The core source for institutional ownership data is derived from form 13F, which is a form filed by large asset managers that manage over $100mm in 13(f) securities (e.g., stocks). Here is a link to the 13(f) security lists.
Many academic studies and marketplace products/strategies rely on 13F filing to establish measures of so-called “institutional ownership.” The research that uses institutional ownership metrics to test an argument is interesting and often compelling, however, few readers ever question the data source that underlies “institutional ownership.” Could there be a case of “garbage in, garbage out” when it comes to this data?
My gut has always said, “Yes.” As a PhD grad student and a former finance professor, I used to be buried in the Thomson-Reuters Institutional Holdings (13F) Database, which is a collection of 13F files for all filers over time. The database is not a pretty sight–data errors are not rare, they are the norm. However, there are two possible reasons for the database issues:
- Thomson-Reuters is doing a bad job
- The underlying source data is no good
13F data stinks?
Turns out #2 might be the culprit. A new paper by Anne Anderson and Paul Brockman, appropriately titled, “Form 13F (Mis)Filings,” highlights the issues with the source 13F filing data submitted by 13F filers. The abstract says it all:
We examine the reliability of Form 13F filings and document the widespread presence of significant reporting errors…Overall, our evidence shows that the widespread reliance on 13F filings for institutional ownership figures is unwarranted.
Ouch. That is quite an indictment!
Here is a table from the paper highlighting the vast differences in shares reported on proxy statements (DEF 14a) and 13F–note the huge differences. Perhaps some of these can be easily explained, but the extraordinary differences are a bit startling…
So what’s the bottomline?
Perhaps we should follow the SEC’s own advice as it pertains to 13F filings:
The reader should not assume that information is accurate and complete.
Readers should make sure they understand their data provider who providers 13F data and also recognize that the underlying data provided to the SEC is not held to the same level of scrutiny as other filings (e.g., financial statement data).
- The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
- Join thousands of other readers and subscribe to our blog.
- This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.