Printer Friendly

Productivity trends in selected industries.

BASED ON LABOR productivity measures compiled by the Bureau of Labor Statistics, this article examines differences in productivity trends over the 1973-88 period for 149 separate mining, manufacturing and service industries. (1) Although this period spans the well-documented productvity slowdown of the 1970s and the subsequent rebound of the 1980s, (2) it appears that productivity trends for separate industries are relatively stable over time with respect to each other. For example, industries with above-average productivity gains during the 1970s were likely to experience above-average gains in the 1980s and industries with below-average changes continued in that vein during both periods. Also, although most industries experienced higher productivity growth rates in the 1980s than in the 1970s, the largest increases occurred in those industries with above-average gains during the 1970s.

Although the reasons for these patterns are not examined in this article, it seems that industries with relatively high productivity gains during the 1970s and 1980s were apt to have experienced declines in employee hours in conjunction with increases in output. Industries experiencing relatively low productivity gains or even productivity declines, on the other hand, were more likely to have experienced changes in employee-hours that matched changes in industry output.


Industry productivity measures form one part of a broad Bureau of Labor Statistics program of productivity measurement. (3) These annual measures are computed by dividing an index of output by an index of aggregated employee hours. (4) For trade and service industries, aggregate hours include the hours of all persons involved in producing the output, including the self-employed and unpaid family workers.

Industry output indexes are based on quantities of products or services provided by the industry. The units are either physical quantities, such as tons of coal, passenger miles, or kilowatt hours, or constant dollar value of production. Because output indexes should reflect differing trends in the output of various products or services produced by an industry, these output measures are constructed from as much detailed data as possible. For example, output of the electrical utility industry is based on the number of kilowatt hours sold to each of seven different types of customers while the output of hardware stores is based on twenty-three different merchandise sale lines.

Industry employment and employee-hour indexes are mainly developed from data compiled by the Bureau of Labor Statistics (BLS). For most industries, BLS publishs monthly and annual data on the total number of employees, the number of production workers, the number of nonsupervisory workers, and the number of women employees as well as average weekly hours data for production and nonsupervisory workers. For industries in which the self-employed are an important part of the work force, indexes are constructed for the hours of all persons, including paid employees, partners, proprietors, and unpaid family workers. Hours of paid employees, both supervisory and nonsupervisory, partners, etc., are treated as homogeneous and additive with no distinction made between hours of different groups of workers.

Published measures used in this study cover 5 mining industries; 111 manufacturing industries; 8 transportation, communications, and utility industries; 16 wholesale and retail trade industries; and 9 financial and service industries. As of 1988, published industries covered about 40 percent of all workers in mining industries; 42 percent of those in manufacturing; 35 percent in transportation; 66 percent in communications; 87 percent in electric, gas, and sanitary services; 63 percent in retail and wholesale trade; 21 percent in finance, insurance and real estate; and 12 percent in services.


During the 1967-73 period, 94 percent of the industries for which separate labor productivity indexes are available showed productivity increases. (5) This proportion fell to 77 percent during the 1973-79 period and then rebounded to 83 percent during the 1979-88 period. While manufacturing, mining and retail trade industries followed this overall pattern, service sector industries experienced a continuing falloff. Prior to 1973, 80 percent of service sector industries registered productivity gains compared to 60 percent in the 1973-79 period and 40 percent in the 1979-88 period.

Even though average annual productivity gains tended to be lower in the 1973-79 period than during the 1979-88 period, most industries maintained their relative positions during both periods (see Table 1 for quartile rankings of industry productivity


changes in both periods). Overall, just over two-thirds of the industries in the top two quartiles in the 1973-79 period (50 of 74) were also in the top two quartiles during the 1979-88 period; conversely about two-thirds of all industries registering productivity changes in the lower half of the range in the 1973-79 period (51 of 75) did so again in the 1979-88 period. (6) There was also a slight tendency for service sector industries to be concentrated in the lower two quartiles of the distribution. Of the 33 service sector industries studied, for example, 16 ranked in the lower half of the distribution in both the 1973-79 and 1979-88 periods.

Interquartile movement was also somewhat limited. As the following tabulation shows, the first quartile in 1979-88 contained 37 industries, of which 15 were in the first quartile during the 1973-79 period, 10 of which were in the second and 12 of which came from the third and fourth quartiles. On the other hand, 20 industries registering productivity changes in the bottom quartile in the 1973-79 period also registered changes in the same quartile during the 1979-88 period.

In percentage point terms, increases in average annual rates of productivity gains during the 1980s tended to be larger for industries ranked in the top two quartiles in both the 1973-79 and 1979-88 periods. Of the 50 industries in the top half of the array in both periods, 30 saw increases in their productivity growth rates, with 12 experiencing gains of 2 percentage points or more. During the 1979-88 period, for example, output per employee hour for the radio and television receiving sets industry increased at an average annual rate of 9.9 percent, compared to annual rate of 5.1 percent during the 1973-79 period.

Table 2

Number of industries ranked by quartiles

quartiles First Second Third Fourth
First 15 10 5 7
Second 13 12 7 5
Third 7 11 13 6
Fourth 2 4 12 20

Among the 51 industries in the lower two quartiles in both periods, 29 showed increases in their productivity growth rates but only 5 reported gains of at least 2 percentage points. In addition, average annual productivity rates either became negative or negative rates increased for 12 industries in the bottom half of the array. Between 1973 and 1979, for example, output per employee hour fell by an average of 0.2 percent a year in gas utilities (SIC 492,93); during the 1979-88 period, the decline averaged 3.4 percent yearly. For the ball and roller bearings industry (SIC 3562), the rates were 0.5 percent and -0.2 percent, respectively.


There does not appear to be a simple reason to explain why some industries consistently experience relatively high productivity gains and some do not. (7) Within each quartile, industries vary widely in terms of output and structure. Ranking in the first quartile during both the 1970s and 1980s, for example, were such diverse industries as fluid milk, synthetic fibers, primary copper, household cooking equipment, telephone communications, and gasoline service stations. It does appear, however, that industries consistently registering relatively high productivity gains generally were able to increase output while at the same time reducing employee hours.

Of the 50 industries ranking in the top two quartiles in both the 1973-79 and 1979-88 periods, just over one-half experienced output increases concomitant with declines in employee hours. Some of these industries, such as wet corn milling (SIC 2046), nonwool yarn mills (SIC 2281) and motor vehicles and equipment (SIC 371), increased output and decreased employee hours in both periods, while others, such as tires and inner tubes, did so in only one period. Output increased faster than employee hours for 16 industries during the 1973-79 period and for 9 industries during the 1979-88 period. Employee hours declined more than output in 8 industries in the 1970s verses 12 industries in the 1980s.

This pattern of output and employee-hour changes was not repeated by industries in the lower two quartiles. In general, changes in output for these industries were almost evenly matched by changes in employee hours. During the 1973-79 period, 34 of the 51 industries in the lower two quartiles in both the 1970s and 1980s experienced increases in both output and employee hours. Representative of this trend were the hand and edge tool industry (SIC 3423), which registered a 3.3 percent average annual gain in output over the 1973-79 period as well as a 2.9 percent increase in employee hours, and the petroleum pipeline industry (SIC 4612,13), which saw output increase by 4.0 percent per year and employee hours by 3.3 percent. These close matches between output changes and employee hours changes were also evident in the 1979-88 period, except that industries were more apt to experience declines in both output and employee hours. During the 1980s, 25 of the industries saw declining output and employee hours. Typical of these industries were the metal cutting machinery industry (SIC 3541) where output declined by 8.4 percent a year over the 1979-88 period and employee hours fell by 9.2 percent, and class I bus carriers (SIC 4111,31,414), which registered a 4.6 percent decline in output and employee hours. Most of the remaining industries experienced output increases again almos exactly matched by employee hour increases.

Within this overall pattern, some individual industries experienced significant changes in productivity that can be partially traced to specific causes. (8) Among mining industries, for example, iron mining (SIC 1011) went from no growth in the 1973-79 period to an average annual gain of 8.8 percent in the 1979-88 period, copper mining (SIC 1021) from 2.3 percent to 9.9 percent and coal mining (SIC 111,21) from a decrease of 3.9 percent to an increase of 8.2 percent. In these cases, productivity was positively affected by the closing of less-efficient facilities, increased demand, and improved mining and processing technologies, such as longwall coal mining and the introduction of leaching techniques in copper mining. (9)

Average annual productivity also increased significantly for the steel industry (SIC 331), from no change in the 1973-79 period to 5.1 percent per year in the 1979-88 period. Among the many factors leading to this increase were a massive restructuring and modernization of major steel mills, the closing of a significant number of older inefficient mills, growth in the number of small specialized highly efficient minimills, widespread diffusion of advanced technology such as continuous casting, and a large reduction in the number of employees. (10) In a similar vein, changes in technology and industry consolidation also spurred productivity gains in the railroad industry (SIC 401). (11) The increase in railroad productivity growth from 1.4 percent in 1973-79 to 8.9 percent in 1979-88 partially reflected the introduction of new technology, such as automatic classification yards and computer-aided centralized traffic control; growth in the use of unit trains and piggyback traffic; mergers and consolidations that led to the abandonment of some rail track and employment declines; and increased output.

In some industries, however, technological changes were more than offset by other factors. During the 1979-88 period, for example, computerized cash registers and optical scanners were widely diffused among grocery stores (SIC 5411). On the other hand, this period also saw the growth of labor intensive departments such as salad bars, delicatessens, and bakeries as well as longer store hours. (12) Overall, productivity declined for the industry by just under 1 percent a year during the 1979-88 period, compared to a 0.3 percent decline during the 1973-79 period.


(1) For more detailed information on these and other industries, including indexes of output, employee hours, and employment, see Productivity Measures for Selected Industries and Government Services, U.S. Department of Labor, Bureau of Labor Statistics, May 1991, Bulletin 2379.

(2) For a discussion of factors affecting overall productivity growth during the 1970s and 1980s, see Edwin Dean and Kent Kunze, "Recent changes in the growth of U.S. multifactor productivity," Monthly Labor Review, May 1988, pp. 12-22.

For a discussion on the dispersion of productivity trends of individual industries and the productivity slowdown of the 1970s, see Peter K. Clark and Jane T. Haltmaier, "The Labor Productivity Slowdown in the United States: Evidence From Physical Output Measures," The Review of Economics and Statistics, August 1985, pp. 504-08.

(3) Other major components are quarterly measures of output per hour for major sectors of the economy, annual measures of multifactor productivity for major sectors, annual measures of multifactor productivity for two-digit manufacturing industries, annual measures of multifactor productivity for selected three-digit industries, international comparisons of labor productivity in manufacturing, and labor productivity measures for selected federal government functions and state and local government activities.

Methods used for these measures differ. For example, industry labor productivity measures use gross output, base year weighted, while multifactor measures use value-added output, base year weighted, and two- and three-digit industry multifactor measures use Tornqvist indexes of gross output.

(4) More detailed information on methods, limitations, and data sources is contained in BLS Handbook of Methods, BLS Bulletin 2285.

For a more detailed description and discussion of service sector industry measures, see Edwin R. Dean and Kent Kunze, "Productivity Measurement in Service Industries," forthcoming National Bureau of Economic Research paper.

(5) Average annual changes in output per employee hour are compound rates of change.

(6) This pattern of relative stability appears to extend back to the 1960s. When comparisons are made for a set of eighty-eight industries for which productivity measures exist from 1960 onward, about three-fifths of the industries in the upper two quartiles during the 1960-73 period were also in the upper half of the array during the 1973-79 and 1979-88 periods (28 out of 44 and 26 out of 44, respectively).

(7) Although labor productivity measures relate output to employment and employee hours, they do not measure the specific contributions of labor, capital, or other factors of production. Rather, they reflect the joint effect of a number of interrelated influences, such as changes in technology, capital investment per worker, level of output, utilization of capacity, layout and flow of materials, managerial skill, and the skills and efforts of the work force.

For an extensive examination of how industries differ, such as in their production functions, and the effect of these differences on productivity growth rates, see Dale Jorgenson, Frank Gollop, and Barbara Fraumeni, Productivity and U.S. Economic Growth, Harvard University Press, Cambridge Massachusetts, 1987.

For a study on the effects of technical efficiency and import competition on long-term productivity growth rates for individual manufacturing industries, see Richard E. Caves and David R. Barton, Efficiency in U.S. Manufacturing Industries, MIT Press, Cambridge, Massachusetts, 1990.

(8) For a more detailed discussion, see Richard B. Carnes, "Productivity in industry and government, 1989," Monthly Labor Review, May 1991, pp. 23-33.

(9) Ibid.

(10) Ibid.

(11) Ibid.

(12) Ibid.
COPYRIGHT 1991 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ardolini, Charles; Sieling, Mark
Publication:Business Economics
Date:Oct 1, 1991
Previous Article:Worldwide convergence of productivity levels: recent empirical evidence.
Next Article:A guide to organizational productivity and quality improvement.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters