Analyzing the recent upward surge in overtime hours

HEADNOTE

During the economic expansion of the 1990s, employers in manufacturing industries were more likely than in previous recessions to increase overtime hours among existing employees than to hire new workers

From

March 1991, the end of the last recession, to early 1997, average weekly overtime in manufacturing increased by 1.6 hours, reaching its highest level-4.9 hours-since Bt.s began publishing the series in 1956.1 Overtime remained at or near this high level over the next year, retreating slightly by the end of 1998. These data are from the BLS Current Employment Statistics (cES) survey, a monthly survey of payroll, hours, and earnings collected from a sample of more than 400,000 of the Nation's employers. The CES program defines overtime as hours for which premiums were paid because they exceeded the number of straight-time workday or workweek hours. Average overtime is computed by dividing the total number of overtime hours in a given industry by the number of production workers in that industry, including those that work no overtime at all.

Historically, average overtime has increased with recoveries and fallen with recessions, with the level never exceeding 4.1 hours. Average overtime fell from 3.7 to 3.3 hours during the 1990-91 recession, but the current expansion has seen overtime reach an unprecendented level. This article analyzes the striking growth in overtime from March 1991 to January 1998 and its relationship to employment.

Overtime growth in the 1990s

Following the 1990-91 recession, cyclical job loss in manufacturing continued until mid-1993. Indeed, after losing 683,000 jobs during the downturn, another 400,000 manufacturing jobs were lost after the recession officially ended in March 1991. Interestingly, however, the cyclical trend in manufacturing overtime hours turned around exactly the same month that the recession ended. By the time that manufacturing employment started its cyclical recovery in July 1993, average overtime had increased from 3.3 to 4.1 hours. (See chart 1.) Overtime hours continued to surge, reaching 4.8 hours in the last quarter of 1994. Manufacturing employment expanded until April 1995, adding a total of 541,000 jobs in a period of less than 2 years.

Average factory overtime fell by 0.6 hour in 1995. In 1996, it started inching upwards again, while employment in the industry experienced a mild downward trend. By December of 1996, average weekly overtime had reached 4.6 hours, after starting the year at 4.2 hours. Employment also started back on a growth trend early in 1996, but at a very slow pace. By March 1997, overtime had reached a record high of 4.9 hours-a level it sustained for the next 2 months and then revisited at the end of the year. In contrast, employment, while still growing, ended 1997 at a level nearly 700,000 lower than its prerecession peak in March 1989.

Sources of overtime growth

Manufacturing's record-setting increase in average weekly overtime is the result of two factors. The first, as shown in table 1, is that, from March 1991 to January 1998, overtime increased in all but one of the component industries in manufacturing. The increases ranged from a notable 3-hour gain in transportation equipment to a relatively slight increase of 0.6 hour in apparel products. As the table illustrates, more than half of the 20 industries within manufacturing had increases of at least 1 hour over the 199198 period. In fact, many of these industries had set records for their overtime series by early 1997.2

IMAGE GRAPH 11

Chart 1.

Some specific industries made exceptional contributions to the growth in overtime hours. Within transportation, for example, overtime in the motor vehicle manufacturing industry jumped by a remarkable 4.4 hours. Similarly, within primary metals, overtime in iron and steel foundries grew by 3.7 hours, and within industrial machinery and equipment, refrigeration and service machinery overtime increased by 2.9 hours.3

The second factor driving the increase in overtime is that the distribution of employment in manufacturing was shifting towards component industries that were adding the most overtime over the 1991-98 period. This effect can be quantified by dividing the industries in the table into two groups. The 10 industries that had the greatest increase in overtime after the recession together averaged 5.2 overtime hours, which was 1.2 hours more than the average for the other 10 industries. At the same time, the 10 industries with the highest average overtime increase also had an accumulated increase in production workers of 11.2 percent, while the bottom 10 lost 4.7 percent of their production workers.

The combined effect of the growth in overtime in nearly every industry and the employment increases in industries with large gains in overtime can be seen in aggregate overtime (the product of production workers and average weekly overtime hours). The top ten overtime gainers accounted for 68 percent of the total manufacturing aggregate overtime in January 1998, compared with 60 percent in March 1991.

Overtime hours also would increase if employment in industries with high overtime levels grew faster than employment in industries with lower levels of overtime. To determine whether this was a factor, manufacturing overtime in 1998 was computed using the employment distribution of March 1991. The results showed that the shift in industry mix contributed little to the increase in overtime, adding just 0.1 hour.

Substituting overtime for employment

Historically, both employment and overtime have increased as the U.S. economy emerged from recessions, with overtime gains generally occurring prior to the employment gains. While this has remained the case since March 1991, employers appeared to rely more heavily on overtime in the current expansion than on hiring new employees. This part of the analysis focuses on overtime growth from the beginning of the current recovery until overtime hours peaked 82 months later, in January 1998, comparing it with employment growth over the same period. The data are compared with two other expansions that lasted at least 82 months. (See table 2.)

When the recoveries that began in March 1961 and December 1982 were 82 months old, they had added 3.5 million and 1.3 million manufacturing jobs, respectively. The peak levels of average overtime associated with those recoveries were 4.1 hours in February through April 1966 and 4.0 hours inFebruary and April 1989. The current recovery's overtime gain of 1.6 hours is slightly below the two previous recoveries; however, because the level at the onset of the current expansion was significantly higher, the peak levels of the two previous recoveries were superseded in less than 2 years. Even with this record-setting strength in overtime, employment grew by only 397,000, or just 17 percent of the average job growth in the two earlier recoveries.

The implied substitution of overtime for hiring can be quantified using full-time equivalents. Full-time equivalents are computed by taking the aggregate overtime and dividing it by 40, the number of hours in a standard workweek. For example, if 20 people worked 6 hours of overtime, the full-time equivalent of that overtime would be J-that is, 3 extra production workers could have been hired rather than existing workers accumulating 120 weekly overtime hours.

From March 1991 to January 1998, the number of production workers in manufacturing increased by 601,000. Over the same period, the full-time equivalent of the aggregate overtime change in manufacturing was 571,000 jobs. (See table 3.) That means that if employers had hired new workers instead of increasing overtime, nearly twice as many production workers would have been hired.

The table also shows where these workers would have been hired. Transportation equipment, which includes auto and aircraft assembly, had an overtime change valued at 107,000 fulltime equivalent jobs, or one-fifth of the total for all manufacturing. Industrial machinery and fabricated metals also would have accounted for a large portion of the hiring during this period. Other industries with relatively large full-time equivalents included rubber and plastics, electronics, and primary metals.

A common factor among the industries that added the most overtime was a highly skilled workforce. The data suggest that when the overall skill level among workers in an industry is relatively high, firms tend to increase overtime during expansions rather than hire new workers. Training highly skilled workers is costly, especially if many of them may be laid off during the next recession. For similar reasons, workers in some highly skilled occupations are in short supply and thus may not be available to the hiring establishment. The 10 industries within manufacturing with the largest overtime gains since the recession had more than 17 percent of their employment in highly skilled positions; the comparable figure for the 10 industries with the least gains is 8 percent.4

Employment and overtime in 1998

After starting 1998 at the record-setting level of 4.9 hours, by the end of the year, average weekly overtime in manufacturing had fallen by 0.4 hour. (See table 4.) Meanwhile, employment in manufacturing declined by 238,000 over the same period, as many export-sensitive industries reacted to the economic crises then occurring in Southeast Asia, Russia, and Brazil. The industry groups with the largest aggregate overtime declines in 1998 included specific (three-digit) industries that were the most export-sensitive of all manufacturing industries, including computers, aerospace, semiconductors, and household audio and video equipment.s The 0.4hour reduction in manufacturing overtime is equal to 157,000 full-time equivalents-that is, had overtime not been reduced by 0.4 hour, employers would have had to lay off an additional 157,000 factory workers in 1998.

IMAGE TABLE 23

Table 1.

IMAGE TABLE 24

Table 2.

UNLIKE IN PREVIOUS EXPANSIONS, manufacturing employers in the 1990s were more likely to increase overtime hours among existing employees than to hire new employees. Despite beginning the current expansion at historically high levels, overtime increased by nearly as many hours as in the earlier expansions of the 1960s and 1980s, bringing the level to a record high (4.9 hours) by the end of 1997. From its low of 3.3 hours in March 1991, overtime increased by 48 percent. The gains in overtime were spread throughout the manufacturing industry groups, with the largest gains occurring in durable goods, especially transportation equipment and primary metal industries.

Meanwhile, employment in manufacturing grew quite modestly during the 1990s expansion, increasing by about 4 percent from its trough in June 1993 to its peak in March 1998. Over comparable periods in the 1960s and 1980s, by contrast, employment increased by 15 percent and 5 percent, respectively. Largely as a result of economic crises abroad, employment began to decline in early 1998, with losses concentrated in export-sensitive industries. But just as overtime had substituted for job gains in the current expansion through 1997, it acted as a cushion against job loss in 1998. In fact, had overtime not been reduced by 0.4 hour in 1998, instead of a loss of nearly a quarter-million jobs in manufacturing, the loss would have been closer to 400,000. Manufacturing employment continued to decline in 1999, while overtime hours held steady, rising slightly by the end of the year.

IMAGE TABLE 31

Table 3.

IMAGE TABLE 32

Table 4.

FOOTNOTE

Notes

FOOTNOTE

1 The "official" starting and ending dates of recessions and expansions are determined by the National Bureau of Economic Research (NBER)-a private, nonprofit, nonpartisan research organization dedicated to promoting a greater understanding of how the economy works. NBER identifies economic turning points-that is, dates when economic activity turned in the opposite direction. For more information, see NBER's website, on the Internet at http://www.nber.org/, accessed February 2000.

2 Industries are defined by the Standard Industrial Classification (stc) system. The 20 component industries within manufacturing are those at the two-digit stc level of aggregation. For more information on the sic system, see Standard Industrial Classification Manual. 1987 (Washington, Dc, Office of Management and Budget, 1987).

3 Overtime data for specific (three-digit) industries within transportation equipment, primary metals, and industrial machinery are not seasonally adjusted. Therefore, to avoid seasonal fluctuations, overtime hours are measured from December 1990 to December 1997.

4 Data from the sc.s 1996 Occupational Employment Statistics program.. Highly skilled positions were defined as engineers, technicians, scientists, and precision workers and assemblers.

5 This analysis is based on the percent of employment tied to exports in 1993; data are from the ass Office of Employment Projections.

AUTHOR_AFFILIATION

Ron L. Hetrick is an economist in the Division of Monthly Industry Employment Statistics, Bureau of Labor Statistics.