Written by Bob Schneider Posted on March 1, 2008

Following the publication of its Draft Decision Memo (which I recently discussed), the Pozen Committee issued its interim Progress Report.

Although the FEI blog is correct in its assessment that the wording of the two reports in the XBRL area show only minor differences, I can’t disagree with the IR Web Reports description of the final recommendations as “watered down”. Notably, while the draft memo stated the SEC should mandate the filing of XBRL-tagged financial statements within a defined time frame, the progress report now substitutes the phrase over the long-term.

That certainly seems less definite.

Here are key aspects of the Committee’s recommendations:

(1) The SEC should phase-in XBRL-tagged financial statements.
(2) In the initial phase, the largest 500 domestic public reporting companies (based on market cap) will be required to furnish XBRL statements.
(3) In the second phase, beginning one year after the start of the first phase, domestic large accelerated filers (as defined by SEC rules) will also be included.
(4) Once the second phase has been implemented and certain preconditions have been met, the SEC should decide whether to move from furnishing to filing of XBRL statements (I’ll get to the distinction in a moment), as well as to extend the requirement to all other reporting companies.
(5) The Committee did not include a specific recommendation on the assurance function.

As CFO.com reports, Appendix A contains a letter of dissent from committee member Peter Wallison of the American Enterprise Institute. The letter criticizes the extended phase-in period, noting that:

Only after the second phase has begun in late 2009 or early 2010 will the SEC (in the Committee’s recommendation) begin to consider whether to require any companies to file (rather than furnish) their XBRL-tagged financial statements. Since the second phase companies will (in the Committee’s recommendation) be permitted to furnish rather than file their XBRL financial statements, that must mean they won’t be required to file their XBRL financial statements until after their 10-Ks are filed in March 2011. That means no company, large or small, will be required to file a 10-K with XBRL financial statements until March of 2012. That’s four years from now, and quite a generous phase-in, considering we are talking about only 2000 or so of the largest and most sophisticated companies in the U.S. When the remaining 13,000 reporting companies will be required to file XBRL financial statements under this mandatory phase-in is anybody’s guess.

Mr. Wallison helpfully explains the difference between furnished and filed, and how he believes that distinction influenced the Committee’s recommendation:

Under the Securities Exchange Act of 1934, companies are absolutely liable for false or misleading material filed with the SEC. However, in the case of material that is merely furnished to the SEC, liability only attaches if it can be shown that the material was intentionally false or misleading. As many readers know, companies face less legal exposure for furnished as opposed to filed information, unless the information was intentionally wrong or misleading. Accordingly, the Committee seems to have adopted the idea of furnishing rather than filing XBRL financial statements because of its concern about the possible cost of auditor assurance. It seems to have reasoned that, if XBRL financial statements were furnished rather than filed, the reduced liability would permit companies to dispense with auditor assurance entirely, and thus to avoid these potential costs.

Mr. Wallison goes on to discuss in detail why he believes the concern about assurance costs is misplaced and ultimately self-defeating. He thinks Mistakes are especially likely if the tagged financial statements are furnished rather than filed. In that case, companies will believe that they don’t have to be particularly careful with the mapping to the XBRL taxonomy, since there will be little likelihood of liability for mere negligence.

Furthermore, he believes Not only was there no need to require the furnishing of XBRL financial statements, allowing XBRL financial statements to be furnished rather than filed will severely impair the value of XBRL for investors and analysts and is an important source of what will be an enormous and unnecessary delay in the adoption of XBRL in the United States.

In essence, Mr. Wallison thinks the Committee wrongly chose furnished over filed statements because of a mistaken belief that assurance costs for the latter would be substantial. There can be little doubt that analysts seek assurance for XBRL statements. In a recent CFAI survey, some 80% of members wanted to see some kind of external audit or review performed on XBRL statements. It would seem reasonable to assume that analysts will be deterred from using (merely) furnished statements, because they lack the auditor’s imprimatur.

And yet

I’m not certain of Mr. Wallison’s contention that companies may be less than fastidious in producing XBRL-enabled statements if they are to be only furnished and not filed. If companies recognize that furnished data will be used to compare their performance against peers, won’t they have a vital interest in ensuring the numbers are accurate? Moreover, if the so-called bolt-on approach for creating XBRL statements is a relatively simple affair — as several VFP participants have indicated — why wouldn’t companies take the trouble to do the job right? Even if the statements are simply furnished, I don’t see companies being unserious about the exercise

Furthermore, while I agree that filed would be better than furnished, in practice I wonder how important the distinction will be to portfolio managers and analysts. As Ralf Frank recently pointed out on this blog, investors have long relied on less-than-perfect, incomplete numbers from data aggregators. Let’s assume that (a) companies prepare their furnished XBRL statements with reasonable care, (b) that analysts get the message that all XBRL statements are based on audited, GAAP data and vary little from the filed statements, and (c) the differences between furnished versus filed statements for key ratios like margins and capitalization are even smaller. If analysts have the opportunity to use easy-to-manipulate XBRL data that is much more complete and accurate than that offered by aggregators, wouldn’t they want to use it?

What concerns me more about the recommendations is that many industry analysts will have XBRL data for only some of the companies they follow. Suppose you’re a sector analyst where your universe comprises, say, five names in the largest 500, two in the top 1000, and two or three small-but-interesting players. Is it worthwhile for you to learn about and use XBRL data when it is available for only part of your coverage? It can be argued that, especially for portfolio managers, having XBRL data for all companies of a certain size is more important than having it for entire sectors. But getting security analysts to not merely acquiesce to XBRL data but to want it is crucial.

And that won’t happen unless they can actually use it.

Add to Del.icio.us | Digg this

This entry was posted on Saturday, March 1st, 2008 at 10:29 am and is filed under General, . You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.