Paper: arXiv 2505.12806
Authors: Cameron Cornell, Lewis Mitchell, Matthew Roughan
Abstract
Causal networks offer an intuitive framework to understand influence structures within time series systems. However, the presence of cycles can obscure dynamic relationships and hinder hierarchical analysis. These networks are typically identified through multivariate predictive modelling, but enforcing acyclic constraints significantly increases computational and analytical complexity. Despite recent advances, there remains a lack of simple, flexible approaches that are easily tailorable to specific problem instances. We propose an evolutionary approach to fitting acyclic vector autoregressive processes and introduces a novel hierarchical representation that directly models structural elements within a time series system. On simulated datasets, our model retains most of the predictive accuracy of unconstrained models and outperforms permutation-based alternatives. When applied to a dataset of 100 cryptocurrency return series, our method generates acyclic causal networks capturing key structural properties of the unconstrained model. The acyclic networks are approximately sub-graphs of the unconstrained networks, and most of the removed links originate from low-influence nodes. Given the high levels of feature preservation, we conclude that this cryptocurrency price system functions largely hierarchically. Our findings demonstrate a flexible, intuitive approach for identifying hierarchical causal networks in time series systems, with broad applications to fields like econometrics and social network analysis.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 6.0/10 (capped from 7.0: no out-of-sample test; see penalty flags)
- Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 7.0)
Why this score: The paper introduces a novel evolutionary algorithm for fitting acyclic VAR processes, involving complex hierarchical graph theory and optimization, while also presenting empirical results on cryptocurrency data and simulation benchmarks, demonstrating backtest-ready methodology.
Research Flowchart
flowchart TD A["Research Goal:<br>Hierarchical Acyclic<br>Causal Networks"] --> B["Methodology:<br>Evolutionary Approach<br>HEAVe Model"] B --> C["Data:<br>100 Cryptocurrency<br>Return Series"] C --> D["Computational Process:<br>Fit Acyclic VARs<br>Preserve Structural Properties"] D --> E["Outcome 1:<br>High Feature Preservation<br>vs Unconstrained Models"] D --> F["Outcome 2:<br>Hierarchical Structure<br>Identified in Crypto System"] E & F --> G["Conclusion:<br>Flexible Method for<br>Time Series Analysis"]
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 evolutionary approach to fitting acyclic vector autoregressive processes and introduces a novel hierarchical representation for causal networks.
Source sentence
We propose an evolutionary approach to fitting acyclic vector autoregressive processes and introduces a novel hierarchical representation that directly models structural elements within a time series system.
Assumes 3
- The analysis is limited to VAR(1) processes, assuming higher-order lags do not occur independently of a lag‑1 dependence. model
Source sentence
To streamline our discussion and estimation of causal networks, we limit our analysis to VAR(1) processes and omit the index p from our discussion, with A = A1 unless otherwise specified.
- Any causal link i→j will first manifest in order‑1 effects; lag p > 1 effects will not occur independently of a p = 1 dependence. model
Source sentence
For the empirical networks in Section 5 we operate under the assumption that any causal link i →j will first manifest in order-1 effects, and that lag p > 1 effects will not occur independently of a p = 1 dependence.
- The error term εt is often assumed to be Gaussian, though this is not a strict requirement. statistical
Source sentence
The εt is often assumed to be Gaussian, however this is not a strict requirement.
Data
cryptocurrency return series · crypto
- cryptocurrency return series crypto-onchain unknown
Source sentence
When applied to a dataset of 100 cryptocurrency return series, our method generates acyclic causal networks capturing key structural properties of the unconstrained model.
- simulated datasets synthetic unknown
Source sentence
On simulated datasets, our model retains most of the predictive accuracy of unconstrained models and outperforms permutation-based alternatives.
Costs
Not applicable
Finds 3
- The identified acyclic networks retain 92% of the predictive accuracy of unconstrained models while containing 90.2% as many links. predictive-accuracy
Source sentence
In an empirical study of 100 cryptocurrencies, the identified acyclic networks retain 92% of the predictive accuracy of unconstrained models while
- HEAVe retains 92% of the predictive power of unconstrained VAR models in simulations, with 5.6‑8.3% improved prediction and 6.1‑14.1% improved hierarchy identification. predictive-accuracy
Source sentence
A simulation study demonstrating the capacity for HEAVe to produce acyclic VAR models that retain 92% of the predictive power of unconstrained VAR. The results show clear advantages to the hierarchical approach, with 5.6-8.3% improved prediction, 6.1-14.1% improved hierarchy identification...
- The hierarchical evolutionary algorithm outperforms permutation‑based evolutionary algorithms in retaining predictive capacity and achieving higher link precision and more accurate hierarchy detection. performance
Source sentence
HEAVe is validated on synthetic networks of varying sizes, where it retains a larger percentage of the unconstrained model’s predictive capacity compared to permutation‑based EAs.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 14 related papers). Coverage is partial, so this is neither replication nor refutation.
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 (87)
- Macroeconomics and Reality (1980)
- Measures of Causality in Complex Datasets with Application to Financial Data (2014)
- Networks: an introduction (2010)
- Bayesian Graphical Models for Structural Vector Autoregressive Processes (2016)
- Clustering in Complex Directed Networks (2007)
- Cryptocurrency market structure: connecting emotions and economics (2019)
- Predicting future stock market structure by combining social and financial network information (2019)
- Dominating Clasp of the Financial Sector Revealed by Partial Correlation Analysis of the Stock Market (2010)
and 79 more
- Statistical analysis of financial networks (2005)
- Clustering and information in correlation based financial networks (2003)
- Structural Entropy: Monitoring Correlation-Based Networks Over Time With Application To Financial Markets (2019)
- Information Network Modeling for U.S. Banking Systemic Risk (2020)
- Econometric Measures of Connectedness and Systemic Risk in the Finance and Insurance Sectors (2011)
- On the Network Topology of Variance Decompositions: Measuring the Connectedness of Financial Firms (2011)
- NETS: Network Estimation for Time Series (2018)
- Bayesian Graphical Models for Structural Vector Autoregressive Processes (2012)
- Network based evidence of the financial impact of Covid -19 pandemic (2021)
- An Analysis of Variance Test for Normality (Complete Samples) (1965)
- Causation Entropy Identifies Indirect Influences, Dominance of Neighbors and Anticipatory Couplings (2014)
- Cryptocurrencies: An unconventional challenge to the AML/CFT regulators? (2014)
- Vector autoregression and causality: a theoretical overview and simulation study (1994)
- Vector Autoregression and Causality: A Theoretical Overview and Simulation Study (1991)
- Emergence of Scaling in Random Networks (1999)
- Granger Causality and Transfer Entropy Are Equivalent for Gaussian Variables (2009) — link
- The New Introduction to Multiple Time Series Analysis (2005) — link
- The Advanced Theory of Statistics, Volume 2: Inference and Relationship (1973)
- Cryptocurrencies, Mainstream Asset Classes and Risk Factors - A Study of Connectedness (2018)
- Modelling Determinants of Cryptocurrency Prices: A Bayesian Network Approach (2023) — link
- Causality and dynamic spillovers among cryptocurrencies and currency markets (2020)
- Causal inference between cryptocurrency narratives and prices: Evidence from a complex dynamic ecosystem (2020)
- NETS: Network estimation for time series (2019) — link
- Econometric Measures of Connectedness and Systemic Risk in the Finance and Insurance Sectors (2011)
- A representation theory for a class of vector autoregressive models for fractional processes (2008)
- Robust Estimation of the Vector Autoregressive Model by a Least Trimmed Squares Procedure (2008)
- Robust estimation for vector autoregressive models (2013)
- A discussion on the robust vector autoregressive models: novel evidence from safe haven assets (2022)
- Cryptocurrencies: market analysis and perspectives (2020)
- Identifying and Analyzing Cryptocurrency Manipulations in Social Media (2021)
- How Elon Musk’s Twitter activity moves cryptocurrency markets (2023) — link
- Vector Autoregression in Cryptocurrency Markets: Unraveling Complex Causal Networks — → 2308.15769
- The Use of the Rank Transform in Regression (1979)
- Rank regression: an alternative regression approach for data with outliers (2014)
- The impact of COVID-19 on the degree of dependence and structure of risk-return relationship: A quantile regression approach (2020) — link
- Does implied volatility (or fear index) affect Islamic stock returns and conventional stock returns differently? Wavelet-based granger-causality, asymmetric quantile regression and NARDL approaches (2022) — link
- CAViaR: Conditional Autoregressive Value at Risk by Regression Quantiles (2004) — link
- Robust Tests for Heteroscedasticity Based on Regression Quantiles (1982)
- Regression Quantiles (1978)
- Calculation of Quantile Regression Standard Errors (1993)
- Econometric Analysis (2012)
- On the Distribution of a Studentized Quantile (1988)
- Efficiency, multifractality, and the long-memory property of the Bitcoin market: A comparative analysis with stock, currency, and gold markets (2018) — link
- The structure and degree of dependence: A quantile regression approach (2013) — link
- Causality in quantiles and dynamic stock return–volume relations (2009) — link
- Does uncertainty move the gold price? New evidence from a nonparametric causality-in-quantiles test (2016) — link
- Distributional predictability between commodity spot and futures: Evidence from nonparametric causality-in-quantiles tests (2019) — link
- Asymmetric dependence between economic policy uncertainty and stock market returns in G7 and BRIC: A quantile regression approach (2018) — link
- Quantile Autoregression (2006) — link
- Network quantile autoregression (2019) — link
- Asymmetric volatility in cryptocurrency markets: New evidence from smooth transition GARCH models (2020) — link
- Volatility dynamics of crypto-currencies’ returns: Evidence from asymmetric and long memory GARCH models (2020) — link
- GARCH Modelling of Cryptocurrencies (2017) — link
- Performance of ARCH and GARCH Models in Forecasting Cryptocurrency Market Volatility (2021) — link
- Statsmodels: Econometric and statistical modeling with python (2010)
- Network vector autoregression (2017) — link
- Rank is all you need: development and analysis of robust causal networks (2024) — link
- Rank Is All You Need: Robust Estimation of Complex Causal Networks (2024)
- Learning Bayesian Networks by Evolution for Classifier Combination (2009) — link
- A Hybrid Evolutionary Algorithm for Bayesian Networks Learning: An Application to Classifier Combination (2010)
- Learning Dynamic Bayesian Networks Using Evolutionary MCMC (2006) — link
- An improved evolutionary approach-based hybrid algorithm for Bayesian network structure learning in dynamic constrained search space (2020) — link
- Towards a More E.cient Evolutionary Induction of Bayesian Networks (2002)
- A Study on the Evolution of Bayesian Network Graph Structures (2007) — link
- Bayesian Networks: A Model of Self-activated Memory for Evidential Reasoning (1985) — link
- Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference (1988)
- Causality: Models, Reasoning and Inference (2009)
- Artificial Intelligence: A Modern Approach (2010)
- Cryptocurrencies, Mainstream Asset Classes and Risk Factors - A Study of Connectedness (2018) — link
- Enhancing Causal Discovery in Financial Networks with Piecewise Quantile Regression — → 2408.12210
- An integrated vector error correction and directed acyclic graph method for investigating contemporaneous causalities (2023) — link
- Vector Autoregressions, Policy Analysis, and Directed Acyclic Graphs: An Application to the U.S. Economy (2003) — link
- Modeling US soy-based markets with directed acyclic graphs and bernanke structural var methods: the impacts of high soy meal and soybean prices (2004)
- Causality and Price Discovery: An Application of Directed Acyclic Graphs (2004)
- Causation, Prediction, and Search (1993) — link
- Introduction To Evolutionary Computing (2003) — link
- Central bank information effects and transatlantic spillovers (2022) — link
- Explaining monetary spillovers: The matrix reloaded (2023)
- Investigating Causal Relations by Econometric Models and Cross-Spectral Methods (1969) — link
2 of 87 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.