With the second quarter Q season in full swing, many filers are noticing for the first time that the SEC has changed their XBRL viewer. Because the SEC released a new version of the viewer between major filing deadlines, the update happened without a lot of notice. In this article, we will cover the changes the SEC made to their viewer and discuss both the improvements that were made as well as some new problems filers might be encountering.
Monday, August 10. 2015
We will start with a little technical background. If you are only interested in what all this means to the end user, you can skip ahead.
The previous SEC viewer was built using Microsoft’s .NET framework, and the SEC’s XBRL validator (Arelle) was built using Python. While these applications were (and still are) open source, the different programming languages and platform dependencies made them both hard to integrate. Also, the .NET framework is only partially cross-platform, which means running the renderer on non-Windows machines was difficult.
The new viewer is also built using Python. This means it can integrate with Arelle easily. In fact, it relies on Arelle to do a lot of the XBRL heavy lifting. From an end user standpoint, none of this means much. From a developer standpoint, the viewer and the validator are now easier to maintain and integrate into other systems.
Now that the technical stuff is out of the way, we can get to what we really want to know. What has changed in the viewer, and why does my report suddenly have problems?
Improvement: The Rendering Log
When the viewer renders an XBRL document, it creates a log of the process. The previous version of the viewer had a rendering log that did not always have messages when items would not appear on presentations as expected or when other issues were encountered.
The new viewer’s rendering log is more detailed and usually contains information when something isn’t rendering the way you’d expect. For example, the old viewer would remove columns from certain presentations and the log would have a message stating a column was removed. The message that appears in the new viewer also states the reason the column was removed. That little change is extremely helpful.
The way the viewer displays units has been dramatically improved. When a fact used a non-monetary (or non-share-based) unit, the old viewer would put the unit name in the column heading. There was no indication of which fact in the column used that unit. Even the order in which the units appeared was inconsistent between columns. Overall, it was not a pretty rendering and the lack of clarity for units meant columns that had multiple unit types were not
The new viewer now marks the unit in the column heading as well as the in the row stub. This clearly marks what items are not shares/monetary and solves a problem that the previous viewer had.
The way the new viewer describes scaling is generally clearer as well. Instead of adding a blurb to the title of the presentation like “in thousands except per share data”, the new viewer will add “$ in thousands, $ / shares in Units” to inform users what the scaling is for other data types. This is much clearer although it doesn’t always display all the data types (for example, if a presentation has monetary, shares and per shares values within it). It also has some issues with certain presentations (more on this later).
Problem Area: Uncategorized Items
The old viewer usually only showed uncategorized facts if the facts could not be displayed on any presentation. The new viewer is far more aggressive when it comes to removing instant facts from presentations that have duration elements. This usually has to do with the Ending/Beginning label roles, but it can happen in other instances.
While checking more aggressively for the appropriate use of these label roles is a good idea overall, the implementation of this change is not in line with other filing changes from the SEC. Because this change affects the way preparers need to create their XBRL reports, it would have been wiser if the SEC changed the EDGAR Filer Manual first, then the XBRL validator, and lastly the viewer.
Doing so would have given filers advance notice of this change as well as documentation of the change. As it was, it appeared to many preparers who were unaware that the viewer was updated that their filings suddenly rendered differently for no reason.
Overcoming issues with uncategorized items
If your filing is suddenly rendering with uncategorized items, the first thing to do would be to check for overuse of the Ending and Beginning label roles. These label roles should only be employed on rollforward tables or on cash flow presentations. If you are using these label roles to alter a duration element to an instant element on other types of presentation, it’s very likely that your element choice is incorrect.
Problem Area: Total Columns
The old viewer only displayed a “Total” column if the presentation was an equity table. The new viewer displays a “Total” column whenever there are multiple contexts that end on the same date on the same presentation. While this change allows for tables other than the equity table to render with a “Total column”, it leads to problems when viewing reports.
For example, the Document and Entity Information presentation can now render with a “Total” column, or Operations tables can even render with two “Total” columns.
Overcoming issues with Total Columns
Unfortunately, there is nothing you can do as a preparer to overcome this problem, provided that you are using the appropriate contexts for each element on your presentations. It is unclear whether this is currently considered a bug by the SEC, but it really needs to be addressed on the SEC’s end.
Problem Area: Bugs
As with all software, there are bugs with the program. These should be less of an issue going forward as the SEC should fix them as time goes on, but they are listed here to provide a heads-up to filers.
1. Presentations that have no monetary facts display a weird value in the heading (“scaling factor is -92233720...”).
2. When a column is removed from a presentation due to lack of facts, the removed facts sometimes get put into the uncategorized section even if they are on another presentation.
As with the Total Columns problem, these issues need to be addressed on the SEC’s end. There is nothing that you can do to prevent these rendering errors, provided that your XBRL is prepared correctly. We would simply recommend that you use available validation tools such as those within GoFiler or other consistency checks to make sure that your XBRL has no errors within it before you preview it or file it to EDGAR.
If you have any questions or have an XBRL report that is not rendering as expected in the new SEC XBRL Viewer, please contact our support team for help. We’ll be happy to take a look at your XBRL and make suggestions so you can get the result you need.