New Robocop Tasked with Busting Fraudulent Financial Report Filers

Experts caution that the automated risk identification and ranking system could inadvertently subject innocent public companies that stand out from peer filers to lingering and costly investigations.

Al Holzinger

In the aftermath of the Enron Corp., WorldCom Inc., and similar turn-of-the-century frauds perpetrated by senior corporate executives, the U.S. Securities and Exchange Commission (SEC) predictably intensified its efforts to punish those filing false or misleading financial statements. In fact, prosecutions of this nature accounted for more than one-fourth of the agency’s caseload during 2003-2005. But then came the financial crisis and, soon thereafter and just as predictably, the SEC turned its attention to reigning in crisis-related malfeasance such as misleading investors about the risks of securities tied to subprime loans.

Consequently, during its 2012 fiscal year that ended last September 30, just 11 percent of SEC enforcement actions concerned purported financial misstatements. But now the pendulum appears to be swinging back with substantial momentum. In early July, SEC Chairman Mary Jo White — who garnered considerable media attention for her dogged pursuit of financial fraud cases during her 1993-2002 tenure as U.S. Attorney for the Southern District of New York — announced the formation of two agency task forces whose stated goals are “increasing prosecution of violations involving false or misleading financial statements and disclosures.” White also announced the launch of a Center for Risk and Quantitative Analytics, which she said will use so-called big data technology to find red flags of both fraudulent and inadvertent financial misstatements. However, she provided few details about how the Center would operate and gave no hints about how to minimize the risk of being singled out for investigation.

An extensive blog post by John Carney — formerly the top securities fraud official at the U.S. Department of Justice and currently a partner at BakerHostetler LLP — and one of his associates seeks to fill in those gaps. “Broadly speaking,” the authors say, the Center’s Accounting Quality Model (AQM) — which he notes has been labeled “not so affectionately” as RoboCop after the part human/part robot police officer in the 1987 science fiction film — “is an analytical tool which trawls corporate filings to flag high-risk activity for closer inspection by SEC enforcement teams.” Registrant companies, they caution, need to understand the logic programmed into the AQM “in order to decrease the likelihood that their firm will be the subject of an expensive SEC audit. “

Within 24 hours of the time a filing is posted to the SEC’s EDGAR system, it will be processed by the fully automated AQM and the results will be stored in a database. The AQM — which the SEC reportedly plans to update frequently to account for moves the agency finds registrants are taking to conceal misstatements — will use industry peer group comparisons to calculate and output a risk score that, in turn, will inform agency auditors of the likelihood the filing is flawed in some way. SEC enforcement staff, in turn, will use this score to prioritize their investigatory work. The agency says it also plans to use those risk scores as a means of corroborating, or invalidating, the approximately 30,000 tips, complaints, and referrals submissions it estimates will be received in future years through its Electronic Data Collection Systems or completed forms TCR.

The blog notes that, as would be expected, the SEC’s AQM builders are keeping their risk determinants “close to the chest.” However, the authors note that SEC Chief Economist Craig Lewis — the principle “builder of RoboCop” — has dropped hints about the types of information most likely to catch the model’s attention. “An accounting policy that could be considered a risk indicator ... would be an accounting policy that results in relatively high book earnings, even though firms simultaneously select alternative tax treatments that minimize taxable income,” Lewis has said, for example. “Another accounting policy risk indicator might be a high proportion of transactions structured as ‘off-balance sheet.’” The authors add that frequent company conflicts with their external auditors, changes in auditors, or filing delays also could be high-risk indicators. Examples of so-called risk inducers — indications of strong incentives for earnings management — could include decreasing market share or declining profit margins.

The authors say the SEC currently is working to develop the AQM’s capability to compare words in the Management Discussion & Analysis (MD&A) sections of registrants’ future annual reports with a list of words and phrasing choices commonly used in the past by fraudulent filers. The SEC’s Lewis recently provided a rationale for this initiative: “So what we’re doing is taking the MD&A section, we’re comparing them to other firms in the same industry group, and we’re finding that in the past, fraudsters have tended to talk a lot about things that really don’t matter much and they under-report all the risks that all the other firms that aren’t having these same issues talk quite a bit about.”

The authors counsel registrant companies by saying the best ways to minimize their risk scores are:

  • Checking filings for “sloppy” XBRL and other data, because the AQM is an automated system that looks for oddities but is unable to account for innocent mistakes.
  • Adopting accounting policies similar to those of industry peers and sticking with the same public accounting firm. “The new path to avoiding SEC investigations is blending in with the competition,” the authors assert.
  • Minimizing off-balance sheet transactions. “While there are plenty of legitimate reasons a company may have significant amounts of off-balance sheet transactions, there are not as many legitimate reasons for this figure to be significantly higher than those in an industry peer group,” the authors conjecture.
  • Making conservative decisions about discretionary accruals. “Any filer pushing the bounds of discretionary accruals should thoroughly explain their decisions in the filing, or they should expect to explain it to an SEC exam team shortly thereafter,” the blog says.

The authors note in closing that “perhaps the comparison of the SEC’s AQM program to Robocop is a bit too fanciful, but RoboCop’s mantra that ‘anything you say can and will be used against you’ still might be good cautionary advice for SEC filers.”

To comment on this article, email FSA Times Editor Shannon Steffee.