Spectrum is allocated in enormous set-piece auctions every few years and then, in most countries, barely moves again. Licences sit underused because the secondary market that regulators authorised on paper never acquired a venue, a price, or a process quick enough to be worth using. This page is about that secondary market — and it opens by saying what Mehrex cannot do for the primary one.
Most regulators have permitted spectrum trading and leasing for years. The volume is negligible, and the reasons are structural rather than legal — there is no venue, no price, and an approval process measured in months against opportunities measured in weeks.
A national award is a combinatorial clock auction designed and run by economists on contract, over days, every few years. It is the right mechanism for the problem, it is extremely well studied, and it has essentially nothing to do with the day-to-day market that ought to exist afterwards.
Two licensees that want to trade must find each other, negotiate privately, and then wait for a regulatory consent process with no fixed clock. By the time it completes the commercial reason for the trade has often expired, which is why most parties do not begin.
A licensee holding national rights it uses in six cities has no practical way to let anyone else use the rest, so it does not. That is not hoarding in bad faith; it is the predictable result of transaction costs exceeding the value of a regional sublease.
A private network for a port, a mine, a stadium or a harvest season is exactly the use case that licensed spectrum should serve. There is no mechanism to buy three months of a specific block in a specific area, so those networks either do not exist or run unlicensed and take their chances.
Once a licence can be sliced by band, geography and term, the resulting units are ordinary tradable instruments — and the value of publishing their price is larger than the trading revenue, because it is the missing input to every private-network business case.
A one-year lease of a defined block in a defined licence area is a book. The technical licence conditions travel with it; what trades is a standard, time-boxed right of access rather than a negotiated bilateral arrangement drafted from scratch.
Nobody building a private network can currently answer what a year of local spectrum costs. A scheduled uncrossing answers it with a transaction, which turns an unfinanceable proposal into a line in a model — and gives the regulator evidence about whether its allocation is efficient.
Where blocks come and go at short notice, a continuous book with IOC and FOK validity lets a buyer take what is available now. Longer terms clear better in scheduled calls. Both run on the same engine, configured per instrument.
Spectrum caps are a standard competition remedy, and in a secondary market they have to bind continuously rather than at award. As pre-trade position limits per participant they are structural: an order that would breach a cap is refused with the limit named.
Every order and every operator action is journaled in one gap-free sequence before it executes. A regulator supervising a market it authorised can reproduce any period exactly, which is a materially different conversation from being shown a report about it.
Instruments sliced by band, area and term; caps that bind before execution; and an operator console a regulator can co-supervise. The engine's own auction is a single-instrument uniform-price call — capable, and deliberately not a package auction.
| Setting | Typical starting point |
|---|---|
| Instrument | Band × licence area × term length |
| Lot size | One block-year, or the smallest sublicensable unit |
| Sessions | Scheduled call auction per term; continuous for short-term blocks |
| Validity | GTD across the approval window; IOC and FOK for immediate availability |
| Limits | Holding caps per participant, enforced pre-trade |
| Approval | Four-eyes on bands, areas, eligibility and manual transitions |
| Settlement | Conditional on regulator consent to the transfer |
Scheduled call auctions per instrument with per-day overrides, continuous trading for short-term blocks, and IOC and FOK validity for buyers taking immediate availability.
Read more →04 · coreOpen-position limits per participant, enforced pre-trade, so competition caps on spectrum holdings bind continuously rather than being reviewed after the fact.
Read more →02 · channelBands, licence areas, terms and eligibility maintained in a browser under maker–checker approval, with every market-affecting change carrying a requester and an approver.
Read more →05 · post-tradeNetting per participant and settlement date, fees per trade, and enriched trade-capture records to drive the licence-transfer filing.
Read more →The first item is the most important thing on this page, and it disqualifies Mehrex from the largest and most visible auctions in this sector. Read it before anything else.
Mehrex runs single-instrument uniform-price call auctions. A national primary award requires package bidding — a bidder needs this block and that block together, or neither, and the mechanism must handle those contingencies across many rounds. That is a combinatorial clock auction, a genuinely different design, and you should buy it from the specialists who build and run them. The fit here is the secondary market, and pretending otherwise would waste everyone's time.
Whether a licence can be traded, leased or sublicensed at all, on what terms, and how quickly a regulator will consent, is the binding constraint on this entire market. A matched trade is conditional until consent arrives. Shortening that process is a regulatory reform project, not an integration.
Guard bands, coexistence with neighbouring licensees, power limits and coordination agreements travel with the right of access. Someone has to verify that a proposed transfer is technically admissible before it is allowed to settle, and that assessment is an engineering function.
A spectrum lease is worthless without radios, sites, backhaul, cores and installers. The reason short-term local licensing has not taken off is at least as much about the cost of deploying for a season as about the cost of the spectrum itself.
Mehrex enforces whatever holding limits you configure. Deciding what they should be, per band and per market, and defending those numbers publicly, is a competition-authority judgement with a large evidential base behind it.
The lease agreement, remedies for interference or non-performance, and the reversion of rights at the end of a term. In a market where the asset is invisible and shared, the dispute framework does more work than usual.
No. National primary awards require a combinatorial clock auction with package bidding, so that a bidder can express that it wants two blocks together or neither. Mehrex runs single-instrument uniform-price call auctions, which is a different mechanism. Its fit is the secondary market — leasing, subleasing and short-term access — not the primary award.
Giving the authorised-but-unused secondary market a venue: standard instruments defined by band, licence area and term; a scheduled uncrossing that publishes a real price for spectrum access; holding caps enforced before an order reaches the book; and a journal a regulator can replay.
Yes, as pre-trade open-position limits per participant, evaluated at clearing-member, firm and client level before an order reaches the book. An order that would breach a cap is rejected and the rejection names the limit and the scope that tripped. Deciding what the caps should be remains a competition-policy question.
Vessel and haulage capacity — perishable, brokered by telephone, and benchmarked by panel assessment rather than by trades.
Read the page →Capacity & accessAccelerator hours and colocation power — extremely perishable, priced bilaterally, and hoarded because there is no spot price.
Read the page →Precious metalsAn app that sells sub-gram quantities of gold to retail buyers, run as a real order book instead of a dealer’s quote.
Read the page →A live demonstration runs about an hour: the order book, the Control Panel, an auction uncrossing, and a replay of the journal that produced it.