Bob Schneider is a Partner in 
Wilson So is the Director of Hitachi America, Ltd. XBRL Can Reduce the Fraud “Expectations Gap”
Written by Bob Schneider Posted February 2, 2007
My favorite line of comedian Fran Leibowitz is “You know you’re an adult when the phone rings and you hope it isn’t for you.”
Here’s another telltale sign of adulthood: You realize the job you do isn’t anything like what people think it is or what you yourself thought it would be.
Auditing is a prime example. Most first-year accounting majors would be surprised to learn that their first job out of school would be to help assure that financial statements are presented fairly. To the extent they know they’ll have to work a while as auditors to get their CPAs, they probably figure they’ll be looking to see if anybody has been stealing money from their clients. In other words, they’ll be looking for fraud, which defined more precisely includes both the deliberate misrepresentation of financial statements as well as the improper use of company resources.
But for auditors, far from being their raison d’etre, fraud is a pain in the butt. Doctors don’t like disease, and exterminators don’t like vermin. But both recognize that they’d be out of a job if their respective nemeses were extinguished. Not so for auditors and fraud. With the exception of forensic auditing specialists, the external auditor would exist happily indeed, much more happily if fraud simply didn’t exist.
Uncovering fraud while performing all the tasks of a financial audit — including big-time frauds where management is on the swindle is really hard. At the same time, fraud creates enormous legal exposure for auditors. If even accounting majors think the independent auditor exists to find fraud, just imagine what the average juror believes. When the prosecutor asks “Where were the auditors?”, their response that “Uncovering fraud is not our primary objective” is greeted with much skepticism.
The heads of the big audit networks addressed these concerns in their recent essay on their vision for the audit profession:
“There are limits to what auditors can reasonably uncover, given the limits inherent in today’s audits. Specifically, unless companies or investors are willing to pay auditors to police all of a company’s transactions, auditors are limited to using indirect means to ascertain whether fraud has occurred.These methods clearly are useful, indeed essential, to preventing and discovering fraud. But they are not foolproof, nor can they be expected to be. Hence, [an] expectations gap arises because many investors, policy makers and the media believe that the auditor’s main function is to detect all fraud, and thus, where it materializes and auditors have failed to find it, the auditors are often presumed to be at fault. Given the inherent limitations of any outside party to discover the presence of fraud, the restrictions governing the methods auditors are allowed to use, and the cost constraints of the audit itself, this presumption is not aligned with the current auditing standards.”
How XBRL Can Help
There is no way to eliminate entirely that “expectations gap.” But I do think the widespread adoption of XBRL can go a long way to reducing it.
In his speech at the recent XBRL conference in Philadelphia, Ian Ball, CEO of the International Federation of Accountants, stated one of the direct benefits of XBRL for auditors: “Simple balance sheet statistical sampling efforts can be replaced by 100% validation routines or by deeper analysis of ledger-level data.” If realized, this could be enormously important to auditors in both discovering fraud and ensuring external audiences they have done everything possible to uncover it.
Statistical sampling is essential to an auditing process that emphasizes gaining reasonable assurance at reasonable cost. But by necessity, only a small portion of all transactions are selected and reviewed, and so the possibility of fraud still looms. From a litigation standpoint, convincing jurors that auditors did their job when fraud existed but relatively few transactions were tested is a tough sell. Even if 100% validation is not achieved, expanding sample sizes and lowering materiality levels increases the chances that fraud will be discovered in the financial audit. To the extent 100% validation is possible, it will be a boon to the auditing profession.
“Deeper analysis of ledger-level data” would also yield huge benefits. The search for fraud is a hunt for the exceptional, and not in the good sense. It’s finding some piece of evidence somewhere that something isn’t kosher. Numbers and ratios that look fine for the entire company for the full year often compiled from disparate accounting systems, with manual data re-entry and sometimes poorly controlled spreadsheets — can mask inconsistencies in the bowels of the organization for shorter periods of time. The introduction of XBRL as a single data standard for the firm enables the analysis of large data pools from deeply drilled levels at low cost, making the search for anomalies much more efficient.
The best solution to fraud, of course, is to make it hard to pull off. XBRL has substantial potential to improve internal control and thereby reduce fraud risk. The seamlessness of XBRL affords full-scale integration of company data with limited human intervention. Business rules both from the perspective of regulating the activities of employees as well as the constraints on the information entered into database systems can be more easily adopted and enforced. Because tagged data allows searches to be done much more quickly, control procedures that were formerly onerous can be instituted and performed without significant cost to the organization.
Finally, I think the benefits of XBRL-GL as a business reporting language as opposed to a mere financial reporting language (as discussed by David vun Kannon in his recent post on this blog) will be significant. Non-financial data — such as whether and when an employee has taken vacation — can be crucial in both detecting and preventing fraud. If this information can be easily expressed and manipulated with the financial data, analysis of employee activities will be enhanced, and discrepancies will be more visible. These advantages will be even more prominent if XBRL adoption becomes worldwide throughout industry, so that combined companies that formerly faced long lag times in merging data systems quickly integrate their data in the XBRL-GL standard.