Collateral eligibility & haircut engine
Tokenized money-market funds are becoming collateral. A data vendor tells you what exists. This does a collateral desk's job: it decides what is eligible and at what haircut, calibrated against what the market already charges to lend against these tokens, then sized for stress. The output is an eligibility schedule, not a catalog. Every figure is live and sourced: onchain reads for supply, DeFiLlama for price, the Morpho API for the LLTV that sets the market haircut.
The deliverable: an eligibility schedule with an assigned haircut per fund, the output a triparty desk produces. The assigned haircut starts from the live market benchmark and is sized up for stress. Move the sliders to re-run it; the schedule and the verdicts update live.
| Fund | Score | Market haircut | Assigned haircut | Verdict |
|---|---|---|---|---|
| BUIDL | 78 | 14.0%on | 4.2% | Eligible |
| USYC | 67 | 14.0%on | 4.8% | Eligible with conditions |
| USDY | 59 | 14.0%on | 6.3% | Eligible with conditions |
| OUSG | 73 | - | 4.7% | Eligible |
| USTB | 74 | 14.0%on | 4.7% | Eligible |
A note on honesty. An LLTV is an immutable onchain parameter, so the implied haircut it yields is verifiable, not a guess. The underlying of a tokenized fund is not onchain, so its composition is issuer-declared and marked as such. The stress parameters are analyst assumptions, shown on the surface so anyone can challenge them. Permissioned registries are not readable here, by design, and that boundary is itself part of the read.