##plugins.themes.bootstrap3.article.main##

This paper examines whether revenue surprises can affect management’s incentive to manipulate their reported earnings to meet or exceed analysts’ earnings forecasts. In particular, it tests whether managers would be more likely to manage earnings upward to achieve earnings expectations depending on the signal of revenue surprises. The empirical analysis using the multinomial logit model shows that managers have the higher likelihood of earnings management to have positive earnings surprises under positive revenue surprises. Further analysis considering all possible situations (R+E+, R+E-, R-E+, R-E-) provides evidence that firms have the strongest incentives to increase their earnings to exceed analysts’ earnings forecasts when they have negative unexpected revenue. These results are consistent with prior research that the major purpose of aligning earnings with market expectations is to avoid asymmetrical negative market responses associated with missing the expected earnings.

Downloads

Download data is not yet available.

Introduction

Previous papers in accounting and literature have provided evidence that a firm’s management kept a very close watch on earnings forecasts (Brown, 2001; Burgstahler & Eames, 2006). In addition, some studies have investigated the market reactions associated with meeting or beating earnings expectations with somewhat contradictory results. The market clearly rewards firms that meet or beat analysts’ earnings forecasts (Bartovet al., 2002; Kasznik & McNichols, 2002) while firms’ small negative earnings surprises are associated with significant stock price declines (Skinner & Sloan, 2002; Kinneyet al., 2002). Furthermore, Matsumoto (2002) and Burgstahler and Eames (2006) provided evidence of purposeful earnings manipulation to avoid missing earnings expectations, which would lead to market punishments. In addition, the reaction of market participants to earnings surprises can vary depending on revenue information, as revenue surprises provide additional insights beyond earnings surprises regarding future profitability (Jegadeesh & Livnat, 2006; Rees & Sivaramakrishnan, 2007). Rees and Sivaramakrishnan (2007) provided evidence that a market response that rewards (penalizes) meeting or exceeding (missing) earnings expectations exists only when the direction of revenue signals is the same as the direction of earnings signals. Hence, this paper investigates more detail of managers’ incentives to meet earnings targets conditioned on whether firms meet or beat revenue expectations.

I focus on management’s incentive to have positive earnings shocks, given positive revenue surprises. Because the market provides larger equity premiums as rewards for meeting both revenue and earnings targets, managers may have strong inclinations to achieve both objectives through upward earnings manipulation. Consistent with our conjecture, results show that given a positive revenue signal, firms are more likely to use more positive discretionary accruals to have positive earnings. Furthermore, the extended analysis is conducted by investigating all possible conditions (R+E+, R+E-, R-E+, R-E-) under which firms’ managers have the strongest incentives to increase reported earnings to meet or exceed analysts’ earnings forecasts. I posit that managers would have greater incentives to manage earnings upward for achieving market expectations of earnings when firms have negative revenue surprises. The further analysis results indicate that firms which meet or beat the earnings targets but fail to meet revenue targets (R-E+) have higher positive discretionary accruals compared to other situations (R+E+, R+E-, R-E-). These findings suggest that firms which have negative revenue surprises (R-E+) are more likely to meet or beat analysts’ earnings forecasts by income-increasing earnings management relative to other situations.

The remainder of the paper is organized as follows. Section 2 discusses the related literature. Section 3 shows hypotheses development. Section 4 provides a brief discussion of the research design. The results from the empirical analysis are reported in section 5. Finally, the conclusion is in section 6.

Related Literature

Earnings Management to Meet or Beat the Market Expectations of Earnings

Prior work has presented an interesting phenomenon: an abnormally large number of firms are meeting or beating market expectations for earnings compared to missing them. Brown (2001) and Burgstahler and Eames (2006) found that there has been an unexpectedly large number of firms which have slightly met or beaten analysts’ earnings forecasts. As the reasons for this phenomenon, Bartovet al. (2002) show the existence of higher market equity premiums in firms that meet or beat analysts’ earnings forecasts than in firms that fail to meet them. Additionally, Kasznik and McNihols (2002) showed that the market clearly rewards firms that meet expected earnings. Furthermore, the market gives more rewards to firms which consistently beat expected earnings (Rees & Lopez 2002). On the contrary, several papers have provided evidence that the market gives more weight to bad earnings surprises. Conradet al. (2002) found disproportionally large declines in stock prices for failure to achieve positive earnings surprises. Skinner and Sloan (2002) also showed the existence of the “torpedo effect,” in which even small negative earnings shocks are associated with significantly large stock price declines. Consistent with prior findings, Dechowet al. (2006) presented that investors appear to react slightly more negatively to firms missing analysts’ earnings expectations, even though the difference may only be significant at the 11% level. They additionally found two significant desires to beat the earnings targets: Delaying bad news and Avoiding the torpedo effect on earnings surprises.

In addition, several papers have suggested that the asymmetric market penalty for unexpected negative earnings surprises could create strong incentives for managers to avoid them through earnings management. Payne and Robb (2000) presented evidence that managers have a greater desire to increase income in order to meet forecasted earnings levels when pre-managed earnings are below market expectations. Matsumoto (2002) and Dechowet al. (2006) also found that, compared to firms reporting earnings below analysts’ forecasts, firms reporting earnings above analysts’ forecasts have higher positive discretionary accruals.

Effects of Revenue Shocks on the Relation between Returns and Earnings Surprises

Even though the impacts of earnings surprises and associated earnings management have been extensively investigated in earlier research, the effects of other accounting signals, such as revenue shocks, on the relation between returns and earnings surprises have not been paid close attention to. Rees and Sivaramakrishnan (2007) focused on the influences of other accounting signals, especially revenue surprises, on the relationship between returns and earnings shocks. By using analysts’ revenues and earnings forecasts as proxies of market expectations of revenues and earnings, they found that the market awards significantly higher (lower) premiums (penalties) to firms meeting or beating earnings forecasts only when the revenue forecasts are also met (not met). However, they did not find significantly different effects on returns when signals of earnings and revenue shocks conflicted with each other. It means that conflicting surprise signals between earnings and revenue result in the elimination of any equity premium or penalty associated with meeting or missing earnings forecasts. They concluded that the market appears to put an equal weight on earnings and revenue surprises to react when both signals are inconsistent with each other.

Hypothesis Development

To develop a hypothesis, this paper considers four situations that firm’s managers might confront before announcing the financial report. The four situations are:

1. both positive revenue and earnings surprises,

2. both negative revenue surprises and positive earnings surprises,

3. both positive revenue surprises and negative earnings surprises, and

4. both negative revenue and earnings surprises.

Case 1: Meeting Market Expectations for Both Revenue and Earnings (Ex Ante R+E+)

This situation represents the ideal scenario for firm managers. However, it is possible that managers might engage in earnings management to amplify the magnitude of earnings surprises if the market rewards meeting earnings expectations based on the size of the positive earnings surprise. Yet, since market reactions to meeting or beating earnings expectations are not related to the magnitude of positive earnings surprises (Rees & Sivaramakrishnan, 2007), managers are unlikely to manipulate earnings to increase the size of these surprises due to the risk of detection. Therefore, once managers achieve the targeted level of revenue and earnings, they have no incentive to engage in earnings manipulation.

Case 2: Failing to Meet the Market Expectation for Revenue but Meeting It for Earnings (Ex Ante R-E+)

Before announcing the financial report, managers may find themselves in a situation where they are meeting earnings expectations but still falling short of revenue expectations. In this case, there is a possibility that firms could receive greater rewards from the market if they manipulate revenue upward to meet or exceed both revenue and earnings expectations. However, revenue manipulation is more difficult than expense manipulation, especially at the last minute before the earnings announcement, and it carries a higher risk of detection by regulators (Ertimuret al., 2003; Dechow & Schrand, 2004). Thus, firms may have less desire to use income-increasing manipulation because they can avoid negative market responses associated with missing the expected earnings. Also, firms could be free from disproportional market penalties for missing one of the targets as long as firms have conflicting signals for earnings and revenue surprises (Rees & Sivaramakrishnan, 2007). Consequently, in this situation, firms are less prone to manage earnings or revenue to meet or exceed market expectations.

Case 3: Meeting Market Expectation for Revenue But Not for Earnings (Ex Ante R+E-)

Managers may confront positive revenue surprises and positive earnings surprises before earnings announcement. Under this situation, firms can avoid asymmetrical market punishment for missing the earnings targets because the market reacts equally to revenue and earnings surprises when there are conflicting signals (Rees & Sivaramakrishnan, 2007). However, they might try to manage earnings upward to meet market expectations for both revenue and earnings, as firms can receive higher market equity premiums for achieving both (Rees and Sivaramakrishnan, 2007). As a result, when firm managers face positive revenue surprises and negative earnings surprises before the announcement, they are more likely to increase earnings to meet or exceed market expectations.

Case 4: Failing to Meet Market Expectations for Both Revenue and Earnings (Ex Ante R-E-)

If firms fail to meet or exceed market expectations for both earnings and revenue, they face more intense market punishments, known as the ‘Torpedo Effect’ (Rees & Sivaramakrishnan, 2007). In this situation, managers have a much stronger incentive to increase reported earnings or revenue compared to other situations they may encounter before a financial report announcement. They have two options: increasing revenue or increasing earnings alone. Since revenue management may not be desirable due to high detection risks, managers are more likely to focus on earnings manipulation to boost reported earnings.

All possible situations which managers can face before announcing the financial report are considered. This study conjectures that firms could have a stronger incentive for earnings management to achieve the targeted earnings when they have both negative revenue and earnings surprises before earnings announcement (Case 4) relative to the other three situations (Case 1, 2, 3). Therefore, the hypothesis is as follows:

• H1: Firms with negative revenue and positive earnings surprises before earnings announcements are more likely to meet or beat analysts’ earnings forecasts by using positive discretionary accruals compared to firms in other situations.

Research Design

Sample Selection

The sample includes December fiscal year-end firms which are traded in the US stock markets between 2009 and 2014. Annual earnings and revenue forecasting data from I/B/E/S summary statistics as proxies for the market’s expectations for them (O’brien, 1988; Bartovet al., 2002; Rees & Sivaramakrishnan, 2007). This paper uses the mean estimate of analysts’ earnings per share (EPS) forecasts as a consensus of market expectations for the month preceding the annual earnings announcement. Annual accounting related data to calculate discretionary accruals and others can be obtained from the COMPUSTAT database. Furthermore, data for financial institutions which have SIC codes between 5999 and 7000 is excluded because these firms are likely to have different incentives for managing earnings from other firms (Burgstahler & Dichev, 1997). The final sample includes 20,734 firm-year observations.

Detection of Earnings Management

Following prior literature (Dechowet al., 1995; Payne & Robb, 2000; Dechowet al., 2006), this study uses changes in discretionary accruals to detect earnings manipulations. Even though the Jones (1991) could identify firms’ earnings management by estimating changes in discretionary accruals, Dechowet al. (1995) recommended the use of a Modified Jones Model that could generate more reliable estimations of discretionary accruals. Accordingly, the Modified Jones Model is adopted to estimate the changes in discretionary accruals. The Modified Jones Model is applied as below.

T A i , t A s s e t i , t 1 = α 1 1 A s s e t i , t 1 + α 2 ( Δ R e v i , t Δ R e c i , t ) A s s e t i , t 1 + α 3 P P E i , t A s s e t i , t 1 + ε i , t

where

TAit – total accruals in year t

Assetit-1 – total assets at the end of year t-1

Δ Revit – change in revenues from year t-1 to year t

Δ Recit – change in receivables from year t-1 to year t

PPEit – gross PP&E in year t

εit – error term

Each coefficient (α1, α2, α3) is obtained from the results of Ordinary Least Squares (OLS) by year and each two-digit SIC code. Then, by using the acquired coefficient, the nondiscretionary accruals are computed in the event year as

N o n D A i , t A s s e t i , t 1 = α 1 1 A s s e t i , t 1 + α 2 ( Δ R e v i , t Δ R e c i , t ) A s s e t i , t 1 + α 3 P P E i , t A s s e t i , t 1

where

NonDAit – nondiscretionary accruals in the event year t

α1’ α2’ α3 – coefficients of α1, α2, α3 acquired from the model (1) regression

Finally, discretionary accruals are calculated from comparing total accruals and nondiscretionary accruals.

D A i , t = T A i , t N o n D A i , t

where

DAit – discretionary accruals for firm i in year t

Multinomial Logit Model for the Empirical Test of Hypothesis

This study examines the relationship between the probabilities of meeting or beating analysts’ earnings forecasts across four cases (Ex post R+E+, R+E-, R-E+, R-E-) and the sign of abnormal discretionary accruals (+DAs and -DAs). Since this setting involves multiple, discrete, and unordered outcomes, the empirical analysis employs a multinomial logit model. Additionally, I anticipate that the effects of the independent variables may vary across different groups. For example, managers may have less incentive to meet or beat earnings benchmarks depending on the situations they face. Therefore, the model allows the parameters of the independent variables to vary across groups. The multinomial logit model used in this study is described in detail.

R E S U P _ D U M = α 0 + α 1 P O D A + α 2 M B + α 3 L T G _ R I S K + α 4 F O L _ A N A L Y S T S + α 5 D I S _ A N A L Y S T S + α 6 S A L E S _ G R + α 7 L O G _ C A P

where the dependent RESUP_DUM is coded the value of 1, 2 or 3 if a firm observation is classified into R+E+, R+E-, R-E-, respectively, and zero for R-E+. The model includes a categorical independent variable representing whether firms have positive or negative discretionary accruals. We put in a value of 1 if a firm has positive discretionary accruals; otherwise, 0. I predict that the coefficient α1 on PODA is negative, which implies that firms in the reference group (R-E+) are more likely to use income-increasing discretionary accruals to achieve positive earnings surprises compared to other groups.

Other independent variables identified in previous research are added to capture possible impacts on the probability of meeting or exceeding the market expectations for earnings. Managers are more likely to report earnings which meet or beat the market expectation under several circumstances: when firms have 1) a high market-to-book ratio (MB) (Skinner & Sloan, 2002), 2) high litigation risk (LTG_RISK) (Sofferet al., 2000), or 3) a large analyst following (FOL_ANALYSTS) (Johnson, 1999). In addition, the dispersion of analysts’ forecasts (DIS_ANALYSTS), which is measured as the standard deviation of analysts’ earnings forecasts, is included (Payne & Robb, 2000). To control for firm performance and size, sales growth (SALES_GW) and the natural logarithm of the market value of common equity (LOG_CAP) are also added.

Empirical Results

Descriptive Statistics for the Variables Used in the Multinomial Logit Model

Table I presents the descriptive statistics (means and medians) for the independent variables used in the multinomial logit regression. More specifically, a total of 28 tests for mean and median difference are conducted because there are 7 variables per all subsamples, including the reference subsample. The results from all tests show that there are 21(20) significant mean (median) differences at statistically significant levels.

Reference var Dependent variables
Variables R-E+ (Code = 0) R+E+ (Code = 1) R+E- (Code = 2) R-E- (Code = 3)
PODA Mean 0.58 *** 0.48 *** 0.49 *** 0.40 ***
Median 0.00 *** 0.00 *** 0.00 *** 0.00 ***
MB Mean 4.46 * 3.92 * 3.11 3.05 *
Median 2.98 * 2.86 * 2.09 2.03 *
LTG_RISK Mean 0.31 0.322 * 0.284 * 0.298
Median 0.00 0.00 * 0.00 * 0.00
FOL_ANAYSTS Mean 6.78 6.83 * 6.21 * 5.91 *
Median 5.60 5.80 * 5.00 * 4.50 *
DIS_ANAYSTS Mean 0.20 * 0.19 *** 0.24 0.25 ***
Median 0.06 0.06 *** 0.09 0.10 *
SALES_GR Mean 0.29 * 0.24 0.26 *** 0.19 *
Median 0.12 * 0.13 0.15 *** 0.11 *
LOG_CAP Mean 6.28 ** 6.32 ** 6.15 ** 6.12 **
Median 6.13 ** 6.23 ** 6.05 ** 5.89 **
N 5868 9600 1928 3338
Table I. Descriptive Statistics

The Association Between The Earnings And Revenue Surprises And Earnings Management

The contingency table in Table II presents the association between each subsample (R+E+, R+E-, R-E+, R-E-) and discretionary accruals. The results indicate that over 58% of firms having positive earnings surprises and negative revenue surprises have positive abnormal accruals. More importantly, the classification table shows that the group (R-E+) contains a larger portion of firm observations having positive abnormal accruals relative to other groups, suggesting that firms with negative revenue and positive earnings surprises have a higher tendency to engage in upward earnings management.

# of observations
R-E+ R+E+ R+E- R-E-
Proxy for earnings management Positive 3423 4632 960 1339
(PODA = 1) (58.34%) (48.25%) (49.78%) (40.12%)
Abnormal discretionary accruals Negative 2445 4968 968 1999
(PODA = 0) (41.66%) (51.75%) (50.22%) (59.88%)
28.30% 46.30% 9.00% 16.09%
Table II. Contingency Table

Multinomial Logit Regression Results

Panel A in Table III reports the results of the multinomial logit regression using all firm observations. The dependent variable is coded 0 if a firm observation contains R-E+, 1 if it contains R+E+, 2 if it contains R+E-, and 3 if it contains R-E-. The coefficients in the result table show the effect of the explanatory variables on coded dependent variables relative to the reference (R-E+). Specifically, the statistically significant coefficient indicates that the corresponding independent variable affects the marginal utility of the relevant coded groups relative to R-E+. A negative sign of estimate implies that firms are more likely to be classified into the reference group (R-E+) than into other groups as the value of those independent variables increases.

Variables Expected sign R+E+ Sig. R+E- Sig. R-E- Sig.
Intercept (?) −0.727 −0.805 −0.856
(−8.34) *** (−7.25) *** (−6.89) ***
PODA (−) −0.063 −0.085 −0.088
(−6.01) *** (−4.62) *** (−5.73) ***
MB (−) −0.002 −0.021 −0.025
(−0.35) (−4.86) *** (−3.38) ***
LTG_RISK (−) −0.096 −0.171 −0.113
(−0.94) (−3.12) *** (−1.11)
FOL_ANAYSTS (−) −0.008 −0.007 −0.005
(−1.92) * (−2.34) ** (−3.56) ***
DIS_ANAYSTS (+) −0.244 0.293 0.304
(−7.92) *** (6.24) *** (5.96) ***
SALES_GR (−) −0.606 −0.723 −0.487
(−13.63) *** (−15.01) *** (−7.65) ***
LOG_CAP (−) −0.156 −0.134 −0.11
(−1.17) (−7.86) *** (−11.34) ***
Chi-square (p-value) 5039.26 (0.000)
Number of observations 20734
Table III. Empirical Results: Panel A: Multinomial Logit Model Results

Panel A in Table II exhibits the negative and highly significant coefficients on PODA for all situations (R+E+, R+E-, and R-E-). These results imply that ceteris paribus, when firms have negative revenue surprises, they are more likely to manipulate the reported earnings upward to meet or beat the analysts’ earnings forecasts compared to other situations. This finding is consistent with the hypothesis. Panel B in Table IV presents the marginal effects of explanatory variables. The results presented in Panel B allow us to assess the simultaneous effect of the explanatory variables on the probability of the four distinct groups. PODA is positively and significantly associated with R-E+ and R+E+, negatively and significantly with R+E- and R-E-. This result indicates that firms are more likely to use income-increasing accruals to achieve positive earnings surprises regardless of revenue surprises. These results also indirectly support the hypothesis that firms having positive revenue surprises are more likely to meet or beat analysts’ earnings forecasts by using income-increasing discretionary accruals. Coefficients on most control variables are consistent to the expectations.

Variables R-E+ Sig. R+E+ Sig. R+E- Sig. R-E- Sig.
Intercept −0.1783 0.1211 0.0087 −0.0485
(3.41) *** (4.56) *** (−4.28) *** (−5.06) ***
PODA 0.0546 0.0340 −0.0240 −0.0646
(5.42) *** (2.98) ** (−3.46) *** (4.61) ***
MB 0.0024 0.0015 −0.0018 −0.0021
(4.50) *** (5.56) *** (−5.56) *** (3.78) ***
LTG_RISK 0.0810 0.0410 −0.0780 −0.0440
(0.92) (1.00) (−2.98) *** (−4.44) ***
FOL_ANAYSTS 0.0017 0.0021 −0.0020 −0.0017
(1.98) * (3.21) *** (−2.45) ** (−3.75) ***
DIS_ANAYSTS −0.1544 −0.1210 0.1317 0.1437
(−7.89) *** (−10.21) *** (4.56) *** (5.01) ***
SALES_GR 0.2288 0.3021 −0.3570 −0.1739
(8.98) *** (7.98) *** (6.54) *** (7.88) ***
LOG_CAP 0.0729 0.0812 −0.0543 −0.0998
(6.13) *** (7.31) *** (9.53) *** (10.31) ***
Table IV. Empirical Results: Panel B: Marginal Effects of Explanatory Variables

Conclusion

Because of differentiated market responses (reward or penalty) associated with earnings surprises conditioned on revenue surprises (Rees & Sivaramakrishnan, 2007), firms’ managers may have various incentives to manage earnings upward to meet or beat the market expectations for earnings given revenue surprises. This paper investigates whether revenue surprises can affect manager’s incentive to manipulate reported earnings to meet or exceed analysts’ earnings forecasts.

By utilizing the multinomial logit model, this study compares the use of income-increasing accruals of the reference group (R-E+) to other groups (R+E+, R+E-, R-E-). The empirical analysis finds that firms with negative revenue surprises (R-E+) are more likely to use positive abnormal accruals to meet or exceed earnings targets relative to other situations (R+E+, R+E-, R-E-). These results suggest that firms have the strongest incentives to increase their earnings to meet or beat analysts’ earnings forecasts when they have negative unexpected revenue. These findings are consistent with prior research that the major purpose of aligning earnings with market expectations is to avoid asymmetrical negative market responses associated with missing the expected earnings.

References

  1. Bartov, E., Givoly, D., & Hayn, C. (2002). The rewards to meeting or beating earnings expectations. Journal of Accounting and Economics, 33(2), 173–204.
     Google Scholar
  2. Brown, L. D. (2001). A temporal analysis of earnings surprises: Profits versus losses. Journal of Accounting Research, 39(2), 221–241.
     Google Scholar
  3. Burgstahler, D., & Dichev, I. (1997). Earnings management to avoid earnings decreases and losses. Journal of Accounting and Economics, 24(1), 99–126.
     Google Scholar
  4. Burgstahler, D., & Eames, M. (2006). Management of earnings and analysts’ forecasts to achieve zero and small positive earnings surprises. Journal of Business Finance & Accounting, 33(5 6), 633–652.
     Google Scholar
  5. Conrad, J., Cornell, B., & Landsman, W. R. (2002). When is bad news really bad news?. The Journal of Finance, 57(6), 2507–2532.
     Google Scholar
  6. Dechow, P. M., Richardson, S. A., & Tuna, A. I. (2006). Are benchmark beaters doing anything wrong?. Ann Arbor, 1001, 48109–41234.
     Google Scholar
  7. Dechow, P. M., & Schrand, C. M. (2004). Earnings quality. Research Foundation of CFA Institute.
     Google Scholar
  8. Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management. The Accounting Review, 70(2), 193–225.
     Google Scholar
  9. Ertimur, Y., Livnat, J., & Martikainen, M. (2003). Differential market reactions to revenue and expense surprises. Review of Accounting Studies, 8(2), 185–211.
     Google Scholar
  10. Jegadeesh, N., & Livnat, J. (2006). Revenue surprises and stock returns. Journal of Accounting and Economics, 41(1–2), 147–171.
     Google Scholar
  11. Johnson, N. S. (1999). Current SEC developments—’Managed earnings’ and ‘The year of the accountant’. In Remarks delivered at the Utah State Bar Mid-Year Convention, St. George, UT. March (vol. 6).
     Google Scholar
  12. Jones, J. J. (1991). Earnings management during import relief investigations. Journal of Accounting Research, 29(2), 193–228.
     Google Scholar
  13. Kasznik, R., & McNichols, M. F. (2002). Does meeting earnings expectations matter? Evidence from analyst forecast revisions and share prices. Journal of Accounting Research, 40(3), 727–759.
     Google Scholar
  14. Kinney, W., Burgstahler, D., & Martin, R. (2002). Earnings surprise ¡°materiality¡± as measured by stock returns. Journal of Accounting Research, 40(5), 1297–1329.
     Google Scholar
  15. Matsumoto, D. A. (2002). Management’s incentives to avoid negative earnings surprises. The Accounting Review, 77(3), 483–514.
     Google Scholar
  16. O’brien, P. C. (1988). Analysts’ forecasts as earnings expectations∗ 1. Journal of Accounting and Economics, 10(1), 53–83.
     Google Scholar
  17. Payne, J. L., & Robb, S. W. G. (2000). Earnings management: The effect of ex ante earnings expectations. Journal of Accounting Auditing and. Finance, 15(4), 371–392.
     Google Scholar
  18. Rees, L., & Sivaramakrishnan, K. (2007). The effect of meeting or beating revenue forecasts on the association between quarterly returns and earnings forecast errors∗. Contemporary Accounting Research, 24(1), 259–290.
     Google Scholar
  19. Skinner, D. J., & Sloan, R. G. (2002). Earnings surprises, growth expectations, and stock returns or don’t let an earnings torpedo sink your portfolio. Review of Accounting Studies, 7(2), 289–312.
     Google Scholar
  20. Soffer, L. C., Thiagarajan, S. R., & Walther, B. R. (2000). Earnings preannouncement strategies. Review of Accounting Studies, 5(1), 5–26.
     Google Scholar