The Impact of Auditor Association on Client Multi-Venue Disclosure Transparency Free

Statement on Auditing Standards (SAS) No. 118 (AICPA 2010b), and its predecessor, SAS No. 8 (AICPA 1975), address the auditor's association with information (e.g., MD&A) in documents (e.g., 10-K reports) containing audited financial statements. Both statements assert that the auditor has no responsibility to corroborate the other information, instead requiring the auditor to carefully read the information and determine whether it contains material inconsistencies with the audited financial statements. Taken literally, the auditor's responsibility extends only to what is said in the MD&A and not to what is not said. This raises the question of whether auditors address the completeness of MD&A disclosures when applying the provisions of SAS Nos. 8 and 118. The issue is important because prior studies show that financial analysts tend to use MD&A content more so than the actual financial statements when making forecasts (Epstein and Palepu 1999; AICPA 2010a) and that users do not accurately recognize which portions of 10-K reports are audited (Bedard, Sutton, Arnold, and Phillips 2012). Adding to the misunderstanding, Libby, Nelson, and Hunton (2006) find that auditors tend to allow greater magnitudes of misstatements in different parts of annual reports.¹

In this paper, we focus on an accounting issue (LIFO liquidations) that can result in required disclosures both in the financial statements and in the MD&A. This issue is ideal for studying disclosure tendencies because its effects are overwhelmingly income-increasing. Combined with the lower level of auditor association with the MD&A, a LIFO liquidation presents an opportunity for management to essentially claim the resulting transitory increase in earnings as operating income (through disclosure omission in the MD&A). GAAP requires at least footnote disclosure of material income effects caused by LIFO liquidations. SEC standards mandate MD&A disclosure of items affecting the material comparability of income amounts between years. LIFO liquidations can materially disrupt this comparability by temporarily altering revenue-to-cost relationships.

These reporting requirements, and the related auditing standards, allow us to address whether the degree of auditor association with financial disclosures affects client companies' disclosure tendencies. Results reveal significantly lower disclosure tendencies for this issue in the unaudited MD&A than in the audited financial statement footnotes. We also find that these lower MD&A disclosure tendencies were not considered by auditors to represent material inconsistencies as defined in SAS Nos. 8 and 118. In light of the noted misperceptions by financial statement users about which parts of financial reports are audited (Bedard et al. 2012), our results provide evidence in support of the PCAOB's current consideration of revising auditing standards regarding the auditor's association with other information in documents containing audited financial statements (PCAOB 2013).

Item 303 of Regulation S-X (SEC 1982) includes these principle objectives for the MD&A:

• to provide a narrative explanation of a company's financial statements that enables investors to see the company through the eyes of management; and

• to provide information about the quality of, and potential variability of, a company's earnings and cash flows, so that investors can ascertain the likelihood that past performance is indicative of future performance (emphasis added).

Other relevant portions of the standards (SEC 2003a; SEC 2003b) state:

• If there is a reasonable likelihood that reported financial information is not indicative of a company's future financial condition or future operating performance … appropriate disclosure in MD&A … may be required. For example … if events and transactions reported in the financial statements reflect material unusual or non-recurring items … provide disclosure where necessary.

and

• If the registrant knows of events that will cause a material change in the relationship between costs and revenues … the change in the relationship shall be disclosed.

The AICPA Task Force on LIFO Inventory Problems (1984) concluded that companies electing a LIFO pricing approach must disclose in the notes to the financial statements any material differences between LIFO inventory costs reported on the balance sheet and what the cost of this inventory would have been using replacement cost (i.e., FIFO).² The dollar difference between these two methods is termed the LIFO Reserve. Assuming consistently rising prices and steady quantities of inventory, this reserve represents the systematic understatement of inventory values (relative to replacement cost) and an indefinite deferral of reported profits. Consequently, LIFO accounting saves income taxes in periods of inflation (Isidore 2009). As evidence, Tipton (2012) shows that LIFO reserves for all oil, gas, and coal companies totaled $372 billion at fiscal 2010 year-end.

Should a LIFO company suddenly decrease its inventory levels, older, presumably lower inventory costs are released into the income statement as cost of goods sold. Current sales prices are then matched with an arbitrarily low cost of sales, resulting in a one-time-higher profit margin that may not be reflective of a company's sustainable earnings. Therefore, a LIFO liquidation represents an unusual disruption in the relationship between costs and revenues, possibly making current earnings less than predictive of future operating performance. Thus, the previously quoted Item 303 of Regulation S-X mandates that management disclose these unusual or non-recurring events, when material, within the MD&A.

In addition, the SEC (2011) requires disclosure, either in the footnotes or parenthetically on the face of the financial statements when a LIFO company “includes a material amount of income in its income statement, which would not have been recorded had the inventory liquidation not taken place.”

A recent SEC complaint demonstrates that companies have used LIFO liquidations to manipulate financial statements and MD&A disclosures. In 2007, Nicor, Inc., a Chicago-based natural gas company, paid a $10 million fine to settle an SEC complaint for using LIFO liquidations to “rig” its reported earnings between 1999 and 2002 (SEC 2007).

SAS No. 8 (AICPA 1975), Other Information in Documents Containing Audited Financial Statements, and SAS No. 118 (AICPA 2010b), which superseded SAS No. 8 under the same title, consider MD&A information to be unaudited. These standards (AU 550 [AICPA 1975] and AU-C 720 [AICPA 2010b]) state that “the auditor's opinion on the financial statements does not cover other information (e.g., MD&A disclosures) and the auditor has no responsibility for determining whether such information is properly stated” (emphasis added). Instead, the standards direct that “the auditor should read the other information of which the auditor is aware in order to identify material inconsistencies, if any, with the audited financial statements” (AICPA 1975, 2010b). If the auditor notes material inconsistencies that are not corrected by the audit client, the auditor is directed to add an explanatory paragraph to an otherwise unmodified audit report to emphasize the disclosure inconsistencies. SAS Nos. 8 and 118 are essentially identical with respect to these requirements.

Regarding implementation, the standards direct the auditor to first read what is in the MD&A and then determine whether it is materially inconsistent with the financial statements. Taken literally, this would minimize or eliminate the need to consider whether the MD&A disclosures exclude important items in the financial statements that affect the company's operating results.

PCAOB standards, which in some cases are taken directly from AICPA standards (AT Section 701) discuss the requirements when a practitioner performs a separate attest engagement (either examination or review) on a client's MD&A. These expanded services are elective and not required for 10-K filings.

In summary, regulatory standards require disclosure of the income effects from a LIFO liquidation in the audited financial statements. Further, SEC standards independently require disclosure in the MD&A of items (e.g., material LIFO liquidations) affecting the comparability of earnings between years. Coupling these disclosure requirements and the different degrees of auditor association with footnote versus MD&A information, we address the following research questions:

• RQ1: Are companies' disclosure tendencies for LIFO liquidations affected by the different degrees of auditor association with the disclosed information?

• RQ2: If so, do auditors' reports recognize these potential disclosure differences as material inconsistencies between the MD&A and the financial statements?

While other potential accounting issues exist to address our research questions (e.g., material extraordinary items, changes in accounting principles), we chose to examine how companies disclosed the effects of LIFO liquidations. Because the effects of these transactions are almost universally income-increasing, they create a clearer temptation to claim credit (through disclosure omission) for the favorable transitory nature of a LIFO liquidation when discussing current operating results in the MD&A. The other possibilities add complexity because of the unpredictable positive or negative effects on earnings.

Using Compustat data, we compared the required LIFO reserve footnote information in pairs of adjacent years. We identified over 900 companies between the years 1993 and 2010 whose LIFO reserves had decreased (presumably creating a LIFO liquidation). We then computed the percentage effect on income by dividing the decrease in LIFO reserve by reported pretax income (since liquidation effects are reported pretax) for the year of the liquidation.

By examining 10-K reports using the SEC's EDGAR database, we determined whether each company had specifically disclosed the income effect of the LIFO liquidation (over and above simply disclosing the LIFO reserves) in its statements/footnotes and/or in its MD&A. As mentioned previously, if the liquidation causes a material transitory effect on income, separate disclosure is required by SEC standards for both the footnotes and the MD&A. The determinations were independently reviewed by two authors and any coding differences were reconciled. We also examined audit reports to determine whether explanatory language had been added to discuss the potential MD&A disclosure inadequacy. We used 5 percent of pretax income as a materiality estimate because prior literature (see Holstrum and Messier 1982; Messier, Martinov-Bennie, and Eilifsen 2005) recognizes that CPA and user perceptions of materiality generally range from 4 to 6 percent of pretax income.

Table 1 shows the disclosure frequencies by income-effect category and disclosure type. Results show a higher (Chi-square = 8.37; p < 0.01) percentage of companies that specifically disclosed the income effects of LIFO liquidations in the audited footnotes than in the unaudited MD&A disclosures. Addressing RQ1, only 27 percent (98 of 362) of companies whose income effects exceeded 5 percent of pretax income disclosed the impact of the LIFO liquidation in the MD&A. By contrast, 57 percent (208 of 362) of the companies disclosed income effects over 5 percent in the footnotes.

Table 1 does not address the potential overlap of these disclosures. That is, do companies that disclose in the footnotes also disclose in the MD&A? Table 2 further addresses RQ1 by showing companies that disclosed in one, neither, or both venues.

The largest clustering of companies (336 of 954, bold and italic upper left cell of Table 2) had presumably immaterial (≤ 5 percent) income effects and, understandably, disclosed the income effect in neither the footnotes nor the MD&A. Conversely, 83 companies (bold and italic in the fifth column) had income effects greater than 5 percent and properly disclosed the LIFO liquidation in both the footnotes and MD&A. Coupling these two groups, 44 percent of the companies (336 + 83 of 954 total) appear to have disclosed as the accounting standard and SEC rule dictate. In addition, 23 percent (215 of 954) of the companies (income effect ≤ 5 percent, shown in italics in the middle three cells of the first row of data) appear to have disclosed above and beyond accounting standards by disclosing, in at least one venue, income effects normally considered immaterial. Combining these results, 67 percent of the companies produced at least the appropriate level of disclosure or nondisclosure. In contrast, 33 percent (155 + 147 + 18 of 954, shown in bold in the > 5% Row of Columns 2, 3, and 4) of the companies did not disclose the income effects of presumably material LIFO liquidations in at least one of the two reporting venues.

Relevant to our first research question, a comparison of two of the bold cells in the > 5 percent Row of Table 2 shows a much greater tendency for companies to specifically disclose presumably material LIFO liquidation effects in the audited footnotes alone (147 observations), as opposed to the unaudited MD&A alone (18 observations).

Another potential confound affecting RQ1 is whether a company has satisfied its financial statement disclosure requirements concerning the liquidation income effects by simply disclosing comparative information about the LIFO reserves in the footnotes. This may explain the result in Table 2 that some companies disclosed in the MD&A, but not in the footnotes. However, as stated earlier, accounting standards clearly mandate specific disclosure of material LIFO liquidation income effects in the footnotes or on the face of the statements. Also, the fact that some companies did specifically disclose the liquidation income effects in the footnotes would argue that the disclosure requirement is sometimes literally followed.

To further address this, Table 3 partitions the Table 2 data to examine only the companies that did disclose the income effects of material liquidations in the footnotes, establishing that the companies recognized the need for disclosure. As seen in this table, 64 percent of companies with existing footnote disclosure of material liquidation income effects did not disclose in the MD&A, while 36 percent disclosed in both venues. This provides further evidence of lower-disclosure tendencies in the MD&A.

Regarding our second research question, we examined audit reports for all sample companies finding that none contained explanatory language to discuss MD&A disclosures. This is consistent with prior research (Eakin, Louwers, and Wheeler 2009; Wheeler, Cereola, and Louwers 2014), and indicates that omitting disclosure of the income effects of LIFO liquidations from the MD&A was not deemed by the auditors to be materially inconsistent with the accompanying financial statements. Our discussions with practicing auditors at two different firms (one international and one regional) indicated that much of the time spent on reading the MD&A consists of verifying that any numbers included in the MD&A can be traced successfully into the financial statements. One indicated that this task is sometimes delegated to lower-level staff. Neither indicated that much time was spent considering what might not have been included. Ultimately, when presented with the issue addressed in this paper, both agreed that not highlighting material LIFO liquidation income effects in the MD&A did represent a potential disclosure inconsistency, but that failure to detect the inconsistency was not in conflict with audit standards (AU 550) because of the direction-of-testing issue.

In this paper, we address whether disclosure transparency is affected by the location of information in financial reports and the degree of auditor association with this information. Using the presumed income effects of LIFO liquidations, we compared audited footnote-disclosure tendencies with unaudited 10-K MD&A disclosure tendencies. One might question the need for duplicate disclosures in the footnotes and the MD&A to an informed reader of the annual report. However, because research suggests that financial analysts rely more heavily on the MD&A than they do on the financial statements, the duplicity of disclosure would seem justified. Furthermore, GAAP and SEC standards independently dictate separate disclosure of these transitory income effects, when material, in the financial statements and in the MD&A.

After partitioning the results using a common materiality threshold, we note significantly lower separate-disclosure percentages within the MD&A than in the financial statement footnotes. Our results provide evidence that differences in the degree of auditor association may affect companies' disclosure transparencies. Because auditors audit the footnote information, but are only required to read the MD&A for material inconsistencies with the financial statements, perhaps companies perceive this as an opportunity to not disclose issues in the MD&A. Beyond speculation, we are unable to explain the MD&A nondisclosure practices of companies with seemingly material LIFO liquidations.

The noted disclosure differences ultimately may be attributable to the way SAS Nos. 8 and 118 frame their requirements. If auditors read the unaudited MD&A disclosures to then determine whether what is said should be deemed materially inconsistent with the audited financial statements, they are understandably less likely to address what is not said in the MD&A.

This poses an interesting question. That is, after an audit client has disclosed the income effects of a LIFO liquidation in the footnotes, is the nondisclosure of the same issue in the MD&A a materially inconsistent client reporting practice? Auditing standards only address whether disclosures that exist in the MD&A are inconsistent with the financial statements. Ironically, a double nondisclosure would not be deemed inconsistent. And, even if a material MD&A disclosure omission is noted, the provisions of SAS Nos. 8 and 118 would not technically require (or allow) an audit report modification because the stated MD&A is consistent with the financial statements. This could explain why none of the audit reports for companies in this study contained explanatory language to emphasize the issue. One cannot infer a deficiency in performance from this finding because the auditors associated with the companies in this study did follow standards as written. However, the combination of the ambiguity of auditing standards with the minimal auditor association with the MD&A provides an opportunity for companies to manage users' perceptions of earnings (e.g., the SEC's settlement with Nicor). Effectively claiming the transitory LIFO liquidation income effects as operating income through disclosure omission is misleading, especially to users who primarily read the MD&A. Therefore, at a minimum, we believe a reconsideration of the word inconsistencies in SAS No. 118 is in order to address the disclosure completeness issue.

Certain factors potentially limit our findings. Reasons other than LIFO liquidations may explain decreases in LIFO reserves between years. Firm-specific price declines could also have caused these decreases, making the low-disclosure rates more understandable. However, for those companies that did specifically disclose the liquidation effects in the footnotes, our results confirm that the MD&A disclosure tendencies were much lower, providing evidence concerning our original research question.

Regarding future research, LIFO liquidation disclosure is only one example of a financial transaction potentially raising multiple disclosure issues. The same transparency issues might have arisen when reporting the cumulative effects of changes in accounting principles in prior years when the retroactive effects of these changes were reported through the income statement (they are now charged to retained earnings). Extraordinary items on the income statement may similarly cause transitory income statement effects for current operations and provide a parallel disclosure transparency question for the MD&A. Further, a change in entity would create comparability issues in potential MD&A discussions of multi-year comparisons of operating results.