Elamur Research⌕
☰
home / papers / 2026 / The Financial Document Causality Detection Shared Task (FinCausal 2026)

The Financial Document Causality Detection Shared Task (FinCausal 2026)

Paper · 2026 · Antonio Moreno-Sandoval, Jordi Porta, Yanco Amor Torterolo Orta et al.

Summary

FinCausal 2026 is the latest round of a long-running shared task held at the Financial Narrative Processing workshop (LREC 2026). It tests whether systems can find cause–effect relations stated inside financial texts. The task keeps the extractive question-answering format introduced in 2025: a system receives a passage and an abstractively worded question and must return the exact answer span from the passage. There are separate English and Spanish subtasks. The data comes from annual reports of IBEX 35 firms in the FinT-esp corpus, with aligned official English translations, plus UK reports from the Lancaster UCREL corpus.

To make the task harder, the organizers:

  • removed trivial or ambiguous items,
  • added over 500 new passages per language built around causal chains of three or more links,
  • reworded 10% of questions to block lexical shortcuts,
  • favored explanatory causality over justificatory causality.

Scoring moved from EM and SAS to an LLM-as-a-judge that assigns 1–5 points by rubric, because correct answers often differ lexically from the references. Nine of 20 registered teams submitted runs. Their scores clustered tightly near the maximum, which the organizers interpret as a mature performance level.

Key points

  • Dataset split: 2,000 training items per language. The test sets have 500 items (English) and 503 items (Spanish), for totals of 2,500 and 2,503. Each item contains an ID, context, question and gold answer, all manually annotated by linguists.

  • Participation: 20 teams registered and 9 submitted official runs. Seven teams entered both subtasks and two entered English only. One further team (YT) described a system without submitting test results.

  • Results:

    • English: HSA CORAL and Sheffield Causal tied for first at 4.814.
    • Spanish: the best score was 4.813, every team exceeded 4.6, and the gap from top to bottom stayed under 0.2.

  • Methods: Most approaches were LLM-based:

    • SFT of GPT-4.1-mini/nano
    • DSPy/MIPROv2 prompt optimization with Gemma3-12B
    • hybrid RAG
    • GPT-5.1 voting ensembles
    • a Qwen2.5 + Gemini candidate/verifier pipeline

    The exception was TU Graz's SpanDiffusion, which applies flow matching on DeBERTa-v3.

  • Leaderboard dynamics: Teams made fast early gains, followed by long plateaus. The Spanish track stabilized early at around 4.75–4.81, while the English track was more volatile near the top.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Universidad Autónoma de MadridMadrid · Spain7 works · 21 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksComputer Vision and Pattern Recognition · 2,869 works

Cite

Cite as Antonio Moreno-Sandoval, Jordi Porta, Yanco Amor Torterolo Orta, Alexia Stanescu, Melina Chatzi and Sofía Roseti. 2026. The Financial Document Causality Detection Shared Task (FinCausal 2026). The 7th Financial Narrative Processing Workshop.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/4zb72h45pvk2

@inproceedings{morenosandoval2026financial,
  title = {{The Financial Document Causality Detection Shared Task (FinCausal 2026)}},
  author = {Antonio Moreno-Sandoval, Jordi Porta, Yanco Amor Torterolo Orta, Alexia Stanescu, Melina Chatzi and Sofía Roseti},
  booktitle = {The 7th Financial Narrative Processing Workshop},
  year = {2026},
  doi = {10.63317/4zb72h45pvk2},
  url = {https://doi.org/10.63317/4zb72h45pvk2}
}

References

25 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Aali Abdullah Alqarni, Mark Stevenson, and Arif Dwi Laksito. 2026. A Comparative Study of RAG Approaches and Fine-Tuning for Causal QA in Financial Text. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Reso… find
  2. Sanae Attak. 2026. Improving Verbatim Financial Causality Extraction with Supervised Fine-Tuning and Prompt Repetition. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Resources Association (ELRA). To appea… find
  3. Ankush Chopra, Shubham Sharma, and Ashmani Kumar. 2026. Extracting Financial Causality: A Multilingual Approach with SLM Fine-Tuning and LLM-Arbitrated Ensembles. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Lang… find
  4. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  5. Markus Freitag, Nitika Mathur, Daniel Deutsch, Chi-Kiu Lo, Eleftherios Avramidis, Ricardo Rei, Brian Thompson, Frederic Blain, Tom Kocmi, Jiayi Wang, David Ifeoluwa Adelani, Marianna Buchicchio, Chrysoula Zerva, and Alon Lavie. 2024. Are LLMs breaking MT metrics? results of the WMT24 metrics shared…
  6. Akash Kumar Gautam, Serhii Hamotskyi, and Christian Hänig. 2026. Causal Connections: Leveraging Multilingual Fine-Tuning for Financial QA@FinCausal 2026. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Reso… find
  7. Aldan Jay, Rafael Berlanda, Yoelvis Moreno, and Vincent Santamarta. 2026. VERSA: Verbatim Extraction via Rephrasing and Self-Aggregation for Financial Causality. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Langu… find
  8. Geza Kovacs, Daniel Deutsch, and Markus Freitag. 2024. Mitigating metric bias in minimum Bayes risk decoding. In Proceedings of the Ninth Conference on Machine Translation, pages 1063–1094, Miami, Florida, USA. Association for Computational Linguistics. find
  9. Dominique Mariko, Hanna Abi Akl, Estelle Labidurie, Stephane Durfort, Hugues de Mazancourt, and Mahmoud El-Haj. 2021. find
  10. The Financial Document Causality Detection Shared Task (FinCausal 2021). In Proceedings of the 3rd Financial Narrative Processing Workshop, pages 58–60, Lancaster, United Kingdom. Association for Computational Linguistics. find
  11. Dominique Mariko, Hanna Abi-Akl, Kim Trottier, and Mahmoud El-Haj. 2022. find
  12. The Financial Causality Extraction Shared Task (FinCausal 2022). In Proceedings of the 4th Financial Narrative Processing Workshop @LREC2022, pages 105– 107, Marseille, France. European Language Resources Association. find
  13. Antonio Moreno-Sandoval, Blanca Carbajo-Coronado, Jordi Porta-Zamorano, Yanco-Amor Torterolo-Orta, and Doaa Samy. 2025. find
  14. The Financial Document Causality Detection Shared Task (FinCausal 2025). In Proceedings of the Joint Workshop of the 9th Financial Technology and Natural Language Processing (FinNLP), the 6th Financial Narrative Processing (FNP), and the 1st Workshop on Large Language Models for Finance and Legal (L… find
  15. Antonio Moreno Sandoval, Ana Gisbert, and Elena Montoro. 2020. FinT-esp: A corpus of financial reports in Spanish. In Miguel Fuster-Márquez, Carmen Gregori-Signes, and José Santaemilia Ruiz, editors, Multiperspectives in analysis and corpus design, pages 89–102. Comares, Granada. find
  16. Antonio Moreno-Sandoval, Jordi Porta-Zamorano, Blanca Carbajo-Coronado, Doaa Samy, Dominique Mariko, and Mahmoud El-Haj. 2023. find
  17. The Financial Document Causality Detection Shared Task (FinCausal 2023). In 2023 IEEE International Conference on Big Data (BigData), pages 2855–2860. find
  18. Georg Niess and Roman Kern. 2026. SpanDiffusion: Flow Matching over Continuous Span Masks for Financial Causal Question Answering. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Resources Association (ELRA…
  19. OpenAI. 2025. gpt-oss-120b gpt-oss-20b model card.
  20. Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing.
  21. Julian Risch, Timo Möller, Julian Gutsch, and Malte Pietsch. 2021. Semantic Answer Similarity for Evaluating Question Answering Models. In Proceedings of the 3rd Workshop on Machine Reading for Question Answering, pages 149–157, Punta Cana, Dominican Republic. Association for Computational Linguisti… find
  22. Bavya Sarda, Pulkit Chatwal, and Sonal Dabral. 2026. QRAFT: QLoRA Retrieval-Augmented Fine-Tuning for Causal Span Extraction in Financial Documents. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Resources…
  23. Zaid Shahrouri, Ayomide Ivienagbor, Idrees Asad, Rijul Shrestha, Yasemin Bal, and Zahaab Nadeem. 2026. LeedsMEng26: Qwen + Gemini for FinCausal 2026 Causality Detection in Financial Narrative Texts. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Pal…
  24. Avinash Trivedi and Mallikarjuna Chindukuri. 2026. Financial Causal QA via Instruction and Prompt Tuning of Gemma3-12B. In Proceedings of the 7th Financial Narrative Processing Workshop (FNP 2026) at LREC-COLING 2026, Palma de Mallorca, Spain. European Language Resources Association (ELRA). To appea…
  25. FinCausal 2026 Rubric find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/sandoval-2026-financial-document-causality-detection-shared/