McVay (2006) concludes that managers opportunistically shift core expenses to special items to inflate current core earnings, resulting in a positive relation between unexpected core earnings and income-decreasing special items. However, she further notes that this relation disappears when contemporaneous accruals are dropped from the core earnings expectations model. McVay (2006) calls for research to improve the core earnings expectations model and to provide additional cross-sectional tests of classification shifting. Using a core earnings expectations model that is not dependent on accrual special items, we show that classification shifting is more likely in the fourth quarter than in interim quarters. We also find more evidence of classification shifting when the ability of managers to manipulate accruals appears to be constrained and in meeting a range of earnings benchmarks. Overall, our evidence provides broad support for McVay’s (2006) conclusion that managers engage in classification shifting. Our study also sheds new understanding of the conditions under which managers are more likely to employ classification shifting.
This analysis identifies a distinct immediate announcement period negative relation between earnings announcement surprises and aggregate market returns. Such a relation implies that market participants use earnings information in forming expectations about expected aggregate discount rates and, specifically, that good earnings news is associated with a positive shock to required returns. Consistent with this interpretation we find that Treasury bond rates and implied future inflation expectations respond directly to earnings news. We also find some evidence that the negative relation between earnings news and market return persists beyond the immediate announcement period, suggesting that market participants do not immediately fully impound these future market return implications of aggregate earnings news.
Prior research addressing questions such as whether investors respond to a hypothesized information event used tests of unusual return and/or trading activity as alternative measures of investor response. We investigate which of these two metrics maximizes the likelihood that a researcher correctly detects the presence or absence of a response. Building on the repeated-sample framework established in Brown and Warner (1980, 1985) and Dyckman et al. (1984), we provide evidence that (1) volume-based metrics, especially measures based on numbers of transactions, provide more powerful tests of investor response to public disclosures than do return-based metrics; and (2) supplementing return-based measures with trading-based measures increases the power of tests designed to detect investor response. Our conclusions are particularly relevant when power is critical (i.e., when sample sizes are small or anticipated investor response is small). Our evidence also suggests that before concluding that investors do not respond to a public disclosure, based on a returns analysis, researchers should confirm the nonresponse inference with trading-based measures.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.