How a market gets listed.
Markets here are manufactured, not created. Nobody types a question into a box. A benchmark is registered once, under a fixed checklist, and windows are generated from it on a calendar. This page is the checklist.
What a benchmark must have.
A number qualifies for listing only if every item below holds. Any failure disqualifies it, whatever the demand.
- An independent administrator publishes it on a stated schedule under a public methodology, and nobody in this market can influence the print.
- A definite observation: one closing level per publication day, with the administrator’s own correction rules.
- Enough public history to calibrate a band from real data, through at least one stress episode.
- A related professional market where the exposure can be hedged, so market makers can quote without holding a view.
- A written fallback ladder covering late publication, corrections, methodology changes, and discontinuation, agreed before listing.
How the band is set.
The floor and cap are outputs of a study, not judgment calls. The rule is mechanical and published with the listing.
- Measure the settlement number over every rolling window of the contract length in the available history, through at least one crisis.
- Place the floor below the quiet drift of the benchmark, wide enough that ordinary windows do not pin at the floor. On a benchmark that drifts negative in calm periods, the floor sits well below zero: our first market’s study showed a floor at zero would have pinned four windows in five.
- Place the cap above the worst outcome on record, so the market keeps pricing power even in a window worse than anything on record.
- Publish the band, the study, and the data sources, then freeze all of it when the window opens.
How settlement resolves, including bad days.
Ordinary settlement is arithmetic on two published prints, averaged over five days at each end of the window. The edge cases follow a ladder written before trading, in order, with no step skippable:
- Late publication: settlement waits a stated number of days for the print. The deadline is in the listing terms.
- Corrections: the administrator’s own correction policy applies, once, within its stated correction window. Revisions after that are not accounted for.
- Methodology change or successor index: the listing names what counts as a successor. If the administrator publishes a continuation, the continuation settles the market.
- Discontinuation with no successor: the listing names the alternate calculation and its data sources, fixed at listing time.
- Nothing on the ladder resolves: the escrow returns to holders in proportion to holdings at the last published observation. There is no discretionary payout and no midpoint award, at any step.
What a listing publishes before trading opens.
- The benchmark, its administrator, and the exact methodology version settlement will read.
- The window’s open date, settlement date, and the five day fixing convention at both ends.
- The floor, the cap, and the calibration study behind them.
- The full fallback ladder above.
- The market maker obligations attached to the listing: committed quoting is a launch condition, not a hope.
What never happens.
- No free text market creation, by anyone, including us.
- No second market on the same benchmark and window: one official pair, one book.
- No parameter changes after a window opens: not the band, not the source, not the dates.
- No vote, committee, or discretionary outcome anywhere in settlement.
The same checklist runs for every new number: inflation, rates, benchmark returns. A listing is the checklist completed, published, and frozen. That is the whole process.