Diskriminant ve Lojistik Regresyon Yöntemleri Kullanlarak Finansal Baúarszlk Tahmini: BIST ømalat Sektörü Örneği

Financial Failure Prediction by using Discriminant and Logistics Regression Methods: Evidence From BIST Manufacturing Sector

Özet

Bu çalúmada 2006-2009 yllar arasnda BIST (Borsa østanbul) ømalat Sanayi Sektöründe faaliyet gösteren 126 iúletmenin finansal baúar/baúarszl÷n tahmin etmek üzere Çok De÷iúkenli Diskriminant Analizi ve Lojistik Regresyon Analizi kullanlarak 1,2 ve 3 yl öncesinden en yüksek tahmin gücüne sahip model belirlenmiútir. Çalúmada kullanlan ba÷msz de÷iúkenler içerisinde bilanço ve gelir tablosundan elde edilen nicel de÷iúkenlerin yan sra KAP’dan (Kamuyu Aydnlatma Platformu) elde edilen 4 ba÷msz nitel de÷iúken kullanlmútr. Analizler sonucunda diskriminant analizi toplam snflandrma do÷rulu÷u, finansal baúarszlktan 3, 2 ve 1 yl öncesinde srasyla %80.16, %83.33 ve %81.75’dir. Lojistik regresyon analizi toplam snflandrma do÷rulu÷u finansal baúarszlktan 3, 2 ve 1 yl öncesinde srasyla %80.16, %87.30 ve %92.86’dr. Lojistik regresyon modeli, finansal baúar/baúarszl÷ 3 yl öncesinde Diskriminant modeli ile ayn snflandrma oran ile tahminlerken 2 ve 1 yl öncesinde Diskriminant modelinden daha yüksek snflandrma performans elde etmiútir.

Abstract

In this study, to predict the financial success/failure of 126 businesses operating in BIST (Istanbul Stock Exchange) Manufacturing Industry Sector between 2006 and 2009, Multiple Discriminant Analysis and Logistic Regression Analysis were used to determine the model with the highest power prediction before 1,2 and 3 years. Among the independent variables used in the study, 4 independent qualitative variables obtained from PDP (Public Disclosure Platform) were used as well as quantitative variables obtained from balance sheet and income statement. As a result of the analyzes, the total classification accuracy of discriminant analysis was 80.16%, 83.33% and 81.75% at 3, 2 and 1 year before financial failure, respectively. The logistic regression analysis total classification accuracy was 80.16%, 87.30% and 92.86% at 3, 2 and 1 year before financial failure, respectively. The logistic regression model predicted the financial success / failure with the same classification rate as the Disriminant model 3 years ago, and achieved a higher classification performance than the Disriminant model 2 and 1 year ago.