Crypto wash trading on centralised exchanges is the industrial-scale version of the oldest trick in markets. Through the 2017 to 2021 cycle, exchange rankings were sorted by reported volume, listing fees flowed to venues with the biggest numbers, and the numbers were self-reported. The predictable result: on many unregulated venues, most of the tape was manufactured.
This page covers the exchange-level phenomenon and how it differs from the trader-level wash trading I study in The Economics of Wash Trading. The two are routinely conflated, and the confusion matters, because they have different perpetrators, different victims, and different fixes.
The benchmark evidence
The reference study is Cong, Li, Tang and Yang, Crypto Wash Trading, published in Management Science. Using statistical fingerprints of authentic trading, first-digit laws, size clustering, tail distributions, they estimated that unregulated exchanges in their sample inflated volumes massively, with wash trading exceeding 70% of reported volume on many venues, while regulated exchanges looked clean. The paper turned a market rumour into a measured fact and forced the data aggregators to rebuild their methodologies.
The mechanism was almost boring: the exchange, or bots it tolerated, traded against itself. No third party needed to be present at all. The victims were downstream: users who chose venues by liquidity, projects that paid listing fees benchmarked to traffic, and anyone whose model consumed the numbers.
Trader-side wash trading is a different animal
In NFT markets the venue usually was not the manipulator. Traders were, and they were responding to prices the venues had posted for volume: trade-to-earn token emissions, rankings, airdrop eligibility. My core finding is that this incentive channel, not buyer deception, explains where NFT wash trading concentrated. Wash volumes showed no significant relationship with genuine future volumes; they tracked the rewards, and on venues like LooksRare the rewards even shaped real activity around them.
The comparison is worth a table:
| Dimension | Exchange wash trading | NFT trader wash trading |
|---|---|---|
| Who manufactures | The venue or tolerated bots | Traders with sibling wallets |
| Payoff | Rankings, listing fees, users | Token rewards, occasionally marks |
| Data visibility | Aggregates only, off-chain | Every trade public, on-chain |
| Detection | Statistical tests on the tape | Funding graphs and round trips |
| Durable fix | Audited or on-chain reporting | Reward designs that ignore self-financed flow |
What changed, and what did not
Aggregators now discount or exclude unverifiable venues, several jurisdictions require surveillance of crypto markets under regimes like MiCA, and enforcement actions have treated manufactured volume as fraud. But the underlying incentive never went away: wherever a statistic routes money or attention and is cheap to fake, it will be faked. The reading list for spotting it is on fake trading volume, and the general-purpose toolkit is on wash trading detection.