| Immense risks and crises exist in the intense market economy as well as the opportunities for development. Especially, it is becoming more and more likely for listed companies to be categorized as "special treatment" or even delisted because of the financial crisis from 2007.. Financial crisis not only threatens companies'own survival and development, but also results in huge damage to their investors and creditors. Therefore, it means a lot to stabilize the security market, the economy and the social development if financial ratio system with abundant mutual information and financial crisis early warning model with high predictability were established.There is abundant research in the field of financial crisis early warning, which obtained good results from diverse perspectives and using various statistical methods. However, there are two obvious problems in the application of such methods:firstly, for many statistical models, financial ratios have to be "simplified"; namely, if some indexes have correlations with each other the most representative ones need to be selected first. Besides, a number of statistical models require same index distributions for normal companies and those with financial crisis. The impossibility of using all financial data during data screening leads to information loss and even bad results of the prediction. Secondly, the aplicable range for each kind of methods is not the same. The hypotheses about data distribution between healthy companies and failed companies for many researches could be unacceptable due to common sense, because they suppose the distribution between some data requirements are obviously unrealistic as researchers have found that normal companies and those with financial crisis early warning bear different structures and distributions of financial data. Therefore, this paper aims to resolve such problems and to offer an intuitive and more visible financial crisis early warning model.The paper is structured as follows. First, methods and results of major research on financial crisis early warning models, as well as the definition of financial crisis and the purposes of financial crisis early warning, are introduced. Second, we present multidimensional scaling, which included the classical MDS, not based-on-distance MDS and the MDS considering the individual differences. Based on rich previous research, financial ratios that had been widely used in researches are selected. Both theoretical and empirical supports have been found for this selection. Then this paper explains how to apply MDS to financial crisis early warning, examines meanings of each dimension, and find some financial ratios which can be used in differing the heathly companies and the failed companies, from the confirmation of dimensions, the buildup and use of space maps, and the special ProFit analysis of MDS. Moreover, it can solve the problem that different model could just be applied in different fields, which means there is no attitude in common. Finally, empirical study is conducted to examine the predictability of the MDS model. MDS is in its early stage of application in financial crisis early warning, although it is not a novel methodology in statistics. This being said, MDS still well maintains financial data information and shows superior visibility, which proves that MDS has broad application and great opportunity of improvement in the field of financial crisis early warning. |
PDF Full Text Request