Print
E-MailThe following post comes to us from Qiang Cheng, Professor of Accounting at Singapore Management University; Ting Luo of the Department of Accounting at Tsinghua University; and Heng Yue, Professor of Accounting at Peking University.
In our paper, Managerial Incentives and Management Forecast Precision, forthcoming in The Accounting Review, we focus on one important characteristic of management forecasts—forecast precision—and examine how managerial incentives affect the choice of forecast precision. We choose to focus on forecast precision (or specificity, as it is sometimes referred to in the literature) for two reasons. First, precision is one of the most important forecast characteristics over which managers have a great deal of discretion. Managers can issue qualitative or quantitative forecasts, and the latter may take the form of point forecasts, range forecasts, or open-ended forecasts. More than 80% of the quantitative forecasts compiled by Thomas Financial are in the range format (i.e., estimates with explicit upper and lower bounds), and there is a large degree of variation in forecast width (i.e., the difference between the upper and lower bounds). One might even argue that managers have greater discretion over the precision of their earnings forecasts than over whether to provide forecasts in the first place (Hirst et al. 2008). Managers cannot always withhold information because it is part of their fiduciary duty to update and correct previous disclosures. Furthermore, withholding information can lead to considerable litigation risks and can cause great damage to a manager’s reputation (Skinner 1994). Second, forecast precision has a significant effect on market reactions to management forecasts. A number of theoretical papers, such as Kim and Verrecchia (1991) and Subramanyam (1996), argue that the magnitude of the market reaction to a disclosure is positively related to its precision, and empirical studies examining the impact of management forecast precision on stock returns and analyst forecast revisions provide support for this argument (e.g., Baginski et al. 1993; Baginski et al. 2007).
Building on prior research, we identify the most frequently investigated managerial incentive in the voluntary disclosure literature, insider trading, and examine whether it provides managers with incentives to choose forecast precision strategically. Given that the precision of management forecasts has a significant effect on stock prices—more precise forecasts have a larger impact on stock prices than vague forecasts—we argue that trading incentives affect forecast precision and that the effect depends on both the sign and magnitude of the news. As managers prefer a higher stock price prior to insider sales, we predict that good news disclosed before insider sales is more precise, and that the more positive the news is, the more precise the forecast is. Similarly, we hypothesize that bad news disclosed before insider sales is less precise, and that the more negative the news is, the more vague the forecast. In other words, we predict a positive association between forecast news and the precision for management forecasts issued before insider sales. Given that prior research finds that there is, on average, a positive association between forecast news and precision (e.g., Skinner 1994; Choi et al. 2010), these arguments imply that the association is more positive for management forecasts issued before insider sales than for those not followed by insider trading. In contrast, as managers benefit from a lower stock price before insider purchases, we predict the opposite for management forecasts disclosed before insider purchases, i.e., a less positive association between forecast news and the precision for management forecasts issued before insider purchases than for those not followed by insider trading.
To test our hypotheses, we examine a sample of 10,799 management earnings forecasts issued in the 1999-2006 period. We use the negative of forecast width (the magnitude of the range for range forecasts and zero for point forecasts) to measure forecast precision. To test our predictions, we regress forecast precision on forecast news, trading incentives (indicators for insider sales or purchases), and their interactions. We also control for other determinants of forecast precision, such as managers’ information uncertainty, market demand for information, the passage of Regulation Fair Disclosure, equity issuance, and the precision of past management forecasts.
Consistent with prior research, we find an overall positive association between forecast news and forecast precision; the more positive the news is, the more precise the forecast is. More importantly, we find that, consistent with our hypotheses, trading incentives systematically affect the association between forecast news and precision. We find that forecast precision is more positively correlated with forecast news for management forecasts issued before insider sales than for other management forecasts. For those issued before insider purchases, we find a less positive correlation between forecast news and precision. To highlight the notion that the direction of insider trading’s effect on forecast precision depends on the sign of the news, in an additional analysis we replace the continuous forecast news variable with indicators for the sign of the news. We find that compared with management forecasts issued at other times, good news issued before insider sales is more precise and bad news before insider sales is less precise, whereas good news issued before insider purchases is less precise and bad news before insider purchases is more precise. Overall, these results indicate that managers choose to issue forecasts in a form that increases these forecasts’ impact if that impact is desirable and reduces it if it is undesirable.
To obtain further support for our main inferences and to provide additional insights, we also examine three conditioning variables that can affect the relation between managerial incentives and forecast precision. First, previous research shows that institutional investors play an important monitoring role and demand more transparent disclosure than individual investors (e.g., Bushee 1998; Ajinkya et al. 2005; Chen et al. 2007). If this is the case, then managers’ strategic behavior is likely to be mitigated by the presence of institutional investors. Consistent with this prediction, we find that the effect of trading incentives on the association between forecast news and precision is weaker when institutional ownership is high than when it is low.
Second, while the risk of strategically changing forecasts precision is lower than the risk associated with other forms of managerial discretion in the voluntary disclosure domain, such as withholding news, it is not risk-free and the extent of the risk varies. We argue that the strategic decision on forecast precision is associated with a higher degree of risk for good news forecasts issued before insider sales and for bad news forecasts issued before insider purchases, because managers have incentives to increase the precision of forecasts in these two scenarios, thus leading to a greater likelihood of forecasts being proven wrong (i.e., a greater likelihood of actual earnings falling outside the forecast range). In contrast, the strategic decision on forecast precision is associated with lower risk for bad news forecasts issued before insider sales and for good news forecasts issued before insider purchases, because managers have incentives to decrease forecast precision in these two scenarios, thereby leading to a lower likelihood of forecasts being proven incorrect. Hence, we posit that managerial incentives are less likely to affect forecast precision in the cases of good news preceding insider sales and bad news preceding insider purchases than in the cases of bad news preceding insider sales and good news preceding insider purchases. Our results are consistent with this prediction.
Third, managers’ ability to choose forecast precision for self-serving purposes depends on investors’ ability to assess the precision of managers’ information. If investors are able to “see through” the precision game and react accordingly, then strategically choosing forecast precision will not benefit managers. Thus, we expect that managers are more likely to strategically choose forecast precision when investors have greater difficulty in evaluating the precision of their information. Using several variables to capture the level of this difficulty, we find results consistent with our prediction.
We also conduct several additional tests to enrich our analyses and to ensure the robustness of our results. First, we validate the assumption that precise forecasts are associated with stronger market reactions than vague forecasts. Second, we find that our results are not driven by the reverse causality (i.e., disclosure precision affecting the existence of insider trading) or self-selection in the issuance of management forecasts. Third, we find that our results are robust to alternative research design choices, such as using the magnitude of insider trading rather than indicators for such trading, and to controlling for the effect of contemporaneous earnings announcements for bundled management forecasts.
The full paper is available for download here.
