Abstract

Considering the deficiencies in the field of researches on financial crisis prediction for the present, this paper build financial early warning model for manufacturing listed companies, using quarterly time serials data three years before special treatment (ST). We find out that the predictive validity of logistic financial early warning system based on time serials data is better than that based on cross-section data; logistic financial early warning is better than fisher multivariate discriminant analysis; corporate profitability, earnings per share (EPS) and general manager stake which significantly affect the financial distress.

Key words: Financial distress, Logistic model, Global principal component analysis.