The applicability of textual disclosures and selected bankruptcy prediction models in assessing the going concern risk of listed firms in Kenya

Abstract

Going concern risk disclosures by companies in financial reports as required by the IFRS helps investors and financial analysts to establish if going concern is in jeopardy. The management complies with IFRS but still companied end up being delisted due to going concern issues. This study was guided by three objectives: to establish if there are textual disclosures on going concern risk by listed firms in Kenya, to determine the extent of prediction of going concern risk using the selected bankruptcy predicting models on listed firms in Kenya and to establish whether the textual disclosures and selected bankruptcy predicting models are statistically significant in assessing the going concern risk of listed firms in Kenya. The population of the study comprised of firms listed at the NSE from 2000 to 2015. Proportionate sampling was adopted where six firms that were either delisted or placed under statutory management were analysed alongside the corresponding thirteen going concern firms from the same sectors. The annual financial reports collected from the Capital Market Authority (CMA) resource centre were analyzed five years prior to going concern risk. Data was presented in tables and statements. In a sample of 13 going concern firms and 6 non going concern firms, the study established that there were textual disclosures for the entire period of analysis. On the extent of prediction of going concern risk using the selected bankruptcy predicting models, the study found that the means of the going concern firms in all the sectors were above the cut off scores at 1.10 for Altman revised four variable model, above 0.862 for Springate model and above 0 for Fulmer model. Sector wise the results showed that the three selected bankruptcy prediction models can, to a significant extent assess going concern risk in the manufacturing, commercial and telecommunication sectors. Results from the regression matrix for both samples, showed a positive relationship between the selected bankruptcy predicting models in assessing the going concern risk. Further, using the T test results failed to accept the null hypothesis at 5% significant level as the p values were all below 0.05, implying that the selected bankruptcy predicting models were not significant in assessing the going concern risk. Finally, results from the T test, provided a p value of 0.902 in a going concern sample and 0.810 in a non going concern sample and showed that textual disclosures are not statistically significant in assessing going concern. The study failed to reject the null hypothesis that textual disclosures were not significant in assessing the going concern risk. This study recommends that textual disclosures be used alongside selected bankruptcy prediction models in financial management decisions.