Risk Methodology
Last updated
Was this helpful?
Last updated
Was this helpful?
Dahlia implements an effective risk parameter methodology with the objective to mitigate the Liquidity Risk and Bad Debt Risk inherent in DeFi lending markets and enable efficient and safe lending markets. The methodology therefore assesses a number of historical data points related to the collateral and loan asset and derives appropriate risk parameters with which a Dahlia lending market is configured.
It is important to understand that Dahlia lending markets are immutable, meaning that their risk parameters are not actively managed but are static. As market conditions change, new lending markets may be created with different risk parameters requiring lenders and borrowers to monitor markets and reallocate positions if needed.
Furthermore, lenders and borrowers should be aware of other types of risks, such as Smart Contract risk, which are discussed here.
The image shows a flow diagram of Dahlia’s risk parameter methodology. It illustrates how both the lending and borrowing assets are evaluated based on a qualitative Asset Rating methodology. The individual asset ratings then feed into a Market Rating capturing the overall risk profile of a Dahlia lending market.
In parallel, quantitative Risk Factors are extracted from relevant historical market data for each lending market, or collateral - loan asset pair. These risk factors are then combined with the market rating to compute the Risk Parameters with which a Dahlia lending market is ultimately equipped.
Assets are classified based on five dimensions: Market Capitalization, Number of Holders, Age of the Asset in terms of price history, 24h trading volume and 24h price volatility. These dimensions combined should allow for a qualitative assessment of the overall maturity and quality of an asset. The Asset Rating is therefore assigned on a four-levels scale ranging from A, the highest quality, to D, the lowest quality.
💡 The Asset Rating captures the overall maturity and quality of collateral and loan assets on a scale of A, the highest quality, to D, the lowest quality.
Dahlia lending markets consist of two assets, the collateral asset and the loan asset. The Market Rating combines the two asset’s individual ratings into a single market rating that captures the market’s overall quality based on the same four-level scale as shown in the table below.
Furthermore, the Market Rating is associated with a Risk Multiplier which serves as a “confidence score” and is used in the computation of the market’s Risk Parameters.
A
If both underlying assets have a rating of A
1
B
If both underlying assets have a rating of at least B
1.5
C
If both underlying assets have a rating of at least C
2
D
If at least one underlying asset has a D rating
3
💡 The Market Rating combines a market’s individual Asset Ratings into a single rating reflecting the lending market’s overall maturity and quality.
The market risk associated with a lending market can be assessed quantitatively in a number of Risk Factors. These factors directly relate to the market’s Bad Debt risk in that they serve as measures for
the likelihood of a market’s relative collateral value to fall to a level where positions would become insolvent (that is, the value of the collateral turns less then the position’s debt given a certain Liquidation Loan-to-Value)
the likelihood of liquidations turn unprofitable for liquidators and thus not being processed anymore
For each market, we therefore collect historical market data to compute the following risk factors:
Measures the “average” price changes within a full day. This translates to the “expected” collateral value drops within a day and thus is inversely related to the market’s LLTV.
Measures uncommon, yet possible, “extreme” price changes within a full day. As such a market’s MDD is always larger then the volatility. The measure translates to the expected “worst case” collateral value drops within a day and thus too is inversely related to the market’s LLTV.
Measures the daily volume of trades in the respective trading pair processed on DEXs. Higher volume translates to more on-chain liquidity and means that liquidations, and in particular atomic, flashloan-powered liquidations, can be processed more efficiently (that is faster and with less slippage). This risk factor thus translates to the liquidation bonus paid to liquidators on a lending market: the higher the available liquidity, the lower the liquidation bonus has to be for timely liquidations to be profitable for liquidators.
Finally, let’s discuss the various risk parameters derived from a lending market’s rating and risk factors. We distinguish two risk parameter sets: the Interest Rate Model Parameters and the Liquidation Model Parameters.
Dahlia uses an adaptive interest rate model that adjusts a market’s interest rate based on short-term utilization changes (regular rate curve), and based on long-term shifts in the market’s supply and demand for the loan asset (curve controller). You can read more about this model here.
The objective of this model is to ensure the market can autonomously and efficiently find an equilibrium borrowing cost while making sure an appropriate level of liquidity is available for lenders to withdraw at any time. It is therefore configured with a number of parameters summarized here below:
Target Utilization: Defines the equilibrium utilization of liquidity in the market and directly translates to the target liquidity reserves available for lenders. Since on Dahlia lending markets the loan and collateral assets are isolated, i.e. rehypothecation is not enabled by design, target utilization can be set relatively high (often at 90%) resulting in high capital efficiency and supply APYs.
Zero Utilization Rate: The interest rate applied for zero utilization in the market. This is a fixed parameter and defines a static lower-end of the utilization-rate curve. It is usually defined as 0%.
Minimal and Maximal Full Utilization Rate: These define the range within which the interest curve controller can adjust a market’s full utilization rate, that is the interest rate for 100% utilization, in any given block. Since the market participants, through their lend & borrow activity, autonomously shift the market’s full utilization rate to a level in this range, the limits of the range are less critical to configure. Nonetheless, the approach taken here is to reflect the Market Rating where a higher rating allows a market to arrive at lower interest rates compared to a market with a lower rating.
Dahlia lending markets implement a simple, yet robust and efficient, liquidation model allowing full liquidations of insolvent positions with a fixed liquidation bonus (more on this here). This model is based on two risk parameters:
Liquidation Loan-to-Value (LLTV): this parameter defines the value a position’s debt cannot exceed relative to it’s collateral value. If a position’s debt value exceed this threshold, the market’s LLTV, it can be liquidated by anyone. The market’s LLTV must be less then 1.0 (100%) and thus includes a buffer safeguarding for expected market price movements during a day. The LLTV is thus derived from the market’s 24hrs Volatility, 24hrs MDD and Market Rating.
Liquidation Bonus: this parameter defines the bonus paid to a liquidator as an incentive to repay an insolvent position’s debt in exchange to receive parts of its collateral. Specifically, the amount of collateral received is computed based on a discounted market price of the collateral where the discount is just the liquidation bonus. Hence ,the liquidation bonus should compensate the liquidator for any slippage of selling the collateral on a DEX and add an incentive on top for timely liquidations. The liquidation bonus is thus computed based on the assumed slippage realized in a “significant” liquidation event (which is defined as the liquidation of 10% of the market’s total debt). Thereby, this slippage is estimated based on the market’s 24hrs DEX Volume risk factor.
