LAUSR

Purpose The aim of this paper is twofold. First, clustering patterns of urban hotels are explored, and, second, clustering effects on performance for upscale urban hotels are estimated. Design/methodology/approach Local indicators of spatial association (LISA) were computed using geographic information system (GIS) techniques. Clustering for the entire population of hotels in Madrid was explored visualizing LISA statistics. Then, a system generalized method of moments regression was applied to test a set of hypotheses about the performance effects of LISA statistics for a sample of upscale urban hotels. Findings Two significantly distinct types of clusters are identified: dense “cold spots” or clusters containing many low-priced hotels and quiet “hot spots” or clusters only containing a few high-priced hotels. And, estimates confirmed two important results: evidence of adverse selection when clustering and evidence of positive location economies for upscale hotels. Practical implications This study has a number of relevant implications for making better hotel location decisions. Specifically, the paper shows the applicability of GIS to find statistically significant clustering in the data. In the hotel sector, knowing exactly where hotel clustering occurs and of what type is of vital importance. Originality/value This paper’s novel application of LISA based on GIS techniques for hotel clustering sheds light on the effects of clustering on performance to convey the subtle nuances of the relationship for upscale urban hotels.