FIAT I was launched as a debt-based stablecoin protocol collateralized by stable fixed term assets. It sought to unlock liquidity for the discounted value of these locked positions. Its launch, however, coincided with an exodus of liquidity that has led to a collapse in DeFi fixed rates. At the same time, the transition of Ethereum to Proof-of-Stake and the wider proliferation of liquid staking derivatives for native network assets has grown the universe of high quality on-chain derivative collateral assets.
FIAT II initially referred to our work on building out a liquidation mechanism that would allow the protocol to onboard volatile fixed term assets as collateral. We realized that our solution to dealing with illiquid, volatile collateral was extensible to any variety of lending arrangements. Coupled with the lack of profitable releveraging opportunities within the handful of fixed rate markets present in DeFi, the vision for a permissionless credit marketplace came into view. FIAT II thus represents an extension of our original FIAT I architecture, itself heavily influenced by MakerDAO’s DSS, coupled with a far more drastic pivot in terms of its total addressable market.
We introduce the concept of a permissionless credit marketplace powered by a dollar-pegged stablecoin. FIAT II maintains protocol solvency through the use of delegated minting capacity for the liquidation of at-risk debt positions. Minting capacity delegation occurs between an elected set of high quality liquid reserve assets **and an unbound set of borrowing vaults **such that “bad” debt can fall back on “good” collateral. Credit delegation is coordinated by delegators seeking out the highest risk-adjusted returns in exchange for underwriting liquidations. We further demonstrate that the cost of minting the stablecoin can thus be derived as a function of the utilization of minting capacity by associated borrowing vaults.
MakerDAO introduced the concept of a debt-based stablecoin backed by exogenous collateral asset(s). Key to the solvency of the protocol is the liquidation of user positions that fall below a minimum collateralization ratio. Its solution, public dutch auctioning of collateral, assumes that there is a sufficiently-liquid secondary market for the collateral asset in order to attract profit-seeking arbitrageurs. This solution cannot preclude the accrual of bad debt, or user positions that fall short of the minimum collateralization ratio, within approved collateral vaults that do not meet this requirement at the scale of their corresponding debt ceilings.
Liquity, another debt-based stablecoin protocol, later introduced the concept of the stability pool as an alternative liquidation mechanism. The stability pool conducts liquidations of user positions that fall below the minimum collateral ratio with a staked stablecoin, specifically LUSD. This solution has found market fit with users keen on purchasing the protocol’s sole collateral asset, Ether, when it experiences high bouts of volatility and liquidations are more likely. However, it is unclear whether users would be willing to bear the opportunity cost of depositing into the stability pool for collateral assets without similar expectations around future price volatility.
FIAT II implements a hybrid solution for liquidating at-risk positions. Instead of requiring users to source already-minted stablecoins in order to participate in automated liquidations, it allows them to commit protocol credit in advance. Users who deposit governance-approved reserve assets **are able to delegate their protocol credit across any number of borrowing vaults, or stablecoin-minting facilities. In the event that a borrowing vault position is flagged as at-risk, credit delegates take on both its outstanding debt and the collateral securing it. They are compensated for their service with a proportionate claim on the accrued interest, liquidated collateral assets, and any subsidies targeting the borrowing vault. The result is a more robust overall collateral composition for the protocol, as the burden of debt repayment is shifted to depositors of reserve assets which themselves can be liquidated via public dutch auction as needed. This allows FIAT II to retain a conservative collateral risk profile even as it extends minting privileges to less liquid assets, as well as outright unsecured or semi-secured borrowers.
To wit, FIAT II offers a market-driven alternative to existing stablecoin and lending protocols which have historically relied on manual governance for collateral onboarding. Its reduction of human intervention allows for a more scalable stablecoin supply while also opening the door for unsecured lending. The remainder of the document describes the processes of credit delegation and borrowing vault liquidation in greater detail; discusses the role of delegated credit (over)utilization in the context of algorithmic interest rate adjustments; and speculates on further potential protocol composability.
FIAT II is an extended collateralized debt position-based (CDP) stablecoin protocol. A typical CDP protocol will entrust the valuation of collateral assets to a permissioned set of smart contracts known as collateral vaults. The protocol can then assign credit and debt balances denominated in the native stablecoin to users who deposit accepted collateral assets and subsequently utilize their minting capacity. Historically, such protocols have not made full use of internal credit transfers between user collateral vault positions. FIAT II does.
FIAT II introduces the concept of reserve vaults. A reserve vault allows users to deposit a specific on-chain asset and create a collateralized credit position (CCP). Each reserve vault is parameterized with key values, namely a credit ceiling ($c^+$) and a minimum collateralization ratio ($r^-$) , the latter of which is then enforced via an approved price feed oracle. Reserve vault assets can be either “stable” or volatile, interest-bearing or not, so long as there is a reasonable expectation of cash-approximate convertibility at the scale of its corresponding credit ceiling. Unlike a traditional collateral vault, a reserve vault cannot itself mint stablecoins and instead can only delegate its determined credit.
The core purpose of delegated credit is to execute liquidations throughout the FIAT II protocol. It is thus imperative that protocol credit remain wholly collateralized by deposited reserve assets. Each reserve vault position ($i$) can thus be characterized by its credit collateralization ratio ($r_i$), or the ratio between collateral value ($a_i$) and associated delegated credit ($c_i$). So long as a position does not violate the minimum collateralization ratio of the reserve vault, it does not pose a risk to protocol solvency.
$$ r_i = \frac{a_i}{c_i}\geq r^- $$