Paper: arXiv 2511.02518
Authors: Paulin Aubert, Etienne Chevalier, Vathana Ly Vath
Abstract
This paper develops a model for option market making in which the hedging activity of the market maker generates price impact on the underlying asset. The option order flow is modeled by Cox processes, with intensities depending on the state of the underlying and on the market maker’s quoted prices. The resulting dynamics combine stochastic option demand with both permanent and transient impact on the underlying, leading to a coupled evolution of inventory and price. We first study market manipulation and arbitrage phenomena that may arise from the feedback between option trading and underlying impact. We then establish the well-posedness of the mixed control problem, which involves continuous quoting decisions and impulsive hedging actions. Finally, we implement a numerical method based on policy optimization to approximate optimal strategies and illustrate the interplay between option market liquidity, inventory risk, and underlying impact.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 6.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced stochastic control, Cox processes, and rigorous analysis of market impact and arbitrage, indicating high mathematical complexity. It also presents a numerical method based on policy optimization and discusses simulation-based experiments, demonstrating substantial empirical implementation efforts.
Research Flowchart
flowchart TD
Start["Research Goal<br>Model Option Market Making with<br>Hedging-Induced Market Impact"] --> Input
subgraph Input ["Data & Inputs"]
I1["Stochastic Option Demand<br>Cox Process Intensities"]
I2["Underlying Asset State<br>Inventory & Price Dynamics"]
end
Input --> Method
subgraph Method ["Methodology Steps"]
M1["Formulate Mixed Control Problem<br>Continuous Quoting + Impulsive Hedging"]
M2["Identify Arbitrage &<br>Market Manipulation Phenomena"]
M3["Implement Numerical Optimization<br>Policy Iteration Algorithm"]
end
Method --> Outcome
subgraph Outcome ["Key Findings & Outcomes"]
F1["Established Well-Posedness<br>of Control Problem"]
F2["Optimal Strategies Reveal<br>Trade-offs: Liquidity vs. Inventory Risk"]
end
Start ~~~ Input
Input ~~~ Method
Method ~~~ Outcome
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 develops a model for option market making in which the market maker’s hedging activity generates permanent and transient price impact on the underlying, and implements a numerical policy‑optimization method to approximate optimal quoting and hedging strategies.
Source sentence
This paper develops a model for option market making in which the hedging activity of the market maker generates price impact on the underlying asset.
Assumes 6
- Option order arrivals are modeled by Cox processes whose intensities depend on the market maker’s quotes and the state of the underlying. model
Source sentence
Instead, we model order arrivals in the options market by Cox processes whose intensities depend on both the quotes offered by the market maker and the state of the underlying, allowing the option price to emerge endogenously as a result of supply and demand.
- The market maker’s hedging activity impacts the underlying price via a permanent and a resilient component. market
Source sentence
This impact is modeled by a permanent and a resilient component.
- The shape of the underlying asset’s order book is time‑invariant. execution
Source sentence
While empirical evidence indicates that order book depth can be highly stochastic, we assume for simplicity that the shape of the order book is time-invariant.
- The order‑book density functions satisfy integrability conditions ensuring finite total quantity and value. execution
Source sentence
Assumption 2.1 (Integrability of the order book shape).
Show 2 more
- Beyond certain price levels no liquidity is available (right‑truncated depth). execution
Source sentence
There exist UA, UB ∈(0, ∞) such that fB(u) = 0 for a.e. u < −UB, fA(u) = 0 for a.e. u > UA.
- The market maker must provide liquidity for a European option while complying with external hedging constraints. market
Source sentence
In this paper, we consider a market maker mandated to ensure the liquidity of a European option, while complying with hedging constraints imposed by a third party (regulator or principal).
Data
Synthetic data only.
Costs
Modeled
Execution cost functions PB and PA are defined via integration over the order‑book shape, linking order size to monetary cost.
Source sentence
Let b > 0 denote the current best bid price. (2.5) By Assumption 2.1, the execution costs (2.4)–(2.5) are finite for all admissible order sizes 0 ≤q−≤ΦB(b) and 0 ≤q+ ≤ΦA(+∞).
Finds 3
- Pre‑trading to influence prices before executing a target trade reduces pure execution P&L (Π(z)T−Π(0)T ≤ −δ²z−c < 0). theoretical
Source sentence
Then, for any admissible z, q, the incremental profit satisfies Π(z) T −Π(0) T ≤−δ 2z −c < 0. The strict inequality holds because δ > 0, which completes the proof.
- The learned neural policy yields a systematic, albeit moderate, improvement over a naive benchmark in terminal cash distribution (e.g., mean 72.16 vs 72.00). performance
Source sentence
Turning to performance, Table 5 show that the learned policy yields a systematic, albeit moderate, improvement over the naive benchmark. Agent Mean Std Median Q5 Q95 Learned 72.16 0.05 72.18 72.06 72.20 Naive 72.00 0.00 72.00 72.00 72.01 Table 5: Terminal cash distribution for the asymmetry intensities configuration.
- Increasing the hedging penalty coefficient κ_hedge leads to more risk‑averse behavior and lower average P&L (e.g., κ=0.5 gives mean −199.72, κ=16 gives mean −263.02). performance figure from elsewhere in the paper
Source sentence
The resulting performance metrics are reported in Table 4, and further insights are provided through the net delta dynamics displayed in Figures 7a and 7b. κhedge Avg. P&L Std. 0.50 -199.72 14.64 15.48 2.00 -223.93 10.70 5.09 4.00 -232.62 8.30 4.28 8.00 -245.89 6.69 3.60 16.00 -263.02 5.07 2.93 Table 4: Sensitivity of the learned policy to hedging penalty coefficient.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 15 related papers). Coverage is partial, so this is neither replication nor refutation.
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 (16)
- Optimal execution and price manipulations in time-varying limit order books (2014) — link
- Dynamic optimal execution in a mixed-market-impact Hawkes price model (2016) — link
- Optimal execution of portfolio transactions (2001) — link
- High-frequency trading in a limit order book (2008) — link
- Algorithmic market making for options (2021) — link
- Deep hedging (2019) — link
- Optimal execution with limit and market orders (2015) — link
- Buy Low, Sell High: A High Frequency Trading Perspective (2014) — link
and 8 more
- A Stochastic Control Approach to Option Market Making (2015) — link
- Dynamical Models of Market Impact and Algorithms for Order Execution (2013)
- Dealing with the Inventory Risk. A solution to the market making problem (2013) — link
- Optimal High Frequency Trading with Limit and Market Orders (2013) — link
- Price manipulation and quasi-arbitrage (2004) — link
- Optimal Execution in a General One-Sided Limit-Order Book (2011) — link
- Market Making via Reinforcement Learning (2018) — link
- Option market making under inventory risk (2009) — link
0 of 16 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.