Valuation of Six Asian Stock Markets: Financial System Identification in Noisy Environments

The open financial economic systems of six Asian countries Taiwan, Malaysia, Singapore, Philippines, Indonesia and Japan - over the period 1986 through 1995 are identified from empirical data to determine how their stock markets, economies and financial markets are interrelated. The objective is to find rational stock market valuations using a country's nominal GDP and a short term interest rate, based on a modified version of the Dividend Discount Model. But our empirical results contradict such conventional financial economic theory. Various methods are used to analyze the 3D data covariance ellipsoids: spectral analysis, analysis of information matrices, 2D and 3D noise/signal determination and ''super-filter'' system identification based on 3D projections. The new ''super-filter'' method provides the sharpest identification of the Grassmanian invariant q of the empirical systems and the best computation of the finite boundaries of the empirical parameter ranges. All six Asian systems are high noise environments, in which it is very difficult to separate systematic signals from noise. Because of these high noise levels, spectral analysis is not reliable. By plotting all 3D q = 2 {Complete} Least Squares projections we find that only Taiwan has a clear q = 2 system, i.e., Taiwan's stock market, economy and financial market are rationally coherent. In contrast, Malaysia, Singapore, Philippines and Indonesia have q = 1 systems, in which stock markets and economies are closely related, but unrelated to the respective domestic financial markets. Several possible economic explanations are provided. We also quantitatively establish the incoherence of Japan's financial economic system. Japan's stock market operates independently from its economy and from its financial market, which are mutually unrelated.