L1. |   .4758457   .4467148    -0.79   0.429     .0755743    2.996111
                   |
      ln_TReserves |
               L1. |   .3638896   .2812422    -1.31   0.191     .0800009    1.655177
                   |
 ForexReserves_GDP |
               L1. |   1.48e-09   5.31e-08    -0.57   0.571     4.16e-40    5.27e+21
                   |
       M2_Reserves |
               L1. |   1.024911    .119144     0.21   0.832     .8160855    1.287173
                   |
    GlobalInterest |
               L1. |   1.108435   .0606569     1.88   0.060     .9957033     1.23393
                   |
   GovExpenditures |
               L1. |   .8919659    .122251    -0.83   0.404     .6818429    1.166842
                   |
         GDPGrowth |
               L1. |   .9214462   .1885935    -0.40   0.689     .6169552    1.376215
                   |
  EquityPriceIndex |
               L1. |   .9738773   .0339534    -0.76   0.448     .9095526    1.042751
                   |
             _cons |   3.96e+09   7.24e+10     1.21   0.226     1.11e-06    1.41e+25
------------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 4 failures and 0 successes completely determined.

Regression output after dropping EquityPriceIndex

Code:

 logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.CA L.Inflation L.NSavings L.DomesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovExpenditures ///
> L.GDPGrowth L.REER_overvaluation ,vce(r)  

Logistic regression                             Number of obs     =        312
                                                Wald chi2(14)     =      37.44
                                                Prob > chi2       =     0.0006
Log pseudolikelihood = -10.881951               Pseudo R2         =     0.7714

------------------------------------------------------------------------------------
                   |               Robust
External_SDebt_Dum | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
 YrAvgCred_GDP_Gap |
               L1. |   .9206694   .0807227    -0.94   0.346     .7753037     1.09329
                   |
                CA |
               L1. |   .7328431    .403305    -0.56   0.572     .2492125    2.155024
                   |
         Inflation |
               L1. |   .8990225   .1064936    -0.90   0.369     .7127572    1.133965
                   |
          NSavings |
               L1. |   1.279389   .3868071     0.81   0.415     .7073793    2.313943
                   |
    DomesticCredit |
               L1. |   .9732705   .0194036    -1.36   0.174     .9359735    1.012054
                   |
     ExportsGrowth |
               L1. |   1.151249   .0841098     1.93   0.054     .9976558    1.328488
                   |
               FDI |
               L1. |   1.314778   .2359267     1.53   0.127     .9249347    1.868934
                   |
      ln_TReserves |
               L1. |   .0060953   .0139758    -2.22   0.026     .0000681     .545381
                   |
 ForexReserves_GDP |
               L1. |   2.85e-25   1.41e-23    -1.14   0.254     1.80e-67    4.52e+17
                   |
       M2_Reserves |
               L1. |   .8413892   .1092195    -1.33   0.183     .6523845    1.085151
                   |
    GlobalInterest |
               L1. |    1.07684   .0867257     0.92   0.358     .9195975     1.26097
                   |
   GovExpenditures |
               L1. |   1.252803   .4548168     0.62   0.535      .614983    2.552128
                   |
         GDPGrowth |
               L1. |   .9530753   .1484941    -0.31   0.758     .7022723    1.293448
                   |
REER_overvaluation |
               L1. |   .8887908   .1039036    -1.01   0.313     .7067898    1.117658
                   |
             _cons |   2.28e+48   1.34e+50     1.90   0.057     .0305961    1.71e+98
------------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 161 failures and 0 successes completely determined.

Tags: None

Jonathan Lau

Join Date: Mar 2018

Posts: 19
#2

21 Aug 2018, 05:21

Does anybody has any ideas what may have caused this to happen?

Thank you very much.
Comment
William Lisowski

Join Date: Dec 2014

Posts: 10150
#3

21 Aug 2018, 06:56

Note the following message at the bottom of your first set of results.

Code:

Note: 251 failures and 6 successes completely determined.

Since only 257 of your observations were included in this model (the remainder must have been left out due to missing values), every observation can be perfectly predicted by the independent variables. This is not a good thing. Note that each set of your results suffers from this problem to a greater or lesser extent.

Here is a Stata FAQ that may help explain your problem. I've run out of time just now so I cannot take this further, perhaps others will advise.

https://www.stata.com/support/faqs/s...ic-regression/
1 like
Comment

Jonathan Lau

Join Date: Mar 2018
Posts: 19

21 Aug 2018, 08:24

Thanks for replying me, William.

Reading the Stata FAQ, maybe it's due to case 2 Hidden Collinearity. After following the steps after Example 2 of the website (under "What to do if this happens" subsection), Stata gave me the following output:

Code:

egen pattern = group(YrAvgCred_GDP_Gap CA Inflation NSavings DomesticCredit Exp
> ortsGrowth FDI ln_TReserves ForexReserves_GDP M2_Reserves GlobalInterest GovExp
> enditures GDPGrowth REER_overvaluation EquityPriceIndex)
(2602 missing values generated)


logistic External_SDebt_Dum L.YrAvgCred_GDP_Gap L.CA L.Inflation L.NSavings L.D
> omesticCredit L.ExportsGrowth ///
> L.FDI L.ln_TReserves L.ForexReserves_GDP L.M2_Reserves L.GlobalInterest L.GovEx
> penditures ///
> L.GDPGrowth L.REER_overvaluation L.EquityPriceIndex,vce(r)

Logistic regression                             Number of obs     =        257
                                                Wald chi2(0)      =          .
                                                Prob > chi2       =          .
Log pseudolikelihood =          0               Pseudo R2         =     1.0000

--------------------------------------------------------------------------------
               |               Robust
External_SDe~m | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------+----------------------------------------------------------------
YrAvgCred_GD~p |
           L1. |   .0002343          .        .       .            .           .
               |
            CA |
           L1. |   1.23e-28          .        .       .            .           .
               |
     Inflation |
           L1. |   18740.06          .        .       .            .           .
               |
      NSavings |
           L1. |   5.31e+13          .        .       .            .           .
               |
DomesticCredit |
           L1. |   18.01216          .        .       .            .           .
               |
 ExportsGrowth |
           L1. |    1933401          .        .       .            .           .
               |
           FDI |
           L1. |   1.80e+33          .        .       .            .           .
               |
  ln_TReserves |
           L1. |   7.75e+54          .        .       .            .           .
               |
ForexReserve~P |
           L1. |          0          .        .       .            .           .
               |
   M2_Reserves |
           L1. |   156.7724          .        .       .            .           .
               |
GlobalInterest |
           L1. |   .0000851          .        .       .            .           .
               |
GovExpenditu~s |
           L1. |   3.05e+34          .        .       .            .           .
               |
     GDPGrowth |
           L1. |   1.16e-17          .        .       .            .           .
               |
REER_overval~n |
           L1. |   .0038201          .        .       .            .           .
               |
EquityPriceI~x |
           L1. |   8.01e-11          .        .       .            .           .
               |
         _cons |          0          .        .       .            .           .
--------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 251 failures and 6 successes completely determined.

. 
end of do-file

. do "C:\Users\u1790263\AppData\Local\Temp\STD1904_000000.tmp"

. predict p
(option pr assumed; Pr(External_SDebt_Dum))
(2,613 missing values generated)

. sort p

. su p

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
           p |        257    .0233463    .1512955          0          1

. sum p,det

                   Pr(External_SDebt_Dum)
-------------------------------------------------------------
      Percentiles      Smallest
 1%            0              0
 5%            0              0
10%            0              0       Obs                 257
25%            0              0       Sum of Wgt.         257

50%            0                      Mean           .0233463
                        Largest       Std. Dev.      .1512955
75%            0              1
90%            0              1       Variance       .0228903
95%     4.48e-17              1       Skewness       6.313259
99%            1              1       Kurtosis       40.85724

The predicted probabilities for most of the observations without missing values seem to be zero.

Code:

tab pattern if p<1e-16

  see notes |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |          1        0.41        0.41
          2 |          1        0.41        0.82
          3 |          1        0.41        1.23
          4 |          1        0.41        1.65
          5 |          1        0.41        2.06
          6 |          1        0.41        2.47
          7 |          1        0.41        2.88
          8 |          1        0.41        3.29
          9 |          1        0.41        3.70
         10 |          1        0.41        4.12
         11 |          1        0.41        4.53
         12 |          1        0.41        4.94
         13 |          1        0.41        5.35
         14 |          1        0.41        5.76
         15 |          1        0.41        6.17
         16 |          1        0.41        6.58
         17 |          1        0.41        7.00
         18 |          1        0.41        7.41
         19 |          1        0.41        7.82
         20 |          1        0.41        8.23
         21 |          1        0.41        8.64
         22 |          1        0.41        9.05
         23 |          1        0.41        9.47
         24 |          1        0.41        9.88
         26 |          1        0.41       10.29
         27 |          1        0.41       10.70
         28 |          1        0.41       11.11
         31 |          1        0.41       11.52
         32 |          1        0.41       11.93
         34 |          1        0.41       12.35
         35 |          1        0.41       12.76
         37 |          1        0.41       13.17
         39 |          1        0.41       13.58
         40 |          1        0.41       13.99
         41 |          1        0.41       14.40
         42 |          1        0.41       14.81
         43 |          1        0.41       15.23
         44 |          1        0.41       15.64
         45 |          1        0.41       16.05
         47 |          1        0.41       16.46
         49 |          1        0.41       16.87
         51 |          1        0.41       17.28
         52 |          1        0.41       17.70
         56 |          1        0.41       18.11
         58 |          1        0.41       18.52
         59 |          1        0.41       18.93
         60 |          1        0.41       19.34
         62 |          1        0.41       19.75
         64 |          1        0.41       20.16
         65 |          1        0.41       20.58
         66 |          1        0.41       20.99
         67 |          1        0.41       21.40
         68 |          1        0.41       21.81
         69 |          1        0.41       22.22
         70 |          1        0.41       22.63
         72 |          1        0.41       23.05
         73 |          1        0.41       23.46
         75 |          1        0.41       23.87
         77 |          1        0.41       24.28
         78 |          1        0.41       24.69
         80 |          1        0.41       25.10
         81 |          1        0.41       25.51
         82 |          1        0.41       25.93
         83 |          1        0.41       26.34
         84 |          1        0.41       26.75
         85 |          1        0.41       27.16
         86 |          1        0.41       27.57
         87 |          1        0.41       27.98
         88 |          1        0.41       28.40
         89 |          1        0.41       28.81
         90 |          1        0.41       29.22
         91 |          1        0.41       29.63
         93 |          1        0.41       30.04
         94 |          1        0.41       30.45
         95 |          1        0.41       30.86
         96 |          1        0.41       31.28
         97 |          1        0.41       31.69
         98 |          1        0.41       32.10
         99 |          1        0.41       32.51
        100 |          1        0.41       32.92
        101 |          1        0.41       33.33
        102 |          1        0.41       33.74
        103 |          1        0.41       34.16
        104 |          1        0.41       34.57
        106 |          1        0.41       34.98
        107 |          1        0.41       35.39
        108 |          1        0.41       35.80
        109 |          1        0.41       36.21
        110 |          1        0.41       36.63
        111 |          1        0.41       37.04
        112 |          1        0.41       37.45
        113 |          1        0.41       37.86
        114 |          1        0.41       38.27
        115 |          1        0.41       38.68
        116 |          1        0.41       39.09
        117 |          1        0.41       39.51
        118 |          1        0.41       39.92
        119 |          1        0.41       40.33
        120 |          1        0.41       40.74
        121 |          1        0.41       41.15
        122 |          1        0.41       41.56
        123 |          1        0.41       41.98
        125 |          1        0.41       42.39
        126 |          1        0.41       42.80
        127 |          1        0.41       43.21
        128 |          1        0.41       43.62
        129 |          1        0.41       44.03
        130 |          1        0.41       44.44
        131 |          1        0.41       44.86
        132 |          1        0.41       45.27
        133 |          1        0.41       45.68
        134 |          1        0.41       46.09
        135 |          1        0.41       46.50
        136 |          1        0.41       46.91
        137 |          1        0.41       47.33
        138 |          1        0.41       47.74
        139 |          1        0.41       48.15
        140 |          1        0.41       48.56
        141 |          1        0.41       48.97
        142 |          1        0.41       49.38
        143 |          1        0.41       49.79
        144 |          1        0.41       50.21
        145 |          1        0.41       50.62
        146 |          1        0.41       51.03
        147 |          1        0.41       51.44
        148 |          1        0.41       51.85
        149 |          1        0.41       52.26
        150 |          1        0.41       52.67
        151 |          1        0.41       53.09
        153 |          1        0.41       53.50
        154 |          1        0.41       53.91
        155 |          1        0.41       54.32
        156 |          1        0.41       54.73
        157 |          1        0.41       55.14
        158 |          1        0.41       55.56
        159 |          1        0.41       55.97
        160 |          1        0.41       56.38
        161 |          1        0.41       56.79
        162 |          1        0.41       57.20
        163 |          1        0.41       57.61
        164 |          1        0.41       58.02
        165 |          1        0.41       58.44
        166 |          1        0.41       58.85
        167 |          1        0.41       59.26
        168 |          1        0.41       59.67
        169 |          1        0.41       60.08
        170 |          1        0.41       60.49
        171 |          1        0.41       60.91
        172 |          1        0.41       61.32
        173 |          1        0.41       61.73
        174 |          1        0.41       62.14
        175 |          1        0.41       62.55
        176 |          1        0.41       62.96
        177 |          1        0.41       63.37
        178 |          1        0.41       63.79
        179 |          1        0.41       64.20
        180 |          1        0.41       64.61
        181 |          1        0.41       65.02
        182 |          1        0.41       65.43
        183 |          1        0.41       65.84
        184 |          1        0.41       66.26
        185 |          1        0.41       66.67
        186 |          1        0.41       67.08
        187 |          1        0.41       67.49
        188 |          1        0.41       67.90
        189 |          1        0.41       68.31
        191 |          1        0.41       68.72
        192 |          1        0.41       69.14
        193 |          1        0.41       69.55
        194 |          1        0.41       69.96
        195 |          1        0.41       70.37
        196 |          1        0.41       70.78
        197 |          1        0.41       71.19
        199 |          1        0.41       71.60
        200 |          1        0.41       72.02
        201 |          1        0.41       72.43
        202 |          1        0.41       72.84
        203 |          1        0.41       73.25
        204 |          1        0.41       73.66
        205 |          1        0.41       74.07
        206 |          1        0.41       74.49
        207 |          1        0.41       74.90
        208 |          1        0.41       75.31
        209 |          1        0.41       75.72
        210 |          1        0.41       76.13
        211 |          1        0.41       76.54
        212 |          1        0.41       76.95
        213 |          1        0.41       77.37
        214 |          1        0.41       77.78
        215 |          1        0.41       78.19
        216 |          1        0.41       78.60
        217 |          1        0.41       79.01
        218 |          1        0.41       79.42
        219 |          1        0.41       79.84
        220 |          1        0.41       80.25
        221 |          1        0.41       80.66
        222 |          1        0.41       81.07
        223 |          1        0.41       81.48
        224 |          1        0.41       81.89
        225 |          1        0.41       82.30
        226 |          1        0.41       82.72
        227 |          1        0.41       83.13
        228 |          1        0.41       83.54
        229 |          1        0.41       83.95
        230 |          1        0.41       84.36
        231 |          1        0.41       84.77
        232 |          1        0.41       85.19
        233 |          1        0.41       85.60
        234 |          1        0.41       86.01
        235 |          1        0.41       86.42
        236 |          1        0.41       86.83
        237 |          1        0.41       87.24
        238 |          1        0.41       87.65
        239 |          1        0.41       88.07
        240 |          1        0.41       88.48
        241 |          1        0.41       88.89
        242 |          1        0.41       89.30
        243 |          1        0.41       89.71
        244 |          1        0.41       90.12
        245 |          1        0.41       90.53
        246 |          1        0.41       90.95
        247 |          1        0.41       91.36
        248 |          1        0.41       91.77
        249 |          1        0.41       92.18
        250 |          1        0.41       92.59
        251 |          1        0.41       93.00
        252 |          1        0.41       93.42
        253 |          1        0.41       93.83
        254 |          1        0.41       94.24
        255 |          1        0.41       94.65
        256 |          1        0.41       95.06
        257 |          1        0.41       95.47
        258 |          1        0.41       95.88
        259 |          1        0.41       96.30
        260 |          1        0.41       96.71
        261 |          1        0.41       97.12
        262 |          1        0.41       97.53
        263 |          1        0.41       97.94
        264 |          1        0.41       98.35
        265 |          1        0.41       98.77
        266 |          1        0.41       99.18
        267 |          1        0.41       99.59
        268 |          1        0.41      100.00
------------+-----------------------------------
      Total |        243      100.00

That means there are 243 numbers to type in the command list x1 x2 if pattern == XXXX // (use the value here from the tab step).

Comment

Jonathan Lau

Join Date: Mar 2018

Posts: 19
#5

21 Aug 2018, 08:26

Does this mean I should scale down the regression model by excluding more regressors, given many of them have some hidden collinearity with each other?

Thank you very much.
Comment
William Lisowski

Join Date: Dec 2014

Posts: 10150
#6

21 Aug 2018, 12:53

Does this mean I should scale down the regression model by excluding more regressors

Almost. You should start with a small model, with just one or two of your independent variables, and add more variables one at a time, carefully comparing the results from run to run looking for anomalies.

Let me add that your variable ForexReserves_GDP is strange in each of the models you presented. Look at the ridiculous sizes of the odds ratio estimates! The odds ratio reflects the ratio of the odds for a one-unit increase in ForexReserves_GDP. With this estimate you have a one-unit increase in ForexReserves_GDP driving the odds to near zero. This suggests the scale of ForexReserves_GDP is such that a one-unit change in the value is inconceivable, or else this is an outshoot of the "completely determined" problem. Based on the model in post #1 "after dropping REER_overvaluation" it may be the case that all four of the completely determined failures in that model lie above a cutoff value of ForexReserves_GDP. (I'm a bit rusty on this logic.)

For extra credit, you can look at the output of help logistic and note the Stored Results at the end of the output. The matrix e(rules) apparently (I've never needed to use it) will provide insight into what combinations of variables completely determine results.

Code:

matrix list e(rules)

will display the matrix. But warning: this whole piece of advice assumes you know something about Stata matrices (see Chapter 14 Matrix Expressions in the Stata User's Guide PDF) and Stata estimation commands (Chapter 20 Estimation and postestimation commands in the same PDF).
1 like
Comment

Jonathan Lau

Join Date: Mar 2018
Posts: 19

22 Aug 2018, 08:15

Thanks for replying me,William. I didn't notice the very small odds ratio estimate for ForexReserves_GDP. Thank you for pointing this out, & it seems the problem really lies on this variable.

When I just ran the regression with another variable & it,

Code:

logistic External_SDebt_Dum  L.YrAvgCred_GDP_Gap L.ForexReserves_GDP  ///
> ,vce(r)

Logistic regression                             Number of obs     =      1,315
                                                Wald chi2(2)      =      67.10
                                                Prob > chi2       =     0.0000
Log pseudolikelihood = -159.18352               Pseudo R2         =     0.2833

------------------------------------------------------------------------------
             |               Robust
External_S~m | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
YrAvgCred_~p |
         L1. |   .9540302   .0055002    -8.16   0.000     .9433107    .9648715
             |
ForexReser~P |
         L1. |   5.33e-11   3.27e-10    -3.86   0.000     3.26e-16    8.73e-06
             |
       _cons |   1.842493   .6913291     1.63   0.103      .883127    3.844046
------------------------------------------------------------------------------
Note: _cons estimates baseline odds.
Note: 16 failures and 0 successes completely determined.

there is already a note below regression output "Note: 16 failures and 0 successes completely determined."

It's very likely this variable is the cause of this problem. Perhaps I should avoid using it in the regression model.

I know nothing about Stata matrices,but may think about reading the 2 chapters of the User guide PDF if possible.

Thank you very much for your help.

Announcement

Missing Standard Errors and Wald Statistics after running logistic regression

Comment

Comment

Comment

Comment

Comment

Comment