The Use of Revenue Disclosures to Inform and Influence the Market

Transcription

1 The Use of Revenue Disclosures to Inform and Influence the Market Presented by Dr Stephen Stubben Associate Professor The University of Utah # 2014/15-09 The views and opinions expressed in this working paper are those of the author(s) and not necessarily those of the School of Accountancy, Singapore Management University.

2 The Use of Revenue Disclosures to Inform and Influence the Market September 2014 Lorien Stice-Lawrence University of North Carolina at Chapel Hill Stephen R. Stubben University of Utah * Preliminary and incomplete. Please do not cite or quote without permission.

3 The Use of Revenue Disclosures to Inform and Influence the Market Abstract: This study addresses whether firms use revenue disclosures to inform and influence the market. We construct a revenue disclosure index of how many commonly discussed revenuerelated topics are included in each firm s 10-K filing. As expected, we find that revenue disclosure is associated with various determinants of corporate disclosure, and it is negatively associated with absolute earnings forecast errors and cost of equity capital. Regarding revenuespecific outcomes, revenue disclosure is associated with a greater likelihood of having a revenue forecast, smaller revenue forecast errors, and a greater weight on revenues in valuation. These results suggest that revenue disclosures provide valuable information to the market about future revenues and influence the market to place a greater weight on revenue in valuation.

4 1. Introduction This study addresses whether firms use revenue disclosures to inform and influence the market. A large literature in accounting has examined corporate disclosure and shows that disclosure can have significant effects on such outcomes as liquidity (Diamond and Verrecchia 1991, Leuz and Verrecchia 2000) and cost of capital (Botosan 1997, Lambert et al. 2012), among others. The combined body of evidence suggests that an increase in information disclosed by firms will benefit market participants. However, the prior literature has largely overlooked specific disclosure topics, instead focusing on more general measures, for example aggregate disclosure indices, management earnings guidance, or accrual quality measures. While useful, these measures ignore the fact that specific types of disclosure convey different types of information and may have different implications for financial statement users. In our paper, we measure dimensions of one specific disclosure type, revenue-related disclosure, and demonstrate that it is informative to financial statement users and affects the way that they use other financial information. We introduce and utilize a new measure of corporate disclosure specifically, revenue disclosure. The measure is a count of how many (out of 14) commonly discussed revenue-related topics are included in each firm s 10-K filing. We first attempt to validate the measure by showing its correlation with disclosure measures commonly used in the accounting literature and find that it is positively associated with the incidence of management earnings forecasts and the length of the 10-K, although it is not significantly associated with the readability of the 10-K (i.e., Fog index). We further validate our measure by documenting its correlation with known determinants of disclosure such as firm size, growth, profitability, and equity volatility, and show that it is higher for firms issuing debt, firms with high litigation risk, and firms engaging in 1

5 acquisitions. We also examine determinants of annual changes in disclosure and find associations between revenue disclosure changes and changes in meaningful economic determinants. Together, these findings suggest that our measure is valid proxy for disclosure that is able to capture both levels and year-by-year changes in revenue disclosure. Next, we use the revenue disclosure measure to replicate and extend prior findings on the consequences of corporate disclosure. Consistent with prior literature, we find that revenue disclosure is associated with smaller analyst earnings forecast errors and a lower cost of equity capital (Lang and Lundholm 1996, Botosan 1997). We extend prior findings on the association between disclosure and cost of capital by distinguishing one-time disclosures from a longer-term commitment to high disclosure. Consistent with the assumption in Diamond and Verrecchia (1991) that firms must commit ex ante to disclosure in order to affect their cost of capital, we document a larger effect of disclosure on the cost of equity capital when disclosure is continually high (i.e., there is a commitment to future high disclosure). Our primary research questions take advantage of the specific nature of the revenue disclosure index and address both the informational role of revenue disclosures themselves and their role in influencing how financial statement users interpret other financial information. First, we assess whether the revenue disclosures we measure are informative to the market by testing whether firms with more revenue disclosure (1) are more likely to have analyst revenue forecasts in the subsequent year and (2) have more accurate revenue forecasts in the subsequent year. We find that among firms with earnings forecasts, those with greater revenue disclosure are more likely to also have revenue forecasts. The association with revenue forecast likelihood holds even after controlling for a broad set of control variables and other general measures of disclosure. We 2

6 also find that firms with greater revenue disclosure have more accurate revenue forecasts, controlling for earnings forecast accuracy and a set of control variables. While these two findings are consistent with prior general findings on the association between disclosure and earnings forecast attributes (Lang and Lundholm 1996, Hope 2003), our focus on revenue disclosure and revenue forecasts (controlling for earnings forecast attributes) allows us to more directly link the information provided by firms and their capital market effects. Together, these findings are consistent with analysts being more willing to issue revenue forecasts for firms with greater available revenue information and with that information aiding analysts in providing accurate forecasts of revenues. Finally, we test whether and how revenue disclosures influence the market. We find that firms with greater revenue disclosure have higher revenue response coefficients, incremental to earnings response coefficients. The higher revenue response coefficient for firms with greater revenue disclosure is evident even though the actual persistence of revenues in earnings prediction does not vary with revenue disclosure. Our findings suggest that investors place a higher weight on revenues when valuing firms that have more detailed revenue disclosures, even though revenues for these firms do not have a higher ability to predict future profitability. Our study makes several contributions to the disclosure literature. First, we present a new measure of disclosure that overcomes some of the limitations of commonly used disclosure measures. Our measure is able to identify specific dimensions of disclosure, similar to studies that use hand-collected metrics (Botosan 1997, Francis et al. 2008), but it overcomes the issue of selection bias present in these studies because it is available for a broad set of firms (all those with 10-Ks posted on Edgar). It is also based on objective criteria and available each year. Second, we provide additional evidence supporting prior findings on the consequences of 3

7 corporate disclosure by documenting effects on analyst forecast accuracy and the cost of equity capital for a new measure of disclosure. Although the literature on the link between disclosure and cost of capital is somewhat contentious and problematic (see Beyer et al for an overview), the effect on analyst forecasts in particular provides evidence that revenue disclosures are informative to financial statement users and not just boilerplate. Third, we provide empirical evidence on the capital market benefits of a commitment to disclose, as opposed to one-time disclosures. Finally, we show that specific disclosures (e.g., those related to revenues) provide specific information to the market and may influence how the market prices the firm s financial information. Our study proceeds as follows. Section 2 summarizes related research and our contributions to the literature and motivates and states our primary predictions. Section 3 describes our measure of revenue disclosure, the models we use in our analysis, and the sample. Section 4 discusses the results, and Section 5 concludes. 2. Related Research and Predictions 2.1 Empirical Measures of Disclosure One contribution of this study is the introduction of a new measure of corporate disclosure that has certain advantages over commonly used measures of disclosure such as selfconstructed measures, AIMR ratings, management forecasts, length of the 10-K, and 10-K readability. As Healy and Palepu (2001) and Beyer et al. (2010) discuss, these measures suffer from a variety of problems, which limit their ability to reliably empirically document relations between disclosure and other determinants and outcomes. 4

8 For example, hand-collected, self-constructed measures, such as those used by Botosan (1997) and Miller (2002), can be labor intensive to assemble. Given the high cost, these measures are usually used on small samples, and they are subject to issues related to selection bias, generalizability of results, and replicability. AIMR ratings are commonly used in the literature as measures of disclosure quality (e.g., Lang and Lundholm, 1993, 1996; Healy et al., 1999). However, these scores may be subjective as they represent analysts perceptions of disclosure quality. Furthermore, they are available only for large firms in the 1980s and 1990s, which raises questions about selection bias and the generalizability of results, especially in more recent periods. Management earnings forecasts are considered an important form of voluntary disclosure and have been used in numerous empirical studies as a proxy for voluntary disclosure (Feng et al. 2009, Beyer et al 2010, Baginski et al. 2012). However, most empirical measures of management guidance are narrowly focused in that they only capture management s forecast of future earnings. Recent improvements in technology have opened the possibility of various machineconstructed disclosure measures. One of the most basic of these measures of disclosure is the length of the 10-K. However, this disclosure measure is ambiguous because it is not clear from the length alone whether a longer 10-K reflects greater disclosure or instead more complexity. The Fog index (Li 2008) digs deeper into the 10-K by examining complexity of disclosures, specifically the complexity of sentences (i.e., number of words per sentence) and the complexity of words (i.e., number of syllables per word). This measure has been used in a number of studies (e.g., Lehavy et al. 2011). However, Loughran and McDonald (2014) discuss issues of 5

9 misspecification and measurement error in the Fog index. They conclude that even the rough measure of the 10-K filing file size is a proxy of readability that outperforms the Fog index. We contribute to the literature on disclosure measures along two dimensions. First, we introduce a revenue disclosure index that can be used as a general measure of disclosure that overcomes many of the empirical difficulties of other disclosure measures. For example, this measure isn t as affected by selection bias because it is available for every firm with an electronic 10-K filed on Edgar. Because the index is the sum of separate disclosure terms, researchers can easily interpret what drives the level of and change in disclosure. The second contribution relates to the specific nature of disclosure our measure captures. Because it is specifically associated with revenues, it allows us to examine the particular effects of revenue-related disclosure on outcomes and gives us more confidence in the empirical links we document because they are directly tied with the nature of the disclosure we measure. Along these lines, Merkley (2014) finds that firms provide more R&D disclosure (which often relates to long-run performance) when current performance is low. This disclosure is positively related to analyst following and earnings forecast accuracy and negatively related to analyst forecast dispersion but negatively associated with earnings disclosure, emphasizing the importance of considering the fact that not all disclosure has the same purpose or effects. In a similar vein, Kravet and Muslu (2013) find that increased MD&A risk disclosures actually increase uncertainty and perceived risk, which is inconsistent with the notion that disclosure in general is always beneficial to the firm. These papers indicate that more research is needed on the determinants and outcomes of specific disclosures, as they can potentially have very different implications than more general measures. 2.2 Determinants of Disclosure 6

10 Demand for corporate disclosure arises from information asymmetry and agency conflicts between managers and outside investors (Healy and Palepu 2001). Most disclosure studies assume that managers have superior information to outside investors about their firms expected future performance. Early full disclosure theories come to the conclusion that sellers will reveal all private information because rational market participants would otherwise interpret nondisclosure as unfavorable news and consequently discount the value of the firm s assets (see Grossman and Hart 1980, Milgrom 1981, and Verrecchia 1983, 2001). However, costs to disclosure can decrease the optimal level of disclosure. For example, managers desire to withhold bad information (Dye 1985, Darrough and Stoughton 1990) or firm-specific costs such as proprietary costs (Verrecchia 1983, Hayes and Lundholm 1996) and agency costs (Baiman and Verrecchia 1996) could offset the net capital market benefit from disclosure and influence a firm to be less than fully forthcoming (Healy and Palepu 2001). An extensive empirical literature in the accounting area has studied many aspects of the determinants of accounting disclosure. These prior studies have shown that the extent, quality and frequency of disclosure is associated with firm size, performance, proprietary costs, firm events such as acquisitions or seasoned equity offerings, litigation risk, volatility, and other drivers (Waymire 1985, Lang and Lundholm 1993 and 2000, Skinner 1997, Botosan and Stanford 2000 and 2005, Ball et al. 2014). Although not the focus of our paper, we find that the drivers of our revenue disclosure measure are consistent with these prior findings, which helps to support the validity of our disclosure proxy. 2.3 Consequences of Disclosure Theoretical studies suggest that greater voluntary disclosure should lower information asymmetry, increase liquidity, and reduce the cost of capital (Glosten and Milgrom 1985, 7

11 Diamond and Verrecchia 1991, Lambert, Leuz, and Verrecchia 2007 and 2012). Other research argues, however, that increasing cost of capital effects may occur if the disclosures themselves lead to a more asymmetric information environment than would exist in their absence (Kim and Verrecchia 1994, Kondor 2012). The empirical evidence, while not conclusive, tends to support a negative association between disclosure and cost of capital. Botosan (1997) provides some evidence consistent with the cost of capital hypothesis. She finds that for firms with low analyst following, there is a negative relation between cost of equity capital and the extent of their voluntary disclosures. Shroff et al. (2013) use the 2005 Securities Offering Reform as a quasi-exogenous shock in voluntary disclosure and find a negative relation between disclosure and the cost of equity capital, while Armstrong et al. (2011) find a positive association between information asymmetry and cost of capital when markets are imperfect, but no relation when markets are perfect. Finally, Botosan and Plumlee (2002) find a negative cross-sectional relation between cost of capital and analyst rankings of annual report disclosures. However, they also find that firms cost of capital is positively related to rankings of quarterly disclosures and not associated with investor relations activities. Francis et al. (2008) measure the disclosure quality of 677 firms in 2001 and show the negative association between voluntary disclosure and cost of capital disappears or is substantially reduced after controlling for earnings quality. They also find that alternative voluntary disclosure proxies are actually associated with higher cost of capital. The conflicting evidence in this literature again underlines the importance of finding clean proxies for disclosure that can lead to clear predictions. Another strand of research on the consequences of disclosure examines more directly the information provided to the market, or the means by which information asymmetry is reduced. 8

12 Lang and Lundholm (1996) find that firms with more informative disclosure policies have larger analyst following, more accurate analyst earnings forecasts, less dispersion of analyst forecasts, and less volatility in forecast revisions. In an international setting, Hope (2003) finds disclosure is associated with more accurate forecasts. Our study adds to this empirical literature by documenting additional evidence of reductions in information asymmetry (i.e., improved analyst forecast accuracy) and capital market benefits (i.e., a lower cost of equity capital) to firms with greater disclosure. 2.4 Disclosure Commitment A crucial assumption in the theoretical models that predict effects of disclosure on outcomes such as liquidity and cost of capital (e.g., Diamond and Verrecchia 1991, Baiman and Verrecchia 1996) is that managers must commit to disclose even when news is bad. As Verrecchia (2001) discusses, only when the manager pre-commits ex ante to an ongoing policy of disclosure in the absence of any prior knowledge of the information is bad news likely to be revealed as well (Verrecchia 2001, pg. 146). Although much of the prior empirical work has measured disclosure as one-time (e.g., Botosan 1997, Botosan and Plumlee 2000, Francis et al. 2008), few studies examine sustained increases in disclosure. Healy et al. (1999) use elevated AIMR scores as a measure of disclosure commitment, while Leuz and Verrecchia (2000) use firms decision to adopt (high quality) international standards such as IAS or U.S. GAAP. However, the generalizability of both studies is limited by the fact that Healy et al. examine only 97 firms and Leuz and Verrecchia s measure of disclosure commitment is adoption of a new set of accounting standards, which also entails many other changes. We contribute to the disclosure literature by empirically testing the effect of continued vs. one-time disclosures on cost of capital for a broad sample of US firms using a broad, well-defined measure of disclosure. 9

13 2.5 Disclosure and Pricing The last literature to which our study is related is that on how disclosure affects the way that financial statement users interpret financial information. On the one hand, disclosure could help financial statement readers to more correctly use and interpret accounting numbers. On the other hand, disclosure could have the effect of drawing a disproportionate amount of attention to certain accounting numbers, potentially resulting in mispricing. Supporting the first theory, Drake et al. (2009) find that firms with better disclosure have stock prices that more accurately reflect the relative persistence of accruals and cash flows, implying that better disclosure leads to more unbiased use of other information. Similarly, an experiment by Hirst and Hopkins (1998) indicates that higher quality of disclosure of comprehensive income can increase financial statement users ability to detect earnings management. Supporting the second theory, Elliot et al. (2010) find that analysts in an experimental setting expect more mispricing of earnings management when disclosure of the earnings management is more transparent, which implies that the transparency of earnings management disclosure can increase perceptions of mispricing above and beyond the actual amount of earnings management itself. Lang and Lundholm (2000) and Jo and Kim (2007) both find evidence that some firms opportunistically increase disclosure around seasoned equity offerings, potentially in an attempt to hype the stock. Overall, it s unclear whether increased disclosure of a specific type will increase overall informativeness or cause investors to fixate unnecessarily on certain pieces of financial information. We address this question in the context of the pricing of revenues of firms with high vs. low revenue disclosure levels. 2.6 Predictions 10

14 We test whether revenue disclosures provide useful information to the market. Using analysts as a proxy for the market, we expect that if revenue disclosures are informative, then analysts would be more likely to issue revenue forecasts and those forecasts would be more accurate. These tests are similar to those in prior research that have examined associations between disclosure and properties of earnings forecasts (Lang and Lundholm 1996, Hope 2003), except that in these revenue tests we attempt to control for the general effects on earnings information and explore the incremental effect of revenue disclosure on revenue information. First, we test whether analysts are more likely to issue revenue forecasts when revenue disclosure is higher. The decision made by an analyst to cover a firm is influenced by many factors (Bhushan 1989). Rather than attempt to identify and control for all possible determinants of an analyst s decision to cover a firm, we focus our revenue forecast analysis on the subset of firms that also have earnings forecasts. The decision to supplement an earnings forecast with a revenue forecast is a choice made by the analyst, and prior research indicates that analysts are more likely to issue an accompanying revenue forecast when they have better information (Keung 2010). We test whether this decision is associated with the firm s level of revenue disclosure and thus whether this disclosure is informative to analysts. H1: Revenue disclosure is associated with a greater incidence of revenue forecasts Next, we test whether revenue forecasts are more accurate for firms that have greater revenue disclosure. If revenue disclosures provide information to analysts that is useful in predicting future revenues, we expect to see a negative association between revenue disclosure and absolute revenue forecast errors. Because our disclosure measure may capture information in 11

15 general, we isolate the revenue-related information by controlling for absolute earnings forecast errors in the test. H2: Revenue disclosure is associated with smaller revenue forecast errors Finally, we test whether revenue disclosure affects the extent to which the market uses revenue information in valuing the company. If higher levels of revenue disclosure lead the market to place a greater weight on revenues in valuation, then we expect to see higher revenue response coefficients for firms with greater revenue disclosure. H3: Revenue disclosure is associated with higher revenue response coefficients in valuation 3. Research Design and Data 3.1 Research Design Revenue Disclosure Index Our measure of revenue disclosure, REV_DISCL, is constructed by counting the number of revenue-related topics discussed in a firm s annual Form 10-K filed with the SEC. In order to identify these disclosure topics, we first identified a set of common revenue disclosure topics and the common phrases and vocabulary associated with them by having an RA manually read the 10-K disclosures of a sample of 25 training firms across several industries. This process allowed us to identify 14 common revenue disclosure categories and a list of key phrases and terminology that were associated with each disclosure topic (see Appendix B for a list of the specific disclosure categories and some example phrases). These are the revenue disclosures that we track in order to calculate REV_DISCL. 12

16 Although our measure is limited in that we will not identify revenue disclosure topics that did not appear in the 10-Ks of our training firms or which used substantially different terminology than our hand-collected sample, we believe that the sample of firms we chose in this initial exercise was sufficiently broad to capture common disclosure across a variety of industries. Even within our diverse sample of training firms, we found that revenue disclosure phrases and terminology were highly standardized. Whether our final measure is a useful measure of revenue disclosure is an empirical matter, and any noise caused by the issues above is likely to prevent us from finding significant results. After compiling the list of phrases associated with each revenue disclosure type, we wrote a program in Perl that generalized these phrases and parsed Form 10-Ks to identify whether each disclosure type appeared in firms disclosure. For some disclosure categories, this was a relatively simple operation; for example, order backlog disclosure was easily identified by searching for whether the phrase backlog occurred in the 10-K. However, other categories were more difficult to identify, for example disclosures of revenue risk. In the end, we categorized disclosures where revenues or sales were discussed in close proximity to such words as risk, sustainability, or competition as revenue risk disclosures, but this ignores revenue risk disclosures related to specific other topics that don t use any of these identifying words. Although our program had a high in-sample success rate for correctly identifying the specific phrases that we collected from annual reports without incorrectly classifying phrases into the wrong revenue disclosure category, there are undoubtedly limitations in the ability of our program to correctly identify the revenue categories we would like to track, especially when it encounters relevant phrases that were not in our training set. Again, this will decrease the ability of our measure to accurately capture the effects that we are trying to document. 13

17 After parsing each 10-K and identifying which revenue disclosure categories are present, we calculate our final revenue disclosure measure, REV_DISCL, by counting the total number of revenue disclosure categories that are present in the firm s 10-K in that year. Because we track 14 unique revenue disclosure categories, the maximum possible value for REV_DISCL is 14 and the minimum is 0. Although some sections of the 10-K are specifically devoted to revenue disclosure (e.g. the accounting policy footnote has a section devoted to revenue recognition), we identify revenue disclosure present in any part of the 10-K because we are interested in all disclosure relating to revenues which is available to financial statement users. 1 If we assume that disclosure present in all sections of the 10-K will be reviewed by at least some financial statement users, then not including revenue disclosure which is found in non-revenue-specific sections would ignore information that is available to investors Regressions We begin by establishing the properties of our revenue disclosure measure and linking it with results in the prior disclosure literature. We first examine the determinants of revenue disclosure. To do so, we estimate Equation (1): 2 REV_DISCL t = b0 + b1 ΔSALES t + b2 ΔSALES t x DECR t + b3 DECR t + b4 SIZE t (1) + b5 ROA t + b6 LOSS t + b7 BM t + b8 LEV t + b9 R&D t + b10 MERGER t + b11 DEBT_ISS t + b12 VOLE t + b13 LITIG t + e t ΔSALES is the annual change in the firm s sales revenue, divided by its total assets, and DECR is an indicator variable that equals one when ΔSALES is negative. Including the 1 Peterson (2011) examines the disclosure about revenue recognition policies in the 10-K accounting policy 2 With the exception of Equation (5), each equation is estimated with year and industry fixed effects and standard errors clustered by year and industry. Equation (5) is estimated using a logistic regression with year and industry fixed effects. 14

18 interaction between ΔSALES and DECR allows us to assess whether a different association with disclosure exists for increases and decreases in sales revenue. SIZE is measured as the natural log of the firm s total assets, ROA is net income before extraordinary items divided by total assets, and LOSS is an indicator variable that equals one when ROA is negative. Lang and Lundholm (1993) find that disclosure increases with firm size and profitability. The book-to-market ratio, BM, is measured as the book value of equity divided by the market value of equity, and leverage, LEV, equals total liabilities divided by total assets. Research and development expense (R&D), calculated as annual R&D expense divided by total assets, is argued to be positively related to proprietary costs (Wang 2007). MERGER is an indicator that equals one if the firm completed an acquisition during the fiscal year (Compustat sales footnote code AA ). DEBT_ISS is the ratio of new debt issued to total assets. Lang and Lundholm (2000) find that disclosure is higher when firms issue securities. VOLE is equity volatility, which captures the precision of information available to management. It is measured as the standard deviation of monthly logged stock returns during the fiscal year. Finally, LITIG measures litigation risk. It is an indicator variable that equals one if the firm operates in a litigious industry (following Francis et al. 1994). Skinner (1994) and Field et al. (2005) find that managers may provide earnings forecasts to avoid costly litigation. 3 We use a similar model to assess changes in annual revenue disclosure, except that all variables but LITIG are measured in annual changes. ΔREV_DISCL t = b0 + b1 ΔSIZE t + b2 ΔSALES t + b3 ΔSALES t x DECR t + b4 DECR t (2) + b5 ΔROA t + b6 ΔLOSS t + b7 ΔBM t + b8 ΔLEV t + b9 ΔR&D t + b10 ΔMERGER t 3 In untabulated analyses, we include earnings quality as an additional control variable (Francis et al. 2008) and none of our inferences relating to revenue disclosure are qualitatively changed. We don t include earnings quality in our tabulated analyses for two reasons: (1) the five-year measurement period reduces the sample size and introduces survivorship bias and (2) VOLE has a strong negative correlation with earnings quality and captures a similar notion of information uncertainty. 15

19 + b11 ΔDEBT_ISS t + b12 ΔVOLE t + b13 LITIG t + e t After examining the determinants of revenue disclosure, we replicate two prior findings on the consequences of disclosure: the association with analysts earnings forecast accuracy and the association with cost of equity capital. Lang and Lundholm (1996) and Hope (2003) find that disclosure is associated with more accurate earnings forecasts. We measure earnings forecast accuracy using EPS forecasts for the subsequent fiscal year that are issued during the seven days following the current year s 10-K filing date. Using this short timeframe helps us to isolate the effect of information provided by 10-K disclosures and avoid confounding subsequently revealed information. The absolute earnings forecast error, EFE, is the absolute difference between the mean of EPS forecasts issued following the 10-K filing and actual earnings per share, scaled by the company s stock price. Control variables are as included in Equation (1) above, which make the incremental effect of revenue disclosure on forecast accuracy, b1, independent of those firm characteristics that are associated with disclosure. 4 EFE t+1 = b0 + b1 REV_DISCL t + b2 SIZE t + b3 ΔSALES t + b4 ΔSALES t x DECR t (3) + b5 DECR t + b6 ROA t + b7 LOSS t + b8 BM t + b9 LEV t + b10 R&D t + b11 MERGER t + b12 DEBT_ISS t + b13 VOLE t + b14 LITIG t + e t In a subsequent estimation of Equation (3), and Equations (4) through (6) that follow, we include three additional measures of disclosure to assess whether REV_DISCL has incremental explanatory power over commonly used disclosure measures. READABLE is financial statement readability, or -1 x the Fog index (Li 2008). LENGTH is the natural log of the number of words 4 Healy and Palepu (2001) note that endogeneity is an important limitation of findings related to disclosure. For example, firms with high disclosure ratings tend to also have high contemporaneous earnings performance (Lang and Lundholm 1993). This may be caused by a self-selection bias firms may increase disclosure when they are performing well. As a result the association between capital market variables and disclosure may be driven by firm performance rather than disclosure per se. Our tests control for performance and other factors that influence disclosure. 16

20 in the 10-K filing, and MF is an index of management earnings guidance. MF equals 2 if the firm releases a point estimate earnings forecast during the year, 1 if the firm releases a range estimate, and 0 if the firm releases only qualitative guidance or no guidance at all. As discussed in Section 2.2, the prior empirical evidence, while not conclusive, tends to support a negative association between disclosure and cost of capital (e.g., Botosan 1997, Botosan and Plumlee 2000). We measure the implied cost of equity capital, ECC, following Claus and Thomas (2001), Gebhardt et al. (2001), Gode and Mohanram (2003), and Easton (2004). Each study s estimate is based on the residual income model, after specifying a relation between equity cost of capital, equity market value, equity book value, and forecasted earnings and dividends. We use the assumptions in Dhaliwal et al. (2005) and Barth et al. (2008). Following Dhaliwal et al. (2005), Hail and Leuz (2006), and Barth et al. (2008), ECC is the mean of these four cost of equity estimates. To mitigate the effects of error in estimating ECC, we eliminate observations for which ECC is less than 0 percent or greater than 50 percent. ECC t+1 = b0 + b1 REV_DISCL t + b2 SIZE t + b3 ΔSALES t + b4 ΔSALES t x DECR t (4) + b5 DECR t + b6 ROA t + b7 LOSS t + b8 BM t + b9 LEV t + b10 R&D t + b11 MERGER t + b12 DEBT_ISS t + b13 VOLE t + b14 LITIG t + e t Again, because we control for determinants of disclosure, the coefficient on REV_DISCL represents the incremental effect of disclosure on the cost of equity capital. We also include in subsequent analysis two additional measures of revenue disclosure: REV_DISCL_HI and REV_DISCL_HI3. REV_DISCL_HI is an indicator variable that equals one if disclosure is at or above the sample median. REV_DISCL_HI3 is an indicator variable that equals one if disclosure has been at or above the sample median in each of the three prior years. We use these two 17

21 indicator variables to differentiate the effect of one-time disclosures from continually high disclosure levels (or a commitment to high disclosure). Next, we assess the revenue-specific information provided by the presence of revenuerelated disclosures. Our first test addresses whether analysts are more likely to issue revenue forecasts for firms with greater revenue disclosures. I_RF is an indicator variable that equals one if at least one analyst issues a revenue forecast for the subsequent year during the seven days following the current year s 10-K filing date. To isolate the specific incremental information provided about revenues, we estimate this logistic regression on the subset of firm-year observations where an earnings forecast is issued. That is, each firm in the subsample has an earnings forecast, and we assess the likelihood that a revenue forecast is also issued in connection with the earnings forecast. I_RF t+1 = b0 + b1 REV_DISCL t + b2 SIZE t + b3 ΔSALES t + b4 ΔSALES t x DECR t (5) + b5 DECR t + b6 ROA t + b7 LOSS t + b8 BM t + b9 LEV t + b10 R&D t + b11 MERGER t + b12 DEBT_ISS t + b13 VOLE t + b14 LITIG t + e t Our second test addresses whether revenue disclosure aids analysts in issuing these revenue forecasts by decreasing their forecast error. We measure revenue forecast accuracy, RFE, as the absolute difference between the mean of revenue forecasts issued in the seven days following the 10-K filing and actual sales revenue, scaled by actual sales revenue. We multiply the ratio by 100 so that it is expressed as a percentage forecast error. We add earnings forecast accuracy to the control variables to allow the assessment of revenue forecast accuracy independent of earnings forecast accuracy. Thus, we focus on forecasting accuracy as it specifically relates to revenues. RFE t+1 = b0 + b1 REV_DISCL t + b2 EFE + b3 SIZE t + b4 ΔSALES t (6) 18

22 + b5 ΔSALES t x DECR t + b6 DECR t + b7 ROA t + b8 LOSS t + b9 BM t + b10 LEV t + b11 R&D t + b12 MERGER t + b13 DEBT_ISS t + b14 VOLE t + b15 LITIG t + e t Finally, we examine the pricing of revenues and expenses conditional on the level of revenue disclosure. If revenue disclosure leads the market to place relatively more valuation emphasis on revenues, then we expect to see a higher revenue response coefficient when disclosure is high. The annual stock return, RETURN, is the monthly compounded stock return, beginning three months into the fiscal year and ending three months after fiscal year end. The three-month lag in computing the annual stock return helps ensure that the price response to 10-K disclosures is included in the return. Annual sales revenue, SALES, and expenses, EXP, are each deflated by total assets. Expenses are measured as the difference between annual revenue and income before extraordinary items. An indicator variable for high revenue disclosure, DISCL_HI, is measured in one of two different ways. First, it is set to one if disclosure is at or above the sample median of REV_DISCL. In a second analysis, DISCL_HI is set to one if abnormal disclosure (i.e., the residual from the estimation of Equation (1)) is greater than zero. 5 As in Equation (4) above, we also include in one analysis REV_DISCL_HI and REV_DISCL_HI3 to differentiate the effect of one-time disclosures from continually high disclosure levels. A significantly positive coefficient on the interaction of sales and the high disclosure indicator variable, b2, indicates that investors place more weight on valuing revenues when revenue disclosure is high. RETURN t = b0 + b1 SALES t + b2 SALES t x DISCL_HI t + b3 EXP t + b4 EXP t x DISCL_HI t (7) + b5 DISCL_HI t + e t 5 Prior research has documented cross-sectional differences in revenue response coefficients, for example growth opportunities (Ertimur et al. 2003) and R&D intensity (Kama 2009). In the second analysis, the disclosure measure we use is orthogonal to a set of control variables, including these two. 19

23 Equation (7) replaces RETURN in Equation (6) with earnings in the subsequent year divided by total assets. This equation allows us to assess whether a difference in persistence of revenues in earnings prediction exists for firms with high or low disclosure. For example, if firms with high revenue disclosure also tend to have more persistent revenues, then a larger revenue response coefficient in Equation (6) is warranted and investors are rationally responding to the signal of revenue persistence provided by revenue disclosures. Otherwise, a larger revenue response coefficient in Equation (6) could represent the market being unduly influenced by a focus on revenues in 10-K disclosures. ROA t+1 = b0 + b1 SALES t + b2 SALES t x DISCL_HI t + b3 EXP t + b4 EXP t x DISCL_HI t (8) + b5 DISCL_HI t + e t 3.2 Sample Our sample spans from 1997 to 2012, and excludes regulated industries (Financial, Insurance, and Utilities). We use accounting data from Compustat, stock market data from CRSP, analysts forecasts from I/B/E/S, and management guidance from Thomson Reuters. 6 We use 10-K filings on Edgar to calculate the revenue disclosure index, 10-K length, and 10-K readability. After requiring data for all variables used in the analysis except management guidance, cost of equity capital, and forecast errors, our sample contains 47,322 observations. 4. Results 4.1 Descriptive Statistics Table 1, Panel A, presents the level of and changes in the revenue disclosure index by disclosure topic. The first column shows the fraction of firm-year observations that discuss each of the 14 individual search terms. The most commonly discussed revenue items are growth in 6 Because our data from Thomson Reuters ends in 2010, MF is not available in the final two years of the sample. 20

24 sales prices (84% of observations), product revenues (83%), revenue forecasts (80%), and revenue from key customers (74%). The least commonly discussed items are organic revenue growth (9%), revenues under percentage-of-completion accounting (18%), and the effect of foreign exchange rates on revenue (26%). After summing the revenue terms, the average firm discloses 6.94 of the 14 items. Annual variation in disclosure exists across the component terms. The percentage of observations with changes in the presence of disclosure range from 4% (order backlog, organic revenue growth, and revenues under percentage-of-completion accounting) to 16% (revenue growth from acquisitions). The aggregate revenue disclosure index changes in 60% of firm-year observations. The relatively high frequency of annual changes supports the notion that the revenue disclosures are not merely boilerplate and copied over from the prior year s 10-K. Table 1, Panel B, presents annual changes in the revenue disclosure index level. For example, the median and modal revenue disclosure index is 7 out of 14 items. Of the 7,371 observations with a revenue disclosure index of 7 in year t-1, 2,935 (40%) also have a revenue disclosure index of 7 in the following year and 7,297 (99%) have a revenue disclosure index that changes by no more than 3 in the following year. The small frequency of extreme changes in the index suggests that noise in the measure is relatively small. Summary statistics for the variables used in subsequent analyses are presented in Table 2, Panel A. As discussed previously, the average firm discloses 6.94 out of 14 revenue items. The range of REV_DISCL from the first to the third quartile is 5 to 9 items. Financial statement readability is multiplied by -1 (i.e., it is the inverse of the Li (2008) Fog index) so that higher values represent more readable disclosure. MF, the management forecast index, has a slightly smaller sample size because we don t have access to data after

25 Panel B of Table 2 presents means and medians of selected variables split by whether or not the firm s disclosures are at or above the sample median. These univariate results show that firms with revenue disclosure at or above the median are more likely to have revenue forecasts (19% vs. 13% of observations, t-statistic = 20.01). They also have smaller earnings forecast errors (0.05 vs. 0.07, t-statistic = 5.19) and smaller revenue forecast errors (0.11 vs. 0.17, t- statistic = 8.52). In each case, the relatively larger effect on revenue forecast frequency and accuracy is consistent with the revenue focus of the disclosure index. Finally, firms with revenue disclosure at or above the sample median have a lower average cost of equity capital in the following year (9.32 vs , t-statistic = 16.25). We supplement these univariate tests with multivariate tests in Tables 3 and 4, but these initial results indicate that the effects of revenue disclosure are economically large, even in relatively blunt (i.e., above vs. below median disclosure) tests. Table 2, Panels C and D present means of select variables by year and by industry, respectively. Panel C reveals that revenue disclosure increases over time, from 5.71 in 1997 to 7.83 in In contrast, other disclosure measures do not exhibit such a strong trend. Financial statement readability, READABLE, increases slightly from in 1997 to by 2001 but then subsequently shows a gradual decrease. Management guidance, MF, also exhibits an increasing trend through 2001 (from 0.11 to 0.30) and a subsequent decline. The length of the 10-K, LENGTH, increases from approximately 25,000 words to 40,000 throughout the course of our sample period. However, we include specifications that control for disclosure to ensure that our results are not being driven by overall disclosure level (interestingly, total disclosure length often has the opposite effect on outcomes as revenue disclosure). Overall, the different trends among disclosure measures suggest that they are not perfect substitutes and each is affected by 22

26 different factors. A thorough analysis of which measure is best is beyond the scope of this study. Panel C also shows a steady increase in the frequency of revenue forecasts over time, consistent with Ertimur et al. (2011). This temporal trend underscores the importance of our inclusion of year fixed effects in our multivariate analyses. Likewise, the industry variation evident in Panel D supports our inclusion of industry fixed effects. 7 The Pearson correlations presented in Panel E reveal some inconsistencies among the disclosure measures. Revenue disclosure is significantly positively associated with 10-K length (0.28) and with management guidance (0.13), and the correlation between 10-K length and management guidance is also positive (0.07). However, readability is negatively associated with revenue disclosure (-0.01) and with 10-K length (-0.40), and it is only slightly positively associated with management guidance (0.01). As discussed in Loughran and McDonald (2014), these mixed correlations raise some questions about the validity of the readability measure. Panel E also shows that revenue disclosure is positively associated with the issuance of revenue forecasts (0.11) and negatively associated with earnings forecast errors (-0.06), absolute revenue forecast errors (-0.11), and the cost of equity capital (-0.12). However, we base our inferences on the multivariate tests that follow Results Table 3, Panel A, presents the determinants of the level of and change in revenue disclosure. Determinants of the level of revenue disclosure appear in the first column of results. The positive coefficient on the change in sales (0.72, t-statistic = 7.73) indicates that revenue growth is associated with higher levels of disclosure. That is, firms discuss revenues more when revenues are increasing. The negative coefficient on the interaction between ΔSALES and DECR 7 We use industry classifications as defined by Barth et al. (2005). 23

27 (-0.91, t-statistic = -6.83) and more specifically the negative total coefficient on ΔSALES when ΔSALES is less than zero (-0.19 = ) indicates that disclosure is higher when revenues are decreasing. In other words, revenue disclosure increases with the absolute change in sales meaning that firms devote more time to discussing revenues when they are changing, although the association between disclosure and sales increases is greater than the association between disclosure and sales decreases. Consistent with Lang and Lundholm (1993) disclosure is positively associated with firm size (SIZE coefficient = 0.30, t-statistic = 36.34) and profitability (ROA coefficient = 0.52, t- statistic = 6.65). Disclosure is higher for loss firms (LOSS coefficient = 0.10, t-statistic = 3.21), suggesting that firms tend to discuss revenues more when profits are low. Revenue disclosure is higher in years that firms close acquisitions (MERGER coefficient = 0.51, t-statistic = 19.86) and in years that firms issue debt (DEBT_ISS coefficient = 0.10, t-statistic = 2.10), and for firms in litigious industries (LITIG coefficient = 0.87, t-statistic = 9.10). The latter result is consistent with firms with high litigation risk disclosing more to limit exposure from potential litigation. Finally, disclosure is positively associated with equity volatility (VOLE coefficient = 1.21, t- statistic = 6.27) but not significantly associated with either leverage (LEV coefficient = 0.02, t- statistic = 0.26) or research and development expenditures (R&D coefficient = -0.11, t-statistic = -0.27). The second column of results relates to determinants of annual changes in revenue disclosure. Sales increases are associated with increases in disclosure, although the relation is not statistically significant (0.07, t-statistic = 1.47). Sales decreases, however, have a stronger positive relation with disclosure changes (incremental coefficient = 0.14, t-statistic = 2.06). The 24

28 positive total coefficient on sales decreases indicates that decreases in sales are associated with decreases in disclosure. ΔSIZE is positively associated with ΔREV_DISCL (0.51, t-statistic = 16.33), suggesting that growing firms increase disclosure. ΔROA is negatively associated with ΔREV_DISCL (-0.16, t-statistic = -3.83) and new loss firms decrease disclosure (0.03, t-statistic = 1.99), indicating that as profits decrease focus on revenues in disclosures increases. Firms also increase disclosure when leverage increases (0.16, t-statistic = 2.37), when R&D expenditures increase (0.35, t- statistic = 2.51), and in years of acquisitions (0.07, t-statistic = 5.46). ΔREV_DISCL is not significantly associated with ΔBM, ΔDEBT_ISS, ΔVOLE, or LITIG. Results on the association between revenue disclosure and earnings forecast accuracy are provided in Table 3, Panel B. REV_DISCL has a negative and statistically significant association with absolute earnings forecast errors (-0.00, t-statistic = -3.02), indicating that forecast accuracy is higher when disclosure is greater. Absolute earnings forecast error is also significantly negatively associated with SIZE (-0.01, t-statistic = -6.05) and ROA (-0.09, t-statistic = -3.26). It is significantly positively associated with LOSS (0.02, t-statistic = 1.91), BM (0.03, t-statistic = 3.72), LEV (0.08, t-statistic = 6.35), R&D (0.19, t-statistic = 4.12), and VOLE (0.39, t-statistic = 7.27). Absolute earnings forecast error is not significantly associated with ΔSALES, ΔSALES x DECR, DECR, MERGER, DEBT_ISS, or LITIG. The second set of results in Panel B reveals that the association between REV_DISCL and EFE is robust to controlling for three other measures of disclosure readability, 10-K length, and management guidance. Of these, each is negatively related to EFE but only readability is marginally statistically significant (-0.00, t-statistic = -1.66). We provide these results to ensure that our revenue disclosure proxy is not simply measuring general attributes of disclosure (in 25