TABLE B.-Persons employed in nonagricultural industries, by hours worked during the survey week: August and September 1958 and September 1957 [Thousands of persons 14 years of age and over]

2

UNEMPLOYMENT AND HOURS OF WORK AMONG HARD GOODS FACTORY WORKERS: SEPTEMBER 1956 TO 1958

(Not adjusted for seasonality. Asterisk (*) denotes enlarged scale to depiot changes more clearly)

PERSONS WORKING 41 HOURS OR MORE
(AS PERCENT OF THOSE AT WORK)

Sept. 1958

June

1958

March 1958

Dec. 1957

March 1958

Sept. 1957

Sept. 1957

Sept. 1957

March 1957

June

June

1957

June 1957

Dec. 1956

Sept. 1956

Sept. 1956

10

10

1 The hard goods labor force comprises the employed plus those unemployed persons whose last job was in hard goods manufacturing.

The number of usual full-time workers whose hours had been cut below 35 because of slack work or other economic reasons was 1.3 million in September, not materially changed over the month but down substantially from its peak for the 1957-58 downturn (2.3 million in March 1958). This group, however, was still about a quarter of a million larger than in September a year ago.

Despite the overall stability in average hours over the month, the proportion on overtime (41 hours or more) in the durable goods manufacturing sector continued to edge upward, reaching 20.4 percent. At the same time, the proportion on shortened workweeks because of slack work and similar factors dipped to 3.8 percent. Early in the year, when the nonfarm workweek was at a low point, there were 15 percent on overtime and 9 percent on part time for economic reasons. The chart on page 4 illustrates the tendency for employers in "hard goods" industries to adjust output by varying the workweek before they make changes in the size of the work force. As indicated, the proportion of wage and salary workers on overtime was being trimmed in late 1956 and early 1957, and the proportion on part time for economic reasons was moving upward, before the percentage of unemployed in the industry group showed any significant change. Similarly, an expansion in hours was already evident in the second quarter of 1958, whereas the rate of unemployment remained at the recession peak until September.

UNEMPLOYMENT

The jobless total fell by 600,000 to 4.1 million in September, matching the large seasonal drop of the previous month. Unlike the July-August changes, however, which were confined entirely to students and recent graduates, the September reduction resulted in large part from the hiring of previously unemployed adult workers. On a seasonally adjusted basis, the September figures reflected the first significant drop in unemployment among adults since the recession began last year.

The number of adult men (25 years and over) out of work fell by 300,000 to 1.8 million, moving under the 2-million mark for the first time this year. Adult men have been more severely affected by unemployment than have other persons in the labor force. In September 1958, they constituted 45 percent of the jobless total in contrast to about 40 percent in September of other recent years.

As noted earlier, the over-the-month decline in unemployment was fairly widespread throughout the durable goods manufacturing sector. In the auto industry, the ratio of unemployed to the total work force dropped from 3 out of 10 to 2 out of 10 workers, its lowest point in 5 months. The unemployment rate in the primary metals group, which includes the steel industry, dropped from 12.4 percent to 10.3 percent. Fabricated metals and the machinery industries also registered declines of about 2 to 3 percentage points, dipping to 8 percent in September.

Elsewhere in nonfarm industries, the rate of unemployment either fell slightly or remained unchanged over the month. In spite of recent improvements, unemployment rates remained significantly above those recorded a year earlier in just about every major industry group.

The number of relatively long-term unemployed (those out of work 15 weeks or longer) was down by 200,000 over the month to 1,450,000, whereas it is usually fairly stable at this time of year. About half of the decrease in long-term unemployment was among durable goods workers. On a seasonally adjusted basis, this was the first significant drop in this category this year; the number of longterm unemployed had continued to increase throughout the spring and summer even after total unemployment began levelling off. This lag is a fairly typical feature in the later stages of a recession, as the group includes some less skilled workers and others with relatively low seniority in their companies. The long-term group in September was still 1 million above its level a year earlier, and constituted 35 percent of total unemployment in contrast to about 20 percent during most postwar years.

New unemployment (the number seeking jobs for 4 weeks or less) was also down over the month both an a seasonally adjusted and an unadjusted basis. The current level-1.6 million-was only slightly above the volume recorded in September a year ago.

TABLE C.-Experienced unemployed persons, by major industry group: August and September 1958 and September 1957

1 Percent of work force in industry who were unemployed. Excludes unemployed persons who never held a full-time civilian job for at least 2 consecutive weeks. Includes forestry and fisheries not shown separately.

SEASONAL ADJUSTMENT OF LABOR FORCE STATISTICS

Employment and unemployment-like most other measures of economic activity-are subject to appreciable seasonal variations, that is, fluctuations which recur fairly regularly at certain times of the year. In recent years, procedures for calculating seasonal adjustment factors and seasonally adjusted data on the Bureau's electronic computers have been successfully applied to the employment and unemployment statistics published in this report.

The seasonal adjustment method programed for the electronic computers is an elaboration of the standard ratio-to-moving-average method, with a provision for "moving" adjustment factors to take account of changing seasonal patterns. Even since World War II, seasonal patterns in several major employment series have changed appreciably. A number of refinements have already been added to the original program prepared in 1954, and the method will be further revised if warranted on the basis of future study and experience. Tests show that in its present form it yields seasonal adjustment factors of a high quality for most time series. A detailed description and illustration. of the method appears in appendixes II and III of Current Population Reports. Series P-50, No. 82, "Seasonal Variations in the Labor Force, Employment, and Unemployment."

Seasonal adjustment factors for the major components of the labor force to be applied specifically to the data for 1957 and 1958 are shown in table D. These factors are consistent with the definitions of employment and unemployment adopted in January 1957. In computing the factors, the pre-1957 data were adjusted insofar as possible to reflect the changes in definitions. In accordance with the Bureau's policy, the factors are recomputed at the conclusion of each calendar year, but those in current use are ordinarily replaced only if a significant change in seasonal patterns is indicated.

Different factors would have been used in adjusting the published series based on the definitions in effect prior to 1957. For major components of the labor force, these factors are provided in table D of series P-50, No. 82.

The factors presented in this report represent approximations of the seasonal component in each of the series shown, and describe the direction and relative magnitude of the seasonal changes during the year. Differences from 100 percent in the factors for a given month represent normal deviations from