Soft Liquidation
Soft liquidation is a gradual collateral conversion mechanism that automatically swaps between collateral and debt assets as prices move, avoiding the abrupt selloff of traditional hard liquidation.
Key Takeaways
- Soft liquidation gradually converts collateral into the borrowed asset as the collateral price falls, then reverses the conversion if the price recovers. This replaces the all-or-nothing approach of traditional hard liquidation.
- The primary implementation is Curve's LLAMMA (Lending-Liquidating AMM Algorithm), which spreads a borrower's collateral across a range of price bands and uses an automated market maker to swap between collateral and debt within each band.
- Soft liquidation is not free: repeated conversions during volatile, sideways markets erode position value through arbitrage spreads and fees, and health factor can still reach zero, triggering a full hard liquidation.
What Is Soft Liquidation?
Soft liquidation is a DeFi lending mechanism that gradually converts a borrower's collateral into the debt asset as the collateral price declines, rather than selling the entire position at once with a penalty. If the price recovers, the mechanism reverses: it converts the debt asset back into collateral, a process known as de-liquidation. The borrower's loan stays open throughout.
In traditional lending protocols like Aave or Compound, a position that breaches its liquidation price is immediately closed by a third-party liquidator who repays the debt and claims the collateral at a discount (typically 5-15%). This hard liquidation is fast but punishing: borrowers lose a significant portion of their collateral in a single transaction and have no opportunity to recover if prices bounce back.
Soft liquidation was designed to solve this problem. By converting collateral incrementally across a defined price range, it gives borrowers a buffer zone where their position absorbs price declines without being forcibly closed. The concept was introduced by Curve Finance as part of its crvUSD stablecoin system and the LLAMMA algorithm.
How It Works
Soft liquidation relies on an automated market maker (AMM) that sits between the borrower's collateral and the debt asset. When a user opens a loan, their collateral is deposited into this AMM and distributed across a series of discrete price bands. As the market price moves through these bands, the AMM automatically swaps collateral for the debt asset (or vice versa).
The Bands System
When a borrower opens a position, LLAMMA distributes their collateral across a configurable number of price bands (the default is 10, adjustable from 4 to 50). Each band covers a narrow price range determined by the protocol's amplification parameter (A). The width of each band follows the formula:
band_width = (1 / A) * collateral_price
Example with A = 100 and collateral price of $3,000:
band_width = (1/100) * $3,000 = $30 per band
10 bands cover a $300 range (roughly 10% of the price)Above the top band, the position holds 100% collateral. Below the bottom band, the position holds 100% debt asset. Within the bands, the position is a mix of both, with the ratio shifting as the price moves.
Conversion Process
As the collateral price falls into a band, the AMM begins selling collateral for the debt asset. This happens through arbitrageurs who trade against the AMM:
- The collateral price drops, creating a discount between the AMM's internal price and the external market price
- Arbitrageurs buy the discounted collateral from the AMM, depositing debt tokens (such as crvUSD) in return
- The band's composition shifts from collateral-heavy to debt-heavy
- If the price continues falling, the process repeats in the next lower band
This is the soft liquidation phase. The borrower's position is being partially liquidated, but the loan remains open and no liquidation penalty is charged.
De-Liquidation (Price Recovery)
If the collateral price rises back through the bands, the reverse happens: the AMM sells the debt asset and buys back collateral. Arbitrageurs again capture a small spread on each trade, but the borrower's position reconverts toward its original collateral composition.
This reversibility is a defining feature. In a hard liquidation model, once a position is closed, the borrower must re-enter the market at current prices. With soft liquidation, a temporary price dip followed by recovery leaves the borrower with their loan intact, though slightly reduced by conversion costs.
Hard Liquidation as Backstop
Soft liquidation does not eliminate the risk of total loss. Every position still has a health factor that accounts for the current value of the collateral mix, accumulated fees, and outstanding debt. If health drops to 0%, a hard liquidation occurs: any user can repay the borrower's remaining debt and claim whatever collateral remains. This acts as the final backstop against bad debt.
LLAMMA: The Primary Implementation
The Lending-Liquidating AMM Algorithm (LLAMMA) is the engine behind Curve's soft liquidation system. It was introduced alongside crvUSD and is also used in Curve's LlamaLend markets.
LLAMMA differs from a standard AMM in several ways:
- Liquidation is per-band, not per-user: the AMM processes conversions at the band level, making the system scalable regardless of the number of borrowers
- External oracle drives the target price: unlike a traditional AMM where the price is determined entirely by pool reserves, LLAMMA uses an external oracle price to set the equilibrium, and arbitrageurs push the internal price toward it
- Borrowers choose their band count: more bands spread conversion across a wider price range, meaning each band converts less aggressively but liquidation begins earlier. Fewer bands allow higher loan-to-value ratios (approximately 93% with 4 bands) but concentrate conversion into a narrower range
Band Count Tradeoffs
| Parameter | Fewer Bands (4) | More Bands (50) |
|---|---|---|
| Liquidation start | Later (closer to liquidation price) | Earlier (wider range above liquidation price) |
| Conversion per band | More aggressive | More gradual |
| Max LTV | Higher (approximately 93%) | Lower |
| Exposure to churning losses | Lower (narrower range) | Higher (wider range, more time in soft liquidation) |
Use Cases
Leveraged Positions with Downside Buffer
Traders who want leveraged exposure to an asset can use soft liquidation to create a graduated exit rather than a cliff. If the market dips temporarily, their position converts partially but recovers when the price bounces. This is particularly useful during flash crashes or short-term volatility where hard liquidation would be premature.
Stablecoin Minting
Curve uses soft liquidation as the core mechanism for crvUSD issuance. Borrowers deposit collateral (ETH, WBTC, or other assets) and mint crvUSD against it. The soft liquidation bands ensure that the stablecoin remains backed even during market downturns, gradually converting collateral into crvUSD to maintain solvency.
Reducing Liquidation Cascades
In hard liquidation systems, a sudden price drop can trigger a cascade: mass liquidations dump collateral onto the market, driving prices lower and triggering more liquidations. Soft liquidation spreads selling pressure over time and across price levels, reducing the feedback loop that causes cascades.
Risks and Considerations
Churning Losses in Sideways Markets
The most significant risk of soft liquidation is value erosion during volatile, sideways price action. When the collateral price oscillates through the bands, the AMM repeatedly converts collateral to debt and back again. Each conversion incurs a small loss because arbitrageurs capture a spread on every trade.
These losses compound over time. A position that stays in the soft liquidation range for days or weeks during choppy markets can lose meaningful value, even if the collateral price ultimately returns to its starting point. Borrowers should monitor their health factor closely during extended periods of soft liquidation.
Not a Safety Guarantee
Soft liquidation delays and softens the impact of price declines, but it does not prevent liquidation entirely. If the collateral price falls far enough and fast enough, health can still reach 0% and trigger a hard liquidation. Borrowers who mistake soft liquidation for liquidation immunity may take on excessive risk.
Oracle Dependency
LLAMMA relies on an external price oracle to determine the target price for each band. If the oracle is delayed, manipulated, or experiences downtime, the AMM may misprice conversions. This is the same oracle risk present in all DeFi lending protocols, amplified by the continuous nature of soft liquidation.
Complexity for Borrowers
Soft liquidation introduces additional parameters (band count, amplification factor, health calculation) that borrowers must understand to manage their positions effectively. The reversible nature of the mechanism can also create a false sense of security: the fact that a position recovered from soft liquidation once does not guarantee it will recover again.
Soft Liquidation vs. Hard Liquidation
| Feature | Soft Liquidation | Hard Liquidation |
|---|---|---|
| Trigger | Price enters band range | Health factor reaches threshold (or 0% in Curve) |
| Conversion speed | Gradual, across price bands | Immediate, full position |
| Penalty | No explicit penalty (but conversion losses occur) | Explicit discount (typically 5-15%) |
| Reversibility | Fully reversible if price recovers | Irreversible: position is closed |
| Loan status | Remains open | Closed |
| Cascade risk | Lower: selling pressure is distributed | Higher: mass sell events compound |
| Best for | Temporary dips, volatile markets | Decisive, sustained price declines |
Why It Matters
Soft liquidation represents a fundamental shift in how DeFi lending handles undercollateralized positions. By replacing the binary liquidate-or-not decision with a continuous conversion process, it aligns more closely with how traditional margin systems work, where positions are gradually reduced as they approach insolvency rather than immediately closed.
For the broader DeFi ecosystem, soft liquidation reduces the systemic risk of liquidation cascades, which have historically caused flash crashes and protocol insolvencies. As overcollateralized lending grows in scale, the choice between soft and hard liquidation models will increasingly affect protocol resilience and borrower experience.
For a deeper look at how stablecoins like crvUSD use these mechanisms, see the research article on stablecoin peg mechanisms and the DeFi protocol risk analysis.
This glossary entry is for informational purposes only and does not constitute financial or investment advice. Always do your own research before using any protocol or technology.