The Settlement Market.

Lending against a private fund is one of the larger credit markets almost no one can see into. Somewhere between 100 and 150 billion dollars of it sits on a handful of bank balance sheets, each loan backed by a fund's own statement of what it is worth, negotiated privately and held to maturity. There is no public price for the risk, no standard way to lay it off, and no way for anyone outside the two parties to a deal to take the other side. The instruments that turned private corporate loan risk into something traded in the open were not built for funds.

This is what we have built in their place. PARI, the Private Asset Risk Index, is a single published figure: the loss lenders actually realized on a cohort of similar facilities, struck by their own auditors rather than modeled. PARIX trades on it, priced the way credit is priced and moving with the losses the index records. A holder takes one of two sides, earning the yield those loans pay or buying coverage that pays out when losses climb, and either is held as easily as a stock. We publish the number and settle disputes over it; we run no market, hold no loans, and take no side. Everything past this page is what others build on the figure.

The two documents behind this one cover the rest. Document 01 sets out how the loss figure is produced and audited, and document 02 sets out the engine that reads the loss positioning each cohort inside its market band. This document takes both as given and describes the market they make possible.

PARI measures one thing per cohort per window: realized loss, the audited shortfall lenders took on that group of facilities divided by what they lent against it. The figure comes from the lenders' own auditors. The funds submit nothing, and nothing is forecast. Each cohort carries its own ticker, so PARI.PC.SR is the reading for US private credit, senior, and a dated series like PARIX.PC.SR.2026 is the contract that settles on a single year of it.

Only a realized credit loss counts: the fund cannot repay, the collateral is worked out below the loan, and the audited shortfall is what PARI records. A holder choosing to sell in the secondary at a discount does not count, nor does a soft writedown, nor anything that fails to breach the conforming line. Each of those is gameable, and an index that included them would drift from the losses lenders actually take. The line itself, the loss level where coverage attaches, is the subject of the next page.

The reading above is honest about being early. Realized loss sits at zero because the public NAV loans we can see are valued at or above cost and none has stopped paying, and the panel is labeled below the minimum constituent count the standard requires. It is the prototype the market grows out of, not yet a tradable index.

Every facility carries a loan to value, the loan over what the fund is worth, so 16 lent against 100 reads 16 percent. The advance rate is what NAV lenders actually do, a market practice band per strategy, driven by illiquidity and soft marks rather than by the realized loss PARI measures. PARI does not derive it. What PARI publishes is the realized loss, and that loss read positions a cohort inside its band, toward the high end when the loss history is clean and toward the low end when the reading is stressed. Senior secured direct lending reads stressed today, so it sits near 16 inside its 15 to 25 band. PARI records a loss only inside that line, and a lender who lent above it keeps the excess on its own book. The engine that reads the loss is document 02.

Different private credit strategies carry different risk and so trade around different market bands. The advance rate is market practice, not a number PARI derives; PARI publishes the realized loss and that read positions a cohort inside its band. Senior secured direct lending is the best supported, because it has the loss history in public filings that positions where inside its band it sits today, which reads stressed and so lands near 16 inside its 15 to 25 band. The others show market practice bands as references until each builds out its own verified loss history.

The exposure, the published loss on a cohort, trades as a dated series. Each one runs for a single period, say 2026 H1, with both sides posting the full amount up front in USDC, so there is no borrowing, no margin, and nothing to liquidate. The series settles once on that period's audited loss and then expires. A lender hedging a known book to a known date holds it to that date.

It does not run forever. A series has a set expiry, it pays out, and it ends. To stay on past that, a holder rolls into the next series by choice, the same way a dated index contract rolls. Nothing renews on its own and no position is carried with no settlement at the end.

Every series pays the same amount per unit against the published index. Neither side holds a custom payout negotiated for one lender, and that uniformity is exactly what lets them trade as instruments rather than sit on a balance sheet as one off private deals.

PARIX trades like any other instrument, and a holder takes one of two sides. Earn the yield and you own the credit, collecting the interest those loans pay and losing only when the cohort's audited losses rise; this is the side that sat behind limited partner status until now, and the one a retail buyer holds. Buy coverage and the position pays out when losses climb, the hedge a lender takes against its own book or the trade an investor puts on when private credit looks stressed. Both sides post their stake in digital dollars up front, so neither can lose more than it puts in.

The lender that buys coverage is also the natural source of the data. In exchange for cover on its own book, a fee, or a share of the data revenue, it agrees to report its audited loss every period, gain or loss, which stops a reporter from showing only its good quarters and keeps a real loss from going unrecorded. Where the lender is a public BDC its filings check the figure for free, and a determination rulebook governs what counts and how a disputed number is resolved.

A holder need not take the whole loss scale evenly. The scale runs from zero loss upward and divides into bands, and capital can sit anywhere along it. Spread thin across the full range, it earns the average yield and absorbs a little of the loss wherever it lands. Concentrated into one narrow band, it earns that band's full yield, which runs richer the nearer the band sits to where loss tends to arrive. The holder who buys the band the loss actually reaches is paid the most and takes the loss in that slice when it comes.

Reward tracks risk down the scale. A band close to where loss usually lands pays the most per dollar and is the most likely to be hit, while a remote band pays little and is rarely touched, and the holder picks where on that tradeoff to sit. The yield arrives up front in USDC, and a loss reaches the holder only if the published number climbs into the band it bought.

This is the instrument as a trader would meet it. A buyer who expects losses to stay low earns the yield; one who expects them to rise buys coverage; both read the same published number and settle on it. The screen looks like any other order book because that is what it is.

Illustrative screen, showing how the published number reads as a tradable instrument, with the two sides settling on the one figure underneath it.

A dated series settles once. When the audited loss for its period is published, the contract pays out in USDC and the position closes, the same amount per unit for every holder, with both sides paid from the collateral they posted up front. To stay on past that, a holder rolls into the next series. Neither the settlement nor the roll leaves room to relitigate the figure, because the oracle has already published it with its full envelope and a stated dispute route.

Because the instrument points at a broad index and pays the same public number to everyone, it reads as a contract on a public outcome rather than a private wager on one borrower or a regulated insurance policy. That breadth is what places it in the index derivative category rather than the single name one. The venue that lists it follows the registered market path, the route Polymarket took when it bought a registered exchange and clearinghouse for 112 million dollars and won approval as a regulated market in November 2025. The publisher stays separate from the venue, and the number stays free to read.

Reaching the economics of a private fund today means becoming a limited partner, an owner of record of a slice of the fund. That carries an accreditation test, a large minimum, a lockup measured in years, and the burden of holding and safekeeping the interest. Those rules attach to ownership of the fund and leave the position with no way to trade out.

The earn the yield side takes the credit without the slice. Every one of those constraints attaches to ownership of the fund, and this contract owns none of it: it tracks the published number and settles in USDC, with no claim on the fund. It is synthetic in the precise sense that it delivers the credit economics through a collateralized contract while granting no fund interest. The return is the carry on the credit risk the holder takes in the contract, not a distribution from the fund.

This is the part of the structure that reaches past fund ownership. The credit economics of private fund lending, normally reachable only by becoming a limited partner, become a collateralized contract on a public number that a holder can take either side of.

A stablecoin holds a fixed value, usually one dollar, and to do that it needs backing that holds its own value through stress. NAV loans fail that test on their own: hard to sell, hard to see into, with no public price and no quick exit in a panic. That is why no one backs a steady token with them today. A traded loss index changes the arithmetic.

Hold a pool of conforming NAV loans for the yield they pay, then pair it with coverage settled on the cohort's published loss, the same instrument this market already trades. When the pool takes losses the coverage pays, so the combined value of loans plus coverage stays bounded and the leftover yield funds the coin's payout. A pool that could fall in value becomes a position that does not.

This is exploratory, and it leans on three things holding: the coverage side funded up front, the audited number staying reliable, and the gap between the pool and the coverage staying narrow. We issue no coin and hold no backing. We supply the rulebook and the published number the coverage references, and others issue and run the rest.

We build the number and the rails to read it: the index engine, the reporter pipeline, the oracle and settlement service, and the publication API, free to read and licensed to settle. The trading venue, the order book, the funding engine, and the contracts are run and capitalized by others, and they reach the published number through thin adapters. Nothing we build requires holding a loan or taking a side, which is the entire point of sitting where we sit. A scorekeeper that places bets is no longer trusted to keep score.

The closest standardized products to ours point at the managers' corporate debt and stock, a proxy on the firms rather than the loss on the fund collateral. One index launched recently to bet against private credit, but it references that same corporate debt alongside a basket of banks, insurers, and REITs, with the fund managers only a thin slice of it. None of it settles on the realized loss of a NAV facility, and the rating agencies that come closest score deals one at a time when hired and keep their loss models private. There is no continuous, published, per cohort reading anywhere. PARI settles on the loss itself, and a party that holds the risk is not the one to publish that number impartially.

The two worlds meet at the figure rather than through a chain of intermediaries. A holder can take coverage or yield on chain and an institution takes the same coverage in a regulated wrapper, and both settle on the identical published PARI number. One reference, read the same way on either side.

Rated NAV loans corroborate the shape of the engine. One agency has rated 86 of these facilities from A to BBB minus, around 51 billion of issuance, and the loss level where senior deals break, near 55 percent loan to value, lands in the same neighborhood the engine draws its line. The existing market points the same direction we do.