Skip to main content
Asset: GHO · Class: CDP · Date: 2023-08-15 Worst spread: −550 bps · Duration: ≈ 8 months below $0.99 Detected live: No (pre-Pegana retrospective)

​
Background

GHO is Aave’s native stablecoin, minted by users depositing supported collateral (mostly ETH and aTokens) and paying a borrow rate. At launch it had no Peg Stability Module — the only way to mint was open a CDP, the only way to retire GHO was repay debt. That asymmetry is a structural depeg trap.

​
Trigger

Within weeks of launch GHO was trading 0.97–0.97–0.98 on Curve and Balancer. The 1.5% borrow rate let arbitrageurs mint GHO at 1,sellat1, sell at 0.98, pocket 0.02minusborrowcost—andthatflowsetthefloor.∗∗Withoutasymmetricmint−to−0.02 minus borrow cost — and that flow set the floor. **Without a symmetric mint-to-1 path from anyone holding GHO, the market couldn’t reach back up.**

​
Cascade

  • Curve’s GHO/3CRV pool stayed permanently imbalanced — depositors collected fees but couldn’t restore the peg.
  • Aave’s borrow rate was too low to choke off new supply; raising it would have hurt legitimate borrowers.
  • Each new mint cap increase moved the price lower as fresh GHO hit thin secondary liquidity.
  • Other CDP stablecoins (crvUSD, LUSD) traded near-parity during the same period — the issue was specifically GHO’s mechanism.

​
Recovery

Aave governance shipped two interventions:
  1. Raised the GHO borrow rate to ~5% to slow mint flow.
  2. Added a GHO Stability Module (GSM) allowing 1:1 swaps with USDC and USDT inside a cap.
Both took until Q1–Q2 2024 to stabilize the peg back near $1.00.

​
What Pegana would have shown

  • August 2023, week 1 — GHO crosses 100 bps spread. PEGGED → DRIFT.
  • August 2023, week 3 — Spread crosses 200 bps. DRIFT → DEPEG.
  • August 2023 to ~Q1 2024 — Spread fluctuates 250 – 550 bps. State stays in DEPEG for ~8 months with brief DEPEG → DRIFT ducks when borrow rate hikes bit, then back to DEPEG.
  • Q2 2024 — GSM lands; spread compresses sustainably below 50 bps. After exit dwell, DEPEG → DRIFT → PEGGED.
A protocol watching GHO’s time_in_DEPEG would have measured 240+ days of sustained under-peg state — a structural signal of a mechanism that was designed wrong, not a shock that needed to pass.

​
Lesson

A stablecoin without a symmetric mint/burn path is a one-way valve. Watching the daily price for a sub-1% deviation misses the structural signal — eight months of $0.97 is a different kind of depeg than a 48h spike. Pegana’s EWMA-smoothed state machine catches both: short shocks emit transitions, and sustained drift accumulates into DEPEG-state durations on the events feed.

​
Sources