Investor Reaction to Strategic Emphasis on Earnings Numbers: An Empirical Study

We analyze the earnings information and stock prices of S&P500 firms and find that investors following S&P500 stocks (i) respond more to pro forma earnings than to GAAP earnings, (ii) respond to an emphasis on pro forma earnings, and (iii) are fixated on pro forma earnings. We provide the first direct evidence that a strategic emphasis on earnings numbers may affect return volatility. Further, our results do not support the argument that a larger investor response to Street earnings might be driven by large differences between the Street numbers and GAAP numbers.


Introduction
This study examines whether a strategic emphasis on earnings numbers in company press releases has any impact on stock returns. We define strategic emphasis as emphasis on one earnings number over another or the exclusion of an earnings number from a press release with the purpose of painting a rosy picture of a firm's financial position. Given that a vast amount of information is constantly flowing into the market, it is very difficult for even astute investors to sift through such an information flow in an organized manner.
Indeed, departing from the traditional analytical modeling of financial reporting that assumes investor rationality and market efficiency with respect to public information, Hirshleifer and Teoh (2003) show that, due to limited investor attention and information processing capacity, informationally equivalent disclosures can have very different effects on investor perceptions, depending on the form of presentation.
Other theoretical studies, such as that of Bloomfield (2002), also suggest that companies can influence investors' judgment by putting earning numbers in a prominent place (headline or lead paragraph) in a press release or by burying them somewhere in a less prominent position. Therefore, it is important to examine whether there is any informational value in emphasizing a preferred earnings metric in companies' press releases. Obviously, investors form expectations, at least partially, about the stocks they follow on the basis of what they read in the earnings press releases.
Companies' earnings press releases usually include two types of earnings numbers: GAAP (Generally Accepted Accounting Principles) earnings and pro forma earnings (popularly known as Street earnings). For GAAP earnings, specific rules are followed to calculate earnings, whereas pro forma earnings are often defined by companies and calculated without following any standard rules. However, the purposeful exclusion releases because they are not subject to the SEC's (Securities and Exchange Commission) scrutiny unless they are fraudulent. Bradshaw and Sloan (2002) provide compelling evidence of managers' tendency to emphasize Street earnings because they present the Street earnings earlier than the GAAP earnings in press releases. The authors relate the tendency of managers to report higher Street numbers to a firm's effort to achieve higher valuations.
Subsequently, studies have identified at least four specific reasons why firms emphasize pro forma earnings over GAAP earnings in their reports. First, firms deliberately report pro forma earnings when their earnings and stock prices start to decline (Bhattacharya, et al., 2004). Second, firms with greater incidences of losses, higher leverage, and higher proportions of special items tend to report pro forma earnings (Lougee and Marquardt, 2004). Third, firms may manipulate earnings to meet or beat analysts' expectations (Doyle, Mc-Nichols, & Soliman, 2013). Finally, senior executives may try to use discretionary components of earnings to inflate a firm's reported earnings and thus its stock price because their compensation packages are tied to the company stock price (Bergstresser & Phillippon, 2006). Empirical evidence also shows that stock returns around earnings announcement days are more responsive to pro forma earnings than to GAAP earnings (Bhattacharya, et al., 2003;Brown & Sivakumar, 2003). Bowen, Davis and Matsumoto (2005) find that a strategic emphasis on an earnings metric (pro forma or GAAP) affects stock market participants, and firms with more media exposure place a greater emphasis on pro forma earnings than on GAAP earnings.
Relatedly, from a behavioral accounting perspective, the functional fixation hypothesis of Hand (1990) suggests that individual investors do not consider the quality of earnings. Experimental research on financial information processing also suggests that the market mechanically capitalizes reported earnings numbers without adjusting for the quality of the earnings (Libby, Bloomfield, & Nelson, 2002). According to Hirshleifer and Teoh (2003), limited attention provides a motivation for such functional fixation on the part of the investors. If the functional fixation holds, a pro forma disclosure may cause investors to perceive an earnings announcement as more favorable, which, in turn, causes them to expect a higher stock price in the period around such an announcement.
When companies tend to purposefully emphasize pro forma earnings and investors care only about the earnings numbers without paying much attention to the underlying accounting procedures and/or have limited information processing capacity, then a thorough analysis of the emphasis on a particular earnings metric becomes an issue of obvious importance to investors and market regulators. In our effort to assess the informational value of a strategic emphasis on earnings numbers, we analyze the earnings information and stock prices of S&P500 firms and address the following three research questions in this paper: a Does a deliberate emphasis on a particular earnings number in company press releases attract a shareholder's attention and affect stock prices? As long as firms use a strategic emphasis and investors perceive pro forma earnings as more value relevant, we should find a positive relationship between returns and such an emphasis unless investors identify that such an emphasis is completely opportunistic. Accordingly, we conjecture that return responses to a pro forma earnings emphasis will be greater than the return responses to a GAAP emphasis.
b Does a managerial emphasis in earnings disclosures affect the volatility of stock returns? Ideally, the very purpose of earnings disclosures in company press releases in general is to accelerate the resolution of investor uncertainty. Accordingly, in the wake of a current period's earnings press release, we would normally expect to observe a decline in uncertainty as measured by the volatility of stock returns. However, an alternative possibility is that an opportunistic emphasis on earnings numbers may magnify the information asymmetry among market participants associated with an earnings surprise and thereby increase the volatility of returns, at least in the short run (see, for example, Kim & Verrecchia, 1994). It is also not unlikely that investors' uncertainty about intrinsic firm value increases because of a deliberate emphasis placed on a particular earnings number.
c Are investors really fixated on pro forma earnings?
The functional fixation hypothesis (Hand, 1990) posits that investors are fixated on an earnings Electronic copy available at: https://ssrn.com/abstract=2362266 Investor reaction to strategic emphasis on earnings numbers: An empirical study metric, and they do not adjust for earnings quality.
On the contrary, however, Abarbanell and Lehavy (2007) Hand (1990) holds even after we recognize the claim made by Abarbanell and Lehay (2007).
Additionally, this paper provides compelling evidence that studies examining the emphasis of alternative earnings metrics on stock return responses should consider the impact of the fourth quarter because most write-offs and special items are recognized in the fourth quarter (Bradshaw & Sloan, 2002).
In this paper, we use a sample of 308 S&P500 firms for the period from 2000: Q1 -2006:Q3. 1 In this particular sample, our results show that investors following S&P500 stocks: (i) respond more to pro forma earnings than to GAAP earnings (evident from the earnings response coefficients); (ii) respond to an emphasis on pro forma earnings in earnings press releases; and (iii) are fixated on pro forma earnings.
The remainder of this paper is organized as follows: Section 2 describes the data used and methodology employed. Section 3 presents and explains the empirical results. Section 4 concludes.

Data and methodology
The final sample used in this study consists of 27 quarters of data for 308 S&P500 companies. We include only those firms in our sample for which at least 16  Bradshaw and Sloan (2002) claim that emphasis on Street earnings is originated from firm managers rather than from forecast data providers such as IBES.

Strategic emphasis on earnings numbers
To identify a strategic emphasis on earnings numbers, we follow Hirshleifer and Teoh (2003) in that investors have limited information processing capacity, which results in a less complete evaluation of quarterly earnings press releases. Elliott (2006); Frederickson and Miller (2004) (2010) finds that, even after the approval of Regulation G in 2002, not all S&P500 firms that disclose a non-GAAP financial measure include a reconciliation in the press release of the quarterly earnings announcement. 2 Accordingly, when one earnings metric in a press release is emphasized before another, we define it as a relative emphasis. When only GAAP or pro forma earnings are mentioned without any immediate information about exclusions in the same paragraph of a press release, we define it as an absolute emphasis. In addition, if reconciliation is provided in a tabular form in a press release, we consider it a relative emphasis. If a press release contains only a narrative description of exclusions and if the description appears somewhere in the release but not immediately after where the pro forma earnings are mentioned, we consider it an absolute emphasis.

Strategic emphasis on earning numbers and stock returns
Our empirical analysis begins with a regression model where we regress standardized cumulative abnormal returns on earnings surprises, emphasis, and control variables such as the number of analysts following a firm, firm size, and a policy dummy. Firm size and the number of analysts following a firm provide predisclosure information about a firm. Smaller (larger) firms are likely to have less (more) pre-disclosure information, leading to more (less) information revealed by earnings announcements, whereas firms that are followed by more (less) analysts are expected to provide large (small) amounts of firm-specific information. To examine the effects of press emphasis on stock returns, we run separate regressions for each emphasis term. Running separate regressions for each emphasis term reduces the possibility of linear dependence between alternative earnings surprises. The regression model used in this study to examine the impact of a strategic emphasis on stock returns is given below: Electronic copy available at: https://ssrn.com/abstract=2362266 Investor reaction to strategic emphasis on earnings numbers: An empirical study In the above regression model, , i t SCAR is the standardized cumulative abnormal return around the event (earnings announcement) window for firm i in quarter t . Defining the event day as 0 d = , we calculate abnormal returns over the event window ( ) Also in Equation (1)

Strategic emphasis on earnings numbers and return volatility
To identify the impact of a strategic emphasis on volatility, we regress the realized volatility of returns around the event window on earnings surprises, emphasis, and the control variables, including abnormal trading volume as an extra explanatory variable in addition to the number of analysts following a firm, firm size, and the policy dummy. Trading volume is a well-established proxy for information arrival in the market (Lamoureux & Lastrapes, 1990;. In addition, trading volume also reflects a lack of consensus among investors. That is, the volume (volatility) reaction to an earnings announcement could be high if there is more disagreement among investors in interpreting the information content of the earnings news. Specifically, we estimate the following regression equation to examine the impact of strategic emphasis on return volatility: In the above regression equation,

Are investors fixated on pro forma earnings?
To check whether the Abarbanel and Lehavy (2007) claim against the functional fixation hypothesis is con-

Estimation Results
Table 2 presents summary information for the dependent and independent variables used in this study. The mean GAAP earnings surprise is slightly higher than that of the pro forma surprise for firms that issued press releases over the sample period. We estimate the models in Equations (1) and (2) running pooled regressions. 4 Dependence in pooled data may arise when (i) residuals are correlated across quarters for a given firm or (ii) residuals for a given quarter are correlated across different firms. When errors are correlated, Ordinary Least Squares (OLS) is still consistent but is biased and typically suggests that standard errors are much too small (Cochrane, 2005).
To ensure that the standard errors are conservative relative to alternative estimation techniques, we estimate the models using pooled OLS with standard errors  Investor reaction to strategic emphasis on earnings numbers: An empirical study Notes: SCAR is the standardized cumulative abnormal return cumulated over four days surrounding the earnings announcement, RV is the log of 4-day realized volatility (around the earnings announcement), E SPF and ES GAAP are pro forma and GAAP earnings surprises, respectively, ANA is the number of analysts following a firm, SIZE is the log of the market value of a firm's equity, and ABVOL is the log of abnormal volume (the ratio of trading volume on the day of an earnings announcement and the average trading volume of the previous 5 days). The sample consists of 5522 observations.

Notes:
The Pearson (Spearman) correlation is shown below (above) the diagonal, and p -values for the null hypothesis of zero correlation are in parentheses. SCAR is the standardized cumulative abnormal return cumulated over four days surrounding the earnings announcement, RV is the log of 4-day realized volatility (around the earnings announcement), ES PF and ES GAAP are pro forma and GAAP earnings surprises, respectively, ANA is the number of analysts following a firm, SIZE is the log of the market value of a firm's equity, and ABVOL is the log of abnormal volume (the ratio of trading volume on the day of an earnings announcement and the average trading volume of the previous 5 days.) The sample consists of 5522 observations.  Hirshleifer and Teoh (2003) and the functional fixation hypothesis of Hand (1990) that investors respond to an emphasis on pro forma earnings because they pay less attention to the appropriateness of the accounting procedures that produce accounting numbers. This result also supports the empirical evidence obtained by Bhattacharya et al. (2003), Brown and Sivakumar (2003), and in experimental studies such as those of Elliott (2006) and Frederickson and Miller (2004) that the market responds more to pro forma earnings than to GAAP earnings. However, the control variables included in the regression between press emphasis and standardized cumulative abnormal returns are not significant at standard significance levels.   ) represents indicator variables used to quantify emphasis on earnings numbers, SIZE is the log of the market value of a firm's equity in the last quarter, ANA is the number of analysts following a firm, and PD is the policy dummy variable (to capture the impact of policy changes) that takes a value of 1 for observations after 2002:Q1 and 0 otherwise. Standard errors clustered by calendar quarters are in parentheses. ** significant at 5% level, *** significant at 1% level.
Investor reaction to strategic emphasis on earnings numbers: An empirical study GAAP earnings, return volatility increases. Given the median value of an earnings surprise in Table 2

Constant
where RV is the log of 4-day realized volatility, indicates pro forma and GAAP earnings surprises, respectively, ) represents indicator variables used to quantify emphasis on earnings numbers, SIZE is the log of the market value of a firm's equity in the last quarter, ANA is the number of analysts following a firm, ABVOL is the log of abnormal volume around the announcement day, and PD is the policy dummy variable (to capture the impact of policy changes) that takes a value of 1 for observations after 2002:Q1 and 0 otherwise. Standard errors clustered by calendar quarters are in parentheses. ** significant at 5% level, *** significant at 1% level.
with more analysts (ANA) following them are expected to produce a large amount of both financial and non-financial information. Therefore, the number of analysts following a firm is expected to reduce volatility. However, if analysts' forecast dispersion around the earnings announcement day increases, volatility may also increase. Abnormal volume (ABVOL) has a significant and positive relationship with volatility, which suggests that any unexpected news (good or bad) is followed by above-average trading activity as the market seeks out a new equilibrium. However, Our regression results reported in Table 6 show that investors respond to pro forma earnings when both the earnings exceed their benchmark and when either pro forma earnings are greater than GAAP earnings or GAAP earnings are greater than pro forma earnings.
This result suggests that S&P500 investors are fixated on pro forma earnings and thus supports the theories regarding investors' inattention to the use of an appropriate accounting procedures and their limited attention in processing earnings information (Hirshleifer & Teoh, 2003;Hand, 1990). However, this result does not support the claim of Abarbanell and Lehavy (2007) that the large differences between Street and GAAP earnings supplied by forecast data providers may lead α β β ε = + + + where SCAR is the 4-day standardized cumulative abnormal return, and ES PF and ES GAAP indicate pro forma and GAAP earnings surprises, respectively. Column (i) reports the results for the full sample when both pro forma and GAAP earnings exceed the benchmark (earnings for the same quarter in the prior year), Column (ii) reports the results for the sample when both pro forma and GAAP earnings exceed the benchmark and pro forma is higher than GAAP earnings, and Column (iii) reports the results for the sample when both pro forma and GAAP earnings exceed the benchmark and GAAP is higher than pro forma earnings. ** significant at 5% level, *** significant at 1% level.
Investor reaction to strategic emphasis on earnings numbers: An empirical study a researcher to wrongly support the notion that investors are fixated on Street earnings.

Impact of extreme observations
To check whether our results are robust to extreme observations, we re-estimate the regression models in Equation (1) and Equation (2) using robust regression.
Our results (not tabulated) for emphasis on competing earnings numbers by companies remain similar to our pooled OLS results. That is, we find that investors respond to an absolute pro forma earnings emphasis in the case of returns regression and respond to a relative emphasis on pro forma earnings in the case of volatility regression.

Impact of the fourth quarter
Fourth quarter earnings differ systematically from other quarter's earnings (Mendenhall & Nichols, 1988;Bradshaw & Sloan, 2002). Bradshaw and Sloan (2002) find that growth calculations based on Street earnings per share are statistically larger than those based on GAAP earnings per share. This difference is especially pronounced in the fourth quarter, where mean growth based on Street earnings per share is approximately three times that calculated based on GAAP earnings per share. Managerial opportunism and accounting year-end adjustments are considered the main reasons for such a difference. From the viewpoint of a strategic emphasis in earnings numbers, the fourth quarter is unique because discretionary write offs and special items are recognized more in the fourth quarter than in other quarters. To account for the unique nature of the fourth quarter and to examine whether it affects our empirical results, we rerun regression models in Equation (1) and Equation (2) excluding the fourth quarter observations from the sample.
Our results (not tabulated) show that when re-running the regressions without fourth quarter observations, the significant return response to an absolute emphasis on pro forma earnings vanishes, whereas the relationship between volatility and pro forma earnings emphasis becomes weak in terms of statistical significance. However, the return and volatility response to pro forma and GAAP surprises remains significant.
This result supports the claim of Bradshaw and Sloan (2002) that the fourth quarter plays a role in results that provide statistical support for a greater investor reaction to pro forma earnings.

Summary and Conclusion
When preparing pro forma earnings statements, managers generally exclude nonrecurring transitory items that are believed to be less relevant to normal ongoing company operations. Because pro forma statements are neither audited nor subject to direct scrutiny by the SEC, companies effectively enjoy the discretion to exclude anything from these financial statements that they believe can affect the accuracy of their financial outlook. Although proponents of pro forma estimates argue that these numbers provide a better idea about a company's actual earnings performance, they leave open the possibility of opportunistically omitting certain important expense items to overstate a company's financial performance.
One of the most controversial uses of pro forma statements is to release pro forma numbers to market participants before the audited GAAP financial statements are made available. Early releases of pro forma numbers without a clear-cut reconciliation of pro forma to GAAP numbers can encourage investors to make investment decisions on the basis of incomplete information.
Using a sample of 308 firms from the S&P500 for the period from 2000:Q1 -2006:Q3, we find that the investors do respond to an emphasis on pro forma earnings in earnings press releases, and they are not as attentive to judging whether exclusions from pro forma statements are appropriate when making investment decisions. Our results also show that when pro forma earnings are emphasized before GAAP earnings, stock return volatility increases in the period surrounding the earnings disclosure in the press releases. These results are robust to extreme observations in the sample.
We also provide empirical evidence that nullifies the claim of Abarbanell & Lehavy (2007) that the reported incremental investor response to Street earnings is in fact driven by the high frequency of large differences between Street and GAAP numbers. Using robust regression estimation, we show that the functional fixation hypothesis propounded by Hand (1990) still holds even if we recognize the claim made by Abarbanell & Lehay (2007).
When fourth quarter observations are dropped from our estimation, return response to an absolute emphasis on pro forma earnings becomes insignificant, and the relationship between volatility and a pro forma earnings emphasis becomes weak in terms of statistical significance. However, the return and volatility response to pro forma and GAAP surprises remains significant.
The results of this paper may have implications for market regulators and policy makers in that they may help to formulate rules to increase the quality and quantity of information by defining what and how information should be presented to investors. If small investors respond to managers' strategic behavior in terms of an emphasis on earnings numbers, policy makers should take action by changing the present policy or by introducing a new policy to control the adverse effect of a strategic emphasis by firms. However, such policy measures may not work in the case of judgmental errors made by rational investors. Therefore, in addition to devising an appropriate policy response, investors also need to be educated (for example, through information campaigns) so that they can identify such strategic behavior of companies and make more rational investment decisions.