Summary
LlamaRisk acknowledges and appreciates the efforts of @bgdlabs in proposing a way forward for the pricing oracle of Pendle Principal Tokens (PTs). Resolving the Oracle question is crucial for onboarding PT assets, representing a significant opportunity for the Aave DAO.
The proposed dynamic linear discount rate oracle is a suitable and well-considered middle-ground solution. It balances the need for a market-reflective price feed without altering trust assumptions on external pricing sources while maintaining governance levers. The dynamic nature of the discount rate adjustment aligns with the broader view that each PT asset requires individualized monitoring and parameter calibration throughout its lifecycle, given the variance in underlying asset volatility and yield expectations.
Synopsis of the Dynamic Linear Discount Oracle
The dynamic linear discount oracle proposed by BGD Labs offers an approach that synthesizes elements from previously discussed models.
- It builds upon the Static Linear Discount concept, where a PT’s price is calculated by applying a predetermined linear discount rate based on its time to maturity. The price converges towards par (1:1) as maturity approaches.
- However, unlike the static model, the BGD proposal introduces dynamic adjustments to the linear discount rate during the PT’s lifetime. This allows the oracle price to adapt indirectly to significant changes in market conditions, underlying asset volatility, or yield expectations, which a static rate cannot account for.
- This contrasts with the Dynamic TWAP Oracle previously proposed, which aimed for a more complex, data transformation-driven pricing mechanism based on observed prices, introducing different complexities regarding potential manipulation vectors and implementation robustness.
The BGD Labs proposal effectively acts as a bridge: it retains the predictability and simplicity of a linear discount framework while incorporating a necessary layer of adaptability managed through governance.
Key Differences in Pricing Approaches:
| Feature | Static Linear Discount | Dynamic Linear Discount | Dynamic TWAP |
|---|---|---|---|
| Discount Rate | Fixed at listing | Adjustable post-listing | Implicit (derived from market) |
| Intervention Frequency | None (time-based only) | Moderate (Risk Steward changes) | High (Edge enforced changes) |
| Update Mechanism | Time progression | Governance (Risk Steward) | Market price changes (TWAP + Thresholds) |
| Complexity | Low | Low/Moderate | High |
General Considerations
Successful implementation relies on several key factors:
- Initial Discount Rate Calibration: Setting appropriate initial
discountRateandmaxDiscountRateparameters at listing is critical. This requires careful analysis of the specific PT, underlying asset, prevailing market conditions, and yield expectations. - Onboarding Timing: We concur that onboarding PT markets only after an initial price discovery phase (e.g., 10-15 days post-launch on Pendle) is prudent. This allows the market to establish a baseline price and volatility profile, informing a more accurate initial
discountRatesetting for the Aave deployment. - Risk Steward: The dynamic adjustment mechanism necessitates robust governance. The proposed use of a Risk Steward, constrained by limits on the magnitude (e.g., 5%) and frequency (e.g., every 2-3 days) of discount rate changes, provides a controlled and transparent adjustment method. All to be announced publicly to make it possible for LlamaRisk and other governance entities to peer-review the changes.
- Overpricing Risk: This approach inherently minimizes overpricing risk compared to potentially lagging market-based oracles, as the price is derived from a controlled discount. While underpricing relative to the secondary market is possible if the discount rate isn’t adjusted promptly, the over-valuation of collateral within Aave is the primary risk mitigated. Diligent monitoring and adjustment by the Risk Steward are key to keeping the oracle price reasonably aligned with fair value.
Future Improvements
The dynamic linear discount model provides a solid foundation. Nonetheless, based on future observations of its performance and the behavior of PT markets within Aave post-listing, we could propose further improvements:
- The effectiveness and optimal frequency/magnitude constraints for discount rate adjustments can be refined over time based on observed market dynamics and operational overhead.
- The linear model offers simplicity, but future iterations could explore a dynamic exponential discount model. As our initial comment on the original thread covered, an exponential curve might more accurately reflect the typical price behavior of zero-coupon bond-like instruments (such as PTs), where price sensitivity to rate changes is higher further from maturity. This could be considered for future PT listings or Oracle upgrades if the linear model shows limitations.
LlamaRisk supports proceeding with the dynamic linear discount rate oracle framework proposed by @bgdlabs, subject to careful parameterization and diligent ongoing management via the designated Risk Steward process. It offers a secure and adaptable solution to onboard PT assets to Aave.
Disclaimer
This review was independently prepared by LlamaRisk, a community-led decentralized organization funded in part by the Aave DAO. LlamaRisk is not directly affiliated with the protocol(s) reviewed in this assessment and did not receive any compensation from the protocol(s) or their affiliated entities for this work.
The information provided should not be construed as legal, financial, tax, or professional advice.