The Basic Comparison, SEC XBRL Filing exampleshows issues relating to comparing SEC XBRL filings. If you understand how to read the XBRL instances and XBRL taxonomies or if you dig into the Measure Relations and Fact Groups (Fact Tables), comparability issues jump out at you.
Now granted, the issues will be better seen within a software application actually trying to do a comparison; but I cannot show such an application currently.
Don't be fooled by the simplicity of this example. It is simple, but it is not simplistic. The information in the XBRL taxonomy and XBRL instance is concise, but it serves its purpose.
For example, imagine that you want to compare the accounting policies of these three basic SEC XBRL type filings. How would you do this? First off, you might try and find the extended link which contains the accounting policies, using that to grab all the accounting policies. Well, that will not work. EFM rules state that company extensions cannot use extended links from the US GAAP Taxonomy. There is an extended link, but each filer needs to create their own extended link. Look closely at the Measure Relations of the ABC Company, XYZ Company, and QQQ Company and you will see what I mean. So, you cannot use extended links for comparisons at all.
OK, the next level down in the taxonomy after the extended link is the [Table]. Just grab the Accounting Policies [Table], use that as the basis for comparison. This is a great idea, but that will not work either. The reason is that there is no Accounting Policies [Table] in the US GAAP Taxonomy. Now, I created one in these three example filings, but although they have the same name, they are from different namespaces (abc:, xyz:, qqq:); they are not the same and will not be seen as the same by software applications. Another way to understand this is to look at the Nonmonetary Transactions, by Type [Table]. Now that [Table] is comparable as (a) it does exist in the US GAAP Taxonomy, and (b) each filer used that [Table]. A rather simple program can be written to get that [Table], all the concepts in that [Table] (as that is clearly defined in the definition linkbase) and you have your comparison.
But you don't have that [Table] to use. You may think that you do because in this example you could ignore the namespace prefix but then you would be missing the point here. Although I named the elements the same in each taxonomy, there is nothing which will ensure that SEC XBRL filers will do this in their filings. So, perhaps there is some concept that everyone uses in their taxonomy, use that and the presentation relations to find all the accounting policies. If you go look at real SEC XBRL filings you will see that that will not work.
The bottom line here is that there is no dependable way to do a comparison of accounting policies. You could spend time and money creating a pretty good software algorithm, but why not make the problem go away so you don't have to spend the time and money to write that algorithm. Just add an Accounting Policies [Table], require that everyone use that [Table], then compatibility can be easy. Same deal for other portions of the taxonomy. This is not to say that everything needs to be required, but base areas such as the accounting policies seem to be something everyone should really have. Same for the balance sheet, income statement, cash flow statement, other specific disclosures, and other areas. What might be contained in those [Table]s may be different, but the place to look in each SEC XBRL filing would be the same. How else would comparability be achieved?
This is just one example, there are many other clues in this basic example. As is said, the proof will be in the pudding: in actual and useful comparability. Where the comparability will exist is a decision for the financial reporting domain. But clearly there should be some comparability at some level.
Have a look at the Basic Comparison example and see what you come up with. Or even better, load these three simple SEC XBRL filings into your favorite comparison tool and see what you come up with.