Customizable industrial taxonomies

The subtleties of XBRL are lost in discussions where some vendors are swaying the "uneducated" to lean on the SEC to prescribe fixed taxonomies. This is unrealistic and impossible to conceive given the market driven economies embraced globally with very few exceptions (Cuba?, Gulf countries?) and the power of XBRL to process and compare within these natural variances (extension taxonomies) from one company to the next.

However, it is worth referring to the cry for standardization:

In response to the Friday March 31st posting entitled "XBRL Update" on the AAO Weblog (http://www.accountingobserver.com/blog/), Eric Linder, CFA, replied:

--------------------------------------------------------------------------------
Jack,

I was forwarded your blog entry about software to read XBRL files (http://www.accountingobserver.com/blog/) from several people as our company, SavaNet, you may know is the only company providing an XBRL analysis application, called the SavaNet XBRL Reader, which is free and available now from www.savanet.net. It is actually even more than a "Reader" application because it also performs professional-level security analysis on information in XBRL format. Although there is a large library of over 100 available Form 10-Ks in XBRL which can be accessed through the Reader's file manager, it doesn't support the filings made to the SEC under its voluntary reporting program. As a former Wall Street equity analyst, XBRL International member and leading XBRL software developer, I can tell you exactly what is going on here:

The problem here is that the XBRL documents filed with the SEC use company-specific XBRL taxonomies which do not allow for the automated processing and comparable analysis that has been promised to the marketplace by XBRL. In essence, companies are creating their own report form and then filling it out, which, as any financial analyst (but not accountant) will tell you, eliminates the ability to automatically process and compare the information because much of the information could be tagged differently by different companies. Even though companies all start with the same base industrial taxonomies in their filings under the voluntary trial program, they have the unlimited ability to add new items and re-do the calculation relationships of existing items, which essentially changes their definition and makes them unusable for analysis.

I often find myself explaining to the non-analysts involved in the XBRL effort that the moment even one item is added to a base taxonomy statement it invalidates most of the rest of the statement for automatic processing and analysis because, in most cases, without manually reviewing the new item, one cannot automatically tell where this new amount was taken out of the existing taxonomy. Many do not understand this conservation rule of "financial physics": since all financial statements add up to a total, something added in one place is necessarily taken out of the location where it was expected. And since this location is unknown, proper financial analysis requires that you do not trust any of the information in statements which is equal to or below the calculation level of the extension or alteration made by a company while performing automated analysis. (Note: since most compaies reported statements are very "flat" these corporate extensions or alterations are usually done at the highest levels of calculation relationships which thus invalidates most of the statement.) So, with company-specific taxonomies that alter or extend base industrial taxonomies, analysts need to go back to manual processing and the majority of the benefits of XBRL are lost.

So the SavaNet XBRL Reader doesn't support the unrestricted company-specific taxonomies used by companies in the voluntary reporting program because these files cannot and should not be used for automated financial analysis, comparison and valuation purposes. (and if all you want to do is view the statements, you can get them off EDGAR in prettier html format). The REAL risk to the financial community is actually that some users (or companies) that don't understand financial analysis DO attempt to perform financial analysis on these files because such an application could only practicably use the items from a base taxonomy for its ratio and valuation analysis, without taking extensions into account, which will lead to erroneous results in many cases.

Most non-analyst and non-accountants don't understand that even though, for example, the Operating Income and Revenue line item elements may still be reported using a company-created taxonomy, that the analysis ratio of Operating Margin (defined as Operating Income / Revenue) should not be calculated if there have been extensions or calculation alterations. This is because companies may have moved items in the calculation relationship which change the definition of Operating Income element and/or, if the application tries to adjust for this by referring to the specific elements that they believe should go into their definition of operating income instead, they will still likely receive an erroneous result because the company could have added extension elements that are not in this formula. You can immediately see the problem for erroneous analysis by the unaware and the opportunity for "gaming the system" by companies who can "create their own" uncategorized extensions to hold their undesirables - making them somewhat invisible to the automated analysis that will inevitability result from the use of XBRL.

Luckily, there is a XBRL implementation method that gets everyone (investors, analysts, companies and the SEC) what they need-- and that is called "customizable industrial taxonomies". Under a customizable industrial taxonomy XBRL implementation, all companies in an industry use the same hyper-detailed taxonomy (up to 3,000 elements in all statements and notes) without extension BUT companies can completely alter the presentation of these items and over-write labels to exactly re-create the current presentation of their As Reported set of financial statements. (Actually, company extensions ARE allowable IF they fall outside of the taxonomy calculation relationships, such as for company-specific notes.) So, the financial statements that investors and analysts see using a customizable taxonomy solution appear exactly as the company desires, but the underlying elements are structured for accurate, hyper-detailed analysis and comparison.

So, anyone who is reading this, here is what you need to know about XBRL: XBRL has absolutely enormous potential to solve a great deal of the reporting issues that investors and financial analysts face today and really can be a once-in-a-career advancement in analyst tools, BUT only if it properly implemented using taxonomies with restricted extensions, such as in a "customizable industrial taxonomy" solution. If extensions and calculation relationship over-rides are not restricted, XBRL will provide little to no value to investors and analysts. But even worse than no benefit is, if companies are allowed to extend taxonomies without restriction, XBRL could, in some respects, even make matters worse for investors who may BELIEVE they are getting an accuracy and reliability they are not, and then rely on erroneous automated analysis, and/or companies make use of uncategorized (or generally categorized) extensions to even further game the system.