In the continuous pursuit of enhancing the security of FiRM, we are introducing a new security feature: Minimum Debt Amounts. This innovative feature is designed to further safeguard the protocol by ensuring that all new borrow positions remain liquidatable, even in high gas environments. Importantly, we want to reassure our existing borrowers that this will not impact their current loans or positions. This document outlines the rationale behind this feature, its benefits, and the methodology behind determining market minimum debt amounts.
In lending protocols like FiRM, the management of borrow positions is a critical aspect of ensuring the system's stability and security. Borrow positions are created when users borrow assets against their collateral, and these positions must remain liquidatable to protect the interests of both borrowers and lenders. Presently on FiRM, users have the flexibility to establish debts of any size. While this flexibility empowers users, it introduces potential challenges related to the profitability of liquidations, especially in scenarios with high gas prices.
Profitability of Liquidations: Liquidations are a fundamental mechanism for maintaining the health of a lending protocol. When a borrower's position becomes undercollateralized, it needs to be liquidated to repay the lender and protect the system. Without a minimum debt threshold, borrowers may have debts that are so small that liquidations would not yield a profit after accounting for gas fees, potentially leaving undercollateralized positions unattended.
Gas Price Sensitivity: Gas prices on the Ethereum network can fluctuate significantly, and during periods of high congestion, executing transactions becomes costly. In such high gas price environments, the profitability of liquidations becomes even more critical. If liquidations do not yield a profit, liquidators are less likely to engage, potentially causing delays in addressing undercollateralized positions.
Borrower Behavior: The flexibility for users to establish debts of any size also raises concerns about borrower behavior. Some users might engage in strategies that result in very small debts, which can be detrimental to the overall efficiency and security of the protocol. For instance, they may repeatedly borrow and repay small amounts, potentially leading to excessive gas costs and making liquidations economically unviable.
To address this concern, we have already implemented checks in the borrowController to ensure that a user's borrow amount plus debt results in an adequate debt size. However, a critical gap remains. A mechanism that would guarantee the profitability of liquidations, even in high gas environments, would address these concerns. Enter Minimum Debt Amounts. Each market can now independently set a minimum debt threshold, providing a clear safeguard against unprofitable liquidations and low-value debts.
The proposed solution is to shift the responsibility of maintaining a healthy debt size from the borrowController to the Market contracts. Under this new framework, each market can have an independently set minimum debt threshold. This threshold will serve as a safeguard against unprofitable liquidations and low-value debts.
When a user initiates a borrow transaction, the market will now check that the borrow amount, when added to their existing debt, exceeds the minimum debt threshold specific to that market.
This feature may require adjustments to existing helper functions used for debt management. We will provide clear guidance and support for developers and users to ensure a smooth transition.
Inverse Finance’s Risk Working Group has conducted a comprehensive analysis to determine the minimum debt amounts in FiRM, the methodology for which is outlined below.
In determining this, one potential consideration is how this feature may impact the behavior of smaller borrowers. While the minimum debt thresholds are designed to safeguard liquidations, they should be set at levels that do not disproportionately affect smaller borrowers. Striking the right balance between security and user-friendliness is crucial.
Gas Spent by Liquidators
We initiated our analysis by estimating the total gas incurred by a liquidator during the process of liquidating a borrower's position in each of the 10 active markets on FiRM. For better accuracy, we assumed this liquidation would be carried out by a bot, and thus would follow a typical three step approach: Acquiring DOLA starting with ETH, performing the liquidation, and finally selling the liquidated asset for ETH.
To estimate the gas spent to liquidate a borrower's position on FiRM we ran multiple simulations on Tenderly adhering to the following steps:
Simulate a user with an existing position, borrowing the maximum possible amount of DOLA.
Deplete the DBR balance in their wallets.
Replenish the user's DBR balance to bring their debt to an unhealthy level.
Execute the liquidation of the borrower's position.
Cost of Liquidation
The total gas spent incurred by the liquidator is the sum of each of the three steps detailed above. To standardize the cost of liquidation, the total gas spent (in gas units) was converted into the Cost of Liquidation, measured in ETH. This conversion was accomplished by dividing the total gas spent by a fixed gas price (measured in gwei) and then multiplying by 10^9 to ensure consistency in units. Our analysis assumed a gas price of 50 gwei, which is twice more than average gas cost since The Merge (see below).
Screenshot from UltraSoundMoney.com
This was then converted to a Cost of Liquidation measured in USD, assuming a price of ETH that reflects the market conditions at the time of the analysis. The end result was an estimated cost of liquidation for each of the 10 markets on FiRM, measured in USD.
Determining Minimum Debt Amounts
With the estimated cost of liquidation for each market and given the market-specific parameters of Liquidation Factor (%) and Liquidation Incentive (%), we proceeded to determine the minimum viable debt amount utilizing the formula:
The objective was to identify the minimum debt threshold that guarantees a profitable liquidation, assuming the liquidation occurs with an assumed Gas Price (in GWEI) and at a given price of ETH. This process resulted in unique minimum debt amounts for each market.
Our preliminary analysis for variable minimum debt amounts leads us to derive the minimum debt amount (in DOLA) for each of the 10 FiRM markets, given the following fixed parameters:
ETH Price (USD): 2000
Gas Price (GWEI): 50
Liquidation Factor: Market-specific, see below
Liquidation Incentive: Market-specific, see below
These minimum debt amounts are tailored to each market's unique characteristics and are set to guarantee profitable liquidations under the given fixed parameters.
However, we can drive this study one step further... In the subsequent analysis for fixed minimum debt amount, our objective is to determine new Liquidation Factors for each of the existing markets on FiRM, given a fixed Minimum Debt Amount of 2000 or 3000 DOLA (depending on the cost of liquidation) and the following fixed parameters:
ETH Price (USD): 2000
Gas Price (GWEI): 50
Liquidation Incentive: Market-specific, see below
In doing so, we are setting a data-driven foundation for determining liquidation factors for individual markets on FiRM. This leads to the following results:
The resulting new liquidation factors are on average lower than what’s presently set. Five of the ten markets would see reductions, leading to a better user experience for users of said markets. The remaining 5 (DAI, CRV, cvxCRV, st-yCRV, and INV) would see increases, and be further secured as a result. However, given borrowing in CRV, cvxCRV, and st-yCRV markets is paused, the RWG would opt to leave parameters for these unchanged. At the same time, given INV liquidity is shallow and borrowing is restrictive, the RWG recommends leaving the liquidation factor for this market unchanged as well. Keeping liquidation factor at 20% (vs. 50.3%) until liquidity deepens substantially greatly reduces the odds of a liquidation cascade, whilst still ensuring a healthy liquidation mechanism.
Given all the above, the RWG favored the option of implementing a uniform minimum debt amount across all markets for the sake of simplicity and effective communication to our users. While market-specific minimum debt amounts offer fine-tuned risk management, a uniform approach streamlines user experience and simplifies our risk administration efforts.
In conclusion, this risk assessment provides a data-driven foundation for determining new Liquidation Factors for each FiRM market as well as setting minimum loan amount of 2000 or 3000 DOLA, depending on market. By ensuring that all new borrow positions are profitable to liquidate, we strengthen the protocol's security, enhance user confidence, and make FiRM a more robust and resilient platform. We are committed to ongoing collaboration with the community to fine-tune and optimize this feature for the benefit of all users. Market conditions, gas prices, and user behaviors can change, and the protocol will evolve to meet these challenges.