GENERALIZED SPECTRUM AND ITS APPLICATIONS IN ECONOMICS AND FINANCE
Hong Yong · Journal of Guangxi Normal University · 2002
This paper reviews some recent development in time series spectral analysis,particularly a new generalized spectral method.The generalized spectrum,proposed in Hong (1999),is based on a data transformation via the characteristic function.It can capture both linear and nonlinear dependencies.The latter may be easily missed by conventional power spectrum or higher order spectra.A variety of applications of the generalized spectrum in economics and finance are discussed.Among them are tests of market efficiency,correct dynamic asset pricing,predictability of the direction of a price change,adequacy of value at risk models,and optimality of probability density forecasts.