A Descriptive Alternative to Cluster Analysis: Understanding Strategic Group Performance with Simulated Annealing

Abstract

Fixed effects models featuring cluster analytical techniques underlie much of the empirical research on strategic groups and their significance in explaining performance differences among competing firms. We propose an alternative “descriptive” approach for analyzing strategic groups and their impact on firms, particularly in multi-industry research settings. Our descriptive alternative utilizes a random effects model and simulated annealing in order to avoid some of the limitations of cluster analytical approaches as well as to provide insight on the relative importance of strategic group-level factors on variance in firm returns. We demonstrate our descriptive approach with simulated annealing of data from the Federal Trade Commission’s Line of Business Program. The demonstration suggests that strategic group-level factors are important for explaining persistent differences in firm performance, but that such factors are fluid and ephemeral, rather than fixed and durable over time.

The authors thank Bernardo Prigge and Ray Willis of the University of Minnesota, and Curtis Wagner and George Pascoe of the US Federal Trade Commission for their assistance. We also thank various participants at the Conference on Statistical Models for Strategic Management, held at Nice, France, for their helpful comments and suggestions in improving the paper. The views presented here are those of the authors and not necessarily those of the US Federal Trade Commission. A review has been conducted to ensure that the data analyzed and presented herein do not identify any individual company’s line-of-business data.