Bias And Random Measurement Error In Accounting-Based Surrogates For Internal Rate Of Return

Abstract

Limitations inherent in alternative profitability measures as estimates of internal rate of return (IRR) often require that managers and researchers employ accounting-based profitability measures. Using published accounting and stock market data, this study models accounting rate of return (ARR) and conditional estimate of internal rate of return (CIRR) as functions of product market risk; growth (g) inventory cost flow assumption (INV), and depreciation method (DE). The models support inferences about the bias and efficiency (i.e. systematic and random error) in the relationships between the two accounting-based profitability measure and IRR, as estimated by the bias and efficiency in their relationships with a factor that is suggested in the finance literature as a determinant of systematic risk (product market risk). The results indicate that ARR estimates IRR with bias attributable to g; however, ARR is unaffected by INV and DEP. Whether g, INV, or DEP affect CIRRs ability to estimate IRR depends on the interval over which CIRR is estimated and the assumed cash flow pattern. On the other hand, CIRR generally estimates IRR with significantly greater efficiency. These results have research design implications, as well as implications for both accounting policy formulation and anti-trust policies.