Probit regression failed completely determination and back up issue

[chris chan]

I am encountering 2 issues when running probit regression on a data sample of corporate bonds. Essentially the data consist; dummy for downgraded bond, lagged short interest, lagged market cap and lagged stock return and a few more. I ran into 2 question. First there shows a note in the regression summary saying 42 failures. I searched for the official explanation but I don't think they are applicable.

Iteration 0: log likelihood = -45979.692
Iteration 1: log likelihood = -44061.666
Iteration 2: log likelihood = -43975.815
Iteration 3: log likelihood = -43975.59
Iteration 4: log likelihood = -43975.59

Probit regression Number of obs = 347,531
LR chi2(3) = 4008.20
Prob > chi2 = 0.0000
Log likelihood = -43975.59 Pseudo R2 = 0.0436

------------------------------------------------------------------------------------------
dummy_down | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------------------+----------------------------------------------------------------
lag_ab_ss | 5.550842 .136764 40.59 0.000 5.282789 5.818895
lag_D_TA | -.0777935 .0274189 -2.84 0.005 -.1315336 -.0240534
lag_stock_monthly_return | -1.805196 .037271 -48.43 0.000 -1.878246 -1.732146
_cons | -1.919995 .0086601 -221.71 0.000 -1.936968 -1.903021
------------------------------------------------------------------------------------------
Note: 42 failures and 0 successes completely determined.

Another issue is when running the regression, the process kept in generating iterations and showed back up the message.

Iteration 0: log likelihood = -46667.784
Iteration 1: log likelihood = -45413.223
Iteration 2: log likelihood = -45403.731
Iteration 3: log likelihood = -45403.724
Iteration 4: log likelihood = -45403.724
Iteration 5: log likelihood = -45403.724 (backed up)
Iteration 6: log likelihood = -45403.724 (backed up)
Iteration 7: log likelihood = -45403.724 (backed up)
Iteration 8: log likelihood = -45403.724 (backed up)
Iteration 9: log likelihood = -45403.724 (backed up)
Iteration 10: log likelihood = -45403.724 (backed up)
Iteration 11: log likelihood = -45403.724 (backed up)
Iteration 12: log likelihood = -45403.724 (backed up)
Iteration 13: log likelihood = -45403.724 (backed up)
Iteration 14: log likelihood = -45403.724 (backed up)

below is a sample of my variable

* Example generated by -dataex-. To install: ssc install dataex
clear
input float lag_stock_monthly_return
   .04285286
 -.032464337
   .04285286
           .
 -.034371644
           .
 -.032464337
    .0700392
           .

  -.15242745
-.0008014426
end

Thanks a lot!

[chris chan]

Oh, i missed one more thing. I came across a research paper stated: " estimated a probit model including year fixed effect". I couldnt find a related command and some suggested probit model cannot have time dummies or time fixed effect.