Alternative Accounting Procedures, Forecast Accuracy, and Value-Relevance of Earnings in the Oil and Gas Industry
Posted: 19 Jun 2000
Date Written: March 2000
Standard setters and professionals have asserted that the successful efforts (SE) accounting method procedure is superior to the full cost (FC) method. Recent studies suggest that the SE firms exhibit higher market-to-book coefficients (Harris and Ohlson 1987), higher coefficients on earnings change variables (Ayres and Raybum 1991), and greater pooled time-series and cross-sectional earnings response coefficients (Bandyopadhyay 1994) compared to FC firms. Previous studies primarily examine the controversy and the debate during a period of high oil prices and exploration activities which could affect their empirical results. This study reexamines the issue primarily during periods of low oil prices and declining exploration activities, to ascertain whether such different economic conditions affect the observed relationships in previous research.
The research investigates whether the FC accounting method provides more or less accurate earnings predictions than the SE accounting method, an issue not examined in previous literature. The paper also examines a related issue of whether the earnings response coefficients (ERCs) derived by one method are greater than those based on the other accounting method. We find that forecast errors for FC firms are significantly smaller than those for SE firms for one-quarter through at least three-quarters ahead forecast horizons. The smaller forecast errors for FC firms are driven by the periods from 1986 through 1989 and 1991 through 1995, when oil prices and exploration activities were low, and by firms in the oil and gas extraction industry (SIC 13) or firms classified as independent type. The estimated coefficients for the variance of EPS (MSTD, a proxy for risk) suggest significant increases in forecast errors as the variance increase for FC and SE firms. However, while these estimated coefficients remain unchanged for SE firms, the change in the estimated slope coefficient of MSTD for FC firms is negative and statistically significant, during periods of declining prices and explorations. These results suggest that risk for FC firms has less impact on forecast errors than risk for SE firms during such periods. In addition, firm-specific ERCs of all FC firms are larger than those of all SE firms, and these statistically significant differences are robust for extraction firms, but not for other oil and gas firms, across ERC models. The results on ERCs contradict findings presented in previous studies. Implications for standards' setters and regulators are briefly discussed.
JEL Classification: G14, L71, M41, M44
Suggested Citation: Suggested Citation