Digital Financial Reporting

I am still experimenting, but what I have been able to do this far is to use the core financial semantics analysis set of 5525 SEC XBRL filings as a base, take the top 100 of those filings, and generate a reorganized version of that information which serves as a set of examples.

You can find these exemplars of the top 100 SEC XBRL financial filings here.

There are a number of different views, so let me explain them a bit:

• List of Companies: This is a list of the top 100 companies in terms of total assets reported. These 100 are a subset of the complete list of 5525 which you can find here. Note that this list does include filings which do not comply with the core financial reporting semantic model.
• Complete taxonomy: This provides a way to look at the complete taxonomy (the presentation relations) for each of those 100 filings.
• Organization of Pieces: This list takes the complete taxonomy, breaks it into individual pieces, and then lets you look at an individual piece by filer.  For example, on the top of that page you see "Document Information (by Filer)".  Click on that an a list of filers appears in the bottom area of the left hand pane.  Click on a company name and you can see the document information portion of the taxonomy for that specific filer.  Click on other filers and compare how each filer reports document information.  Click on other things such as balance sheet, income statement, income tax disclosure, etc. Nice way to compare.
• RDF List: My list is driven by this XML based list of companies, a pointer to their XBRL instance and more importantly a pre-processed XML infoset of the relations within their taxonomy.  This is imortant because a computer can understand this list and read it.
• Taxonomy Relations Infoset: This is a rendering of the XBRL taxonomy presentation relations for the taxonomy of the filing for a filer.  This rendering is not intended for humans, it is intended for computers.  It is not XBRL, rather it is an infoset expressed in an easier to understand and work with XML rendering of that information.  In this case it is for the first entry in that RDF file.

My HTML rendering is simply a recasting of the information from the SEC XBRL financial filings in the EDGAR database into a form which is useful for a specific purpose: to compare filings and provide examples. I am showing all these steps so that people can understand how the HTML is generated and how the information is derived.

More importantly, imagine a software application being able to read those XBRL files.  Imagine all this information being available within a software application for creating SEC XBRL financial filings.  Why would you need that? Why do people use the AICPA publication Accounting Trends and Techniques? Same reason.

But there is a big difference. The AICPA version is a survey of 600 companies.  My list is of 100, but that is only for my testing, I will expand this to all 5525 in my set.  The AICPA only does 600 because they create that publication manually.  My process will be fully automated.  The AICPA version is in the form of a book with a limited number of pages.  My version will have no limits as it exists on the Internet. I can provide much more information such as specializations by industry, specific accounting topics, etc.

The biggest difference is, however, is that my list will provide metadata for software applications. The AICPA version is for humans to read.

So basically, imagine Accounting Trends and Techniques tightly integrated into a software application used for creating financial statements.  Seems pretty useful to me.

What do you think of the idea? The entire SEC EDGAR XBRL financial filings database as a helpful resource for creating financial statements. Sort things by industry, by financial reporting topic, by whatever one might find useful.

I am updating my lists of exemplars, cycling back and making everything consistent. If you don't know what I am talking about when I say exemplar, search this blog site and/or read Modeling Business Information Using XBRL.

More to come.

So I am digging into income from continuing operations before income taxes and I am uncovering some additional interesting stuff. I decided to look at all the filings for one SIC code and picked this one 7011 Hotels and Motels to take a look at.  Here is a list of what I found. The two interesting columns are the label used by the filer and the concept used to report that line item.  Of 29 filers looked at, 12 created extension concepts and 17 used a US GAAP Taxonomy concept for what seems to be exactly the same thing. Here are the labels and extension concepts for the 12 which created extension concepts for this line item:

• Loss before income tax (benefit) provision | asca:IncomeLossFromContinuingOperationsBeforeIncomeTaxes
• Income (loss) from continuing operations | hot:IncomeFromContinuingOperationsBeforeTaxes
• Income before income taxes  | wynn:IncomeLossFromContinuingOperationsBeforeIncomeTaxesAndNoncontrollingInterest
• Income before income taxes  | chh:IncomeLossFromContinuingOperationsBeforeIncomeTaxes
• Income from operations before income taxes  | penn:IncomeLossFromContinuingOperationsBeforeIncomeTaxes
• Income before income taxes | rlh:IncomeLossBeforeTaxes
• Income before income taxes  |mar:IncomeLossFromContinuingOperationsBeforeIncomeTaxes
• Loss before income taxes  | mgm:IncomeLossFromContinuingOperationsAll
• Income before income taxes | h:IncomeLossBeforeIncomeTaxes
• Loss before income tax expense | mhgc:IncomeLossBeforeIncomeTaxes
• Income (loss) before income taxes and discontinued operations  | get:IncomelossBeforeIncomeTaxesAndDiscontinuedOperations
• Income before income taxes | lvs:IncomeLossFromContinuingOperationsBeforeIncomeTaxesAndMinorityInterest

Now here are the labels for SEC filers who did NOT create an extension concept, rather they use one of four US GAAP Taxonomy concepts:

• Income (loss) from continuing operations (HAS NO TAXES)
• Net earnings from continuing operations
• Net loss from continuing operations
• Income (loss) before income tax provision (benefit)
• Income (loss) before income taxes
• Income (loss) from continuing operations before income taxes
• Income (loss) from continuing operations before income taxes
• Income (loss) from continuing operations before income taxes
• Income before income taxes
• Income before income taxes
• Loss (income) before income taxes
• Loss before income taxes
• Loss before income taxes and discontinued operations
• Net income before taxes
• Earnings (loss) before income taxes
• Income (loss) before income taxes
• Loss before taxes

Not a lot of difference between the line items really. 

Some say "earnings" some "Net income" some "Loss" some "Net earnings" some "Income (Loss)" some "Net loss".  But, all those terms refer to the same thing in essence: Income (Loss). 

Some refer to continuing operations specifically, others say "before discontinued operations", others do not use the term continuing operations, but they all actually refer to: continuing operations.

Some say "taxes" some say "income taxes" some say "tax provision (benefit)".  Again, they mean: income tax (benefit).

Further, here is a link to the SEC Division of Corporate Finance, Financial Reporting Manual. I did a quick search on "income from continuing operations" and came up with 22 hits.  This is a screen shot of one:

I would also point out that the terms "revenues", "gross profit" and "net income" are referred to also. Core financial reporting semantics.

So what is the point? It seems to me that the 12 filers on this list would be hard pressed to justify creating an extension concept for what amounts to "income from continuing operations before taxes", a term which 17 others here did us, and over 70% of all filers report.  Further, because of the high percentage of inappropriate extension concepts I am finding, that percentage of 72% will likely get higher if those extenision concepts are replaced with the proper concepts.  Plus, the fact that the SEC seems to require this concept to be reporting is pointing to this concept being part of the core financial reporting semantics.

Initially I said that 5151 SEC XBRL financial filings had a "5 Star" rating. It was pointed out to me that I had changed the evaluation criteria, dropping XBRL Cloud's EFM validation as a constraint and thus minimizing EFM validation. This was not my intent.

My intent was to focus on the core financial semantics of the filings.

However, I do accept this criticism and have provided a revised set of filings which both support the core financial semantics which I am trying to point out, but also have no XBRL Cloud EFM validation errors reported.  This revised list can be found here. With this adjustment, the SEC XBRL financial filings which pass both criteria drops to 4875 filings which is 88% of all the filings analyzed.

This is still a good increase.

This list of 5151 SEC XBRL financial filings, 93% of the 5525 filings I analyzed, get 5 stars for meeting the core financial semantics constraints.

Granted, it is a small set of constraints, 9 in total, but it is a start.  The interesting thing to me is that I could do this at all.  The information on that rather large HTML page is extracted from SEC XBRL filings without even using an XBRL processor.

So, a beach head has been established. It seems to me that XBRL can in fact work to deliver on the promises made by its advocates (of which I am a member).  I propose that these 5151 filings and the 9 constraints offer clues as to how to get the complete information set for all filers working like a Swiss watch.

The list of 5 star filings has grown from one, to 24, to 92, to 5151. Granted, not a certifiable scientific analysis but positive signs none-the-less.