Simple Summary
Gauntlet’s analysis of the Compound v3 absorption and liquidation mechanism reveals that:
- The Compound v3 liquidation mechanism presents more risk than the v2 liquidation mechanism, but also greater upside.
- In v3, a collateral asset’s Liquidation Penalty (LP), combined with the global Store Front Price Factor, is more crucial than its Liquidation Factor (LF) in safeguarding against insolvency.
- In v3, for many assets, LPs and LFs can be increased, for better risk management and capital efficiency. Factors to consider when determining which levers to increase include the asset’s liquidity, supply cap, and empirical demand.
Analysis
Compound v2’s Liquidation Process
In Compound v2, the liquidation process is designed to minimize the risk of insolvency by providing a cushion through the Collateral Factor (CF). A lower CF creates a larger buffer against price drops, protecting the protocol, while a higher CF increases capital efficiency but offers less of a buffer.
Assume the following two scenarios:
| Parameter | Scenario A (Moderate WBTC CF) | Scenario B (Elevated WBTC CF) |
|---|---|---|
| WBTC Collateral Factor | 70% | 90% |
| WBTC Liquidation Penalty | 5% | 5% |
| Close Factor | 50% | 50% |
Let’s examine how a lowly collateralized user position is liquidated under both scenarios:
Block 1: Initial State (WBTC Price at $100):
| Action | Scenario A (WBTC CF 70%) | Scenario B (WBTC CF 90%) |
|---|---|---|
| WBTC Supplied | $100 | $100 |
| USDC Borrow Limit | $70 (70% of supplied WBTC) | $90 (90% of supplied WBTC) |
| USDC Borrowed | $70 | $90 |
Block 2: Subsequent State (WBTC Price drops by 2% to $98, triggering liquidation):
| Action | Scenario A (Lower WBTC Collateral Factor) | Scenario B (Higher WBTC Collateral Factor) |
|---|---|---|
| WBTC Supply | $98 | $98 |
| USDC Borrow Limit | $68.60 (70% of $98) | $88.20 (90% of $98) |
| USDC Borrowed | $70 | $90 |
| WBTC Price Drop to cause insolvency | 28.6% | 8.2% |
| USDC Repaid by Liquidator | $35 | $35 |
| WBTC Received by Liquidator | $35 * 1.05 = $36.75 | $35 * 1.05 = $36.75 |
| New WBTC Supply | $61.25 | $61.25 |
| New USDC Borrowed | $35 | $55 |
| New WBTC Price Drop to cause insolvency | 42.9% | 10.2% |
In both scenarios, a subsequent significant drop in WBTC price presents risks, which are then mitigated by a liquidator. Scenario A (lower CF) has a greater buffer (42.9%), thus providing more protection. Scenario B (higher CF), while more capital efficient, has a thinner margin (10.2%) for further price drops, increasing the risk of insolvency.
Compound v3’s Liquidation Process
Compound v3 liquidation mechanism differs from the one in v2. In v3, the LF determines when an account is eligible to be absorbed. When an account is absorbed, the entire collateral is taken over by the protocol, which then settles the user’s debt based on the LP. The absorbed collateral is then put up for sale, with a sale price determined by the LP and the Store Front Price Factor.
The following analysis uses similar LF comparisons as the previous v2 CF example.
| Parameter | Scenario A (Moderate WBTC LF) | Scenario B (Elevated WBTC LF) |
|---|---|---|
| WBTC LF | 70% | 90% |
| WBTC LP | 5% | 5% |
| Store Front Price Factor | 60% | 60% |
Let’s examine how a lowly collateralized user position is absorbed under both scenarios:
Block 1: Initial State (WBTC Price at $100):
| Action | Scenario A (WBTC LF 70%) | Scenario B (WBTC LF 90%) |
|---|---|---|
| WBTC Supplied | $100 | $100 |
| USDC Borrow Limit | $70 (70% of supplied WBTC) | $90 (90% of supplied WBTC) |
| USDC Borrowed | $70 | $90 |
Block 2: Subsequent State (WBTC Price drops by 2% to $98, triggering absorption):
| Outcome | Scenario A (LF 70%) | Scenario B (LF 90%) |
|---|---|---|
| WBTC Absorbed by Compound | $98 | $98 |
| USDC Owed to Borrower (after LP) | $98 × (1 - 5%) = $93.10 | $98 × (1 - 5%) = $93.10 |
| USDC Previously Borrowed | $70 | $90 |
| Additional USDC Paid by Compound | $23.10 | $3.10 |
| Break-even WBTC Sale Price for Compound | $93.10 | $93.10 |
The example illustrates that when a user’s position is absorbed, the collateral transfers to the comet’s ownership. The absorbed user then receives a USDC payment based on the collateral’s price at the time of absorption, according to the following formula:
In both scenarios, the absorbed collateral was $98, and the Liquidation Penalty was 5%, so the protocol paid the user:
The collateral becomes available for purchase immediately after absorption. The Store Front Price factor determines the percentage of the Liquidation Penalty allocated to buyers of the absorbed collateral.
The protocol sets the sale price of the absorbed collateral according to the following formula:
In both scenarios, the protocol will initially quote the collateral for:
Naturally, the price of absorbed collateral can fluctuate from one block to the next. If the collateral’s price decreases by y% between its absorption and sale, the protocol must meet the following condition to avoid financial loss from the absorption.
Refactoring the equation to solve for y, we get:
In both scenarios, the WBTC LP is 5% and the Store Front Price Factor is 60%, so the WBTC price drop would have to decrease by less than 2.04% between absorption and selling in order to avoid losing money on the absorption. Updating the table:
| Outcome | Scenario A (LF 70%) | Scenario B (LF 90%) |
|---|---|---|
| WBTC Absorbed by Compound | $98 | $98 |
| USDC Owed to Borrower (after liquidation penalty) | $98 × (1 - 5%) = $93.10 | $98 × (1 - 5%) = $93.10 |
| USDC Previously Borrowed | $70 | $90 |
| Additional USDC Paid by Compound | $23.10 | $3.10 |
| Break-even WBTC Sale Price for Compound | $93.10 | $93.10 |
| Post-absorption WBTC Price Drop to cause insolvency | 2.06% | 2.06% |
In both scenarios, irrespective of WBTC LF, Compound pays out the same USDC amount to the borrower post-absorption, and the absorbed WBTC amount remains constant. The break-even sale price for WBTC and the post-absorption WBTC price drop to cause insolvency are also identical. The primary risk differential arises if the WBTC price drop from Block 1 to Block 2 is so significant that the USDC already borrowed exceeds the USDC owed post-absorption. To safeguard against this, the protocol must validate the following at Block 2:
Reformulating this for Block 2 conditions:
Adjusting for a percentage drop (x%) in WBTC value from Block 1 to Block 2:
Simplifying further:
Therefore, given a Liquidation Penalty of 5% and a max WBTC inter-block price drop of 2%, we derive:
Therefore, with a maximum WBTC price drop per block of 2%, and a WBTC LP of 5%, the optimal WBTC LF should be set at 93.1%. A lower LF, while maintaining the same risk level, would reduce capital efficiency.
In v3, as long as the Liquidation Factor remains below or equal to the necessary Liquidation Penalty buffer, reducing the Liquidation Factor further doesn’t enhance insolvency protection for the protocol. In v3, it is the Liquidation Penalty and Store Front Price Factor that primarily guard against insolvency, as evidenced by the breakeven sale price and its role in motivating liquidators to engage in arbitrage.
Additionally, it’s important to observe that in both v3 scenarios, the post-absorption WBTC Price Drop required to trigger insolvency (2.06%) is significantly lower than that in the more conservative v2 scenario (10.2%). This difference arises because a 5% WBTC Liquidation Penalty, reduced by a 60% Store Front Price Factor, offers a smaller insolvency buffer compared to even a 90% Collateral Factor in v2, irrespective of the Liquidation Factor. Consequently, current Liquidation Penalty values in v3, ranging from 5% to 12%, present greater risk to the protocol. However, the absorption mechanism also offers increased potential for gains, as the protocol can accrue positive reserves if the collateral price rises post-absorption.
Updating Compound v3 parameters accordingly
In v3, for many assets, LPs and LFs can be increased, for better risk management and capital efficiency. Factors to consider when determining which levers to increase include the asset’s liquidity, supply cap, and empirical demand. The table below presents the optimal LPs for each asset in the v3 comets with constant LF, alongside the optimal LFs for each asset when maintaining a constant LP.
| Comet | Asset | Current LF | Current LP | Current Store Front Price Factor | Optimized LF (holding LP constant) | Optimized LP (holding LF constant) | Optimized LF Increase (%) | Optimized LP Increase (%) | Max Post-Absorption Price Drop (%) to avoid insolvency | Max Liquidator Slippage (%) |
|---|---|---|---|---|---|---|---|---|---|---|
| Ethereum v3 USDC | WBTC | 77% | 5% | 60% | 93.10% | 21.43% | +16.10% | +16.43% | 2.06% | 3.0% |
| Ethereum v3 USDC | WETH | 90% | 5% | 60% | 93.10% | 8.16% | +3.10% | +3.16% | 2.06% | 3.0% |
| Ethereum v3 USDC | LINK | 85% | 7% | 60% | 91.14% | 13.27% | +6.14% | +6.27% | 2.92% | 4.2% |
| Ethereum v3 USDC | COMP | 70% | 12% | 60% | 86.24% | 28.57% | +16.24% | +16.57% | 5.17% | 7.2% |
| Ethereum v3 USDC | UNI | 75% | 7% | 60% | 91.14% | 23.47% | +16.14% | +16.47% | 2.92% | 4.2% |
| Polygon v3 USDC | WBTC | 75% | 5% | 60% | 93.10% | 23.47% | +18.10% | +18.47% | 2.06% | 3.0% |
| Polygon v3 USDC | stMATIC | 65% | 7% | 60% | 91.14% | 33.67% | +26.14% | +26.67% | 2.92% | 4.2% |
| Polygon v3 USDC | WMATIC | 70% | 7% | 60% | 91.14% | 28.57% | +21.14% | +21.57% | 2.92% | 4.2% |
| Polygon v3 USDC | MaticX | 60% | 7% | 60% | 91.14% | 38.78% | +31.14% | +31.78% | 2.92% | 4.2% |
| Polygon v3 USDC | WETH | 83% | 5% | 60% | 93.10% | 15.31% | +10.10% | +10.31% | 2.06% | 3.0% |
| Base v3 USDC | cbETH | 80% | 7% | 60% | 91.14% | 18.37% | +11.14% | +11.37% | 2.92% | 4.2% |
| Base v3 USDC | WETH | 84% | 5% | 60% | 93.10% | 14.29% | +9.10% | +9.29% | 2.06% | 3.0% |
| Arbitrum v3 USDC (Native) | WBTC | 77% | 5% | 80% | 93.10% | 21.43% | +16.10% | +16.43% | 1.04% | 4.0% |
| Arbitrum v3 USDC (Native) | WETH | 85% | 5% | 80% | 93.10% | 13.27% | +8.10% | +8.27% | 1.04% | 4.0% |
| Arbitrum v3 USDC (Native) | ARB | 60% | 7% | 80% | 91.14% | 38.78% | +31.14% | +31.78% | 1.48% | 5.6% |
| Arbitrum v3 USDC (Native) | GMX | 45% | 7% | 80% | 91.14% | 54.08% | +46.14% | +47.08% | 1.48% | 5.6% |
Deciding whether to raise an asset’s LF, LP, supply cap, or a mix depends on its liquidity and demand within the protocol. Assets that have high demand and low liquidity are good candidates to increase LP, in order to decrease risk. Conversely, assets that have low demand and high liquidity are good candidates to increase LF, in order to increase demand.
Gauntlet will continue refining these parameters, balancing these considerations with empirical data on liquidity and demand.