Model exit in a vertically differentiated market: Interfirm competition versus intrafirm cannibalization in the computer hard disk drive industry

Considerable controversy exists about the appropriate way hospital competition should be measured and whether findings are accurate if certain methods are employed. Data from the Healthcare Cost and Utilization Project (HCUP), the American Hospital Association (AHA), and other supplemental data sources are used to create and evaluate hospital competition measures. Correlation coefficients of these measures are assessed. Moreover, each measure is independently included as an explanatory variable in otherwise identical hospital cost function regressions. Their corresponding parameter estimates are then compared. Most measures are highly correlated. Inferences about the effect of competition on hospital cost remain the same when alternative hospital competition measures are employed. We caution researchers against using this finding to arbitrarily select a competition measure when the magnitude of the estimates is important.