Financial performance prediction and a research on enterprises in BIST SME Industry Index

In the field of finance, there are many studies conducted with the definitions such as financial failure prediction, financial distress prediction, and bankruptcy prediction model as failure prediction studies of enterprises. The reason for the intensive financial failure studies is that the results of business failures concern the entire society at both micro and macro levels. It is thought to be the right proportion between the development and economic robustness of countries and the numbers of financially successful businesses. Therefore, the phenomenon of bankruptcy of enterprises, which is a result of the process of financial failure, reveals the importance of early warning models. The aim of the study is to develop reliable prediction models with DA, LRA, YSA and C5.0 models for one year before the financial failure year of the enterprises covered by the BIST SME INDUSTRY INDEX and to determine the financial ratios or ratios that are important in determining the financial successful business and unsuccessful businesses. In the first part of the study, the definitions of SMEs and the importance of SMEs in the country's economies were discussed. The second part includes the definition of financial failure concept, its types, causes and forecasting methods used in financial failure prediction. In the third section, financial failure prediction models have been developed with the data set created from the five-year financial data of 38 enterprises within the scope of SME INDUSTRY INDEX. The performances of the prediction models, which were developed taking into consideration the year of financial failure of the enterprises, were compared. Independent variables that are important in the classification of businesses as financial successful or unsuccessful were determined according to the prediction models used. The difference of the study from others, i) In addition to the financial ratios commonly used in failure studies, the use of cash flow data from the cash flow statements of businesses, business, investment and financing activities as categorical independent variables, ii) The Intellectual Value Coefficient (VAIC) developed by Ante Pulic and its components are calculated and added for each business. 70% of the data set used in the study was divided into training, 15% into testing and 15% into validation subsets; the models developed with the training set were tested with data sets that prediction models had never seen before. Confusion matrix, Receiver Operating Characteristic (ROC AUC) and Press's Q test were used in the evaluation of the prediction models. In addition, the financial ratios of the businesses included in the data set were also tested with the Z Score model and the corrected Z' Score model developed by Altman for manufacturing enterprises. In the study, -ANN model has a higher degree of predictive ability than other models, -The most important independent variable is gross profit margin, -Efficiency of the capital used (CCE) positively affects business performance, -Cash flow from operating activities (IFI) reduces the risk of failure, -Altman Z Scores produced significant results as of the study period, -The results were obtained that the leverage ratio and the ratio of short-term debts increase the risk of failure.