Spatial Demand Decisions in the Pacific Northwest

Abstract

Individual transportation demand decisions form the basis of aggregated demand models that the U.S. Army Corps of Engineers (ACE) uses to evaluate infrastructure improvements. The treatment by ACE planning models has been criticized on the grounds that the demand expressions are inappropriate. Specifically, the demand models used are perfectly inelastic (to a threshold point) and ignore the spatial market setting in which transportation decisions are made. This paper reports the results of a survey of shippers located on and off the Columbia–Snake waterway in the Pacific Northwest to estimate shipper-level demand decisions. The model addresses the concerns about the current ACE planning models by estimating elasticities rather than assuming them and by allowing the probability of using the river to be explained in terms of the spatial setting of the demander. The survey provides choices made in both revealed and stated preference settings. The choice attributes are rate, transit times, and reliability. The revealed and stated preference data are combined to estimate demand functions. The choice model is then used to simulate the likelihood of using the river as a function of shipment distances. To perform these simulations, rates and transit times are fit to distances. These relations are used with the choice model to explain decisions to use truck–barge as a function of distance to the waterway. The choice model provides statistically significant effects of rate, time, and reliability on mode choice, and the simulation suggests that as distance from the waterway increases, the likelihood of using the waterway falls.