Alternative Approaches to Forecasting: The Recent Work of the National Bureau
Tig HE appearance of Business Cycle Indicators pTovides a good opportunity to review some of the National Bureau's recent work on business cycles and forecasting.' It is more than a decade since the Bureau published Moore's Statistical Indicators of Cylical Revivals and Recessions, a paper which attained a degree of popularity unusual for National Bureau publications. 2 Since then, Moore's list of indicators has been kept up to date and widely used. The papers brought together in the present volume represent a comprehensive report on the recent work of Moore and his colleagues on cyclical indicators.3 Business Cycle Indicators consists of two volumes, the second of which is an appendix which describes and gives the monthly or quarterly data for a long list of series. Volume I is in three parts. The first and most important includes excerpts from annual reports of the National Bureau by Fabricant and Burns, Moore's earlier Occasional Papers on statistical indicators and several other papers by Moore, the original paper by Mitchell and Burns on indicators published in I938, a paper by Frank Morris on the predictive value of the leading series, and one by W. A. Beckett reporting on a set of indicators for Canada. Moore's revised list of 26 series is presented in Chapter 3. Part II contains six papers, some previously unpublished, on individual leading indicators -profits, business failures, new incorporations and new business firms, manufacturers' new orders, hours worked, and other labor-market series. Of these, the paper by Zarnowitz on new orders will probably be of greatest interest. None of these chapters is likely to excite the reader; they are all essentially descriptive; and some are outright stodgy in their style and treatment of their subject matter. Part III is concerned with methods of using the individual indicators and diffusion indices on a current basis. Two papers by Shiskin report on the work he has done with the indicators at the Census Bureau. Two others are by Moore. One describes an amplitude adjustment for the leading indicators, and the other deals with the average duration of run as a way of summarizing the current behavior of a group of series. I shall make no attempt to review each of these papers. Instead, in the rest of this article, I shall (i) compare the National Bureau's approach to business-cycle research with that which rests on the use of aggregative models and (2) attempt a general evaluation of the usefulness of economic indicators in short-run forecasting.
- DOI
- 10.2307/1926399
- Volume
- 44 (3)
- Pages
- 284
- Export
- BibTeX
- Sources
- openalex crossref