Abstract:
We note that the existence of the maximum likelihood estimates for Poisson regressiondepends on the data configuration. Because standard software does not check for thisproblem, the practitioner may be surprised to find that in some applications estimation of thePoisson regression is unusually difficult or even impossible. More seriously, the estimationalgorithm may lead to spurious maximum likelihood estimates. We identify the signs of thenon-existence of the maximum likelihood estimates and propose a simple empirical strategyto single out the regressors causing this type of identification failure.