John Sutton

Published January 2004

---

This paper introduces a novel analysis of the classic ?persistence of leadership? question, and applies it to a newly constructed dataset for Japanese manufacturing. The analysis rests on an appeal to an empirical ?scaling relationship? between current market share and the variance of changes in market share. This relationship provides a powerful ?model selection criterion? for candidate models of market share dynamics. It also makes it feasible, even in small datasets, to test directly for the properties of the ?first passage times? corresponding to loss of leadership.