As many readers know, Walter Hamscher is an XBRL pioneer and one of the smartest people in the field. He recently gave us a wide-ranging interview in which he discussed a host of XBRL topics, including the SEC’s proposed mandate, the pace of XBRL adoption in the US and abroad, assurance issues, and Inline XBRL. The third part of the interview, which begins with question (6), appears below. The first and second parts were published over the past two weeks; we’ll publish two more installments over the next two weeks. 

(6) There’s been much discussion about the use of extensions in XBRL. From one point of view, they allow companies to adopt the language to their particular circumstances; from another, they allow companies to indulge in creative accounting and make financial statements less comparable.

The Pozen Committee’s Progress Report stated that, because of the development of US GAAP taxonomies, there will be less need for customized extensions. Do you agree with that assessment? Do you think the extension issue poses the major problem for preparers and users of financial statements? Or will it now be only relevant for a limited number of companies?
 
Actually, I disagree with a premise of the question. When I hear that it’s going to make financial statements less comparable — less comparable than WHAT? Less comparable than the mess that you have today? No, you can’t make it any worse. So it may be the same, it may be better but it can’t be any worse.

That said, I do think that the Pozen Committee is absolutely right. We invested a huge amount of effort in the US GAAP taxonomy with a large vocabulary with that intent and architecturally designed it in such a way that the extensions would be not only less necessary but a lot more predictable, easier to do them the right way, if you will. Does the extension issue pose a problem for preparers? No. It’s actually the solution, not the problem.

There’s a very interesting phenomena I’ve observed which is if you talk to people who are in the business of supplying data or tools to investment analysts, they say extensions are a problem.

Yet I have not heard any sophisticated, professional analyst, who is actually doing financial modeling, who is actually making business decisions, who has ever indicated that that was actually a problem. On the contrary, they see it as actually a huge step forward from the mess that we have today.

So I think you have got to be cautious when you hear people who are not professional analysts speak on behalf of analysts and give you the hearsay this will make analysis more difficult. First of all, ask if that person is really an analyst or a software vendor supplying to an analyst — that’s a very different perspective. Then ask whether widely held stocks are the ones that will have the most extensions. If you actually look at the last five years of data — and like all computer people that’s kind of the way I look at things, I always want to see the data first — look at the distribution of unique line items (never mind tags, but just unique line items used in financial statements), it’s actually the smaller companies that have much more need for extensions than the large ones.

Here’s why. If I’m Van Camp and my business is frozen foods, my revenues are so aggregated that you can say, yeah, this is the food industry and these are foods — you don’t need to distinguish between the different types of fish. If I am a small salmon importer, however, I may very well need to distinguish revenues from West Coast salmon and East Coast salmon.

I think we may be seeing people commenting on XBRL and extensions without really every having done anything with it. They have never actually tried modeling companies, and they’ve never actually read, for example, the US Taxonomies Preparers Guide, which spells out what constitutes a good extension. How do you do it so as to minimize your own effort and maximize the usefulness of your statements? There’s a lot of good information in there, and I confess that it frustrates me that we have people commenting on things who haven’t taken the time to actually go and read how it’s done or tried it.

I find it very interesting that Jeff Diermeier, head of the CFA Institute completely understands why it’s going to make things easier than they are, whereas software vendors who are worried about “this is not normally the way that XML works” or “this is not a flat finite table of standardized line items” are just missing the point. They don’t understand how much of an improvement it is over the current text-based mess.

(7) On May 14, the SEC voted to propose a rule for the adoption of XBRL for financial reporting. Was the proposal more, less, or about as aggressive as you had expected? Was there anything in the proposal that surprised you? [NOTE: The interview with Walter was conducted before publication of the proposed rule on May 30.]

I guess it was more a sense of what was left somewhat open in the proposal. I wouldn’t say it was a surprise; it was more of a recognition that they are seeking to fully understand all the implications of alternative proposals. For example I think people are going to be very dissatisfied with text blocks as a method for notes, even if the phase-in schedule allows it only for the first year of a company’s filings. I don’t think that’s going to last very long. Preparers are, frankly, more likely to make tag selections for the face of their financials that might have to change later to line up with tag choices in the notes, and that’s not good. Most definitely, analysts certainly don’t get any benefit from block text. That’s a part that I hope will evolve in the next few months to a more refined, sophisticated notion of how it can be done.

My understanding was that block tagging would be for the first year, but then for the second year you would have detail tagging.

Yes, I guess what I am driving at is that feedback will drive the SEC to a clearer understanding that will evolve to the point where they will be able to be a lot more specific about what those detail tags would look like, how they should be done. What I would predict is that there will be commentators who will start to look at the new filings, look at the text blocks and say, well, this is dumb, and they will be right.

I also was more surprised it didn’t say very much at all about attestation standards or assurance standards around the XBRL document. It seems perfectly clear to me what to do, and I guess there’s just been a communication gap or something, and a lot of people attending to other things. But I think it’s a solvable problem and I think it would tremendously increase the value of XBRL documents if they did have an assurance standard associated with them and a clearly defined mechanism for doing that. So, they are seeking input, and I believe it is another thing where the SEC needs to have a solution; because I think people will point at it and say “I can’t use this XBRL data — it does not have an opinion on it.”

(8) Could you describe that assurance mechanism, and how you think it could be accomplished?
 
The basic idea here is, if you look at some of the proposals that have been made about how to indicate assurance in an XBRL document, they have a major flaw: they presume that the person preparing the document is the one saying that he has assurance on its contents. That flies in the face of the whole notion of assurance; assurance is a comment on somebody else’s document. I have a webpage hosted which shows exactly this.

Basically, you would take the opinion and have it point to all and only the facts in another report to which your assurance should apply. So when you say this report taken as a whole [presents fairly the financial position…], there is a very specific technical notion of what it means to “take it as a whole.” You point to all of the facts that are included, and you don’t point at any that are not included. You can take that document containing the opinion — it lives separately from the original, you can sign it, you can encrypt it — and do all those fancy things that you want to. But it’s under the control of the auditor rather than under the control of the publisher of the document.

So what I did was create a little example in which I simply took a hypothetical attestation that someone might give on one of the Microsoft filings. I posted it on my website and it has pointers to all of the facts in a Microsoft filing on the EDGAR site.

So that way a user can look at the document and follow every one of those links and basically construct just a document with just the numbers and the facts and the assertions that the management has actually made to which I am actually attesting.

So it’s a very important notion here. You don’t want preparers to be applying a tag that says that this was or was not audited…it’s not their job to do that.

 

Add to Del.icio.us | Digg this

This entry was posted on Friday, June 27th, 2008 at 6:59 pm 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.