Application of intellectual capital in SME bankruptcy.

Previous studies indicate that applying solely financial ratios (FR) provides limited SME bankruptcy prediction performance. On the other hand, the application of non-financial features is cost-ineffective for SMEs. Intellectual capital (IC) features provide a meaningful alternative to analyse SMEs' financial performance since companies with higher IC regularly achieve consistently higher sales growth. This paper aims to examine the possibilities of applying intellectual capital features in predicting SME bankruptcy. 14 IC features and 27 FR of 54,003 SMEs from 2016 to 2021 were collected from financial statements. Three groups of XGBoost and CatBoost models were developed – with only IC features, with only FR and combining FR and IC features. The results show that IC features are a practical addition to FR, with the best AUC equal to 89%, and their combination outperforms models using only FR by an average of 2.3%. Moreover, IC features such as capital employed efficiency and structural capital reduced the likelihood of SMEs' bankruptcy. The implication of this study is that SME bankruptcy models perform better using IC features without significantly increasing financial cost or processing time, which can be helpful for financial institutions. Additionally, this can contribute to developing other methods of measuring IC. [ABSTRACT FROM AUTHOR]