Somewhat Random but New Artifact Soon

Almost a year ago I described our date search capability (using the date search). The post was prompted by a question from one of our clients. Soon after it went live I received an email from Professor Matt Ege at TAMU who wondered about the misspelling of Arthur Andersen (I spelled ANDERSON). If you are a loyal follower you know I am grammatically challenged (their versus there). My excuse is I have one brain cell and as I am writing these posts I am thinking about something else I have to do. The truth of the matter is I misspelled their name but had results.

Fast forward to today. After the original request for strategies related to identifying auditors we received another set of related questions in January and then I again in March. At that point we had been working on a new tagging strategy for some time and it just seemed to me that we should tag the 10-K with the auditor.

This morning Manish Pokhrel (who is our data guru/manager) sent me a compilation of some examples he has found during this process.

Deloitte’s signature was in error in Dime Bancorp’s 10-K filed in 2015 (10-K Filing):

Another example of Arthur Andersen’s signature with an o from Green Mountain Power’s 2001 10-K (10-K):
KPMG had a name mishap in the 10-K filed by Heritage Financial Corporation in 2002 (10-K)

Just concentrating on the ANDERSON case – there are 511 10-K filings with that spelling (in many cases the spelling is not attached to the audit report)

All of that is interesting (maybe only to me). However, Manish has made significant progress on collecting and normalizing the name of the auditor for all of the 10-Ks. While we intend to tag this information into the 10-K as additional metadata – that is another step that will take some time.

To make this information available to you sooner we are going to create a new data artifact in the next couple of weeks – AUDITOR. We will have the balance sheet date, (CDATE), the auditor name, the date of the report and the office location (CITY/STATE (or COUNTRY for non-US offices). You will request it like all of our artifacts – CIK – DATE pairs in a request file. This is a small artifact but it is cumbersome to adjust size limits per-artifact so we will still have a 20,000 CIK-YEAR pair limit PER request file.

I will provide more details – but if you have a research question where you wanted to search for disclosures by auditor – the auditor data result file will merge trivially with the search result file because of all of the other metadata (CIK-RDATE-CDATE-FVAL). This is imperfect but the matching will be easy.

If not before I hope to push this out over Thanksgiving weekend.

Leave a Reply