Subject
The relationship between market share and ROI.

Type of Report
Analysis of PIMS data on market share and accounting profits.

Objective
To illustrate a method for using accounting identities in the analysis and interpretation of studies dealing with market share-ROI relationships.

Audience
Academics and managers with an interest in strategic planning and analysis of data containing accounting identities.

Main Points
As statistical analysis of the relationship between market share and profitability has become more complex, there has been a tendency to lose track of the basic accounting definitions and relationships involved. By breaking return on investment into its component parts, we are able to show the influence of certain accounting identities as distinct from causal, antecedent relationships. The procedure that we follow avoids four potentially serious errors in other models purporting to explicate the relationships between market share and profitability:

1. Under-specification: the exclusion of at least one component from the relevant ROI identity.
2. Component-antecedent specification: an inappropriate hybrid of both accounting components (e.g., turnover defined as sales divided by investment) and causal antecedents (e.g., relative price).
3. Functional form misspecification: specifying non-linear accounting as a linear function.
4. Specifying ROI to be linearly related to the inverse of an accounting component (e.g., regressing ROI on turnover despite the fact that, by definition, ROI is the difference between turnover, or sales divided by investment, and total cost divided by investment).

To illustrate both the necessity and the value of sorting out sources of variation in basic accounting identities and relationships before proceeding to look for antecedent causal relationships, we performed some simple statistical analysis on a PIMS sample of 2,124 businesses. The results were both straightforward and provocative, showing that almost all of the variation in return investment associated with market share is due to variations in return on sales (ROS), not turnover (S/I), and that the largest portion of variation in return on sales is associated with the ratio of purchasing expenditures to sales, reconfirming a result from the earliest studies of the PIMS data.

Implications
The implications for researchers and managers are far from trivial. Researchers need to get back to some accounting basics as the first step in specifying models of the market share-profitability connection. Managers need to be extremely cautious in interpreting and applying the published results of such research, especially by taking time to understand potential sources of error in the specification of the underlying model and looking for the confusion of accounting components and antecedent causal variables.

Market share is highly correlated with ROI in the PIMS/database. Understanding this relationship in terms that correspond to plans for increasing or maintaining profitability is important. A change in market share that is expected to lead to a change in profitability must affect the accounting components of profitability. Thus any examination of the relationship must consider the relevant accounting components of profitability.

About the Authors
Paul Farris is Professor of Business Administration at the University of Virginia. Mark Parry is Assistant Professor of Business Administration at the University of Virginia. Fred Webster is the E. B. Osborn Professor of Marketing at Dartmouth College and a former executive director of the Marketing Science Institute.