Non-Stationary Covariance Matrices And Noise

Abstract

The exact meaning of the noise spectrum of eigenvalues of the covariance matrix is discussed. In order to better understand the possible phenomena behind the observed noise, the spectrum of eigenvalues of the covariance matrix is studied under a model where most of the true eigenvalues are zero and the parameters are non-stationary. The results are compared with real observation of Brazilian assets, suggesting that, although the non-stationarity seems to be an important aspect of the problem, partially explaining some of the eigenvalues as well as part of the kurtosis of the assets, it can not, by itself, provide all the corrections needed to make the proposed model fit the data perfectly.