First things first, we appreciate the time and effort spent on assessments for both Prime and Select. At the same time, we want to clarify points regarding the v-wmtUSDC market and seek to continue the discussion (for example, in the context of DAO-managed fleets), even where it may not change the conclusion.
Liquidity
Providing liquidity through AMM pools for private credit products is not very sustainable as it incurs a high cost of capital. Hence, the liquidity is provided through RFQ/solvers, like UniswapX. Currently, the capacity is at ~500k wmtUSDC under minimal (sub 5 bps) slippage. In case of potential insolvency one could expect secondary creditor claims markets to emerge.
Saying all this, we want to note that Wintermute doesn’t have any admin functionality for the instant burning of debt tokens in its possession; hence, providing a sizeable instant exit opportunity at any time will lead to unnecessary capital inefficiency. Armitage is also in consideration of alternative designs for this specific asset, by tranching or looping through protocols like 3F.
Earlier markdown
While the general logic could apply to certain onchain private credit products, the devil is in the details. Wildcat, as a protocol, pushes borrowers to honour redemptions by imposing a delinquency penalty. For wmtUSDC, the grace period before delinquency is 48 hours, so holding redemptions open for 2 days straight increases the APR by 5%. That’s enough to make the loan expensive for the borrower quickly, though not by itself sufficient to derisk the Morpho market.
Market derisking is handled by another safeguard implemented at the vault level: the bespoke market oracle for Wildcat markets. The oracle operates on the delinquency time accumulator from the wmtUSDC market and applies a price discount proportional to the duration of the delinquent state. So, aggressive loopers could be derisked pretty quickly given oracle parameters.
We’d note this addresses the “no markdown before default” concern directly for the delinquent case, but we recognize it doesn’t cover credit deterioration. That gap is inherent to any point-in-time credit product without continuous external pricing, not specific to this market.
Same risk counted twice
On the concern that v-wmtUSDC collateral used to borrow more USDC double-counts the same credit exposure: this is accurate in the sense that both legs share the same ultimate counterparty risk. We don’t think this is unique to this market structure, given it’s true of any leverage looping involving correlated risk.
Supply and borrowing demand
We work on reducing the supply concentration risk caused by seed capital that does not leave the vault at the bootstrap stage. Regarding being the only lender for this market, it was deployed by Armitage, incl. the mentioned oracle. We have mid-term plans to increase liquidity by onboarding additional curators. The main roadblock is determining the form that private credit should prevail (e.g., tranching or 3F) or if it worth for them to coexist, rather than growing this market first and pushing for the migration later.
In regards to caps:
- $25M as an absolute cap is more than necessary, and that’s the one we set in expectation of growth and the fact that we’ll be the only curator for some time. The idea here is that raising the cap is more expensive in terms of time (because of the timelock) than decreasing it. We’d be glad to consider decreasing it by a factor of 3 to make the exposure risk more controllable.
- 100% relative cap follows similar logic, and relative caps are implemented internally within our allocation engine. The relative cap we use internally is 17.5%, with a bit of extra space if needed, given being the only lender for v-wmtUSDC market.
Looping strategy is an expected use case for this market, which makes economic sense given the positive carry spread, similar to the AA_FalconXUSDC/USDC market, or existing Pendle markets. We would like to see more third-party usage in the form of repo deals, but the market is still nascent in this regard, and the spread makes looping lucrative, similar to the 3rd-party examples mentioned.
On the topic of borrower concentration, we want to note that two of the mentioned loopers are known to us (experienced DeFi farmers without any ties to Armitage/Wintermute) and performed looping in a way that was convenient for them, given the lack of current support for this market on platforms like DeFi Saver. There is additional demand for looping in this market; we don’t plan to serve it at the moment to support a more responsible rollout.
Liquidations
Concerning a potential insolvency event (the same logic applies to depeg events, which we witness quite often nowadays), liquidator identity is of no importance as nobody would take the underlying asset on the balance sheet at the value assumed by the market oracle. We acknowledge that the liquidator may not appear to be a strong backstop compared to other markets in the event of insolvency for v-wmtUSDC. However, the oracle was designed to decrease the price each day until it reached $0.01, while anyone could buy claims on this collateral at any discount and repay the debt at least partially.
Outside the insolvency case, at 100% utilization and/or an extended delinquency state flagged by the oracle, speculatively liquidating to buy debt at a discount and redeem it is also economically viable for third parties. This incentive creates the backstop rather than any specific liquidator identity on this market.
