A SPECTRUM ANALYSIS OF SEASONAL ADJUSTMENT,
Abstract
The methods of cross-spectrum analysis are applied to the problem of the seasonal adjustment of economic time series. In all, four methods are analyzed by means of empirical analysis of the results of applying the methods to artificially generated time series of known composition. Three of the four analyzed seasonal adjustment methods are known in the literature while the fourth method was developed on the basis of analysis of the other three. The three methods are: Hannan's method, Wald's method, and the Census Method. The fourth method is termed the rational-function method. It is found that there exist considerable differences in performance between the various methods when the seasonal varies with time. Hannan's method estimates (in an efficient, unbiased way) a constant seasonal, and thus is unsuited to the estimation of variation in the seasonal. The Wald and rational-function methods are able to estimate relatively more rapid changes in the seasonal amplitude than is the case with the Census Method. However, the Census Method and the Rational-function method both estimate changes in seasonal pattern which Wald's method takes as constant. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1964
- Accession Number
- AD0436769
Entities
People
- Herman F. Karreman
- Michael D. Godfrey
Organizations
- Princeton University