Spatial Econometric Modeling of Origin‐destination Flows*

Abstract

Standard spatial autoregressive models rely on spatial weight structures constructed to model dependence among n regions. Ways of parsimoniously modeling the connectivity among the sample of N = n 2 origin-destination (OD) pairs that arise in a closed system of interregional flows has remained a stumbling block. We overcome this problem by proposing spatial weight structures that model dependence among the N OD pairs in a fashion consistent with standard spatial autoregressive models. This results in a family of spatial OD models introduced here that represent an extension of the spatial regression models described in Anselin (1988).

Incorporating Transportation Network Structure in Spatial Econometric Models of Commodity Flows

“…LeSage and Pace (2005) provide a parsimonious way to structure the connectivity of origindestination regions in a fashion consistent with conventional spatial autoregressive models where each observation represents a region rather than an origin-destination pair. This seems to have been the stumbling block to extending conventional spatial econometric methods to origin-destination flow situations.…”

“…The focus of this study is on a formal method for adjusting the spatial weights introduced by LeSage and Pace (2005) to reflect a general dependence structure between origin and destination regions that incorporates the nature of the transport network infrastructure. We are also interested in whether this type of adjustment will improve the estimates, inferences and predictions of the model.…”

“…We note that numerous other flows such as telecommunication, airline passengers, train travel and shipping, and automobile and truck traffic are also heavily dependent on the transport network infrastructure used. LeSage and Pace (2005) make the intuitively plausible argument that: 1) large commodity flows from region A (origin) to region Z (destination) might be accompanied by similarly large flows from neighbors to region A to region Z; 2) large commodity flows from region A to region Z might be accompanied by similarly large flows from region A to neighbors to region Z; and 3) large commodity flows from region A to region Z might be accompanied by large flows from neighbors to region A to neighbors of region Z.…”

“…Our approach would accommodate either type of network, focusing only on the presence or absence of a network route in the regions under study. Following LeSage and Pace (2005), our methodology allows for three types of spatial/network connectivity between origin and destination regions. We overlay information regarding the network structure and regions, providing an extension of the LeSage and Pace (2005) methodology.…”

“…The model accounts for spatial dependence in the origin-destination flows by introducing a spatial connectivity matrix that allows for three types of spatial dependence in the origins to destinations flows. We modify this origin-destination connectivity structure that was introduced by LeSage and Pace (2005) to include information regarding the presence or absence of a major highway/train corridor that passes through the regions. Empirical estimates indicate that the strongest spatial autoregressive effects arise when both origin and destination regions have neighboring regions located on the highway network.…”