Paper: arXiv 2312.16185

Authors: Haochun Ma, Davide Prosperino, Alexander Haluszczynski, Christoph Räth

Abstract

Identifying and quantifying co-dependence between financial instruments is a key challenge for researchers and practitioners in the financial industry. Linear measures such as the Pearson correlation are still widely used today, although their limited explanatory power is well known. In this paper we present a much more general framework for assessing co-dependencies by identifying and interpreting linear and nonlinear causalities in the complex system of financial markets. To do so, we use two different causal inference methods, transfer entropy and convergent cross-mapping, and employ Fourier transform surrogates to separate their linear and nonlinear contributions. We find that stock indices in Germany and the U.S. exhibit a significant degree of nonlinear causality and that correlation, while a very good proxy for linear causality, disregards nonlinear effects and hence underestimates causality itself. The presented framework enables the measurement of nonlinear causality, the correlation-causality fallacy, and motivates how causality can be used for inferring market signals, pair trading, and risk management of portfolios. Our results suggest that linear and nonlinear causality can be used as early warning indicators of abnormal market behavior, allowing for better trading strategies and risk management.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced nonlinear causality metrics (transfer entropy, convergent cross-mapping) and statistical physics tools like Fourier surrogates, indicating high mathematical density. It also applies these methods to extensive historical financial data (DAX and Dow-Jones) with rolling-window backtests, demonstrating significant empirical implementation.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Assess linear & nonlinear causality in financial markets"] --> B["Data Input:<br>Stock Indices (Germany & U.S.)"];
  B --> C{"Methodology Setup"};
  C --> D["Transfer Entropy<br>Convergent Cross-Mapping"];
  C --> E["Fourier Transform Surrogates<br>Isolate Linear vs. Nonlinear"];
  D --> F["Computational Process:<br>Quantify Causality & Correlation"];
  E --> F;
  F --> G{"Key Findings/Outcomes"};
  G --> H["Nonlinear Causality Exists<br>Linear correlation underestimates total causality"];
  G --> I["Applications:<br>Early warning, Pair trading, Risk management"];

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 general framework to assess co‑dependencies by identifying and interpreting linear and nonlinear causalities in financial markets.

Source sentence
In this paper we present a much more general framework for assessing co-dependencies by identifying and interpreting linear and nonlinear causalities in the complex system of financial markets.

Assumes 4

  • Survival bias is negligible for the analysis. data
    Source sentence
    We would like to note that the survival bias16 is negligible for our analysis.
  • Logarithmic returns are stationary. statistical
    Source sentence
    To ensure stationary time series, we convert the stock prices pt to logarithmic returns: xt = log pt −log pt−1 .
  • The analysis uses daily closing prices of the selected stocks. data
    Source sentence
    our data consists of the daily closing prices of all stocks that were in the index through April 20, 2022
  • Correlation serves as a good proxy for linear causality. model
    Source sentence
    we find that while correlation is a good proxy for linear causality

Data

  • German DAX and U.S. Dow‑Jones indices (30 highest‑capitalized stocks) market-prices public
    1973-01-19 to 2022-04-20 · 30 highest‑capitalized companies in Germany and the U.S.
    Source sentence
    For our real‑world analysis, we select a subset of stocks from the DAX and Dow‑Jones indices that represent the 30 highest capitalized and thus most influential companies in Germany and the U.S., respectively. Beginning on January 19, 1973, our data consists of the daily closing prices of all stocks that were in the in

Costs

Not modeled A trading or execution result is reported without costs.

Finds 3

  • Stock indices in Germany and the U.S. exhibit a significant degree of nonlinear causality. statistical
    Source sentence
    We find that stock indices in Germany and the U.S. exhibit a significant degree of nonlinear causality
  • Correlation is a good proxy for linear causality but neglects a significant amount of nonlinear causality. statistical
    Source sentence
    while correlation is a good proxy for linear causality, nonlinear effects are disregarded and thus, significant amount of nonlinear causality is neglected
  • Integrating causality measures into market signal inference, pair trading, and portfolio construction yields superior results. performance
    Source sentence
    we propose a simple integration of causality measures into market signal inference, pair trading, and portfolio construction routines and show that they yield superior results

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 27 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 4 papers in the archive (4 by independent authors): Causal Hierarchy in the Financial Market Network -- … (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.

Cited References (57)

  1. Embedding Riemannian manifolds by their heat kernel (1994)
  2. Detecting causality in complex ecosystems (2012)
  3. Convergent cross mapping: theory and an example (2018)
  4. Linear and nonlinear market correlations: Characterizing financial crises and portfolio optimization (2017)
  5. Surrogates with random Fourier phases (2009)
  6. Generating surrogate data for time series with several simultaneously measured variables (1994)
  7. Causal network reconstruction from time series: From theoretical assumptions to practical estimation (2018)
  8. Measuring information transfer (2000)
and 49 more
  1. Granger causality and transfer entropy are equivalent for Gaussian variables (2009)
  2. Determining embedding dimension for phase-space reconstruction using a geometrical construction (1992)
  3. Improved surrogate data for nonlinearity tests (1999) — link
  4. Generating surrogate data for time series with several simultaneously measured variables (1994) — link
  5. rEDM: An R package for Empirical Dynamic Modeling and Convergent Cross Mapping
  6. Essays in econometrics: collected papers of Clive WJ Granger (2001)
  7. Convergent cross-mapping and pairwise asymmetric inference (2014)
  8. Wiener–Granger causality: a well established methodology (2011)
  9. To Understand the Wild U.S. Stock Rally, Just Forget About 2020 (2020) — link
  10. Revisiting algorithms for generating surrogate time series (2012)
  11. Information theoretic test for nonlinearity in time series (1993)
  12. Directionality of coupling from bivariate time series: How to avoid false causalities and missed connections (2007)
  13. Reliability of inference of directed climate networks using conditional mutual information (2013)
  14. Estimating good discrete partitions from observed data: Symbolic false nearest neighbors (2003)
  15. Non-linear dependence and teleconnections in climate data: sources, relevance, nonstationarity (2014)
  16. Evaluation and Extension of the Transfer Entropy Calculus for the Measurement of Information Flows Between Futures Time Series During the COVID-19 Pandemic (2021)
  17. Introduction to econophysics: correlations and complexity in finance (1999)
  18. Scaling behaviour in the dynamics of an economic index (1995)
  19. When financial economics influences physics: The role of econophysics (2018)
  20. Turbulent cascades in foreign exchange markets (1996)
  21. Options, futures and other derivatives (2019)
  22. Pearson correlation coefficient (2009)
  23. Survivorship bias in performance studies (1992)
  24. Sample size requirements for estimating Pearson, Kendall and Spearman correlations (2000)
  25. Nonlinear dynamics in financial markets: evidence and implications (1995)
  26. Hierarchical structure in financial markets (1999)
  27. Foundations of portfolio theory (1991)
  28. Empirical properties of asset returns: stylized facts and statistical issues (2001)
  29. Measuring skewness: a forgotten statistic? (2011)
  30. The coupled logistic map: a simple model for the effects of spatial heterogeneity on population dynamics (1995)
  31. Random matrix theory analysis of cross-correlations in the US stock market: Evidence from Pearson’s correlation coefficient and detrended cross-correlation coefficient (2013)
  32. Correlation structure and evolution of world stock markets: Evidence from Pearson and partial correlation-based networks (2018)
  33. A short note on the size of the dot-com bubble (2006)
  34. The VIX ``Volmaggedon’’, with Exchange-Traded Notes Destabilizing the Market (2021)
  35. Rolling analysis of time series (2003)
  36. Detecting strange attractors in turbulence (2006)
  37. Dynamic causality analysis using overlapped sliding windows based on the extended convergent cross-mapping (2021)
  38. On the efficacy of state space reconstruction methods in determining causality (2015)
  39. Effects of noise on transfer entropy estimation for damage detection (2009)
  40. Inferring Direct Causality from Noisy Data using Convergent Cross Mapping (2019)
  41. The fast Fourier transform and its applications (1988)
  42. On the use of r and r squared in correlation and regression (2019)
  43. A review of the Granger-causality fallacy (2015)
  44. Calculation of average mutual information (AMI) and false-nearest neighbors (FNN) for the estimation of embedding parameters of multidimensional time series in matlab (2018)
  45. Pairs Trading: quantitative methods and analysis (2004)
  46. Options futures and other derivatives (2003)
  47. VII. Note on regression and inheritance in the case of two parents (1895)
  48. Causal inference from noisy time-series data–testing the convergent cross-mapping algorithm in the presence of noise and external influence (2017)
  49. An overview of value at risk (1997)

0 of 57 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.