1989•SSRN Electronic JournalOpen access

The Estimation of Multinomial Probit Models: A New Calibration Algorithm

Wagner A. Kamakura

Abstract

This study proposes the estimation of Multinomial Probit models using Mendell-Elston's approximation to the cumulative multivariate normal for the computation of choice probabilities. The accuracy of this numerical approximation in computing probabilities is compared with other procedures used in existing calibration programs. Finally, the proposed estimation procedure is tested on simulated choice data.

About this research paper

What this paper is about

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Key concepts: Multinomial probit, Multinomial distribution, Computation, Multivariate probit model, Calibration, Estimation, Probit, Econometrics