Academic journal article Business Economics

Predicting Turning Points in Economic Activity with Indexes of Economic Indicators: Improved Reliability Using a Logistic Model

Academic journal article Business Economics

Predicting Turning Points in Economic Activity with Indexes of Economic Indicators: Improved Reliability Using a Logistic Model

Article excerpt

THOSE RESPONSIBLE for business cycle dating have always had the thankless task of evaluating economic data and declaring the direction of economic activity. The role of the Business Cycle Dating Committee of the National Bureau of Economic Research (NBER) is similar to that of an economic commentator. In fulfilling that role, the Dating Committee is merely expressing expert opinion on momentum shifts in economic activity.

The political ramifications of this role are better understood in light of the recent experience in calling an end to the 1990-91 recession. The Dating Committee waited approximately eighteen months before they were willing to specify March 1991 as the trough in the downturn that began in August 1990. The timing of the decision was questioned in Republican circles, since it was made after the November election. Of course, an earlier decision would have been interpreted by the Democrats as an attempt to help reelect the incumbent. Given the long delay, information on the dating of the recession's trough was of little use to policy makers and business planners. While it is doubtful that the delay had any significant effect on business activity, an earlier call might have saved President Bush from the sharp criticism of the media when he claimed in early 1992 that the recession was over.

Looking back on the events, this eight-month contraction places it in a second place tie for the shortest recession in the postwar period. Only the six-month decline in the first half of 1980 was shorter. Indeed, one might reasonably argue that this half-year downturn was merely the initial decline in a protracted recession that lasted through 1982.

The task of dating the peaks and troughs is not a simple one. In order to avoid the embarrassment of making a premature judgment, members of the Dating Committee have chosen to wait until they are certain that all relevant data are signaling a turning point. While not the official definition, the frequently quoted standard for recession is two consecutive calendar quarters of decreasing real gross domestic product (GDP)(1). In practice, the committee considers a great deal more economic data before officially declaring troughs in business activity (see Hall, 1991/92). As a result, the economy will have been in a state of decline for several months before the committee officially declares a recession.

The prediction of cyclical turning points is a TABULAR DATA OMITTED common goal of business and government policy makers. The purpose of this paper is to demonstrate the usefulness of the three composite indexes in predicting turning points in economic activity. While this information has been used by other researchers, our approach to the problem exploits the statistical advantages of a logistic regression model.


Since their introduction in 1938, researchers have used the composite indexes of economic indicators, particularly the index of leading indicators, to provide advance information on changes in economic activity. This practice is based on the expectation that macro aggregates will change systematically as the economy enters a different phase in the business cycle. Because the components of the composite indexes have been chosen to represent and monitor business activity, it is reasonable to assume that the indexes themselves would serve as reliable predictors of economic activity.

While there is no well-prescribed economic theory linking the composite indexes to the business cycle, there is an economic rationale behind the use of the indexes in forecasting turning points in business activity. Ratti (1985) provides a detailed description of the composite indexes and an analysis of their usefulness in forecasting. The components of the index of leading indicators are included because they represent commitments to future economic activity; (housing starts, new orders for consumer goods, net business formation). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.