Application Of Maximum Likelihood To A Bivariate Two-Limit Tobit Model For Estimation Of Rural Retail Sales Potential

A bivariate likelihood function for a two-limit tobit model is developed and applied to the estimation of rural resident shopping patterns. Comparison of results from the univariate and bivariate models shows that parameters are estimated with more precision, and derivatives of conditional expectations are larger under the bivariate procedure. Results suggest rural commercial sector development plans should address a broad array of retail stores in order to capture the interdependencies of retail purchases.