Learn

The 45-day lag in 13F filings

Every 13F you read is a snapshot of what a hedge fund held six-to-eleven weeks ago. This is the single most important constraint to understand before using 13F data — and once you do, it stops being a limitation and starts being an asset.

The timeline

The SEC's Form 13F rule (§240.13f-1 of the Securities Exchange Act) gives institutional investment managers 45 calendar days after the end of each quarter to file. That means:

Quarter ends Mar 31 → filings due by May 15

Quarter ends Jun 30 → filings due by Aug 14

Quarter ends Sep 30 → filings due by Nov 14

Quarter ends Dec 31 → filings due by Feb 14

Most funds file on the last possible day — why disclose sooner than you have to? So when you read a Q1 filing on May 14, you're looking at positions as they were on March 31, 45 days ago, in a world where the market has moved 10-15% in either direction since.

Why the delay exists

It's not sloppy regulation. The 45-day window is a negotiated compromise with three constituencies:

The funds get time to accumulate or unwind positions without broadcasting their intentions real-time. If disclosures were instant, front-runners would trade against the fund as soon as a new position appeared.
The SEC gets structured disclosure of institutional exposure — the whole point of the Securities Exchange Act of 1934. Without 13Fs, nobody outside the fund would know what $100M+ of US equities is doing.
Retail investors get a lagged but honest view of smart-money positioning. Stale data is better than no data, if you know how to use it.

What the lag destroys (and what survives)

A lot of what retail investors try to do with 13Fs is actively destroyed by the lag. But not all of it.

What the lag destroys

Front-running. By the time you read that Ackman bought $420M of Chipotle, Ackman has either added more (cheap signal), trimmed (bad copy), or exited entirely (catastrophic copy). You cannot out-trade the man who already moved.
Swing trading. 13F data is useless for entries timed in days or weeks. It's calibrated for quarters or longer.
Catalyst-driven plays. If a drug trial reads out in April and Burry sold before it, you find out in mid-May — after the stock has already moved 40%.

What survives the lag

Conviction clustering. When 5-8 tracked managers all buy the same name in the same quarter, that's a strong signal regardless of 45-day lag. The consensus is the content, not the timing. See /consensus.
Long-horizon theses. Buffett holding Coca-Cola since 1988 doesn't care about 45 days. Patient positions show up in 13Fs just fine.
Manager fingerprinting. Watching what a specific manager does across many quarters reveals their style (concentration, sector tilt, turnover) — and style is stable. See any manager page.
Sector rotation. When the aggregate of tracked managers moves from financials to energy in a quarter, the direction of flow is still useful even at 45-day lag. See /rotation.
Idea generation. 13Fs are terrible trade signals and excellent research starting points. "Five smart people I respect are buying this — should I spend an hour understanding why?" is valid. "Five smart people bought this, I'll buy it too" is not.

How HoldLens handles the lag

We make the lag visible, not hidden:

Every move row shows the actual filing date inline, not just the "quarter" — so you can see how old the data is at a glance.
We render price delta since filing next to every move (e.g. "+18.3% since filing") — so you can see whether the market has already priced in the news. If the stock is up 20% since a superinvestor bought, your entry price is NOT their entry price.
Our ConvictionScore doesn't promise to predict next week. It tells you what the aggregate of tracked managers believes as of the most recent filings.
On every quarter digest page we show the next filing deadline prominently — so you know when the next wave of data arrives.

The honest take

13F data is 45 days lagged. That's a fact. Sites that try to sell you "real-time hedge fund signals" are either lying or reselling press-release coverage. The legitimate use of 13F data is long-horizon, pattern-based, and honest about its limits.

If you want to use it well: accept the lag, cluster the conviction, respect the concentration, and don't mechanically copy anyone.