Knowledge that Transforms

To make high-quality research more accessible and easier to explore.

Fields:
188 results ✕ Clear filters

Relative Forecast Accuracy and the Timing of Earnings Forecast Announcements.

The Accounting Review 1986 61(1), 58-75
Previous studies concerning the relative accuracy of management and financial analyst earnings forecasts have produced conflicting conclusions. Resolving this conflicting evidence is important because of the role relative accuracy potentially plays in such issues as the imposition of mandatory management forecast disclosures, the information content of management versus analyst forecast releases, and the motive of management in providing earnings forecasts to the market. Using weekly consensus (mean) financial analyst earnings forecasts, we document a close association between relative forecast accuracy and the timing of the release of the forecasts. We find that management forecasts issued subsequently to, coincidentally with, and up to four weeks prior to analyst forecasts are significantly more accurate than the analysts' estimates. The consensus analyst forecasts are more accurate beginning the ninth week after the release of the management forecast. Thus, by more closely controlling for the timing of the analysts' forecasts, we are able to provide a more precise picture of the dynamic nature of relative forecast accuracy.

Measurement Error and Statistical Sampling in Auditing: The Potential Effects.

The Accounting Review 1986 61(3), 379-399
The auditing profession, in its use of existing statistical estimators and statistical procedures, has been assuming implicitly that the effect of nonsampling error on these estimators and procedures is negligible. This study investigates the potential effect of one type of nonsampling error, errors of measurement on the part of the auditor, on the behavior of alternative estimators and statistical sampling procedures in a Monte Carlo simulation study. The results indicate that the potential existence of independently occurring measurement error can significantly affect the auditor's statistical results and that the actual warranted level of confidence obtained by a statistical procedure may often be far below the desired level of confidence.