Paper: arXiv 2404.15478

Authors: Alexander Barzykin, Philippe Bergault, Olivier Guéant

Abstract

The primary challenge of market making in spot precious metals is navigating the liquidity that is mainly provided by futures contracts. The Exchange for Physical (EFP) spread, which is the price difference between futures and spot, plays a pivotal role and exhibits multiple modes of relaxation corresponding to the diverse trading horizons of market participants. In this paper, we model the EFP spread using a nested Ornstein-Uhlenbeck process, in the spirit of the two-factor Hull-White model for interest rates. We demonstrate the suitability of the framework for maximizing the expected P&L of a market maker while minimizing inventory risk across both spot and futures. Using a computationally efficient technique to approximate the solution of the Hamilton-Jacobi-Bellman equation associated with the corresponding stochastic optimal control problem, our methodology facilitates strategy optimization on demand in near real-time, paving the way for advanced algorithmic market making that capitalizes on the co-integration properties intrinsic to the precious metals sector.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 4.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper employs advanced stochastic control theory with nested Ornstein-Uhlenbeck processes, Hamilton-Jacobi-Bellman equations, and Riccati equations, indicating high mathematical density. However, the summary and excerpt focus on theoretical modeling and optimization without presenting backtested results, implementation code, or specific data, placing it in the realm of theoretical research rather than immediate practical deployment.

Research Flowchart

  flowchart TD
  A["Research Goal"] --> B["Market Making for Precious Metals"]
  B --> C["Key Challenge: Spot-Futures Liquidity Gap"]
  C --> D["Data Input: EFP Spread Time Series"]
  D --> E["Methodology: Nested Ornstein-Uhlenbeck Process"]
  E --> F["Computation: Approx. Hamilton-Jacobi-Bellman"]
  F --> G["Outcome: Optimized Strategy for P&L & Risk"]
  G --> H["Near Real-time Algorithmic Market Making"]

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 nested Ornstein‑Uhlenbeck model for the EFP spread and a computationally efficient stochastic optimal control method to optimize market‑making strategies in spot precious metals.

Source sentence
We model the EFP spread using a nested Ornstein-Uhlenbeck process, in the spirit of the two‑factor Hull‑White model for interest rates.

Assumes 6

  • Transaction likelihood depends only on the distance between quoted price and reference spot price. model
    Source sentence
    Transaction likelihood is assumed to depend solely on the distance between the proposed prices and the reference spot price.
  • Bid and ask intensity functions share the same functional form Λ(z,δ)=λ(z)f(δ). model
    Source sentence
    We assume that the functions Λb and Λa take the form
  • Spot‑EFP correlation is zero in the numerical example. data
    Source sentence
    Spot‑EFP correlation is typically small and assumed to be zero in this example.
  • Permanent market impact is ignored for simplicity. execution
    Source sentence
    We have not considered permanent market impact here for the sake of simplicity.
Show 2 more
  • Execution rates for spot and futures are modeled by stochastic processes (vS_t) and (vF_t). execution
    Source sentence
    The execution rates associated with the market maker’s activities are modeled by the processes (vS_t)_{t≥0} and (vF_t)_{t≥0} for spot and futures, respectively.
  • Linear permanent market impact would not alter the effectiveness of the model. model
    Source sentence
    Under the assumption of linear permanent market impact, our model remains effective.

Data

  • Spot gold (XAUUSD) market data with futures liquidity market-prices unknown
    Gold spot market
    Source sentence
    For illustration, we consider market making in spot gold (XAUUSD) with access to futures liquidity.

Costs

Modeled Costs are part of the reported headline results.

Instantaneous market impact is modeled as LS(v)=ψS|v|+ηS v^2 and LF(v)=ψF|v|+ηF v^2 with ψS=0.4 bp, ψF=0.2 bp, ηS=7·10⁻⁸ bp·day·oz⁻¹, ηF=3·10⁻⁸ bp·day·oz⁻¹.

Source sentence
Standard functional dependence of instantaneous market impact on execution rate is assumed: LS(v) = ψS|v|+ηSv2 and LF (v) = ψF |v|+ηF v2 with ψS = 0.4 bp, ψF = 0.2 bp, ηS = 7·10−8 bp·day·oz−1 and ηF = 3 · 10−8 bp · day · oz−1.

Finds 4

  • Introducing futures hedging improves the Sharpe ratio of the OTC spot strategy by at least 30%. performance
    Source sentence
    Introducing futures hedging into risk management of OTC spot improves Sharpe ratio by at least 30% (not shown), despite the additional round‑trip cost of futures trading.
  • Optimal quotes are tighter when futures are available, reducing bid‑ask spreads across inventory levels. performance
    Source sentence
    It also allows to provide better prices to liquidity takers, as shown in Figure 10 where we compared the quoted bid‑ask spreads in the absence and presence of a futures market.
  • The optimal strategy keeps the EFP position mean‑reverting, reflecting the EFP spread’s dynamics. theoretical
    Source sentence
    EFP position clearly echoes the mean‑reverting nature of EFP – the dealer aims to keep EFP risk against the direction of EFP deviation.
  • Volume share of futures execution significantly exceeds that of spot due to lower cost. performance
    Source sentence
    Volume share of futures execution significantly exceeds that of spot due to lower cost.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 30 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers 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.