RELATION BETWEEN EARNINGS RESTATEMENTS AND CONSERVATISM IN THE US

Abstract

This study examines the relation between earnings restatement and the accounting conservatism of the firm. This study is the first extensive study to examine the relation between earnings restatements and accounting conservatism. I investigate earnings restatements announced by U.S. listed firms between 1 January 1990 and 31 December 2005, which were reported in FACTIVA. The results show there is no evidence that a less conservative firm is more likely to restate their earnings. The results hold for both the subsamples of firms which restate their earnings upwards and downwards. Instead, I find evidence that firms which are more conservative are more likely to restate, contrary to my hypothesis. I attribute this to the failure of the conservatism measures to detect conservatism where it is likely to exist, which emphasizes the need for the development of alternative conservatism measures. In addition, I find differences in the sub-samples of firms which restate their earnings downwards and firms which restate their earnings upwards. There is evidence that firms which restate their earnings downwards have higher leverage and have higher anticipated earnings growth compared to non-restatement firms. However, for firms which restate their earnings upward, I find no evidence they have higher leverage and there is limited evidence they have higher anticipated earnings growth. Future studies should differentiate between the two sub-samples in examining earnings restatements, as the results suggest different firm characteristics and motivations for earnings management. Lastly, I find that firms which restate their earnings after being prompted by outside parties such as regulators and auditors are not less conservative than firms which internally detected the need for a restatement. Instead, I find evidence that they are more conservative, reinforcing that the conservatism measures are unable to detect accounting conservatism.