A Comparative Study of Distress Prediction Models for SMEs Start-ups in Renewable Energy

The purpose of this paper is to identify the suitable variables for the financial distress prediction models and compare the accuracy of MDA (multiple discriminant analysis), LA (logit analysis) and ANNs (artificial neural networks) for the early warning signal of SMEs (small and medium enterprises) start-ups in renewable energy industries. The research methods are MDA, LA and ANN which have been widely used in the world. Dataset was composed of 100 renewable energy SMEs in KOSDAQ and under the act of external audit of stock companies. The financial data of each company over the period from 2006 to 2013 were collected from KIS-Value. We found the result that 4 financial ratios were statistically significant and the accuracy of MDA is 92.9%, while that of LA and ANN are 94.3%, 87.5% respectively for the one year before the bankruptcy in 2013. The accuracy rate of bankruptcy prediction for the “one year before(T-1)” is better than for the “two year before (T-2)” to all 3 models. The importance of this study was to demonstrate empirically that financial distress prediction models are applicable to the SMEs start-ups in Korea renewable energy industry as an early signal of bankruptcy.