Interest rates in leveraged lending markets
Core mechanism
Leverage looping lets a borrower increase exposure to a high-yield asset by repeatedly borrowing non-interest-bearing (NIB) stablecoins against it and reinvesting the proceeds. The borrower receives the residual return after financing costs and bears the asset's downside. The tradeoff is that if the collateral value ever crosses the liquidation threshold, the position is liquidated; liquidation penalties and adverse execution can consume most or all of the borrower's equity.
The lender earns interest for supplying the borrowed asset. Their main economic risk is that collateral falls through the liquidation threshold faster than it can be sold at a fair price, leaving bad debt. Smart-contract, oracle, and governance risks may also apply.
Call the borrow rate X.
Two common sources of lending-market inefficiency separate the borrow rate from the lender's realized rate:
Idle liquidity. A liquid reserve of 10% in the lending protocol means that only 90% of supplied capital earns the borrow rate.
Protocol or curator fee. Curators generally charge a fraction of the borrow rate, most typically 10%. Lending protocols without a curator also charge this, referring to it as the reserve factor, and use it to fund protocol reserve to cover bad debt. As a result of a 10% curator/reserve fee, suppliers only retain 90% of the interest.
Under those assumptions, the lender rate is 0.90 * 0.90 * X = 0.81 * X.
As an aside, the curator fee arguably should be benchmarked to the risk-free rate (i.e. the Treasury-bill rate), since then it would be more like a performance fee, but this is not currently the market standard. Of course then maybe the rate would be higher.
Borrower economics
Let the collateral asset yield Y and support a 90% maximum loan-to-value ratio. The theoretical maximum loop is approximately 10x gross asset exposure: 10 units of asset deposited in the lending protocol, with 9 units of debt. Then the annualized return is 10 * Y - 9 * X = Y + 9 * (Y - X).
If the borrower chooses 5x gross exposure instead, consisting of 5 units of asset deposited in the lending protocol with 4 units of debt, then their annualized return is 5 * Y − 4 * X = Y + 4 * (Y − X).
I like to write the return this way because it becomes clear that each additional unit of leverage adds Y - X of return. For example, if the asset yield exceeds the borrow rate by one percentage point, or Y - X = 1%, then 5x leverage adds four percentage points to the borrower's unlevered yield. The borrower earns Y + 4% while accepting the larger drawdown and liquidation risk.
Implied lender rates
To compute boundary conditions for market feasibility, let's assume that:
utilization is 90%; and
the curator or protocol takes 10% of borrower interest; and
X = Y - 1%, leaving the borrower a one-percentage-point spread on each borrowed turn.
The lender therefore receives 0.81 × (Y − 1%).
Then, for a given value of Y, here is the lender yield:
| Asset yield (Y) | Borrow rate (X = Y − 1%) | Lender yield (0.81 × X) |
|---|---|---|
| 5% | 4% | 3.24% |
| 5.5% | 4.5% | 3.65% |
| 6% | 5% | 4.05% |
| 6.5% | 5.5% | 4.46% |
| 7% | 6% | 4.86% |
| 7.5% | 6.5% | 5.27% |
| 8% | 7% | 5.67% |
| 8.5% | 7.5% | 6.08% |
| 9% | 8% | 6.48% |
| 9.5% | 8.5% | 6.89% |
| 10% | 9% | 7.29% |
| 10.5% | 9.5% | 7.70% |
| 11% | 10% | 8.10% |
| 11.5% | 10.5% | 8.51% |
| 12% | 11% | 8.91% |
This result is a feasibility threshold under the stated assumptions. A lender may be willing to accept a lower rate, allowing the borrow rate to fall and leaving more spread for the borrower. But if lenders require a return above the table's implied lender yield, the protocol cannot satisfy them while preserving the borrower's assumed one-percentage-point spread. The loop would need a higher asset yield, lower idle liquidity, a lower protocol or curator fee, or a smaller borrower spread.
This result tells us whether a particular lender-borrower market will clear. For example if a lender says they would be willing to lend against a 9% asset, but they want 7% yield, then this market will not clear. From this information alone we don't know why the market doesn't clear - the lender might be asking for too much, or the lender might be pricing it correctly and the asset is overly risky for the rate offered, or the hurdle rate from curator fee and idle liquidity might be too high.
Role of incentives
Incentives can temporarily bridge the gap between the lender yield generated by the market and the lender's reservation price, meaning the minimum return the lender requires to supply capital. For example, if the market produces a 5% lender yield and lenders require 6%, a one-percentage-point incentive can clear the market while it is active.
Incentives should fund a transition with a credible path to unsubsidized market clearing. By the time incentives end, either the sustainable lender yield must rise or the lender reservation price must fall. The lender yield could rise through a higher asset yield and borrow rate, greater utilization, or lower protocol and curator fees. The reservation price could fall as liquidity deepens, price behavior proves orderly, liquidation performance establishes a track record, or perceived smart-contract and operational risk declines.
If neither change is likely, incentives create capital that leaves when the subsidy ends. The relevant decision is whether the expected transition is plausible within the proposed incentive period and whether post-incentive retention can be measured.
Market-clearing question
The market works only when lenders accept the resulting yield as sufficient compensation for the collateral's liquidation risk and the protocol's other risks. The relevant comparison is the lender's net yield against the risk-free rate.
Lender comfort depends heavily on the orderliness of collateral price movements:
Continuous, liquid price movement gives liquidators time and market depth to close positions near the expected price.
Jump risk, thin liquidity, market closures, or impaired redemption can move collateral below the liquidation threshold before an orderly sale is possible, creating bad debt.
The same stated LTV can therefore support very different lender rates across collateral types. A complete assessment should consider liquidation depth and slippage at stressed prices, oracle behavior, liquidation incentives and latency, collateral concentration, withdrawal liquidity, and historical gap risk.
The hurdle to clear
The hurdle in the above formulas comes from two components, the curator/protocol fee, and the idle capital in the lending protocol. The first one goes to the curator/protocol, while the second one goes to the stablecoin issuer.
Lower curator/protocol fees, or lending arrangements denominated in currencies where the stablecoin issuer shares NIM back to the market, result in lower hurdle rates for leverage.
Reusable formulas
For utilization U, fee rate F, gross asset exposure L, asset yield Y, and borrow rate X:
Lender return: U × (1 − F) × X
Borrower return: L × Y − (L − 1) × X = Y + (L − 1) × (Y − X)
These formulas are a simplified steady-state model. Actual realized returns depend on variable rates, compounding, leverage maintenance, liquidation penalties, transaction costs, incentives, and changes in collateral value.
Originally published on X on September 13, 2026.