Paper: arXiv 2609.37872

Authors: Andrey Itkin

Abstract

We model market impact as the response to submitted order flow net of counterflow from latent traders, activated when price displacements from the level that would prevail without the order exceed individual thresholds. Order flow depletes this pool, and a generalized Langevin equation governs its recovery over several time scales. Its memory kernels are finite sums of exponentials, so its Markovian lift is exact rather than an approximation. For an undepleted pool, aggregation under explicit assumptions on individual trading responses yields an intermediate square-root regime between linear small- and large-order limits, without imposing a square-root impact law. Scaling thresholds and responses with price noise makes impact in this regime proportional to volatility, and thresholds that grow with the execution horizon make it independent of duration. With constant displayed depth, expected round-trip costs are nonnegative under the log-price convention, independently of the memory. Numerical experiments show that depletion narrows the square-root range and that memory spectra producing similar single-order impacts can respond differently after substantial prior trading. Calibration to market data is left to a companion paper.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: This paper presents a highly mathematical and novel approach to modeling market impact using a generalized Langevin equation. While the theoretical framework is robust, the empirical rigor is currently low as calibration to market data is explicitly left for a companion paper, limiting immediate practical validation.

Research Flowchart

  flowchart TD
    A["Research Goal: Model Market Impact"] --> B{"Key Methodology: Generalized Langevin Equation"};
    B --> C{"Inputs: Latent Traders & Order Flow Assumptions"};
    C --> D["Computational Process: Aggregation & Numerical Experiments"];
    D --> E{"Outcomes: Concave Price Impact & Square-Root Regime"};
    E --> F["Further Outcomes: Volatility/Duration Dependence, Nonnegative Round-trip Costs"];

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 generalized Langevin model of latent liquidity that captures market impact as the response to order flow net of a counterflow from latent traders, with exact Markovian lift and finite-sum exponential memory kernels.

Source sentence
We model market impact as the response to submitted order flow net of counterflow from latent traders, activated when price displacements from the level that would prevail without the order exceed individual thresholds.

Assumes 6

  • Counterflow is a signed function of displacement with the same sign as the displacement (A odd, |A| nondecreasing). model
    Source sentence
    The counterflow of a fresh pool is a signed function of the displacement, qcf = A(D), sgn A(D) = sgn D, (19) with A odd and |A| nondecreasing in |D|.
  • A counterparty acts only when the magnitude of the displacement exceeds an individual threshold χ ≥ 0. model
    Source sentence
    A counterparty acts only when the magnitude of the displacement exceeds an individual threshold χ ≥0.
  • The latent pool dynamics follow a generalized Langevin equation with finite-sum exponential memory kernels and bounded Lipschitz amplitudes. model
    Source sentence
    Assume the potential Eq. (18), the finite kernels Eq. (10) and Eq. (11) with bounded Lipschitz amplitudes cj, a bounded pool map ρ with bounded derivative, and a normalized threshold measure ˆΠ of finite mass.
  • The reference log-price mt is a martingale. model
    Source sentence
    The proof of Proposition 3 uses the martingale property of the log reference mt in Eq. (3).
Show 2 more
  • Round‑trip schedules have zero net position at the horizon (∫0^H q_t dt = 0). execution
    Source sentence
    (34) A round trip is a schedule with zero net position at the horizon H, that is, R H 0 qt dt = 0.
  • The depth L0 is constant. model
    Source sentence
    With constant displayed depth, expected round-trip costs are nonnegative under the log-price convention, independently of the memory.

Data

No dataset named in the text.

Costs

Not applicable The paper reports no trading or execution result; the cost the extractor found is a model parameter.

The paper-text extraction found a cost parameter inside a model (Round‑trip costs are shown to be non‑negative under the log‑price convention, independent of the memory spectrum.); the classifier found no reported trading result, so costs are not applicable.

Finds 5

  • Round‑trip trading costs are non‑negative for any bounded predictable schedule with zero net position, regardless of memory kernels or threshold measures. theoretical
    Source sentence
    Then C[q] = L0 2 E D2 H  + E "Z H 0 qcf t Dt dt # ≥0.
  • An intermediate square‑root impact regime emerges between linear small‑ and large‑order limits without imposing a square‑root law. theoretical
    Source sentence
    For an undepleted pool, aggregation under explicit assumptions on individual trading responses yields an intermediate square-root regime between linear small- and large-order limits, without imposing a square-root impact law.
  • Impact in the square‑root regime scales with volatility when thresholds and responses are scaled with price noise, and becomes independent of execution duration when thresholds grow with horizon. theoretical
    Source sentence
    Scaling thresholds and responses with price noise makes impact in this regime proportional to volatility, and thresholds that grow with the execution horizon make it independent of duration.
  • Depletion of the latent pool narrows the range over which square‑root impact holds, and memory spectra that give similar single‑order impacts can behave differently after substantial prior trading. theoretical
    Source sentence
    Numerical experiments show that depletion narrows the square-root range and that memory spectra producing similar single-order impacts can respond differently after substantial prior trading.
  • The memory spectrum cannot create a profitable round‑trip; neither the displacement nor the counterflow term can be negative. theoretical
    Source sentence
    Neither term can be made negative by the evolution of the latent state, so the memory spectrum cannot create a profitable round trip.

Checked by others

Supported by

  • Price impact in equity auctions: zero, then linear (2023)
    On the question: How does price impact scale with order size in equity auctions?
    Compare the findings

    This paper: An intermediate square‑root impact regime emerges between linear small‑ and large‑order limits without imposing a square‑root law.

    That paper: Post‑clearing instantaneous impact is sub‑linear, transitioning from linear to square‑root behavior as order size grows.

    Both report a zero‑impact start, then a linear regime, followed by an intermediate square‑root impact; A observes the transition, B derives the square‑root without imposing it.

  • The two square root laws of market impact and the role of sophisticated market participants (2023)
    On the question: How does price impact scale with order size or participation rate?
    Compare the findings

    This paper: An intermediate square‑root impact regime emerges between linear small‑ and large‑order limits without imposing a square‑root law.

    That paper: For a fixed executed volume Q, impact is proportional to √γ when the participation rate γ is large, and becomes linear in γ for small γ.

    Both report a √ scaling regime (A: √γ for large γ, B: intermediate √ impact) and linear scaling for small orders; B derives the √ regime without assuming it.

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 (171)

  1. A fractional reaction–diffusion description of supply and demand (2018)
  2. Market impact with multi-timescale liquidity (2018) — link
  3. Agent-based models for latent liquidity and concave price impact (2014)
  4. Continuous auctions and insider trading (1985)
  5. Optimal execution of portfolio transactions (2001)
  6. No-dynamic-arbitrage and market impact (2010)
  7. Anomalous price impact and the critical nature of liquidity in financial markets (2011)
  8. Fluctuations and response in financial markets: the subtle nature of `random’ price changes (2004)
and 163 more
  1. The long memory of the efficient market (2004)
  2. How markets slowly digest changes in supply and demand (2009)
  3. Capital Market Equilibrium with Transaction Costs (1986) — link
  4. Portfolio Selection with Transaction Costs (1990) — link
  5. Price Manipulation and Quasi-Arbitrage (2004)
  6. Optimal Trading Strategy and Supply/Demand Dynamics (2013)
  7. Optimal Execution Strategies in Limit Order Books with General Shape Functions (2010)
  8. Order Book Resilience, Price Manipulation, and the Positive Portfolio Problem (2012)
  9. Optimal Trade Execution and Price Manipulation in Order Books with Time-Varying Liquidity (2014)
  10. Optimal Execution and Price Manipulations in Time-Varying Limit Order Books (2014)
  11. How efficiency shapes market impact (2013) — link
  12. The two square root laws of market impact and the role of sophisticated market participants — → 2311.18283
  13. Crossover from linear to square-root market impact (2019) — link
  14. Optimal trade execution and absence of price manipulations in limit order book models (2010) — link
  15. A L angevin approach to stock market fluctuations and crashes (1998) — link
  16. The price impact of order book events (2014) — link
  17. Strong convergence rates for Markovian representations of fractional Brownian motion (2021) — link
  18. Heston’s stochastic volatility model implementation, calibration and some extensions (2003)
  19. Simple and efficient simulation of the H eston stochastic volatility model (2008)
  20. The Heston Model with Time-Dependent Correlation Driven by Isospectral Flows (2021)
  21. The H eston Model and Its Extensionsin in VBA (2015)
  22. On the distribution of extended CIR model (2018)
  23. Pricing derivatives in stochastic volatility models using the finite difference method (2002)
  24. Algorithm AS 155: the distribution of a linear combination of chi-square random variables (1980)
  25. Computing the distribution of quadratic forms: further comparisons between the Liu-Tang-Zhang approximation and exact methods (2010)
  26. Option Valuation Using the Fast Fourier Transform (1999)
  27. Pricing options under stochastic volatility: a power series approach (2009)
  28. Financial modelling with jump processes (2004)
  29. Statistical inference for time-changed Levy processes via Mellin transform approach (2016)
  30. The H eston stochastic volatility model with piecewise constant parameters - efficient calibration and pricing of window barrier options (2018)
  31. On pricing barrier options and exotic variations (2018) — link
  32. On the numerical solution of Fredholm integral equations utilizing the local radial basis function method (2019)
  33. Radial Basis Functions with Applications to the Geosciences (2015)
  34. Rough Volatility Literature (2022) — link
  35. Rough multifactor volatility for SPX and VIX options (2021) — link
  36. The SINC way: a fast and accurate approach to Fourier pricing (2021) — link
  37. Pricing under rough volatility (2016)
  38. Asymptotic analysis for stochastic volatility: martingale expansion (2011)
  39. Rational approximation of the rough Heston solution. (2019)
  40. Malliavin Calculus in Finance: Theory and Practice (2021)
  41. ADOL - Markovian approximation of rough lognormal model (2019) — link
  42. ADOL - Markovian approximation of rough lognormal model (2019)
  43. Things we think we know (2019) — link
  44. Makovian approximations of stochastic Volterra equations with the fractional kernel (2021) — link
  45. The step stochastic volatility model (2021)
  46. Extreme at-the-money skew in a local volatility model
  47. private communication (2020)
  48. Representation of a fractional Brownian motion in terms of an infinite–dimensional Ornstein–Uhlenbeck process (2011)
  49. First Baruch Volatility Workshop (2015) — link
  50. Short–time near–the–money skew in rough fractional volatility models (2018)
  51. On the short-time behavior of the implied volatility forjump-diffusion models with stochastic volatility (2007)
  52. Rough volatility: fact or artefact? (2022) — link
  53. Volatility is rough (2018)
  54. On the short-time behavior of the implied volatility for jump-diffusion models with stochastic volatility (2007)
  55. Short-time at-the-money skew and rough fractional volatility (2017)
  56. The microstructural foundations of leverage effect and rough volatility (2018)
  57. Convergence of heavy-tailed H awkes processes and the microstructure of rough volatility — → 2312.08784
  58. No-arbitrage implies power-law market impact and rough volatility (2020)
  59. Multifactor approximation of rough volatility models (2019)
  60. Affine V olterra processes (2019)
  61. Markovian approximations of stochastic V olterra equations with the fractional kernel (2023)
  62. Log-modulated rough stochastic volatility models (2021)
  63. The rough H awkes H eston stochastic volatility model (2024)
  64. Nonequilibrium Statistical Mechanics (2001)
  65. The generalized L angevin equation in harmonic potentials: anomalous diffusion and equipartition of energy (2022)
  66. Equality connecting energy dissipation with a violation of the fluctuation-response relation (2005)
  67. Energy flow, partial equilibration, and effective temperatures in systems with slow dynamics (1997)
  68. Volterra Integral and Functional Equations (1990)
  69. Short time behavior of the ATM implied skew in the ADO-H eston model (2024) — link
  70. ADOL : M arkovian approximation of the rough lognormal model (2019)
  71. Does the term-structure of equity at-the-money skew really follow a power law? (2023)
  72. Rough volatility: fact or artefact? (2024)
  73. Malliavin Calculus in Finance: Theory and Practice (2021)
  74. Joint SPX-VIX calibration with G aussian polynomial volatility models: deep pricing with quantization hints (2025)
  75. Pricing and calibration in the 4-factor path-dependent volatility model — → 2406.02319
  76. Data-driven parameterization of the generalized L angevin equation (2016)
  77. Efficient B ayesian estimation of the generalized L angevin equation from data (2024) — link
  78. Error analysis of generalized L angevin equations with approximated memory kernels (2025) — link
  79. The moment formula for implied volatility at extreme strikes (2004)
  80. Leverage effect in financial markets: the retarded volatility model (2001)
  81. Forward start volatility swaps in rough volatility models (2025) — link
  82. Estimating the H urst parameter from the zero vanna implied volatility and its dual — → 2510.26310
  83. Estimating the H urst parameter from short term volatility swaps: a M alliavin calculus approach (2019)
  84. An intuitive introduction to fractional and rough volatilities (2021)
  85. Time reversal invariance in finance (2009)
  86. Market heterogeneities and the causal structure of volatility (2003)
  87. The fine-structure of volatility feedback I : multi-scale self-reflexivity (2014)
  88. Quadratic H awkes processes for financial prices (2017)
  89. From quadratic H awkes processes to super- H eston rough volatility models with Z umbach effect (2021)
  90. Marketron games: self-propelling stocks vs dumb money and metastable dynamics of the good, bad and ugly markets — → 2501.12676
  91. Marketron Through the Looking Glass: From Equity Dynamics to Option Pricing in Incomplete Markets — → 2508.09863
  92. Indifference pricing and hedging in stochastic volatility models (2007)
  93. Utility indifference pricing: an overview (2009)
  94. An example of indifference prices under exponential preferences (2004)
  95. Beyond rough volatility: decoupling memory and scaling via a generalized Langevin equation (2026) — link
  96. A generalized Langevin model for volatility II : incomplete-market pricing and joint SPX-VIX calibration (2026)
  97. Rough fractional diffusions as scaling limits of nearly unstable heavy tailed H awkes processes (2016)
  98. The fluctuation-dissipation theorem (1966)
  99. The stable states picture of chemical reactions. II . Rate constants for condensed and gas phase reaction models (1980)
  100. Brownian motion in a field of force and the diffusion model of chemical reactions (1940)
  101. Reaction-rate theory: fifty years after Kramers (1990)
  102. Diagonal Frog: High-order positivity-preserving FD schemes for anisotropic Fokker-Planck equations (2026) — link
  103. Flux-Corrected Diagonal Frog: second order and positivity at all time steps (2026) — link
  104. Statistical Theory of Open Systems: Volume 1: A Unified Approach to Kinetic Description of Processes in Active Systems (1995)
  105. Benchmark dataset for mid-price forecasting of limit order book data with machine learning methods (2018)
  106. Multifractal random walk (2001)
  107. Fluctuations and response in financial markets: the subtle nature of ‘random’ price changes (2003)
  108. The Volatility Smile Problem (2002)
  109. The characteristic function of rough Heston models (2018) — link
  110. Volatility is Rough (2014) — link
  111. The quadratic rough Heston model and the joint S&P 500/VIX smile calibration problem (2020) — link
  112. Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations (2017) — link
  113. Solving high-dimensional partial differential equations using deep learning (2018) — link
  114. Convergence of the Deep BSDE Method for Coupled FBSDEs (2018) — link
  115. Overcoming the curse of dimensionality in the numerical approximation of semilinear parabolic partial differential equations (2020) — link
  116. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations (2019) — link
  117. Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks (2021) — link
  118. Algorithms for Inverse Reinforcement Learning (2000)
  119. The Joint S&P 500/VIX Smile Calibration Puzzle Solved (2022)
  120. Joint Modeling and Calibration of SPX and VIX by Optimal Transport (2022) — link
  121. The VIX-Derived Volatility Model: A VIX-first Joint SPX-VIX Framework (2026) — link
  122. Calibrating Inelastic Markets to Options: The Lean Marketron and the Generalized Langevin Equation (2026) — link
  123. Dispersion-Constrained Martingale Schr "o dinger Bridges: Joint Entropic Calibration of Stochastic Volatility Models to S&P 500 and VIX Smiles (2025) — link
  124. Joint calibration to SPX and VIX options with signature-based models — → 2301.13235
  125. Neural joint S&P 500/VIX smile calibration (2023) — link
  126. Fast Exact Joint S&P 500/VIX Smile Calibration in Discrete and Continuous Time (2022)
  127. Stochastic Calculus for Finance II : Continuous-Time Models (2004)
  128. Brownian Motion and Stochastic Calculus (1991)
  129. Mixture Kalman Filters (2000) — link
  130. Sequential Monte Carlo Methods in Practice (2001) — link
  131. Marginalized particle filters for mixed linear/nonlinear state-space models (2005)
  132. Local Stochastic Rough Volatility: Pathwise Filtering and the Conditional Density Equation (2026) — link
  133. Calibrating Inelastic Markets to Options: The Lean Marketron and the Generalized Langevin Equation (2026) — link
  134. From rough SDE to rough PDEs : a LSV perspective (2026)
  135. Rough stochastic filtering (2025) — link
  136. Uniqueness of the solution of the filtering equations in spaces of measures (2024) — link
  137. On Stochastic Partial Differential Equations and their applications to Derivative Pricing through a conditional Feynman-Kac formula (2021) — link
  138. Valuing American options and Flexible Forwards contracts in time-dependent models (2026) — link
  139. Rough stochastic differential equations (2021) — link
  140. Nonlinear rough Fokker-Planck equations (2025) — link
  141. Rough PDEs for local stochastic volatility models — → 2307.09216
  142. The smile calibration problem solved (2011)
  143. Volatility is (mostly) path-dependent (2023)
  144. Leverage effect in financial markets: the retarded volatility model (2001)
  145. Mimicking an I t ^o process by a solution of a stochastic differential equation (2013)
  146. Mimicking the One-Dimensional Marginal Distributions of Processes Having an It ^o Differential (1986)
  147. Calibrating Volatility Surfaces via Relative-Entropy Minimization (1997)
  148. Markovian Approximations of Stochastic V olterra Equations with the Fractional Kernel (2023)
  149. On Approximation of Functions by Exponential Sums (2005)
  150. The Laplace Transform (1941)
  151. Bernstein Functions: Theory and Applications (2012)
  152. Data-Driven Learning of the Generalized Langevin Equation with State-Dependent Memory (2024)
  153. Stochastic Flows and Stochastic Differential Equations (1990)
  154. Some extensions of It ^o ’s formula, stochastic flows, and stochastic differential equations (1981)
  155. Equations of Cartoon-type or equations of the theory of conditional Markov processes (1965)
  156. Direct estimation of equity market impact (2005)
  157. The non-linear market impact of large trades: evidence from buy-side order flow (2013)
  158. A fully consistent, minimal model for non-linear market impact (2015)
  159. Hawkes processes in finance (2015)
  160. Transport, Collective Motion, and Brownian Motion (1965)
  161. Quantitative Portfolio Management: The Art and Science of Statistical Arbitrage (2021)
  162. A GLE – LSV model with state-dependent leverage and joint calibration to SPX historical and options data and VIX instruments (2026)
  163. GLE Volatility Models under Incomplete-Market Pricing: From Historical Dynamics to SPX and VIX Derivatives (2026)

8 of 171 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.