A New Approach for Modeling Economic Count Data Forthcoming : Economics Letters

A new parametric model for the econometric analysis of non-negative integers is proposed. Its distinguishing feature is that it allows for more flexible variance-mean relationships than the models used hitherto. Estimation with maximum likelihood is illustrated using a dataset on ship damage incidents. A New Approach for Modeling Economic Count Data Forthcoming: Economics Letters Rainer Winkelmann University of Munich Klaus F. Zimmermann University of Munich and CEPR July 1991 Abstract A new parametric model for the econometric analysis of non-negative integers is proposed. Its distinguishing feature is that it allows for more flexible variance-mean relationships than the models used hitherto. Estimation with maximum likelihood is illustrated using a dataset on ship damage incidents.A new parametric model for the econometric analysis of non-negative integers is proposed. Its distinguishing feature is that it allows for more flexible variance-mean relationships than the models used hitherto. Estimation with maximum likelihood is illustrated using a dataset on ship damage incidents. *Financial support from the German Science Foundation (DFG) is gratefully acknowledged. Correspondence to: Prof. Dr. Klaus F. Zimmermann SELAPO, University of Munich Ludwigstr. 28 RG D-8000 München 22 Germany