Digital Financial Reporting

The document Core Financial Reporting Semantics points out the key financial reporting semantics which I see in a balance sheet, income statement, cash flow statement, and overview of information about how a company reports its financial information.

All of these core financial reporting semantics are supported, and in fact determined by, a set of 5525 SEC XBRL financial filings which I examined.  You can find the complete data set here.

Jean Louis Rodolpe Agassiz is attributed with the following quote:

The facts will eventually test all our theories, and they form, after all, the only impartial jury to which we can appeal.

The following is a summary of the core semantics which SEC XBRL filings show. SEC XBRL financial filings would find it hard to justify not following these core financial reporting semantics in my view:

• Balance sheets report assets.
• Balance sheets report liabilities and equity.
• Balance sheets report equity.
• Balance sheets balance.
• Cash flow statements report net cash flow.
• Net cash flow foots (perhaps in two ways, perhaps only in one).
• Income statements report net income (loss).
• OPTIONAL, certainly a best practice: Income statements report income (loss) from continuing operations before taxes.
• OPTIONAL, certain industries do not apply: Balance sheets report current assets and current liabilities.

Are their other rules? Likely.  Those 9 rules would be very, very hard to dispute.

While the US GAAP Taxonomy can be ambiguous about how this information is reported, it is the case, as proven by SEC XBRL financial filings themselves, that this information must be reported it seems to me.  The US GAAP taxonomy does have concepts for every one of these 9 rules.  As such, an SEC filer would be hard pressed to justify creating an extension concept for any of these items, it seems to me. 

I unofficially declare that using the information from the income statement of SEC XBRL financial filings maddening.

While it is true that I am not a professional programmer, in fact I am not very good and have been referred to as a "dinker"; I do understand some of the things which are helpful in using information and things which get in the way of using information.

Now, the balance sheet and cash flow statement are a cake walk. The reason is that there are totals and subtotals which you can use to check to be sure you are grabbing the right information on both the balance sheet and cash flow statement. The structure of balance sheets and cash flow statements between filings is very consistent. You can calculate or impute information and there is generally something to check what you impute against to be sure you are getting things right.

There is little of these types of things when it comes to the income statement.  What is more there are other things which just get in the way.  For example, I thought I was getting Net Income (Loss) correct because I was at least finding the numbers in the filing.  But what I realized is that there is duplication of concepts in the SEC XBRL financial filings.  Duplicate values for Net Income (Loss) exist if the filer has a noncontorlling interest. Seeing as 21% of all filers have a noncontrolling interest, I cannot be sure I am grabbing the correct Net Income (Loss) concept.  There is a big difference between getting some number and getting the right number.

It seems as though the details are there.  If the roll up of the income statement is correct and if the detail is there (it matches the HTML so it likely is), then you can get to the pieces you want.  But as far as I can see, this could be a task which needs to be done for each and every SEC filing to make comparisons work well and to be able to provide, say, an aggregate value for revenues, net income (loss) or other such reported facts across all SEC XBRL financial filings or some subset of filings.

The good news is that it does look like it is quite possible to use this information.

The income statement is a bear. The balance sheet and cash flow statement were rather straight forward as compared to the income statement.  Might need to take a different approach.

There are two primary reasons the income statement is a bear.  The first is that there really is no prescribed form or set of totals which one can latch onto.  For the balance sheet and cash flow statement there are "buckets" which everything fits into, totals are generally reported for these buckets or enough totals exist so that the totals (such as liabilities) which are not reported can be easily imputed (i.e. because equity is required to be reported and liabilities and equity is reported, finding total liabilities is a cake walk). The second issue is the number of permutations/combinations which are possible due to the existence or non-existence of equity method investments, taxes, discontinued operations, extraordinary items, noncontrolling interest, preferred shares. Picking the correct "subtotals" can be a challenge.

What I can say about the income statement for the 5525 filings in my test set is this:

• Net income (loss):  5390 of the set (98%) reported one of three concepts to express what amounts to Net Income (Loss): us-gaap:ProfitLoss, us-gaap:NetIncomeLoss, us-gaap:NetIncomeLossAvailableToCommonStockholdersBasic.  What I am saying is that the US GAAP term "Net Income (Loss)" is being reported using three DIFFERENT concepts.  Basically, there is ambiguity as to which of the three of those concepts to use in different reporting scenarios.  I think I have the right algorithm for detecting the correct concept but I cannot be sure as I have not tied out the income statement like I have tied out the upper level of the balance sheet which proves my balance sheet numbers are correct.
• Income (loss) from continuing operations before taxes: So far I have 3938 of the set (71%) who report what amounts to "Income (loss) from continuing operations before taxes". A wide variety of labels are used by filers for this concept, but the essence of what is being reported is precicely that: income or loss from continuing operations before taxes. Interestingly, lots of filers are creating an extension concept for this and even calling the concept exactly this.  There seems to be two reasons for this.  First, as I understand it in an earlier version of the US GAAP taxonomy did not have this concept, filers created an extension concept, and then they did not realize that a new concept has been added.  Second, the names of the concept are so strange and which concept is used is hard to figure out because of the different permutations/combinations (whether you do or do not have certain line items) that filers seem to simply give up and create an extension.  That is the sense I get from looking at filings.  Of the 30 companies which make up the Dow Industrials, 100% report this line item, 22 use the same concept, and 8 created an extension concept for what appears to be exactly the same line item.
• Revenues:  I was able to locate the line item(s) which make up revenues for 4624 (84%) of the filings, but that is not to say that revenues does not exist for others.  Understanding which concepts make up revenues is another story, this is likely the hardest task in reading an SEC XBRL financial filing. This is because there are so many different ways filers report revenues.  Some provide a helpful total for revenues, some do not.  There are all sorts of combinations so making sure you have total revenues can be a challenge, particularly if there is no total.  One company of the 30 companies which make up the Dow Industrials, Exxon Mobil, saw fit to create an extension concept for revenue. Yet Exxon Mobil did not create an extension for cost of revenues.
• Cost of revenues: Same deal as revenues, all over the board.
• Gross profit: While reporting gross profit is used by some industries, other industries do not report it. Of my set of filings 1846 (33%) do report gross profit. Having the concept gross profit can be quite helpful in figuring out which concepts make up revenues and cost of revenues.
• Operating income (loss): 3797 (69%) of filers report operating income (loss), all of these use the exact same concept.
• Nonoperating income (expenses): Did not look for this yet, I suspect that not a lot of filers provide this total and there is lots of variety in how the details are reported.
• Interest and debt expense: Did not look for this yet, again, I suspect a fair amount of variety here making this challenging to detect.
• Income (loss) from equity method investments: 877 (16%) of filers report equity method investment income. Clearly if a filer does not have equity method investments, which is often the case, this line item would not be reported.
• Income tax expense (benefit):  Gotta pay the tax man!  3732 (68%) do so, I guess the rest have moved off shore to avoid paying taxes (humor).
• Income (loss) from discontinued operations: I found income from discontinued operations for 3811 (69%) filings.  That could be off though, I was only looking for the total, might be higher. Not every filer has discontinued operations.
• Extraordinary items: Only 6 of the filings reported extraordinary items. As expected, this is very rare.
• Income (loss) from noncontrolling interest: 1139 (20%) of filers reported income or loss from a noncontrolling interest. Not every filer has a noncontrolling interest.
• Preferred stock dividends and other adjustments: Did not look at this yet.

I have not given up on being able to sort this out and automate the comparison of these high-level items between filers, clearly where it makes sense.  Comparing a retailer who uses gross profit to a bank who does not use gross profit can be done, it is just done at a different level of the income statement, say income (loss) from continuing operations before taxes, which they both will have.

Personally, what I would like to see done is certain line items should be reported whether you have them or not. What I am trying to achieve is two things: automatable comparison and/or aggregation across all filings and "ease".

For example, if a filer does not have any revenue, from my vantage point I feel that it is better to report revenue of zero than it is to drive software vendors nuts trying to write reliable algorithms to accurately compute say an aggregation of revenues across all SEC filings.  Same deal for a few other line items, maybe "income (loss) from continuing operations before taxes", "net income (loss)", perhaps a few others.

Trying to do this today is possible, but it will be brittle and fragile, not robust, reliable, accurate, and predictable which is what we need.

Filers are not giving up any flexibility to report their operations how they feel they should be reported within US GAAP.  All I am talking about here is a few subtotals, make sure they are there.  It is not disputable where income from equity investments, income from discontinued operations, income from noncontrolling interest, nonoperating income and expenses, revenues, gross profit, taxes, extraordinary items are reported.  That does not change.  All I am saying is that adding a few subtotals will make it easier for those trying to grab this data, expanding the set of those who can make use of this information.

From a creation perspective, my real interest really, these subtotals do exist in the taxonomy you are working with.  The creators of information know the relations which they want to express and can go into the taxonomy to find the pieces, leveraging these semantic relations which do exist.  That is not an issue.  But for those using the information, if they don't have something to grab ahold of, they will have to unravel the income statements of filers one-by-one.  This very possible, but time consuming, expensive, and brittle.

Continuing on my journey to discover the high level semantic model of financial statements, I dug a layer deeper into the cash flow statement.

As I pointed out, of 5525 filings, 5304 (96%) reported either the concept us-gaap:CashAndCashEquivalentsPeriodIncreaseDecrease or us-gaap:CashPeriodIncreaseDecrease. Why two different concepts?  Why would different concepts be needed for this fundamental piece of information? Personally, I see no need.  One concept such as "us-gaap:NetCashFlow" or something like that would be sufficient in my book. Another thing which is unclear is why a filer would need to extend that concept. A small number of them do.

Digging deeper, 4933 (89%) had the detailed concepts to arrive at what amounts to Net Cash Flow and the computation of the sum of those detailed concepts agreed with the total. Net Cash Flow is comprised of:

• Net cash flow from operating activities of continuing operations
• Net cash flow from financing activities
• Net cash flow from investing activities
• Net cash flow from discontinued operations
• Effect of exchange rate changes

Add these up, the agree to the total for 4933 in my test set. Now, filers can report the discontinuing operations in a couple of different ways but all that is taken into consideration in my computation.

There is something interesting which I would point out. Note the bullet for "Effect of exchange rate change" in the list.

A very small minority of filers, 118 (2%) do not include the "Effect of exchange rate changes" in Net Cash Flow; rather they put it in the roll forward computation: Beginning Cash + Net Cash Flow + Effect of exchange rate changes = Ending Cash.

In querying some accountants about this the bottom line is that there is no specific industry taking this approach and there seems to be no "business reason" or "unique reporting need" to take this approach. It is not illegal to do this per the financial reporting standards, it is certainly not best practice as most people take the approach to include this in Net Cash Flow.  This is an example of an ambiguity in the financial reporting standards.

On to the income statement....

I am now aware of three different sources of an automatically generated lists of the 30 companies which make up the Dow Industrials, each generated from using the SEC XBRL financial information reported by these companies:

• My list
• University of Texas (Dallas) student list
• PrimeAim list

My list and the University of Texas student's list agree, we reconciled them to make sure we got the numbers correct.  The list is as of a specific point in time.  It is automatically generated, but these two lists are not dynamic, meaning updated for new information which is received. Two different code bases were used to generate the information and the same numbers were derived, which is a good sign.

The PrimeAim list is dynamic.  A few of the companies which make up the Dow have submitted XBRL information to the SEC since my list and the University of Texas student created our list.  As such, they don't agree.

Clearly a dynamically updated list has its advantages.