Digital Financial Reporting

More and more scientific journals are "marking up" research, published articles, and other knowledge of domain using controlled vocabularies, taxonomies, and ontologies for that domain.

The accounting community should do the same.  I have no idea to what extent this is occurring, but the lack of a financial report ontology which is necessary to do this correctly leads me to believe that very little, if anything, has been done. Perhaps some journals are creating silos of information.

Biology and medicine seem to be leaders in marking up research and published articles and creating the needed underlying ontologies. This link takes you to the Open Biological and Biomedical Ontologies (OBO Foundry).

Microsoft released an add-in for Word back in 2008 which enables one to mark up documents in this way.  This video walks you through marking up documents, gives you an idea of how marking up documents works, etc.  It is worth watching to understand this opportunity.

This blog post provides additional information on this free and open source Microsoft Word Ontology Add-in. You can download this add-in here. (This is the home page of this Microsoft External Research project.)

The advantage of organizing accounting literature in this way should be apparent.

For more information about the possibilities here, watch the first video, Ontologies as a Branch of Philosophy, in this Introduction of Ontology.  In particular, pay attention between minutes 23:30 and about 35:00 in the video.

How about this bold undertaking which shows the possibilities here: The Encyclopedia of Life. This shows vision. Go to that encyclopedia and search on "Cats" and this is what is returned. Look through the tabs provided, particularly the "Resources" tab which takes you to other information about "cats".

Everyone knows how much fun accounting is. This sort of global accounting knowledgebase would go a long way toward improving financial literacy, making this complex domain just a little easier to navigate, and working in this profession just a little easier.

There is an interesting feature of the XBRL Cloud Evidence Package: you can embed the pieces into other HTML documents.  For example, this is an embedded: (not that great an example due to the limitations of the width of my blog)

Here is a link to an embedded balance sheet, income statement, and cash flow statement prototype. (This works the best using Google Chrome.)

Interesting! Note that you can click on the pieces and the details of the report elements and the facts will appear in a popup.

This seems incredibly useful.

In a prior blog postI mentioned an SEC XBRL financial filing comparison tool which I created. That tool was mostly manual in terms of selecting the entities to compare and it pointed to XML files on my computer.

I made some significant improvements to that comparison tool.  You can watch a video which shows the basic functionallity of this improved version of the tool here.

The primary improvement to the tool was an interface for selecting the companies to compare and rather than using local files for the comparison, I use commercially available information.  Basically, I rebuilt the back end of my tool to use the XBRL Cloud web service of SEC XBRL filing information.

This XBRL Cloud web service provides the list of entities, the submission information, and the information about the XBRL filing. All this information is organized by my, via my Excel application interface which you can see below:

Comparison tool selection form 

I take the information from the web service, load it into Excel, and then I generate the comparison using Excel.  You can see that below or download the resulting Excel spreadsheet here:

Comparison result 

Until now I have been focused my analysis of SEC XBRL financial filings on concept usage. I am now expanding that to have a closer look at how filers use [Axis] within SEC XBRL financial filings to express characteristics used to describe reported facts. For this blog post I am focusing on balance sheets.

This is a list of 82 SEC filers who each provide balance sheets. Now, because XBRL requires an entity, a period, and a concept; every filer has to provide at least these 3 characteristics.  The range of characteristics goes from 3 to 7.  Why?

First off, let me quickly explain the terminology which I am using.  (All my terms are defined on the Financial Report Ontology wiki.)

• Fact. A fact defines a single, observable, reportable piece of information contained within a financial report, or fact value, contextualized for unambiguous interpretation or analysis by one or more characteristics.
• Characteristic. A characteristic provides information necessary to describe a fact. A fact may have any number of characteristics.
• Axis. An axis is a means of providing information about the characteristics of a fact reported within a financial report. An [Axis] is a class of report element used by the US GAAP Taxonomy and therefore in SEC XBRL financial filings.  (The term used in XBRL for [Axis] is dimension which is defined by the XBRL Dimensions Specification as, "Each of the different aspects by which a fact may be characterized.")

So why the variation between what characteristics SEC XBRL filers report? Well, let's have a look.

As I said three characteristics are required: Entity, period, and concept.  XBRL requires this.  Entity, or more appropriately reporting entity or the CIK number of the filer is used on 82 out of 82 SEC XBRL financial filings.  Period or again more precisely calendar period is likewise used on all 82 filings I looked at. The concept is also required and used by all 82 filers for all reported facts. So, all 82 filings consistently use those characteristics on every balance sheet fact.

Focus on that statement for a moment, "every balance sheet fact." That is important to understand. Characteristics defined relate to every fact within that component.

A popular characteristic is "Legal Entity [Axis]", it is used by 20 of the 82 SEC filings I looked at?  why only 20, why not all 82? What is the legal entity of the other 62 reporting entities?  What does this mean, that the others don't have a legal entity?  No.  If a legal entity axis is not provided, then the legal entity of "consolidated entity" is assumed per SEC XBRL filing rules in the EFM. So whether you do, or do not, provide the legal entity [Axis], users of the information will either explicitly or implicitly understand which legal entity a filer is talking about.

Another popular [Axis] is the "Class of stock [Axis]" (us-gaap:StatementClassOfStockAxis). Of the 82, a total of 13 used that [Axis]. Remember what I said about the characteristic expressed by the [Axis] for a component (i.e. within a [Table]) being applicable for all the facts within that component/[Table]?  Well, in my view the way the class of stock [Axis] is being used is a mistake.  Cash and cash equivalents, total assets, long-term debt, and other stuff found on those balance sheets should not be characterized by the class of stock [Axis].  A breakdown of the details of some category of stock using this axis, sure.  But not the entire balance sheet.  This is a modeling mistake. Chrysler, which has capital units, did a similar thing as those who modeled class of stock, they just used Capital Units [Axis] (us-gaap:CapitalUnitsByClassAxis).

The next most popular [Axis] is the "Reporting Scenario [Axis]" (us-gaap:StatementScenarioAxis).  11 filers used this [Axis]. Interestingly, of the 11; all 11 used the member of the [Axis] (or the value of the characteristic) "scenario unspecified".  What is the reporting scenario which should be assumed per SEC EFM filing rules?  Nothing is said.  In my view, using this [Axis] is a modeling mistake.  It helps nothing.

Three filers used the "Variable Interest Entities, by Classification of Entity [Axis]" (us-gaap:VariableInterestEntitiesByClassificationOfEntityAxis).  This seems reasonable, however the name of the [Axis] is not reasonable.  The phrase "...by Classification of Entity..." is not necessary in the name of the [Axis].  That explains what the [Axis] does, not what the [Axis] is.  "Variable interest entity [Axis]" would be more appropriate in my view.  This is a common issue on [Axis] within the US GAAP Taxonomy.  But the [Axis] itself does seem reasonable.  What is assumed if it does not appear?  No variable interest entities exist which need to be reported. Comparability should be fine.

One filer, American Express, created an extension to define the axis "Error Correction And Prior Period Adjustments Restatement By Restatement Period And Amount [Axis]".  This is an error in my view.  First, for something as common as a restatement, no one should need to create an extension concept.  Second, the "thing that moves" in a restatement is the report date.  Therefore, the report date [Axis] is what should have been used.  The "Report Date [Axis]" already exists in the US GAAP Taxonomy. In my view, this is a modeling mistake.

Berkschire Hathaway created an extension axis "Financial Segments [Axis]" but does not define what a financial segment is as compared to a Business Segment [Axis], which does exist in the US GAAP Taxonomy. Not sure what to think about that.  I can sort of see the rational.  The information is comparable and properly modeled.  So I guess I would see no problem with this. But hold on a minute; in the business segment disclosure, they also used that same extension axis.  Nope, this is a modeling error.  Should have used the existing business segment axis.

Ford created an extension axis "Business Sectors [Axis]".  This seems to be the same deal as the extension created above.  Looking at the disclosures, business segment breakdown...Ford uses the same extension axis.  Yeah...no; this is a modeling error.  No need for that extension axis.  Besides, if you look at the balance sheet in the HTML document and compare that to the XBRL, I am not sure what the creators of this balance sheet were thinking.  I would avoid using the Ford balance sheet as an example.

Again, have a look of the 82 SEC filers which I took a look at for yourself.  I am going to do a more comprehensive analysis of [Axis] for the balance sheet and other components after all the 10Ks get filed.

The bottom line for me is that the SEC filers have a lot of adjustments which need to be made with their [Axis] use.  Use of the information by analysts is really the ultimate test.  As long as more information is provided and as long as how the [Axis] are used do not hinder analysis; then I think more detail really does not hurt anything.  But, using existing [Axis] is far better than creating extension report elements.