Duke Economics Working Paper #98-10
A new class of regression type models termed essentially linear models is proposed. The class is characterized by geometric considerations. Within the class the distribution of the maximum likelihood estimator is easily approximated by a natural extension of the pstar-formula even though the MLE itself is not typically achievable in closed form. The method has many econometric applications and works well in practice.
Key words: Likelihood, Sufficiency, Statistical geometry, Zipf-Estoup regression, pstar-formula
JEL: C10, C20, C51
Retrieve document:
24 pages