Spectral Methods in Econometrics.
This study, describing the spectral methods of time series analysis and their use in econometrics, is intended to serve as an introduction for graduate students and econometricians who wish to familiarize themselves with the spectral, or frequency domain, approach. The theory of covariance stochastic processes is introduced, with special emphasis on its application in econometrics. Univariate, bivariate, and multivariate spectral analysis are discussed, the latter two in terms of distributed lag models. Estimation procedures and approximating sampling distributions applicable to the spectrum, the cross spectrum, and the bivariate frequency response function are also described. In the spectral approach to estimation and testing for a system of simultaneous distributed lag equations, extensive use is made of the properties of a Toeplitz matrix, which leads to a number of simplifications in the statistical inference. The final chapter of the study illustrates the application of spectral techniques to an econometric problem: Income and consumption time series are examined using their sample spectra, cross spectrum, and an adaptive distributed lag model. 291 pp. Bibliog.