Multi‐party computation mechanism for anonymous equity block trading: A secure implementation of turquoise plato uncross

Abstract

Dark pools are financial trading venues where orders are entered and matched in secret so that no order information is leaked. By preventing information leakage, dark pools offer the opportunity for large volume block traders to avoid the costly effects of market impact. However, dark pool operators have been known to abuse their privileged access to order information. To address this issue, we introduce a provably secure multi-party computation mechanism that prevents an operator from accessing and misusing order information. Specifically, we implement a secure emulation of Turquoise Plato Uncross, Europe's largest dark pool trading mechanism, and demonstrate that it can handle real world trading throughput, with guaranteed information integrity.

Agents and Artificial Intelligence

“…In real financial markets, order book information is known to be strategically useful, as it exposes the trading intentions of market participants. In order to hide one's intention to trade, some trading venues, described as dark pools, do not reveal market quotes and all order information remains hidden (see, e.g., [5,6], for a summary of dark pools and methods for implementing cryptographically secure dark pool mechanisms using multi-party computation (MPC)). This ensures no information "leakage", and can therefore result in a better execution price.…”

Private Liquidity Matching Using MPC

“…Prior usage of MPC in financial applications has mainly focused on auctions, such as [5,8,9,29], for one shot auctions, and [4,13,14] for auctions running in Dark Markets. MPC was also used for privacy-preserving financial data analysis, such as [7] to conduct statistics over the performance of companies throughout the year or to compute the systemic risk between financial institutions such as in [2] and [25].…”

Kicking-the-Bucket: Fast Privacy-Preserving Trading Using Buckets

“…Yet, dark pools persistently suffer from negative reputation as some operators have taken advantage of their privileged access to the non-displayed orders in their systems. Indeed, between 2011-2018, dark pool operators paid more than $217 million to the SEC in penalty settlements for misusing customer order information or operating the dark pool in a way that disadvantaged their customers [CST20]. In the shadowy world of the dark pool, it is easier for a market manipulator to hide.…”

“…In [CST19], Cartlidge et al used MPC to present a proof-of-concept implementation of three dark pool trading mechanisms, showing that "volume matching" can be viably executed in a privacy-preserving manner with order throughput similar to that required by a real world dark pool trading venue. Further, in [CST20], Cartlidge et al demonstrated how to use MPC to run multiple auctions in parallel, offering simultaneous trading across thousands of stocks such that the identity of the stock being traded is also hidden and secure. The throughput per MPC engine is however significantly lower than that of the volume matching from [CST19] due to the use of a more complex matching algorithm.…”