Quote-driven vs order-driven markets

Dealer markets quote you a price; order-driven markets match your order against a public book. Most modern venues (equities CLOBs, crypto exchanges, Polymarket) are order-driven, but the distinction decides who you trade against.

The idea

What is a quote-driven (dealer) market?

In a quote-driven market, a defined set of dealers (market makers) stand ready to buy and sell, quoting a bid and an ask on request. You trade against the dealer: the dealer is your counterparty, takes the other side of your trade, and warehouses the resulting position. Pricing is frequently bilateral and on-request (a request-for-quote, or RFQ) rather than discovered in a public book. The dealer's reward is the spread it quotes; its risk is being left holding inventory that moves against it.

The intuition is that a foreign-exchange desk or a bond dealer is a shop, not a noticeboard. You do not see a public queue of other customers' orders; you walk up and ask "what is your price in ten million?" and the dealer quotes you a two-way price. Liquidity is provided by the dealers, not discovered in a shared book, so pre-trade transparency is low: you see the quotes you ask for, not the whole market. The classic homes are spot FX, much of the corporate and sovereign bond market, interest-rate swaps and many bespoke derivatives. The institutions that intermediate this world (prime brokers and dealer banks) are central to it, which is exactly the friction that the newer order-driven crypto and prediction-market venues remove.

What is an order-driven market?

In an order-driven market, liquidity comes from participants' own orders posted into a shared central limit order book (CLOB). Anyone can be a maker (posting a resting limit order) or a taker (crossing the spread to consume one). There is no designated counterparty: you trade against whoever else is resting in the book, and the venue's matching engine, not a dealer, pairs buyers and sellers by price-time priority. The price is discovered openly, in the aggregate of everyone's order flow.

The intuition is that instead of asking a dealer, you join or hit a public queue. This is the world the rest of this atlas describes in detail: a limit order book you can post into, cancel from, and trade against at any instant. Market-maker roles may still exist (incentivised liquidity providers, designated market makers with quoting obligations) but they compete inside the same book rather than quoting bilaterally. The classic homes are lit cash equities, listed futures and options, and most centralised crypto exchanges. Pre-trade transparency is high (the book is published) which is precisely why microstructure signals like order-flow imbalance and the microprice are even computable here: they are read off a book that a dealer market never exposes.

What about hybrids and the AMM?

The clean two-way split is a teaching abstraction; most real markets are hybrids. Equity venues blend a continuous CLOB with periodic auctions and designated market makers with quoting obligations. FX and bond platforms layer RFQ, streaming dealer quotes and central books on top of one another. And decentralised crypto adds a genuinely new third pole, the automated market maker (AMM): a smart contract that quotes algorithmically from a pooled inventory, with no order book at all.

Even the canonical order-driven market has quote-driven and auction elements bolted on. A lit equity venue runs a continuous book plus open and close auctions, designated market makers, and off-book RFQ and dark venues alongside it. On the dealer side, FX and fixed income are electronifying unevenly: a continuum from voice and RFQ, through streaming dealer prices, to all-to-all central books in liquid pairs and on-the-run government bonds. The dealer model is eroding toward order-driven, but at different speeds in different products (as of 2026). The AMM is the most novel: on a decentralised exchange a single mechanism quotes from a bonding-curve formula (the constant-product rule $x \cdot y = k$ in the Uniswap-style design) against a pooled inventory. It is "quote-driven" in that one mechanism is always your counterparty, but the quote is a deterministic function of pool state, not a dealer's discretion. That is a genuinely different design, expanded in crypto market making.

Where does each asset class sit?

As a 2026 rule of thumb: equities, listed futures and options, and centralised crypto are order-driven CLOBs; spot FX, corporate bonds and OTC swaps are predominantly quote-driven dealer or RFQ markets, though electronifying; decentralised crypto runs on AMMs; and prediction markets such as Polymarket are order-driven CLOBs on thin, bounded-payoff contracts. The categories are stable, but which asset sits where is drifting, so always verify the specific venue.

The placement below is the map that lets you port the rest of this topic onto your own market correctly. Read it as "what am I actually modelling: a dealer's quote, a book's queue, or an AMM's curve?"

The portability lesson is the point of the whole exercise. The maths of microstructure (spread capture, queue value, adverse selection, inventory risk) transfers across all of these. What changes is only the object you point it at: a dealer's quote, a book's queue, or an AMM's curve. That is exactly the bridge the markets topic builds.

Worked example

A client wants to buy 10,000 units, shown both ways, as of 2026. Everything here is synthetic. First the quote-driven (RFQ) route: the client requests a quote in 10,000 from three dealers. Dealer A quotes 50.00 / 50.06, Dealer B quotes 50.01 / 50.05, and Dealer C quotes 50.00 / 50.07. The client buys 10,000 from B at 50.05: a single clean fill, with no market impact visible to anyone else, but a four-tick effective spread paid to the dealer, who now warehouses $-10{,}000$ and must hedge it.

Now the order-driven (CLOB) route. The same 10,000 as a market buy walks the public book: 3,000 at 50.01, then 4,000 at 50.02, then 3,000 at 50.03, for a volume-weighted average of about 50.02, tighter than the dealer quote on a deep, liquid book. But the sweep is visible on the feed and lifts the touch, leaking information the RFQ avoided entirely.

The trade-off the example makes concrete: the dealer market swaps transparency for discretion and risk transfer (no leakage, but a wider quoted spread and dependence on a counterparty who must be willing to warehouse your size). The order-driven market swaps discretion for transparency and competition (tighter prices in liquid names, but your sweep is public and moves the price). Flip the size to a block far larger than the book's depth and the ranking reverses: the dealer's risk transfer becomes cheaper than crawling up a thin book in full view. The numbers here are synthetic and rounded; real spreads, dealer behaviour and book depth are venue- and instrument-specific and must be verified and dated.