26% of rulebook changes are already live before you see them

A weekly review doesn't just run late on these. It misses them entirely, by design.

Across 37 exchanges and clearing houses, we looked at every notice with a stated effective date - 1,955 of them - and did nothing more than arithmetic on the exchanges' own publication and effective dates. No scoring, no assessment, just: when was it published, and when did it take effect. 511 notices were effective on or before the day they were published.

Why speed doesn't fix this

The instinct is to treat this as a latency problem: review more often, staff up, cut the turnaround time between publication and review. That instinct is wrong here. No review cadence - daily, hourly, even real-time reading of every circular the moment it lands - closes a gap that was never open. If a rule was effective before it was published, the exposure existed before the notice did. There was no window to review into.

This is the distinction between a rulebook change you missed and one that was never yours to catch on a schedule at all.

Why this is a trading question, not a compliance one

A backdated change to a margin requirement, a circuit breaker threshold, or an execution rule on a covered venue doesn't wait for anyone's review cycle. If the rule moved before it was published, any trading activity in that window happened against the new rule, whether or not the desk knew it existed. That's an ability-to-trade exposure - a question of whether the firm was operating within current rules - not a question of whether someone read a notice on time.

Retrospective monitoring can document that this happened after the fact. It can produce an audit trail. What it cannot do is prevent the exposure, because by construction the notice arrived after the fact.

What actually closes the gap

The only structural answer is continuous ingestion of exchange circulars as they publish, paired with immediate classification of what changed, which covered venues it touches, and who owns acting on it. Not a queue that waits for next week's review meeting - a system that assigns the change to a named owner the moment it's identified, so implementation starts before the backdating compounds the exposure further.

That's the practical difference between monitoring and control. Monitoring tells you, eventually, what happened. Control assigns ownership and tracks implementation from the moment a change is identified - including the moment it's identified as already having taken effect.

26% of changes arriving pre-empted isn't a reason to review faster. It's a reason to stop relying on review cadence as the control at all.