A note on causality and the structure-performance controversy

Interindustry, cross-sectional studies of structure and performance assume, according to the market power doctrine, that structural variables are exogenous to performance, though this notion has been questioned at different times by a number of industrial organization economists. This study uses a time series approach to test the validity of this notion at the line of business level. Sims' causality tests were performed for three separate firms and the interrelationships over time between the firm's market share and rate of return were estimated (i.e., do changes in a firm's market share occur prior to changes in the firm's rate of return or do changes in the firm's rate of return occur prior to changes in a firm's market share). Two of the three firms examined failed to exhibit the unidirectional causality assumed by the conventional structure-performance paradigm. These results, even though derived from a limited number of undiversified firms, suggest that the issue of causality must be settled before the functional form of the model can be specified.