For the second time this year, crypto markets have experienced a flash crash that brought the major exchanges to their knees – and with them, the client-facing trading platforms of many banks. Here is how these platforms can pass the next such stress test.
For bankers tasked with overseeing digital asset trading services, it must have felt like a recurring nightmare. The same day that El Salvador became the first country to make Bitcoin legal tender, exchanges around the world saw mass sell-offs, and many failed to keep up. Unfilled orders, app errors, outages and general chaos had knock-on effects for third-party trading portals, leaving traders frustrated whether they were trying to cut their losses or buy the dip. And all this came just a few months after the May flash crash, when the same issues came with widespread panic and slashed Bitcoin by a third.
The technical causes of congested blockchains
What causes all this disruption during crypto sell-offs? There are two main components. Firstly, although the blockchain networks that underpin digital assets are decentralized, most trading takes place through centralized exchanges. Such venues rely on conventional communication protocols such as REST and WebSocket APIs to maintain connectivity with the order and execution management systems (OEMS) of banks and brokers. Just as with any web-based service, when a huge volume of traders simultaneously attempts to sell their assets, the system sometimes cannot cope with the number of requests.
The second component is the nature of the blockchain networks themselves. On the blockchain, each transaction needs to be checked and approved by a group of network validators or miners. While the specifics vary for different cryptocurrencies, the miner who first proposes the next block of valid transactions typically earns a fixed “block reward” and a variable transaction fee for their work, the latter fluctuating according to supply and demand. During a huge sell-off, the number of pending transactions on the blockchain network spikes but the number of miners remains largely unchanged. This can lead to slower confirmation times and higher fees for each transaction.
Take Ethereum, for example. In the September 7 rush, average transaction fees quadrupled to $21.29, up from $5.53 the day before. Looking back at May 30, Bitcoin’s average confirmation time climbed to over 28 hours, a shocking increase on the usual 40-200 minutes.
The implications for banks trading crypto and digital assets
For banks providing digital asset trading services to their clients, the network congestion that can be triggered by large sell-offs gives rise to some formidable operational challenges. For instance, to execute client orders across multiple venues with a smart order router, banks need to continuously pre-fund these transactions by holding balances on multiple exchange-owned hot wallets. If on-chain transfers suddenly spike due to a rapid fluctuation in price, further cryptocurrency deposits to these exchange accounts are likely to be delayed, resulting in higher transaction fees.
This is particularly likely if many market participants are trying to transfer funds to exchanges in an effort to front-run each other with higher mining fees, and thereby capture a better exchange price than everyone else. While some banks may seek to use the credit lines offered by crypto brokers to avoid prefunding, they need to make sure they have a way to instantly transfer collateral to brokers. Otherwise, they will be hampered by the same problem and might incur the double blow of transaction fees and transaction costs. Ultimately, these issues can paralyse a bank’s crypto trading platform just when clients need it most.
Preparing for a future-proof digital asset banking
Despite the recurring volatility, most banks engaged in the digital asset market remain undeterred. The growing influence of institutional investors notwithstanding, crypto remains a largely retail-driven market, and is likely to remain a volatile asset class for the time being. So what can banks do to prevent service disruptions in the next big sell-off?
- Establish diversified connectivity:To maintain uninterrupted service when exchanges go offline, banks need to be able to route client orders to an array of venues. It is advisable to maintain business relationships with at least five major exchanges and two brokers that offer credit lines. In addition, to enable instant deposits, banks can work with a depository bank which maintains accounts for both counterparties (the venues and the bank). Furthermore, a smart order router (SOR) can be configured to automatically relay orders from unresponsive exchanges to responsive ones.
- Minimize on-chain transfers: For SOR to be effective, orders need to be pre-funded so that they can be executed immediately. By using a third-party, off-exchange settlement network like Fireblocks in conjunction with depository banks such as Signature or Silvergate, the execution timing risk arising from slow on-chain transfers can be avoided. This enables banks to almost instantly top up their exchange balances for client buy orders, or their credit lines with brokers for client sell orders. While this does not eliminate the counterparty risk inherent in prefunding third-party hot wallets, it greatly reduces it throughout the trading life cycle by automating asset transfers, thereby reducing operational risk. Furthermore, it gives banks the ability to use low-fee exchanges when price volatility is relatively low, while offering a fast exit via more expensive brokers when needed.
- Create synergies with brokers: During mass liquidation events, brokers tend to be short of liquidity to meet spiking client demand, which can result in lost commissions and unsatisfied clients. To further bolster their available options, banks can provide brokers with a continuous stream of liquidity to execute third-party client trades. In return, the banks can benefit from larger credit limits and hence serve a greater number of clients by executing orders via the broker. Bear in mind that brokers usually accept multi-currency margin collateralization. This can be achieved by transferring fiat between the vostro accounts of the broker and third-party depository banks such as Silvergate or Signature Bank.
- Create an automated first response system: Finally, banks can establish algorithmic tools to help respond to and prepare for market fluctuations. For example, using a platform like WIRESWARM by AlgoTrader, banks can establish a quantitative model that uses event-driven decision-making technology to automatically increase prefunding with exchanges and collateral deposits with brokers, for instance when market signals indicate that a period of volatility is approaching. In cases where the bank adopts a principal brokerage model – meaning that it executes client orders against its own inventory – quant models can be used to continuously neutralize exposure to crypto derivatives such as futures, perpetual swaps or options.
The past 12 months have provided evidence of two enduring trends in the digital asset industry. On the one hand, established financial institutions and banks – including Goldman, JPMorgan, Citi and BNY Mellon – continue to build digital asset services for their clients. On the other hand, the market remains strongly influenced by retail investor sentiment, as panic buying and selling during bouts of volatility are highly lucrative business opportunities for professional traders.
To prevent service disruptions for clients during sell-offs, banks need to establish broad-based connectivity to exchanges and fully automate trade life cycles in order to optimize funding and settlement mechanisms in terms of risk, timing and cost. This will enable them to avoid the pitfalls of network congestion and continue to provide a seamless service to their clients, even in the most challenging of circumstances.