“Regulation of non-GAAP reporting and investors’ perceptions”
- The effect of the updated C&DIs issued by the SEC in May 2016-
Name: Justine Oudshoorn Student number: 11425059
Thesis supervisor: dhr. drs. J.F. Jullens Date: June 20, 2018
Word count: 18,970
MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam
Statement of Originality
This document is written by student Justine Oudshoorn who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
Over time regulation of non-GAAP reporting has increased. The SEC is still concerned about the misleading potential of these non-GAAP measures. In May 2016, the SEC published the updated Compliance and Disclosure Interpretations (C&DIs) in the United States. The expectations of the SEC are that because of the updated C&DIs non-GAAP measures will be presented less prominently and will be less capable of misleading investors. This thesis uses a large hand-collected database of non-GAAP earnings to examine the effect of the updated C&DIs. I find that after the updated C&DIs non-GAAP numbers are less prominently presented. In the quarter before the updated C&DIs approximately 36% of the non-GAAP EPS numbers are presented with higher prominence than the GAAP equivalent. This prominence disappeared in the quarter the C&DIs are updated, only 4% of the non-GAAP numbers is still prominent. Moreover, I find that the difference between non-GAAP and GAAP earnings does not decrease after the updated C&DIs. Therefore, I find no evidence that the non-GAAP measures are constructed in a less misleading manner after the updated C&DIs. This implies that this expectation of the SEC is not met. Lastly, I find no evidence that returns are more driven by non-GAAP earnings in the post-C&DIs period. This suggests investors do not focus and react more on non-GAAP earnings after the C&DIs are updated. In all, I find limited evidence that the updated C&DIs have a big impact on reporting content and investors’ perceptions.
Keywords: non-GAAP earnings, GAAP earnings, updated C&DIs, prominence, misleading, investors, returns.
Table of Contents
1. Introduction ... 5
2. Literature review and theoretical background ... 9
2.1 Theoretical background ... 9
2.1.1 Corporate disclosure theory ... 9
2.1.2 Voluntary disclosure theory... 10
2.1.3 The role of regulation ... 11
2.1.4 Non-GAAP measures: what is it? ... 12
2.2 Non-GAAP measures: two motives ... 14
2.2.1 Informative motive ... 14
2.2.2 Opportunistic motive ... 16
2.3 Regulation ... 20
2.3.1 SOX & Regulation G ... 20
2.3.2 Compliance and Disclosure Interpretations (C&DIs) ... 21
2.4 The effect of Regulation ... 23
3. Hypothesis development ... 26
4. Research Design ... 30
5. Sample selection ... 32
6. Data collection... 33
6.1 Non-GAAP data hand-collection ... 33
6.2 Data collection other variables ... 36
7. Results ... 40
7.1 Descriptive statistics ... 40
7.2 Non-GAAP prominence ... 44
7.3 Non-GAAP misleading potential ... 46
7.4 Non-GAAP informativeness ... 49
7.5 Additional analysis ... 53
Conclusion... 55
References ... 59
1. Introduction
Concerns about non-Generally Accepted Accounting Principles (non-GAAP) disclosures are still current practice. Non-GAAP financial figures are measured and created by managers but these measures are not created in accordance with Generally Accepted Accounting Principles (GAAP). Reporting of extra, non-statutory financial information might serve two purposes. First, non-GAAP earnings can provide valuable information to shareholders, investors, and other potential shareholders. In this sense, non-GAAP metrics might reflect a better picture of the company’s core business than the reported GAAP numbers. Mostly, non-recurring items, such as restructuring charges are intended to be excluded from non-GAAP earnings, as this will give a better picture of future profitability (Black and Christensen, 2009). On the other hand, non-GAAP numbers might also be used to report opportunistically. Managers might overstate the non-GAAP number, by, for example, also excluding recurring items and with this strategically meet earnings benchmarks. If this is the case, investors might be misled by the reported non-GAAP number (Doyle et al., 2011).
Due to increased discussion about the use of non-GAAP measures and accounting scandals, non-GAAP reporting started to become regulated in some parts of the world around the 2000s. Non-GAAP reporting has been regulated in the United States by the U.S. Securities and Exchange Commission (SEC). The SEC was worried that the non-GAAP numbers were too misleading. In July 2002, the Sarbanes-Oxley Act (also known as SOX) was introduced. Because of SOX, the SEC had to introduce regulation about non-GAAP reporting. As a result, the SEC implemented Regulation G in March 2003. This regulation prohibited the non-GAAP measure to be misleading. First, it requires public companies to disclose the most directly comparable GAAP financial measure next to the non-GAAP measure. Moreover, a reconciliation from the non-GAAP financial measure to the most directly comparable GAAP financial measure is required (SEC, 2002). This created a different environment for non-GAAP reporting.
Over time, the SEC still worried about the misleading potential of non-GAAP numbers. In 2010, the SEC labeled non-GAAP earnings as ‘fraud risk factors’ (Leone, 2010). They also created a guidance of Compliance and Disclosure Interpretations (C&DIs) in the same year. On May 17th, 2016, the SEC updated the C&DIs about non-GAAP reporting. This update was published to provide additional guidance on what the SEC expects from non-GAAP reporting. The SEC’s main goal was to promote changes in non-GAAP reporting with the updated
C&DIs. The expected changes were mainly related to make non-GAAP reporting less misleading and give non-GAAP reporting less prominence in press releases. The updated C&DIs give relevant examples of non-GAAP measures that might be misleading. An example is when normal, recurring operating expenses are excluded. Examples are also given of non-GAAP measures which have greater prominence than the comparable non-GAAP measure (SEC, 2017). An example of this is when only the non-GAAP number is presented in the headlines while the comparable GAAP number is omitted from the headline (Deloitte, 2017b).
It is interesting to investigate whether firms adapt their non-GAAP reporting after the C&DIs are updated. This is to see whether regulation helps to reduce the likelihood that non-GAAP measures are opportunistically constructed. Next to that, it is also interesting to investigate whether and how investors incorporate those changes in their evaluation of companies. Investors might think that after the update the non-GAAP earnings are more reliable and less misleading and therefore they will rely more on this earnings number than before. On the other hand, investors might anticipate that non-GAAP numbers are constructed opportunistically and that the updated C&DIs will not have an effect on this. In this case, investors will not rely more on non-GAAP earnings after the update.
The purpose of this thesis is to examine the effect of the published C&DIs. More specifically, this thesis will investigate whether firms adapt their non-GAAP reporting and whether investors change the assessment of non-GAAP earnings after the updated C&DIs. Therefore, the research question of this thesis is:
Research question: “Does the update of the Compliance and Disclosure interpretations change non-GAAP reporting behavior and investors’ assessment of non-GAAP reporting?”
It is important to both examine whether firms adapt their non-GAAP reporting and whether investors adapt their assessment after the updated C&DIs. If investors adapt their assessment but nothing changes in the reporting investors might only be more misled. If the non-GAAP reporting changes because of the updated C&DIs and investors do not, one can wonder if investors are missing out or still think the measures are misleading. The optimal outcome of regulation is when both the non-GAAP reporting changes and investors anticipate this. In this case, non-GAAP measures are seen as less misleading and therefore investors focus more on non-GAAP reporting.
To see whether firms adapt their non-GAAP reporting after the updated C&DIs this thesis will look at the two specified goals of the SEC. I will investigate whether firms adapt their non-GAAP reporting to be both less misleading and less prominently presented after the updated C&DIs. If this is the case it will lead to non-GAAP measures that are more consistently applied and will create more confidence in non-GAAP measures under investors and stakeholders. This might result in an increased reliance on non-GAAP numbers by among others, investors. An answer to this research question shows important evidence for the policymakers in the U.S. to show whether regulation of non-GAAP reporting is effective. Furthermore, this thesis will show whether regulation influences investors’ responses to earnings announcements where non-GAAP measures are included. These findings should also interest regulators of other countries that consider implementing regulation to improve the credibility of financial reporting, especially non-GAAP reporting.
To examine whether the goals of the C&DIs are met I hypothesize that after the updated C&DIs the non-GAAP earnings are presented in a less misleading and less prominent manner. Misleading focuses on the composition of the non-GAAP number while prominence focuses on the placement of the number in the press release. Intuitively one might think that investors prefer to follow GAAP earnings since managers have a lot of discretion and therefore anticipate that non-GAAP earnings are easily manipulated. However, Bradshaw and Sloan (2002) find that investors gained the preference to follow the non-GAAP earnings number instead of the GAAP earnings number. The non-GAAP number replaced the GAAP profit number as the primary determinant for stock prices within their sample period. Evidence is also found that SOX and Regulation G have resulted in bringing non-GAAP disclosures to investors’ attention and improved the perceived quality of these disclosures (Black et al., 2012). Therefore, I hypothesize that investors will rely more on non-GAAP measures after the updated C&DIs because the non-GAAP measures might be more transparent, more credible and a better representation of reality.
I will use a sample of hand-collected non-GAAP earnings of 2,892 observations from S&P500 firms and benchmark this with GAAP earnings. The sample consists of a pre- and post-C&DIs sample, both samples contain four quarters. The updated C&DIs are published in the second quarter of 2016. My pre-C&DIs sample runs from Q2-2015 to Q1-2016 and the post-C&DIs sample from Q2-2016 to Q1-2017. To test whether the non-GAAP measures are presented in a less prominent way I examine whether firms make changes in their reporting habits in press releases after in post-C&DIs period. To see whether non-GAAP earnings are
less misleading I examine the difference between non-GAAP and GAAP earnings numbers in both the pre- and the post-C&DIs period. Lastly, I will measure market reactions to earnings announcements and find out whether investors rely more on GAAP or non-GAAP measures and if this changed due to the C&DIs. To measure this, I use the model as is used in Bhattacharya et al. (2003); Black et al. (2012) and Bradshaw et al. (2017). To my knowledge, no research exists on the impact of the C&DIs in general and on investors reactions. Therefore, I will contribute to the literature by providing insights into this. The second contribution is that I create a large hand-collected database for non-GAAP EPS. In total I did 4,148 observations and find 2,892 non-GAAP EPS numbers.
I find that firms immediately adapt their non-GAAP reporting after the updated C&DIs. More specifically, non-GAAP earnings numbers are presented less prominently from the quarter the updated C&DIs took place. I do not find evidence for the second hypothesis. The difference between the non-GAAP EPS number and the GAAP EPS number does not increase after the updated C&DIs. Thus, there is no evidence that the measures are less misleading after the update. With regards to the last hypothesis, I find that investors focus both on non-GAAP and GAAP numbers, and focus more on non-GAAP numbers in both the pre- and post-C&DIs period compared GAAP numbers. However, I do not find evidence that returns are more driven by non-GAAP forecast errors in the post-C&DIs compared to the pre-C&DIs period. Thus, investors do not rely more (or less) on non-GAAP earnings after the C&DIs update took place. To give a concrete answer to my research question I can state the following. Firms do partly adapt their non-GAAP reporting behavior after the updated C&DIs by giving the non-GAAP numbers less prominence. However, the way the non-GAAP numbers are constructed does not change. The numbers do not change in their misleading potential. Furthermore, investors do not change their assessment of non-GAAP numbers after the updated C&DIs.
The remainder of this thesis proceeds as follows. The next section describes my theoretical framework and literature review. I will discuss corporate and voluntary disclosure literature first. After this, I examine non-GAAP reporting literature and previous findings on how investors react to non-GAAP disclosures. The hypotheses will be developed in Section 3. Section 4 describes my research design. Section 5 and 6 will show how the sample is constructed and the data collected. The results are presented in Section 7. Lastly, I conclude.
2. Literature review and theoretical background
This chapter starts with an overview of corporate disclosure and voluntary disclosure theory. This is a good starting point because the theory shows in a general way what arguments firms have to disclose certain information, either voluntarily or regulated. Non-GAAP reporting is both a corporate and a voluntary disclosure. Second, the two motives for reporting non-GAAP measures are explained. Third, I will show how investors are influenced by these motives. Lastly, I will present the regulatory actions of the SEC in the United States and show how these regulations have an impact on non-GAAP reporting and investors’ perceptions. For an overview of the most important findings from the empirical studies I refer to Appendix A.
2.1 Theoretical background
2.1.1 Corporate disclosure theory
Both disclosures of company information and financial information are seen as important sources for outside investors to be informed about the company’s performance and governance. This makes disclosures important for an efficient capital market (Healy and Palepu, 2001). Corporate disclosures can be done through regulated financial reports or voluntary communication. Regulated financial reports include financial statements, footnotes, management discussion and analysis and regulatory filings. Voluntary communications are for example management forecasts, analysts’ presentations and conference calls, press releases, internet sites and other corporate reports. Lastly, disclosures about firms can also be done via information intermediaries, these are for example financial analysts, industry experts, and the financial press (Healy and Palepu, 2001).
In a capital market economy, the optimal allocation of savings to investments opportunities is an important challenge. Business entrepreneurs want to attract household savings to fund their business ideas. Matching savings to potential investments is complicated and therefore information is necessary. The demand for accounting information by outsiders arises for two reasons. Firstly, managers within the firm typically have more information about expected profitability of the firm and the future of investments than outsiders. This is known as information asymmetry. Information asymmetry makes it difficult for outside capital providers (such as investors) to assess the profitability of the firm’s investment opportunities (Beyer et al., 2010). Secondly, the demand for accounting information arises from a separation of ownership and control. In modern economies capital providers do not have full
decision-making rights. These decision-decision-making rights lay at management. After investors have invested, the self-interested manager has an incentive to make decisions to expropriate investors’ resources. This is known as the agency theory (presented by Jensen and Meckling, 1976). An agency problem arises between managers and outside investors (Beyer et al., 2010; Healy and Palepu, 2001). Proposed solutions to the agency problem are for example to provide optimal contracts between firms and investors. These contracts force companies to disclose information to make sure managers take the best actions for their firm (Healy and Palepu, 2001). Moreover, a board of directors might function as a monitoring mechanism to act on behalf of the shareholders. To reduce the agency problem and information asymmetry corporate disclosures are necessary.
Corporate disclosures serve two roles in the capital market: a valuation role and a stewardship role (Beyer et al., 2010). The valuation role is to reduce information asymmetry, which helps investors to choose among alternative investments (i.e. reduce estimation risk). The stewardship role is to align incentives better, leading to a better capital allocation within the firm. Non-GAAP disclosures can be seen as a voluntary disclosure that reduces information asymmetry between the firm and the capital market (i.e. investors).
2.1.2 Voluntary disclosure theory
It is often claimed that if managers do not disclose any news and investors know there is news investors take the conclusion that the information would be bad news. This would have caused them to revise their expectations downwards. Therefore, managers have incentives to disclose information to distinguish themselves from other managers with less favorable information. Beyer et al. (2010) identify conditions under which firms disclose all their private information including (1) disclosures are costless; (2) investors know the firms have private information; (3) all investors interpret the firms’ disclosures in the same way and firms know how investors will interpret the information; (4) managers want to maximize their firms’ share prices; (5) firms can credibly disclose their private information; and (6) firms cannot commit ex-ante to a specific disclosure policy.
However, these conditions only hold in a perfect market which is not the case in modern market economies. Because of this, full disclosure is not likely to occur. Firstly, there is a cost to disclosure. Companies will only disclose information from which the benefits outweigh the costs. Disclosure can be costly when the information in the disclosure informs the competitors. These costs are known as proprietary costs. So non-disclosure does not mean that there is only
bad news, it can also be good news but with a high cost of disclosure (Beyer et al., 2010). Second, if investors do not know that the manager has private information, the firm would be viewed the same as firms without any private information. Lastly, voluntary disclosures are not necessarily credible. Based on the agency theory managers can have incentives to make self-serving voluntary disclosures instead of reducing information asymmetry and therefore only mislead the capital market (Healy and Palepu, 2001).
Firms can benefit from voluntary disclosures. First, voluntary disclosures can help to reduce information asymmetry among informed and uninformed investors. This leads to higher stock liquidity and lowers cost of capital. This is because investors are confident that transactions occur at a “fair price”, which in turn will increase the liquidity of the stock. Moreover, investors ask a risk premium for bearing information risk. Voluntary disclosures will reduce the information risk and therefore the cost of equity capital (Healy and Palepu, 2001).
To disclose information voluntarily is a decision made by managers themselves, this makes the information less credible. To enhance the credibility of these disclosures intermediaries can give a guarantee over managements’ disclosures or the quality of the voluntary disclosure can be compared with the audited financial reports (Healy and Palepu, 2001). Healy and Palepu (2001) conclude that with the interference of regulators, standard setters, auditors and other capital market intermediaries it is possible to increase the quality of voluntary disclosures.
2.1.3 The role of regulation
As is explained above, managers do not voluntarily disclose all their private information. Therefore, all over the world corporate governance reporting and disclosure is regulated. In the United States accounting standards are created by the Financial Accounting Standards Board (FASB) for capital market participants (SEC, 2000). These accounting standards are known as GAAP and help companies with the creation of reports by making sure that these reports have “transparent, consistent, comparable, relevant and reliable information” (SEC, 2000). The purpose of financial reporting according to the FASB is to provide an objective look at a company’s financial situation (FASB, 2017). They recognize that financial reporting is necessary for the functioning of a capital market. With regulation of financial reporting, investors and other creditors can compare financial reports. This leads to a cost saving for both prepares and users of financial reports (FASB, 2017). By regulating disclosures,
the firm is forced to reduce the information asymmetry between the informed and uninformed (Healy and Palepu, 2001).
Beyer et al. (2010) state two main reasons for regulating disclosures. First, the agency problem leads to misaligned incentives between the firm and its investors. This might lead to information transfers that are not credible. Regulation is a mechanism that allows firms to commit to certain levels of disclosures and to improve the quality of the disclosures. Second, the public good aspect of disclosures results in free-rider problems, this is when people enjoy a service without paying anything. In a free market, if enough people can enjoy a good without costs there is a danger that the good will be underprovided or even not provided. In this case, additional information might improve social welfare but only managers’ incentives to disclose information is not sufficient, therefore regulation is required. Leuz (2010) adds another reason to justify regulation of company’s financial reports and disclosure activities. He states that there will be cost savings for companies. Since information is required to be disclosed, more information is available and this will reduce the cost of capital. Therefore, regulation might be desired. The regulation of disclosures will depend upon the part of the information that a firm voluntarily discloses or that can be produced by other market participants (Beyer et al., 2010). Non-GAAP reporting is a voluntary disclosure decision to reduce information asymmetry between the firm and the capital market. Non-GAAP information might give a better picture of future profitability and/or has the potential to mislead market participants. Because of the misleading potential and the potential of value-added information, non-GAAP reporting is regulated in the U.S. One of the reasons for this regulation is to protect investors. The regulation of non-GAAP reporting does not make non-GAAP reporting mandatory. However, the regulation shows how to construct non-GAAP measures if they are disclosed, this is to reduce the potential of being misled by these numbers.
2.1.4 Non-GAAP measures: what is it?
Generally Accepted Accounting Principles (non-GAAP) are often used. Non-GAAP financial figures are created by managers but not according to Generally Accepted Accounting Principles (GAAP). Non-GAAP earnings arise from exclusions made from GAAP earnings that managers believe are irrelevant for determining the financial performance of the firm (Entwistle et al., 2006).
Managers might believe that GAAP numbers on its own do not provide a full picture of their business and financial performance and therefore non-GAAP measures are added as supplements (Deloitte, 2017a). The SEC defines a non-GAAP measure as:
“A numerical measure of a registrant’s historical or future financial performance, financial position or cash flows that:
(i) Excludes amounts, or is subject to adjustments that have the effect of excluding amounts, that are included in the most directly comparable measure calculated and presented in accordance with GAAP in the statement of income, balance sheet or statement of cash flows (or equivalent statements) of the issuer; or
(ii) Includes amounts, or is subject to adjustments that have the effect of including amounts, that are excluded from the most directly comparable measure so calculated and presented” (SEC, 2002).
Management of the company is responsible to construct the non-GAAP measures, the numbers are shown in press releases and are not audited by auditors (Black et al., 2017a; Chen et al., 2012). For a better understanding of what a non-GAAP disclosure is and how it is incorporated in an earnings announcement examples of Marriott International and Regeneron are given in Appendix B and C. Since non-GAAP numbers are reported voluntarily, firms disclose non-GAAP measures not in a consistent or comparable way. Therefore, it might be hard to compare the non-GAAP numbers of different firms. In (almost) all press releases that report non-GAAP measures a caution statement is included to make users aware of this inconsistency. This statement makes clear that the numbers are presented not within a specific reporting regime and are thus constructed based on management judgment. Regulation G requires that companies show the reconciliations from the non-GAAP reported number to the closest GAAP equivalent, this is also stated in the caution statement. There is no change in the caution statement after the updated C&DIs. An example of Marriott International is provided:
“In our press release and schedules, and on the related conference call, we report certain financial measures that are not required by, or presented in accordance with, United States generally accepted accounting principles (“GAAP”). We discuss management’s reasons for reporting these non-GAAP measures below, and the press release schedules reconcile the most directly comparable GAAP measure to each non-GAAP measure that we refer to (identified by a double asterisk on the preceding pages). Although management evaluates and presents these non-GAAP measures for the reasons described below, please be aware that these non-GAAP measures have limitations and should not be considered in isolation or as a substitute for revenue, operating income, income from continuing operations, net income, earnings per share or any other comparable operating measure prescribed by GAAP. In addition, we may calculate and/or present these non-GAAP financial measures differently than measures with the same or similar names that other companies report, and as a result, the non-GAAP measures we report may not be comparable to those reported by others” (Marriot International, 2016a).
2.2 Non-GAAP measures: two motives
Already at the end of the last century, non-GAAP measures were used. From 1998 onwards, non-GAAP measures are increasingly used over time (1998 – 2006) except for a two-year dip around SOX and Regulation G (Black and Christensen, 2009). This number only increased further over time. In the fourth quarter of 2016, 96% of the S&P500 used non-GAAP measures in their press releases (Coleman and Erickson, 2017). Moreover, the reported non-GAAP earnings number is almost always higher than the reported GAAP earnings number. In the period 1986-1997, there was an increase in the number of exclusions made in non-GAAP earnings measures (Bradshaw and Sloan, 2002). This means there was a growing gap between the GAAP and non-GAAP numbers. Often non-recurring, transitory expenses are excluded in calculating the non-GAAP number (Black and Christensen, 2009). Two motives can dominate as the reason why firms would disclose non-GAAP information. First, the informative motive states that firms disclose non-GAAP measures to give better information to investors and therefore reduce the information asymmetry. Under this motive, non-GAAP earnings are also seen as a better predictor of future profits. Second, the opportunistic motive entails that managers disclose non-GAAP information opportunistically to potentially mislead investors. Which motive dominates is still highly debated. Now, I turn to evidence found for both motives.
2.2.1 Informative motive
The main argument managers use to report non-GAAP numbers is that these numbers provide a clearer picture of “core earnings”. Evidence is found that shows that non-GAAP earnings are more informative, more permanent and a better predictor of future performance (Bhattacharya et al., 2003; Entwistle et al., 2010; Weil, 2001a). Market participants believe non-GAAP earnings are a better representation of core earnings compared to the GAAP numbers (Bhattacharya et al., 2003). Non-GAAP measures exclude certain items that are included under GAAP measures. These exclusions are generally income statement or balance sheet items. Evidence shows that mostly the non-recurring or transitory items such as restructuring charges are excluded (Black and Christensen, 2009). Other examples of potential exclusions are non-cash items such as asset impairments and amortization of intangibles (Larcker and Tayan, 2010). Firms believe the non-GAAP earnings show the persistent part of earnings and will give investors better information to predict future performance (Philips et al., 2002).
Curtis et al. (2014) and Bhattacharya et al. (2003) investigate if the primary motivation is to mislead or to inform, by investigating how management responds to a transitory gain. Since non-GAAP earnings will exclude transitory items to reflect core earnings, a transitory gain will lower the non-GAAP number. Although managers’ disclosure behavior differs across firms, they find that the informative motive dominates. Thus, managers are also willing to exclude positive transitory items, this highlights that managers’ true motivation is to show core earnings instead of overstating current operating performance (Bhattacharya et al., 2003; Black et al., 2009; Curtis et al., 2014).
More evidence of the informative motive comes from Leung and Veenman (2018). This research shows that non-GAAP reporting is especially informative for loss firms. They find that the expenses excluded from GAAP earnings to arrive at non-GAAP earnings are of not very informative. The information excluded from the non-GAAP number is so noisy that this distorts the informativeness of the GAAP number. They suggest that a loss firm that reports non-GAAP earnings will have significantly better future performance than firms only presenting GAAP earnings. This is because non-GAAP disclosures are more informative and predictable for investors. Therefore, investors are better able to forecast and value loss firms that report non-GAAP disclosures (Leung and Veenman, 2018).
2.2.1.1 Investors
Investors are, together with analysts, the most studied group regarding earnings announcements. Although investors are free to ignore non-GAAP earnings, plenty of literature find that investors do not (Weil, 2001b). There is a large literature that examines investors’ preference for GAAP and non-GAAP earnings. This is mostly done by examining associations between stock prices and GAAP and non-GAAP earnings surprises. Bradshaw and Sloan (2002) are among the first ones who compare GAAP and non-GAAP earnings relative to stock prices. Even though there was an increasing gap between GAAP and non-GAAP earnings around the end of last century the explanatory power of non-GAAP became significantly higher than GAAP earnings. They find evidence that investors gained the preference to follow the non-GAAP earnings number instead of the GAAP earnings number. The non-GAAP number replaced the GAAP profit number as the primary determinant for stock prices (1985-1997) (Bradshaw and Sloan, 2002).
Both Bhattacharya et al. (2003) and Lougee and Marquardt (2004) follow the same line of reasoning and conclude that investors see the non-GAAP earnings number as more informative and closer to core earnings of firms. Bhattacharya et al. (2003) find that 66% of the non-GAAP earnings announcements found a profit while only 52% of GAAP earnings were profitable. Investors do not discount non-GAAP numbers and do not think they are less reliable (Bhattacharya et al., 2003). Evidence shows that investors see non-GAAP earnings disclosures as far more value relevant than GAAP earnings disclosures. Furthermore, they are also more closely related to stock prices than GAAP earnings (Entwistle et al., 2010). Moreover, Black et al. (2012) find evidence that investors focus on both GAAP and non-GAAP earnings numbers but they find a higher magnitude for non-GAAP earnings. This means investors pay more attention to, an put more weight on non-GAAP earnings.
Johnson and Schwartz (2005) find evidence that despite concerns of regulators about non-GAAP disclosures, investors are not misled by the disclosure of non-GAAP earnings. Based on experimental findings, Elliot (2006) confirms the above finding. Specifically, Elliot (2006) states that the presence of a reconciliation led analysts and investors to view non-GAAP earnings as more reliable. Such a reconciliation is mandatory after SOX and Regulation G. Therefore, investors and analysts increasingly rely on non-GAAP numbers when making judgments of firms. However, the emphasis management places on non-GAAP disclosures can affect the judgment of non-professional investors (Elliot, 2006).
Bradshaw et al. (2017) confirm above findings, although they state that previous studies are subject to measurement error. Bradshaw et al. (2017) conclude that these studies have a downward bias in investors’ responses to GAAP earnings. Fortunately, they find that this bias is not big enough to overturn prior conclusions. Moreover, they also find, while correcting for this measurement error, that investors respond more to non-GAAP earnings relative to GAAP earnings. This concludes that non-GAAP earnings are seen as very informative by investors.
2.2.2 Opportunistic motive
The other stream of literature argues that managers report non-GAAP earnings opportunistically. Many papers found evidence about the opportunistic motive of management, see for example Barth et al. (2012); Black and Christensen (2009); Bowen et al. (2005); Brown et al. (2012a) and Doyle et al. (2003).
In contrast to what is found by Larker and Tayan (2010) other studies conclude that managers exclude items opportunistically (Black and Christensen, 2009; Brown et al., 2012a; Whipple, 2015). Whipple (2015) divides the exclusions made between ‘special items’ and ‘other items’. ‘Special items’ are one-time non-recurring items. These items are excluded from non-GAAP earnings to give a better understanding of the company’s core performance. ‘Other items’ are recurring expenses, the explanation from management to exclude these items is less clear (Whipple, 2015). Whipple (2015) find evidence that 78% of the companies in his sample excludes ‘other items’. These exclusions are susceptible to be misleading and are seen as opportunistic disclosures. Black and Christensen (2009) find that the most frequently-used exclusions are recurring items such as depreciation and amortization, R&D expenses and stock-based compensation. These exclusions are not on a one-time basis and therefore non-GAAP earnings will probably not reflect recurring income but can easily mislead investors. Barth et al. (2012) find evidence that managers opportunistically exclude expenses to increase earnings, smooth earnings and meet earnings benchmarks and they did not find evidence that the non-GAAP earnings measure better predicts future performance.
Moreover, Curtis et al. (2014) find that even though the informative motive dominates in their study, there is also evidence for the opportunistic motive with regards to transitory gains. A portion of the firms they investigated seem to be opportunistic in the sense that they only disclose non-GAAP earnings information when it increases investor’s perceptions of core operating earnings. These firms only disclose non-GAAP earnings in transitory loss quarters, but not in transitory gain quarters. A fraction of their sample discloses non-GAAP earnings that exclude transitory losses but include transitory gains. This can be seen as misleading.
Another indicator that is often investigated to show opportunistic behavior of management is the strategic use of exclusions to meet the analyst’s earnings benchmark. Non-GAAP earnings are more likely to meet or beat analysts’ forecasts than the Non-GAAP number of the same firm (Bhattacharya et al., 2003; Black and Christensen, 2009). Bhattacharya et al. (2003) results’ indicate that over 80% of the non-GAAP numbers meet the earnings benchmark while only 39% of the GAAP measures meets or beats analysts’ forecasts. This suggests that non-GAAP numbers might be constructed in a misleading manner to meet or beat the analyst forecast.
Furthermore, Isidro and Marques (2009) investigate five earnings benchmarks and the effect on non-GAAP reporting. They find that all five (analyst forecast, industry performance, last years’ earnings, avoid losses and higher performance) have a positive effect on the propensity of managers to disclose non-GAAP figures. The analyst forecast is the strongest motivation to report the non-GAAP profit. This shows that managers can use opportunistic non-GAAP earnings numbers when they want to meet earnings benchmarks. Non-GAAP earnings are seen as a tool for earnings management to meet analyst forecasts. Managers define the non-GAAP number opportunistically to exceed analysts’ forecasts and analysts do not fully incorporate this behavior in their forecasts (Black et al., 2017a; Doyle et al., 2013).
2.2.2.1. Investors
Evidence shows that investors do not fully understand that certain expenses excluded in non-GAAP earnings are associated with lower future firm performance. Therefore, investors can possibly be misled by non-GAAP metrics (Brown et al., 2012a; Brown et al., 2012b; Doyle et al., 2003). Doyle et al. (2013) find evidence that managers opportunistically define non-GAAP earnings to meet analyst’ expectations. Investors only partly anticipate this behavior by discounting positive earnings surprises. This shows the market is partially efficient in identifying and penalizing firms that meet earnings benchmarks because of opportunistic behavior.
Brown et al. (2012a) find evidence that managers opportunistically time their earnings releases that contain a non-GAAP earnings number in an effort to alter investors’ perceptions of firm performance. When the earnings release contains a non-GAAP measure, managers accelerate the timing of the earnings announcement relative to when the earnings release does not contain a non-GAAP measure. The acceleration of earnings release increases with the amount of recurring expenses being excluded. They find evidence that often, recurring items are excluded. This is evidence for the opportunistic motive. Moreover, they also find evidence that investors fail to fully anticipate the predictive power of the recurring item exclusions that are incorporated in the non-GAAP measure. Another research by Brown et al. (Brown et al., 2012b) investigates investor sentiment on non-GAAP disclosures. They suggest that if investor sentiment is high firms are: (1) more likely to disclose non-GAAP earnings; (2) more likely that the non-GAAP number exceeds the GAAP number; (3) the number of recurring exclusions and (4) the non-GAAP number is placed with higher prominence. They find substantive evidence that managers create non-GAAP numbers in an opportunistic manner and this increases with investor sentiment. This opportunism is used to mislead investors.
Plenty of literature makes a distinction between less-sophisticated investors and sophisticated investors, see for example Allee et al. (2007); Bhattacharya et al. (2007) and Christensen et al. (2014). Evidence shows that primarily the less-sophisticated investors trade on non-GAAP earnings. Less-sophisticated investors lack the necessary sophistication and knowledge to understand the accuracy and reliability of the non-GAAP disclosures. Therefore, less-sophisticated investors are most likely to be misled. On the other hand, short sellers are able to incorporate manager’s opportunistic behavior. Short sellers are a group of extremely well informed sophisticated investors. Some short sellers see non-GAAP disclosures as a signal of decreased reporting quality or overvaluation and also as an opportunity to gain an informational advantage. Short sellers are able to see through non-GAAP earnings and target firms with aggressive exclusions (Christensen et al., 2014). Less-sophisticated investors are most likely to be misled by the opportunistically derived non-GAAP earnings number (Allee et al., 2007; Bhattacharya et al., 2007).
In sum, evidence is found for both motives. While some research claims that investors can anticipate on the misleading potential of the non-GAAP measures other evidence states that investors can only partly anticipate this and will potentially be misled. To note, it is not the goal of this thesis to examine which motive dominates.
2.3 Regulation
2.3.1 SOX & Regulation G
In 2001, the Securities and Exchange Commission (SEC) issued a warning that non-GAAP reporting in certain circumstances “can mislead investors if it obscures GAAP results” (SEC, 2001). The SEC recognized that although non-GAAP numbers can add value there is a potential to mislead the market and thus investors. This is mainly because non-GAAP financial information has no defined meaning and no uniform characteristics (SEC, 2001). Due to increased scrutiny, the SEC introduced the Sarbanes-Oxley Act (SOX) in 2002. The widespread public concern and scandals such as Enron led to the introduction of SOX. SOX tried to improve the accuracy and reliability of corporate disclosures and tried to restore public confidence. Section 401(b) of SOX assigns the SEC to introduce regulation about non-GAAP reporting.
As a result, the SEC issued Regulation G on March 28, 2003. Regulation G has a general disclosure and a specific content requirement. The general disclosure requirement states the following:
“A registrant or a person acting on its behalf, shall not make public a non-GAAP financial measure that, taken together with the information accompanying that measure, contains an untrue statement of a material fact or omits to state a material fact necessary in order to make the presentation of the non-GAAP financial measure, in light of the circumstances under which it is presented, not misleading.” (SEC, 2002).
The specific content requirement states that companies should present the most directly comparable GAAP financial measure. Moreover, a quantitative reconciliation between the non-GAAP and the most comparable non-GAAP number needs to be provided (SEC, 2002).
Next to this, the SEC also made amendments to Regulation S-K item 10(e). Item 10(e) expands regulation G with the following requirements. Firstly, the prominence of the most directly comparable GAAP measure must be equal to or greater than that of the non-GAAP measure. Secondly, there should be a statement of the author indicating reasons why the non-GAAP measure provides useful information to investors. Lastly, companies should also disclose a statement which presents additional purposes (if any) for which the firm will use the non-GAAP measures (SEC, 2002). In addition to the regulation, the SEC published 33 Frequently Asked Questions (FAQs) to help users comply with the rules.
2.3.2 Compliance and Disclosure Interpretations (C&DIs)
In 2010, regulators were still concerned with the potential of non-GAAP disclosures to mislead investors. In 2010, the SEC labeled non-GAAP earnings measures as ‘fraud risk factors’ (Leone, 2010). In the same year, the SEC changed these FAQs into a guidance of Compliance and Disclosure Interpretations (C&DIs) (Deloitte, 2017b).
At the end of 2015, SEC officials started to discuss non-GAAP measures at various public venues. The SEC worried about companies’ extensive use of non-GAAP measures. The renewed focus on non-GAAP reporting was because of several concerns: (a) increased use and prominence of the measures, (b) the nature of the adjustments, (c) an increasingly large difference between the GAAP and non-GAAP numbers (Deloitte, 2017a).
In response to these concerns about non-GAAP measures, the SEC updated the Compliance and Disclosure Interpretations (C&DIs) in May 2016. The SEC presents the new C&DIs based on questions and answers and is updated regularly (SEC, 2017). The SEC expected that the updated C&DIs would promote changes in the use of non-GAAP measures. The SEC has two specific expectations:
- After the updated C&DIs non-GAAP measures will be less misleading.
- After the updated C&DIs non-GAAP measures will be presented less prominently (Deloitte, 2017a).
The first expectation is based on a change in the composition of the non-GAAP numbers, i.e. how the measures are constructed. The second one expects that there is a change in the place of the non-GAAP measures in the press release. The C&DIs are specified into four different topics: prominence, misleading measures, per share measures and tax effects. These topics will now be discussed one by one. For a more complete picture of the updated C&DIs, I added an explanation that shows an overview of the new requirements of the updated C&DIs per topic. For this, I refer to Appendix D.
As is stated above, the most comparable GAAP measures with equal or greater prominence should be presented together with a non-GAAP measure. However, until the new C&DIs, there was no formal guidance on this which resulted in diversity. The updated guidance gives examples of disclosures when a non-GAAP measure will be more prominent than the most comparable GAAP measure and thus fails to meet the prominence requirement. Examples given are omitting comparable GAAP measures from headlines or captions and presenting a
GAAP measure using a style of presentation (e.g., bold, larger font) that emphasizes non-GAAP over the comparable non-GAAP measures (PWC, 2016).
One of the most outstanding themes of the whole regulation of non-GAAP measures is to prevent these measures to be misleading. The updated C&DIs provide guidance on presentations that the SEC might see as misleading. Examples are presenting non-GAAP measures that exclude normal, recurring items; non-GAAP measures excluding non-recurring expenses but including non-recurring gains; present non-GAAP measures inconsistently over time (Deloitte, 2017b).
The third theme discussed in the C&DIs is per share measures. A non-GAAP measure should be defined as a performance measure or a liquidity measure. This determination influences the item the non-GAAP measure should be reconciled to. Performance measures are usually reconciled to net income while liquidity measures should normally be reconciled to cash flows from operations. The updated C&DIs clarify that non-GAAP liquidity measures are not allowed to be presented on a per share basis (PWC, 2016). It is allowed to disclose a non-GAAP per share performance measure as long as it can be determined as a performance measure. The updated guidance states that when non-GAAP measures are analyzed they will focus on the substance of the non-GAAP measure, not on the characterization given (Deloitte, 2017b).
Lastly, non-GAAP measures might be disclosed after tax and thus show the tax adjustments when reconciling from a GAAP measure. The C&DIs described criteria how income tax effects of non-GAAP measures should be calculated and presented. The tax expense of a performance measure should be consistent with the amount of non-GAAP income. The C&DIs also show that all the adjustments made should be shown gross of tax and there should be a separate adjustment and explanation for the tax expense (Deloitte, 2017b; PWC, 2016). Figure 1 shows a timeline of the SEC guidance.
After the C&DIs were updated, the SEC increased the focus on how non-GAAP measures are used and presented. This is done during routine filing reviews of annual and quarterly reports and press releases. The non-GAAP measures are the second most frequent topic of comment for the SEC. The SEC started to target reviews of earnings releases that are furnished to the SEC, this was specifically on 8-K releases (Deloitte, 2017b). Overall, this leads to increased attention from the SEC for non-GAAP disclosures.
2.4 The effect of Regulation
Multiple studies investigated the impact of regulation, both SOX and Regulation G, on non-GAAP reporting. In the short term there is a decline in non-GAAP earnings disclosures in the period from 2001-2003 (Entwistle et al., 2006; Heflin and Hsu, 2008; Marques, 2006). In 2001, over 77% reported non-GAAP earnings, this was only 54% in 2003 (Entwistle et al., 2006). Moreover, next to the frequency also the magnitude of non-GAAP exclusions decreased. The non-GAAP earnings are reported in a less income-increasing and thus less misleading measure after regulation. In 2001, 85% of the firms had a bigger non-GAAP number than the GAAP equivalent, in 2003 this was only 67% and there was also a sharp decline in the difference between the two numbers (Entwistle et al., 2006). Furthermore, the non-GAAP number is presented in a less prominent manner. Lastly, there is a modest decrease in the likelihood that non-GAAP earnings meet or beat analyst forecasts (Heflin and Hsu, 2008). Thus, in the short term SOX led to a decrease in opportunistic reporting and thus non-GAAP reporting that is less likely to mislead investors.
Looking at the somewhat longer term, non-GAAP reporting increased over time from a period from 1998 to 2007, except for a two-year period that Entwistle et al. (2006) reported about (Black et al., 2012). Thus, although reporting requirements increased significantly over time, firms keep on using non-GAAP numbers. This might support the informative motive for managers to report a non-GAAP earnings number. SOX and Regulation G require companies to show how they construct their non-GAAP earnings number, this gives investors more information and they are therefore less likely to be misled after regulation. It appears that
managers are less likely to use non-GAAP exclusions aggressively after SOX. Whipple (2015) finds that after regulation, excluding ‘other items’ is highly informative because these are non-cash items that are discounted by investors when firm performance is priced. The exclusions made in the non-GAAP numbers after regulation are comparable to analysts’ earnings forecasts. This gives investors comparable information and is therefore very useful. However, Whipple (2015) also find some evidence that opportunism can still be the motivation to exclude ‘other items’ but he states that regulation causes the primary motive for the adjustments to be informative.
Evidence is found that, after SOX was implemented, investors pay attention to both GAAP and non-GAAP earnings numbers, but investors pay more attention to non-GAAP earnings disclosures (Black et al., 2012; Marques, 2006). Moreover, investors discount earnings announcements in which managers make aggressive exclusions in their non-GAAP earnings number more after SOX. SOX and Regulation G have, to a certain degree, resulted in bringing non-GAAP disclosures to investors’ attention and improved the perceived quality of these disclosures (Black et al., 2012). Black et al. (2012) states that it appears that Regulation G and SOX have achieved its intended purpose. They find that investors pay more attention to and react more upon non-GAAP earnings in after SOX. Brown et al. (2012a) find that SOX reduces aggressive non-GAAP reporting. Moreover, they state that SOX increased investors’ ability to recognize and adjust for the recurring exclusions that are in the non-GAAP earnings announcements.
Again, transitory gains can reveal something about the opportunistic motives of managers. A transitory gain would result in a lower non-GAAP earnings number than the GAAP earnings number since transitory items are normally excluded from the non-GAAP number. Baumker et al. (2014) find that management often disclose a transitory gain but they are less likely to exclude this gain from the non-GAAP number. Prior to Regulation G 62% of firms excluded transitory gains from the non-GAAP earnings number, after Regulation G this was only 34%. Baumker et al. (2014) state that this might be an unintended consequence of regulation. Moreover, transitory gains are less likely to be incorporated in the non-GAAP number when there are no transitory losses. This shows that managers are likely to prefer non-GAAP earnings that improve performance. Thus, one can see that there is still an opportunistic motive on non-GAAP reporting even after regulation.
Recent evidence comes from Black et al. (2017b). They show that research about non-GAAP disclosures is still relevant. They also looked at the influence of SOX on non-non-GAAP reporting. They found that managers exclude fewer recurring items on which analysts disagree after SOX. Managers are also less likely to make recurring exclusions to meet or beat analyst’ forecasts on non-GAAP when firms would have missed their GAAP EPS forecast. However, the evidence is not all in the same direction. While managers are less likely to exclude amortization or depreciation expenses in their non-GAAP earnings number, they also find that some firms still exclude recurring items such as interest, tax and stock-based compensation in the post-SOX period. This suggests that regulation has influenced aggressive non-GAAP reporting but some firms still seem to encounter aggressive non-GAAP exclusions and thus can still be misleading.
To conclude, the evidence found above is important evidence for regulators in the United States since this tells something about the effectiveness of regulation. Moreover, it can also be of relevance to regulators of other countries because they see if and how regulation can be effective. Above literature suggests that regulation has partly achieved its goal by reducing opportunistic reporting but it still has the potential to be misleading.
26
3. Hypothesis development
The literature review shows that non-GAAP reporting still has the potential to be misleading (Black et al., 2017b). The SEC is also still worried about the misleading potential of non-GAAP reporting and therefore updated the C&DIs. The SEC had the expectations that the updated C&DIs would promote changes in the use of non-GAAP measures, these expectations are mainly related to potentially misleading measures and misplaced prominence of the non-GAAP measures (Deloitte, 2017a). The influence of SOX and Regulation G is extensively investigated, the impact of these updated C&DIs is not.
One of the concerns the SEC had was the prominence of non-GAAP numbers in press releases. The SEC expected that the updated C&DIs will promote changes in the use of non-GAAP measures, under which the misplaced prominence of the non-non-GAAP measures (Deloitte, 2017a). Therefore, a different presentation of the non-GAAP number in press releases is expected after the updated C&DIs. Placed with less prominence entails that the non-GAAP earnings number in the press release is presented after the GAAP earnings number. Moreover, the non-GAAP number is not allowed to be disclosed in the headlines when the GAAP equivalent is omitted from the headlines. The first hypothesis examines whether this expectation of the SEC is met (see Figure 2 for Libby Boxes). It is stated as follows:
Hypothesis 1: The non-GAAP EPS number gained less prominence in the press releases after the updated C&DIs.
Updated Compliance and Disclosure Interpretations (C&DIs)
Prominence of non-GAAP EPS in press release
(Place in press release) Non-GAAP EPS one quarter
before and one quarter after the updated C&DIs
Prominence of
non-GAAP measures Conceptual
Operational
Another concern of the SEC was the misleading potential of non-GAAP measures. The increased difference between the amounts reported under non-GAAP earnings and GAAP earnings gained specific attention (Deloitte, 2017a). Non-GAAP earnings are often higher than the GAAP earnings since non-recurring costs are excluded in the non-GAAP earnings number. But as stated in the literature review, not only non-recurring items are excluded, also recurring items such as stock-based expenses are excluded even after regulation (Black et al., 2017b). The updated C&DIs more specifically state which items are allowed to be excluded from the non-GAAP earnings. Therefore, less opportunistic exclusions are expected. This brings the non-GAAP earnings number closer to the GAAP earnings number. This suggests that the difference between the two earnings numbers should be decreased after the updated C&DIs and therefore less misleading. The updated C&DIs also state that non-recurring gains should be included. Baumker et al. (2014) found that after SOX this was not always the case, but the C&DIs provide clearance that transitory gains should be included. The SEC expects that the measures are less misleading after the updated C&DIs. To test this expectation, the second hypothesis is stated as follows (see Figure 3 for Libby Boxes):
Hypothesis 2: The difference between non-GAAP and GAAP EPS number decreased after the updated C&DIs.
Updated Compliance and Disclosure Interpretations (C&DIs)
Difference between the GAAP EPS and non-GAAP EPS Two samples: pre-C&DIs
and post-C&DIs
Misleading potential of
non-GAAP measures Conceptual
Operational
To note, the first hypothesis focuses on prominence, which is determined by the place of the non-GAAP number in the press release. On the other hand, the second hypothesis focuses on misleading measures, this entails that it focuses on the content of the non-GAAP measures, how the measures are constructed instead of its place.
As is shown in the literature review, investors focus more on non-GAAP earnings than on GAAP earnings and this is even more after regulation (Black et al., 2012; Bradshaw et al., 2017; Marques, 2006). I expect that the updated C&DIs will affect investors’ perceptions of non-GAAP reporting. The updated C&DIs’ intention is to increase the informativeness and comparability of non-GAAP measures. The non-GAAP measures should be constructed in a way that has less potential to mislead investors. For example, it is expected that due to the updated C&DIs only special items are excluded. Furthermore, investors can also better identify and understand when a measure is constructed in an opportunistic way. Therefore, they might be less scared to rely to non-GAAP measures after the new regulation. Since the updated C&DIs will give more clarity and transparency to the non-GAAP measures, investors could be more confident and feel more comfortable to rely on those measures than before. Thus, it is expected that investors react more to non-GAAP earnings after the updated C&DIs. Investors determine returns and therefore returns are the variable of interest in this research. The third hypothesis states:
Hypothesis 3: Returns are more driven by non-GAAP earnings after the SEC updated the C&DIs guidelines than before the C&DIs were updated.
However, we must be careful with this hypothesis. What should be kept in mind is the difference in impact between SOX & Regulation G and the updated C&DIs. SOX and Regulation G were the first concrete regulation in place in the United States, which had a big impact and changed a lot about the non-GAAP reporting. The C&DIs are solely updated guidelines based on existing regulation and therefore the impact might be less prominent or even absent. This hypothesis is written in alternative form, the corresponding null hypothesis is that returns are not more driven by non-GAAP earnings after the updated C&DIs compared to the period before the C&DIs. To visualize I added Libby Boxes that explain the third hypothesis (see Figure 4).
Independent Dependent
FEGAAP FEnon-GAAP
Abnormal returns (CAR)
Figure 4: Libby Boxes Hypothesis 3
Investors’ perceptions GAAP and non-GAAP
earnings
Impact of new C&DIs
Dummy variable: POSTCDI. 0 for pre-C&DIs sample 1 for post-C&DIs sample
Conceptual
4. Research Design
To test the first hypothesis, I will observe whether firms present their non-GAAP number less prominent after the updated C&DIs. This is examined during the non-GAAP hand- collection. I will observe if the prominence of the non-GAAP EPS number differs between the first quarter of 2016 and the second and third quarter of 2016. I determine a non-GAAP measure to be prominent if it is placed before the GAAP equivalent. The focus here is on the short-term. This is to make sure companies change their reporting outlook because of the updated C&DIs. There is only a one-time change in the placement of the non-GAAP number required, from being more prominent to being less prominent than the GAAP equivalent. Therefore, I think focusing on the short term is more appropriate.
To test the second hypothesis, I will calculate whether the difference between non-GAAP and GAAP information decreased after the updated C&DIs. For this, both non-GAAP and GAAP EPS information should be obtained. The sample will be divided into pre-C&DIs and post-C&DIs. The change took place in Q2 of 2016 and therefore pre-C&DIs sample is from Q2 2015 to Q1 2016 while the post-C&DIs sample is from Q2 2016 till Q1 2017. The difference between the GAAP EPS number and the non-GAAP EPS number will be calculated. After this, I test whether the difference in the post-C&DIs will be smaller or bigger than the difference in the pre-C&DIs.
To test the third hypothesis, which suggests that investors’ reaction to non-GAAP earnings numbers are different in the pre- and post-C&DIs, I will investigate whether investors react to the updated C&DIs by looking at returns. Because the unit of analysis is quite similar to Black et al. (2012), I follow their approach. The research will be based on quarterly earnings announcements and quarterly stock prices. This is because the update was only in 2016, looking at yearly data will not result in a reliable research since not enough information is available. The analysis will contain a cross-time setting, by investigating four quarters prior to the updated C&DIs and four quarters after the updated C&DIs. I will use a pre-C&DIs sample and post-C&DIs sample and estimate the following regressions. Model 1-3 will be estimated twice, both for the pre- and post-C&DIs sample. The models are as follows:
Model 1: 𝐶𝐴𝑅 = 𝛼0+ 𝛼1 𝐹𝐸𝐺𝐴𝐴𝑃 + ℇ Model 2: 𝐶𝐴𝑅 = 𝑏0+ 𝑏1𝐹𝐸𝑛𝑜𝑛𝐺𝐴𝐴𝑃 + ℧ Model 3: 𝐶𝐴𝑅 = 𝑐0+ 𝑐1𝐹𝐸𝐺𝐴𝐴𝑃 + 𝑐2𝐹𝐸𝑛𝑜𝑛𝐺𝐴𝐴𝑃 + 𝜆 Model 4: 𝐶𝐴𝑅 = 𝑑0+ 𝑑1𝐹𝐸𝐺𝐴𝐴𝑃 + 𝑑2𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝑑3𝐹𝐸𝐺𝐴𝐴𝑃 𝑥 𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝛾 Model 5: 𝐶𝐴𝑅 = 𝑒0+ 𝑒1𝐹𝐸𝑛𝑜𝑛−𝐺𝐴𝐴𝑃 + 𝑒2𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝑒3𝐹𝐸𝑛𝑜𝑛−𝐺𝐴𝐴𝑃 𝑥 𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝛼 Model 6: 𝐶𝐴𝑅 = 𝑓0+ 𝑓1𝐹𝐸𝐺𝐴𝐴𝑃 + 𝑓2𝐹𝐸𝑛𝑜𝑛−𝐺𝐴𝐴𝑃+ 𝑓3 𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝑓4 𝐹𝐸𝐺𝐴𝐴𝑃 𝑥 𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝑓5 𝐹𝐸𝑛𝑜𝑛−𝐺𝐴𝐴𝑃 𝑥 𝑃𝑂𝑆𝑇𝐶𝐷𝐼 + 𝜇 Where:
CAR Abnormal returns cumulated over the 3-day window surrounding the quarterly earnings announcement.
FEGAAP GAAP operating earnings per share minus the mean analysts’ GAAP
earnings forecast per share scaled by stock price five days before the earnings announcement.
FEnon-GAAP Non-GAAP earnings per share minus the mean analysts’ non-GAAP
earnings forecast per share scaled by stock price five days before the earnings announcement.
POSTCDI Dummy variable that equals 1 if the quarter is post-C&DIs and 0 if the quarter falls before the C&DIs update.
The models described above examine if returns are driven on both non-GAAP and GAAP earnings and which of these earnings numbers dominates. The first two models look at the informativeness of GAAP earnings and non-GAAP earnings respectively. Model 1 tests whether GAAP earnings drive returns, Model 2 does the same for non-GAAP earnings. The third model combines both earnings numbers and tests whether returns are driven on both earnings numbers. The magnitude of the different earnings might differ. These models are estimated both in the pre- and post-C&DIs period to see whether returns are driven on non-GAAP and non-GAAP earnings in both periods. To say, with these models I will test whether investors pay attention and react to both non-GAAP and GAAP earnings numbers.
Model 4-6 adds an indicator variable to address the difference between the pre- and post-C&DIs periods. The POSTCDI variable test whether the magnitude of market reactions to earnings announcements that include a non-GAAP number is different in the post-C&DIs period. The models test whether there is a difference in the extent to which investors react to GAAP earnings and non-GAAP earnings in the post-C&DIs period.
5. Sample selection
I collect a large sample of quarterly earnings press releases that contain a non-GAAP earnings number. Many studies use I/B/E/S earnings as a proxy for non-GAAP earnings. These are street earnings reported by analyst forecast tracking services. However, this method is criticized. Entwistle et al. (2010) state that non-GAAP earnings from press releases are far more value relevant than the I/B/E/S number. Next to that, Bhattacharya et al. (2003) find a statistically significant mean difference of approximately 4 cents per share between non-GAAP earnings identified in press releases and the I/B/E/S number. Moreover, the paper Bentley et al. (2018) criticize the use of I/B/E/S as a proxy for non-GAAP EPS. They state that if I/B/E/S is used to identify manager’s non-GAAP disclosures, the aggressiveness of reporting is underestimated. Often the I/B/E/S number gives the same number as the number in the press release. However, using I/B/E/S as a proxy for non-GAAP disclosure systematically misses some manager-disclosed non-GAAP numbers and incorrectly includes some analyst-adjusted metrics that managers do not explicitly report (Bentley et al., 2018). Therefore, hand-collecting seems to be the most reliable and relevant method.
I collect data from the S&P 500 firms, the S&P 500 is chosen because of the economic importance in both the United States and globally and it is often used in other non-GAAP literature (Baumker et al., 2014; Entwistle et al., 2010; Marques, 2010). Moreover, Coleman and Erickson (2017) find that 96% of the S&P 500 firms presented a non-GAAP number in their press releases during the fourth quarter of 2016. Therefore, I expect that non-GAAP earnings numbers are largely available. Because I am primarily interested in the effects of the updated C&DIs, I separate the sample into pre-C&DIs and post-C&DIs. The pre-C&DIs sample is defined as earnings from the second quarter of 2015 to the first quarter of 2016. The post-C&DIs sample ranges from the second quarter of 2016 to the first quarter of 2017. However, it is initially not clear if the C&DI change reporting immediately. Therefore, I decided to collect all the data till the second quarter of 2017. This gives the ability to choose the exact composition of the pre- and post-C&DIs sample after I have done some analysis. It might be the case that it takes some time for firms to adapt to the updated C&DIs. After the hand-collection process, I find that it is accurate to start the post-C&DIs sample in the second quarter of 2016 because firms seem to adapt immediately after the update.
