A recent podcast on derivatives deserves unpacking
A point inspired by a recent podcast on crypto derivatives deserves unpacking.
https://t.co/ygbP15p515
ADL (auto-deleveraging) isn't a feature. It's a doomsday scenario.
When your hedge disappears at 3am because the venue/counterparties ran out of margin buffer, that's not "working as designed." That's systemic risk with good PR.
The conventional fix is bigger capital buffers. DCOs. Incentivized liquidity pools. Third-party insurance. Essentially: the TradFi playbook, adapted for crypto rails.
The problem is mutualization itself. For this to work the capital cover has to be huge. It is the reason FCMs are so sticky.
Every participant in a shared default fund inherits exposure to every other participant's risk behavior. You're not just managing your counterparty risk - you're exposed to counterparties you never chose, positions you can't see, and risk profiles you never underwrote. Whether that is ADL or Default waterfall.
It's Quanto-like behavior: your P&L becomes dependent on variables you didn't trade and can't hedge.
The 1010 liquidations made this visible. Cryptos implementation shows that even those with good positions and data can't win. Market makers and prop traders got ADL'd on hedges that were working.
Retail got liquidated on positions that should have survived. Not because of their risk management - because of someone else's.
There's an alternative architecture that doesn't require 100m USD of default fund before trade 1.
Bilateral credit intermediation with securitized tail risk:
- Each counterparty relationship is structurally isolated
- Tail risk transferred via bond issuance to investors who price it, not pooled across participants who inherit it
- Off-chain real-time risk calculation, on-chain margin and settlement
- Margin posted to a capitalized legal entity with actual recovery resolution
- No shared fund to raid. No cascade across unrelated counterparties. No ADL.
Institutions won't scale exposure to venues that can unilaterally tear up their hedges. The risk infrastructure needs to match the risk appetite.
Settlement rails are necessary. But they're not sufficient to scale.
I am very much enjoying people I respect discussing the issues here. I am sure we have a solution.