Incorporating financial reports and deep learning for financial distress prediction: empirical evidence from Chinese listed companies

This study conducts a comparative study on text information processing methods for financial distress prediction. Word2vec and bidirectional encoder representations from transformers (BERT) are employed to convert financial reports into vectors. Weighted word2vec and BERT-sentence are also used for enhanced text processing and report quantification. Experimental results based on a data set of 62 312 Chinese listed companies from 2000 to 2021 show that weighted word2vec achieves an average prediction accuracy of 85.27% in cross-validation and 84.67% in sliding-time- window validation. The findings indicate that incorporating semantic information from management discussion and analysis (MD&A) significantly improves the performance of distress prediction models for listed companies, regardless of the text-processing technique used. Text-based features become comparable with financial indicators and even surpass them as the prediction horizon extends. Combination features offer greater enhancement than financial indicators, especially for longer prediction horizons. We therefore offer a comprehensive validation of the MD&A for the purpose of predicting financial distress, and we firmly believe that it serves as a valuable tool in mitigating risk within financial risk management.