31 May 2011

XBRL: Meeting the Needs of the Financial Analyst

Today we have a guest post from one of the finest voices in the world of XBRL. Bob Schneider is a San Francisco-based business writer. He was formerly editor of Data Interactive, the Hitachi XBRL blog. The views expressed are his alone, as is the style, making reading Bob's views an ongoing pleasure.

XBRL: Meeting the Needs of the Financial Analyst

I’ve just finished reading The Leopard by Giuseppe di Lampedusa, a superb novel  (made into a wonderful Burt Lancaster movie) about a Sicilian aristocrat at the time of the Risorgimento. Written in Italian about events remote both in time and place, there must surely have been numerous words and passages that were extraordinarily difficult to render in English. Yet the book’s translator, Archibald Colquhoun, rarely succumbs to the temptation to merely italicize the original Italian, leaving the reader to decipher these unwanted “extensions” to his personal dictionary. No doubt there are imprecisions, but the prose remains smooth and graceful throughout, and the reader’s pleasure uninterrupted.

The tasks of the literary translator, and those of the accountant who maps financial statement items to XBRL taxonomies, have little in common. But with respect to word choice by the former and element selection by the latter, their marching orders are similar. Both should choose from the existing glossary the term that most nearly matches the intended meaning; only when no term corresponds should the lexicon be expanded.

The art, of course, is deciding where nuance requires neologism. I remember a TV host asking the awesome wordsmith William F. Buckley Jr. – who once wrote a column titled I Am Lapidary But Not Eristic When I Use Big Words – why he so often violated the admonition to writers to use familiar terms. Buckley replied that the interviewer had only stated the first half of the rule; the full injunction is to use the simplest word that most completely conveys the writer’s meaning. (Buckley’s response is more fully and elegantly conveyed in the cited article.). There’s a constant tension between familiarity and precision and anyone who puts fingers to keyboard will make his choices (eg, IMHO, there’s little to be gained by using eristic instead of controversial or disputatious.)

But writers and translators know that, even if eristic isn’t in our personal dictionary, we can find it in Webster’s. The Webster’s of the US financial reporting world is the 2011 US GAAP Financial Reporting Taxonomy which has some 15,000 elements. Whether this number is small or large, adequate or inadequate, can be debated (as can the 15,000 estimate itself). What is indisputable is the large number of extensions to the taxonomy that companies are using in their XBRL exhibits. In a March 21 blog post, Dan Roberts reviewed some 1,500 filings and noted that 23 “have over 50% extensions, with the highest being 67% which translates to two out of every three data points reported and tagged in XBRL were extensions.” In a recent Compliance Week article, Dan states that “extensions so far represent about 11 percent of all data reported into the XBRL system.”

Imagine trying to read The Leopard with one of every ten words in Italian!

Let’s assume – wrongly – that all of these extensions provide more precision and information than existing elements in the taxonomy. Taken as a whole, might extensions represent significant added value that analysts can utilize, offsetting any added difficulties in comparability? 

Financial statement data is essential to the work of the securities analyst, but it’s also important to understand its limitations for his needs. The data is expressed in the language of complex and voluminous accounting rules that can offer its authors wide latitude in methodology. It is historical data, and while it contains clues about the future, it’s mostly about the past. It’s completely public and available to everyone, which means the analyst’s ability to gain any marginal advantage from it – as opposed to proprietary information that can be gleaned -- is difficult and limited.  By “proprietary,” I don’t mean anything nefarious, like unauthorized insider information; rather, it is all the nonfinancial knowledge – management quality, new technology, industry structure, customer sentiment, the regulatory environment, and so forth -- that analysts most vitally bring to bear.

Moreover, there are analysts, and then there are analysts. I once attended a career planning session that included a panel discussion with investment professionals, among them two securities analysts. They were each asked how important accounting was in their work. One said that accounting was the language of business -- if you didn’t like accounting, forget about a career in the financial world. The other, a technology analyst, hesitatingly said he thought there might be someone over at Bear Stearns following accounting issues.

Both responses were perfectly reasonable: for some industries, like banking, the reported numbers – and the rules under which they are determined -- are crucial, while in others (certainly technology) stock prices are driven by “proprietary” information. Of the thousands of equity research reports I’ve read, only a minority made any mention of accounting standards, and often just to clarify year-over-year comparisons.

The relative usefulness of financial statement footnotes – which promises to turn the steady flow of extensions into a tidal wave – is brought into sharp relief by the results of a FINRA-sponsored study published last August in the Journal of Accountancy. Examining how “investment professionals” (including both buy-side and sell-side analysts) and “nonprofessional investors” use the company annual report, the study found that while 68% of professionals viewed at least one of the 21 footnotes, 12 of the 21 footnotes were viewed by fewer than 20% of professional analysts.  The highest percentage of professionals viewing any single footnote was 37% (the disclosure of “subsequent events”).

A possible counterargument of XBRL proponents would be that this low utilization is exactly the reason we have the mandate, which will soon allow investors to penetrate opaque footnotes and unlock their analytical value. Perhaps, but a recent letter to the SEC by the Committee on Corporate Reporting of Financial Executives International does not indicate latent demand for any XBRL data:
Of those companies that track the usage of this information [ie, XBRL filings] on their corporate websites, none reported more than a slight interest in this information from the investor community. The number of hits ranged between 3 and 20 hits, per quarter and some of those may have been either employees of the company or the service provider verifying the accuracy of the final XBRL data posted on the websites.

Whatever efforts one might make toward mitigation – the difficulty of finding XBRL data on corporate websites, the lack of a “critical mass” of XBRL data that would make it worthwhile, alternative sources for XBRL exhibits – a neutral observer must admit these are very low numbers indeed.

Ralf Frank worries that XBRL “still looks and feels more like a concept car than one [investment professionals] can drive off the lot.” In a recent interview with Hitachi’s Data Interactive blog, he ruefully complains “What about downstream processes? The corporate reporting supply chain – that is how those of us in the investment world would see it – does not end in the database of a regulator.” While recognizing the value of the “real assets” of the IFRS and US GAAP taxonomies, he extols the virtues of a “Core Financials” approach:
We have always been in favor of a strategy for XBRL that rests on spreading it as quickly as possible….Under a “Core Financials” approach reporting would be wide, not deep. XBRL instances containing broad, core financial summary sheets for a large number of companies would be provided rather than detailed, granular instances on fewer companies. So investment analysts would have, say, 50 items from each of 10,000 companies instead of 1,500 items from each of 1,000 companies.

But has this current lack of utility of XBRL exhibits made me an apostate? Not at all. I still believe the global data standard for business information, when deployed with converged IFRS/national GAAP accounting principles and rules, will offer extraordinary benefits for both investors and the broader class of economic actors. But the focus must shift to the needs of end-users, the consumers of financial information. A 10-K will never be as much fun to read as The Leopard; but through the efforts of the XBRL community it can be made more transparent, understandable, and rewarding.  

1 comment:

  1. Thanks Bob for a thoughtful and though-provoking piece.

    Perhaps Inline XBRL will be the production vehicle based on the XBRL concept car chassis... It naturally provides for the "Core Financials" wide-not-deep approach to reporting.

    In the UK Statutory Accounts for all 1.6m companies are required to be filed online to HMRC in Inline XBRL format (that's as "wide" as it could possibly be under the circumstances), but those Accounts typically contain a relatively shallow 2-300 marked up facts drawn from a concept vocabulary that it initially smaller than, and ultimately no larger than, the extent of the standard UK Accounts Taxonomies. There's no need to mark-up disclosable detail with non-comparable private extensions simply to convey the information. If a human ever needs to delve into this detail they can, in the traditional manner.

    Admittedly this data is currently consumed by HMRC and never sees the light of day again, but the UK company registrar, Companies House, is following on HMRC's coat-tails and is mandating the exact same information, in the exact same format, for the public record, starting in 2013. It will then be available to anyone equipped with some simple web-based, Inline XBRL-aware, data-mining tools.