We construct an empirically founded model of a repo trade intermediated by two broker-dealers and prove multiple equilibrium and the existence of equilibrium at the joint profit maximizing volume of trade. We then present a smart contract that resolves multiple equilibrium by requiring each broker-dealer to report its client schedule and its minimum hurdle spread, and implementing a selection rule that filters out hurdle-infeasible outcomes. Whenever there exists an equilibrium that exceeds both hurdle spreads, the protocol selects the joint profit maximizing feasible trade and thereby avoids a collapse to no trade. The smart contract is a machine executed algorithm which eliminates the need for trust. Hardware and cryptography are used to prevent leakage of broker-dealer client trade schedules, and to enable privacy-protected auditing with zero-knowledge proofs of the integrity of computations. The outcome can be implemented by a myopic strategy where a broker-dealer truthfully reports its own variables without anticipating its counterparty's reports. This minimizes cognitive and computational complexity, thereby making our smart contract suitable for real-world deployment.

翻译：我们构建了一个基于经验数据的回购交易模型，该交易由两家经纪商中介，并证明了多重均衡的存在性，以及在联合利润最大化的交易量上存在均衡。随后，我们提出了一种智能合约，通过要求每家经纪商报告其客户计划表及最低门槛利差，并实施一个过滤掉不可行门槛结果的筛选规则，从而解决多重均衡问题。当存在同时超过两家经纪商门槛利差的均衡时，该协议会选择联合利润最大化的可行交易，从而避免交易崩溃至零交易量。该智能合约是一种机器执行的算法，无需依赖信任。我们利用硬件和密码学技术防止经纪商客户交易计划表泄露，并通过零知识证明实现计算完整性的隐私保护审计。该结果可通过短视策略实现，即经纪商在不预期对手方报告的情况下真实报告自身变量，从而最小化认知和计算复杂度，使我们的智能合约适用于实际部署。