Risk metrics and formulas

The current implementation scores five inputs, then combines them into a final integer risk score from 0 to 10.

Utilization

Utilization measures the share of available liquidity that is borrowed:

$$\text{Utilization} = \frac{\text{Total Liabilities}}{\text{Total Assets}}$$

Liquidation at risk

Liquidation at risk is zero at or below the safe-utilization threshold. Above that point it increases with excess utilization:

$$\text{LaR} = \text{Total Liabilities} \times \frac{\text{Utilization} - \text{Safe Utilization}}{\text{Collateral Factor}}$$

The current implementation uses a safe-utilization value of 0.4 and a collateral factor of 1.

Value at Risk

The current VaR estimate applies a 5% volatility factor and a 95% confidence Z-score of 1.65:

$$\text{VaR} = \text{Total Assets} \times 0.05 \times 1.65$$

Borrow usage

Borrow usage compares liabilities with available assets:

$$\text{Borrow Usage} = \frac{\text{Total Liabilities}}{\text{Total Assets} \times \text{Collateral Factor}}$$

Higher usage maps to a higher risk band, reaching the maximum score when liabilities equal or exceed assets.

Rate spread

The spread is the difference between the borrow and supply rates:

$$\text{Spread} = \text{Borrow Rate} - \text{Supply Rate}$$

Spreads below 2% receive the lowest implemented spread score. Scores increase at 2%, 5%, 8%, 9%, and 10% boundaries.

Final score

The engine combines the component scores with fixed weights:

$$\begin{aligned} \text{Final Risk} ={}& 0.30(\text{Utilization Score}) \\ &+ 0.20(\text{LaR Score}) \\ &+ 0.20(\text{VaR Score}) \\ &+ 0.15(\text{Borrow Usage Score}) \\ &+ 0.15(\text{Spread Score}) \end{aligned}$$

The result is converted to an integer and used by the allocation solver and fund-level risk controls.