The Accuracy of Cash Flow Estimation Procedures

Abstract

Prior research uses mechanical procedures to estimate cash flow data. This study examines the accuracy of these procedures by measuring errors between estimated cash flows and reported cash flows. The results indicate that mechanical rules provide poor estimates for reported cash flows. We also show that the errors between cash flow estimates and reported cash flows can be reduced by adjustments made from footnote disclosures. However, large errors remain, even after adjustments are made from footnote disclosures. These remaining errors are correlated with firm specific characteristics such as sales (firm size), extraordinary items, foreign currency gains and losses, and changes in inventory.

Direct or Indirect Cash Flow Statements?

“…Austin and Bradbury (1995) show that mechanical procedures for estimating cash flow data provide poor estimates of reported cash flow numbers (i.e., cash flow from operations and direct cash flow components). For example, the error in cash from customers (payments to suppliers) is, on average 5% (6%) but ranges from 0% to 71% (0% to 75%).…”

An analysis of the reasons for the asymmetries surrounding earnings benchmarks

“…receivables, inventory, payables, and other current items). We use cash-flow statement data to minimize the estimation error in direct cash-flow accruals (Austin and Bradbury, 1995). In our ratios analysis, we incorporate cash-from-customers and payments-to-suppliers line items from the direct cash-flow statement.…”

“…This evidence suggests that firms above the benchmark are taking real actions by generating more cash from customers rather than managing accruals. 22 Austin and Bradbury (1995) find that 'other current items' have a significant impact on estimates of operating cash flow and payments-to-suppliers in particular. 23 Taking first differences before scaling reduces the influence of scaling when more than half of the observations for these items are zero.…”

“…4 This resulted in 1471 observations over the period 1991-2002. For missing data, we prefer not to estimate cash-flow numbers because Austin and Bradbury (1995) demonstrate that mechanical procedures to construct cash-flow numbers, typically derived by using the change in balance sheet data to reverse the accruals in earnings, provide poor estimates of reported cash-flow numbers. Furthermore, they show that the errors from constructed cash-flow numbers are correlated with firm-specific attributes (e.g.…”

“…The direct cash-flow statement also enables us to use line items such as cash-from-customers and payments-to-suppliers to extend the number of financial ratios and the number of potential explanations for earnings management. 2 Furthermore, the use of reported cash-flow statement data minimizes the errors in cash-flow data and therefore the estimates of accruals (Austin and Bradbury, 1995;Hribar and Collins, 2002).…”