Public financial data is not a live feed. It is a series of documents published on deadlines, and every deadline puts distance between an event and the moment anyone outside the company can see it. That distance is the disclosure lag.
The lag is a rule, not a failure
Nobody is hiding anything when a fund's holdings arrive six weeks after the quarter ends. The reporting regime says forty-five days, and forty-five days is what you get. The same is true of every other form:
- Insider transactions are normally reported within two business days of the trade.
- Some insider transactions are exempt from that and only appear in an annual catch-up filing, up to a year later.
- Institutional holdings arrive within forty-five days of the quarter end they describe.
- Ownership disclosures by large shareholders run on their own clocks, and the lighter form has the longer one.
- A sanctions list is exactly as current as its last publication, which may be days old.
Why one date is not enough
A page showing only the publication date invites you to read an old fact as new. A page showing only the event date invites you to think the information was available at the time. Both are wrong in the same way: they hide the age of the fact.
That is why every record here carries both dates, and why the number of days between them is computed and shown. The lag is not metadata; for a lot of questions it is the answer.
What the lag does to comparisons
The moment you compare two things with different lags, the comparison acquires a time distortion that has nothing to do with the subject.
A fund's reported position and an insider's reported trade in the same company describe moments that can be six weeks apart. A "current" holder list built from quarterly reports is a photograph of a date already past. A sanctions screening result reflects the list version you searched, not the authority's live register.
None of this makes the data less useful. It makes it data about specific dates, which is what it always was.
What we do about it
We store the event date and the publication date separately for every record, we never substitute one for the other, and we never carry a value forward to fill a gap. Where a source publishes nothing for a date, the answer is that nothing was published — not the previous value repeated.