Bloomberg reported this week that the U.S. Justice Department has opened a criminal probe into whether traders are manipulating the price of Bitcoin and other digital currencies, specifically by spoofing – entering fake orders to trick other traders into buying or selling.
Spoofing was once common in the stock and commodities markets, but has since been curtailed by robust trade surveillance programs at both the exchange and brokerage levels. Could similar trade surveillance programs prevent spoofing in the crypto markets? Perhaps, but there are some unique features of the current crypto trading market structure that make trade surveillance a challenge. To name a few:
Fragmentation. Like the stock markets but unlike the futures markets, the same crypto currencies can be traded on multiple venues. This creates a trade surveillance challenge because a spoofer can enter fake bids on Venue A, wait for professional traders who consume quote data from multiple venues to react, and then trade opposite those professional traders on Venue B. In this scenario, neither Venue A nor Venue B, acting separately, has enough data to identify the subject activity as spoofing.
In the stock markets, Venue A and Venue B solve this problem by agreeing to share their quote data with a common host (currently FINRA and soon to be the Consolidated Audit Trail) who can stitch their respective fragments of the market together, and can then use the resulting pooled dataset to identify cross-venue spoofing.
No such common host exists yet for the crypto markets. This leaves each venue able to identify only spoofing occurring entirely within each venue, but not spoofing occurring between or among two or more venues. This is a significant blind spot, but is one that may be corrected soon. One crypto exchange, Gemini, has proposed creating a new “Virtual Commodity Association” to do exactly that.
Linking Accounts. Even if crypto exchanges did pool their quote data, they would still need to overcome the additional challenge of linking accounts across multiple venues. Each venue identifies customer bids and offers by an account number that is only meaningful to that venue. If the same trader opened both Account A at Venue A and Account B at Venue B, and Venue A and Venue B agreed to pool their order data with a common host, the common host would not know that Account A and Account B were related unless it had some additional account linking information.
The most obvious way to link accounts would be to use the account holder’s social security number. This solution was proposed in the blueprints for the Consolidated Audit Trail. But due to data breach concerns, that proposal was torn to shreds by regulators and even members of Congress last year, and as of this writing a suitable alternative has yet to be proposed. It is unlikely that any similar proposal in the crypto space would fare any better.
Another potential piece of account linking information could be the public key for the crypto wallet of the account holder. This key is unique to each wallet and is used to record transfers of title on the blockchain. The same trader could use the same public key to record deposits and withdrawals of cryptocurrencies to and from multiple venues. If that public key was linked to each of his bids and offers across multiple venues, it would suffice to allow a common host of quote data to identify cross-venue spoofing.
The problem with this approach is that crypto wallets are cheap enough to obtain to be the equivalent of burner phones for would-be manipulators. It would be trivial for a spoofer to identify himself to Venue A as Wallet A and to Venue B as Wallet B, thus successfully concealing any link.
Direct Market Access. Unlike on the stock exchanges, individual investors can trade on crypto exchanges directly without holding an account at an intermediary broker. This bypasses an important layer of trade surveillance. In the stock markets, brokers can (and do) effectively screen their own customers’ order flow for spoofing even when those customers are spraying bids and offers to multiple different exchanges.
In futures markets, individual investors can access the exchange directly, but the instruments traded are monopoly products that can only be traded on one exchange. Since 100% of the market’s bids and offers for each instrument arrive at one exchange, that exchange is able to effectively screen for spoofing despite the absence of intermediary brokers.
By combining both direct market access and fragmented multi-venue trading, the crypto markets miss both of the layers of trade surveillance that effectively constrain spoofing in the stock and futures markets.
None of the above challenges are reason not to implement the spoofing controls that are within reach. Each venue can and should conduct trade surveillance of the trading within its control. Perfect should not be the enemy of the good. But to reach the level of global market integrity enjoyed by the stock and futures markets, some structural changes are needed for the crypto markets.