David Burgstahler

Earnings Management to Avoid Earnings Decreases and Losses

Anecdotal evidence and recent research suggest there are incentives to avoid earnings decreases and losses. This paper provides systematic evidence that firms increase reported earnings to achieve these goals. Specifically, using all available observations on the annual industrial and research Compustat database for the years 1976-1994, cross-sectional distributions of scaled annual earnings changes and earnings are bell-shaped and fairly smooth except in the regions near zero. For both earnings changes and levels distributions, there is a trough immediately to the left of zero and a peak immediately to the right of zero which are inconsistent with the overall shape of the remainder of the distribution, i.e., the frequencies of small decreases in earnings and small losses are unusually low and the frequencies of small increases in earnings and small positive income are unusually high. Exploring the methods of earnings management, we find evidence that two components of earnings, cash flow from operations and changes in working capital, are used to achieve the manipulation of earnings. Finally, we present two theories about the motivation for avoidance of earnings decreases and losses. We find that the main results of this paper are consistent with both prospect theory and managerial opportunistic behavior motivated by stakeholder use of earnings-related heuristics.

The Importance of Reporting Incentives: Earnings Management in European Private and Public Firms

A mounting arrangement for a plurality of electrical components is disclosed wherein these components may be included in cases which are stacked in sets on a level-by-level basis. These cases include resilient mounting tabs on each end thereof which cooperate with mounting plates and allow a plurality of cases to be stacked in each layer and subsequent additional layers stacked on top of the first layers.

Earnings Management to Avoid Losses and Earnings Decreases: Are Analysts Fooled?*

This paper explores whether analyst forecasts impound the earnings management to avoid losses and small earnings decreases documented in Burgstahler and Dichev 1997, whether analysts are able to identify which specific firms engage in such earnings management, and the implications for significant forecast error anomalies at zero earnings and zero forecast earnings. We use data from Zacks Investment Research 1999 and find that analysts anticipate earnings management to avoid small losses and small earnings decreases. Further, analysts are much more likely to forecast zero earnings than firms are to realize zero earnings, and analysts are unable to consistently identify the specific firms that engage in earnings management to avoid small losses. This latter inability contributes to significant forecast pessimism associated with zero reported earnings and significant forecast optimism associated with zero earnings forecasts.

Do Stock Prices Fully Reflect the Implications of Special Items for Future Earnings

Previous research (Rendleman, Jones, and Latane [1987]; Freeman and Tse [1989]; Bernard and Thomas [1990]; and Ball and Bartov [1996]) indicates that security prices do not fully reflect predictable elements of the relation between current and future quarterly earnings. We investigate whether this finding also holds for the special items component of earnings. Given that special items are prominent in financial analysis and are assumed to have relatively straightforward implications for future earnings (special items are assumed to be largely transitory), one might expect that prices would fully impound the implications of special items for future earnings. Based on the “two-equation” approach used in Ball and Bartov [1996] and other studies (e.g., Abarbanell and Bernard [1992]; Sloan [1996]; Rangan and Sloan [1998]; and Soffer and Lys [1999]), we find that while prices reflect relatively more of the effects of special items compared to other earnings components, we still reject the null hypothesis that prices fully impound the implications of special items for future earnings. The “two-equation” approach assesses the consistency of coefficients in a pair of prediction and pricing equations, and thus depends on an assumed functional form. However, a less structured abnormal returns methodology like that used in Bernard and Thomas [1990] also supports the conclusion that the implications of special items are not fully impounded in prices. Specifically, a trading strategy based only on the sign of special items earns small but statistically significant abnormal returns during a 3-day window four quarters subsequent to the original announcement of special items.

Changes in the probability of bankruptcy and equity value

Abstract Work in finance suggests that changes in the probability of bankruptcy likely affect firm value by affecting perceptions of the cost of transacting with the firm. In this paper, accounting information from the balance sheet as well as the income statement is used in conjunction with Ohlsons (1980) bankruptcy prediction model to calculate unexpected changes in the probability of bankruptcy. For a sample of firms with a large estimated probability of bankruptcy sometime during the ten-year study period, unexpected changes in the probability of bankruptcy are useful in explaining security returns even after controlling for unexpected earnings.

Detecting Contemporaneous Security Market Reactions to a Sequence of Related Events

This paper describes results of a study of the empirical properties of two statistics, which extend security-return-based test statistics used in previous studies of contemporaneous market reactions to events.1 The first, called the H statistic, is a generalization of the Cumulative Average Residual (CAR) statistic. The procedure employed to test the significance of the H statistic extends the generalized least squares approach of Collins, Rozeff, and Salatka [1982] and Collins and Dent [1984]. The second, called the G2 statistic, is a generalization of the idea first introduced by Beaver [1968], who squared residuals in order to detect market reactions even if the average reaction was zero. The procedure we employed to test the significance of the G2 statistic has its closest antecedents in Patell [1976] and in Gonedes [1975].2 We might also note here that Schipper and Thompson [1983; 1985] propose an alternative framework within which tests similar to ours can be conducted. In our study, empirical distributions of H and G2 statistics were generated under the null hypothesis that there was no market reaction.

Earnings Precision and the Relations between Earnings and Returns

Formal and heuristic valuation models suggest that changes in firm value associated with a revision in expected earnings should be a multiple on the order of typical earnings capitalization factors between 10 and 30. However, empirical estimates of the earnings response coefficient (ERC) have usually been in a range an order of magnitude lower, between 1 and 3 (Kothari 2001). This paper uses a simple Bayesian model to integrate previous theoretical and empirical results and bridge this large gap. In the Bayesian model, low precision implies a low weight on new earnings information. As a result, the observable earnings surprise is much larger than the unobservable earnings revision, which in turn implies a smaller coefficient on earnings surprise than on the revision in expected earnings. Although the model uses statistical measures, the Bayesian concept of precision closely parallels the accounting concept of earnings quality, i.e., the extent to which current earnings is informative about expected future earnings. Precision summarizes the effects of multiple determinants of earnings quality.Consistent with the model, two proxies for precision, forecast dispersion and absolute magnitude of earnings surprise, explain a broad empirical range of coefficients on earnings surprise ranging from near 0 up to 30. We reconcile the surprisingly large proportion of observations falling in the upper end of the range with the surprisingly low estimated ERCs typically reported in previous research. Finally, we demonstrate in the appendix how the precision proxies can be used to design tests that focus on the majority of observations with larger surprise coefficients and thereby allow researchers to detect the effect of determinants of surprise coefficients that would otherwise be empirically undetectable.

Size Management by European Private Firms to Minimize Disclosure and Audit Costs

We examine size management by European private firms for which disclosure requirements increase at size thresholds. Our estimates suggest at least 8% of firms near thresholds that impose income statement disclosure manage size downward, and the average firm that manages size sacrifices more than 6% of its assets. We find that multiple determinants of proprietary costs predict this behavior, and that size management to avoid mandatory audits, which are similarly imposed at size thresholds, is of comparable magnitude. Our results triangulate the economic significance of proprietary costs in a setting largely without confounding capital market, agency, or compliance costs.

Disclosure Incentives and Data Availability for Private Firms: Implications for Comparisons of Public and Private Firm Financial Reporting Quality

For private firms, public disclosure of financial information is often at management’s discretion. We argue that differences in incentives to publicly disclose result in differences in data availability for private firms, which in turn can affect the conclusions of studies that rely on private firm data. We examine this point in the context of the literature documenting mixed evidence on differences between public and private firm financial reporting quality (FRQ). Using data surrounding a 2006 regulation change that dramatically strengthened enforcement of public disclosure requirements for German private firms, we compare the FRQ of three groups of firms: Private firms that voluntarily disclose financial statement information (“private voluntary” firms), private firms that disclose only due to effective enforcement of mandatory disclosure requirements (“private mandatory” firms), and public firms that are all subject to mandatory disclosure requirements. We find little or no evidence that private voluntary firms have different FRQ from public firms. However, we find consistent evidence that private mandatory firms have lower FRQ than both public firms and private voluntary firms. The results suggest that understanding the effect of disclosure incentives on private firm data availability is critical to interpreting the results of studies that rely on private firm data.