A new method to estimate the noise in financial correlation matrices

Abstract

Financial correlation matrices measure the unsystematic correlations between\nstocks. Such information is important for risk management. The correlation\nmatrices are known to be ``noise dressed''. We develop a new and alternative\nmethod to estimate this noise. To this end, we simulate certain time series and\nrandom matrices which can model financial correlations. With our approach,\ndifferent correlation structures buried under this noise can be detected.\nMoreover, we introduce a measure for the relation between noise and\ncorrelations. Our method is based on a power mapping which efficiently\nsuppresses the noise. Neither further data processing nor additional input is\nneeded.\n