Speech by SEC Staff:
Remarks Before the 13th XBRL International Conference
R. Corey Booth
Chief Information Officer and Director
Office of Information Technology
U.S. Securities and Exchange Commission
May 16, 2006
Thank you very much for the introduction and thanks for inviting me to address the group today. This is my first time attending an XBRL International conference, and it's exciting to see so much interesting work happening around the world. As many of you know, the SEC has been explicitly involved in XBRL-related work for over a year now, and it's certainly helpful to see our peer regulators and many others moving in similar directions. I'd like to spend some time today discussing what we're doing, how it fits into what other groups are doing, and how we can work together going forward. Before getting too far, though, I should remind you that my comments today represent my personal opinions only, and do not represent the views of the Commission, the Commissioners, or the staff of the SEC.
I think it's now clear that there are many at the SEC who are excited about what we typically call "interactive data" that is, improving corporate reporting through technology. Chairman Cox virtually addressed the 12th XBRL International Conference in Tokyo, as well as the US group's last meeting in San Jose. Other Commissioners and staff members have hit the speaking circuit, and the SEC has issued a number of press releases on this topic. In all of those cases, a recurring theme is that using modern data formats is a critical step in bringing corporate disclosure into the 21st century.
First, let me say a few words about the SEC's regulatory charter and its disclosure program. We have over 15,000 publicly traded companies registered with the agency, as well as about 35,000 mutual fund and other investment company portfolios. All of them file financial reports and other information through our EDGAR system creating a total volume of about 750,000 filings annually. The primary goal of collecting these filings is to make them available for public access. Our capital markets are large and deep, and one of the SEC's most deeply held principles is to provide the investing public with extensive and detailed disclosure of company information.
I think the agency can be proud of its use of electronic filing and information distribution. But we can aim higher. Today, the vast majority of EDGAR documents are filed in ASCII text, and another large fraction in HTML. That's fine for reading about a company's strategy and general issues, but if you want to do financial analysis or compare accounting policies between companies, you then have to do a lot of printing, searching, data entry, text parsing, and other mechanical work. Or, you can go to a third-party data provider, who can provide you with a database of financial information but the data provider will have made a number of assumptions to simplify and standardize the financial information, and it may no longer be consistent with how the company intended to present its financials. And you won't get any of the valuable information from the footnotes.
Since you're at this conference, I know you can all envision the attractive alternative posed by XBRL and interactive data, so I won't belabor the point. The potential benefits are persuasive enough greater transparency of financial information, reduced costs for investors and analysts, potentially even deeper coverage of midcap companies by analysts, and ultimately more efficient markets.
Let me paint what I think is an interesting scenario. Wall Street types have been talking for a couple of years about algorithmic trading basically, using computers to process real-time streams of market data and making fast, automated trading decisions. Today, that market data is mostly about stock prices and volumes, since that's what's available in real time. But at some point in the not-distant future, I envision a hedge fund starting to algorithmically trade with XBRL-based balance sheet and P&L data in real-time as it's disclosed by companies. At that point, we will all know that interactive data has won the day.
We're not there yet, but a lot of progress has been made. The SEC, as well as others in this room, have been hard at work for the last few years. In early 2004, the agency launched a task force to look at XBRL and interactive data. That effort resulted in the launch of the XBRL voluntary filing program, with the explicit goal of conducting a real-world test of the interactive data concept. Under this program, any public company or mutual fund can file their financial information in XBRL format in addition to their standard filings. Our requirements are quite flexible companies can file their full GAAP financials including footnotes in XBRL format, and incorporate a variety of extensions to represent their reports the same way as they appear in paper. Thus, our program differs from some of the other programs around the world, in that we have not historically had primarily highly structured forms for reporting, and we do not have them under XBRL either.
Overall, the voluntary filing program has perhaps been an unusual one for the agency. Often, we have been late adopters rather than early adopters of new technologies and approaches, but the voluntary program has placed us closer to the leading edge. This is very exciting, although not always what we first expected.
What didn't we expect? To be clear, the voluntary filing program is an important initiative, and I think we have pursued it appropriately. I must admit, though, I had hoped that we would have had more filer volume. We celebrated the one-year anniversary of the program on April 4, and at that point we had 27 filings from ten companies most of which were technology-oriented companies with their own commercial interest in the success of XBRL. Taking the long view, I think that's understandable. Technology and information interchange standards can take a long time to adopt. Color TV was invented in the 1940's. But it wasn't until 1972 that color TVs outsold black-and-white sets in the US for the first time. It took about thirty years for the US Department of Defense's ARPANET to become the Internet on everyone's desk. But most of us can also observe that the cycle time for technology adoption seems to be moving faster these days, and certainly most of us in this room are interested in the adoption of XBRL sooner rather than later. So, let's examine for a few minutes what some of the barriers to broad adoption seem to be in my mind, there are three broad categories.
First, the preparation of XBRL statements is still perceived to be difficult, and I believe there is also reality behind that perception. There are a lot of companies in the room who are working on software to aid in preparation, and many of these products are very effective and intuitive, but that's not the whole solution. The taxonomies are still the subject of vigorous discussion and further development, and it's clear that they will need to continue to evolve. Our examination of the filings received to date reveals significant variation in how companies are translating their numbers into XBRL, even on relatively basic issues. For instance, filers differ in how negative numbers get represented, or whether line item captions are edited to correspond to those found in the official filings.
Those kinds of issues suggest that there is still a lot of room for technical judgment and interpretation in how to apply the taxonomies to a particular situation. That in turn implies that you need to have significant subject matter knowledge in order to effectively create an XBRL document. In other words, the preparer him-or-herself needs to be comfortable with both the technological aspects and the accounting aspects of the standard. And that kind of subject matter expertise isn't as easy to find as we would like remember, earlier I said we have 15,000 corporate registrants. So, these are the kinds of challenges I see that affect the "supply" of XBRL submissions from companies.
The second issue is on the "demand" side we have not yet seen large growth in demand for XBRL information by the investor community. As with the supply side, there are a lot of great minds at work in this room who are thinking hard about this problem, and certainly we have seen some interesting early efforts to leverage interactive data. I think the problems on this side are less inherently difficult for instance, I don't think every retail investor or Wall Street analyst needs to understand the complexity of XBRL in order to make use of it. Instead, there just needs to be a core group of technology providers and market data providers who understand it, and can provide the information and tools in useful ways. Again, tools exist to slice and dice XBRL data but they are not yet easy to use. We also haven't seen very many truly compelling demos or proofs-of-concept to inspire the potential customers for the data and tools. All in all, we are still not seeing the kind of demand pull that would truly catalyze the market, and in turn make more companies interested in solving the supply-side issues.
The third category of challenges affects both supply and demand, and really is about the core XBRL standards and processing XBRL documents. One important example: the taxonomies for representing US companies need to be evolved and refined, and that will likely involve substantially more staff resources than have been applied so far. I am very interested in finding a way to ensure that taxonomy development is adequately funded and effectively overseen, and also making sure that the taxonomy development process is transparent and able to accommodate input from all interested parties. Other problems are more technical and less organizational, such as rendering: it remains a difficult problem to effectively render XBRL instance documents in human readable format, particularly if extension tags are involved. As another example, validation standards can differ from product to product, which can raise questions as to whether someone's XBRL document is truly compliant with the standards.
There are also some significant conceptual questions about the future roadmap for the XBRL standards themselves, and how they will evolve. In addition to the core XBRL specification, there are a number of supplemental modules on various development tracks; these include specifications for versioning, dimensions, formulas and functions. These developments in turn will impact future taxonomy efforts, and the creation of XBRL enabled software. So, the market needs clear information about the future development of the XBRL specifications, so decision makers and developers can make the right investments to support these future capabilities.
The SEC will continue to look hard at these challenges across all three dimensions of supply, demand, and the core technology, and we will continue to monitor how the market is evolving to address them. I know that many people in this room, as well as in the various countries involved in XBRL, are thinking about many of the same challenges, and I'm sure that collectively we have some great ideas for how to proceed. Since I have the podium for now, though, let me share a few thoughts for where I think we ought to be heading.
I'll start at home with the SEC. I think the agency has an important role to play in the world of interactive data, and XBRL in particular. I intend to work with the Chairman, the Commissioners, and the staff to continue helping play that role. But I should underscore two items which I think should serve as guiding principles for the agency's ongoing efforts.
First, I believe that our role as a government agency should focus for now on supporting adoption, not dictating it. I am often asked whether the agency will quickly move towards mandatory XBRL filing. In its rule proposals last year, the Commission stated that one purpose of the Voluntary Filing Program is to be able to evaluate better whether mandatory filing might make sense in the future. But my personal belief is that a mandatory requirement ought to happen only after voluntary adoption spreads further. I can point to many examples around the world where prematurely setting technology standards can result in failed initiatives or longer-term problems whether in financial regulation, telecommunications regulation, or even private sector standards setting. I believe we have not yet reached an inflection point where there is significant critical mass and knowledge of the XBRL standard, and so for now I think we can do better by aiming for a market-driven transition to XBRL rather than a regulatory mandate.
The other guiding principle I think the SEC should continue to pursue is to continue giving registrants as much flexibility in the presentation of their financial statements as they currently do under US GAAP accounting. This is a different position from that taken by our colleagues at the FDIC, where their call reports tend to require that member banks report their numbers in more of a fill-in-the-blank approach. Of course, many other securities regulators globally have historically had a fill-in-the-blank approach, and so it makes sense that they would do so under XBRL. But our approach has always been based on full GAAP accounting treatment, in order to give a full and nuanced view of a company's financial position, and I don't see that changing. The fact is that the universe of SEC registrants is extremely diverse, and we need to account for that. Unfortunately, GAAP accounting has gotten rather complicated in the several hundred years since the medieval Italian merchants invented double-entry bookkeeping. While I'm sure there are opportunities to reduce the complexity of GAAP and IFRS, I think the SEC's XBRL needs to conform to the accounting standards, not to force a simplification of them. So, these requirements for flexible accounting create the need for flexible taxonomies and extensions. Those in turn create challenges for validation, rendering, analysis, and other technologies. But if we want to accurately present company financials as audited and as reported, which has always been the SEC's focus, this is what will be needed.
Besides posing challenges for the group, though, the SEC can also be helpful. I think we have a valuable role to play in working with the various players in the XBRL "ecosystem" to track and encourage its evolution. I think we can also play a role in facilitating dialogue around the specifications and taxonomies, and help bring the Consortium, the accounting standards bodies, and other players together. We are making another big move in this area beginning in June, when we will host the first of a series of roundtables on "giving investors and analysis better financial data via the Internet". This should be an exciting forum looking at all sides of the issues I've outlined today, and is open to the public. We certainly encourage any of you who happen to be in the US at that time to come to the event; we also are accepting public comments via our website, and we would encourage your thoughts on the session.
Besides what we do and say in public, I also think the SEC should continue to make selective investments in technology for our own purposes to better validate, analyze, store, disseminate, and present XBRL information to the public. For instance, we recently introduced a new RSS feed for the XBRL documents we post to the SEC website. That will enable interested people to more easily locate XBRL filings, or take other steps to make the information more accessible.
In addition, we are being more aggressive in expanding the voluntary filing program and cultivating a dedicated base of filers. In March, we announced a new "test group" of 17 companies across a range of industry sectors who have committed to filing in XBRL format for a year. In return, we are offer some incentives related to our reviews of their EDGAR filings. Participating companies will receive expedited reviews of securities registration statements. If the participant is what we call a "well-known seasoned issuer", we will inform the company whether or not the staff will select their annual reports for review, within 30 days after filing. The test group is still open to other interested companies, and we would welcome any foreign registrants who currently file US GAAP financials with the agency, so any suggestions and referrals for volunteers would be appreciated. Besides building up additional filing volume, our intention is to work with the test group to better understand the issues and effort associated with encoding and filing XBRL documents with the SEC.
Obviously, the SEC is only one of the players on the field. There are many important things others are doing and should continue to do to help work on the challenges I've outlined. Technology-oriented contributors should continue to refine their products, and help work on some of the unsolved problems associated with rendering, human readability, and other "tough issues". Some of these can be solved through individual firms' innovations; in other cases, it probably makes more sense to build them into the core set of standards promulgated by the Consortium.
With regard to those core standards the XBRL spec, the linkbase definitions, the taxonomies, and other definitions that ought to be universal my high-level guidance is that we need to strive for focus and simplicity where possible. The world of financial information is hugely complex, and covers the waterfront from general ledger accounting to tax preparation to internal control audits to disclosure to equity analysis. Working XBRL into all of the specific niches will take a long time, and require a lot of work to effectively encompass the various requirements. But the question for now ought to be: where are the most critical leverage points across the whole chain, where a relatively simple implementation of XBRL can create real value in the shorter term? And given that, what can we here do to make the standard successful and broadly adopted in that niche?
Here's one example: I think the idea of using XBRL to encode earnings release information is a very powerful one, whether it takes place through SEC 8-Ks and 10-Qs or through the press release services. This type of information is often relatively simple and condensed, and shouldn't require massive taxonomies. But it's also very valuable and time-sensitive information, and analysts are keenly interested in quickly and accurately digesting it. So, any efforts by the news and data services to promote the distribution and consumption of earnings information in XBRL format would be a major step forward in driving demand for XBRL among investors and analysts.
Another example of simplicity might be found within the SEC's own voluntary filing program. Our rules require participants to file the main financial statements in XBRL format, but they are not similarly required to file the notes in XBRL. Now, I think the notes are where a lot of the most useful information lies, but companies also seem to be facing a disproportionate amount of effort and challenge in encoding their footnotes. One approach could be to just file the main financial reports in XBRL the first time out, and begin providing XBRL footnotes a quarter or two later. Maybe this kind of simpler, walk-before-you-run approach would induce more companies to stick their toes in the water.
It's also worth thinking about mutual fund disclosure in terms of simplicity and accessibility. Unlike investors in equities, mutual fund investors are most interested in a narrow set of information investment objectives, costs and fees, performance, holdings, managers and today our disclosure model requires them to root through a bulky text-based prospectus to find it. While this is not financial reporting data in the same way I have been talking about, it's clearly highly amenable to interactive data concepts, and I am hopeful that the Commission will find ways of creating a simpler way for mutual fund investors to find, analyze, and compare the information they need through improved mutual fund disclosure.
Finally, more broadly, let me raise the important question of building a critical mass of XBRL expertise. I think it's too much to ask in the shorter term that each of the SEC's 15,000 registrants have staff who are comfortable preparing XBRL filings, which means that help is more likely to come from external service providers. Over the past several months, a number of firms as well as the XBRL-US Consortium have offered a variety of free and commercial services for helping companies create and file XBRL documents. That is extremely helpful, and we're looking forward to seeing more takeup on those offers. Beyond that, though, there is another very large body of professional services talent in the accounting firms and law firms who help companies in financial reporting, and I believe they have an opportunity to provide help as well. Somehow, we need to figure out how to get past the fact that the number of people at XBRL conferences is far smaller than the number of companies who can be XBRL-izing their information.
So, to close out my portion of the day: I've wanted to celebrate the truly impressive progress that has been made, and also to give some food for thought to the global XBRL leadership team about how we can continue that progress going forward. Make no mistake: interactive data in general, and XBRL in particular, is a great opportunity to help investors take full advantage of the information they are given, and to improve the efficiency of the financial markets. I look forward to continuing to work with all of you to make this a reality. Thanks again.