Written by Bob Schneider    Posted on April 26, 2007

“Transforming Financial Information: Use of XBRL in Federal Financial Management” is a white paper recently released by the Industry Advisory Council and freely available online. In my post of April 20, I wrote about the paper’s discussion of the information supply chain and the tradeoff between information reach and richness. In this second and final post, I’ll discuss why the authors believe XBRL can substantially improve financial management at Federal agencies.

On page 32 of the report, the authors begin a detailed overview of the Federal budget process, whose participants are the individual agencies, the executive office, and the legislative branch. Not surprisingly, the process includes a stunning number of reports whose information greatly overlaps.

Throughout the fiscal year, agencies are required to submit a series of reports to various entities like OMB that wind up being consolidated into the Financial Report of the United States Government. On page 36, ten of the primary reports in the process are described; but the authors warn that they represent “only a sampling” of financial information exchanges and the reporting responsibilities of agencies. Indeed, one Federal employee says “There are so many reports involved, pulling data through such a wide set of systems, that no one person really understands it in totality.”

There is a mind-boggling chart on page 38, reproduced below, that demonstrates the enormous amount of overlap in Federal financial reports. (You can click the image for a better view.) The reports are often in formats like PDF which, while easy to read, are not easily reusable, or flexible for computer-based analysis.

The authors comment:

In the end, the magnitude of cost created through redundant reporting provides limited value, since the information is reported in a wide variety of formats using a copious number of reports, technologies, and systems, making it costly to obtain a complete picture of a single agency and difficult to appreciate the Federal government as a whole.

On pages 39 to 45 of the report, the authors outline areas of opportunity where XBRL adoption would achieve vast improvement. As I read through the details, I noted the many benefits mentioned were those consistently championed by XBRL supporters:

• Consolidate many similar information requests into one, thus greatly cutting overall costs;
• Automate manual processes;
• Eliminate reconciliation points;
• Standardize terminology while keeping it flexible;
• Deal successfully with non-quantitative information;
• Share data among reporting entities;
• Monitor capabilities more effectively;
• Reduce the number of data calls.

The authors end with a series of recommendations for implementing XBRL and getting started. I suppose it was beyond their mandate and scope, but I would have appreciated some discussion of the winners and losers in XBRL adoption at the Federal level and what the political ramifications might be. In other words, does the political class want this? Does the bureaucracy? I also wished more attention had been paid to costs of implementation and the roadblocks, and the likelihood that actual improvements would be achieved.

Nevertheless, this is an extremely useful study. The paper is another reminder that, while interest in interactive data remains substantially focused on financial reporting, the potential of XBRL for transforming how public and private organizations work both internally and in combination with external audiences is enormous.

Written by Bob Schneider     Posted April 20, 2007

Any writer who puts “white paper” in his opening sentence has to worry that readers are already suppressing yawns. Tell them that the paper he wants to discuss is exciting and intriguing, and the yawns may turn to smirks.

Admittedly, members of the Paris Hilton Admiration Society are unlikely to be enthralled by the recently published “Transforming Financial Information - Use of XBRL in Federal Financial Management” (freely available at the site of the Industry Advisory Council, under whose auspices it was written). But for those interested in how XBRL can revolutionize information collection, processing, and analysis by Federal agencies, this paper indeed provides stimulating reading.

In this first of two posts, I’ll discuss two salient points the authors make, namely, (a) why the information supply chain is so important, and (b) the tradeoff between richness and reach in information. In the second post, I’ll look at why XBRL holds such promise for Federal financial management. These brief capsules won’t do justice to the authors’ argument, but I hope they will encourage you to read the entire paper.

On page 17 of the report, the authors describe a general model for reporting, which can be viewed simply as “a pattern of pairing requests and responses between organizations.” The authors say that it is often assumed financial information requests are completely and accurately understood.

What actually happens is that (a) responders (ie, suppliers of information) can’t decide what’s needed and requesters (ie, collectors of information) have to help them figure it out; or (b) responders wrongly assume they know what’s needed and respond incorrectly. Either way, there’s more expense for both requester and responder..

In traditional, paper-based reporting, these costs rapidly multiply because each report has a single, narrow audience with a single narrow purpose. Each time you add a new audience or purpose, you add a new report producing more back-and-forth that generates more cost.

In contrast, consider the information supply chain in a digital environment, where information is passed along a chain of interested parties so many organizations can share the same information amongst themselves. The authors use the diagram below, taken from page 22 of the report, as an example (click the thumbnail to see the image):

” Information that is requested of one body (#1) may in turn become a request to another party (#2) with the response (#3) being passed, in part or in whole, back along to the original requestor (#4).Ultimately, the information may be shared out, in part or in whole, with other interested parties (#5). When the pairs of requests/responses are chained together they can be viewed as an information supply chain. The concept of an information supply chain emphasizes the notion that information can be reused, which is to say, it is not just gathered and provided once, but is passed along a chain of interested parties.”

With the information supply chain, many related requests can be combined into one, and the responses can be shared among many parties. The numerous audiences for this one body of information can receive all or just part of it. And this one body of information can have many purposes. Thus costs are reduced both by eliminating redundancy and sharing expense across the supply chain.

XBRL underpins these efforts, because at each step of the chain data can be viewed, analyzed, and manipulated without compromising data integrity. XBRL automates the processing of business information so that manual processes of validation, re-entry, and comparison are eliminated. And XBRL allows data to be provided once and reused again and again, so it can be shared across a broad base of users and for many purposes.

Another powerful concept the authors introduce is the trade-off between information richness and information reach. Richness is the quality of information, while reach is the number of users that can share it. Richness includes the amount of information; the ability to customize it; interactivity; reliability; security; and timeliness. The diagram below, taken from 28 of the report, shows the tradeoff between richness and reach (click the thumbnail to see the image).

Traditionally, rich information required people to be physically close to one another, but the Internet allows much greater reach for much richer information.  As a semantic technology, XBRL helps associate different types of information to enable much broader and richer information discovery and exchange, and pushes out the ceiling at which the tradeoff must be made still further. The ultimate result is much richer information for a much wider audience at reasonable cost.

In my next post, I’ll discuss specifically the authors main focus of how, by adopting the model of an information supply chain that can supply richer information to a wider audience at relatively low cost, XBRL can substantially improve Federal financial management.

Written by Bob Schneider     Posted on April 15, 2007

“There are eight million stories in the Naked City. This has been one of them.”

I was reminded of that tag-line from the famed TV crime drama when I recently re-read Max Rottersman’s post on XBRL and Mutual Funds Compliance in which he wrote:

“I’ve witnessed millions of dollars and careers that have been ruined from incorrect or cavalier interpretations of the rules.  I’ve never heard of someone losing their job over bad data. Therefore, XBRL will have a hard time selling itself to the fund community based on arguments for efficiency and data integrity.  These promises are out-of-touch, if not irrelevant.”

Since Max has been the chief compliance officer of a mutual fund, I’m certain he knows what he’s talking about and I have no reason to contradict anything he says.

But I would like to suggest that, outside the mutual funds neighborhood, there are stories in the Compliance City — maybe not eight million, but more than a few that carry a slightly different message. Here’s one of them.

Long ago, for several years, I helped assure that the equity research of a large investment bank complied with Rule 472 of the New York Stock Exchange governing communications with the public by member firms. The rule has been changed and strengthened considerably since I helped administer it. But first as staff to the department’s Supervisory Analyst and later as an SA myself, I had two types of responsibilities.

First, I was to ensure the analyst had a reasonable basis for making his recommendation (don’t ask me how an SA could sign off on an Internet buy call based on the number of clicks its website got I don’t know). Second, I had to make sure the analyst didn’t write things like “This winner is the Alice Kramden of all stocks — it’s going to the moon, baby!”

Because most heads of research don’t want to read analyst reports all day, the SA function is often delegated to people with (1) the (minor) analytical abilities necessary to pass the required Series 16 SA exam, and (2) some writing skills, so the report can be both SA’d and edited by only one set of eyes. (Of course, the lawyers reviewed all reports for legal matters, like Chinese Wall issues; but they didn’t care about anything else.)

In any organization, most of the prestige and power go to the people who bring in the money. At a big brokerage, that’s the investment bankers and traders. Although the words of big-name research analysts can move stock prices, the research product itself (mostly) commands only so-called soft dollars, which is not nearly as much fun as the hard stuff. In short, investment research is staff, not line. And if you’re a staff person in a staff department like I was, you really are a nobody.

But a nobody with an additional responsibility beyond simple compliance: to make sure the firm published a quality editorial product. And that’s where’s data accuracy comes in.

Because even though it wasn’t my job to check the analysts’ numbers, it was my job to make sure readers could trust what they were reading, and that the firm’s brand was protected. Because if an EPS estimate is shown as actual, or if 200X and 200Y data are reversed, or if pages 3, 16, and 22 of a report each have different numbers for cash flow, then readers begin to lose faith in what they’re reading –not just for this one analyst but for all your other publications as well. In other words, whatever else the rest of the firm was up to, I was working in or at least considered myself to be working in — financial publishing.

I don’t want to make it sound like I stood there with sword unsheathed, fighting for data accuracy against a horde of sloppy and indifferent analysts. Obviously, no analyst wants a mistake-ridden report. But as analysts, they simply had different priorities  99% accuracy, especially on non-essential data, was certainly good enough. Of the thousands of reports I SA’d, I had analysts yell at me dozens of times for not getting their research published yesterday, but never once about messing up their numbers.

Most of the work I did made little difference. In retrospect, a lot of the time I was overzealous about producing a first-class editorial product when I should have been more attentive to analyst priorities, not to mention profit enhancement.

But very occasionally an SA erred in the other direction, and it cost us  big time. One SA screw-up especially comes to mind (no, it wasn’t me, but there but for the grace of God). An analyst did a report on a Swedish company, and somehow during the production process, krona-denominated figures were wrongly presented in US dollars. Upon reading the published report, the company, which had been an investment banking client, figured our firm didn’t know what it was doing and took its business elsewhere.

What points am I trying to make? First, the compliance function may be used for non-compliance needs. The risk of sanction by the NYSE was small; but the necessary compliance function underpinned the objective of creating a first-class editorial product, which was deemed (rightly or wrongly) by management as a significant selling point for the firm’s research. Making very bright, aggressive research analysts submit to an editorial process that wasn’t somehow coupled with a compliance requirement would have been very difficult.

That the editorial process may have gone too far, and may have even introduced errors into the process, is cautionary, but irrelevant to my point management wanted significant control over its research product. And I’m certain compliance is used in similar fashion by other powerless entities in organizations to maintain some degree of control over powerful people who might otherwise be difficult to tame.

Second, although it’s impossible to know if the adoption of XBRL would have assured that the Swedish company’s data would have been published correctly, my sense is that interactive data will go a long way to reducing the possibility of such mix-ups in currencies as well as mistakes like wrongly labeled columns for yearly and quarterly data, forecasts presented as actuals, and so on.  That’s probably not a big selling point for XBRL among individual analysts, but it is certainly a positive for any organization that publishes financial information and wants to maintain its credibility in the marketplace.

Written by Dr. Robert Eccles     Posted on April 10, 2007

Dr. Robert Eccles, an internationally recognized expert on corporate reporting, is a founder and managing director of The Perception Group.  He has been a professor at the Harvard Business School, where he served on the faculty for 14 years. Dr. Eccles is an advisor to the Enhanced Business Reporting Consortium.

It looks like XBRL is finally getting good traction in the United States thanks to the SEC’s interactive data initiative.  The new XBRL-US organization is working on completing a taxonomy for financial statements with a deadline in the summer of 2007.  Ultimately this taxonomy is intended to include footnotes, although I’m not sure of the timing. This will clearly be a big step forward. In addition, the Investment Company Institute recently released a taxonomy for the Risk/Return Summary of the mutual funds prospectus, and Broadridge, a spin-off of ADP, is working on developing a taxonomy for the Notice and Proxy Statement.  Certainly the capabilities of XBRL extend beyond financial data to cover virtually any type of regulatory filing, although as the data becomes more qualitative and narrative in form, the challenge becomes greater.

Another example of expanding XBRL beyond financial statements is the work being done by a nonprofit group called the Enhanced Business Reporting Consortium.  The EBRC is attempting to organize a market-based initiative involving the collaboration of companies, investors, and a wide range of intermediaries like sell-side analysts, rating agencies, and accounting firms.  The charter members of the EBRC are the American Institute of Certified Public Accountants, Grant Thornton, Microsoft, and PricewaterhouseCoopers.  On December 6, 2007, members of this group met with SEC Chairman Christopher Cox to talk about how XBRL could be extended into broader narrative reporting, such as for the MD&A and 10K, and for industry-specific key performance indicators.  These ideas were well received and the EBRC is now working on putting together a detailed action plan.  It should be emphasized that this is a market-driven initiative that has SEC support, rather than being an SEC-driven initiative.

Another important point is that the EBRC seeks to make this a global initiative.  Although the original members are all U.S.-based organizations, the EBRC is already collaborating with groups in Japan and Europe.  Why repeat the experience with accounting standards and have multiple frameworks?  After all, the type of information contained in a 10K such as that on the business environment, strategies, and risks — is relevant to companies and investors all over the world. Similarly, industry-specific key performance indicators are relevant to every company in an industry. The geographic location of its headquarters is irrelevant, particularly in a global economy where the Internet is becoming an increasingly common location where business takes place.

A very rough draft XBRL taxonomy for the Enhanced Business Reporting Framework Version 2.0 has been developed. However, this taxonomy was produced before it became clear that there is sufficient support for the EBRC to be a global initiative.  The next step is to convene representatives from Japan, Europe, and the U.S. to develop EBRC Version 3.0.  An XBRL taxonomy will be developed for this, hopefully using a collaborative taxonomy building tool.  We will keep Data Interactive readers informed of the EBRC’s progress in these areas.

Written by Bob Schneider     Posted April 5, 2007

As reported in the Chicago Sun-Times, new SEC rules compel companies to disclose much more information about executive compensation. The highly detailed executive compensation data disclosed in annual proxy statements will include a consolidated table that shows total annual compensation for top managers. For the interactive data community, the notable element of the new rules is that data for hundreds of the largest firms will be available online in XBRL. SEC Chairman Christopher Cox commented on the subject in his recent address to the Corporate Counsel Institute:

” Even before interactive data becomes the norm for all reporting companies, we’re going to tag the executive compensation data for you using XBRL. And we’re going to put an interactive data web tool on the SEC’s website, to let users slice and dice the executive compensation data any way they like or do industry comparisons, or even do analyses of particular forms of compensation, such as stock options. We’re going to do this for at least several hundred of the largest public companies in America and we expect to have it available in June. The truth is, investors and their representatives on the compensation committees of boards of directors have a right to this information. And they have a right get it in a form they can really use.”

I don’t want to get into a debate about executive compensation levels, fascinating as that would be, or the SEC’s new disclosure rules themselves. (For those interested, SEC Commissioner Roel Campos’s recent speech at the Summit for Executive Compensation highlights important issues of the argument.)

I think it’s uncontroversial to say that the enormous rise in executive compensation relative to employee earnings which according to one (perhaps overstated but indicative) estimate rose from 42:1 in 1982 to over 400:1 in 2004 — has to give even the most vigorous free-market advocate pause. I also think it merely betrays a flair for the obvious to suggest that the new rules should be given a chance to see if they have a benign effect on executive compensation levels. And putting  executive compensation data online in XBRL format indisputably represents another step forward in interactive data adoption that XBRL advocates can celebrate.

But I have to admit to some reservations. I find it disconcerting that investors will have easy access via XBRL to fairly arcane data on executive compensation before they have the same privilege for items like sales and profits. Moreover, I’m not sure how smaller investors will use this information to make better investment decisions.

As a citizen, I want executive compensation to be reasonable. As an investor, I want the greatest return on my capital. I much prefer a very richly compensated CEO who delivers strong earnings and stock price appreciation that exceeds market and sector averages to a fairly compensated executive who produces poor results and below-average returns.

I recognize that no one is telling investors to use executive compensation to the exclusion of all other factors in making investment decisions. But comparing compensation to company and stock performance under a particular management team is a complicated and difficult task, given the trade-offs between achieving long-term company objectives and short-term stock price movements. Institutions and other sophisticated investors have the smarts to use (or ignore) compensation data as they see fit; but I fear some small investors could be lured into making simplistic judgments based on executives’ total compensation levels.

Another concern I have is that XBRL will become associated in CEO minds as simply a way to browbeat them about how much money they make. This is the second time in recent months that Chairman Cox has celebrated XBRL for its use in gaining greater transparency in executive compensation, the other being disclosures about the backdating of stock options.

There can be no dispute that bringing such malfeasance to light is welcome. Still, it is another instance where XBRL is being used to disclose data that may have greater implications for corporate governance than stock performance. Interactive data needs the support of company managements to be a success. If it’s seen primarily as a useful vehicle for those with a political agenda against CEOs in particular and large companies in general, I fear the business community will be less likely to offer its support.

More broadly, I recognize that reporters have a limited amount of real estate for their stories, and that those articles reflect the news of the day. Still, it would be nice if there was more mention in the press that XBRL is a truly international standard that many countries besides the US are adopting for financial reporting, and that its uses extend well beyond those of corporate policeman. Indeed, XBRL provides outstanding advantages for the exchange and integration of both financial and nonfinancial data for everyday business activities, both within and outside the organization.

Despite these misgivings and objections, the implementation of XBRL for executive compensation disclosure represents a significant step forward. I continue to be pleasantly surprised at the extent of Chairman’s Cox’s commitment to XBRL adoption. As he champions interactive data in his speeches and looks to new avenues to adopt XBRL, I hope he’ll focus on the advantages that interactive data can provide to all investors as well as the broader business community, as opposed to those greeted most warmly by corporate watchdogs.