Abstract

The literature to date on estimating government preference functions has followed the development of the multinominal logit model, employing data on discrete choice situations to elicit information on decisionmaking criteria. Most data available to economists on government decisionmaking are in the form of expenditure share allocations. Making use offamiliar utility concepts, this article presents a formal modelfor handling expenditure share data by paying attention to the condition that the share lies between zero and one and to the condition that all shares sum to unity. The model is illustrated with a case of federal government allocations to states for education. * In earlier work McFadden (1975, 1976) examined models for revealing an organization's decisionmaking criteria by examining the decisionmaker's choice when confronted with discrete competing projects. The model envisaged decisionmakers undertaking an implicit benefit-cost comparison of alternatives, so that, after stochastic elements have been accounted for, there is some probability of selecting each one of the possible alternatives. The formal development of this approach led to the now-familiar multinomial logit choice model. McFadden illustrated the model with empirical evidence on the highway routing decisions of the California Division of Highways. In the same spirit Barton (1979) has undertaken an investigation of the FCC's decisions in comparative broadcast licensing cases to assess whether FCC decisions are consistently related to the

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call