How to Actually Read a 13F Filing

From raw document to insight

Most investors never look at an actual 13F filing — they rely on third parties to interpret the data. That's legitimate, but once you've walked through a filing yourself, you understand the data much better.

Step 1: Find the filing

Go to EDGAR (sec.gov/edgar), search for the fund name or CIK, and filter for "13F-HR". You usually see multiple entries per quarter — the HR filings (originals) and HR/A filings (amendments / corrections) are what matter.

Step 2: Understand the structure

A 13F consists of two parts:

• Cover Page: Manager identification, total value, position count
• Information Table: The actual holdings list as XML

Each row contains: issuer, CUSIP, security class, market value (in $1,000s!), share count, investment discretion, voting authority, and optionally put/call designation.

Step 3: Compare quarter-over-quarter

The real value emerges only in comparison. Load two consecutive filings, join them on CUSIP, and look for:

• New positions (present in Q2, missing in Q1)
• Complete sells (present in Q1, missing in Q2)
• Adds (same CUSIP, more shares)
• Trims (same CUSIP, fewer shares)

Step 4: Watch for stock splits

A common rookie mistake: with a 3:1 split, it looks like the manager tripled the position — when in reality only the share count increased mathematically. Always compare split-adjusted.

Step 5: Treat options separately

The "Put/Call" column indicates whether it's an equity or options position. Options should never be added to stock positions — they have a completely different risk profile.

Step 6: Add context

A naked filing says little. Ask yourself:

• How large is the position relative to the total portfolio?
• Is it a new buy or an add?
• Does it fit the manager's historical strategy?
• What did other top managers do in the same quarter?

Tools that help

Obviously, you don't have to do all this manually. Platforms like Whaleboard handle the heavy lifting — you can parse, compare, and visualize filings. But understanding the underlying data remains essential.