SEC Litigation and Revenue Recognition Disclosures Study

Corporate Disclosure Response to SEC Litigation Pertaining to Revenue Recognition Benjamin Vaughan Susan B. Hughes Barbara Arel The University of Vermont 1

Accounting and Auditing Enforcement Release Hired New CFO Channel Stuffing Litigation Settlement 1999 2006 2009 2001 - 2002 2

Purpose of the study: To determine if firms revenue recognition disclosures change over the time period of suspected and detected revenue misstatement.

Why its important Understanding how disclosures change could lead to better detection of fraudulent revenue recognition practices by regulators, auditors and informed users. Pertinent to the FASB Disclosure Framework & the SEC s Plain English Initiative 4

Literature Review INVESTORS AVOID THE LITIGATED. Less trading in firms with AAERs, regardless of previous infringements. (Feroz et al 1991) LITIGATION HURTS BUSINES. Firms lose 38% of FV upon litigation release. (Karpoff et al 2008) 38% FV Loss Breakdown (Karpoff et al 2008) LITIGATION HURTS MANAGERS. 90% of culpable managers lose their jobs, own 9.7% of the firm s equity and face an average of $8 million in fines and the possibility of jail time. (Karpoff et al 2008) 9% Fines and/or Penalties 24% Reputational Damages 67% Market Adjustment

Literature Review (cont.) The Incomplete Revelation Hypothesis predicts users read less complex disclosures when analyzing financial statements. (Bloomfield 2002) MORE COMPLEX DISCLOSURES: Result in lower trading volume and consensus by small investors. (Miller 2010) Take longer to be fully represented in the market price. (You & Zhang 2008) More complex Revenue Recognition Policies decreases chance of AAER. (Peterson 2011)

Literature Review (cont.) CHANGES DURING AAER Emphasize growth strategies and product lines rather than info related to fraudulent activity. (Hoberg & Lewis 2013) ALTER DISCLOSURES PRE-LITIGATION More readable language, cautionary language, year- to-year disclosure changes (Nelson & Pritchard 2007) OBFUSCATED DISCLOSURES Lower earnings less readable annual reports Increases earning more readable disclosures (Li 2008)

Research Questions Revenue related Infringement and litigation will trigger changes in footnote significant accounting policy (SAP) RQ1: Disclosure Readability RQ2: Disclosure Redundancy between disclosures year-to-year

Sample Selected from the 518 AAER from 2009-2012 listed on the SEC website. 9

Accounting and Auditing Enforcement Release Hired New CFO Channel Stuffing Litigation Settlement A.P. I.P. 1999 2006 2009 2001 - 2002 10

Selections from Sample Profile Company Infringement AAER Settlement Data Range Infraction Arthrocare Corp 2005-2008 2011 2011 2004-2012 Channel Stuffing Recognized revenue on invalid contracts and contingent sales. Aspen Tech 1999-2002 2009 2007 1998-2008 Bally Tech. 2003-2005 2011 2008 2002-2009 Improperly recognized bill-and-hold sales. Recognized revenue on sales with full right of return. Improperly recognized consignment sales, and generated sham transactions. Improperly recognized revenue on sales with side agreements or full rights of return. Misclassified bulk sales as operating revenues. Arranged sales with concurrent buy back provisions to boost revenues. Improperly recognized on lease agreement, improper use of Bill-and-hold. Basin Water 2006-2007 2011 2011 2006-2008 Cabletron 2001-2002 2011 2007 2000-2005 Cardinal Health 2000-2004 2009 2007 1999-2008 Delphi 2000-2004 2010 2006 1999-2007 Diebold 2002-2007 2010 2010 2001-2011 11

Methods Obtained 10-Ks from year before infringement year after settlement. Extracted SAP disclosures pertaining to revenue recognition. Relied on Microsoft Office Readability Suite Captured Word Count and Flesch Reading Ease Index (FRE) scores Compared disclosures year over year Analyzed word count and FRE data with paired t-tests. 12

SAP Word Count Distribution Column A B C Period Pre-Infringement Infringement Post-Infringement n 29 30 30 Mean 261 330 382 Range Std. Dev. 19 - 756 35 - 1176 35 - 959 201 280 251 13

SAP Word Count Test Results Paired t-Tests using the logged data (t, 2-tailed p value) A-B -2.317, p=.028 B-C -2.602, p=.014 A-C -4.082, p=.000 Announcement period results NOT significant 14

SAP FRE Scores Distribution Column A B C Period Pre-Infringement Infringement Post-Infringement n 29 30 30 Mean 18.96 18.20 19.86 Range 2.0-48.0 3.1-36.1 8.0-38.6 Std. Dev. 9.7 8.6 8.2 15

SAP FRE Scores Results Column Period Paired t-Tests using the logged data (t, 2-tailed p value) B-A C-B C-A A B C Pre-Infringement Infringement Post-Infringement -.375, p = .710 1.591, p = .122 .818, p = .420 Wilcoxon Signed Ranks Test Z, Asymp. Sig. (2-tailed) B-A C-B C-A -.973, sig. = .330 1.875, sig. = .061 -.706, sig. = .480 16

Results: Word Counts & FRE Scores Significant increases in the word counts throughout the infringement period. Significant change in the readability of disclosures after the year of infringement. 17

Number of Sentences: LN Infringement Column A B C Period Pre-Infringement Infringement Post-Infringement n 29 29 30 30 Mean 2.09 2.40 2.29 2.41 Range 1 34 2 - 53 2- 37 Std. Dev. .867 .858 .871 .847 18

Number of Sentence Results Paired t-Tests (t, 2-tailed p value) B-A 2.580, p = .015 C-B 2.288, p = .030 C-A 3.226, p = .003 Wilcoxon Signed Ranks Test Z, Asymp. Sig. (2-tailed) B-A 2.549, sig. = .011 C-B 2.213, sig.=.027 C-A 2.902, sig. = .004 19

Redundancy Methods Using Word Compare identified and tallied the number of sentences deleted, added, changed or retained SAP SAP Determined redundancy percentages = (New, Modified or Carried Yr. 1) / (Total Sentences Yr. 2) (Deleted Yr. 2) / (Total Sentences Yr. 1). Redundancy data was analyzed with the Wilcoxon Signed Ranks Test. Pre-infringement Infringement, Infringement Post-infringement. 20

M D D C C D 21

Number of Deleted Sentences Pre- Infringement to Infringement Announcement to Post- Announcement Infringement to Post-Infringement Pre-announcement to Announcement Period n 29 30 24 24 Mean % 12.51 24.10 12.57 11.26 Range % 0.00 - 57.14 0.00 - 100.00 0.00 - 47.83 0.00 - 88.57 Std. Dev. 17.92 29.27 17.34 22.17 Wilcoxon Signed Ranks Test (post to pre) 0.032 Asymptotic Sig. 0.586 No significant change in the number of new or carry-forward sentences. 22

Results Total words increase significantly throughout the infringement period: p = 0.000 Disclosures became significantly more readable from the year of infringement year after: Sig. = 0.061 Firms added new sentences throughout the infringement period: Sig. = 0.001 No significant changes in word count or readability in the announcement period. 23

Conclusion Firms undertake significant disclosure creation around the period of infringement, but not around the period of litigation settlement announcement. Further investigation around the period of infringement may yield more disclosure trends. Increasing sample size and adding control variables may clarify the results around announcement period. 24

Further Work Multivariate Tests Panel data including firm size, profitability, leverage, operating cash flows, changes in auditor, CEO/CFO and GAAP. Extended Redundancy Disclosure Analysis to time between the alleged infraction and the settlement of litigation. Disclosure Coding examine disclosures for % of boilerplate-type disclosures, method disclosures, tonal variation, other changes. 25

THANKYOU 26