Effective Expectations: A Binomial Approach Toward Predicting Auto Registrations

Abstract

Automotive manufacturers sell their products to consumers through dealerships in a successive market structure. Manufacturers have a vested interest in both monitoring and influencing the behavior of the dealers who sell their products, but the methods by which this monitoring is accomplished can be improved by accounting for variation in local markets. Indeed, US courts have found that the current methods used to monitor selling behavior, called sales effectiveness, are unreliable and ill-suited for monitoring automotive dealerships. In this paper, a method is presented which provides predicted automotive registrations of particular manufacturers for local markets by employing a log-odds regression with demographic covariates. The methodology provides a marked improvement from the methods which manufacturers have historically employed.