Paper: arXiv 2401.05414
Authors: Xinshuai Dong, Haoyue Dai, Yewen Fan, Songyao Jin, Sathyamoorthy Rajendran, Kun Zhang
Abstract
Financial data is generally time series in essence and thus suffers from three fundamental issues: the mismatch in time resolution, the time-varying property of the distribution - nonstationarity, and causal factors that are important but unknown/unobserved. In this paper, we follow a causal perspective to systematically look into these three demons in finance. Specifically, we reexamine these issues in the context of causality, which gives rise to a novel and inspiring understanding of how the issues can be addressed. Following this perspective, we provide systematic solutions to these problems, which hopefully would serve as a foundation for future research in the area.
Complexity vs Empirical Score
- Math Complexity: 7.0/10
- Empirical Rigor: 6.5/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper presents advanced mathematical derivations and causal graph formalisms (SCM, Markov conditions, faithfulness) alongside empirical validation on S&P 100 data with proposed methods like CD-NOD, striking a balance between theoretical depth and data-driven implementation.
Research Flowchart
flowchart TD A["Research Goal<br>Address Three Demons in Finance<br>(Time Resolution, Nonstationarity, Latent Factors)"] --> B["Causal Perspective<br>Framework Selection"] B --> C["Data Input<br>Financial Time Series Data"] C --> D["Methodology<br>Causal Analysis of Demons"] D --> E["Computational Process<br>Addressing Time Resolution Mismatch"] D --> F["Computational Process<br>Handling Nonstationarity & Latent Factors"] E --> G["Key Findings<br>Systematic Causal Solutions"] F --> G G --> H["Outcome<br>Foundation for Future Research<br>in General Financial Markets"]
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 rank‑based latent causal discovery method that leverages rank constraints of the cross‑covariance matrix to uncover underlying causal structure, including latent variables, and provides a systematic solution that works when time‑resolution mismatch, non‑stationarity, and latent
Source sentence
Rank Constraint and Motivation. Rank-based Latent Causal Discovery Algorithm.
Assumes 6
- The underlying data generation follows a linear instantaneous causal model so that log returns can be treated as linearly aggregated i.i.d. observations. model
Source sentence
To deal with time resolution, mismatch, we assume the underlying procedure follows Eq. 1 and we use the log return of each stock, formulated as returnt = log(pricet −pricet−1).
- Noise terms in the structural causal model are non‑Gaussian to enable the use of the Generalized Independent Noise (GIN) condition for direction identification. statistical
Source sentence
Here we aim at making use of Generalized Indepen-dent Noise (GIN) condition (Xie et al., 2020) to decide the directions, by assuming the noise terms ϵVi in Definition 1 are non-gaussian.
- Within each segment identified by change‑point detection the data are stationary; otherwise rank‑deficiency constraints would be invalid. data
Source sentence
5.2 Baysian Change Point Detection To utilize our rank-based latent causal discovery method we need to assume that our data is stationary; otherwise the nonstationary would induce rank deficiency constraints that cannot correctly reflect the underlying causal structure.
- Latent variables exist and act as common children of observed variables (e.g., X1‑X4 share a latent L1). model
Source sentence
9 Dong, Dai, Fan, Jin, Rajendran, and Zhang Example 2 Suppose we observe four variables X1, X2, X3, X4 as common children of a latent variable L1.
Show 2 more
- Observed variables are jointly Gaussian when estimating causal coefficients, allowing a factor‑analysis‑like likelihood formulation. statistical
Source sentence
Similar to the typical factor analysis setting (Gorsuch, 2014), we assume that ϵV are all gaussian and thus X are jointly gaussian, and we find the causal coefficients A by maximizing the log-likelihood of observational data: arg max A,Ψ −n|X| 2 log 2π −n 2 log |ΣX| −n 2 tr((ΣX)−1 ˆΣX), (6) s.t., diagonal entries of ΣL are all 1, (7) where n is the number of i.i.d. observations of X, tr(˙) returns the trace of a
- The latest data are more informative for handling concept drift. market
Source sentence
To handle concept drift, prior works usually rely on the assumption that the latest data contains more useful information.
Data
- SP100 stock price data (log daily and hourly returns) market-prices unknown
Source sentence
4.4 The Whole Procedure and Potential Usage We will apply the proposed rank-based latent causal discovery method to SP100 stock price data, which suffers from the three demos - time resolution mismatch, non-stationarity, and the existence of latent variables. The CD-NOD analysis was conducted on these ten selected stocks, utilizing their log daily return data spanning from 2019 to 2023.
Costs
Not applicable
Finds 5
- The rank‑based latent causal discovery algorithm successfully recovers the Markov equivalence class of the underlying causal graph despite the presence of latent variables. theoretical
Source sentence
4.3 Determining All Directions and Estimating Causal Coefficients Our rank-based latent causal discovery algorithm is able to output the Markove equivalence class of Omin(Os(G)).
- Empirical validation on SP100 data shows that the method identifies four stocks (PFE, BA, AAPL, NFLX) whose causal generating mechanisms change over time. statistical
Source sentence
The discovered causal graph indicates four stocks whose causal generating mechanisms are subject to changes over time: Pfizer (PFE), Boeing (BA), Apple (AAPL), and Netflix (NFLX).
- Applying the method to different time windows (2017, 2019, 2021) reveals how change points such as Covid‑19 affect the underlying causal graph. performance figure from elsewhere in the paper
Source sentence
Our first experiment is to show how the change point, e.g., Covid-19, influences the underlying graph.
- The approach can handle all three demons—time‑resolution mismatch, non‑stationarity, and latent factors—simultaneously, as demonstrated in the empirical study. theoretical
Source sentence
Based on this understanding we propose rigorous methods to address each issue as well as a systematic solution that works even when the three issues are present simultaneously. • Our empirical study validates the effectiveness of our proposed methods in real-life scenarios in finite sample cases.
- Estimated causal coefficients exhibit meaningful relationships, e.g., high coefficient (0.84) between energy companies XOM and CVX, and shared latent structure among finance stocks BAC, USB, JPM. statistical
Source sentence
This also implies that BAC, USB, JPM share something in common that is not shared by other financial stocks. As for the energy companies XOM and CVX, we observe that they are highly related with a high coefficient 0.84.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 21 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 (70)
- Factor analysis: Classic edition (2014)
- Exact and efficient Bayesian inference for multiple changepoint problems (2006)
- Bayesian online changepoint detection (2007) — link
- Modeling changing dependency structure in multivariate time series (2007)
- A Versatile Causal Discovery Framework to Allow Causally-Related Hidden Variables (2023) — link
- Digesting anomalies: An investment approach (2015)
- Calculation of entailed rank constraints in partially non-linear and cyclic models (2013)
- A five-factor asset pricing model (2015)
and 62 more
- A comparison of new factor models (2017)
- Multifactor explanations of asset pricing anomalies (1996)
- On persistence in mutual fund performance (1997)
- Approximating discrete probability distributions with dependence trees (1968)
- Probabilistic R easoning in I ntelligent S ystems: N etworks of P lausible I nference (1988)
- Regime-switching stochastic volatility: Evidence from the crude oil market (2009)
- Learning under concept drift: A review (2018)
- An overview of concept drift applications (2016)
- A survey on concept drift adaptation (2014)
- Paired learners for concept drift (2008)
- Adaptive random forests for evolving data stream classification (2017)
- Mining high-speed data streams (2000)
- DDG-DA: Data Distribution Generation for Predictable Concept Drift Adaptation (2022)
- Addressing Non-Stationarity in FX Trading with Online Model Selection of Offline RL Experts (2022)
- Investigating causal relations by econometric models and cross-spectral methods (1969)
- Temporal aggregation and spurious instantaneous causality in multiple time series models (2002)
- A correspondence principle for simultaneous equation models (1970)
- Causal discovery from temporally aggregated time series (2017)
- Concept drift detection via competence models (2014)
- Inside the black box: A simple guide to quantitative and high frequency trading (2013)
- Asymptotic behaviour of temporal aggregates of time series (1972)
- Systematic sampling and temporal aggregation in time series models (1984)
- Some consequences of temporal aggregation in empirical analysis (1999)
- Temporal aggregation, cointegration and causality inference (2008)
- Further methods of derivation of quarterly figures from annual data (1967)
- A methodological note on the disaggregation of time series totals (1986)
- Estimating the underlying change in unemployment in the UK (2000)
- Temporal disaggregation by state space methods: Dynamic regression methods revisited (2006)
- Causality (2009)
- Causal discovery from heterogeneous/nonstationary data (2020)
- Causation, prediction, and search (2000)
- Optimal structure identification with greedy search (2002)
- A linear non-Gaussian acyclic model for causal discovery. (2006)
- Detecting causal associations in large nonlinear time series datasets, Sci. Adv., 5, eaau4996 (2019)
- Models, reasoning and inference (2000)
- Causal inference in the presence of latent variables and selection bias (2013) — link
- On Reichenbach’s common cause principle and Reichenbach’s notion of common cause (1999)
- A tetrad test for causal indicators. (2000)
- Learning high-dimensional directed acyclic graphs with latent and selection variables (2012)
- Recursive causal structure learning in the presence of latent variables and selection bias (2021)
- Estimation of causal effects using linear non-Gaussian causal models with hidden variables (2008)
- Learning linear non-Gaussian causal models in the presence of latent variables (2020)
- Learning the Structure of Linear Latent Variable Models. (2006)
- Generalized measurement models (2005)
- Trek separation for Gaussian graphical models (2010)
- Calculation of entailed rank constraints in partially non-linear and cyclic models (2013) — link
- Learning linear bayesian networks with latent variables (2013)
- Causal clustering for 1-factor measurement models (2016)
- Triad constraints for learning causal structure of latent variables (2019)
- Generalized independent noise condition for estimating latent variable causal graphs (2020)
- Latent hierarchical causal structure discovery with rank constraints (2022)
- Estimation of linear non-Gaussian acyclic models for latent factors (2009)
- Identification of partially observed linear causal models: Graphical conditions for the non-gaussian and heterogeneous cases (2021)
- Independence Testing-Based Approach to Causal Discovery under Measurement Error and Linear Non-Gaussian Models (2022)
- Identification of linear latent variable model with arbitrary distribution (2022)
- Learning latent causal graphs via mixture oracles (2021)
- Discovery and visualization of nonstationary causal models (2015) — link
- Causal Factor Investing: Can Factor Investing Become Scientific? (2022)
- Causality: Models, Reasoning and Inference (2009)
- The theory of two factors. (1914)
- Kernel-based conditional independence test and application in causal discovery (2012) — link
- Independence, invariance and the causal Markov condition (1999)
0 of 70 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.