Rented liquidity,
measured directly.
How much of a pool's liquidity is only there for the subsidy, estimated from how deposits actually behaved the last time the reward rate moved.
What TVL cannot tell you.
Two pools report the same number. One grew on its own, the other is renting every dollar of it. Nothing an on-chain consumer reads separates them.
A campaign expiry date is public, but a date is not a durability estimate.
Capital enters faster than it leaves, so one elasticity figure is the wrong number.
New pools have no post-campaign history, which is the one feature that would settle it outright.
How this works
Emission rates change discretely and observably. Every rate change is a dated natural experiment in how much of the deposit base is actually attached to the subsidy.
01
Finds the rate changes
Every emission step, decay point, and campaign boundary is a dated, public observation.
02
Controls for the market
Aggregate stablecoin supply, chain TVL, and realised volatility are controlled for, so the estimate is pool-specific rather than beta to the broader market.
03
Separates up from down
Only the downward coefficient enters the hazard. Ignoring that asymmetry systematically overstates outflow speed, which is the failure mode most likely to make an oracle loudly and repeatedly alarmist.
What it reads, derives,
and does with it.
What it reads
Every observable rate change.
What it derives
The coefficients it fits per pool.
What it does with it
Where each coefficient lands.
“The scoring model is publishable and copyable, but the accuracy record is not. A competitor who reimplements the ensemble from this paper starts with an identical model and a calibration history of zero.”
Thirty-six surfaces across contract, API, agent, and automation
Durability attestation for autonomous capital. Read the horizon before the capital moves, not after.