Paper: arXiv 2401.03305

Authors: Xue Cheng, Peng Guo, Tai-ho Wang

Abstract

The paper addresses the problem of meta order execution from a broker-dealer’s point of view in Almgren-Chriss model under execution risk. A broker-dealer agency is authorized to execute an order of trading on some client’s behalf. The strategies that the agent is allowed to deploy is subject to a benchmark, referred to as the reference strategy, regulated by the client. We formulate the broker’s problem as a utility maximization problem in which the broker seeks to maximize his utility of excess profit-and-loss at the execution horizon, of which optimal feedback strategies are obtained in closed form. In the absence of execution risk, the optimal strategies subject to reference strategies are deterministic. We establish an affine structure among the trading trajectories under optimal strategies subject to general reference strategies using implementation shortfall (IS) and target close (TC) orders as basis. Furthermore, an approximation theorem is proposed to show that with small error, general reference strategies can be approximated by piece-wise constant ones, of which the optimal strategy is piece-wise linear combination between IS and TC orders. We conclude the paper with numerical experiments illustrating the trading trajectories as well as histograms of terminal wealth and utility at investment horizon under optimal strategies versus those under TWAP strategies.

Complexity vs Empirical Score

  • Math Complexity: 8.0/10
  • Empirical Rigor: 4.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper presents a stochastic control formulation with multiple stochastic differential equations, deriving closed-form solutions and establishing an affine structure theorem, which indicates advanced mathematical density. However, it lacks any mention of real-world data, backtests, or implementation code, relying instead on theoretical numerical examples for illustration.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Optimal Execution Subject to<br>Reference Strategy"] --> B["Methodology:<br>Broker Utility Maximization<br>(PnL at Horizon)"]
  B --> C["Model Setup:<br>Almgren-Chriss Model<br>with Execution Risk"]
  C --> D["Key Theoretical Result:<br>Affine Structure:<br>Strategy = α × IS + (1-α) × TC"]
  D --> E["Computational Process:<br>Solve Optimization for<br>Coefficient α"]
  E --> F["Findings/Optimal Trajectory:<br>Subject to Reference Strategy<br>(vs. TWAP Benchmark)"]
  F --> G["Outcomes:<br>1. Closed-form Solution<br>2. High Utility/PnL<br>3. Risk-Adjusted Performance"]

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 formulates the broker's order execution problem as a utility maximization problem and derives closed‑form optimal feedback strategies that are affine combinations of implementation‑shortfall (IS) and target‑close (TC) orders.

Source sentence
We formulate the broker's problem as a utility maximization problem in which the broker seeks to maximize his utility of excess profit-and-loss at the execution horizon, of which optimal feedback strategies are obtained in closed form. We establish an affine structure among the trading trajectories under optimal strategies subject to general reference strategies using implementation shortfall (IS) and target close

Assumes 6

  • The price impact follows the Almgren–Chriss model with permanent impact γ and temporary impact η. model
    Source sentence
    The price dynamic is assumed to follow the Almgren–Chriss model [4] [1] [5]. The term γ(xt −x0), for γ ≥0, is usually referred to as the permanent impact.
  • Execution risk is introduced via a diffusion term m(v) driven by a correlated Brownian motion Zt. execution
    Source sentence
    To account for this uncertainty, we introduce a noise component driven by a correlated Brownian motion Zt into the dynamic of the position xt as dxt = −vtdt + m(vt)dZt. The diffusion term m(v) characterizes the magnitude of execution risk.
  • The broker maximizes an exponential (CARA) utility of excess P&L with risk‑aversion parameter θ. other
    Source sentence
    The broker’s problem is to determine an optimal admissible strategy v∗that maximizes his expected exponential utility at the horizon T, i.e., sup v∈A E h u  ˜Π i , (2) where the utility function u(x) := 1 θ(1 −e−θx), θ > 0 represents the broker’s preference following CARA (Constant Absolute Risk Aversion) preference and θ > 0 is the risk-aversion parameter.
  • When the execution‑risk parameter m0 equals zero, the optimal trading trajectories become deterministic. model
    Source sentence
    In the absence of execution risk, the optimal strategies subject to reference strategies are deterministic. 4 Zero execution risk For the purpose of better understanding as well as visualizing the optimal strategy obtained in (5), we present in this section the trading trajectory xt under optimal strategy in the case when m0 = 0 for various reference strategies.
Show 2 more
  • Reference strategies (benchmarks) are given by IS and TC orders that the broker must track. market
    Source sentence
    For instance, implementation shortfall (IS) orders [1] are frequently employed by managers for the purpose of short-term alpha pursuit. IS orders are constructed with a pre-trade benchmark price in mind, aiming at executing orders at an average price that remains relatively close to the market price at the beginning of the trade.
  • The underlying asset price follows a Brownian motion with drift μ and volatility σ. market
    Source sentence
    The fair price St is driven by the SDE dSt = µdt + γdxt + σdWt, or equivalently, St = S0 + µt + γ(xt −x0) + σWt, where µt describes the tendency of the stock and (Wt)t∈[0,T ] is a standard Brownian motion.

Data

  • Synthetic Monte‑Carlo simulations of price paths and execution outcomes synthetic unknown
    Source sentence
    Parameters chosen as base case are (m0, η, ρ, θ, σ, β) = (0.05, 10, 0, 0.002, 200, 1000). Table 1 summarizes the parameters that vary, while the others are held fixed, in each case and their corre-sponding figures.

Costs

Modeled Costs are part of the reported headline results.

Permanent impact γ and temporary impact η are incorporated into the transacted price eSt = St − η vt.

Source sentence
The term γ(xt −x0), for γ ≥0, is usually referred to as the permanent impact. The slippage −ηvt, for η ≥0, is also referred to as the temporary impact.

Finds 5

  • Closed‑form optimal feedback strategies are obtained for the utility maximization problem. theoretical
    Source sentence
    We formulate the broker's problem as a utility maximization problem in which the broker seeks to maximize his utility of excess profit-and-loss at the execution horizon, of which optimal feedback strategies are obtained in closed form.
  • In the absence of execution risk, optimal strategies are deterministic and can be expressed as affine combinations of unit IS and unit TC orders. theoretical
    Source sentence
    In the absence of execution risk, the optimal strategies subject to reference strategies are deterministic. We establish an affine structure among the trading trajectories under optimal strategies subject to general reference strategies using implementation shortfall (IS) and target close (TC) orders as basis.
  • Numerical experiments show the optimal strategy consistently achieves higher average terminal wealth and utility than TWAP, with lower variability. performance
    Source sentence
    We observe that the optimal 22 Table 2: Stress testing parameters m0 η ρ θ σ β Baseline 0.05 10 0 0.002 200 1e6 Scenario 1 (large σ) 0.05 10 0 0.002 2000 1e6 Scenario 2 (large β) 0.05 10 0 0.002 200 1e7 Scenario 3 (large m0) 5 10 0 0.002 200 1e6 Scenario 4 (small η) 0.05 0.1 0 0.002 200 1e6 strategy not only consistently achieves higher average terminal wealth and utility than those under TWAP but also bears lower
  • Increasing price volatility σ makes the optimal strategy track the reference strategy more closely. performance
    Source sentence
    Thus, to mitigate the risk incurred from the price volatility during execution, the optimal strategy tracks more closely to the reference strategy.
  • Stress‑test results indicate the optimal strategy remains robust under extreme parameter scenarios, outperforming TWAP in both mean P&L and volatility. performance
    Source sentence
    Apparently, in all the scenarios the histograms under optimal strategies are concentrated around zero whereas those under TWAP strategies are much more widely spreading. We conclude that, even under extreme situations, the optimal strategies outperform TWAP strategies not only in higher averaged P&L but also with lower volatility risk, which shows the robustness of our results.

Checked by others

Contradicted by

  • Optimal execution on Uniswap v2/v3 under transient price impact (2026)
    On the question: How does increasing price volatility σ affect the optimal execution schedule?
    Compare the findings

    This paper: Increasing price volatility σ makes the optimal strategy track the reference strategy more closely.

    That paper: When volatility becomes large, the optimal strategy converges to a fully front‑loaded schedule.

    A: higher σ → strategy tracks reference more closely. B: higher σ → strategy becomes fully front‑loaded, contradicting A.

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 (34)

  1. Stochastic controls: Hamiltonian systems and HJB equations (2012)
  2. Linear-exponential-quadratic Gaussian control (2013)
  3. Optimal execution of portfolio transactions (2001)
  4. Execution costs (2010)
  5. Target close execution strategies — link
  6. The Financial Mathematics of Market Liquidity: From optimal execution to market making (2016)
  7. Optimal execution of a VWAP order: a stochastic control approach (2015)
  8. VWAP execution and guaranteed VWAP (2014)
and 26 more
  1. Value under liquidation (1999)
  2. Optimal execution with nonlinear impact functions and trading-enhanced risk (2003) — link
  3. Optimal control of execution costs (1998)
  4. Adaptive arrival price (2007)
  5. Optimal trade execution: a mean quadratic variation approach (2012)
  6. Mean–variance optimal adaptive execution (2011)
  7. Risk aversion and the dynamics of optimal liquidation strategies in illiquid markets (2009)
  8. Optimal basket liquidation for CARA investors is deterministic (2010)
  9. Comparison between the mean-variance optimal and the mean-quadratic-variation optimal trading strategies (2013)
  10. Optimal execution with uncertain order fills in Almgren–Chriss framework (2017)
  11. Optimal execution with dynamic risk adjustment (2019)
  12. Optimal execution with non-linear transient market impact (2017)
  13. Transient linear price impact and Fredholm integral equations (2012)
  14. No-dynamic-arbitrage and market impact (2010)
  15. Optimal execution with quadratic variation inventories (2023)
  16. A closed-form execution strategy to target volume weighted average price (2016)
  17. Optimal execution in Hong Kong given a market-on-close benchmark (2018)
  18. Constraint Ornstein-Uhlenbeck bridges (2017) — link
  19. The self-financing equation in limit order book markets (2019) — link
  20. A new risk-sensitive maximum principle (2005)
  21. Unwinding stochastic order flow: When to warehouse trades — → 2310.14144
  22. Accelerated share repurchase: pricing and execution strategy (2015)
  23. Accelerated share repurchases (2011)
  24. Brokers and informed traders: dealing with toxic flow and extracting trading signals (2022)
  25. CFA Program Curriculum 2018 Level II (2017)
  26. Optimum consumption and portfolio rules in a continuous-time model (1975)

1 of 34 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.