Analysis of Financial Correlation Matrix Using Random Matrix Theory

We analyzed the distribution of the eigenvalues of the empirical cross correlation matrix constructed from the data of Chinese stock market and the distribution of the components of some eigenvectors corresponding to certain eigenvalues. Compared with the analytical random matrix theory results, we found that a small part about 11%of eigenvalues are out of the range of random matrix theory results and that the distribution of the components of the eigenvector corresponding to the largest eigenvalue is evidently different from that of random matrix theory. These results demonstrate that chinese stock market contains 11% non efficient part