Fair point on the looper economics, and you’re right that I didn’t address it directly. Let me work through it.
You’re correct that PT markets as currently structured depend on 90%+ LTV to generate meaningful looped yield. Drop to 75–80% and the recursion ceiling collapses — spot sUSDe becomes the better trade and PT deposits dry up. The May 2025 → March 2026 decline from ~$4.6B to ~$515M happened for exactly this reason: implied yields compressed, loopers exited, and the collateral base reflected that.
But I want to push back on the implicit conclusion. “The product doesn’t exist at safer parameters” is an observation about the product, not a justification for the parameters. It means PT lending collateral is structurally dependent on aggressive LTV — which is actually a really important framing. PTs aren’t passive collateral in the sense ETH is. They’re looper infrastructure. The lending protocol isn’t underwriting “an asset”; it’s underwriting a specific leveraged strategy.
That reframing matters because the risk question changes. The question isn’t “what LTV keeps PTs viable” — it’s “should lending protocols underwrite looper strategies at parameters that assume no tail event over the tenor window?” For a 14-day PT, the answer might still be yes at 91%. For a 365-day PT, the framework says probably not, because 12 months is enough time for a regulatory action, a hedge-desk failure, a protocol exploit, or an AMM liquidity collapse to be non-negligible.
Two things that might reconcile this:
- Duration-graded LTV still lets near-maturity PTs loop at ~91%. The product works in the final 30 days. The constraint kicks in on fresh long-dated issuance.
- The discount curve already prices duration into the asset — but not into the LTV parameter applied to the mark. A 365d PT at $0.85 and a 14d PT at $0.99 both get 91% LTV against their respective prices, which means the position’s liquidation buffer is the same fraction of current mark regardless of tenor. Duration is priced in one place (the mark) but not the other (the LTV), and that’s the gap.
If the conclusion is “lending protocols should accept that PT markets aren’t commercially viable at risk-calibrated parameters,” that’s a real argument. I’d just want it stated explicitly, because it’s a different claim than “the framework is wrong.”