Warehouse rental market segmentation using spatial profile regression

Abstract

Warehouse rental markets can be segmented into multiple submarkets in which rental storage units share similar structural characteristics in that they are reasonably close substitutes for one another with their geographical proximity. Improved understanding of warehouse rental market segmentation enables warehouse owners to effectively formulate marketing strategies and warehouse renters to reduce search costs. Previous studies either assumed a priori submarkets or used cluster analysis to delineate submarkets based entirely on the similarity of rental prices. However, such approaches have limitations in addressing associations between warehouse rents and their determinants because of the potential spatial autocorrelation and multicollinearity in warehouse rent data sets. In the present study, we address the gap in the literature by introducing a method known as Bayesian spatial profile regression for warehouse rental submarket segmentation. This approach allows us to assess meaningful relationships between warehouse rents and their determinants as a unique profile for each submarket, while accounting for spatial autocorrelation in warehouse rents and multicollinearity among their determinants. In a case study, we demonstrated an application of spatial profile regression to a warehouse rent data set for the Seoul Metropolitan Area (SMA) of South Korea and identified two submarkets: high-rent and low-rent groups. The high-rent group was strongly associated with proximity to the urban center in Seoul and Incheon Port, higher floor area ratio, relatively older building age, higher land price, transportation, and automated warehousing services. The associations for the low-rent group were the opposite of the high-rent group and featured proximity to industrial complexes away from the urban center. The results reflected the highly polarized segmentation of the warehouse rental market in the SMA.