Notes ·

Field notes: the earthquake that triggered two contracts on two different clocks

A remote communications tower on a rocky, overcast mountain ridge.

A 6.1 hit at shallow depth, about forty kilometers from the nearest instrumented station with a long enough record to matter. Two parametric contracts sat over the same footprint. Neither counterparty was wrong, exactly. They were just running on different clocks.

Contract one was a single ground-motion station, PGA threshold, automatic trigger inside six hours of the event timestamp. Contract two referenced the regional seismological agency's bulletin magnitude, no station, no instrument, just whatever the agency published and however many times they revised it.

The station fired first. PGA cleared the attachment threshold at hour four, clean trigger, no ambiguity, payout notice went out before the first news wire even had a confirmed death toll. That contract's clock was fast and it was final. Good for the buyer, in this case, because the shaking at that one point happened to be severe even though the broader region came through lighter than the instrument suggested.

Contract two was still waiting on the agency. Initial bulletin had the quake at 6.3. Six hours later it dropped to 6.1 on a depth revision. At hour thirty, a third bulletin nudged it to 6.2 after they reprocessed with two more regional stations. That contract's trigger threshold was magnitude 6.2. So for a day and a half, the second contract was alternately triggered, not triggered, and triggered again, depending on which version of the bulletin you were reading when you checked.

This is the part of the job nobody puts in the slide deck. The basis risk conversation usually happens before the event, in the index design meeting, where everyone nods about single-station risk and bulletin revision risk and agrees the probability is low. Then the quake happens and you are on a call at 2 a.m. with a counterparty's claims team, and the question on the table is simple: did this event happen, in the way the contract defines "happen," or didn't it.

The station contract didn't have that problem because the instrument doesn't revise itself. It read what it read. The index contract had the problem because the index was built on an institution's editorial process, and editorial processes get updated as better data comes in. The agency's bulletin is a different kind of truth than a sensor on a pole, built to improve with revision rather than lock at first read.

A before/after multispectral pass over the affected districts settled the second contract, not the fourth bulletin revision. It showed a footprint matching what a 6.2, not a 6.1, would do to roof lines and road surfaces in that terrain. The pull came from outside both counterparties, timestamped against the event, and it didn't care which bulletin version was live when someone checked. It was the thing both sides could look at and stop arguing about which clock was the right clock. That's the role an independent read of the footprint itself is built to fill, available the moment the event qualifies, instead of a race to see whose instrument or whose bulletin gets cited at the negotiating table.

Station triggers and bulletin triggers solve different problems, and each carries its own failure mode. A station is fast and final but blind to everything outside its radius. A bulletin covers the whole region but moves on you for days after the ground stops shaking. Writing the next contract means adding a third observation that doesn't belong to either clock, something that reads the footprint once the dust settles and says, independently, what happened and where.

If your next trigger definition needs that kind of outside read built in before the next event, not after, it's worth having one lined up.

Start a pilot

Photo: Dženis Hasanica / Pexels