Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

Composite Models Reduce Forecast Errors: The Case of New Automobile Sales 1979-1983

Howard Keen Jr.
Business Economics
Vol. 19, No. 4 (July 1984), pp. 47-50
Stable URL: http://www.jstor.org/stable/23483742
Page Count: 4
  • More info
  • Cite this Item
Composite Models Reduce Forecast Errors: The Case of New Automobile Sales 1979-1983
Preview not available

Abstract

Most business forecasters attempt to incorporate all relevant available information into their projections. One systematic way of doing this is by combining predictions from two or more sources into a formal composite model. This paper reports the results of combining projections of new car sales from Chase Econometrics with those of a univariate Box-Jenkins model. Although Box-Jenkins models tend to produce relatively less accurate predictions in highly volatile periods, such as that of 1979-1983, combining those predictions with the Chase forecasts nevertheless resulted in smaller errors than those of Chase alone.

Page Thumbnails

  • Thumbnail: Page 
47
    47
  • Thumbnail: Page 
48
    48
  • Thumbnail: Page 
49
    49
  • Thumbnail: Page 
50
    50