Paper: arXiv 2408.15846
Authors: Ruijie Tang
Abstract
This study investigates the application of causal discovery algorithms in equity markets, with a focus on their potential to build investment strategies. An investment strategy was developed based on the causal structures identified by these algorithms. The performance of the strategy is evaluated based on the profitability and effectiveness in stock markets. The results indicate that causal discovery algorithms can successfully uncover actionable causal relationships in large markets, leading to profitable investment outcomes. However, the research also identifies a critical challenge: the computational complexity and scalability of these algorithms when dealing with large datasets. This challenge presents practical limitations for their application in real-world market analysis.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 6.0/10 (capped from 9.0: transaction costs only partially modelled; see penalty flags)
- Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 9.0)
Why this score: The paper employs advanced statistical models (VarLiNGAM, TiMINo) with substantial mathematical formulation and proofs, meeting the high math complexity threshold; it also demonstrates high empirical rigor by using real-world multi-market data, implementing a full backtesting workflow, and providing open-source code for reproducibility.
Research Flowchart
flowchart TD
A["Research Goal: Apply Causal Discovery to Equity Markets for Profitable Strategies"] --> B["Data Input: Time Series of Equity Market Returns"]
B --> C["Methodology: Causal Discovery Algorithms"]
C --> D{"Computational Process: Scalability & Complexity"}
D -- Issues with Large Datasets --> C
D -- Successful Identification --> E["Outcome: Identified Causal Structures"]
E --> F["Application: Investment Strategy Construction"]
F --> G{"Evaluation: Profitability & Effectiveness"}
G -- Yes --> H["Key Finding: Actionable Causal Relationships & Profits"]
G -- No/Partial --> I["Key Finding: Limitations in Scalability"]
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 workflow that uses causal discovery algorithms to build a trading strategy based on identified causal structures in stock price data.
Source sentence
An investment strategy was developed based on the causal structures identified by these algorithms.
Assumes 6
- The market is liquid and short selling is permitted without borrowing costs. market
Source sentence
For simplicity, we assume a liquid market where short selling is permitted without incurring borrowing costs.
- A fixed daily transaction cost of 0.1% is applied to trades. execution
Source sentence
we impose a fixed daily transaction cost of 0.1% [7].
- The trading strategy is dollar‑neutral with equal weighting on long and short sides. execution
Source sentence
our trading strategy employs a dollar‑neutral portfolio, ensuring that the dollar amounts of long and short positions are balanced
- Linear regression is used for one‑day‑ahead predictions. model
Source sentence
using the one‑day‑ahead predictions obtained through regression
Show 2 more
- VarLiNGAM assumes linear relationships, non‑Gaussian independent errors, and no latent confounders. model
Source sentence
The LiNGAM algorithm assumes that each variable is a linear function of its causes plus an error term ei, which follows a non‑Gaussian distribution with non‑zero variance. Additionally, the error terms ei are independent across variables, and the system is assumed to contain no latent confounders.
- Time lag attributes are removed from tsFCI and VarLiNGAM causal graphs for analysis. data
Source sentence
the causal graphs generated by tsFCI and VarLiNGAM are compressed by removing the time lag attributes for subsequent analysis.
Data
- SP500 component stocks market-prices public
Source sentence
The data, sourced from [15], covers a period of 10 years (2009.09.01 - 2019.12.31) and includes prices for SP500 and CSI300 component stocks.
- CSI300 component stocks market-prices public
Source sentence
The data, sourced from [15], covers a period of 10 years (2009.09.01 - 2019.12.31) and includes prices for SP500 and CSI300 component stocks.
- Nancy Pelosi portfolio market-prices public
Source sentence
We collected stock prices from her current portfolio, comprising 12 stocks, over the past five years from Yahoo Finance.
Costs
Unclear - sources disagree
Source sentence (the authors' words)
we impose a fixed daily transaction cost of 0.1% [7].
Finds 5
- VarLiNGAM is the most effective algorithm, handling large datasets well and generating the highest returns across all datasets. performance
Source sentence
Among the three algorithms, VarLiNGAM is the most effective, handling large datasets well and generating the highest returns on all datasets.
- Incorporating a causal discovery phase generally improves portfolio returns compared to a self‑cause‑only baseline. performance
Source sentence
Causal discovery proves more effective than the self‑cause‑only setting. The portfolio supported by causal discovery significantly outperforms the self‑cause‑only portfolio.
- Causal discovery is feasible for large asset classes (SP500, CSI300) but ineffective for the small Pelosi portfolio. negative
Source sentence
causal discovery proves ineffective when applied to smaller asset classes like Pelosi, where the true driving forces of its component stocks fall outside the analysis scope.
- VarLiNGAM is the only algorithm that can process 400–500‑stock datasets; tsFCI and TiMINo are limited to much smaller datasets. performance
Source sentence
VarLiNGAM is the only feasible option for analyzing the stock market with 400–500 stocks, while tsFCI and TiMINo are constrained to processing much smaller datasets.
- Using a smaller time lag and selecting 1%‑6% of stocks as winners yields the best returns for the trading strategy. performance
Source sentence
For our trading strategy, setting the number of winner stocks to 1%-6% of the size of the original asset class yields the best returns.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 35 related papers). Coverage is partial, so this is neither replication nor refutation.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (24)
- Detecting the lead–lag effect in stock markets: definition, patterns, and investment strategies (2022)
- Causal inference on time series using restricted structural equation models (2013)
- Survey and evaluation of causal discovery methods for time series (2022)
- Returns to buying winners and selling losers: Implications for stock market efficiency (1993)
- Algorithmic trading and quantitative strategies (2020)
- Momentum: what do we know 30 years after Jegadeesh and Titman’s seminal paper? (2023)
- Causal discovery from heterogeneous/nonstationary data (2020)
- Behind distribution shift: Mining driving forces of changes and causal arrows (2017)
and 16 more
- Do industries explain momentum? (1999)
- Time series momentum (2012)
- Quantitative equity portfolio management: An active approach to portfolio construction and management (2006)
- Value and momentum everywhere (2013)
- Testing for causality: A personal viewpoint (1980)
- Some recent development in a concept of causality (1988)
- On causal discovery from time series data using FCI (2010)
- Estimation of a structural vector autoregression model using non-gaussianity. (2010)
- A linear non-Gaussian acyclic model for causal discovery. (2006)
- Trading costs (2018)
- Deep learning with long short-term memory networks for financial market predictions (2018)
- Evaluating multiple classifiers for stock price direction prediction (2015)
- Elements of causal inference: foundations and learning algorithms (2017)
- Causation, prediction, and search (2001)
- GPU acceleration for information-theoretic constraint-based causal discovery (2022)
- Accelerating constraint-based causal discovery by shifting speed bottleneck (2022)
0 of 24 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.