An urban transportation demand model is described that has a number of features available in evaluating the effects of alternative transportation systems. The model is derived directly from the theory of consumer demand in the economic literature. A discussion of the reasoning underlying the model leads to a presentation of the general specification of the model, including a description of the relevant variables, the mathematical form taken by the model, and the statistical techniques used in estimating its parameters. Empirical estimates of model's parameters were obtained using data for the Boston metropolitan areas. Selected results are presented and their implications for transportation investment planning are discussed. The model was estimated by means of constrained multiple regression analysis. This method of estimation consists of estimating parameters by minimizing the sum of squared deviations as with ordinary least squares but performing this minimization while satisfying certain pre-specified conditions derived from a priori information. Constrained regression was used to treat collinearity by imposing appropriate sign constraints on the direct elasticities and cross- elasticities of the system variables. Discussion is presented of the statistical reliability of the estimates.
Abstract