A 13F tells you what the smart money owned — as of a date that is already at least 45 days in the past, and often closer to four months. The form is the backbone of institutional holdings data, filed every quarter by more than 5,000 managers, and it is structurally, permanently late. If you are building an AI agent to read institutional positioning, the fix is not to pretend the lag away or to buy a faster 13F — there is no faster 13F. It is to stop treating the quarterly snapshot as your only clock, and layer disclosures that arrive in days on top of the one that arrives in months.
TLDR:
- A 13F is due 45 calendar days after quarter-end; a position opened early in the quarter isn't public for ~135 days — and its new-buy alpha decays with a ~4-month half-life.
- Faster forms exist: Form 4 insider trades in 2 business days, 8-Ks in 4, Schedule 13D in 5 — and opportunistic insider buys carry real predictive value.
- An agent closes the gap by joining the slow 13F to these fast signals in one normalized layer. It narrows the lag; it does not erase it, and honest tools say so.
The 45-day rule is a floor, not the delay
Managers with at least $100 million in US equities must file Form 13F within 45 calendar days of each quarter's end — roughly February 14, May 15, August 14, and November 14. That 45 days is the part everyone quotes, but it is the smaller half of the delay. A 13F reports holdings as of the last day of the quarter, so the real lag depends on when in the quarter a position was actually established.
Buy a stock in the first week of a quarter and it sits invisibly for the ~85 remaining days of that quarter, then another 45 before the filing is due: about 130–135 days from trade to disclosure. A late-quarter buy surfaces in about 48. The 13F is also a long-only, quarter-end snapshot — no shorts, no intra-quarter round trips, no cash — so even on the day it lands, it describes a portfolio the manager may have already changed.
Compare that against the other things a company and its insiders are required to disclose, and the 13F is by far the slowest clock in the building — see Figure 1.
| Disclosure | What it reveals | Deadline | Effective lag (trade → public) |
|---|---|---|---|
| Form 4 | Insider (director/officer/>10%) trades | 2 business days | ~2 days |
| Form 8-K | Material corporate events | 4 business days | ~4 days |
| Schedule 13D | Activist >5% stake | 5 business days | ~1 week |
| Schedule 13G (QII) | Passive >5% stake | 45 days after quarter-end | up to ~135 days |
| Form 13F-HR | Long institutional holdings | 45 days after quarter-end | 45–135 days |
Sources: SEC FAQ About Form 13F; Form 4 deadline; Form 8-K; 13D/13G 2024 amendments.
Figure 1: Lag from trade to public disclosure, by form. Sources: SEC; Investor.gov; Skadden.
Why the lag actually costs you
Late data would be fine if the information kept. It doesn't. Research on institutional positioning finds that the incremental alpha on a manager's newly purchased stocks starts around 36 basis points in the first month and decays with a half-life of about four months. Put the two facts together and the problem is obvious: a mid-quarter buy first becomes visible in a 13F around month five — right as roughly half the edge has already bled out (Figure 2).
| Months after the buy | Approx. new-buy alpha | State when it's a mid-quarter 13F buy |
|---|---|---|
| Month 1 | ~36 bps | Not yet public |
| Month 3 | ~25 bps | Not yet public |
| Month 5 | ~18 bps | ~First visible in the 13F |
| Month 9 | ~9 bps | Fully public, mostly decayed |
Source: decay profile from the Novus / Barclays Public Ownership Index (36 bps month 1, ~4-month half-life); "first visible" timing derived from the 45-day rule.
Figure 2: The signal decays before the filing lands. Source: Novus / Barclays.
This is also why "just copy the 13F" backtests disappoint. The systematic alpha that survives the lag is concentrated in low-turnover, high-conviction managers whose theses play out over quarters, not days. For a high-turnover fund, the 13F is a photograph of a portfolio that no longer exists.
Closing the gap: layering faster signals onto institutional holdings data
The move is not to replace the 13F — it's the most complete public map of institutional ownership there is — but to date it correctly and surround it with disclosures that arrive while the information is still worth something. Two are especially useful because they are both fast and documented.
Form 4 insider trades. Directors, officers, and 10%-plus holders must report their own transactions within two business days — a deadline Sarbanes-Oxley cut from ten days precisely to close the information gap. And insider buying, filtered correctly, predicts returns. The cleanest result comes from Cohen, Malloy and Pomorski, who separated routine from opportunistic insiders and found routine trades carried near-zero signal while opportunistic purchases earned about 82 basis points per month, roughly 10% annualized. Earlier work by Lakonishok and Lee found insider purchases predict 4–8% abnormal returns over 6–12 months, concentrated in smaller firms.
Schedule 13D and 8-K. An activist crossing 5% must file a 13D in five business days (amendments in two); a material corporate event triggers an 8-K within four. Neither tells you what a fund's whole book looks like, but both tell you something changed long before the next 13F confirms it — which is exactly what you can read against a portfolio, as covered in monitoring 8-K events with agents.
| Signal | Lag | Documented edge |
|---|---|---|
| Opportunistic insider purchase (Form 4) | 2 business days | ~82 bps/month, ~10% annualized |
| Insider buys, smaller firms (Form 4) | 2 business days | ~4–8% over 6–12 months |
| Activist stake (Schedule 13D) | 5 business days | large announcement-window returns |
| 13F new-buy | 45–135 days | ~36 bps/month, ~4-month half-life |
Sources: Cohen, Malloy & Pomorski via Quant Decoded; Lakonishok & Lee; Novus / Barclays.
For an agent, "closing the gap" is a data-join problem. When a fund's 13F finally shows a new position, the agent asks: has an insider at that company been buying in the weeks since the quarter closed? Has a 13D or a material 8-K landed? Each fast signal either corroborates the stale snapshot or contradicts it, and a contradiction ("the fund bought, but three insiders have since sold") is often the more valuable read. The hard part is not the reasoning; it's having 13F holdings, Form 4 trades, and 8-K events sitting in the same normalized, entity-resolved, citation-backed data layer so the agent can join them by company in one step instead of reconciling three inconsistent scrapers.
The honest limit
None of this makes 13F data real-time, and any tool that implies otherwise is selling something. Insiders trade for reasons that have nothing to do with a thesis; 13Ds are rare; 8-Ks are noisy. Layering faster disclosures narrows the timeliness gap and adds independent corroboration — it does not turn a quarterly snapshot into a live position feed. The right goal for an agent is a better-dated, cross-checked view of institutional positioning, with every number carrying the filing and date it came from, so a human can see exactly how fresh — or stale — each piece of the picture really is.
FAQ
Why is 13F data always at least 45 days old?
Form 13F must be filed within 45 calendar days after the end of each calendar quarter, and it reports holdings as of the quarter's last day. Because a position may have been opened early in that quarter, the real lag from trade to disclosure runs from about 48 days up to roughly 135.
How stale is a 13F position by the time it's public?
For a position opened in the first weeks of a quarter, about 130–135 days. Since the new-buy alpha signal decays with a half-life of roughly four months, a mid-quarter buy has already lost around half its edge by the time it first appears in a filing.
What faster SEC filings can an agent use to supplement 13F data?
The most useful are Form 4 insider transactions (2 business days), Form 8-K material events (4 business days), and Schedule 13D activist stakes (5 business days). Each arrives while the information is still fresh and can corroborate or contradict a stale 13F snapshot.
Do insider purchases actually predict returns?
Filtered ones do. Opportunistic insider purchases earned about 82 basis points per month in Cohen, Malloy and Pomorski's work, while routine trades carried near-zero signal; Lakonishok and Lee found 4–8% abnormal returns over 6–12 months, concentrated in smaller firms. The filter matters more than the raw signal.
What is the best SEC filings API for combining 13F and insider data for AI agents?
The requirement is one normalized layer that carries 13F holdings, Form 4 insider trades, and 8-K events together, entity-resolved and cited back to each source filing so an agent can join them by company. FocusAlpha's SEC filings API is built for exactly this join; the practical test for any provider is whether every returned value links to the filing and date it came from.
What is FocusAlpha?
FocusAlpha is a SEC filings API and agent-ready financial data layer: it turns SEC filings (10-K, 10-Q, 8-K, 13F), earnings-call transcripts, and other trusted company communications into structured, normalized data where every value keeps its citation back to the source document. AI agents connect via API or MCP to research public companies from complete, trusted information.