The Consortium of EDGAR Filing Agents and Software Developers

October 29, 2004

Mr. Jonathan G. Katz
U.S. Securities and Exchange Commission
450 5th Street, N.W.
Mail Stop 6-9
Washington, D.C. 20549-0609

Re: S7-35-04 Proposed Rule: XBRL Voluntary Financial Reporting Program on the EDGAR System

Dear Mr. Katz:

The Consortium of Filing Agents and Software Developers appreciates the opportunity to provide comments on the subject rule proposal.

General Comment on the Proposed XBRL Pilot

We commend the SEC for initiating the discussion of XBRL outlined in the Proposed Rule. Given XBRL's general complexity as well as the current state of the technology, we share your views that a Pilot project is the best way to afford all participants an opportunity to fully assess the technical challenges presented by XBRL, its cost of implementation, as well as XBRL's utility.

Responses to Specific Request for Comment

    1. Is the proposed rule permitting volunteer filers to furnish financial information in XBRL appropriate? Is there a better way to accomplish testing and analysis of XBRL data?

Given the variables involved and the current state of XBRL tools and technology, we believe the proposed Pilot is the best available approach for the SEC at this time. Although there exist today datasets of XBRL-tagged financial statements, a project that does not involve issuers directly would fall far short as an adequate test of the technology.

It is worth noting, however, that while the proposed Pilot project will give any issuer wishing to participate an excellent view of the technology, how it works, and the cost and level of effort currently involved in creating and filing an XBRL instance document, the Pilot will, in all likelihood, not give either the SEC or the document user community an optimal opportunity to fully explore the benefits of XBRL-tagged data from an analytical perspective. Any successful analytical program will need to rely heavily on a comprehensive set of corporate financials that is complete not only with respect to the documents that are tagged, but also provides data across a reasonable number of years. The fact that a very limited number of issuers will provide tagged instances for what is likely to be an incomplete set of financial statements, calls into question a major component of the SEC/analyst side of the Pilot.

Issuers will benefit from the pilot by virtue of the fact that they can explore the technology as extensively or as superficially as they choose. The SEC, however, will be disadvantaged by those same factors. For the SEC to adequately assess the potential XBRL may hold on the analytical side, the SEC will most likely need to go outside the Pilot to properly develop its analytical routines and capabilities in order to accurately assess XBRL's ultimate utility.

    2. For purposes of the program, volunteers can furnish in XBRL format, among other types of financial information, a complete set of financial statements. Are there special issues or difficulties raised by providing notes to financial statements in XBRL format? If so, should we permit volunteers to furnish financial statements in XBRL format if they omit the related notes? Should we allow volunteers to furnish in XBRL format some but not all financial statements (e.g., only a balance sheet)? Should we also allow tagging for other items, such as Management's Discussion and Analysis or Management's Discussion of Fund Performance that are part of existing taxonomies?

With respect to MD&A, we recommend the Pilot be limited to tagging financial statements.

Getting financial statement footnotes tagged in XBRL presents a considerable challenge, since, as we understand the current state of affairs, it is not possible to tag footnotes in a manner that would tie a specific financial data element in a footnote to the correct caption within the statement. We recommend, however, that the footnotes be included in the XBRL instance document, even if footnote tagging takes place at the paragraph level or higher (e.g., a single tag that precedes all of the footnotes for a specific document or statement). Such a decision to tag at the higher level will not degrade the value of the Pilot project. The issue of extending the tagging requirement to all financial statements is an especially difficult one for the SEC. Extending tagging to all financials will almost certainly reduce issuer participation in the Pilot. Not having a consistent dataset across participants or across time for the same participant, presents the SEC with the problem discussed in our response to Question 1 above. One compromise might be for the SEC to require that all participants, if they elect to participate in the Pilot, furnish, at a minimum, the same statement(s) each time they submit an XBRL instance.

Alternatively, the SEC may wish to discuss Pilot participation with each issuer and negotiate the base statement to be provided by that specific issuer.

    3. Are the standard taxonomies in the voluntary program sufficiently developed? If not, explain what further development would be necessary. Please address taxonomies with respect to specific industries or types of companies if you have information or views on these. Is the taxonomy builder software sufficiently developed that volunteers would be able to create extensions as needed?

Regardless of how complete XBRL taxonomies are (or may become), it appears evident that no standard taxonomy will prove sufficient to accommodate the accounting captions of the majority of firms using it. Furthermore, given XBRL's "extensibility," it should not be necessary to construct taxonomies that are complete for every issuer. In fact such an effort defeats the underlying purpose of XBRL.

If the SEC elects to implement a Pilot that fully explores XBRL, company-specific extensions will be at the heart of the experiment. If there are problems with this technology, they will show up in XBRL's ability to handle extensions. From our perspective, only by having issuers create company-specific extensions, will the SEC, the technology community and pilot participants be able to gauge the accuracy and efficacy of extension-related software.

Extensions will test just how easy or how difficult it truly is to integrate "non-standard" data into effective XBRL creation, validation, presentation, calculation, and reference routines.

    4. What specific criteria should be applied to determine the adequacy of the standard taxonomies?

As we suggest above, the focus should probably not be on the adequacy of the standard taxonomies, but on the ease or difficulty of integrating extension elements into the standard XBRL components. Issuers, the SEC, software vendors, and analysts will all take from the Pilot views of how XBRL works with and integrates extensions. This experience will provide valuable guidance for any SEC XBRL initiative beyond the Pilot.

    5. Should we include other standard taxonomies in the voluntary program? If so, specify which ones and explain why you believe such taxonomies are sufficiently developed.

With the exception of the non-U.S. GAAP taxonomies referenced below in Question 6, we are not aware of additional taxonomies the SEC should consider for this Pilot.

    6. Should we allow foreign private issuers or foreign governments who use non-U.S. GAAP standard taxonomies to participate in the voluntary program? If so, how should this be implemented? What adaptations, if any, would be needed? How would U.S. GAAP reconciliations be handled in a voluntary XBRL submission?

For the moment, we suggest the SEC Pilot be limited to the U.S. GAAP taxonomies proposed. However, we also suggest the SEC consider expanding the pilot at a later date to include IFRS taxonomies as a means of encouraging non-U.S. issuers to participate.

    7. We plan to permit all filers to furnish XBRL data as an exhibit to Exchange Act and Investment Company Act filings so long as they use one of the specified standard taxonomies and form types. Should we further limit participation, such as by size or specific industry? Should we allow volunteers to furnish XBRL data with Securities Act filings?

The Consortium does not recommend limiting participation by company size or industry grouping.

    8. We have proposed that XBRL data furnished by volunteers must be the same financial information as in the corresponding portion of the HTML or ASCII version. Should we allow volunteers to present less detailed financial information in their XBRL data?

The question of "level of detail" in an XBRL instance document will ultimately be of great importance to the SEC should it elect to move forward with this technology. The Pilot is the best place for the SEC to determine its ability to validate that issuers are conforming to a requirement that there be conformance between the official statement and its associated XBRL instance.

Our recommendation is, of course, predicated on the Pilot allowing extensions. If the SEC pilot is structured on a fill-in template or requires issuers to stay within the confines of an existing taxonomy, this issue becomes moot.

    9. In order for the XBRL version of the financial statements to have the same level of detail as the HTML or ASCII version, we expect most companies would file extensions to the standard taxonomy. If you expect that companies would file extensions to the standard taxonomy, explain why extensions would be necessary. Would there be some companies that do not expect to file extensions? If not, explain why. Would the use of extensions harm the comparability that otherwise would exist among volunteers that use the same standard taxonomy?

It is our understanding that extensions will be necessary for all, or nearly all, participants. As we discuss in items 3 and 4 above, the success or failure of this experiment will depend, in very large measure, on the quality and reliability of software implementations associated with integrating such taxonomy extensions.

Leaving extensions out of the Pilot, would almost certainly lead the most sophisticated financial data analysts to find serious fault with the Pilot as failing to provide them with test data at an adequate level of detail.

    10. Are there any confidentiality concerns regarding submitting extensions? If so, what are they?

To the extent that extensions are nothing more than company-specific efforts to make a standard taxonomy conform to the issuer's accounting presentation needs, there should be no questions of confidentiality. XBRL extensions simply allow for an XBRL tag to be created that replicates an existing line item in a company's financial statement where no standard XBRL tag is available.

    11. We are contemplating allowing volunteers to submit XBRL data as an amendment to their filings or with a Form 8-K or Form 6-K that references the filing that contains the financial information to which the XBRL data relates. Should we require volunteers to submit XBRL data at the same time or within a specified number of days from the time they submit their official filing? Would this present difficulties for volunteers? Should we require volunteers to submit XBRL data only as an exhibit to the filing to which the XBRL data relates (i.e., remove the option to submit the XBRL data as an exhibit to an otherwise unrelated Form 8-K or Form 6-K)?

Allowing volunteers to submit their XBRL-tagged financial statements after they submit their official 10-Ks or 10-Qs will be of critical importance to securing Pilot volunteers. Pilot participants will have steep learning curves as well as a critical need to apply a wide assembly of corporate participants to any participation in the XBRL Pilot. It is essential the SEC recognize the need every participant will have for time to learn, understand, create, and review each instance document they prepare. Hence, allowing the XBRL instance to be filed via 8-K subsequent to the filing to which the data relate is essential.

    12. We plan to develop and provide via our website an application for a standard template to render the XBRL information in human readable form. What are the advantages and disadvantages of our requiring the use of such a standard template? For example, could a standard template prevent a volunteer from presenting its XBRL data in as much detail as, and in a manner substantially similar to, the financial statements in its official filing? Should we only develop standard templates for certain industries? Instead, should we allow each volunteer to submit its own template for rendering the XBRL data?

As we have suggested above, we believe a Pilot that is based on issuers' replicating in their XBRL instances the detail provided in their official statements is essential for a true test of the XBRL technology. It is also our understanding that almost every participant will be required to create some number of extensions in order for them to match the detail they provide in their official statements. Consequently, a "standard style sheet" would appear to be of no particular value.1

As important as it is to render XBRL into human-readable form, the SEC's pilot can proceed without such capability. The Pilot is primarily a test of XBRL document creation and XBRL-tagged data analysis. It is not essential that the instance document be converted to an easily-read format for this Pilot experiment.

We ask the SEC to be explicit in the final rule regarding it's decisions pertaining to the rendering and display of participant XBRL. If the SEC takes the position that display is not essential, we suggest the SEC permit participants to include a stylesheet for their particular submission should they wish to do so.

In the event the SEC elects to implement a Pilot that utilizes either a fixed reporting template or prohibits the use of extensions, then a standard template (or stylesheet) used for rendering XBRL instances readable makes sense. The advantages of such an approach to the Pilot are:

  • Effort and cost on the part of issuers would be significantly reduced

  • Effort and cost to the SEC would be reduced

  • The Pilot would almost certainly attract more volunteers than it otherwise would

  • The SEC and all other data users would have a much more uniform and consistent set of tagged data to experiment and work with. (The SEC and the analyst community would need to decide if the level of detail available were adequate.)

But again, limiting the Pilot to a fill-in template or captions already existing inside a taxonomy, will fail to test the most difficult component of XBRL – extensions.

A final thought regarding the usefulness of a restricted Pilot project is that such an implementation could always be the SEC's fallback position. In the event issuers and the SEC conclude from a test of the complete XBRL technology suite, that XBRL is either not ready, is too expensive, or is too complex for widespread, mandatory use at the conclusion of the Pilot, the SEC could almost certainly implement a template solution with relative speed and ease.

    13. As to the voluntary program, we propose to exclude XBRL-Related Documents from the certification requirements of Rules 13a-14 and 15d-14 under the Exchange Act and Rule 30a-2 under the Investment Company Act and we state that the XBRL-Related Documents should omit audit opinions and review reports. For purposes of the voluntary program, should officers of the company certify the XBRL data? If so, what should the certification criteria be? Should auditors be required to attest to the data? If so, what should their attestation requirements be? What are the advantages and disadvantages of requiring certification and attestation? What complications would arise if a volunteer presented an audit or review report in its XBRL-Related Documents?

The financial data on which the XBRL-tagged documents are based is already certified. Extending the certification requirement to the Pilot, XBRL instance will only serve to dissuade issuers from participating.

Please note, the Consortium of EDGAR Filing Agents and Software Developers has not provided responses to Questions 14 and 15.

    16. How should we determine how useful the tagged data is to users of the information?

The great promise of XBRL is that all users of SEC disclosure information will benefit from the existence of tagged data. Determining the extent to which XBRL's promise can be glimpsed from Pilot experiences will require the SEC to solicit extensively the views of all Pilot participants and their partners, but especially the views of the analyst, accounting and audit communities. Although the information presented to the SEC through such discussions will consist primarily of subjective judgments drawn from rather limited experience, it should be sufficient to accurately inform the SEC's assessment and evaluation processes.

A second component of the assessment process should be based on the SEC's own, internal use of the Pilot data submitted by participants. Internal efforts to analyze XBRL data will not only provide the SEC first-hand experience with using the data, but give the SEC a reference against which to judge the experiences related by others.

Given the variety of entities that will undoubtedly be anxious to test their ability to make effective use of the Pilot XBRL data, the SEC could assist the effort greatly if it adopted some means of making it easy to locate filings containing an XBRL Exhibit 100 document. Some commercial disseminators may well make the investment that would allow subscribers to select on "Exhibit 100" so as to produce an appropriate answer set, but others may not.

Since being able to locate XBRL instances will be so important to the Pilot, the SEC might consider modifying the form types referenced in the proposing release in a way that communicated the existence of an Exhibit 100 (e.g., 10-K/XB).

Alternatively, to avoid confusion with the rest of the documents of that specific form type, the SEC might consider adding to Form 8-K, Item 8.02, to be used solely for submitting an XBRL instance document.

A submission header element identifying the existence of an XBRL instance could also assist user efforts to locate pilot XBRL instances. One last suggestion in this regard: the SEC, having processed all incoming filings, might consider posting a spreadsheet of accession numbers of all submissions containing XBRL exhibits.

    17. What specific steps can we take to encourage registrants to participate in the voluntary program?

One thing the SEC might consider doing would be to staff a technical helpdesk that participants or potential pilot participants could call for advice and consultation. Because XBRL is so complex, it would be extremely helpful if there were SEC personnel with sufficient technical expertise to provide guidance, explanations or suggestions to all parties that will be involved with creating and filing XBRL instance documents.

Such a group not only helps filers, it has the added benefit of providing the SEC with continuous feedback during the Pilot as to exactly what issues, problems or concerns pilot participants are experiencing.

Comment on Concerns Associated with Item IV. D.

The Proposing Release at IV. D. – Receipt and Acceptance says that if SEC validation determines an XBRL instance fails validation, "…all XBRL-Related Documents would be removed and the official filing would be accepted and disseminated…"

We are concerned about the situation wherein an 8-K is comprised of nothing but the 8-K Cover page and an XBRL Exhibit 100. In such circumstances, if the exhibit is removed, the surviving, disseminated cover page will contain no useful information. It is our recommendation that in such circumstances, the SEC suspend the entire filing.

Additionally, we request the SEC make widely available whatever validating parser and customized SEC routines it uses to validate XBRL submissions. Having the parser (and its SEC custom additions) outside the EDGAR test system will improve the ability of all participants to submit documents that will pass SEC XBRL validation routines.

Again, thank you for the opportunity to respond to the rule proposal. As we have in the past, our consortium is prepared to assist the SEC with additional discussion or testing should the agency conclude it would be helpful.

Sincerely and on behalf of the Consortium of Filing Agents and Software Developers,

Richard K. Bonaparte


American Financial Printing
Bowne & Co., Inc.
Capital Printing
Command Financial Press
CT Corporation
Davis Polk & Wardwell
Merrill Corporation
Pacific Financial Printing
RR Donnelley
St Ives Burrups

1Our understanding is such that we do not believe it is possible to handle presentation in the manner suggested in SEC footnote 60: i.e. by combining a standard style sheet with additional Presentation Linkbase information.