13F Institutional Holdings History Database Data: Dataset Structure and Field Coverage

Form 13F is the SEC's quarterly disclosure for institutional investment managers with discretion over $100 million or more in US-traded equities: once a quarter, each manager files every long position it held at quarter end. A 13F institutional holdings history database takes those filings and rebuilds them as one row per filer-security-quarter position, so a researcher can ask not just "who owns Apple today" but "how did this fund's Apple stake move across eighty quarters." That structure is exactly how Datadory catalogs WhaleWisdom 13F Aggregator: all SEC Form 13F filings since Q1 2001, restructured into searchable positions, with Schedule 13D/G filings and insider transactions layered on from 2006 onward.

Coverage depends entirely on which source you pick, and the gaps are structural:

1. Q1 2001 to the latest filed quarter - WhaleWisdom 13F Aggregator's 13F position history. Internally each filing date carries a numeric quarter id (id 1 = 3/31/2001, id 40 = 12/31/2010), and the quarters command lists every date your account can reach. Free accounts see only the past two years - roughly eight completed quarters - and never the current quarter.
2. 2006 onward for event-driven ownership - Schedule 13D/G filings and insider transactions start there in the same account, so activist stakes and 5% holder changes sit beside the quarterly snapshots.
3. 2019 Q4 through 2026 Q2 - the SEC Division of Economic and Risk Analysis harvest of Form N-PORT data sets indexed by the Data.gov Securities Portal: 27 quarterly ZIP resources, downloadable directly from sec.gov. N-PORT replaced much of the analytical detail 13F omits, but only from late 2019.

The documented API surface is deliberately small - four commands cover the whole lookup-to-positions chain:

• quarters lists every filing date in the database and which your account can access.
• filer_lookup resolves a filer id from a partial name or CIK.
• stock_lookup resolves an internal stock id from a ticker or company name and returns ACTIVE or DELISTED status.
• holdings returns the positions themselves.

A concrete first call looks like /shell/command.json?args={"command":"stock_lookup","symbol":"aapl"}, which returns the stock_id that holdings queries need. Requests authenticate with a session cookie or an HMAC-SHA1 signature computed over the args object, a newline and an ISO timestamp, capped at 20 requests per minute, and output switches between JSON and CSV by changing the request extension. Formats listed for the record run json, csv, html and xlsx via the Excel add-in.

One verification caveat worth quoting: during checks on August 21, 2026, the unauthenticated /filer/holdings page returned only {"total":0,"records":0,"page":1,"rows":[]} because the site renders holding tables client-side. The full holdings-response field list - position values, share counts, portfolio percentages - was not observable without credentials, so confirm it against your own account before committing a build.

Datadory ranks investors & quant researchers first among personas for this record at relevance 3 of 3, with the stated use case of tracking hedge-fund position changes quarter over quarter and screening for crowding across 13F filers. Data scientists & ML engineers score the same 3/3, pulling filer-security-quarter rows back to 2001 as ownership-flow features through the documented JSON API. Market researchers & consultants (2/3) summarize institutional ownership shifts into client-facing research, and competitive intelligence & product teams (1/3) check which funds sit behind rivals.