Institutional Market Makers
Institutional guide to crypto market making, inventory risk, quote skew, adverse selection, hedging, latency and liquidity withdrawal.
Reading progress — saved on this device
Institutional market makers provide executable two-sided liquidity while managing inventory, adverse selection, latency and hedge risk. The spread is compensation for those risks, not guaranteed profit.
Learning objectives
- Understand why market-making spreads compensate for more than fees.
- Analyse inventory, adverse-selection, latency and cross-venue hedge risk.
- Interpret quote skew and liquidity withdrawal during volatility.
How institutional market making works
A market maker posts bids and offers on an order book or answers RFQs while trying to earn spread, rebates and related execution revenue within risk limits. In quiet markets fair value changes slowly and hedges are usually accessible, allowing tight spreads. During fast markets, stale quotes are more dangerous and expected hedge slippage rises, so rational makers widen spreads or reduce displayed size.
Inventory management is central. If clients repeatedly sell into the market maker’s bid, the maker becomes long inventory. It can respond by lowering its bid, improving its offer, reducing bid size, increasing offer size, hedging with futures or trading another venue. Those adjustments are called inventory or quote skew and are intended to encourage flow that brings risk back toward target.
Adverse selection is the risk that counterparties know or react to information faster than the maker. A quote can earn a nominal spread and still be a bad trade if fair value moves against the maker before it can hedge. This is why sophisticated pricing models consider short-horizon volatility, trade toxicity, queue position and latency rather than using a fixed spread.
Crypto fragmentation creates both opportunity and risk. A maker may quote BTC on one exchange, hedge on another and finance inventory using a third venue. Economic delta can look neutral while operational exposure remains concentrated in one stablecoin, custodian or exchange. Withdrawals or API failures can turn a hedge problem into a balance-sheet problem.
Latency includes more than network speed. Market data can be stale, order acknowledgements delayed and exchange risk engines inconsistent under load. A maker that receives a fill on Venue A but cannot place the hedge on Venue B for several seconds is exposed to directional movement during the delay. The expected cost of that delay belongs inside the quoted spread.
Market makers also manage capital allocation across products. Quote size should reflect the depth of the hedge market and the amount of collateral available, not just local order-book demand. A maker may deliberately withdraw from a thin token even when spread is wide because the inventory cannot be reliably hedged or liquidated.
For liquidity consumers, the lesson is that displayed liquidity is conditional. A normal-time order book does not guarantee depth during liquidation cascades, news or venue outages. Institutional execution should examine historical stress depth, quote resiliency and the number of genuinely independent liquidity providers rather than count every displayed order as permanent capacity.
Worked example
A maker quotes BTC £49,990 bid and £50,010 offer around a £50,000 fair-value estimate. The full spread is £20, or 4 basis points. A client sells 100 BTC at £49,990, leaving the maker long about £4.999 million.
If the maker can immediately sell or short the hedge at £50,000, it realises approximately £10 per BTC of positive hedge P&L relative to the £49,990 client purchase, or about £1,000 on 100 BTC. If the best available hedge is only £49,985, it instead loses about £5 per BTC, or £500 on 100 BTC. If the hedge market has already fallen to £49,900 by the time the fill is known, adverse selection has cost roughly £90 per BTC relative to the client fill, or about £9,000 on 100 BTC, far more than the nominal spread. A maker’s profitability depends on realised hedge and inventory outcomes, not the posted spread alone.
Institutional analysis workflow
- Measure quoted spread, displayed depth and actual executable size.
- Map inventory limits and the venues/instruments used for hedging.
- Measure post-fill fair-value movement to estimate adverse selection.
- Track hedge latency, failed orders, transfer constraints and venue outages.
- Compare quote resiliency and depth during volatility, not only in normal markets.
Common mistakes
- Treating quoted spread as guaranteed market-maker profit.
- Ignoring hedge-market depth, latency and transfer constraints.
- Assuming exchange diversification removes stablecoin, custodian or infrastructure correlation.
- Interpreting quote withdrawal as manipulation when it may reflect rational risk limits.
- Using normal-time displayed depth as a stress-liquidity assumption.
Knowledge checkpoint
- What is adverse selection in market making?
- How does quote skew help manage inventory?
- Why can a delta hedge still leave operational risk?
- Why do spreads generally widen during volatility?
FAQs
❓ Why do market makers widen spreads when markets move quickly?
Expected adverse selection, inventory variance and hedge slippage increase, so the quote must compensate for greater risk.
❓ What is quote skew?
It is the adjustment of bid/offer price or size to influence incoming flow and manage an inventory imbalance.
❓ Are market makers directional traders?
They can carry temporary directional inventory, but many institutional strategies aim to keep that exposure within explicit limits and hedge it.
❓ Can liquidity disappear suddenly?
Yes. Risk limits, exchange outages, rapid price discovery and balance-sheet constraints can cause displayed liquidity to retreat quickly.
Summary
Institutional market making is a continuous pricing-and-risk-management process. Spread capture must compensate for inventory, adverse selection, latency and hedge cost, while liquidity consumers should assume that quoted depth becomes less reliable during the very events that create the greatest execution urgency.
Want this in a personalised order?
Take the crypto assessment and get a custom path of 10 modules matched to what you already know. Free, no card required.
Build my path →