This study explores the communication patterns of Slovak banks with stakeholders through mandatory disclosures mandated by Basel III's Pillar 3 framework and annual reports in 2007-2022. Our primary objective is to identify key topics communicated by banks and analysing the sentiment of this communication during turbulent periods (i.e., alternating periods of stability and crisis) in 2007-2022. Textual data was collected from Pillar 3 disclosures, annual reports, and additional regulatory reports. A hybrid model was developed to extract the most important keywords from each collected document chapter. This hybrid model (model combining multiple approaches) combines elements of statistical approaches to keyword extraction, (keyword frequency dictionary), linguistic approaches (pair-of-speech tagging in order to select noun-phrases), and machine-learning based approaches (BERT) to extract meaningful keywords. Subsequently, a sentiment analysis was performed on the extracted keywords using a Loughran-McDonald lexicon (list of words labelled with sentiment) specially designed for financial texts. Based on the adjusted univariate results, we can reject the global null hypothesis of independence of the sentiment category of keywords from time for negative sentiment at p = 0.0000 for positive sentiment at p = 0.0005, and for neutral sentiment at p = 0.0000 significant level. The multilevel comparison revealed that negative sentiment was most frequent during the global financial crisis and the COVID-19 pandemic, likely impacting stakeholder confidence and trust. Conversely, positive sentiment dominated during periods of financial stability, potentially enhancing stakeholder satisfaction and investment decisions. This research points out that the sentiment of the selected commercial bank documents changes depending on the years. A commercial bank can use this knowledge and include sentiment information as predictors when modelling financial distress. For bank management of selected commercial bank the examined documents are an important communication tool, the wording of which can have a significant impact on stakeholder behaviour towards the bank, their styling is very important.

Comenius University in Bratislava Bratislava Slovakia

Department of Computer Science Constantine the Philosopher University in Nitra Nitra Slovakia

Science and Research Centre University of Pardubice Pardubice Czech Republic

Zobrazit více v PubMed

Munk M, Pilkova A, Benko L, Blažeková P. Pillar 3: market discipline of the key stakeholders in cee commercial bank and turbulent times. Journal of Business Economics and Management. 2017;18(5):954–73. doi: 10.3846/16111699.2017.1360388 DOI

Pilková A, Munk M, Benko Ľ, Blažeková P, Kapusta J. Pillar 3: Does banking regulation support stakeholders’ interest in banks financial and risk profile? PLoS One. 2021;16(10):e0258449. doi: 10.1371/journal.pone.0258449 PubMed DOI PMC

Nopp C, Hanbury A. Detecting Risks in the Banking System by Sentiment Analysis. Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Lisbon, Portugal: Association for Computational Linguistics; 2015. p. 591–600. doi: 10.18653/v1/D15-1071 DOI

Correa R, Garud K, Londono JM, Mislang N. Sentiment in Central Bank’s Financial Stability Reports. Internationalfinance. 2017;2017.0(1203):1–46. doi: 10.17016/ifdp.2017.1203 DOI

Dong M, Jondeau E, Rockinger GM. Textual Analysis of Banks’ Pillar 3 Documents. SSRN Journal. 2019. doi: 10.2139/ssrn.3365005 DOI

Kubaščíková Z, Tumpach M, Juhászová, Zuzana. Contextual non-financial information analysis of annual reports. 2021.

Kubaščíková Z, Juhászová Z, Tumpach. Annual Reports Non-Financial Textual Analysis. Prague; 2021.

Kubaščíková Z, Tumpach M, Juhászová Z, Zeytinoglu E, Chyzhevska L. Contextual Non-financial Analysis of Annual Reports of Slovak, Turkish and Ukrainian Companies. SHS Web of Conf. 2021;92:02033. doi: 10.1051/shsconf/20219202033 DOI

Sharma P, Li Y. Self-Supervised Contextual Keyword and Keyphrase Retrieval with Self-Labelling. Mathematics & Computer Science. 2019. doi: 10.20944/preprints201908.0073.v1 DOI

El-Beltagy SR, Rafea A. KP-Miner: A keyphrase extraction system for English and Arabic documents. Information Systems. 2009;34(1):132–44. doi: 10.1016/j.is.2008.05.002 DOI

Campos R, Mangaravite V, Pasquali A, Jorge A, Nunes C, Jatowt A. YAKE! Keyword extraction from single documents using multiple local features. Information Sciences. 2020;509:257–89. doi: 10.1016/j.ins.2019.09.013 DOI

Pasi G. Advances in Information Retrieval: 40th European Conference on IR Research, ECIR 2018, Grenoble, France, March 26-29, 2018, Proceedings. Cham: Springer International Publishing; AG; 2018.

Rose S, Engel D, Cramer N, Cowley W. Automatic Keyword Extraction from Individual Documents. 1st ed. In: Berry MW, Kogan J, editors. Text Mining. 1st ed. Wiley; 2010. p. 1–20. doi: 10.1002/9780470689646.ch1 DOI

Mehrabi E, Mohebi A, Ahmadi A. Improved Keyword Extraction for Persian Academic Texts Using RAKE Algorithm; Case Study: Persian Theses and Dissertations. jipm. 2021;37(1):197–228. doi: 10.52547/jipm.37.1.197 DOI

Zhang L, Wang W, Ma J, Wen Y. IWF-TextRank Keyword Extraction Algorithm Modelling. Applied Sciences. 2024;14(22):10657. doi: 10.3390/app142210657 DOI

Zhang M, Li X, Yue S, Yang L. An Empirical Study of TextRank for Keyword Extraction. IEEE Access. 2020;8:178849–58. doi: 10.1109/access.2020.3027567 DOI

Bennani-Smires K, Musat C, Hossmann A, Baeriswyl M, Jaggi M. Simple Unsupervised Keyphrase Extraction using Sentence Embeddings. arXiv; 2018. doi: 10.48550/ARXIV.1801.04470 DOI

Huang Q. Research on Keywords Extraction of Film Reviews Based on the KeyBERT Model. TCSISR. 2024;5:732–8. doi: 10.62051/1zpndy68 DOI

Zhang J. Hotel review analysis based on LDA model and KeyBert model. TE. 2024;1(8). doi: 10.61173/rqsp2623 DOI

Sun Y, Qiu H, Zheng Y, Wang Z, Zhang C. SIFRank: A New Baseline for Unsupervised Keyphrase Extraction Based on Pre-Trained Language Model. IEEE Access. 2020;8:10896–906. doi: 10.1109/access.2020.2965087 DOI

Joseph J, Vineetha S, Sobhana NV. A survey on deep learning based sentiment analysis. Materials Today: Proceedings. 2022;58:456–60. doi: 10.1016/j.matpr.2022.02.483 DOI

Karanikola A, Davrazos G, Liapis CM, Kotsiantis S. Financial sentiment analysis: Classic methods vs. deep learning models. IDT. 2023;17(4):893–915. doi: 10.3233/idt-230478 DOI

Araci D. FinBERT: Financial sentiment analysis with pre-trained language models. arXiv. 2019. doi: 10.48550/ARXIV.1908.10063 DOI

Kelebercová L, Zozuk NČ. Sentiment classification of annual reports in Slovak language using symbolic, subsymbolic and statistic methods. Procedia Computer Science. 2023;225:3508–16. doi: 10.1016/j.procs.2023.10.346 DOI

Bodnaruk A, Loughran T, McDonald B. Using 10-K Text to Gauge Financial Constraints. SSRN Journal. 2013. doi: 10.2139/ssrn.2331544 DOI

Loughran T, McDonald B. Measuring Firm Complexity. SSRN Journal. 2020. doi: 10.2139/ssrn.3645372 DOI

Yang Y, Zhang K, Fan Y. Analyzing Firm Reports for Volatility Prediction: A Knowledge-Driven Text-Embedding Approach. INFORMS Journal on Computing. 2022;34(1):522–40. doi: 10.1287/ijoc.2020.1046 DOI

Li X, Wu P, Wang W. Incorporating stock prices and news sentiments for stock market prediction: A case of Hong Kong. Information Processing & Management. 2020;57(5):102212. doi: 10.1016/j.ipm.2020.102212 DOI

Mullings R. Do central bank sentiment shocks affect liquidity within the European Monetary Union? A computational linguistics approach. The European Journal of Finance. 2022;29(12):1355–81. doi: 10.1080/1351847x.2022.2124530 DOI

Taylor-Neu K, Rahaman AS, Saxton GD, Neu D. Tone at the top, corporate irresponsibility and the Enron emails. AAAJ. 2024;37: 336–64. doi: 10.1108/AAAJ-12-2023-6792 DOI

Tetlock PC, Saar‐Tsechansky M, Macskassy S. More Than Words: Quantifying Language to Measure Firms’ Fundamentals. The Journal of Finance. 2008;63(3):1437–67. doi: 10.1111/j.1540-6261.2008.01362.x DOI

Stone P, Dunphy D, Smith M, Ogilvie D. The General Inquirer: A Computer Approach to Content Analysis. MIT Press; 1966.

Consoli S, Barbaglia L, Manzan S. Fine-Grained, Aspect-Based Sentiment Analysis on Economic and Financial Lexicon. SSRN Journal. 2021. doi: 10.2139/ssrn.3766194 DOI

Bos T, Frasincar F. Automatically Building Financial Sentiment Lexicons While Accounting for Negation. Cogn Comput. 2021;14(1):442–60. doi: 10.1007/s12559-021-09833-w DOI

Rizinski M, Peshov H, Mishev K, Jovanovik M, Trajanov D. Sentiment Analysis in Finance: From Transformers Back to eXplainable Lexicons (XLex). IEEE Access. 2024;12:7170–98. doi: 10.1109/access.2024.3349970 DOI

Park J, Lee HJ, Cho S. Automatic construction of context-aware sentiment lexicon in the financial domain using direction-dependent words. arXiv. 2021. doi: 10.48550/ARXIV.2106.05723 DOI

Vine S. Pdfplumber. 2023. Available: https://github.com/jsvine/pdfplumber

Qi P, Zhang Y, Zhang Y, Bolton J, Manning CD. Stanza: A Python Natural Language Processing Toolkit for Many Human Languages. arXiv. 2020. doi: 10.48550/ARXIV.2003.07082 DOI

Devlin J, Chang M-W, Lee K, Toutanova K. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. arXiv; 2018. doi: 10.48550/ARXIV.1810.04805 DOI

Conneau A, Khandelwal K, Goyal N, Chaudhary V, Wenzek G, Guzmán F, et al. Unsupervised Cross-lingual Representation Learning at Scale. arXiv; 2019. doi: 10.48550/ARXIV.1911.02116 DOI

Pikuliak M, Grivalský Š, Konôpka M, Blšták M, Tamajka M, Bachratý V, et al. SlovakBERT: Slovak Masked Language Model. arXiv; 2021. doi: 10.48550/ARXIV.2109.15254 DOI