Cascade protection
This is the systemic argument for locking. Locked positions don't just earn you a bonus — they protect everyone in the pool from liquidation cascades. Whether they protect you depends on aggregate adoption, which is a coordination game.
What's a cascade?
A liquidation cascade is the feedback loop:
- Prices drop. Some borrowers' health factors fall below 100%.
- Liquidators clear those positions, taking collateral.
- Liquidators sell the collateral on the open market for stablecoins.
- The selling pressure pushes prices lower.
- More borrowers' positions become underwater.
- Loop.
Cascades have been responsible for the largest losses in DeFi history — March 2020's "Black Thursday" cost MakerDAO $8.3M (USD) in bad debt; numerous other protocols have lost large amounts in similar circumstances.
How locks attenuate the loop
Locked supply can't be redeemed for the underlying asset. When a locked position is liquidated, the position tokens transfer to the liquidator — but the underlying tokens stay in the vault.
This breaks step 3 of the loop. The liquidator can hold the locked tokens until expiry, or sell them at a discount on the secondary market — but they can't immediately dump them on the spot market. The selling pressure that drives the cascade is bounded by the unlocked fraction of the pool.
The protocol formalises this as the Cascade Attenuation Theorem:
In a pool where fraction ϕ of supply positions are permanently locked, the maximum cascade amplification factor is bounded by (1 − ϕ).
So:
- 0% locked → cascade unaffected (same as traditional protocols).
- 50% locked → cascade depth halved.
- 100% locked → cascade essentially eliminated.
Simulation results
The whitepaper simulates 1,000 borrowers under different lock fractions and price shocks. Selected results (linear price impact, k = 1.5×10⁻⁴):
| Price shock | 0% locked | 25% locked | 50% locked | 75% locked | 100% locked |
|---|---|---|---|---|---|
| 10% | 8.6% | 8.2% | 7.8% | 7.8% | 7.4% |
| 15% | 19.8% | 16.5% | 14.4% | 13.5% | 12.7% |
| 20% | 36.5% | 30.6% | 26.7% | 22.4% | 19.8% |
| 25% | 85.7% | 55.8% | 37.1% | 33.3% | 29.4% |
| 30% | 99.1% | 84.9% | 61.2% | 47.9% | 38.6% |
The dramatic difference at the 25% shock — 85.7% vs 29.4% — illustrates the kind of regime change locks can create. Cascade dynamics are non-linear: a small attenuation can prevent the runaway loop.
The coordination problem
Cascade protection is a public good. Locking your position protects the whole pool, including everyone who didn't lock. So why would you lock?
Three answers:
- The bonus. Locked suppliers earn extra interest; locked borrowers pay less. The protocol pays you to provide the public good.
- Self-interest in stable markets. Cascades hurt everyone with collateral exposure. Even from a purely selfish view, supporting cascade protection is rational if the bonus exceeds the discount cost.
- Counter-cyclical equilibrium. When utilization is high (and cascade risk is highest), the bonus is most attractive. Lock adoption rises naturally during periods when it matters most.
What governance can and can't do
Governance can adjust the bonus (LOCK_BONUS) and the malus (LOCK_MALUS), within the constraint that both are bounded by the spread s. Increasing the bonus encourages more locking but reduces protocol margin.
Governance cannot force users to lock. The mechanism is purely incentive-based.
Governance also cannot force-unlock anyone — locks are credible commitments, not parameters that can be retroactively changed.
What it doesn't fix
Three caveats:
- Cascades among unlocked positions still happen. The (1 − ϕ) bound applies to the unlocked fraction. If 20% of the pool is locked and 80% is unlocked, cascade dynamics among that 80% are unchanged.
- Adoption is not guaranteed. The Nash equilibrium analysis identifies two self-reinforcing equilibrium regions — low adoption (locked fraction
< 15%) and high adoption (> 40%) — with selection depending on path. A protocol that bootstraps poorly may stay in the low equilibrium. - Locked supply tokens received in liquidation create their own headache for liquidators. A liquidator who absorbs a fully-locked victim adds debt without gaining redeemable collateral. This may slow the rate of liquidation, which is desirable for cascade protection but reduces keeper economics. See For keepers.
Where to go next
- Bonus and malus — what the protocol pays you to lock
- Theory & proofs — the formal analysis of the equilibrium
- Liquidation risk — what cascade protection means for your individual position