> For the complete documentation index, see [llms.txt](https://initiativ.gitbook.io/initiativ/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://initiativ.gitbook.io/initiativ/mm_internal/4.-trading-mechanisms/4.2-request-for-quote.md).

# 4.2 Request for Quote

{% hint style="warning" %}
**Pending acceptance and review.** New information has surfaced since this section was written and the RFQ product design is being reconsidered. The section below records the design as it currently stands and is included for review rather than as a settled position. Material aspects of it may change.
{% endhint %}

RFQ is a competitive quoting channel for trades in institutional size. An initiator requests firm quotes for a named instrument, direction and size; one or more responders return firm prices in competition; the initiator executes against them. It runs in two modes: anonymous, the default, where the request is broadcast to eligible responders with no identities shown on either side, and disclosed, where the initiator hand-picks who to ask and both sides see each other. RFQ has its own interface; its backend shares trade registration and clearing with OTC and Block Clear where sensible, rather than being an extension of the OTC screen.

RFQ serves two purposes for the Exchange. First, it carries block interest that would otherwise move the screen, complementing the order book rather than fragmenting it, with a minimum size per product keeping ordinary flow on the book. Second, it is a liquidity-building gateway: it lets the Exchange offer instruments such as TAS from launch, before on-screen depth exists to support them, with the intention of migrating that flow onto the order book once liquidity is established.

At launch RFQ carries EUA futures outright, EUA futures calendar spreads, and TAS on futures. Spot is not offered through RFQ at launch: it trades on the central order book, with spot TAS routed to the Trade Scheduler (Section 4.4). Trade at Auction (TAA) is under consideration but not confirmed for launch (see open items).

## Core mechanics

#### Firm quotes only

Responders submit firm, binding prices. There is no indicative pricing, no negotiation and no counter-proposal loop.

#### Asymmetric binding

The initiator specifies instrument, direction and size, never a price, and is bound only when they actively accept a quote. A responder is bound the instant they submit a quote and cannot withdraw it. One narrow exception on the initiator side: in anonymous mode they may optionally attach a Limit price to the accept action itself, not to the request, which bounds the sweep to quotes at or better than that price. This asymmetry is what lets counterparties compete on price while the initiator carries no obligation until they choose.

#### Multiple execution

A single request can be filled across several responders’ quotes rather than requiring one winner. Each resulting fill is a separately reported trade.

#### Cancellation

The initiator may cancel an outstanding request at any time before acceptance, in either mode, with no Trading Limit consequence. Responders cannot cancel a submitted quote; their exposure is bounded by the validity window instead.

#### Quote validity window

The initiator sets the window per request, bounded by a per-product default and maximum. For the launch products the default is 1 minute and the maximum 5 minutes. For TAS on futures the window can never extend past the 17:15 settlement cutoff.

## Modes and allocation

{% tabs %}
{% tab title="Anonymous (broadcast)" %}
The initiator has a single accept action, optionally carrying a Limit price. Accept with no limit triggers a price-time priority sweep across all available quotes, filling down the priority order until the size is met or quotes run out. Accept with a limit runs the same sweep but only over quotes at or better than the limit, stopping before anything worse; quotes not reached go to Not selected and any shortfall is a partial fill.

Either way allocation stays fully mechanical: price-time priority decides who fills, and the limit is only the initiator’s own commercial boundary, the same role a limit price plays on the order book. A partial fill on available liquidity is possible.
{% endtab %}

{% tab title="Disclosed (targeted)" %}
The initiator accepts individual quotes one at a time, in any order and combination, allowing a deliberate split across counterparties. Allocation is manual, reflecting the initiator’s own choice of whom to ask.
{% endtab %}
{% endtabs %}

## Counterparty selection

Initiators may hand-pick which counterparties receive a disclosed request, chosen fresh each time; there is no venue-provided persistent target or block list. The fair and non-discretionary duty binds Initiativ as operator, not a client’s own commercial choice of whom to ask or which quote to accept.

## Minimum size

A minimum size applies per product, set case by case. At launch:

| Product              | Minimum size                                               |
| -------------------- | ---------------------------------------------------------- |
| EUA futures outright | 50 lots (matching ICE’s block convention for the contract) |
| EUA calendar spreads | Legs summing to 50, that is 25 lots per leg                |

## Eligibility

Not every market can be offered over RFQ. A market qualifies only if it is liquid enough that a single order cannot move or game it. The RFQ trade itself is private and feeds no calculation; the concern is what follows. Whoever takes on a large position from an RFQ trade typically hedges or unwinds it in the visible market afterwards, and in a thin market that activity moves the price, including the daily settlement print that marks every holder of the contract. For TAS the concern is sharper, since a TAS holder benefits directly from where the settlement window lands and so has an added incentive to trade into it; a thin market makes that cheap, a deep one makes it expensive and obvious.

At launch the test is that a product’s minimum trade size must be no more than roughly 10 to 20 percent of representative volume in a window comparable to the settlement window (about 10 minutes), measured on the settlement window for TAS and on the thinner leg for spreads. Ahead of launch, with no live intraday data, this is assessed against an interim proxy (average session volume spread evenly across the session), replaced by measured window-level volume once trading begins.

This launch test is a known simplification, not the end state. Minimum trade size is a commercial decision, while eligibility is meant to be a pure risk control, so tying the two together risks the control drifting whenever the commercial number changes for unrelated reasons. It is accepted for launch only because a properly decoupled test needs real trade-level data that does not exist yet. Once live, eligibility will be set from the market’s own trade-size distribution against window volume (a percentile-based test), ideally paired with a breadth measure such as the number of distinct trades or counterparties in a comparable window, since size alone can miss a market whose volume is dominated by one or two large trades.

## Validation and risk

Validation timing mirrors the binding: a responder is validated in full at quote submission (account status, eligible markets, Trading Limit, maximum quota, participant sub-limit, and the fat finger and price collar checks), and the initiator at acceptance. A responder’s quote reserves Trading Limit from submission until it is cancelled, expired, executed or not selected. The price collar applies to every quote, with the reference built per instrument type: the order book price for an outright, the implied spread on the thinner leg for a spread, and a differential band around the expected basis for a TAS quote. Self-trade prevention is the same Member-level automated block used on the order book, with beneficial-owner analysis handled by the MAR Article 16 surveillance layer. Full check-by-check logic is in the Order Validation and Entry Controls Spec (Section 7.3). RFQ request and quote statuses form part of the platform status vocabulary, consolidated in the Order and Trade Lifecycle section.

## Transparency and regulatory grounding

RFQ relies on the pre-trade transparency carve-out for RFQ systems under MiFIR (RTS 2), confirmed by the 2024 MiFIR review. Post-trade transparency applies in full to every executed RFQ trade; large trades may qualify for deferred publication once an AMF-approved deferral regime is in place. Running RFQ alongside the order book under a single MTF licence does not create reclassification risk provided allocation stays non-discretionary, which is why anonymous mode uses an objective price-time sweep rather than free initiator choice. All quotes, including losing, expired and withdrawn ones, are retained for five years for regulator access (MiFIR Article 25).

## Open items

* The Trading Limit and margin mechanism for futures exposure on RFQ (outright futures, futures spreads, TAS on futures) is not yet defined and depends on the CCP and clearing model; the spot-style Trading Limit concretely applies only to the spot leg of a spot-versus-future spread, itself gated on CCP acceptance.
* The post-trade deferral regime for large trades requires design and AMF approval.
* Eligibility to be decoupled from minimum trade size once live (percentile-based test on actual trade-size distribution, ideally paired with a breadth measure).
* Whether TAS on spot should also be offered via RFQ in large size, rather than remaining exclusive to the Trade Scheduler.
* Whether TAA is offered at launch at all.