EconBase
← All papers

Transformer-based CoVaR: Systemic Risk in Textual Information

Junyu Chen, Tom Boot, Lingwei Kong, Weining Wang

arXiv 13 Feb 2026 · Econometrics

arXiv:2602.12490 · PDF · DOI · OpenAlex · Extracted main text

Abstract

Conditional Value-at-Risk (CoVaR) quantifies systemic financial risk by measuring the loss quantile of one asset, conditional on another asset experiencing distress. We develop a Transformer-based methodology that integrates financial news articles directly with market data to improve CoVaR estimates. Unlike approaches that use predefined sentiment scores, our method incorporates raw text embeddings generated by a large language model (LLM). We prove explicit error bounds for our Transformer CoVaR estimator, showing that accurate CoVaR learning is possible even with small datasets. Using U.S. market returns and Reuters news items from 2006--2013, our out-of-sample results show that textual information impacts the CoVaR forecasts. With better predictive performance, we identify a pronounced negative dip during market stress periods across several equity assets when comparing the Transformer-based CoVaR to both the CoVaR without text and the CoVaR using traditional sentiment measures. Our results show that textual data can be used to effectively model systemic risk without requiring prohibitively large data sets.

Citation extraction

83
references
140
in-text mentions
83
distinct cited
2
self-citations
28,522
main-text words

appendix boundary found by none_found · 100% of the source is main text. Read the extracted text to check this.

Most heavily cited references

The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.

ReferenceIntensityMentionsSectionsMain text
1Edelman, Benjamin and Goel, Surbhi and Kakade, Sham and Zhang, Cyril (2022) Inductive Biases and Variable Creation in Self-Attention Mechanisms1.000205100%
2Adrian, Tobias and Brunnermeier, Markus K (2016) CoVaR1.00083100%
3Schmidt-Hieber, Johannes (2020) Nonparametric regression using deep neural networks with ReLU activation function1.00064100%
4Wolfgang Karl Härdle and Weining Wang and Lining Yu (2016) TENET: Tail-Event driven NETwork risk self0.92843100%
5Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Ja… (2017) Attention is All You Need0.92843100%
6Oscar Hernan Madrid Padilla and Wesley Tansey and Yanzhen Chen (2022) Quantile regression with ReLU Networks: Estimators and minimax rates0.92843100%
7Tetlock, Paul (2007) Giving Content to Investor Sentiment: The Role of Media in the Stock Market0.84333100%
8Loughran, Tim and McDonald, Bill (2011) When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks0.84333100%
9Keilbar, Georg and Wang, Weining (2022) Modelling Systemic Risk Using Neural Network Quantile Regression self0.84333100%
10Trauger, Jacob and Tewari, Ambuj (2024) Sequence Length Independent Norm-Based Generalization Bounds for Transformers0.81142100%

Showing the top 10 of 83 scored citations.