CEX-to-CEX Arbitrage
Learn how cross-exchange crypto arbitrage compares executable prices, pre-funds inventory and manages transfer, fee, latency and counterparty risk.
Reading progress — saved on this device
CEX-to-CEX arbitrage seeks to buy an asset where it is cheaper and sell where it is more expensive, but the quoted spread is only useful if both legs can be executed and settled.
Learning objectives
- Calculate executable rather than headline cross-exchange spread.
- Distinguish pre-funded simultaneous execution from transfer-first execution.
- Account for fees, depth, latency, withdrawal and counterparty risk.
What it is
Two centralised exchanges can quote different prices because their user bases, fiat rails, inventory and liquidity differ. Arbitrage trading helps reconnect those fragmented markets.
A robust desk often pre-funds both venues: cash/stablecoins on the cheap venue and asset inventory on the expensive venue. It can then buy and sell nearly simultaneously and rebalance inventory later.
How it works
Headline best bid/offer can be misleading. A £100,000 order may consume several book levels, so the relevant comparison is volume-weighted executable price on both venues.
Trading fees differ by tier and maker/taker status. Withdrawal fees and blockchain costs matter for subsequent inventory rebalancing.
A transfer-first retail approach—buy, withdraw, wait, deposit, then sell—carries price risk during the delay. By arrival, the spread may have closed or reversed.
A persistent premium may signal capital controls, withdrawal problems or counterparty risk. The spread can be compensation for being unable to move value out of the venue.
How to analyse and apply it
| Check | Why it matters | What to verify |
|---|---|---|
| Depth/VWAP | Determines real price at size. | Simulate both order books for intended quantity. |
| Fees | Can consume small spreads. | Use actual account fee tier. |
| Settlement | Controls inventory rebalancing. | Check deposits, withdrawals, network and limits. |
| Counterparty | Explains some persistent premia. | Apply venue exposure limits; do not assume parity implies solvency. |
A strategy is not complete until the signal, sizing, execution, invalidation and review process are explicit. Any discretionary override should be recorded so it can be separated from the tested rule set.
Worked example and thought exercise
Exchange A shows BTC ask £50,000 and Exchange B bid £50,300: headline spread £300. For a 2 BTC trade, book depth makes actual buy VWAP £50,080 and sell VWAP £50,240, leaving £160/BTC gross.
If taker fees total £70/BTC across both legs and expected rebalancing cost is £20/BTC, net spread is about £70/BTC, or £140 for 2 BTC before operational risk.
Thought exercise: why might a 5% premium on an exchange with suspended withdrawals be less attractive than a 0.4% premium on a healthy venue?
Common mistakes and practical workflow
- Comparing top-of-book prices without depth.
- Buying first and hoping the spread remains during transfer.
- Ignoring account-specific fee tiers and limits.
- Treating persistent premiums as free money rather than possible credit-risk signals.
Practical workflow
- Check venue status and withdrawal/deposit functionality.
- Calculate VWAP on both books for the same quantity.
- Subtract all known friction.
- Execute both legs only within latency/risk limits.
- Rebalance inventory separately with transfer and counterparty controls.
✅ Knowledge checkpoint
- Why is top-of-book spread insufficient?
- What advantage does pre-funding both venues provide?
- How can withdrawal suspension explain a persistent premium?
- What turns transfer-first arbitrage into a directional trade?
FAQs
❓ Why do exchange prices differ?
Crypto markets are fragmented by venue, user base, capital and settlement rails.
❓ Can arbitrage remove all price differences?
No. Costs, latency, limits and credit risk create bands within which differences can persist.
❓ Do I need APIs?
Not conceptually, but fast and reliable execution becomes more important as spreads shrink.
❓ What is the biggest non-price risk?
Often venue/counterparty and settlement risk, especially when capital must remain pre-funded.
📋 Summary
CEX-to-CEX arbitrage monetises fragmentation only when the spread is executable, net of friction and supported by reliable settlement. Pre-funding reduces latency risk but increases capital tied to exchange counterparties, so credit limits are integral to the strategy.
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 →