A Nearest-Neighbor Spatial-Association Measure for the Analysis of Firm Interdependence

Retail locational factors and strategies have been a focus of research among geographers, location analysts, marketing analysts, and regional scientists. Well known in the literature are the cumulative-attraction and interceptor-retail-location strategies. To analyze these strategies or forces of firm interdependence, one approach taken has been to describe the spatial pattern of the firms by certain probability models and the nearest-neighbor statistics. However, when studying the spatial association between retail firms from two different chains or between two different types of firms, the familiar Clark and Evans nearest-neighbor statistic no longer applies. What is needed is a method that is capable of measuring the spatial association between two distributions of points. More specifically, the method should determine whether the spatial association between two groups of points is one of aggregation, independence, or avoidance, as a result of the location strategy employed. The nearest-neighbor method developed in this paper for such a task is a modification of that of Clark and Evans. The probability distribution, the mean, and the variance of the nearest-neighbor statistic are derived, and several empirical analyses of the retail-location strategies among convenience-food stores are provided.