A SPECTRAL STUDY OF 'OVERADJUSTMENT' FOR SEASONALITY.

Abstract

The usual model for an economic time series contains an irregular factor assumed (except for strikes, etc.) to be random. A random series, of course, will have components of roughly equal magnitude at all frequencies; thus in particular a seasonal pattern and trend will by chance occur in such a series. Hence decomposition techniques should allow a certain amount of seasonality and trend to remain in the estimated irregular factor. This fact has not been taken into account by methods used to date, so it is not surprising that noticeable troughs appear in the spectrum of estimated irregular factors. This phenomenon is called 'overadjustment'. The present paper proposes a method for recovering the overadjustment and gives some preliminary empirical results. The procedure of overadjustment and recovery is regarded as a logical one in the absence of a method of decomposition using the spectrum directly. (Author)

Document Details

Document Type
Technical Report
Publication Date
Nov 11, 1964
Accession Number
AD0608567

Entities

People

  • Nigel F. Nettheim

Organizations

  • Stanford University

Tags

DTIC Thesaurus Topics

  • Air Force
  • Chemical Reactions
  • Decomposition
  • Diffraction
  • Dissociation
  • Frequency
  • Recovery
  • Spectra

Fields of Study

  • Mathematics

Readers

  • Logistics and Supply Chain Management.
  • Mathematics or Statistics
  • Statistical inference.