"GAAP is designed to allow companies to add individual line items if they feel there is something unusual about their company," observes PR Newswire's Michelle Savage, who is vice-chair of the XBRL-US Adoption Committee. "XBRL is designed to mirror the flexibility of U.S. GAAP." Savage welcomes the momentum that the SEC and FASB are likely to provide, parrying the suggestion that their involvement will lead to new regulation. "Does that mean they will start dictating how taxonomies will be developed, [or] will try to standardize all of the tags?" she asks. "FASB hasn't done that with U.S. GAAP, why would they do it with XBRL?"

Yet even in its nascent state, XBRL is already sparking debate over the degree to which companies should be allowed to customize — or extend — the official taxonomies.

"XBRL documents filed with the SEC use company-specific XBRL taxonomies which do not allow for the automated processing and comparable analysis that has been promised to the marketplace by XBRL," complained Eric Linder in a recent post he sent CFO.com that had appeared in the Analyst's Accounting Observer blog.

Linder, a member of XBRL International, is president of SavaNet, which makes XBRL reader software. He says that even though the companies in the SEC's pilot start with the same base taxonomies, "they have the unlimited ability to add new items and re-do the calculation relationships of existing items." That, he says, renders the data useless for the kinds of analysis Cox has been promising.

Of course, as a maker of software that would compare XBRL financial data from different companies, Linder has a vested interest in a higher level of standardization, than, say Grant Thornton's Roberts. "The 'X' stands for extensibility," says Roberts. "We are not going to give people a standard chart of accounts."

Then again, as an accountant, Roberts too has a vested interest in how XBRL develops. He says it will be up to the accounting profession to determine whether a company's XBRL extensions legitimately represent an item unique to that company's finances or whether they were created simply to obscure the item. "Then it would be incumbent upon the auditor to say this does not 'fairly represent,' because it specifically uses XBRL to inject opacity into an otherwise highly transparent environment."

Neither Linder nor Roberts represent preparers or end-users of financial statements. As more companies begin using XBRL data, and as analysts seek compare them, the debate over what is and is not a legitimate extension is only likely to grow louder. Whether the SEC and FASB can — or should — stay out of that debate remains to be seen.

• « previous
• 1
• 2