Skip to main content
Asset: USDC · Class: Fiat stable · Date: 2023-03-11 Worst spread: −1300 bps · Duration: ≈ 36 h outside PEGGED Detected live: No (pre-Pegana retrospective)

​
Background

Circle disclosed late on Friday March 10, 2023 that ~3.3BofUSDCreservessatatSVB,whichhadbeenplacedintoFDICreceivership.TheintrinsicvalueofaUSDCtokenstillreflected3.3B of USDC reserves sat at SVB, which had been placed into FDIC receivership. The intrinsic value of a USDC token still reflected 1 of redemption claim — but the secondary market refused to price it that way until guarantees arrived.

​
Trigger

The first sub-0.99printslandedwithinanhouroftheSVB−exposuredisclosure.WithinfourhoursUSDCwastrading0.99 prints landed within an hour of the SVB-exposure disclosure. Within four hours USDC was trading 0.96 on Coinbase and 0.92onDEXs,withworst−caseprintsnear0.92 on DEXs, with worst-case prints near 0.87 in thin Sunday hours.

​
Cascade

  • DAI lost its peg simultaneously — its reserves were USDC-heavy. (See DAI / SVB contagion.)
  • MakerDAO emergency-paused USDC swaps via the PSM.
  • DEX pools rebalanced violently — large LPs printed losses on every imbalance.
  • Borrowing markets that priced collateral at $1.00 of USDC liquidated otherwise-healthy positions.

​
Recovery

Treasury, the Fed and the FDIC announced full backstop on Sunday March 12. USDC re-pegged in under 4 hours of US market open Monday. The full episode lasted ~60 hours.

​
What Pegana would have shown

  • Friday ~22:30 UTC — First Curve 3pool imbalance prints USDC at $0.989. Smoothed spread crosses 30 bps after dwell. PEGGED → DRIFT transition fires.
  • Saturday ~03:00 UTC — Spread crosses 50 bps. DRIFT → DEPEG.
  • Sunday ~04:00 UTC — Thin-Sunday DEX prints push spread above 5%. DEPEG → CRITICAL.
  • Monday ~14:00 UTC — Sustained recovery; exit dwell expires. CRITICAL → DEPEG → DRIFT → PEGGED over ~3 hours of sustained sub-threshold readings.
Each transition would have carried both intrinsic_usd (≈ $1.00 throughout) and market_usd (the actual print), letting downstream consumers reason about which side moved.

​
Lesson

Pegana’s two-numbers framing is built for exactly this case. Intrinsic was ~$1; market was not. A monitor that only watches one side prints either a false negative or a false positive — neither is useful when liquidation engines are pulling the trigger.

​
Sources