Temporal Variation in Consumer Spatial Behavior in Shopping Streets

Abstract

Modeling consumer behavior in shopping streets to support urban planning decisions has a long tradition in a variety of disciplines. Surprisingly, temporal variation in such a shopping process has received only scant attention in these models. This paper therefore proposes a multinomial logit model, in which real time is part of the utility function, implying that temporal variation in consumer preference structures can be analyzed. A grid-based estimation procedure is developed to estimate the unknown temporal variability from consumer choice data. The goodness-of-fit of the estimated model is very satisfactory and most parameters are statistically significant and in anticipated directions. Time-varying effects on consumer behavior are shown to be fairly strong. Some factors are weakened over time while others are strengthened. This finding suggests the existence of systematic variations in consumer behavior, which could enable retailers, urban planners, and real estate developers to evaluate their plans or to make decisions in more specific spatio-temporal contexts.