How managerial discretion in SAB 74 disclosure timing mitigates uncertainty under ASC 606

With a constantly evolving economy and a desire for improved reporting, the Financial Accounting Standards Board (FASB) and International Accounting Standards Board (IASB) work together to develop and adopt new standards for financial reporting. New standards tend to materially impact reported financial statements for some firms, but not others, creating uncertainty about upcoming financial reports. Additionally, capital market participants must adjust their interpretation of similar numbers reported under different standards. To assist investors, the Securities and Exchange Commission (SEC) requires Staff Accounting Bulletin (SAB) 74 disclosures, which report the anticipated effect of upcoming changes to accounting standards. While SAB 74 disclosures are required, management has significant discretion over the content and timing of these disclosures. We study the determinants and effects of SAB 74 disclosure timeliness related to the implementation of Accounting Standard Codification (ASC) 606 in the United States. This study aims to enhance our understanding of managerial incentives to disclose and analysts’ ability to interpret accrual information when new accounting standards create uncertainty.

After several years of debate, the FASB finalized ASC 606 in May 2014. The new standard dramatically changed revenue recognition to have a focus on contracts with customers and eliminated industry-specific standards that previously existed. ASC 606 emerged from a joint project between the FASB and the IASB aimed at improving the consistency and transparency of revenue recognition practices. Under prior US Generally Accepted Accounting Principles (GAAP), industry-specific standards often produced inconsistent treatment of economically similar transactions, leading to a lack of comparability and creating opportunities for revenue management. ASC 606 introduced a single, principles-based framework centered on the transfer of control in contracts with customers, with the goal of standardizing revenue reporting across industries. This framework was designed to eliminate inconsistencies, enhance comparability and reduce managerial discretion over revenue timing, thereby improving financial reporting transparency and investor decision-making.

Companies had over three years to analyze its effects, disclose anticipated changes and implement the new revenue recognition standard. We analyze firms’ SAB 74 disclosures to measure how long before adoption they disclose the expected effects of ASC 606. SAB 74 is an unusual disclosure in that it is mandatory, but managers have significant discretion over the timing, allowing us to study the timing of disclosure while removing the choice to disclose.

When disclosing unfavorable earnings information, managers have incentives to disclose early to reduce litigation costs or to maximize stock-based compensation incentives (Skinner, 1997; Aboody & Kasznik, 2000), as well as incentives to disclose later due to career or reputational concerns (Kothari, Shu, & Wysocki, 2009). It is unclear whether these incentives apply to disclosures about future accounting changes; the effects of a standard change clearly do not have the same benefits and consequences as earnings news. To our knowledge, no researcher has studied the timeliness of disclosures of reporting changes that are largely outside managers’ control. Thus, whether and how managers’ incentives impact disclosure timing is an empirical question.

We use the time between a company’s first non-boilerplate SAB 74 disclosure and its ASC 606 implementation as a measure of timeliness. We use the realized effect of ASC 606 in the first period of implementation as a proxy for the manager’s expectations and examine whether managers delay this disclosure based on the new standard’s net income effect. Results indicate that when the new revenue recognition standard results in lower net income than under the previous standard, managers are more likely to disclose the expected effects later compared to firms that report higher net income under the new standard. This finding is consistent with managers delaying negative information, even when the information implies a non-economic change to income that is outside the manager’s control.

We next turn our attention to sell-side analysts. Analysts are sophisticated users of financial statements and project financial statement variables as part of their valuation process (Schipper, 1991; Jiao, Koning, Mertens, & Roosenboom, 2012). The estimates they produce are often used as measures of managerial performance and for valuation purposes by institutional investors. Analysts also tend to have a deep understanding of the firms they cover. Understanding the underlying economic processes in a firm should allow analysts to continue forecasting for a firm, even under new standards. However, this process is not costless. Greater time and higher-quality information from management should improve analysts’ ability to perform their responsibilities. We expect that analysts can understand the changes under ASC 606 and that earlier SAB 74 disclosure would allow more time to build the changes into their forecasting models.

We first examine analysts’ earnings forecasting accuracy in the first quarter following the adoption of ASC 606. We find an overall reduction in accuracy following the implementation of the new standard. The reduction in accuracy is significantly larger for firms that we identify as materially affected by ASC 606. Even within materially affected firms, analysts’ accuracy was worse for firms that were more strongly affected by the new standard.

We next examine the effect of SAB 74 timeliness on analyst earnings forecast dispersion and forecasting accuracy. The absolute effect of ASC 606 on net income is positively associated with analyst forecast dispersion, indicating that analysts are capable of recognizing that there will be an affect, but they do not agree on the magnitude. However, more timely disclosure of SAB 74 mitigates the association. These results indicate that analysts do not have an initial consensus on the effects of ASC 606 but are able to come closer to consensus with guidance from the firm and more time to analyze the information. Similarly, more timely SAB 74 disclosure mitigates the negative impact of material changes on revenue recognition on analysts’ forecasting accuracy.

Together, these results indicate that managers delay the disclosure of negative accounting standard change news and that timelier SAB 74 disclosure reduces uncertainty for sophisticated market participants. When managers disclose earlier, they can control market expectations and reduce disagreement around the effect. Earlier disclosure also increases overall analyst forecast accuracy under new accounting standards.

We make three contributions to the disclosure and financial reporting literature. First, we contribute to the voluntary disclosure timing literature by examining a setting in which the disclosed information is largely outside of management’s control and has no clear, direct impact on firm value. SAB 74 disclosures offer several important distinctions relative to commonly studied voluntary disclosures such as earnings guidance or conference calls. Litigation risk is unlikely to be a primary concern, as we are unaware of lawsuits based on SAB 74 disclosures. Career concerns are muted, as the effects stem from externally imposed accounting changes rather than management performance. The information disclosed pertains primarily to accrual recognition, which should not fundamentally affect firm value by reducing stock price incentives. Moreover, SAB 74 disclosures are triggered solely by new standards, abstracting away from persistent disclosure patterns (“stickiness”) observed elsewhere. Finally, the quarterly disclosure pattern enables a low-frequency measure of disclosure timeliness compared to traditional intra-period studies that focus on days or hours. In this unique setting, we find that managers delay disclosure when the anticipated effect of ASC 606 on earnings is negative.

Second, we contribute to the literature on analysts’ interpretation of accounting changes. Prior research documents that analysts’ forecasts improve following mandatory improvements in accounting comparability and transparency (e.g. Jiao et al., 2012; Tsang, Xiang, & Yu, 2023). Our findings complement this work by showing that, in contrast, the adoption of ASC 606 – despite improving revenue recognition standards – complicated analysts’ earnings forecasts for affected firms. In particular, delayed and less informative transition disclosures were associated with greater forecast errors and dispersion. Our results also complement the recent evidence by Hao and Pham (2023), highlighting the challenges analysts faced during ASC 606 adoption.

Third, we extend the literature on accounting change disclosures and accounting policy transparency. Prior studies (e.g. Hope, 2003; Paprocki & Stone, 2004) found that high-quality disclosures about accounting policies reduce analyst errors and information asymmetry. Others (e.g. Alexander, Ettredge, Stone, & Sun, 2011; Noland, Pasewark, & Strawser, 1998) show that firms sometimes use accounting change estimates opportunistically. We build on this line of work by demonstrating that even in a setting of quasi-mandatory disclosure, managers exercise discretion over the timing of information provision, but early disclosure materially reduces uncertainty faced by analysts. Together, these contributions deepen our understanding of how managers use discretion over transitional disclosure timing and its implications for financial market participants.

SAB 74 US public companies are required to disclose impending accounting changes in their quarterly and annual reports. Managers are required to explain the impact of new accounting standards on future financial statements. These disclosures should become more detailed and quantitative as the effective date of the accounting change nears, including comparisons of accounting policies, implementation status and potential financial effects. Research indicates, however, that managers and stakeholders may be uncertain about the impact of new regulations, providing an opportunity for the disclosures to be strategically timed or biased (Berger & Hann, 2003; Alexander et al., 2011; Menon, 2020). More broadly, research shows that managers have incentives to withhold bad news, particularly when seeking to support stock prices, although this tendency can be mitigated by higher litigation risk (Bao, Kim, Mian, Su, & Nancy, 2019). This general tendency toward strategic disclosure is likely to influence SAB 74 practices as well.

SAB 74 disclosures are distinctive in that their existence is mandatory, but the content and timing of information provided are at the company’s discretion. They have similarities with other disclosures like Management Discussion and Analysis and Risk Factor disclosures, which are often criticized for being boiler-plate and uninformative (Brown & Tucker, 2011; Beatty, Cheng, & Zhang, 2019). Despite this, the SEC has recently heightened scrutiny on the SAB 74 disclosures, particularly regarding new standards on revenue recognition, leases and credit losses, urging companies to disclose the effects of future accounting changes earlier (Orenstein, 2016).

Despite the regulatory emphasis on timely and detailed disclosures of impending accounting changes, research on the actual timing, usefulness and impact of these disclosures is limited. Previous studies suggest that firms often provide minimal information until closer to the adoption date, and the actual effect of recent SEC scrutiny on the usefulness of these disclosures remains unclear (Berger & Hann, 2003; Orenstein, 2016).

Enache, Griffin, and Moldovan (2025) studied the adoption of ASC 842, the new leasing standard. They found that most companies were slow to provide any information in SAB 74 disclosures, waiting until the final year, despite four years of warning time. They also find that disclosures become less readable as the adoption date approaches, delaying analysts’ forecasts. They conclude that the SAB 74 disclosures may favor more sophisticated investors, going against the SEC goals. Overall, although firms meet the SEC’s minimum SAB 74 requirements, further work is needed to examine how their timing and content choices influence the disclosures’ effectiveness and broader consequences.

This study uses the new revenue recognition standard, ASC 606, to study the effects of disclosure timeliness and the general effects of disclosure on analyst forecasting. The new standard superseded ASC 605 and standardized revenue reporting across industries by focusing on contracts with customers and performance obligations. ASC 606 also removed industry-specific guidance that existed under ASC 605 ^[1]. Importantly, revenue that was deferred under old standards might be accelerated and recognized earlier under the new standard, or vice versa. Similarly, related contract expenses that were initially amortized over the life of a contract might now be expensed in full at the start of the contract, or vice versa.

Overriding industry-specific guidance makes ASC 606 a more principles-based approach. Lee, Lee, and Sadka (2024) and Lee and Lee (2020) both find that ASC 606 adoption increased uncertainty around firm revenues, as reflected in higher analyst forecast errors (AFE), greater dispersion, more managerial discretion in earnings and a higher cost of debt. Greater uncertainty around the adoption of a principles-based standard gives managers leeway to obscure unfavorable financial information.

The SEC allows two adoption methods for a new standard to mitigate the effects of uncertainty. First, the full retrospective approach requires firms to retroactively adjust prior periods to report each period as if the new standard had been in effect, which allows for comparisons within a firm over time. The second approach, the modified retrospective approach, is more common and provides less total information. This approach requires firms to disclose the current period’s results under the new standard and the old standard, with no retrospective adjustment to prior periods. Under this approach, it is possible to measure the effect of the standard on net income by comparing the reported earnings to the comparative earnings under the old standard; however, it is more difficult for financial statement users to compare within the same firm over time, since previous periods are still only reported under the previous standards.

Hao and Pham (2023) found that ASC 606 generally improved revenue informativeness but decreased analyst forecast quality. They also find that when companies use the full retrospective method, revenue informativeness is more improved and analyst forecasts exhibit smaller errors. This study differs from Hao and Pham (2023) and Lee et al. (2024) by studying SAB 74 disclosures and information that was available to analysts prior to the first periods that were reported under ASC 606. We complement the current literature by providing a more complete picture of information available to analysts and the net effects of that information.

Managers weigh the perceived costs and benefits when deciding whether, when and what to disclose. Whether or not managers will strategically delay the expected impact of ASC 606 is unclear, ex ante. First, managers may not believe that this disclosure impacts firm value or is used to assess their performance, since the information disclosed is about accounting changes outside of their control with no underlying economic consequences and disclosure choice is multifactorial (Tang, 2005). Second, the implementation of a new accounting standard is not a trivial task for a company. Companies often use external consultants, such as an advisory accounting firm, to assist with the implementation (Cai, Chen, & Zhou, 2020). It often requires new or updated systems, different tracking of information differently and staff training. We expect the level of complexity of the new standard, specific to that firm, to be associated with the timeliness of disclosure. Thus, the relationship between the direction of news and the disclosure timing may be dominated by nonstrategic factors.

It is generally accepted in the literature that bad news is released later than good news, first shown by Kross (1981) and later explored by Bagnoli, Kross, and Watts (2002). Many studies examine the strategic disclosure of information for reasons such as proprietary costs (Verrecchia, 1983), litigation costs (Skinner, 1997; Donelson, McInnis, Mergenthaler, & Yu, 2012), career concerns (Kothari et al., 2009), compensation incentives related to the firm’s stock price (Aboody & Kasznik, 2000) and uncertainty about information endowment (Dye, 1985; Menon, 2020). Proprietary costs arise when disclosures aid competitors and significantly influence firms’ disclosure choices (Verrecchia, 1983; Ellis, Fee, & Thomas, 2012). Career concerns and investor perception also play a role in disclosure choices, as managers often delay bad earning news to maintain a favorable image (Kothari et al., 2009; Baginski, Campbell, Hinson, & Koo, 2017). This delay allows for the possibility of offsetting bad news with future good news or using earning’s management techniques (Graham, Harvey, & Rajgopal, 2005). Managers might delay announcing an accounting change that could lower net income, to avoid perceptions of previously inflated earnings, thus affecting firm value estimations.

Litigation risk is the primary condition that incentivizes early disclosure of bad news. Studies by Skinner (1997) and Baginski, Hassell, and Kimbrough (2002) illustrate that firms tend to disclose earlier when litigation risk is high and that this reduces the eventual costs of litigation. While these costs may not relate specifically to SAB 74 disclosures, firms clearly face regulatory costs based on the quality of SAB 74 disclosures. The SEC has made a point of encouraging timely and complete disclosure in SAB 74 releases and has issued several comment letters related to poor disclosure (Orenstein, 2016). Overall, we seek to understand the average effect of news on disclosure timing in this setting. Because litigation risk is unlikely to be a dominant factor, we expect managers to delay negative news, leading to our first hypothesis.

We next turn our attention to the information effects of ASC 606 implementation for materially affected firms. We examine analyst forecasts as analysts are sophisticated market participants who are likely aware of the new accounting standards and how they affect the company. In concurrent research, Chen, Kimbrough, Lee, and Lee (2023) provide evidence that management guidance becomes less frequent following the implementation of ASC 606. Management guidance informs analysts’ estimates, so forecasting quality may reduce overall. Jiao et al. (2012) showed that International Financial Reporting Standards (IFRS) improves analyst forecasting ability. The switch to IFRS and switch to ASC 606 both represent moving to more principal-based standards, so we may expect improved performance postimplementation. However, Barth, Gow, and Taylor (2012) provide evidence that street earnings forecasts often do not consider GAAP changes. The totality of existing evidence suggests that materially affected firms are likely to involve more uncertainty and complexity in the new estimates, meaning that analysts will have less accurate predictions for more affected firms.

Our final hypothesis examines the effects of more timely disclosure on market uncertainty. We expect that analysts are more likely to consider the effects of an accounting change when they have the information about the expected change, specific to the firm they are forecasting, in multiple consecutive quarterly filings (i.e. timelier disclosure), since it could be easily overlooked due to information overload. Analysts tend to view early disclosure favorably (Son & Crabtree, 2011). Analysts can research GAAP changes and financial statement effects, but it often requires insider knowledge of the firm’s systems and contracts to assess the idiosyncratic effects. If managers’ pre-implementation disclosure helps analysts forecast earnings more accurately in the period of an accounting change, we expect more timely information to be associated with more accurate and more consistent earnings’ forecasts by analysts.

Our main interests regard the timeliness of SAB 74 disclosures. Therefore, we constructed a sample of US firms that were all affected by ASC 606 but exercised discretion in the timing of their SAB 74 disclosures. We start with the intersection of Compustat, The Center for Research in Security Prices (CRSP) and I/B/E/S, where we can match the first quarter using ASC 606 to the quarterly filing in the SEC’s Electronic Data Gathering, Analysis, and Retrieval (EDGAR) database. We have two main restrictions on the sample: 1) we restrict our focus to firms that have been materially affected by the new revenue recognition standard, and 2) the firms should adopt ASC 606 using the modified retrospective approach so that we can calculate the effect of the standard on earnings in the period of implementation (see Appendix B for an example of this disclosure and the hand-collection procedures for determining the effect of the standard). Using textual analysis, we remove firms who adopt ASC 606 using the full retrospective approach ^[2], since we will not be able to calculate the effect of the new standard in the period of adoption for these firms. By excluding early adopters, we have reduced the potential for selection bias between these early adopters and other firms, as Amir and Ziv (1997a, 1997b) show that firms who have positive private information about accounting changes are more likely to adopt the standard early. Using this sample of 1,944 firms, we examine the financial statements for the first quarter of adoption and collect the impact of the standard on total revenues and net income.

Table 1 details the sample selection. We motivate our determination of materiality based on a common audit approach, using both a percentage of revenues and a common-sized net income approach. We first calculate the net income impact in two slightly different ways: 1) (NI_606 – NI_605)/Rev_606 and 2) (NI_606 – NI_605)/Rev_605, where NI_606 and NI_605 are net income for the adoption period under ASC 606 (the new standard) and ASC 605 (the old standard), and scalars are either total revenues under ASC 606 (Rev_606) or ASC 605 (Rev_605). We then calculated the revenue impact of ASC 606 as (Rev_606 – Rev_605)/Rev_605. Both the net income and the revenue impact measures represent common-size percentages. We restrict our sample to firms where the absolute value of the net income impact is greater than 1% of sales or the absolute value of the revenue impact is greater than 5%. We identify 168 firms that are materially affected by ASC 606. Requiring all control variables and three or more analysts to measure forecast dispersion reduces our primary sample to 131 observations.

While this is a relatively small sample size, it is not unprecedented in accounting research (see Zechman, 2010; Michels, 2017). In fact, Bamber, Christensen and Gaver (2000) argue that proper investigation of a phenomenon requires examining a setting in which it is likely to affect decision-making. However, this does affect the generalizability of our results to a setting where a new standard has less material impact on a firm’s reported financials.

Our first hypothesis regards the timeliness of SAB74 disclosure. We measure the timeliness of the disclosure based on how early the firm provides its information to the market on a quarterly basis. Based on an initial review of 50 firms’ SAB 74 disclosures around ASC 606, the earliest that we observe any firm discussing idiosyncratic information regarding the standard is six quarters before the adoption date (i = −6, third quarter 2016 for calendar-year end firms for whom the effective date of adoption is January 1, 2018). As such, quarter −6 is considered the baseline “boilerplate” disclosure where firms simply disclose that there is an upcoming accounting change, with no firm-specific information. We verify this assumption when hand-collecting the disclosures. In each period between the 2016 Q3 10-Q disclosure and the 2017 10-K disclosure, firms generally keep the initial boilerplate disclosure and optionally add more (or remove) information until “full disclosure” is reached in the last period before the new guidance is adopted (i = −1).

Some firms may initially increase the disclosure to include several possible expectations of the new standard’s effects and then reduce the disclosure to “full disclosure” in the final quarter before implementation. We created a measure to capture the amount of full disclosure that was available during the 6 quarters preceding implementation. Specifically, DisclosureTimeliness is measured as

We make two adjustments to control for firms, including more information than what is ultimately disclosed. If the numerator inside the summation is greater than one, we make an adjustment to capture the distance from full disclosure as follows:

For example, if the numerator is 1.3 (including 30% more words than what is in the final quarter before implementation), an adjustment was made to make the numerator 0.7, which is equidistant from 1. This construct is similar to other intra-period timeliness measures that use stock price changes instead of textual disclosure changes (see Blankespoor, deHaan, & Zhu, 2018; Butler, Kraft, & Weiss, 2007; Twedt, 2016). Finally, we truncate the values of DisclosureTimeliness to have a minimum of zero ^[3]. A higher measure of DisclosureTimeliness is consistent with the firm arriving at full disclosure in earlier quarters and the information being disclosed in multiple consecutive quarterly reports. A smaller DisclosureTimeliness measure is consistent with a firm arriving at full disclosure later, possibly only in a single filing before implementation.

We also present an alternative measure, DisclosureJumpQuarter, which identifies the quarter (−5 through −1) with the largest increase in disclosure length. DisclosureJumpQuarter takes values from 1 to 5 with a higher value indicating that the largest jump was earlier. Appendix C includes a more thorough description and an example of calculations of the timeliness variables.

We examine analyst behavior following ASC606 implementation for hypotheses 2 and 3. We follow established literature by measuring accuracy using AbsAFE, which is the median ^[4] analyst forecast minus the actual earnings according to I/B/E/S, scaled by the stock price at the most recent quarter end (Kirk, Reppenhagen, & Tucker, 2014; Hutton & Stocken, 2021). AnalystDispersion is the standard deviation of analyst forecasts made within 90 days of the earning announcement for the first quarter of ASC 606 adoption and measures the consistency with which analysts interpret the information from the firm. Hypothesis 2 asks whether a material impact affects analysts’ accuracy. As such, we tested the treated firms against the available intersection of Compustat, CRSP and I/B/E/S. The sample includes five quarters: Q1 2017 to Q1 2018, where Post = 1 for Q1 2018 (the quarter of ASC 606 adoption) and 0 otherwise. This resulted in 9,611 observations for the main test.

In our tests for H1, we use a binary variable, Negative606Impact, which is equal to 1 if ASC 606 decreased net income. We combine this with ASC606ImpactMagnitude to examine how managers disclose negative news and how this disclosure changes with the existing degree of uncertainty. In an alternative measure, we rank the net income effect of ASC 606 from most negative to most positive and assign each to a ranked quintile (Quintile606Impact). We test our second hypothesis with difference-in-differences design and use standard treat and post variables, with treatment being a firm that is materially affected by ASC 606. Tests of Hypothesis 3 use DisclosureTimeliness as an explanatory variable along with an interaction with ASC606ImpactMagnitude, to examine the effect of disclosure timeliness on analyst forecast characteristics.

Table 2, Panel A, presents the descriptive statistics of our sample. ASC606ImpactMagnitude is on average just under 2% of revenues, with the mean of 21% suggesting some large outliers justifying our use of a binary variable in tests. ASC 606 decreased net income compared with ASC 605 in the quarter of adoption for 47% of sample firms (Negative606Impact), indicating that the new standard had varying effects on firms.

We also included a robust set of control variables in line with prior research on disclosure and analysts’ forecasts. The firm’s complexity is likely to affect ASC 606 implementation and timeliness of related disclosures. First, we include the standard deviation of sales for the previous eight quarters before adoption ^[5], SalesVolatility, assuming that projecting the impact of the standard on sales will be more difficult if sales are more volatile. We next include the number of business segments (NumBusSegs) assuming that a higher number of business segments requires more separate assessments of the impact, which would take more time to complete. The net income ASC 606 effect magnitude (ASC606ImpactMagnitude) is also included, assuming that larger effects are more complex and take more time to assess. Lastly, this standard specifically affects contracts with customers that have certain provisions in them (Ernst & Young, 2019). We include indicator variables equal to one if the 10-K prior to implementation includes the following phrases: “multiple element arrangement”, “rebate” or “warranty” given that the existence of those items in customer contracts will require more planning on behalf of the firm. In our regressions, we included each of these “complexity” variables separately as controls. We also use principal component analysis (ComplexityPCA) to reduce the dimensionality of these controls in some specifications.

Managers have more than one mechanism to inform market participants about the effects of accounting changes. One mechanism would be to provide a forecast of earnings for the adoption period, which may include information on accounting changes to explain any increase or decrease from previous expectations. As such, we include a control variable, GuidedEPSQ12018, equal to one if the firm provided Earnings Per Share (EPS) guidance for the first period under the adoption of ASC 606 ^[6]. Lastly, while our hypotheses are on the relative timing of information in the disclosure, the shear amount of information or the existence of certain qualitative information could be the driving factor in our tests. Thus, we include ICompareAccountingPolicies, ImplementationStatusDisclosure and QualitativeImpactDisclosure ^[7]. Lastly, SAB74FinalWordCount controls for the number of words in the SAB 74 ASC 606 disclosure in the period prior to adoption (i = −1).

We follow Blankespoor et al. (2018) and control for a number of factors related to firm disclosure. We proxy the market’s demand for disclosure using analyst coverage from I/B/E/S and institutional investor ownership percentage from Thompson Reuters, on the premise that greater analysts and institutions may increase the demand for timely information. We control for the overall disclosure quality of the firm with an indicator if the firm has released guidance in the year prior to the implementation date (GuidancePastYear) with data from I/B/E/S and an indicator if the firm has had a material weakness or material misstatement in the previous three years (RestOrWeaknessPast3Years) with data from audit analytics. We include firm size (LogMarketCap) and firm age (LogFirmAge) to control for differences in resource availability to spend on the accounting function. We include growth potential (BookToMarket) to control for differences in disclosure styles for high- and low-growth firms, and Standard Industry Classification (SIC) two-digit industry fixed effects (FE) to control for disclosure differences and ASC 606-effect differences across industries.

We also include controls related to firm fundamentals, as they may affect the manager’s choice to disclose information or the complexity of implementing ASC 606 (Christensen, Davis, Rodriguez-Vasquez, & Wu, 2025). The company’s return on assets is included because profitability may affect the incentive to withhold bad news or accelerate good news disclosure. We include the company’s Tangibility and R&D Intensity (RD) because they are related to both the firm’s reporting complexity and other opportunities for earnings management. We include a measure of the firm’s market competition (Concentration) because competitive dynamics have many effects on disclosure. Finally, we include CEOChairDuality, BoardSize and Big4Auditor as controls for governance and the scrutiny surrounding the firm’s financial reporting. All variables are defined in Appendix A.

One caveat to this study is its generalizability, given that we focus on a subset of companies that have the largest income statement effects from ASC 606 and do not early adopt or use the full retrospective method. Table 3, Panel B, compares our sample to a larger sample of firms at the intersection of Compustat, I/B/E/S and Thompson Reuters. Past sales for our sample have less variability than the larger sample of firms. Firms in our sample also have higher institutional ownership and analyst following, suggesting that the firms in the sample are more visible than the average firm.

Table 4, Panel C, compares the Fama French 12 industry makeup of the sample compared to the Compustat population in the first fiscal quarter after December 15, 2017. ASC 606 had a larger effect on software companies and entertainment companies given that those firms are more likely to have multiple element arrangements with customers, which have different recognition procedures under the new standard (Ernst & Young, 2019). Our sample therefore contains a larger proportion of Business Equipment: Computers, Software and Electronic Equipment (Industry 6) and Other: Mines, Construction, Building, Transportation, Hotels, Business Services, and Entertainment (Industry 12). The sample also underrepresents financial firms; however, most accounting research papers do not include these firms due to the heavily regulated disclosure environment. Lastly, the sample slightly overweights Telephone and Telephone Transmission (Industry 7) and Wholesale, Retail and Some Services (Industry 9), and underrepresents the other industries. Thus, we control for two-digit SIC code fixed effects in our analyses to address the disclosure or accounting impact differences between industries.

We implement ordinary least squares (OLS) to test the associations between the dependent and independent variables of interest. OLS provides the best linear unbiased estimates of coefficients and is the standard for concurrent research in accounting with similar dependent variables (Wooldridge, 2010; Du, 2023; Chen, 2023). To test the first hypothesis, whether the timing of the SAB 74 disclosure is associated with the net income effect after including controls, we estimate the following model:

Our main test regresses either DisclosureTimeliness or DisclosureJumpQuarter on explanatory variables measuring the impact of ASC 606 on a given firm. We include the magnitude of the effect and an interaction with Negative606Impact (equal to one if ASC 606 net income is less than ASC 605 net income for the first period under the new standards) to allow the magnitude of the effect to vary across positively and negatively affected firms. We control for firm characteristics described in the previous section and include industry fixed effects.

If managers delay the provision of information when the effect of the new standard is negative on net income, then $β1$ will be negative, providing support for our first hypothesis. We expect $β2$ <0, consistent with the notion that firms more affected by the accounting standard face more difficulty in assessing the firm-specific impact and are slower to disclose. If managers provide less timely information for larger negative effects, we expect $β3$ <0 as well. Alternatively, managers may want to reduce extreme negative uncertainty, which would be supported by $β3$ >0. In separate regressions, we include Quintile606Impact as the independent variable of interest, expecting a positive coefficient, consistent with positively affected firms disclosing about the effects earlier.

The expected relationship between timeliness and the control variables for complexity is less straightforward. If managers proactively assess and disclose their information as it becomes known, they will disclose it earlier when the complexity of the accounting change is higher, suggesting a positive relationship. On the other hand, a more complex implementation could take more time to gather any information, and the amount of information known in the period before adoption could still be low based on the uncertainty and complexity of the change. This thought process predicts a negative relationship between ComplexityPCA and DisclosureTimeliness.

To reduce concerns of selection bias, we implement a propensity-score-matched (PSM) sample among the firms that were materially affected by ASC 606, matching the sample of negatively affected firms (the treatment group) to positively affected firms (the control group) on all control variables used in the analysis. We employ a matched sample of materially affected firms to test H1, which posits that negatively impacted firms respond differently than positively affected firms ^[8]. This reduces our sample size to only 84 observations but helps to mitigate concerns about non-random assignment to positive or negative impacts of ASC606 ^[9].

Our second hypothesis predicts that analysts will have trouble predicting earnings from materially affected firms. First, we examine analysts’ accuracy by looking at absolute errors in the first quarter following ASC 606 adoption. We estimate standard difference-in-differences regression using firms that we identify as materially affected as the treatment group.

In this specification, Treat is an indicator for firms that we identified as materially affected by ASC 606, and the control group is the universe of untreated firms with available data. We implemented OLS for both the entire available sample and used PSM to identify a more closely aligned sample as a control group. For the PSM-Difference in Differences (DID) tests in Table 9, we begin with the full sample of firms with the necessary control variables available. Firms with material ASC 606 effects form the treatment group, while firms without a material ASC 606 effect form the control group. We then estimate propensity scores using a logit model with firm size (log market value of equity), market-to-book ratio, analyst following, loss indicator and two-digit SIC fixed effects as predictors. One-to-one nearest-neighbor matching with replacement is performed, imposing a common support restriction to ensure comparability across groups. To assess matching quality, we evaluated covariate balance before and after matching using standardized mean differences and overall fit statistics. The diagnostics indicate that matching substantially improves balance: the mean bias across covariates decreases from 20.1% pre-matching to 3.3% post-matching, and the pseudo R² falls from 0.037 to 0.001. These results confirm that the matched treatment and control groups are well balanced, providing a reliable basis for the subsequent DID analysis.

We test our third hypothesis and examine whether timelier SAB 74 disclosures help analysts forecast earnings under the new standard with the following regression:

We limit the sample for this analysis to materially affected firms because we do not have data on disclosure timeliness for non-affected firms. AnalystDispersion is the standard deviation of analyst forecasts made within 90 days of the earnings announcement for the first quarter of ASC 606 adoption, and AbsAFE is the median analyst forecast minus the actual earnings according to I/B/E/S, scaled by the stock price at the most recent quarter end. Based on the results in Table 9, we expect β₂ > 0 because analysts are less capable of accurately predicting earnings for more materially affected companies. An estimate of $β3$ <0 would indicate that more timely disclosure mitigates the negative effect of major ASC 606 effects and provides support for H3.

We first provide some graphical, univariate evidence for our first hypothesis whether firms’ directional effect of ASC 606 impacts the timeliness of their disclosure about the estimated effects in the quarters leading up to the accounting change. Figure 1 graphs the percentage distance from full disclosure $(SAB74wordcounti−SAB74wordcount−6SAB74wordcount−1−SAB74wordcount−6)$ for periods i = −6 through −1 ^[10]. The graph presents the median firm’s DisclosureTimeliness separately for firms who are negatively affected (NI_606 < NI_605) and positively affected. The line for positively affected firms is above negatively affected firms each quarter, suggesting that the median positively affected firm is faster to reach full disclosure than the median negatively affected firm. Overall, Figure 1 provides some univariate evidence against the null for Hypothesis 1, suggesting that negatively affected firms are slower to disclose the effects of the new standard.

Table 5 presents t-tests of the means of DisclosureTimeliness, confirming that the difference is statistically significant at the 1% level. Interestingly, t-tests of the differences in the lengths of the quarter −1 disclosures (SAB74FinalWordCount) and the percentage of firms that guide earnings under the first period of adoption (GuidedEPSQ12018) are not significantly different across the effect of the standard, suggesting that the complexity of the change is not different across positively and negatively affected firms. However, AFE are significantly higher for negatively affected firms, indicating that analysts’ forecasts were more optimistic for these firms. This pattern suggests that analysts may have underestimated the magnitude of the negative impact of ASC 606 on earnings, consistent with analysts facing challenges in adjusting their expectations when new accounting standards adversely affect reported performance.

Tables 6 and 7 present the results for H1. Panel A uses DisclosureTimeliness as the dependent variable, while Panel B uses DisclosureJumpQuarter. Columns 2 and 4 in each panel replace individual complexity controls with ComplexityPCA to reduce noise.

In Panel A, the coefficient on Negative606Impact is negative and significant at the 10% level in column 1 and at the 5% level in column 2. This indicates that firms expecting a negative net income impact from ASC 606 tend to disclose the expected effects later (i.e. more slowly – than firms expecting a positive impact). Notably, ASC606ImpactMagnitude is not statistically significant in either panel, while $β3$ is significantly positive only in column 2 of Panel B. This provides limited evidence that among negatively affected firms, larger impacts are associated with earlier disclosure.

Columns 3 and 4 use Quintile606Impact as the main independent variable, and the coefficients are positive and significant across both panels. This supports the finding that firms are more adversely affected by ASC 606 delay or are slower to fully disclose transition effects. For example, the coefficient on Negative606Impact in column 1 of Panel A (−0.083) implies that, all else equal, negatively affected firms disclose about 20% more slowly than positively affected firms (relative to the mean DisclosureTimeliness of 0.407). Among the control variables, SalesVolatility is significantly negative, suggesting that firms with more volatile revenues delay disclosure. The coefficient of ComplexityPCA is positive but statistically insignificant, possibly reflecting limited power due to sample size.

Overall, these results reject the nullity of H1, indicating that disclosure timing differs based on whether ASC 606 has a positive or negative income effect. One possible interpretation is that small negative impacts may be deferred while managers pursue offsetting actions – such as operational improvements, real earnings management or accruals adjustments – to smooth earnings (Roychowdhury, 2006; Dechow & Dichev, 2002; Graham et al., 2005). In such cases, disclosure may be delayed to allow bundling of bad news with more favorable developments. In contrast, when the anticipated negative impact is large, managers may accelerate disclosure, either because mitigating the effect is infeasible or due to concerns about litigation risk or the demand to reduce uncertainty.

This interpretation aligns with the directionally positive interaction coefficients across models and with significant results in Panel B using DisclosureJumpQuarter. The pattern is consistent with Enache et al. (2025), who found that ASC 842 transition disclosures became more detailed and complex closer to adoption, likely in response to regulatory pressure for completeness. Similarly, firms facing larger ASC 606 impacts may feel compelled to disclose earlier and in more detail as adoption nears. We find consistent results in the PSM sample (see Table 8).

Table 9 shows the results of our estimations. Column 1 focuses on a PSM sample as the control group, while column 2 uses the full sample as the control group. The coefficient on Post is negative in column 1 and positive in column 2, providing mixed evidence about the overall effect of ASC 606 adoption on analysts’ forecast accuracy in the first quarter following adoption for unaffected firms ^[11]. The positive and significant estimate for β₃ in column 1 indicates that the difficulty in forecasting earnings is larger for firms that were materially affected by 606. In column 2, we replace Treat with ASC606ImpactMagnitude, a continuous treatment variable, and obtain consistent results, finding that analyst forecasts are less accurate when companies are more affected by ASC 606 ^[12]. Overall, these results are consistent with analysts being unable to fully determine the effects of the new accounting standard change and incorporate the expected effects into their earnings forecasts for materially affected firms ^[13].

Table 10 shows that both dispersion and absolute forecast errors are positively associated with ASC606ImpactMagnitude, consistent with our results in Table 9, showing that analysts have a more difficult time forecasting earnings when the accounting standard change effect is larger. However, the relationship is significantly mitigated when the firm discloses the expected effects in a more timely manner (negative coefficient on DisclosureTimeliness×ASC606ImpactMagnitude). In summary, Table 10 provides evidence that analysts set more consistent and more accurate expectations of earnings when managers reduce uncertainty early by providing more timely insight into the effect of ASC 606 changes.

While our main results provide evidence in support of all three hypotheses, it is possible that the estimation technique, variable measurements, or sample selection bias the results. Thus, we employed a battery of tests to strengthen the causal interpretation of our findings. First, we repeat our analyses on the PSM samples presented above. Second, in testing our third hypothesis, to address potential endogeneity in our timeliness measure (DisclosureTimeliness), we implemented two-stage least squares (2SLS) estimation. In the first stage, we regress DisclosureTimeliness on our instrument Negative606Impact – an indicator equal to one if ASC 606 decreases net income – with our full set of controls. Negative606Impact yields a highly significant coefficient and a first-stage F-statistic exceeding 15, confirming instrument relevance (Larcker & Rusticus, 2010). Its exogeneity follows from the exogenous adoption of ASC 606, which alters net income through a mandatory accounting change rather than managerial choice. In the second stage, we replace DisclosureTimeliness with its predicted values from the first stage and estimate the impact on AFE and dispersion. The instrumented timeliness measure remains negatively and significantly associated (p < 0.01) with both outcome variables, indicating that earlier SAB 74 disclosures improve forecast accuracy and reduce dispersion. The 2SLS results, reported in Table 11, align with our prior findings and further support H3.

In unbalanced tests, we apply different measures of DisclosureTimeliness with relaxed restrictions on negative values and the “distance” adjustment described above. Changing these assumptions does not alter any of the conclusions. In our primary tests, we employed OLS as the primary estimation technique. While it is the primary estimation technique in the accounting literature, it suffers from potential issues of multicollinearity and overfitting, and our small sample size may cause issues with convergence. Hence, we employ Least Absolute Shrinkage and Selection Operator (LASSO) regression and Bayesian regression as alternative estimation techniques with similar results (see Muth, Oravecz, & Gabry, 2018; Ranstam & Cook, 2018). Additionally, we implement a placebo test in which we randomize assignment of Negative606Impact and, as predicted, fail to find results ^[14].

This study provides evidence that managers are more likely to delay SAB 74 disclosures related to ASC 606 when they anticipate a negative earnings impact. Although such discretion occurs within the bounds of compliance, it appears to serve strategic purposes – consistent with bad news hoarding behavior documented and delayed disclosure under uncertainty in prior work (e.g. Bao et al., 2019; Kothari et al., 2009; Menon, 2020). This finding contributes to the literature on discretionary disclosure timing and the effects of ASC 606 (Doyle & Magilke, 2009; Hao & Pham, 2023).

We show that even in quasi-mandatory settings, managers strategically time adverse disclosures, and that earlier and more gradual disclosure is associated with significantly lower analyst forecast error and dispersion – particularly among firms most affected by ASC 606. These findings suggest that timely transition disclosures help reduce the informational uncertainty caused by new accounting standards. Our results are robust across multiple specifications, including alternative timeliness measures, instrumental variable and Bayesian estimation and adjustments for sample selection using propensity score matching.

This study makes three key contributions. First, we extend the literature disclosure timeliness by demonstrating that timing discretion – even when content is ostensibly required – has real consequences for market participants (Lee & Lee, 2020; Enache et al., 2025). Second, we add to the literature on analysts’ interpretation of accounting changes by providing evidence that ASC 606 adoption complicated analysts’ forecasts, despite improving revenue recognition overall (Jiao et al., 2012). Third, we contribute to the literature on disclosures surrounding mandatory changes in accounting standards (Hope, 2003; Paprocki & Stone, 2004). We provide evidence that managers exercise discretion in the timeliness of these mandatory disclosures and that the timeliness has economic consequences for market participants.

We acknowledge limitations to our study. Our focus on firms with material ASC 606 effects enhances precision but limits generalizability. Our timeliness measure relies on word counts rather than qualitative assessments, and we do not directly examine investor reactions. However, prior research suggests that obfuscated or delayed disclosures can affect market efficiency. These limitations offer several directions for future work. Researchers could examine how SAB 74 timing influences stock price volatility, bid-ask spreads or institutional trading; study disclosures related to other standards (e.g. ASC 842); or compare US and international transition regimes. Future studies could also use textual analysis tools to assess the tone, clarity or specificity of transition disclosures and their effects on investor behavior. Overall, our findings underscore that when firms disclose mandated information, it can be as important as what they disclose. Regulators and standard setters should consider not only the content of transition guidance but also how timing discretion may introduce new forms of information asymmetry – even in contexts intended to promote transparency.

Aaron gratefully acknowledges financial support provided by the Woody L. Hunt College of Business. Jesse gratefully acknowledges financial support provided by the Charles R. and Dorothy Carter Chair in Business Administration Fellowship and the Woody L. Hunt College of Business. We want to thank Nathan Marshall, Henry Laurion, Nicholas Jasa, Andrew Stephan, Tathagat Mukhopadhyay, Kevin Dow and workshop participants at the University of Colorado Boulder for their helpful comments and suggestions.

The supplementary material for this article can be found online.