Assessing the Accuracy of Cash Flow Models: The Significance of Payment Conditions

Cash flow forecasting methods have evolved to allow detailed predictions for individual projects. These methods, principally the cost-schedule integration (CSI) technique, make extensive use of project estimate and schedule data. An implicit assumption of these methods has been that accuracy is largely a function of the quality of data available to the model. To the writers' knowledge, there has been no assessment of the ability of project specific cash-flow models to accurately predict cash flows given accurate input data. This paper makes two contributions. First, two complementary methods are presented—pattern matching logic and factorial analysis—that provide an ability to assess the accuracy of cash flow models. Second, through demonstration of these methods using data from two projects, a critique is made of the ability of existing CSI models to accurately predict cash flows. The paper concludes by recommending extensions of CSI models to include more detailed payment conditions, including differential payment lags, components for materials and labor, and payment frequency. A further conclusion is the call for more research to better understand the balance between managers' need for information and the ability of predictive models to provide that information.