Paper: arXiv 2409.04233
Authors: Lukasz Szpruch, Marc Sabaté Vidales, Tanut Treetanthiploet, Yufei Zhang
Abstract
We study the loan contracts offered by decentralised loan protocols (DLPs) through the lens of financial derivatives. DLPs, which effectively are clearinghouses, facilitate transactions between option buyers (i.e. borrowers) and option sellers (i.e. lenders). The loan-to-value at which the contract is initiated determines the option premium borrowers pay for entering the contract, and this can be deduced from the non-arbitrage pricing theory. We show that when there are no market frictions, and there is no spread between lending and borrowing rates, it is optimal to never enter the lending contract. Next, by accounting for the spread between rates and transactional costs, we develop a deep neural network-based algorithm for learning trading strategies on the external markets that allow us to replicate the payoff of the lending contracts that are not necessarily optimally exercised. This allows hedge the risk lenders carry by issuing options sold to the borrowers, which can complement (or even replace) the liquidations mechanism used to protect lenders’ capital. Our approach can also be used to exploit (statistical) arbitrage opportunities that may arise when DLP allow users to enter lending contracts with loan-to-value, which is not appropriately calibrated to market conditions or/and when different markets price risk differently. We present thorough simulation experiments using historical data and simulations to validate our approach.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced financial mathematics, including non-arbitrage pricing theory, nonlinear PDEs for options with rate spreads, and barrier option frameworks, indicating high mathematical complexity. It also demonstrates high empirical rigor through the development of a deep neural network-based hedging algorithm, validation via historical data and simulations, and discussion of real-world application to decentralized lending protocols.
Research Flowchart
flowchart TD
A["Research Goal<br>Price & Hedge Decentralised Lending Contracts"] --> B{"Key Methodology"}
B --> C["Derivative Pricing Theory<br>Loan = Option, LTV = Premium"]
B --> D["No-Friction Model<br>Derive Null Optimal Exercise"]
B --> E["Deep Neural Network<br>Learn Hedging Strategy"]
C & D --> F["Computational Process<br>Simulation & Data Replication"]
E --> F
F --> G{"Key Outcomes"}
G --> H["Hedge Lender Risk<br>Beyond Liquidations"]
G --> I["Identify Arbitrage<br>Mispriced LTVs"]
```mermaid
flowchart TD
A["Research Goal: Pricing & Hedging<br>Decentralised Lending Contracts"]
A --> B["Methodology: Derivative Framework"]
B --> C["Model Loan as Option<br>LTV determines Premium"]
B --> D["No-Friction Analysis<br>Null Optimal Exercise"]
C & D --> E["Deep Neural Network<br>Learning Hedging Strategies"]
E --> F["Computational Process<br>Simulation & Historical Data"]
F --> G{"Outcomes"}
G --> H["Hedge Lender Risk<br>(Beyond Liquidations)"]
G --> I["Arbitrage Detection<br>Mispriced LTVs"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a deep neural network‑based algorithm to learn trading strategies on external markets that replicate the payoff of decentralised lending contracts and hedge lender risk.
Source sentence
we develop a deep neural network-based algorithm for learning trading strategies on the external markets that allow us to replicate the payoff of the lending contracts
Assumes 4
- No market frictions and no spread between lending and borrowing rates. market
Source sentence
when there are no market frictions, and there is no spread between lending and borrowing rates, it is optimal to never enter the lending contract
- Existence of an external market where the agent can borrow/lend USDC at riskless rates and trade the risky asset without transaction costs or price impact. market
Source sentence
Assumption 2. We assume that there exists an external market where: • The agent can borrow and lend any amount of USDC at the riskless rates rb,D and rc,D, respectively. • The agent can buy and sell any amount of risky asset with price P... The
- Once a position is open for liquidation it disappears from the agent balance sheet. model
Source sentence
For simplicity, we assume that once the position is open for liquidation, it disappears from the agent balance sheet.
- Trades can only happen on a discrete time grid. execution
Source sentence
assuming trades can only happen on a discrete time grid
Data
Synthetic data only.
Costs
Modeled
c is a constant fee paid for portfolio rebalancing
Source sentence
Define the total cost of the trading strategy π as Cτ(π) := ∑ t∈Πn,t≤τ c(πt −πt−∆t), where c : R →R+ is a constant fee paid for portfolio rebalancing.
Finds 2
- In the frictionless case with no rate spread, it is optimal to never enter the lending contract. theoretical
Source sentence
it is optimal to never enter the lending contract
- Deep‑hedge pricing yields a lending option price that varies with the rate spread, as shown in the simulation figure. pricing-calibration
Source sentence
Price of the lending option derived from the deep hedge for fixed parameters µ = 0,σ = 0.8,θ 0 = 0.83, 0 transaction costs, and variable rate spread ∈[0,0.5].
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 14 related papers). Coverage is partial, so this is neither replication nor refutation.
1 paper in the archive
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (56)
- The valuation of American barrier options using the decomposition technique (2000)
- Mathematical models of financial derivatives (2008)
- Optimal stopping and free-boundary problems (2006)
- Systemic fragility in decentralized markets (2022)
- Mathematical methods for financial markets (2009)
- Brownian motion and stochastic calculus (1991)
- Short Communication: A Primer on Perpetuals (2023) — link
- Robust pricing and hedging via neural SDEs (2020) — link
and 48 more
- Stock loans in incomplete markets (2013)
- The paradox of adversarial liquidation in decentralised lending (2023)
- The Economics of Interest Rate Models in Decentralised Lending Protocols (2023)
- A primer on perpetuals (2023)
- Fundamentals of Perpetual Futures — → 2212.06888
- On stock-based loans (2022)
- The decentralized financial crisis (2020)
- Stock loans (2007)
- Reap the harvest on blockchain: A survey of yield farming protocols (2022)
- Sok: Decentralized exchanges (dex) with automated market maker (amm) protocols (2023)
- From banks to DeFi: the evolution of the lending market (2022)
- Liquidations: DeFi on a Knife-edge (2021)
- An empirical study of defi liquidations: Incentives, risks, and instabilities (2021)
- Conditional sig-wasserstein gans for time series generation (2020) — link
- Arbitrage-free neural-SDE market models (2021) — link
- Short squeeze in defi lending market: Decentralization in jeopardy? — → 2302.04068
- Backward stochastic differential equations in finance (1997)
- Leveraged Trading Via Lending Platforms (2024)
- An optional decomposition of Y g, $\xi$- submartingales and applications to the hedging of American options in incomplete markets (2019) — link
- An optional decomposition of Y g, $\xi$- submartingales and applications to the hedging of American options in incomplete markets
- Morpho Optimizing Decentrlized Liquidity Protocols (2022)
- Uniswap v3 Core (2021)
- Perpetual dual American barrier options for short sellers (2019)
- Aave (2020)
- A general version of the fundamental theorem of asset pricing (1994)
- DeFi and the Future of Finance (2021)
- Smart contracts and decentralized finance (2023)
- Decentralized Finance: On Blockchain- and Smart Contract-Based Financial Markets (2021) — link
- Methods of mathematical finance (1998)
- Compound (2018)
- Mitigating decentralized finance liquidations with reversible call options (2023) — link
- Auto.gov: Learning-based On-chain Governance for Decentralized Finance (DeFi) — → 2302.09551
- Market risk assessment. An analysis of the financial risk to participants in the Aave protocol (2022)
- Sok: Decentralized finance (defi) attacks (2022)
- Deep hedging (2019)
- Deep-learning based numerical BSDE method for barrier options (2019) — link
- Monte Carlo methods in financial engineering (2004)
- Solving path dependent PDEs with LSTM networks and path signatures (2020) — link
- Stochastic calculus for finance II: Continuous-time models (2004)
- Inefficiency of CFMs: hedging perspective and agent-based simulations — → 2302.04345
- The costs of swapping on the uniswap protocol — → 2309.13648
- Equilibrium in a defi lending market (2023)
- On the inherent fragility of defi lending (2022)
- Interest rate rules in decentralized finance: Evidence from compound (2023)
- Defi leveraged trading: Inequitable costs of decentralization (2023)
- DeFi lending: intermediation without information? (2022)
- Attacks on dynamic defi interest rate curves (2023) — link
- Agents’ Behavior and Interest Rate Model Optimization in DeFi Lending (2024)
5 of 56 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.