Currently, the DeFi lending landscape is dominated by Peer-to-Pool, variable rate models characterized by governance-heavy* architectures where most risk-related and other critical parameters (such as interest rates) are determined by governance in a top-down (and at times reactionary) fashion. While these models were a 0 to 1 innovation in DeFi and still offer (and will probably continue to offer) significant utility to users, we believe there’s an opportunity to expand the current lending landscape in two particular areas:
- Fixed rates: The value of fixed rates is obvious. They would allow borrowers to predictably understand the costs of a given loan. This is in stark contrast to the status quo, where the cost of a given loan can significantly change over time, ultimately degrading UX.
- Market-based parameters: Rather than all critical parameters being determined by governance, we believe there’s an opportunity to gradually move to a more market-based approach, where (some) parameters are freely set by market participants. While ideally all parameters would be fully market-driven, this level of granularity would probably lead to significant liquidity fragmentation, ultimately degrading UX. Thus, we believe a hybrid approach, where some parameters are still governance defined while others are fully market-driven makes sense. Specifically, we think an ideal parameter to implement in a fully market-driven way is interest rates, given their importance within lending protocols and the large positive impact this change could have for both lenders and borrowers.
To achieve the above objectives we propose an overcollateralized, zero coupon bond-based P2P lending protocol (inspired by OG fixed lending protocols like Yield and Notional on Ethereum and newer players like Jet v2 on Solana). Let’s unpack what we mean. The heart of the protocol would be zero-coupon bonds and the system to mint and redeem them. Zero-coupon bonds are a debt instrument that do not pay interest but rather trade at a discount, generating a profit for their holder at maturity when they are redeemed.
Within the proposed protocol, to mint these bonds (which would be done by interested borrowers), users would need to lock collateral in ratios pre-defined by governance. After minting them, these users would be able to sell these bonds on the market* for the token they were looking to borrow. Lenders, on the other hand (who are the buyers of these bonds on the market), would receive these bonds and be able to exercise them for their face value at maturity.
Let’s explore an example:
- Alice wants to borrow 1,000 USDC for 1 year.
- The current market price for 1,000 USDC is 1,100 1-year USDC bonds.
- Alice provides enough ATOM collateral to mint 1,100 1-year USDC bonds.
- She then proceeds to sell them in the market* for 1,000 USDC.
- Bob, who wanted to lend 1,000 USDC, had previously placed a limit order on the market to buy 1,100 1-year USDC bonds for 1,000 USDC.
- When Alice sells the bonds, Bob receives the 1,100 1-year USDC bonds which, at maturity, would be redeemable for their face value: 1,100 USDC. If we assume this transaction happened at the start of the 1-year period, then the implied interest rate for this loan would be 10%.
At maturity, borrowers would have a window to repay their loan or the system could automatically roll their maturity one additional period. This mechanism would guarantee that there’s always the necessary principal for lenders to redeem. Note that this system is overcollateralized and borrowers are still subject to liquidations whenever their collateralization ratio drops below certain governance-defined thresholds.
There are two main components that must be built to achieve the desired design:
- A bond minting mechanism that accepts collateral and handles liquidations when collateral value falls relative to what is borrowed. We prefer to leverage existing risk frameworks and liquidation infrastructure of e.g. Mars protocol, as the overlap in functionality is significant.
- A venue for price discovery of the bonds as time passes and they approach maturity. We believe the ideal venue here is a hybrid orderbook and AMM. Orderbooks would allow lenders to granularly express their views about interest rates. However, a weakness of this design is that it would almost certainly require sophisticated, active market-makers to help supply liquidity since they must constantly move their midquote as time passes and bonds approach liquidation. Algorithmically supplied liquidity would help bootstrap this and reduce reliance on centralised MMs.
Where: TBD.
Hackathon MVP: The bond minting/redemption system for one principal, one maturity and one collateral asset. The Rollover mechanism would not be part of the MVP.