What to do when a parametric index and ground reports disagree
Every parametric program runs on the assumption that the index is a fair stand-in for the loss. Most of the time it is. Then a storm hits, the adjuster's notes describe downed structures and flooded fields, and the index comes back below trigger. Now you're holding an index mismatch and a counterparty who's pointing at the contract language, not the weather.
This happens more at the edges of a peril footprint than in the center. A station fifty kilometers from the worst damage reads a weaker signal. A grid cell that aggregates wind speed over too wide an area smooths out a localized spike. The index measured a different point, a different averaging window, or a different threshold than the one the ground crew was standing in.
Start with what the index measured, not what it was supposed to measure
Before you call it a dispute, pull the index methodology and figure out what observation it's built on. A rainfall index driven by a single gauge network has a different failure mode than a wind index built from reanalysis data. Ask:
- What's the spatial resolution of the underlying measurement, and where does your loss sit relative to the nearest data point?
- Is the index a point observation, an areal average, or a modeled reconstruction?
- When was the data captured relative to the event, and does that window cover the peak of the peril?
A lot of disputes die here because the gap turns out to be real. The index measured the wrong square kilometer, or it sampled a window that missed the peak, and that's basis risk doing what basis risk does: the ordinary cost of using a proxy instead of an on-site read. If the gap looks more like a data gap than an honest miss, you've got grounds to push.
Build a dispute file a counterparty can't wave off
A phone call saying "our ground team saw worse" doesn't move a settlement. What moves it is a record that's independent of both your field report and the counterparty's own index provider. The strongest version of that record is a before and after comparison of the footprint itself: what the ground looked like right before the event and what it looked like right after, pulled from a source neither party controls.
That's the piece most disputes are missing. Ground reports are useful but they're yours, so a skeptical counterparty can discount them as advocacy. The index is theirs, so you can't verify it without their cooperation. What settles the argument is a third read that doesn't belong to either side: a wide-swath scan of the affected area showing the footprint as it changed, timestamped to the event window. An independent before/after read of the peril footprint gives you something to put next to the index number instead of just a counter-claim.
Once you have that, the dispute file should walk a reviewer through three things in order: what the index showed, what the ground crew documented, and what the independent observation confirms about the extent of the event. If the independent read lines up with the ground reports and diverges from the index, you've isolated exactly where the index fell short, whether that's a coverage gap, a timing gap, or a resolution problem. If it lines up with the index instead, you've saved yourself from pushing a dispute that was never going to hold.
Submitting the dispute
Most parametric contracts have a defined window for contesting a trigger determination, often short, sometimes as little as a few weeks from the index publication date. Check that first. A dispute built on solid evidence still dies if it misses the contractual clock.
When you submit, lead with the observation gap, not with the dollar amount. A reviewer who sees "the index data point was 40 km from the loss and the independent scan shows the footprint extending past that point" is looking at a measurement question. A reviewer who sees "we think we're owed more" is looking at a negotiation. Frame it as the former even when the stakes are the latter.
When the index and the ground genuinely disagree for good reason
Sometimes the honest answer is that the contract's trigger definition was too narrow for the peril that occurred, and no amount of additional evidence changes the settlement for this cycle. That's a product design conversation for the next renewal, not a dispute you can win now. Worth noting in the file anyway, because it's the detail that gets the index definition rewritten before the next event makes the same gap again.
If you're heading into a dispute window and need a read that doesn't depend on either party's own data, that before/after scan is worth having in hand before the clock runs out.