Publication: Finansal Başarısızlığı Belirlemede İstatistiksel Yöntemlerin Sınıflandırma Performanslarının Karşılaştırılması: Borsa İstanbul'da Bir Uygulama
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Abstract
Teknolojinin hızla gelişmesi ve rekabet koşullarının zorlaşması nedeniyle işletmelerin varlığını devam ettirebilmelerinin ve faaliyetlerini sürdürebilmelerinin gün geçtikçe zorlaştığı görülmektedir. Piyasa koşullarının zorlaşması nedeniyle ve ekonomideki dalgalanmaların etkisiyle krizler yaşanmaktadır. Bu yüzden, işletmeler finansal başarısızlıkla karşı karşıya kalmaktadır. Ancak, işletmeler ani bir şekilde finansal başarısızlığa maruz kalmamakta, öncesinde bu duruma ilişkin bazı sinyaller vermektedir. Bu durum, finansal başarısızlığın önceden belirlenmesinin ve bu belirlemeyi sağlayan önem seviyesi yüksek göstergelerin tespitinin önemini ortaya koymaktadır. Finansal başarısızlık hem işletmelerle yakından ilgili grupları hem de genel ekonomiyi doğrudan etkilemesi nedeniyle önemli bir kavram olarak karşımıza çıkmaktadır. Dolayısıyla söz konusu durumun nedenlerinin doğru bir şekilde tespit edilmesi ve gereken önlemlerin zamanında alınması işletme yönetimi, ülke ekonomisi, yatırımcılar, kredi kuruluşları ve toplum açısından oldukça yararlı olacaktır. Bu bağlamda çalışmanın amacı, finansal başarısızlık sinyallerinin önceden tespit edilmesini sağlayacak göstergeleri belirlemek ve bunu sağlayacak analizlerin hangisinin daha güçlü sınıflandırma yeteneğine sahip olduğunu ortaya koymaktır. Çalışma kapsamında 2013–2016 yılları arasında Borsa İstanbul (BİST)'da işlem gören, imalat sanayii sektöründe yer alan ve bilanço ile gelir tablolarına ulaşılan 156 işletme incelenmiştir. Veriler, Kamuyu Aydınlatma Platformu (KAP)'nun internet sitesinde yer alan tablolar yardımıyla ve Excel'de yapılan hesaplamalar sonucunda sağlanmıştır. Böylece, hesaplanan finansal oranlar bağımsız değişkenleri ve finansal başarı durumu bağımlı değişkeni oluşturmuştur. Yöntem olarak lojistik regresyon analizi, kısmi en küçük kareler diskriminant analizi, CHAID analizi ve yapay sinir ağları tercih edilmiştir. Lojistik regresyon ve CHAID analizleri SPSS 25.0, kısmi en küçük kareler diskriminant analizi MATLAB R2012b ve yapay sinir ağları SPSS Modeler 1.0 paket programları kullanılarak yapılmıştır. Bu dört yöntemle de elde edilen doğru sınıflandırma oranlarının oldukça yüksek çıktığı görülürken, en yüksek sınıflandırma gücüne sahip analizin üç dönem için de yapay sinir ağları olduğu tespit edilmiştir. Bunun yanı sıra finansal başarısızlığın belirlenmesinde en etkili oranlar net kâr marjı, esas faaliyet kâr marjı, kaldıraç oranı, duran varlıklar/devamlı sermaye oranı, stok bağımlılık oranı ve duran varlıklar/öz kaynaklar oranı olarak belirlenmiştir. Kârlılık oranlarının ve kaldıraç oranının işletmeler için önemli olduğu dikkate alındığında, önem derecesi yüksek değişkenler arasında yer almaları beklentiyi karşılamıştır. Anahtar Sözcükler: Finansal Başarısızlık, Lojistik Regresyon Analizi, Kısmi En Küçük Kareler Diskriminant Analizi, CHAID Analizi, Yapay Sinir Ağları.
Due to the rapid development of technology and the difficult conditions of competition, it seems that it is becoming increasingly difficult for businesses to continue their existence and their operations. Crises are being experienced because of the difficult market conditions and the effects of economic fluctuations. As a result, businesses face financial failures. However, businesses are not suddenly exposed to financial failures, but before that they give some signals about this situation. This situation demonstrates the importance of predicting financial failures in advance and identifying high-severity indicators that provide this determination. Financial failure emerges as an important concept because it directly affects both the groups closely related to the businesses and the overall economy. Therefore, determining the causes of the aforementioned situation correctly and taking the necessary precautions at the right time will be quite useful in terms of business management, the country's economy, investors, credit institutions and society. In this context, the aim of this study is to identify the indicators that will enable us to detect the signals of financial failure in advance and to exhibit which analyzes will provide the best predictability and classification capability. Within the scope of the study, 156 businesses operating in the manufacturing industry, which were operand in Istanbul Stock Exchange Market (BİST) between 2013 and 2016 and whose balance sheet and income statements are accessible, have been examined. The data have been obtained from the tables on the website of the Public Disclosure Platform (PDP) and from the calculations made in Excel. Thus, while the calculated financial ratios have constituted the independent variables, the financial success situation has constituted the dependent variable. Logistic regression analysis, partial least squares discriminant analysis, CHAID analysis and artificial neural networks have been preferred as methods. Logistic regression and CHAID analyzes were performed by using SPSS 25.0, while for partial least squares discriminant analysis and artificial neural networks, MATLAB R2012b and SPSS Modeler 1.0 packet programs are employed respectively. While the correct classification ratios obtained by these four methods were seen quite high in each analysis, it was determined that the analysis with the highest estimation power was artificial neural networks for three periods. Besides, the most effective ratios in forecasting financial failure were determined as net profit margin, operating profit margin, leverage, fixed assets / total capitalization, inventory dependency ratio and fixed assets / equity ratio. Given that profitability ratios and leverage are important for businesses, it has met the expectation that they are statistically significant variables. Key Words: Financial Failure, Logistic Regression Analysis, Partial Least Squares Discriminant Analysis, CHAID Analysis, Artificial Neural Networks.
Due to the rapid development of technology and the difficult conditions of competition, it seems that it is becoming increasingly difficult for businesses to continue their existence and their operations. Crises are being experienced because of the difficult market conditions and the effects of economic fluctuations. As a result, businesses face financial failures. However, businesses are not suddenly exposed to financial failures, but before that they give some signals about this situation. This situation demonstrates the importance of predicting financial failures in advance and identifying high-severity indicators that provide this determination. Financial failure emerges as an important concept because it directly affects both the groups closely related to the businesses and the overall economy. Therefore, determining the causes of the aforementioned situation correctly and taking the necessary precautions at the right time will be quite useful in terms of business management, the country's economy, investors, credit institutions and society. In this context, the aim of this study is to identify the indicators that will enable us to detect the signals of financial failure in advance and to exhibit which analyzes will provide the best predictability and classification capability. Within the scope of the study, 156 businesses operating in the manufacturing industry, which were operand in Istanbul Stock Exchange Market (BİST) between 2013 and 2016 and whose balance sheet and income statements are accessible, have been examined. The data have been obtained from the tables on the website of the Public Disclosure Platform (PDP) and from the calculations made in Excel. Thus, while the calculated financial ratios have constituted the independent variables, the financial success situation has constituted the dependent variable. Logistic regression analysis, partial least squares discriminant analysis, CHAID analysis and artificial neural networks have been preferred as methods. Logistic regression and CHAID analyzes were performed by using SPSS 25.0, while for partial least squares discriminant analysis and artificial neural networks, MATLAB R2012b and SPSS Modeler 1.0 packet programs are employed respectively. While the correct classification ratios obtained by these four methods were seen quite high in each analysis, it was determined that the analysis with the highest estimation power was artificial neural networks for three periods. Besides, the most effective ratios in forecasting financial failure were determined as net profit margin, operating profit margin, leverage, fixed assets / total capitalization, inventory dependency ratio and fixed assets / equity ratio. Given that profitability ratios and leverage are important for businesses, it has met the expectation that they are statistically significant variables. Key Words: Financial Failure, Logistic Regression Analysis, Partial Least Squares Discriminant Analysis, CHAID Analysis, Artificial Neural Networks.
Description
Tez (yüksek lisans) -- Ondokuz Mayıs Üniversitesi, 2018
Libra Kayıt No: 122764
Libra Kayıt No: 122764
Keywords
İstatistik, İşletme, Borsa İstanbul, CHAID Analizi, Diskriminant Analizi, En Küçük Kareler Yöntemi, Finansal Başarısızlık, Lojistik Regresyon Analizi, Performans, Performans Değerlendirme, Statistics, Yapay Sinir Ağları, Business Administration, İstatistiksel Yöntemler, İstanbul Stock Exchange, Chaid Analysis, Discriminant Analysis, Least Squares Method, Financial Failed, Logistic Regression Analysis, Performance, Performance Evaluation, Artificial Neural Networks, Statistical Methods
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
222
