Printer Friendly

A new leading indicator: workers recently laid off.

A new leading indicator: workers recently laid off

Layoff rates have long been used as leading indicators in business cycle analysis. The layoff rate in manufacturing was initially selected as a leading indicator of business cycles in 1960.1 In 1961, it became 1 of 12 leading indicators published by the Bureau of Economic Analysis in Business Conditions Digest. The rate was derived from the labor turnover survey of manufacturing establishments, conducted by the Bureau of Labor Statistics and discontinued at the end of 1981. BEA subsequently replaced layoffs with initial claims for unemployment insurance.

Prior to 1967, the Current Population Survey collected data on laid-off workers on a very limited basis and only indirectly. This group included only those persons who were neither working nor looking for work, but responded that they had a job from which they were temporarily laid off and expected recall within 30 days. Since 1967, nonworking survey respondents have been asked directly whether they were on layoff. These workers are counted as unemployed regardless of their job search activity and form a subgroup of the job losers category. Further, workers on layoff are classified by the number of weeks since they were laid off.

At the Center for International Business Cycle Research, we used the number of all job losers on layoff, together with temporary layoffs prior to 1967, as a component of our leading employment index.

Recently, we observed that a better leading indicator could be obtained from data on recent "job losers on layoff,' rather than all workers on layoff. The recent jobless consist of those who were laid off within the last 5 weeks and are still unemployed at the time of the household survey. This group would seem to correspond closely to those included in the reports by employers on the number of workers laid off during the past month. The category can be converted to a layoff rate by dividing by total civilian employment. The result is a new leading indicator available currently.

The new indicator's lead-lag record during the business cycle from 1969 to 1982 is shown in table 1, together with the records of the related series. The layoff series, which we have seasonally adjusted, and the unemployment claims measure performed identically at all four troughs. The new series led at all four peaks, while initial claims led at three of the four.

Compared with the total layoff rate, the new indicator has longer leads at several turns, as would be expected because the new series reflects recent actions, whereas total layoffs include many who were laid off months earlier. Relative to the manufacturing layoff rate during the overlapping period of 1969-81, the new indicator shows longer leads at three of the four peaks, and about the same timing at the troughs. The manufacturing layoff rate, in turn, has a somewhat better leading record than temporary layoffs before 1969. Therefore, it seems reasonable to join the manufacturing layoff rate before 1969 to the new rate after 1969 to form a longer series with one break in coverage. It leads at 14 of the 16 business cycle turns from 1948 to 1982, with coincident timing in the other two turns and an overall average lead time of 6 months. The combined series is a more consistent leader than initial claims for unemployment insurance, which has coincident timing four times and lags twice during the same period.

So far as other important indicator characteristics are concerned, such as prompt availability and freedom from extra cycles and erratic movements, the new layoff rate stands up reasonably well to its competitors. Because it is a product of the household employment survey, the figures for a previous month are normally available on the first Friday of the following month. These figures are subject to revision annually, when seasonal factors are changed. Initial claims are available weekly, with a 2-week delay, which puts them on a par with the new layoff rate, although the monthly average is not available until the middle of the following month. Erratic movements in the new layoff rate are relatively large, however, as the following measures show:2

In all three series, it takes a span of 3 months for the average cyclical change to exceed the average irregular change.

The layoff rate and initial claims series tend to lead at business cycle peaks by much longer intervals than at troughs. In this respect, they are similar to the total unemployment rate, which leads at peaks but usually lags at troughs. The primary reason for this asymmetry is that business cycle dates are based upon data that reflect the long-run growth of the economy, whereas layoff and unemployment rates are relatively "trendless.' A trendless series tends to reach earlier peaks and later troughs than a series with a rising trend. When the turns in the layoff rates are matched with those in the total unemployment rate, rather than the business cycle, the leads are more nearly symmetrical. (See table 2.) The new layoff rate series leads the downturns in unemployment by an average of 5 months and the upturns by 7 months, for an overall average lead of 6 months.

Compared with employment, the new layoff rate again leads at both peaks and troughs, but by much longer intervals at peaks. This is to be expected, because nonfarm employment is virtually coincident with the business cycle, and reflects the growth trend of the economy.

In view of the record of the new layoff rate as a leading indicator, the Center for International Business Cycle Research has revised its leading employment index to include the new layoff rate since 1969 and the manufacturing layoff rate prior to 1969. At some future date, the new indicator might be considered a candidate for the Bureau of Economic Analysis' composite leading index, replacing initial claims for unemployment insurance.


ACKNOWLEDGMENT: We are indebted to Chantal Dubrin and Marcus Yumane for the statistical work on this report, and to John Stinson of the Breau of Labor Statistics for providing the new data on layoffs.

1 Geoffrey H. Moore, Business Cycle Indicators (Princeton University Press for the National Bureau of Economic Research, Princeton, NJ, 1961), p. 64.

2 For an explanation of the ratio of irregular to cyclical change and months for cyclical dominance, see the Handbook of Cyclical Indicators, Bureau of Economic Analysis, 1984, pp. 167-68.


Table: 1. Leads and lags of layoff rates at business cycle turns, 1948-82

Table: 2. Leads and lags of layoff rates at turns in employment and unemployment, 1948-82
COPYRIGHT 1986 U.S. Bureau of Labor Statistics
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1986 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Moore, Geoffrey H.; Cullity, John P.
Publication:Monthly Labor Review
Date:May 1, 1986
Previous Article:Short-time compensation: the AFL-CIO perspective.
Next Article:Union response to changes in printing technology: another view.

Related Articles
Store cuts: Circuit City lays off 11 in N.H.
Economic indicator remains flat in February.
Factory lays off agency workers.
Architect ranks shaken to the foundation.
Washington Post, New York Times seek new cost cuts.
With jobless fund under stress, state hikes tax: growing unemployment payouts leave state looking for options.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters