پیش بینی های مدیریت درآمد و اظهارات آینده نگرانه
Abstract
1- Introduction
2- Prior literature and hypothesis development
3- Sample selection, measurement, and validation
4- Empirical results
5- Summary
Reference
Abstract
We identify forward-looking statements (FLS) in firms’ disclosures to distinguish between “forecast-like” (quantitative statements about earnings) and “other”, or non-forecast-like, FLS. We show that, like earnings forecasts, other FLS generate significant investor and analyst responses. Unlike earnings forecasts, other FLS are issued more frequently when uncertainty is higher. We then show that earnings-related FLS are more sensitive to uncertainty than quantitative statements, suggesting that managers are more likely to alter the content than the form of FLS when uncertainty is higher. Our study indicates that incorporating other FLS into empirical measures provides a more comprehensive proxy for firms’ voluntary disclosures.
Introduction
For several decades, management earnings forecasts have been the focus of significant academic interest. Initially, that interest was driven by the concern that forecasts were not credible – so much so that firms were prohibited by the SEC from providing such forecasts in their securities filings.1 Following substantial evidence that investors do respond to earnings forecasts (e.g., Patell, 1976; Waymire, 1984; Baginski et al., 1994; Hutton et al., 2003; Rogers and Stocken, 2005), that view has changed considerably. In fact, more recent studies have shown that management forecasts provide more information to investors than any other accounting source (Beyer et al., 2010). A common theme among these studies is the focus on quantitative earnings forecasts. In this paper, we use textual analysis to study forward-looking information disclosed by managers. We examine both the quantitative estimates of future earnings that are typically treated as forecasts in disclosure research, and the set of qualitative and/or non-earnings forward-looking statements often overlooked in disclosure research.2,3 To distinguish between different types of forward-looking statements, we identify, for each statement, whether the statement is quantitative or non-quantitative and whether it explicitly refers to earnings or not. We refer to forward-looking statements as “forecast-like” if they are both quantitative and refer explicitly to earnings. We refer to all remaining forward-looking statements as “other”, or “non-forecast-like”, forward-looking statements.4 Isolating the different types of forward-looking statements, and when managers issue those statements, allows us to investigate why managers make the disclosure choices we observe. While several studies have examined certain types of non-earnings forward-looking statements individually (e.g., revenue or cash flow projections), our goal is to study a comprehensive sample of forward-looking statements (both forecast-like and other) to assess how and why forecast-like statements differ from other forward-looking statements in terms of both determinants and consequences.