This paper examines bond-for-bond refundings and their effects on stock returns. Refundings can affect the reported income, cash flows (including taxes), dividend constraints and financial ratios of firms. For a sample of 36 NYSE and ASE firms that performed refundings between 1971 and 1980, stock returns were significantly higher than predicted (only) around the release of the quarterly earnings announcement that included the refunding's effects. While the refundings were found to have many characteristics that were hypothesized to benefit shareholders, only the change in earnings per share was found to be associated with the prediction errors. Further, there appears to be no refunding-related information released in the quarterly earnings announcement, except for the refunding gain. These results imply that a portfolio of refunding firms can be created in advance of the quarterly earnings announcement that will generate abnormal returns around the earnings announcement. Because trading rules are inconsistent with the concept of an efficient capital market, these results constitute an anomaly.

We investigate the reliability of mandatory annual fair value estimates for UK investment property. We find that appraisal estimates understate actual selling prices and are considerably less biased and more accurate measures of selling price than respective historical costs. Investigations of managerial discretion over fair value reporting reveal that managers select among permissible accounting methods to report higher earnings, time asset sales to smooth reported earnings changes, smooth reported net asset changes and boost fair values prior to raising new debt. Finally, we find that the reliability of appraisal estimates increases when monitored by external appraisers and Big 6 auditors.