Hello,

I am missing running a series of models using eprobit, entreat in Stata 16.1. My model is:
Code:
eprobit recidivism $covariates $years if $restriction1 & $restriction2, entreat(eviction = i.instrument $covariates $years) vce(cluster household)


At the bottom of the output, Stata lists a coefficient and related stats (Std error, z, p, and ci) for the correlation of the error terms of the dependent variable and the independent variable: corr(e.evict, e.recidivism).

For some, but not all models, the coefficient is present, but the related statistics are missing. There are simply dots where the values should be. In most models, the correlation is extremely high.
Code:
corr(e.evict, e.recidivism) | -.9999715 |   .   |   .   |   .   |   .   |   .   |


I have read through the ERM documentation, and I cannot find a reason why this information would be missing. Based upon the output from the models that include these stats and from non-interacted versions of the model (eteffects and eprobit, entreat, nointeract) I am concluding that the correlation is statistically significant, though the coefficient seems very high. Does anyone know why this this might be happening?

Thank you!