(Prediction Model for Delistings: Empirical Evidence from Borsa İstanbul)
Keywords:
Involuntary Delisting, Voluntary Delisting, Early Warning SystemAbstract
Purpose – The aim of the research is to develop a model to predict involuntary delisting from Borsa Istanbul and to test its predictive power. Design/methodology/approach – The scope of the research consists of 79 delisted firms in Borsa Istanbul in the period of 2000-2018. Sample of the research, covering control group, is 147 firms. The model is based on the literature of failure. The variables selected for the model are classified in six dimensions: “liquidity”, “profitability”, “leverage and solvency”, “efficiency”, “market multiples” and “size”. In the research, considering the limited size of data in the research sample, the logistic regression method is applied. While determining the variables to be included in the logistic model, in addition to correlation coefficients, the dimensions of the variables and the number of uses in the literature are also taken into account. At the end of the selection process among 22 financial ratios, 11 financial ratios are used in the model as independent variables. Findings – Model results are found to be compatible with expectations and literature. According to the most appropriate cut-off value of 0.5, the model correctly classifies 83% of the observations. 79% of the signals given by the model regarding delisting is realized. On the other hand, delisting that the model could not predict (missing signals) is limited at 13%. Discussion – When the accurate classification of the models in the studies using the method of logistic regression in the literature of failure (between 77% and 95%) is taken into consideration, the success level of the applied model (83%) is acceptable. The findings reveal that early warning models can be used for predicting delisting.
Downloads
Published
How to Cite
Issue
Section
License
![Creative Commons License](http://i.creativecommons.org/l/by-nd/4.0/88x31.png)
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.