An Automated Market Maker Algorithm for Fixed-Rate Trading with Flexible Maturities
摘要
This paper aims to build an automated market maker (AMM) protocol to realize fixed-income products in decentralized finance. Compared to typical AMMs for cryptocurrency exchange, fixed-rate AMMs are much more difficult to design due to the inter-dependency of interest rate, loan price, and, especially, time factor. The literature lacks scientific developments. Existing efforts are either too simplistic, focusing more on computational automation and less on economic effect, or too engineering-crafted, using ad hoc formulations without economic justification. Indeed, today’s primal fixed-rate AMMs, Yield AMM and Notional AMM, are too financially-risky for liquidity providers (Yield AMM) and too expensive for borrowers and lenders (Notional AMM). In this paper, we propose BondMM, a novel fixed-rate AMM protocol with two major advances. First, BondMM is superior regarding properties that reflect the financial strength of a lending institution: 1) better price impact to attract borrowers and lenders, 2) better capital efficiency to attract liquidity providers (LPs), and 3) better financial stability for both users and LPs. Second, BondMM allows for 4) issuance and trading of loans having flexibly different maturity dates using only one liquidity pool; the maturity can be arbitrarily requested, long or short term. This is an innovative feature unseen in both the literature and the practice of DeFi. Existing fixed-rate protocols would dedicate a separate liquidity pool to each maturity date, hence causing fragmented liquidity. By having a single pool shared by all maturities and LPs, BondMM improves capital efficiency, increases liquidity, and decreases computational costs. Traders can easily swap cross-maturity and increase leverage by requesting long maturities. In terms of blockchain feasibility, BondMM is constructed, whose correctness provable, using elegant closed-form mathematical expressions, hence efficient for smart contract implementation. The properties and performance of BondMM are validated with our theoretical analysis and evaluation study simulating real-world and synthetic datasets.