STATISTICAL CAUSALITY AND STRATEGIC BEHAVIOR IN INDUSTRIAL-MARKETS

This paper proposes a simple test of the leader-follower model of stra tegic behavior. This test relates the temporal notions of leadership c entral to such models to the empirical methods of statistical causalit y. This test is performed using data from the US softwood plywood indu stry of the last three decades. Others have productively explored the spatial pricing practices of this industry by applying a leader-follow er model. Similarly, we find that a leader-follower model explains wel l the temporal relations between key strategic variables (prices) in t he industry. We conclude that the leader-follower model imposes meanin gful restrictions on observable time-series data and that statistical causality is a useful method for testing these restrictions.