Paper: arXiv 2510.06879

Authors: Natascha Hey, Eyal Neuman, Sturmius Tuschmann

Abstract

We introduce an offline nonparametric estimator for concave multi-asset propagator models based on a dataset of correlated price trajectories and metaorders. Compared to parametric models, our framework avoids parameter explosion in the multi-asset case and yields confidence bounds for the estimator. We implement the estimator using both proprietary metaorder data from Capital Fund Management (CFM) and publicly available S&P order flow data, where we augment the former dataset using a metaorder proxy. In particular, we provide unbiased evidence that self-impact is concave and exhibits a shifted power-law decay, and show that the metaorder proxy stabilizes the calibration. Moreover, we find that introducing cross-impact provides a significant gain in explanatory power, with concave specifications outperforming linear ones, suggesting that the square-root law extends to cross-impact. We also measure asymmetric cross-impact between assets driven by relative liquidity differences. Finally, we demonstrate that a shape-constrained projection of the nonparametric kernel not only ensures interpretability but also slightly outperforms established parametric models in terms of predictive accuracy.

Complexity vs Empirical Score

  • Math Complexity: 8.0/10
  • Empirical Rigor: 8.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced nonparametric estimation, concave propagator models, and rigorous theoretical analysis of price manipulation, indicating high math complexity. It demonstrates high empirical rigor through implementation on both proprietary and public datasets, with statistical validation, confidence bounds, and backtest-ready predictive comparisons against parametric benchmarks.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Estimate multi-asset<br>market impact nonparametrically"] --> B["Methodology:<br>Offline nonparametric estimator<br>for concave propagator models"]
  B --> C["Data & Inputs"]
  C --> D["Proprietary CFM Metaorders<br>& Public S&P Order Flow<br>+ Metaorder Proxy"]
  D --> E["Computational Process:<br>Shape-constrained projection<br>of nonparametric kernel"]
  E --> F["Key Outcomes"]
  F --> G["1. Self-impact is concave<br>& decays via shifted power-law"]
  F --> H["2. Cross-impact adds<br>explanatory power; concave > linear"]
  F --> I["3. Asymmetric cross-impact<br>driven by liquidity differences"]
  F --> J["4. Nonparametric model<br>outperforms parametric benchmarks"]

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 an offline nonparametric estimator for concave multi-asset propagator models that avoids parameter explosion and provides confidence bounds.

Source sentence
We introduce an offline nonparametric estimator for concave multi-asset propagator models based on a dataset of correlated price trajectories and metaorders.

Assumes 5

  • The noise terms are conditionally sub‑Gaussian. statistical
    Source sentence
    ... each ϵ(n) is R‑conditionally sub‑Gaussian with respect to Fn‑1.
  • Impact function h is concave (h(x)=sgn(x)|x|c with c∈(0,1]). model
    Source sentence
    where hc(x) := sgn(x)|x|c with c ∈(0, 1] is a concave impact function
  • The true propagator matrix lies in the admissible kernel set Gad. model
    Source sentence
    Suppose that the true propagator matrix G* is contained in the set of admissible kernels Gad from (2.5)
  • Price impact is concave and follows a shifted power‑law decay. market
    Source sentence
    In particular, we provide unbiased evidence that self‑impact is concave and exhibits a shifted power‑law decay
Show 1 more
  • Cross‑impact can be asymmetric due to relative liquidity differences. market
    Source sentence
    We also measure asymmetric cross‑impact between assets driven by relative liquidity differences.

Data

  • CFM metaorders · Capital Fund Management other proprietary
    2012 to 2022 · energy and agricultural futures (oil, gas, metals, corn, wheat, soybeans)
    Source sentence
    We start by integrating proprietary execution data from Capital Fund Management (CFM), consisting of 12,000 metaorders across energy and agricultural futures,
  • Synthetic metaorder proxy for corn futures synthetic unknown
    Corn futures (front, 1‑quarter, 2‑quarter contracts)
    Source sentence
    ... This synthetic enrichment ensures consistency with observed market microstructure features
  • S&P 500 order flow imbalance · LOBSTER other public
    2024 (237 trading days) · 197 constituents of the S&P 500
    Source sentence
    ... includes detailed order book events for all NASDAQ‑traded stocks, in particular, all executed orders.

Recognised sources: LOBSTER

Costs

Unclear - sources disagree The platform classifier and the paper-text extraction give different answers, so no value is asserted.

Platform classifier: Not applicable

Paper text extraction: Modeled - Execution costs are defined as CT(Q)=½∑_{i,j} Q_i^T G_{i‑j} h(Q_j).

Source sentence (the authors' words)
the total execution costs corresponding to Q caused by the self‑ and cross‑impact associated with G are given by CT(Q) :=

Finds 5

  • Self‑impact is concave and decays across multiple timescales, yielding a 94.4% R² improvement over linear models. statistical
    Source sentence
    Self‑impact is shown to be concave and to decay across multiple timescales (a relative R2 improvement of 94.4% from a linear to a concave model
  • Cross‑impact is concave and improves predictive R² by 1.74% compared to linear cross‑impact. other
    Source sentence
    Table 5 highlights that cross‑impact is concave (a relative R2 improvement of 1.74% from linear to concave cross‑impact).
  • The shape‑constrained projected nonparametric kernel slightly outperforms established parametric models in predictive accuracy. other
    Source sentence
    Finally, we demonstrate that a shape‑constrained projection of the non‑parametric kernel
  • Introducing cross‑impact provides a significant gain in explanatory power, with concave specifications outperforming linear ones. statistical
    Source sentence
    We find that introducing cross‑impact provides a significant gain in explanatory power, with concave specifications outperforming linear ones
  • For aggregate order‑flow data, a shifted power‑law kernel yields the best fit and improves out‑of‑sample R² by up to ~89% versus linear models. other
    Source sentence
    ... a relative out‑of‑sample R2 improvement of 84.39%−89.05% depending on the prediction horizon compared to the linear case

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 45 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers in the archive (2 by independent authors): Price manipulation in nonlinear transient impact models: … (uses its method or data)

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.