Abstract
This short note drows attention to an obvious although so far only partly noted relation between an inequality involving the hazard rate of the normal distribution and maximum likelihood estimation in the binary probit and in the Tobit model. Global concavity of the likelihood function for both the binary probit and the Tobit model can be proved by means of an inequality concerning the hazard rate of the standard normal distribution. As not yet noted in the literature this inequality may also be used to show that the hazard rate is monotonically increasing.
Similar content being viewed by others
References
Amemiya, T. 1985Advanced Econometrics. Oxford: Basil Blackwell.
Barlow, R.E. and Proschan, F. 1975. Statistical Theory of Reliability and Life Testing. Probability Models. New York: Holt, Rhinehart and Winston.
Daganzo, C. (1979).Multinomial Probit. The Theory and Its Aplication to Demand Forecasting. New York: Academic Press.
Kaufmann, H. (1988). “On Existence and Uniqueness of Maximum Likelihood Estimates in Quantal and Ordinal Response Models”.Metrika 35, 291–313.
Maddala, G.S. 1983.Limited-Dependent and Qualitative Variables in Econometrics. Cambridge: Cambridge University Press.
Olson, R.J. (1978). “Note on the Uniqueness of the Maximum Likelihood Estimator for the Tobit Model”.Econometrica 46, 1211–1215.
Tobin, J. (1958). “Estimation of Relationships for Limited Dependent Variables”.Econometrica 26, 24–36.
Author information
Authors and Affiliations
Additional information
Results of this paper are related to research which is financially supported by Thyssen-Stiftung. Comments on an earlier version by an anonymous referee are gratefully acknowledged.
Rights and permissions
About this article
Cite this article
Ronning, G. Binary probit, tobit and hazard rate. Statistical Papers 31, 291–294 (1990). https://doi.org/10.1007/BF02924702
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02924702