Affordable Access

An Appraisal of Short-Term Economic Forecasts

  • Economics


SUMMARY MEASURES AND TYPES OF ERROR This PDF is a selection from an out-of-print volume from the National Bureau of Economic Research Volume Title: An Appraisal of Short-Term Economic Forecasts Volume Author/Editor: Victor Zarnowitz Volume Publisher: UMI Volume ISBN: 0-87014-445-6 Volume URL: Publication Date: 1967 Chapter Title: SUMMARY MEASURES AND TYPES OF ERROR Chapter Author: Victor Zarnowitz Chapter URL: Chapter pages in book: (p. 31 - 59) 4 SUMMARY MEASURES AND TYPES OF ERROR AVERAGE ERRORS OF ANNUAL FORECASTS OF GNP AND INDUSTRIAL PRODUCTION The over-all averages in Tables 1 and 2 indicate that the errors in the annual forecasts of gross national product averaged about $9 to $10 billion disregarding sign. They appear small—about 2 per cent— compared with the average level of GNP, but they are big enough to make the difference between a good and a bad business year. The aver- age year-to-year change in GNP over 1953—63 amounted to approxi- mately $22 billion. Thus the errors were, in terms of absolute averages, not quite one-half the size of the errors that would have been produced by assuming that next year's GNP would be the same as last year's. On the other hand, it is important to realize that aggregates as com- prehensive and complex as GNP are beset by conceptual and estimation difficulties that make their measurement quite difficult and imprecise. When the series to be predicted is subject to substantial errors of obser- vation, forecasting becomes particularly hazardous. A mean forecasting error of $10.0 billion does not appear very large relative to a mean of $5.5 billion in the revisions of the given series (see the absolute aver- ages in Table 1, column 13). In the index of industrial production, short-term fluctuations play a greater role and trend a lesser one than in GNP, which should make forecasting the index more difficult. On the other hand, revisions are mu

There are no comments yet on this publication. Be the first to share your thoughts.


Seen <100 times