Crypto Wash Trading: Fake Volume and How It Is Detected

The U.S. Commodity Futures Trading Commission defines wash trading as entering, or appearing to enter, transactions that make purchases and sales seem to occur without incurring market risk or changing the trader's market position. Securities regulators commonly describe the related concept as self-trading or trades without a meaningful change in beneficial ownership that create a false appearance of activity.¹⁶

The economic feature is artificial activity, not merely buying and later selling the same asset. Two independent parties can legitimately trade frequently. A market maker can buy and sell throughout the day while accepting inventory and price risk. By contrast, coordinated sides of a trade can neutralize ownership or risk while producing prints that look like independent demand.¹

Wash trading is also different from a tax 'wash sale,' a jurisdiction-specific rule about realizing a loss and reacquiring an asset. This guide concerns market activity and volume integrity. It does not address tax-loss rules.

Volume is used as a shortcut for attention, liquidity, and market quality. Traders choose venues and assets partly from activity rankings; token projects seek listings; index and data providers filter markets; and exchanges advertise market share. Artificial volume can improve those visible positions and create social proof that draws real orders.²³

The shortcut fails because executed volume is only one dimension of liquidity. A market can print a large notional amount between coordinated accounts while offering little independent depth. A real trader then encounters a wide spread, thin book, high price impact, delayed withdrawal, or an inability to sell at the displayed price.³⁴

Artificial prints can also contaminate reference prices, volatility estimates, venue rankings, empirical research, and risk systems. The harm extends beyond one exchange when aggregators or automated strategies treat self-reported volume as comparable across venues without quality controls.²⁴

On a centralized exchange, the venue controls the matching engine, customer accounts, and internal ledger. A customer may coordinate two accounts, a market-making service may self-trade through algorithms, or the venue may publish data that cannot be reconciled to genuine orders. Public observers usually lack beneficial-owner and account-control records, so definitive attribution often requires venue data or investigative authority.⁶⁷³

On a decentralized exchange, transactions and wallet addresses may be public, making circular fund flows, repeated counterparties, and common funding sources more observable. But a wallet address is not automatically one person: one actor can control many addresses, custodians can control pooled addresses for many users, routers split trades, and arbitrage can create repeated patterns without manipulation.⁵

In the U.S. Department of Justice's 2024 NexFundAI operation, charging documents alleged that purported market makers used multiple wallets and self-trading bots to manufacture apparently organic volume. Those allegations illustrate the evidence investigators seek—control, instructions, code behavior, communications, and purpose—beyond a public anomaly score.⁷⁶

A 2019 Bitwise submission connected to a proposed bitcoin exchange-traded product examined trade and order-book patterns across more than 80 venues reporting meaningful volume. It argued that roughly 95% of the reported spot bitcoin volume in its March 2019 sample was fake or non-economic and identified a smaller group whose patterns appeared consistent with real markets.³⁴

The SEC did not simply adopt that conclusion as a universal fact. In its order on the proposal, the Commission examined whether the sponsor's method reliably separated real from fake volume and whether the identified segment was insulated from manipulation elsewhere. That regulatory scrutiny is a useful warning against turning a vendor screen into a final legal finding.⁴

Cong, Li, Tang, and Yang later studied 29 exchanges using first-significant-digit distributions, trade-size rounding, and transaction-tail patterns. Their NBER working paper, subsequently published in Management Science, reported that estimated wash trading averaged more than 70% of reported volume on the unregulated exchanges in their sample and linked fabricated volume to rankings and short-lived price effects.²

These results document serious historical integrity problems and useful methods. They do not establish that 70% or 95% of all crypto volume in August 2026 is fake. Venue rules, regulation, market composition, data quality, pair selection, incentives, and surveillance change. A current estimate requires a new timestamped sample and published method.²³

A retail observer cannot reproduce regulator-grade surveillance from a 24-hour leaderboard. The goal is to identify mismatches and reduce reliance on one number. Compare the same liquid pair across several venues at the same moment, and record the result rather than relying on a screenshot selected after the fact.

The CFTC's statutory wash-sale prohibition directly addresses covered futures, options, and swaps; securities laws and antifraud provisions can apply to covered securities activity; and other criminal, commodities, consumer-protection, or market-manipulation theories may apply depending on the conduct. A spot crypto asset's legal treatment and the regulator with authority cannot be determined from this page.¹⁶⁷

The 2024 U.S. cases against purported crypto market makers show regulators and prosecutors focusing on self-trading algorithms and services marketed as artificial volume generation. Those matters involve allegations and case-specific evidence; they should not be generalized to every market maker or venue.⁶⁷

In the European Union, MiCA establishes market-abuse obligations for in-scope crypto assets and trading platforms, while ESMA has issued supervisory guidance addressing the cross-border and social-media features of crypto markets. Exact obligations depend on the asset, actor, venue, date, and applicable law.⁸⁹