Applying Grey Relation Analysis to Establish the Financial Distress Prediction Model for Electronic Companies in Taiwan

The majority of past research has focused on the use of literature feedback or factor analysis as metrics for financial distress prediction. The theoretical basis for the former is relatively thin, while the latter is severely limited by data requirements. As such, this paper will instead use grey relation analysis to determine several indices with high levels of relation, and to select several representative indicators. This method will provide the indicators under a more sound theoretical basis, while also overcoming issues of time constraints on data collection or unknown data distribution types. Additionally, unlike previous financial distress prediction models which have frequently overlooked the differences between industries, this paper will use logistic regression analysis in the use of 26 electronics companies as research subjects, and after removing from the sample those with inadequate data, a total of nine companies will be analyzed according to applied finance and corporate governance indicators, building a financial distress prediction model for the electronics industry and then comparing the rate of error in both this and the traditional document-based model. Results show that 9 indicators are applicable to financial distress prediction in the electronics industry— 7 financial indicators and 2 corporate governance indicators. In terms of rate of error, over one year the two models’ determinations of financial distress are minimally different, but going back over two years, the GRA model is less likely to be incorrect. As the number of years involved increases, so does the difference in error rate.