Identifying dynamics in strategic groups

Abstract

Purpose The purpose of this paper is to characterize the situation in the Spanish banking industry through the identification of strategic groups based on a set of variables. Design/methodology/approach To do so, the authors use a 13-year data set and a time inhomogeneous hidden Markov model (HMM) in which the time variable transition matrix captures institutions’ group switching behavior to identify these strategic groups. In fact, the authors consider a mixture model is the data generation process. Findings Two groups are identified. These groups are primarily characterized by size and other strategic variables. The probability of remaining in a group is generally high: 87.28 per cent for SG1 and 61.84 per cent for SG2. The probability of switching groups is low: 12.72 per cent probability of switching from SG1 to SG2 and 38.16 per cent probability of switching from SG2 to SG1. Banks in SG1 seem more stable over time; they have low levels of switching behavior and well-defined long-term behavior. Banks in SG2 seem to evolve in terms of group membership. Originality/value Using an inhomogeneous HMM with time-variable transition matrix, this paper allows for time-varying parameters in the distributions to analyze the evolution of strategic group membership in this industry to detect changes in group strategy, changes in membership and the stability of groups over time.