Accounting Quality Model: So What Exactly Is the SEC Looking For?
Four years into the XBRL phase-in, the SEC has now collected one year of XBRL-detailed tagging data for all public companies and has a foundation for streamlining and improving its processes for examining and analyzing corporate disclosures.
In his keynote address at the XBRL International Conference in Dublin, Ireland, on April 16, 2013, Craig Lewis, Chief Economist and Director of the SEC’s Division of Risk Strategy and Financial Innovation, spoke about the latest innovations in analysis tools, specifically, the Accounting Quality Model (AQM) that his team is building. The AQM uses SEC EDGAR XBRL filings from more than 8,000 registrants to detect accounting fraud. The AQM also allows the SEC to review each reporting company at least once every three years and more frequently for a significant number of companies.
So what does all of this mean? In a nutshell, the SEC is building an “electronic exam” that speeds up the process for identifying financial statements that may contain fraud. When measured against industry peers, the risk score allotted can help the SEC develop a pipeline of reports to be sent to the Corporate Finance department for additional analysis. So rather than waiting for returns to be viewed within a three-year exam cycle, the SEC can begin an electronic exam of a filer’s financial data within 24 hours of a document having been received by the SEC!
One challenge facing the SEC– and any other user of XBRL data– is the quality of the XBRL filings themselves. If the SEC is using XBRL documents to help detect accounting fraud, companies must deliver quality filings to the government in the first place. Because XBRL has become an integral part of the financial reporting process, the XBRL component should be taken as seriously as the official HTML document.
RR Donnelley is urging all companies to have a formal process for assessing the completeness, mapping, consistency and structure of all XBRL submissions. For many companies, improving XBRL submissions hinges on:
• minimizing the use of extended custom elements
• reviewing that appropriate signage is applied to all elements (for instance, negative values shouldn’t be reported for concepts that are expected to have positive values)
• ensuring that missing or incorrect calculation linkbases are not used