APPLICATION OF MULTIVARIATE ANALYSIS ON THE SINGAPORE'S STOCK DATA

Abstract

This study looked at the use of multivariate analysis on the Singapore's stock market. The techniques used are: 1) Hotelling T2 test 2) Discriminant analysis 3) Principal components analysis and 4) Factor analysis. It is hoped that this cross-sectional study will provide a complementary tool to the already numerous amount of time series analyses done on the stock market. The focus was mainly on the comparison between trustee and non-trustee stocks. A total of eleven variables from 94 listed companies incorporated in Singapore were used. We found that there was a significant difference in the vector means of the variables between trustee and non-trustee stocks. Statistical discrimination between these two kinds of stocks was good: 74.60% and 83.87% of trustee and non-trustee stocks respectively were classified correctly. The analysis of the structure of the stock data as a whole gave meaningful and useful interpretations to the components retained. In the case of analysing only trustee stocks, we also obtained good results. However, the same cannot be said for non-trustee stocks. All in all, the application of multivariate techniques on the Singapore's stock data had provided much insights to the understanding of the stock pricing mechanism of the Singapore's stock market.