Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
labor Supply and Demand Within the U.S. Economy As we noted in Chapter 1, the level of total employment within the U.S. economy is determined primarily by nontechnological factors. Techno- logical change can, however, affect the demand of individual sectors or industries for labor. Process innovations that increase labor productivity reduce the amount of labor (and potentially the amount of other inputs) required per unit of output. But this reduction need not and in fact has not translated into increased total unemployment in the United States. Instead, the employment impacts of any reductions in the amount of labor required for each unit of a product typically are offset by increases in demand for the product in response to lower prices or increases in the demand for other commodities. In addition, product innovations can create jobs in entirely new industries. To maximize the employment- expanding influence of technological change within the economy, workers must be able to move from sectors of declining labor demand to those in which employment opportunities are expanding. Indeed, the technologi- cal and other factors that have altered the structure of the U.S. economy and the demand for different types of labor during the past two decades appear to have increased the need for such movement. THE U.S. ECONOMY: CHANGES IN STRUCTURE AND PERFORMANCE SINCE THE 1960s Both the structure and performance of the U.S. economy have changed greatly since the 1960s. The share of private nonagricultural employment 51
52 TECHNOLOGY AND EMPLO YMENT accounted for by manufacturing, which stood at roughly 36 percent in 1966, has continued to decline, to roughly 24 percent in 1985 (U.S. Bureau of Labor Statistics, 1985a, 1986a). At the same time, international trade has expanded and now plays a much more significant role within the economy. Imports were 5.1 percent of GNP in 1966 and 11.4 percent in 1986; the share of GNP represented by exports expanded from 6 percent in 1966 to 8.9 percent in 1986 (President's Council of Economic Advisers, 1987, Tables B-1 and B-991.~ Sales in the industries associated with (as it was known in the 1960s) "automation" that is, instruments, office equipment, computers, and electronic components have more than doubled, growing from $53.4 billion (in 1986 dollars) in 1967 to more than $114 billion in 1985 (U.S. Bureau of the Census, 1982b, 1985a, 1985b).2 Although the pace of technological change may not have increased greatly since the 1960s, "automation" technologies now are applied more widely. Twenty years ago, automation was viewed as applicable mainly in the manufacturing sector; now, both the manufacturing and nonmanufacturing sectors make extensive use of information and computer-based technologies. As we noted in Chapter 2, changes in the international economic environment since the mid-1960s have narrowed the technological "gap" between the United States and other industrial economies. The dwindling technological lead of the United States, along with many other factors (e.g., the Vietnam war, oil price increases by the Organization of Petroleum Exporting Countries), has contributed to a deterioration in the performance of the U.S. economy. Unemployment levels throughout the 1970s and 1980s have remained well above those of the 1960s (Podgursky, 19841.3 The average annual rate of growth in real average hourly earnings during 1970-1986 was -0.4 percent, following average annual growth rates of 2.7 and 1.7 percent, respectively, during the 1950s and 1960s (President's Council of Economic Advisers, 1987, Table B-411. Finally, labor productivity growth in the U.S. nonfarm sector has declined since the 1960s, reaching an average annual rate of less than 1 percent during 1973-1986 (President's Council of Economic Advisers, 1987, Table B-441. Growth in imports and exports of goods relative to total U.S. production of goods during this period is even more dramatic. "Merchandise" imports and exports, which include agricultural products, automotive goods, petroleum products, and industrial supplies and materials, respectively, grew from 6.6 and 7.6 percent of U.S. goods production in 1966 to 20.2 and 12.8 percent in 1985 (President's Council of Economic Advisers, 1987, Tables B-6 and B-99). 2This calculation does not include telecommunications equipment sales, which have also grown rapidly since the mid-1960s. 3Podgursky noted that unemployment rates at the business cycle quarterly peaks in 1969, 1973, 1979, and 1981 were 3.6, 4.8, 6.O, and 7.4 percent, respectively.
LABOR SUPPL Y AND DEMAND 53 TABLE 3-1 Sectoral Composition (percent) of New Jobs in the U.S. Economy, 1955-1965, 1965-1975, and 1975-1985 Period Sector1955-1965 1965-1975 1975-1985 Manufacturing11.9 1.9 4.8 Mining and construction2.0 2.5 6.3 Transportation and utilities-1.0 3.1 3.4 Wholesale and retail trade21.8 26.5 29.0 Finance, insurance, and real estate6.9 7.4 8.7 Services27.7 30.2 39.1 Government31.7 28.4 8.2 NOTE: There were 10.1 million new jobs created during 1955-1965, 16.2 million during 1965-1975, and 20.7 million during 1975-1985. Percentages may not total 100 due to rounding. SOURCE: U.S. Department of Commerce, International Trade Administration (1987), Table 9, p. 12. Despite its disappointing performance in these areas, the U.S. economy expanded employment opportunities rapidly during 1975-1985. Indeed, more than 20 million jobs were created during the 1975-1985 decade, an expansion that allowed the labor force to absorb both the "baby boom" cohort and greater numbers of women seeking employment. The share of the 20.7 million new jobs created during 1975-1985 accounted for by the private nonmanufacturing sector was greater than in previous decades of the post-World War II period (Table 3-1~. Four industry groups (whole- sale and retail trade; transportation and utilities; finance, insurance, and real estate; and services) collectively accounted for approximately 80 percent of the new jobs created during 1975-1985, a substantial increase from these groups' shares of roughly 55 percent during 1955-1965 and 67 percent during 1965-1975. Although the manufacturing sector's share of the jobs created during 1975-1985 (4.8 percent) was well below its share during 1955-1965 (11.9 percent), the 1975-1985 share was higher than that of 1965-1975 (1.9 percent). Many more jobs were created in man~ufactur- ing during 1975-1985 than during 1965-1975. Several studies (e.g., Bluestone and Harrison, 1986) have noted that many of the jobs created during 1979-1984 paid relatively low wages. Bluestone and Harrison's empirical results, however, are sensitive to the temporal end point of their analysis. Extending their analysis beyond 1984 to cover 1985 substantially increases the share of high-wage jobs that were created during the entire period because 1981 and 1982, years that include a severe recession, receive less weight in the longer time series.
54 TECHNOLOG Y AND EMPLO YMENT Other evidence suggests that the jobs created during 1975-1985 were not uniformly "low-quality" jobs (i.e., low-wage jobs with minimal skill requirements and prospects for advancement). Leon's (1982) study of the occupational structure of the jobs created during 1972-1980 indicates that, during most of this period, the occupations exhibiting the highest rates of growth were professional, technical, and managerial positions. Moreover, according to Rosenthal (1985), the distribution of weekly earnings within the overall occupational structure in 1982 was slightly less skewed toward low-wage occupations than in earlier years. Nonetheless, McMahon and Tschetter (1986) argue that, within the high-wage occupations whose growth has been rapid, there is a tendency for the jobs created since 1973 to occupy a relatively low position in the intraoccupational earnings distribution. The facts of job expansion, earnings growth, and occupational shifts are less in dispute than the interpretation of the various trends. Some analysts (e.g., Bluestone and Harrison, 1986) view recent trends as evidence that equality of economic opportunity has declined because of technological or structural change for example, increased international competition or the growth of service sector employment. Others (e.g., Blackburn and Bloom, 1987; Rosters and Ross, 1987; Lawrence, 1984; Levy, 1987; McMahon and Tschetter, 1986; Rosenthal, 1985) see these trends as the result of demographic factors, combined with the severe recession of the early 1980s, slow growth in the overall economy, and low productivity growth. These researchers suggest that rapid growth in the U.S. labor supply coincided with slow (or negative, during 1981-1982) economic growth, expanding the supply of job seekers relative to the number of openings and placing downward pressure on wages. Although the various trends require continued monitoring and assessment, slow economic growth and low rates of productivity increase, rather than technological change in the U.S. economy, appear to be the key factors in explaining these shifts. The causes of the slow economic growth, high unemployment, and declining productivity growth rates that have afflicted the United States as well as other industrial nations since the 1960s are not well understood. Some portion of these developments can be ascribed to the disruptions in the global economy that occurred when oil prices rose in 1973 and 1979. But even when the erects of these disruptions are discounted, the record of U.S. demand management policies (i.e., fiscal and monetary policies) since the mid-1960s has been disappointing. Many economists and policy analysts now are more skeptical about the theory and practice of macroeconomic policy than they were in the 1960s. The ability of policymakers to reduce or eliminate unemployment through careful fiscal and monetary policies appears to be more limited than originally sus
LABOR SUPPL Y AND DEMAND 55 pected, due in part to the structural changes that have occurred in the U.S. and global economic systems. TRENDS IN U.s. UNEMPLOYMENT Aggregate Unemployment One of the most important factors contributing to public concern over the effects of technological change in the U.S. economy is the growth, noted earlier, in aggregate unemployment above the average levels of the 1950s and 1960s. There are no widely accepted explanations of the post-1973 growth in unemployment, although slow economic growth and frequent recessions (there have been four during this period) have contributed to it. Two of the recessions (1974-1975 and 1981-1982) were the most severe economic downturns the country has experienced since the Great Depression. In addition, the 1970s and 1980s saw no lengthy economic expansion comparable to that of the 1960s, although the current growth cycle may yet prove equally durable. Economic growth during 1966-1985 was much lower than that during 1948-1966 (a more detailed discussion of slower output growth follows). The frequently recessionary condition and slow growth of the U.S. economy since 1970, as well as the "oil shocks" of 1973 and 1979, all reduced the demand for labor, placing stronger upward pressure on aggregate unemployment rates than during the 1960s. Has the entry into the labor force of the huge baby boom cohort and a growing number of women, which increased the labor supply, contributed to higher aggregate unemployment during the 1970s and 1980s? According to Podgursky (1984), the baby boom cohort contributed significantly to aggregate unemployment in the early and mid-1970s, but its contribution has since declined. Instead, higher unemployment in the early 1980s resulted in large part from permanent job losses, which were concentrated among workers in manufacturing, mining, and construction. The higher unemployment rates of the 1980s appear to reflect longer spells of unemployment for a relatively small portion of the work force, rather than an increase in the share of the work force experiencing short spells of unemployment (Podgursky, 1984~.4 A portion of the higher post-1973 unemployment, especially the unem- ployment of the 1980s, appears therefore to be structural; that is, it reflects mismatches between worker skills or worker locations and job 4Average unemployment duration increased from 9.3 weeks in 1970 to 17.5 weeks in 1982 (Podgursky, 1984).
56 TECHNOLOGY AND EMPLOYMENT openings for a relatively small group of unemployed workers who experience lengthy spells of unemployment, rather than a cyclical down- turn in the overall economy that results in short spells of joblessness for a very large number of workers (Summers, 1986~. Such mismatches may come from technological change, but they also reflect rapid structural change of all types within this economy since 1970, including dramatic increases in international trade and recent significant import penetration in numerous manufacturing industries.5 Although aggregate unemployment has been high, relative to post-1945 averages, during the 1970s and 1980s, U.S. unemployment rates recently have fallen below those of many Western European nations, including West Germany, France, and Great Britain. Moreover, the average duration of U.S. unemployment now is shorter than that observed in a number of Western European nations. In 1985 slightly more than 15 percent of the unemployed population in the United States had been out of work for more than 6 months; in Great Britain, 60 percent of those unemployed had been out of work that long, whereas in West Germany this figure stood at 55 percent.6 The lower average rate and shorter duration of U.S. unemployment reflect the high rate of job creation and loss in the U.S. economy, as well as the relatively high geographic mobility of U.S. workers. Leonard (1986), in his study of Wisconsin, estimated that, each year, the jobs created and lost equaled nearly 14 percent and 11 percent, respectively, of the previous year's jobs.7 The U.S. labor market thus is extremely fluid and dynamic, attributes that should ease the adjustment of workers to new technology. How does high or low aggregate unemployment affect unemployment within different groups of U.S. workers? Blue-collar workers accounted for a disproportionate share of the growth in U.S. unemployment during 1970-1982. With 31 percent of the 1982 labor force, their share of the 5Costrell's (1987) measure of structural change, which is based on changes in the employment shares of the 12 economic sectors discussed in greater detail later in this chapter, suggests that such change has accelerated during 1979-1985, compared with 1973-1979 and 1966-1973. Rissman (1986) obtained similar results. 6See "The Supple Rigidity of America's Job Machine," The Economist 302 (February 7, 1987):28-29. The entry into the labor force of the baby boom cohort is partly responsible for recent high European unemployment rates. The European baby boom postdates that of the United States by 5-10 years; Western European economies now are absorbing large increases in their labor forces that exert upward pressure on unemployment rates (Norwood, 1983). 'Leonard's study examined the 1977-1982 period, which covers years of both economic expansion and of recession. His results therefore should not be biased by the state of the business cycle.
LABOR S UPPL Y AND DEMAND 57 1970-1982 increase in the unemployment rate was 47 percent (Podgursky, 19841. Unemployment is substantially lower among whites than among blacks; black male unemployment rates are from two to three times higher than those of white men (President's Council of Economic Advisers, 1987, Table B-38; U.S. Bureau of the Census, 1986, Table 6621. In general, young people and minorities bear the brunt of cyclical downturns in the economy because of labor market imperfections (e.g., discrimina- tion) and skill differentials. Conversely, a full-employment economy (in which unemployment is restricted largely to those individuals voluntarily engaged in job search) particularly benefits disadvantaged workers (see Chapter 51. Displacement One group of the unemployed whose situation has attracted consider- able attention and concern (Flaim and Sehgal, 1985; Podgursky, 1987; Secretary of Labor's Task Force on Economic Adjustment and Worker Dislocation, 1986; U.S. Congress, Office of Technology Assessment, 1986a, 1986b) comprises experienced workers who suffer permanent job loss or "displacement." The U.S. Bureau of Labor Statistics (BLS) 1984 survey of displaced workers focused on workers who had lost jobs because of plant shutdowns, an employer going out of business, or permanent layoffs resulting from other factors (Flaim and Sehgal, 19851.9 Displaced workers are widely cited (U.S. Congress, Office of Technology Assessment, 1986b) as victims of technological change. How large a share of the unemployed population comprises displaced or dislocated workers? Any estimate of this share depends on the definition of dislocation used. Counting all workers suffering from perma- nent job loss as displaced, BLS estimated that 11.5 million workers were displaced during 1979-1983, yielding an average annual flow of 2.3 million workers. (Unpublished BLS data cited in U. S. Congress, Office of Technology Assessment, 1986a, suggest that the flow ranged from 1.2 million workers in 1979 to 3.3 million workers in 1983.) When the definition of a displaced worker is restricted to "experienced" workers sThe BLS administers the quarterly Current Population Survey (CPS), gathering data on the employment and earnings of respondents. In January 1984 and January 1986, 13LS added a special supplement to the CPS to obtain data on the former earnings and employment status of workers (20 years of age or older) who had lost their jobs. These data are the basis of the analyses by Flaim and Sehgal (1985), Podgursky (1986), and others. 9Estimating the size of the displaced worker population is a complex problem. The Secretary of Labor's Task Force on Economic Adjustment and Worker Dislocation (1986) developed seven definitions of displaced workers, each of which yielded a different estimate of the displaced worker population.
58 TECHNOLOGY AND EMPLOYMENT workers with at least 3 years' experience in the job from which they were laid ok the BLS survey data suggest that a total of 5.1 million workers were displaced during 1979-1983 (Flaim and Sehgal, 1985), which implies an annual average flow of slightly more than 1 million workers. (BLS data cited in U.S. Congress, Office of Technology Assessment, 1986a, esti- mate that this flow ranged from 0.6 million workers in 1979 to 1.4 million workers in 1983.) Podgursky (1987) also analyzed the number and characteristics of workers displaced during 1979-1982, restricting his population to full-time nonagricultural workers. He estimated that 6.4 million workers were displaced during this period, yielding an average annual how of 1.6 million displaced workers.~° Significantly, Podgur- sky's comparative analysis of data from the 1984 and 1986 displaced worker surveys revealed little decline in the average rate of displace- ment 1.5 million workers per year during 1981-1984. Although the BLS estimate of the number of experienced workers displaced annually (regardless of the causes of displacement) is no more than 10-13 percent of the total unemployed population at any point in time, if one considers the share of unemployment accounted for by all displaced workers, the annual flow of such workers increases to 20-31 percent. Moreover, the contribution of displaced workers to increases in unemployment since 1980 has been unusually high (Summers, 1986~. The Secretary of Labor's Task Force on Economic Adjustment and Worker Dislocation (1986) estimated that more than 50 percent of the increase in unemployment during the 1981-1982 recession resulted from permanent job loss a substantially higher figure than the permanent job loss share of the unemployment increase (roughly 37 percent) in the three prior recessions. Once displaced, how long are workers unemployed? According to the 1984 BLS survey, nearly 25 percent (1.3 million) of the 5.1 million workers displaced during 1979-1983 were still unemployed in January 1984, and more than 13 percent (700,000) of those 5.1 million had left the labor force during 1979-1983 (Flaim and Sehgal, 19851. As of January 1984, 60 percent of the experienced workers displaced during the previous 5 years had found employment, albeit at wages that may have been lower than those of their previous jobs; 67 percent of the displaced workers surveyed in January 1986 had found jobs (Flaim and Sehgal, 1985; U.S. Bureau of Labor Statistics, 1986d). Podgursky's analysis (1987) revealed '°Unlike BLS, however, Podgursky did not use job tenure to further restrict his population of displaced workers. He did restrict his analysis to workers who had lost their jobs at least 12 months prior to the date of the 1984 and 1986 BLS surveys in order to better assess the postdisplacement unemployment history of survey respondents and minimize the share of workers who were in fact laid off temporarily.
LABOR SUPPLY AND DEMAND 59 little if any change in the median duration of unemployment following displacement in his samples of the displaced worker populations from the 1984 and 1986 surveys; median weeks of unemployment for blue-collar males fell from 26 weeks (in the 1984 survey, based on 1979-1982) to 20 weeks (from the 1986 survey, covering 1981-1984), whereas median female blue-collar unemployment increased from 40 to nearly 48 weeks. The median duration of white-collar workers' unemployment following displacement fell from 14 weeks to 12 weeks during this period. Bendick and Devine (1981) found that the geographic region within which displacement occurred was more significant in explaining the duration of unemployment than the industry from which workers were displaced. Indeed, according to the Secretary of Labor's Task Force on Economic Adjustment and Worker Dislocation (1986), each additional percentage point in the regional unemployment rate added 1-4 weeks to the average duration of unemployment for displaced workers in that area. Other analyses of displaced workers (Flaim and Sehgal, 1985) found that the earnings losses associated with reemployment after displacement were largest in areas of high unemployment. What are the financial consequences of displacement? Of the 5.1 million experienced workers identified as displaced in the January 1984 BLS survey, a sizable portion 1.6 million-did not receive unemployment benefits. Moreover, nearly 50 percent of the 3.5 million experienced workers who received benefits had exhausted them by January 1984. About 60 percent of the workers still unemployed in January 1984 who were covered by health insurance in their previous jobs (a total of 1 million) had lost health insurance coverage. The median ratio of earnings in new jobs to earnings in previous jobs for displaced workers who previously were full-time workers and eventually found other full-time employment (59 percent of blue-collar and 65 percent of white-collar workers in the 1984 survey) in Podgursky's sample was 93 percent for blue-collar workers and 99 percent for white-collar workers (Podgursky, 19874. This median disguises considerable variance, however, as 30 percent of blue-collar workers and 24 percent of white- collar workers found jobs that paid less than 75 percent of the wages they received in their previous jobs. For experienced workers previously employed in durables manufacturing who were displaced during 1979-1983, median weekly earnings declined by more than 20 percent, from $344 to $273 (Flaim and Sehgal, 19851. Many displaced workers thus face considerable income losses as a result of layoffs, although they may own substantial assets (primarily homes). Data from the 1984 BLS survey (Flaim and Sehgal, 1985) suggest that the goods-producing sector accounted for the majority of displacements (60 percent), although it employed less than one-third of the total U.S.
60 TECHNOLOGY AND EMPLOYMENT work force. Durables manufacturing, with 12 percent of nonfarm employ- ment, accounted for 33 percent of total displacements. The majority of displacements during 1979-1983 occurred in blue-collar occupations specifically, among operators, fabricators, and laborers (Flaim and Sehgal, 1985; Podgursky, 19871.~ Displaced workers tended to be younger and were more likely to be female or black than employed workers. The duration of unemployment was positively associated with age: older displaced workers experienced longer periods of unemploy- ment. Race was the single most significant characteristic in explaining the duration of displacement; displaced black workers endured significantly longer periods of unemployment (Podgursky, 1987~. Few studies of displaced workers analyze the relationship between education and displacement. Those that do (e.g., Flaim and Sehgal, 1985) have found that better-educated workers fare better after layoffs. Accord- ing to Flaim and Sehgal (1985), "about 75 percent of those who had been in managerial and professional jobs were back at work when interviewed fin the BLS survey of January 19841. In contrast, among the workers who had lost low-skill jobs as handlers, equipment cleaners, helpers, and laborers, less than one-half were working in January 1984" (p. 61. Podgursky (1987) also found that higher educational attainment was associated with shorter spells of unemployment after displacement. The evidence that higher levels of skill are associated with shorter unemployment is significant in view of the fact that many displaced workers have serious deficiencies in basic skills. Bendick (1982) found that 34 percent of those workers from declining industries who were unemployed for 8 weeks or more did not have high school diplomas. Moreover, 49 percent of workers with such limited educational attainment were functionally illiterate. The U.S. General Accounting Office (1987a), citing unpublished data from the January 1984 BLS survey, reported that 32 percent of the dislocated workers unemployed as of January 1984 were high school dropouts and thus may have had serious basic skills deficiencies. Among the causes of recent worker displacement, domestic technolog- ical change appears to be a relatively minor factor. Although the 1984 and 1986 surveys of displaced workers did not determine the causes of worker displacement, a 1986 U.S. General Accounting Office (GAO) study did pursue this question. GAO surveyed approximately 400 establishments to assess the reasons for plant closures and permanent layoffs, the events "The census category of "operators, fabricators, and laborers" includes machine operators; assemblers; inspectors; welders; motor vehicle operators; operating engineers; freight, stock, and equipment movers; and general laborers. For a complete list of the detailed occupational categories included in blue-collar employment, see U.S. Bureau of the Census, 1980, pp. xvi-xviii.
LABOR SUPPL Y AND DEMAND 61 responsible for most worker displacement. The most significant cause of these events, cited by 70 percent of the respondents, was reduced product demand. Increased competition, high labor costs, and the high value of the U.S. dollar accounted for 69, 57, and 32 percent, respectively, of the responses. Those causes of displacement that appeared to be directly related to technological change facility obsolescence and production automation accounted for 23 percent (lOth out of 14 causes) and 16 percent (12th out of 14), respectively, of the responses. Although these GAO data are subject to recall bias and differing interpretations of the meaning of the various causes of layoffs and closures, they suggest that technological change is not one of the primary causes of worker displacement. Nonetheless, technological change in foreign firms or nations often underpins trade-related displace- ment. When we view the issue in such global terms, it increases the likelihood that technological change in other nations may play a significant role in the displacement of American workers. Although technological displacement is not a large problem for the U.S. economy in the aggregate, for those workers experiencing pro- longed unemployment, the financial and emotional costs of technologi- cal displacement are enormous. We believe that the costs of displace- ment, regardless of its cause, are often so high that ameliorative policies are needed. (See Chapter 7 for a discussion and critique of current pub- lic policies for displaced workers; Chapter 10 presents policy recom- mendations and options for adjustment assistance for displaced work ers.) TRENDS IN LABOR SUPPLY Long-Term Growth LABOR FORCE GROWTH, 1947-1986 The supply of labor in the economy is an important influence on aggregate unemployment. Periods of rapid growth in the labor supply, other things being equal, will exhibit higher rates of unemployment. The level of aggregate unemployment in turn strongly affects the ease with which displaced workers find new jobs, which influences the duration of unemployment they face and the level of wages associated Leonard (1986) notes that during a period of rapid growth in the labor force (1979-1982) in Wisconsin, an average annual rate of decline in employment of less than 1.2 percent doubled the state's unemployment rate in 3 years, from 5 percent in 1979 to 10 percent in 1982.
62 TECHNOLOG Y AND EMPLO YMENT 200 180 160 140 120 80 60 40 20 Population _ _ SCOOT _ ~ o - 1 1 _~ 1 1 TOO _~ ,~ - Labor Force 1947 1953 1959 1965 1971 YEAR 1977 1983 1989 1995 FIGURE 3-1 U.S. noninstitutional population 16 years of age and older and the civilian labor force, 1947-1995. SOURCE: President's Council of Economic Advisers (1987) and U.S. Bureau of Labor Statistics (1986b). with reemployment. The employment prospects of labor force entrants also are affected by conditions of labor supply and demand. Trends in labor supply thus are important to predicting the ease with which labor markets adjust to the adoption of new technologies. The U.S. civilian labor forced grew at an average rate of 1.9 percent per year during 1947-1986, whereas the civilian noninstitutional U.S. population 16 years of age and older grew 1.6 percent per year (Figure 3-11. Labor force growth was higher than the rate of growth of the U.S. population during this period due to the entry of the baby boom cohort into the labor force, as well as to increases in the proportion of the female population active in the labor force. During 1970-1986, the U.S. civilian labor force grew by 2.3 percent per year on average, compared with a growth rate of 1.7 percent per year for the civilian noninstitutional population (President's Council of Economic Advisers, 19871. i3The civilian labor force consists of individuals, excluding members of the active-duty armed forces, 16 years of age and older who are employed or seeking employment. '4This population excludes individuals in prisons, hospitals, and mental institutions.
LABOR SUPPLY AND DEMAND 63 PROJECTED GROWTH, 1984-1995~5 The U.S. Bureau of Labor Statistics (1986b) computes projections of labor force growth by forecasting labor force participation rates and applying those rates to the Census Bureau's population projections. BLS's projections for "low," "moderate," and "high" rates of labor force growth are based on different assumptions about the characteristics of the population and the labor force. The "moderate" projections indicate that the total civilian labor force will grow much more slowly in the future by comparison with 1970-1984. Projected average annual growth rates in the labor force that is, rates of growth in the supply of labor are 1.3 percent during 1984-1990 and 1 percent during 1990-1995, substantially below the rate during 1970-1984. The corresponding pro- jected annual growth rates for the civilian population are 0.9 percent for 1984-1990 and 0.8 percent for 1990-1995. Labor force growth rates are projected to remain higher than population growth rates because of continued growth in labor force participation. The most dramatic change forecast for 1984-1995 is the 16 percent reduction in the size of the 16- to 24-year-old entrant cohort, from 24 million in 1984 to 20.2 million in 1995 (U.S. Bureau of Labor Statistics, 1986b).~6 This cohort also will experience minimal change in gender and racial composition between 1984 and 1995. Although women and minorities account for approximately 75 percent of the projected growth in the labor force through 1995 (Figure 3-2), the total projected growth in the labor force during 1984-1995 of ap- proximately 14 percent is not large enough to cause significant changes in the gender or racial composition of the overall work '5The most recent detailed 1995 labor force projections of the U.S. Bureau of Labor Statistics (1986b) use 1984 as the base year. '6The projected declines in the pool of labor force entrants have raised concerns about the adequacy of the future supply of scientists and engineers to meet the economic and technological challenges faced by this nation. Both the supply of and demand for scientists and engineers are influenced by a wide range of variables, however; consequently, they exhibit considerable flexibility and responsiveness. It is therefore not clear that a reduction in the number of 18- to 24-year-olds will result in a decline in the number of engineers and scientists. Immigration flows, increases in college enrollments among older members of the population, and changes in the share of the population with science and engineering undergraduate degrees that become practicing members of these professions (a decision influenced heavily by salary outlooks) can all offset the impact of declines in the size of the entrant pool (U.S. Congress, Office of Technology Assessment, 1985a). In addition, as the National Research Council's Panel on Engineering Labor Markets (1986) noted, employers can adjust to changes in the supply of scientific personnel in many ways, thus enhancing the flexibility of the overall system.
64 TECHNOLOGY AND EMPLOYMENT Nonwhite Women (15.5%) In,; ~ . . . ~1 Nonwhite Men (10.3%) White Women (49.7%) White Men (24.5%) FIGURE 3-2 Composition of 1984-1995 labor force growth. SOURCE: U.S. Bureau of Labor Statistics (1986b). force.~7 Overall, the gender and racial composition of the 1995 labor force will closely resemble that of the 1984 labor force (Table 3-21. In 1995, BLS projects that nonwhites (a category that includes blacks, Asians, and nonwhite Hispanics) will account for roughly 15 percent of the labor force a modest increase from their current share of 13 percent. The BLS projections incorporate very conservative estimates of the annual How of illegal immigrants. If this source of labor continues to grow, the BLS projections of labor force growth may be low; however, the labor market impacts of such immigration should be significant in a few regions, rather than nationally. The reductions in labor force growth projected by BLS should lessen the labor market pressures that have been partly responsible for high rates of aggregate unemployment during the past decade. Structural and cyclical unemployment, however, will not vanish; as we note in Chapters 4 and 5, those workers who lack basic skills whether they are labor force entrants or experienced workers who have been displaced are likely to face employment problems in the future. One indicator of basic skills preparation is educational attainment. lathe share of 1984-1995 labor force growth accounted for by women and minorities (75 percent) is an increase over 1970-1980, when women and minorities accounted for approximately 65 percent of labor force growth (U.S. Bureau of the Census, 1987, Table 639). The projected change is largely due to an increase in the share of nonwhite men and women, accompanied by a decrease in the share of white men.
LABOR 5 UPPL Y AND DEMAND 65 TABLE 3-2 Composition of Labor Force by Gender and Race (percent), 1984 and 1995 Year Civilian Change Labor Force 19841995a in Share Men 56.253.6 -2.6 Women 43.846.4 2.6 White 86.785.2 - 1.5 Men 49.446.4 -3.0 Women 37.438.9 1.5 Nonwhite 13.314.8 1.5 Men 6.87.3 0.5 Women 6.47.5 1.1 NOTE: Percentages may not total 100 due to rounding. aProjected values. SOURCE: U.S. Bureau of Labor Statistics (1986b). We discuss current levels of attainment in the U.S. work force and projected future trends next. Educational Attainment of the U.S. Labor Force, 1959-1990 Table 3-3 depicts changes in the educational attainment of the civilian labor force since 1959. The median number of school years completed has increased only slightly: from 12 years in 1959 to 12.8 years in 1986. TABLE 3-3 Educational Attainment (percent) of the Civilian Labor Force for Selected Years, 1959-1986 Less Than High College Median No. of Total No. High School School School Years Year of Workers Graduate (4 Years) 1-3 Years 4+ Years Completed 1959 65,842 50.3 30.7 9.3 9.6 12.0 1970 78,955 34.8 39.0 13.3 12.9 12.4 1975 92,328 29.3 39.6 15.5 15.7 12.5 1980 105,449 23.8 40.1 17.9 18.2 12.7 1984 111,943 19.5 40.7 19.0 20.9 12.8 1985 114,256 19.2 40.2 19.5 21.1 12.8 1986 116,087 18.5 40.4 19.9 21.3 12.8 NOTE: These data include members of the labor force who are enrolled in school. SOURCE: U.S. Bureau of Labor Statistics (1985b, Table 61); U.S. Bureau of Labor Statistics, Office of Employment and Unemployment Statistics (1985, 1986).
66 TECHNOLOGY AND EMPLOYMENT The data, however, suggest considerable shifts during this period in the distribution of educational attainment within the work force. The share of the labor force without a high school degree declined from 50.3 percent in 1959 to 18.5 percent in 1986; the share with at least a college degree more than doubled. These changes are particularly noteworthy in view of the dramatic growth in the labor force during 1959-1986. Much of the increase in median educational attainment reflects the entry into the work force of younger, more highly educated individuals combined with the retirement of older workers with lower levels of educational attainment (Barrow, 1985~. In view of the evidence cited earlier in this chapter suggesting that better-educated workers experience shorter periods of unemployment after job loss, these data support guarded optimism about the ability of the U.S. labor force to adjust to future technological change. Other data reveal significant gender-based and racial differences in educational attainment. According to BLS, 20.3 percent of the male labor force had not completed high school as of 1986, a proportion slightly larger than the 16.1 percent of the female labor force lacking a diploma. Men exhibited higher levels of postsecondary educational attainment in 1986; 42 percent of men had attended college for at least 1 year, versus 40.2 percent of women in the labor force. Blacks and Hispanics continued to lag behind whites in educational attainment. In 1986, 73.9 percent of the black members of the labor force and 56.8 percent of Hispanics had high school diplomas; both of these figures are substantially lower than the 82.4 percent of white members of the labor force with diplomas. In addition although 22 percent of white labor force members had completed at least 4 years of college, only 12.3 percent of blacks and 8.9 percent of Hispanics had done so. Nonetheless, the educational gap between black and white members of the labor force has narrowed significantly since 1959; in that year only 25.5 percent of blacks in the labor force had completed high school, a figure well below the 52.6 percent of white workers with high school degrees (U.S. Bureau of Labor Statistics, Office of Employment and Unemployment Statistics, 1986; U.S. Bureau of Labor Statistics, 1985b). Projections of future levels of educational attainment are subject to considerable uncertainty and are heavily influenced by gender-based, racial, and ethnic differences in educational attainment and by changes in these attainment levels. Barnow (1985) used the BLS labor force projections for 1990 to forecast changes in the educational attainment of different groups within the labor force through that year. Barnow, whose projections incorporated changes in educational attainment levels in male and female workers, forecast that in 1990 86.2 percent of adult (25-64 years old) male workers would have a high school diploma,
LABOR S UPPL Y AND DEMAND 67 slightly lower than the 89.5 percent of female workers with diplomas. As in 1984, the postsecondary educational attainment levels of male work- ers were projected to exceed those of female workers, with 26.7 percent of adult male workers and 22.8 percent of female workers having college degrees. It is difficult to extend this analysis of educational attainment to 1995 and broaden it to incorporate changes in the racial composition of the work force because of the lack of data to support projections of changes through 1995 in the educational attainment levels of white and nonwhite workers or of men and women. (Barrow's projections end in 1990 and are not broken down by race.) However, projections of the 1995 educational attainment of the U.S. labor force that hold current levels of attainment constant among men and women and white and nonwhite workers suggest that changes in the racial or gender composition of the future U.S. work force will have a minimal impact on aggregate levels of secondary or postsecondary educational attainment. ~8 The changes in aggregate attain- ment levels due to changes in labor force composition are modest because the U.S. labor force is projected to grow relatively slowly through 1995. Although the 1984-1995 cohort of labor force entrants is projected to include a larger share of women and minorities than did the 1970-1980 cohort, as was noted above, these projected changes in the composition of the entrant cohort imply minimal change in educational attainment levels for labor force entrants through 1995. Although average levels of educational attainment in the U.S. labor force are not likely to change dramatically in the near future, educational attainment appears to be increasingly significant to the economic welfare of an individual. Census data on individual incomes (Table 3-4) reveal that the economic returns from schooling, measured as the differences in median annual incomes for individuals with different educational attain- ment levels, increased during 1973-1984, especially for men and women between the ages of 25 and 34 (these data are discussed in greater detail by Levy, 19871.~9 Measured in constant dollars, the ratio of the median annual income for males between 25 and 34 who completed high school to the median annual income of males who completed only 1-3 years of high school grew by more than 25 percent during 1973-1984 and registered a comparable gain for women. With the exception of high school comple- tion for women over 25 years of age, large increases are apparent between '8These projections combine data from BLS on forecast labor force growth and composition with data from the CPS on current educational attainment rates. '9The effects on incomes of slow economic growth, high inflation, and a rapidly expanding labor force are also revealed in the consistent declines in median annual earnings during 1973-1984 in all groups in Table 3-4.
68 TECHNOLOG Y AND EbIPLO YMENT TABLE 3-4 Education and Median Individual Total Money Income, 1973 and 1984 (in 1982 dollars) 1-3 Years of 4 Years of 4 Years of High School High School College Ratio Ratio Ratio Group (A) (B) (C) (B/A) (C/A) (C/B) Men 25+ years old 1973 17,383 21,839 28,103 1.26 1.62 1.29 1984 11,590 17,414 26,093 1.50 2.25 1.50 Women 25+ years old 1973 5,718 8,004 12,528 1.40 2.19 1.57 1984 5,142 7,252 12,622 1.41 2.45 1.74 Men 25-34 years old 1973 17,032 20,470 23,681 1.20 1.39 1.16 1984 10,081 15,754 21,759 1.56 2.16 1.38 Women 25-34 years old 1973 6,270 7,954 13,196 1.27 2.10 1.66 1984 4,526 7,375 13,228 1.63 2.92 1.79 NOTE: Dollar amounts were adjusted using the implicit price deflators for personal consumption expenditures (President's Council of Economic Advisers, 1987, Table B-3). SOURCE: U.S. Bureau of the Census (1975, Table 58; 1986, Table 33). 1973 and 1984 in the income effects of greater educational attainment in all of the comparisons in Table 3 4.2° These increases reflect the fact that changes in the structure of the economy, new technology, and increased international competition have expanded the returns to individuals (in terms of income) from higher educational attainment, especially within the cohort that entered the labor force during the 1970s. Such gains, however, also mean that continuing racial or ethnic differences in educational attainment will widen the economic gaps among these groups in the U.S. economy. Comparing the Educational Attainment of U.S. Labor Market Entrants with Those of Other Nations The importance of international trade and competitiveness to U.S. living standards, as well as the evidence that other nations may be developing and adopting some new technologies more rapidly than U.S. firms, draws attention to the relative levels of educational attainment of the U.S. and foreign labor forces. How does the educational attainment of Increases in the returns from college education during this period were partially offset by increases in the direct costs of a college education.
LABOR SUPPE Y AND DEMAND 69 the U.S. labor force compare with that of the labor forces of other industrial nations? There are few reliable data on the educational attain- ment of the overall labor force in other industrial nations. For this reason, as well as the importance of labor force entrants' attainment for the projection of future trends in attainment levels, we present comparative data only on entrants in seven other industrial nations. It is difficult to develop international comparisons of levels and trends in educational attainment because educational systems and policies differ markedly across nations. Consider the provision of education and training after the completion of compulsory schooling. Higher education and training in countries such as Japan, the United States, and Canada follow the "schooling model" in which such offerings are integrated into the formal educational system. Other countries for example, West Ger- many follow the "dual model," which is characterized by a strong and highly developed apprenticeship sector. The "mixed model," found in the United Kingdom, places greater emphasis on the informal sector (on-thejob training outside of an apprenticeship system) for education and training beyond the compulsory level. Table 3-5 shows the rates of enrollment in education and training by age (ages 16-19) for several countries. To minimize the differences across the various models, both full-time and part-time participants are included. In all of the countries included in the table, more than two-thirds of the 16-year-olds are in some form of secondary or postsecondary education and training; in four countries the proportion is above 90 percent. For those countries following the schooling model (e.g., the United States), the sharpest drop in participation rates occurs between the ages of 17 and 18, corresponding to the change from secondary to postsecondary edu- cation. For countries following the dual model (e.g., West Germany), the distinction between full-time and part-time participants becomes particu- larly important, depending on the weight assigned to apprenticeship programs. In such countries the bulk of the 16- to 19-year-old age group attends school on a part-time basis. The United States has the largest percentage of 16- and 19-year-olds enrolled in full-time education; the rates for 17- and 18-year-olds are second only to Japan and the Netherlands, respectively. Individuals entering higher education both within and outside of the formal university structure account for a larger share of the relevant cohort in the United States (60 percent) than in any other nation for which data are available (Organisation for Economic Co-operation and Development, 1984~. Although certain educational gaps between the United States and other nations have narrowed in recent years, the United States remains among those countries with the highest levels of participation in secondary and postsecondary education and training. There are only limited data to
70 TECHNOLOG Y AND EMPLO YMENT TABLE 3-5 Enrollment Rates (percent) in Education and Training, 16- to 19-Year-Olds, for Eight Industrial Nations 16-year-olds 17-year-olds 18-year-olds 19-year-olds PT FT PT FT PT FT PT FT Francea 83.9 73.3 68.9 60.0 45.2 42.2 30.0 29.4 West Germanya 92.1 57.3 89.3 38.9 71.5 29.6 41.8 21.3 Italyb 69.1 54.5 70.3 47.4 51.3 36.2 29.4 18.5 JapanC 94.0 94.0 94.0 94.0 n.a. n.a. n.a. n.a. The Netherlands 97.8 92.0 84.7 74.8 62.8 53.4 43.9 34.6 Swedene 87.4 86.7 78.4 76.7 44.7 40.3 23.5 16.9 United Kingdoma 68.0 48.3 52.8 30.2 37.0 16.5 28.5f 13.4 United Statesg 94.3 94.3 87.1 86.6 54.7 51.2 40.9 37.3 NOTE: FT = full-time enrollment; PT = part-time enrollment; n.a. = not available. SOURCE: Organisation for Economic Co-operation and Development (1985). aThe last year for which data are available is 1981. bHigher education is not included, but regional vocational training is included. The last year for which data are available is 1981. CStatistics are from the Ministry of Education; data are for 1980. Figures refer to the proportion of young people completing lower secondary education (approximately 99 percent of the cohort) who continue at the upper secondary level. Participation rates for 17-year-olds are similar to those of the 16-year-old group given the very low dropout rate in Japan. Data are for 1982. eThese figures include the different types of adult education; data are for 1980. fThe figure includes both 19- and 20-year-olds. "Data are for 1982. support assessments of the quality of the education received by U.S. labor force entrants. Recent time series evidence, in the form of scores on standardized tests, suggests that the quality of the educational preparation of entrants may increase in coming years. Since the 1970s successive cohorts of children entering school have scored higher on standardized tests. By some measures, achievement in elementary grades is at its highest level in three decades.2~ In addition, although the gaps in test scores between minority and nonminority groups remain large, they are narrowing. College test scores remain low, but they may 2~Test score data for the state of Iowa provide annually standardized data extending over three decades. Data from 1984 suggest that the median 3rd grader (the 50th percentile) scored better than roughly 68 percent of his or her counterparts in 1954 (Congressional Budget Office, 1986). Although Iowa is not representative of the nation as a whole, these state-level findings are corroborated by the results of the Congressional Budget Office's analysis of a much wider range of tests.
LABOR SUPPLY AND DEMAND 71 also increase as the cohorts with higher test scores move through high school (Congressional Budget Office, 1986~. Nevertheless, although this evidence suggests that future U.S. entrants to the labor force may have better basic skills than those who are currently seeking employ- ment, other evidence indicates that other nations' labor force entrants currently are better prepared in such skills. McKnight et al. (1987) and Lee et al. (1987) suggest that the quality of basic skills training in the United States lags behind that of other industrial nations such as Japan. LABOR DEMAND The level of demand for labor in the United States is determined primarily by the rate of growth of the entire economy, which in turn is affected by a wide range of influences including government policy, external "shocks" (e.g., the oil price increases of 1973 and 1979), and business cycle fluctuations. Technological change has little impact on aggregate labor demand. At the level of individual sectors or industries, however, the demand for labor is affected by the rate of growth in out- put and the level of wages, both of which may be influenced by technological change. As was noted previously, by reducing the cost of output, domestic technological change often contributes to increased domestic and international demand for output, offsetting all or much of the impact on labor demand of any reductions in the amount of labor required per unit of output that result from the use of new technology. In some cases, however, technological change may reduce U.S. em- ployment. U.S. industries in which rates of technological change, productivity growth, and output cost reduction lag behind those of foreign firms may experience employment losses as U.S. firms lose export markets and domestic sales to the products of foreign, techno- logically superior competitors. Significant lags in U.S. technological performance therefore may contribute to erosion in employment and wages. As we noted in Chapter 1, an alternative means of reducing the cost and price of U.S. goods is reductions in U.S. wages, which can occur through cuts in the dollar value of U.S. real wages or by reductions in the value of the dollar in relation to foreign currencies. These alternatives, of course, may reduce the standard of living in the United States. Moreover, as rapid rates of international technology transfer continue, the level of techno- logical sophistication within relatively low-wage nations will increase- which means that, in the absence of technological change, U.S. wage cuts or dollar devaluations will have to be significant for this nation to compete successfully with other countries.
72 TECHNOLOGY AND EMPLOYMENT TABLE 3-6 U.s. Postwar Productivity Growth (percent) in the Nonfarm Business Sector (average annual growth rates in output/hour) 194~1985 194~1957 1957-1966 1966-1973 1973-1979 1979-1985 Sector Nonfarm business 1.8 2.5 2.8 188.8.131.52 Manufacturing 2.6 2.7 2.9 184.108.40.206 Nonfarm, nonmanufacturing 1.5 2.3 2.7 1.3-0.1-0.1 SOURCE: Calculated from unpublished data, U.S. Bureau of Labor Statistics, Office of Productivity and Technology. Growth in Labor Productivity and Output Labor productivity growth provides one index of the rate of technolog- ical change within an economy. This measure admittedly is imperfect, as Chapter 2 noted; technological change need not be laborsaving in character (and thus an agent of labor productivity enhancement), and other important nontechnological influences (e.g., changes in the scale of production establishments, the rate of capital formation, improvements in the education of the work force) also affect the rate of advance in labor productivity. Nevertheless, this datum provides a crude index of the rate at which technological change is spreading throughout the economy. As the summary of this chapter and Chapter 4 discuss at greater length, growth in labor productivity, which frequently results from technological change, supports growth in real wages and international competitiveness. Aggregate productivity growth, which is the weighted average of labor productivity growth in different sectors of the economy (weighted by the sectors' shares of total output), has remained below the average rates of the 1950s and 1960s during most of the 1970s and 1980s. Table 3-6 presents growth rates in nonfarm labor productivity over five postwar intervals, beginning and ending at comparable points in the business cycle. These figures show high rates of productivity growth during 1948-1966, which were followed by a decline during 1966-1973, further serious deterioration during 1973-1979, and a modest recovery during the most recent business cycle (1979-19851. Consistent with the discussion in Chapter 2, recent productivity growth trends do not support the hypothesis that domestic technological change has accelerated in the overall economy. Moreover, trends in labor productivity growth and unemployment strongly suggest that, consistent with the previous discussion of the effects of technological change on employment, productivity growth is not associated with higher aggregate unemployment. Figure 3-3 displays annual rates of productivity growth in
LABOR S UPPL Y AND DEMAND 73 9 8 o I O ~ ~ a: Z Z I ~ [L O J Z ~ Z O 7 6 5 44 3 2 l -1 , - --~] Unemployment Rate ~] b\~; Percent Change In Ou~uVHour . 1 ~ _ _ 1 1 1949 1954 1959 1964 1969 1974 1979 1984 YEAR FIGURE 3-3 Percentage change in output per hour (labor productivity growth) and the annual unemployment rate, 3-year moving averages, 1949-1985. SOURCE: President's Council of Economic Advisers (1986). the nonfarm business sector, along with annual unemployment rates, in 3-year moving averages (to reduce the effects of business cycles on long-run trends) for 1949-1985. Throughout the post-1949 period, but especially after 1973, the trends in these two series diverge unemploy- ment has climbed, whereas productivity growth has declined. Baumol (1986) presents similar data for longer time periods. Table 3-6 also distinguishes trends in manufacturing labor productivity from those in nonmanufacturing and shows that rates of labor productiv- ity growth in manufacturing and in nonmanufacturing have behaved quite differently in recent years. U.S. productivity in the nonmanufacturing sector began to decrease during 1966-1973; during the 1973-1979 business cycle, deterioration in productivity growth occurred within both sectors. During the most recent business cycle (1979-1985), productivity growth rates have increased only in manufacturing. Thus, much of the slowdown in measured U.S. productivity growth that has occurred during the last two decades is located within nonmanufacturing industry. Two important points must be noted in any discussion of productivity trends. The first concerns the difficulty of measuring productivity in the nonmanufacturing sector of the economy. The data in Table 3-6, which suggest that productivity growth in the nonmanufacturing sector has been
74 TECHNOLOGY AND EMPLOYMENT low, conflict with anecdotal evidence from industries such as financial services in which product innovation and productivity growth, much of which are based on information and computer technologies, appear to be considerable. If product innovation has in fact been particularly rapid in financial services, telecommunications, and other nonmanufacturing in- dustries in recent years, the quality of productivity data for these industries may have declined because of the problems posed by product innovation for the measurement of output and productivity (see Chapter 21. An unknown portion of the low productivity growth measured in nonmanufacturing industry also may reflect problems of measuring inputs and outputs in this sector. Finally, the quality of employment and output data for the service industries is impaired by the classification schema used for these data (Kendrick, 1986; Marimont and Slater, 1986; see also Chapter 81. Rates of productivity growth in the nonmanufacturing sector thus may be understated in the available public data. A second important point concerns the lack of explanations for the general decline of productivity growth rates in the United States and in other industrial nations since the early 1970s. Despite extensive research (summarized in Baily, 1986, and Wolff, 1985), there is no widely accepted explanation for the post-1973 decline. Studies have considered lower investments in R&D or physical capital, intersectoral shifts in labor between manufacturing and nonmanufacturing, government regulation, and lower labor force quality as contributors to changing productivity growth rates, but most analysts have yet to accept any one factor or combination of factors as a satisfactory explanation. In addition, although foreign productivity growth rates typically have exceeded U.S. perform- ance, all industrial nations have experienced declines in productivity growth rates since 1973 (U.S. Bureau of Labor Statistics, 1986c). Many of the causes of slower productivity growth therefore are not unique to the United States but have affected all industrial nations. Changes in the Sectoral Composition of Output and Employment The top portion of Table 3-7 gives the share of total private nonfarm business output accounted for by each of 12 major U.S. economic groups during 1948-1985. The manufacturing sector's share ("Durables" and "Nondurables") of total output has been remarkably constant throughout this period in 1948, it was 27.9 percent; in 1985, it stood at 28 percent. The table also highlights the growth during this period in the shares of total output for the finance, insurance, and real estate group, which increased from 9.5 to 11.3 percent; for the services group (including business and health services), which increased from 10.7 to 15 percent;
LABOR S UPPL Y AND DEMAND 75 TABLE 3-7 Percentage of U.S. Gross Domestic Product Originating in Industry Group Group1948 1957196619731979 1985 Durables17.0 17.818.517.817.2 17.2 Nondurables10.9 10.611.011.511.1 10.8 Mining8.5 220.127.116.11.3 4.7 Construction10.6 12.211.58.07.0 5.8 Transportation9.0 18.104.22.168.6 4.5 Communications1.1 22.214.171.124.9 3.5 Utilities1.6 126.96.36.199.4 3.7 Wholesale trade6.5 188.8.131.52.8 9.6 Retail trade12.4 12.311.812.111.9 12.3 FIRE9.5 10.210.110.911.8 11.3 Services10.7 9.710.511.813.3 15.0 Gov't. enterprises2.3 184.108.40.206.6 1.6 Growth Rates (percent) of Gross Domestic Product Originating in Group 1948- 1957-1966-1973-1979- 1948-1966- 1948 Group1957 1966197319791985 19661985 1985 Durables4.0 220.127.116.11.1 4.32.3 3.2 Nondurables3.2 18.104.22.168.6 3.82.6 3.2 Mining3.2 2.02.1-0.40.1 2.60.7 1.6 Construction5.1 3.5- 1.90.3- 1.3 4.3- 1.0 1.6 Transportation-0.2 22.214.171.124- 1.4 1.31.4 1.4 Communications6.8 6.37.86.05.0 6.56.3 6.4 Utilities8.5 126.96.36.199.5 7.24.0 5.5 Wholesale trade4.1 5.15.13.03.4 4.63.9 4.2 Retail trade3.4 188.8.131.52.6 3.52.9 3.2 FIRE4.4 184.108.40.206.4 4.13.3 3.7 Services2.5 220.127.116.11.0 3.74.5 4.1 Gov't. enterprises0.8 18.104.22.168.7 2.41.9 2.2 Total3.5 4.03.32.52.0 3.82.7 3.2 NOTE: Gross domestic product calculations based on constant 1982 dollars. FIRE = Finance, insurance, and real estate. SOURCE: Calculated from U.S. Bureau of Labor Statistics, Office of Productivity and Technology, unpublished data developed by BLS from Department of Commerce (Bureau of Economic Analysis national income and product account data and the Federal Reserve) Index of Industrial Productivity for Durable and Nondurable Manufacturing. and for the communication services group, which grew from 1.1 to 3.5 percent. The bottom portion of the table shows the rates of growth in the output of each of these groups. Two conclusions are obvious. Output growth rates in most groups were higher during 1948-1966 than during 1966-1985. Slower
76 TECHNOLOG Y AND EMPLO YMENT aggregate output growth during 1966-1985, combined with rapid growth in the labor supply, contributed to higher aggregate unemployment during this period. The second conclusion concerns the relative rates of output growth in the manufacturing and nonmanufacturing sectors. During 1948-1966, the rate of growth of output for manufacturing was above the economy-wide average; it fell below the average during 1966-1985. Output growth in most nonmanufacturing industries (with the exception of mining, construction, and transportation), on the other hand, remained strong after 1966. Intersectoral differences in productivity growth, combined with intersectoral differences in output growth, have affected the level of demand for labor in the manufacturing and nonmanufacturing sectors of the U.S. economy. Because sectoral employment growth is the difference between growth in sectoral output and growth in sectoral productivity, these differences affect the level of labor demand. Within the nonmanufacturing sector, low productivity growth and rapid output growth (outside of mining, transportation, and construction) have re- sulted in a strong demand for labor. The manufacturing sector, on the other hand, has experienced a resurgence of productivity growth since 1979 to levels comparable to or greater than those of the l950s and early 1960s. Slow growth in output, however, also has characterized this sector since 1979 because of import penetration of U.S. markets for manufactured goods and the slow growth or collapse of foreign markets for U.S. manufactured exports. During 1980-1984 alone, Davis (1986) estimated that declines in U.S. merchan- dise exports resulted in the loss of as many as 1.8 million jobs, many of which were in manufacturing. According to Davis (1986), "tE]xport- related jobs accounted for 80 percent of the total 1980-1984 decrease in manufacturing employment Efrom 20.3 million in 1980 to 19.4 million in 19841" (p. 92~.22 Hight (1986) estimated that increased imports during 1982-1984 cost nearly 800,000 U.S. jobs in mining and manufacturing, 77 percent of which were in 14 (out of a total of 80) manufacturing industries.23 Growth in demand for the output of U.S. manufacturing from both domestic and foreign markets supported employment growth during 22Pollock and Almon (1986) also present data suggesting that the negative employment impacts of increased imports and declining exports during 1980-1985 substantially exceeded those of technological change in all 35 of the manufacturing industries they examined. 23Apparel, motor vehicles, computers and office equipment, electronic components, leather products, radio and television receivers, primary metals, radio and communications equipment, industrial chemicals, furniture and fixtures, general industrial machinery, electrical machinery, sawmills and planing, and basic steel.
LABOR SUPPL Y AND DEMAND 77 the 1970s, but the collapse of export markets and surging imports contributed to declines in the rate of growth or reductions in the level of employment in many manufacturing industries in the 1980s. Manufacturing's share of U.S. private nonfarm employment has been declining gradually since 1919. As U.S. Bureau of Labor Statistics (1985a) figures show, during 1919-1948, this share declined from 44 percent to 40 percent; since 1948, the rate of decline has increased, particularly since 1966. The rate of decline in employment within manufacturing, relative to the rest of the private nonfarm economy, reached an average annual rate of 2 percent after 1966 and then increased to 3 percent per year during 1979-1985. Thus, the share of private nonfarm employment accounted for by manufacturing stood at 36, 28, and 24 percent, respectively, in 1966, 1979, and 1985. Groups registering the most dramatic gains in their shares of employment since 1966 include wholesale and retail trade, services, and finance, insurance, and real estate. Despite these declines in its share of employment, manufacturing had 19.3 million employees in 1985, versus 19.2 million in 1966 (President's Council of Economic Advisers, 1987, Table B-401. Growth in the nonmanufacturing share of total employment since 1966 reflects more rapid growth in this sector, rather than absolute declines in manufacturing employment. Since 1979 resurgent productivity growth and stagnant output growth in manufacturing, combined with rapid output growth and stagnant productivity growth in nonmanufacturing industry, have accelerated longstanding trends of decline in the manufacturing sector's share of total employment. What role has domestic technological change played in these trends? Certainly, the resurgence in productivity growth within manufacturing must be taken as a partial indicator of improved domestic technological performance. The reasons for this resurgence, however, are no better understood than the reasons for the decline in manufac- turing productivity growth during 1973-1979. Moreover, improved do- mestic productivity growth during 1979-1985 did not translate into growth in manufactured exports and employment. In assessing the effect of productivity growth on employment, we must also consider the reasons for the decline in manufactured exports and growth in imports after 1980. INTERNATIONAL TRADE, TECHNOLOGICAL CHANGE, AND U.S. EMPLOYMENT International Trade and Employment Manufacturing industry is more exposed to international competition than most other nonagricultural industries by virtue of the internationally
78 TECHNOLOGY AND EMPLOYMENT "traceable" character of its outputs.24 Manufactured exports currently dominate U.S. nonagricultural exports, accounting for nearly $182 billion in 1984 (nonagricultural merchandise exports) and substantially exceeding total services exports of $69 billion-$91 billion (President's Council of Economic Advisers, 1987, Table B-100; U.S. Congress, Office of Tech- nology Assessment, 1986c). Exports of manufactured goods also support numerous nonmanufactur- ing jobs. The U.S. Department of Commerce's International Trade Administration (1983) estimated that merchandise exports in 1982 (total ing $211 billion) supported nearly 5 million jobs, of which more than 4 million depended on the export of manufactured goods. Of these 4 million jobs, slightly more than one-half, or 2.3 million, were located in manu- facturing; 1.7 million jobs were in the nonmanufacturing sector of the economy. During 1979-1985, a number of factors were responsible for reduced U.S. manufactured exports and increased U.S. imports of manufactured goods. These influences included the appreciation of the U.S. dollar, which was associated with the combination of large federal budget deficits and tight monetary policy that characterized the U.S. economy in the early 1980s (Fieleke, 19841. Appreciation of the dollar depressed the U.S. price of imports of foreign goods while increasing the price of U.S. exports, all of which had a considerable erect on the U.S. trade balance (Figure 3-4~. Imports increased from $332 billion25 in 1980 to more than $521 billion in 1986; U.S. exports declined from $389 billion in 1980 to $371 billion in 1986 (President's Council of Economic Advisers, 1987, Table B-201. Economic growth also was weak during the early 1980s in many of the countries that are important U.S. export markets, thereby reducing the possibilities for increased exports. Furthermore, the under- lying competitiveness of U.S. manufactured products, which is revealed in product quality and price/performance characteristics, may have declined during the past decade, as suggested in a number of studies (Finan et al., 1986; President's Commission on Industrial Competitive- ness, 19851; this issue is discussed in greater detail below. Technological Change and U.S. Exports U.S. exports since 1945 from the manufacturing and nonmanufacturing sectors alike have been goods whose production depended on large Widespread application of information and other computer-based technologies, how- ever, is gradually changing the extent to which services for example, business, financial, and communications-can also be traded internationally. 25All figures are in 1982 dollars.
LABOR SUPPL Y AND DEMAND 79 ~ o 45 lo ~ 11 o CO C] ~ o - X ~ - llJ - Z ~ 160 140 120 100 80 60 - 20 o O LU ~ Cal CO Z ~ ~ 0 C] o Hi: - G m - - - - - Nominal Dollars ~ -20 -40 -60 _ -80 1 1 1 1 1 1 -~ 1979 1980 1981 1982 1983 1984 1985 1986 Trade Balance\ - ~O" YEAR FIGURE 3-4 Changes in the U.S. trade balance and the value of the dollar, 1979-1986. SOURCE: President's Council of Economic Advisers (1987). investments in R&D and on skilled, relatively high-wage labor. Numerous studies have documented a significant relationship between the high skill or R&D content of manufactured products and the role of those products in U.S. exports (see Gruber et al., 1967; Keesing, 1967~. U.S. exports also are more heavily dependent on it&D-intensive industries than are the exports of other industrial nations (Organisation for Economic Co- operation and Development, 1986b). Bartel and Lichtenberg (1987), among others, argue that this nation has specialized in the export of manufactured goods embodying advanced technologies, the development and initial production of which are relatively intensive in their use of skilled labor and scientific talent (for reasons noted in Chapter 21. U.S. export-intensive industries are large employers of skilled and professional labor when compared to all U.S. manufacturing industry (U.S. Congress, Office of Technology Assessment, 1986b; U.S. International Trade Com- mission, 19831. The Office of Technology Assessment (1986c) study of international trade in services corroborates this analysis for the nonmanufacturing sector; services exports tend to support high-wage, high-skill employment in comparison to overall U.S. nonmanufacturing employment. During the 1960s and 1970s, the manufacturing workers displaced by
80 TECHNOLOGY AND EMPLOYMENT increased imports of foreign goods were employed in relatively low-wage, low-skill jobs (Aho and Orr, 19811. As other nations continue to develop their technological and manufacturing capabilities, lower-skill, lower- wage U.S. manufacturing jobs will continue to be threatened. Moreover, the nonmanufacturing sector will feel the effects of increased import penetration as well; workers in that sector who are displaced because of increased imports are also likely to be employed in relatively low-skill, low-wage jobs. A significant difference between the 1980s/1990s and the 1960s/1970s is that much low-wage foreign competition in manufacturing no longer is low-productivity competition. In part because of more rapid rates of technology transfer, as well as increased technological sophistication in many foreign economies, production and product technologies in some industries within many low-wage competitor nations now approach or exceed those of the United States in quality and product sophistication. This changing international environment is likely to increase the impor- tance of investments by U.S. firms and public institutions in the skills of the labor force and in the R&D necessary to generate and adopt advanced technologies in both the manufacturing and nonmanufacturing sectors. The Competitiveness of U.S. Industry In view of the importance of international trade for U.S. employment and wages, recent signs of declining U.S. competitiveness are a cause for concern. International industrial competitiveness-that is, the ability of U.S. products to preserve or increase their share of international mar- kets subsumes a number of factors, among them product quality (in- cluding technological sophistication and design quality), product service, and price. As we noted in Chapter 1, a great many nontechnological factors also affect competitiveness, including the rate of domestic savings and capital formation, other nations' trade and financial policies, and the exchange rate of U.S. and foreign currencies. Because the appreciation of the U.S. dollar during 1980-1985 affected the price of U.S. goods in international trade, regardless of changes in their quality, at least some part of U.S. industry's competitiveness problems is related to the dollar's behavior during the first half of the 1980s. Other evidence, however, suggests that declining competitiveness in some U.S. manufacturing industries predates the high dollar exchange rates of 1980-1985. Import penetration in 28 of 40 major U.S. manufac- turing industries increased during 1972-1982, a decade that spanned a period of undervalued as well as overvalued U.S. dollars (President's Commission on Industrial Competitiveness, 19851. In addition, the U.S.
LABOR SUPPL Y AND DEMAND 81 balance of trade in high-technology products, historically a U.S. export stronghold, has been deteriorating since the late 1970s; the 1986 balance of trade in these items yielded a deficit of $2.6 billion (based on unpublished 1986 data from the U.S. Department of Commerce, Interna- tional Trade Administration). According to Finan et al. (1986), much of the deterioration in the U.S. high-technology trade balance reflects a combination of stagnant exports of U.S. goods and increases of more than 40 percent in imports in some sectors. In response to these trends, U.S. firms in several high-technology industries, especially electronics, have moved a larger share of their production to "offshore" locations: This movement to offshore sourcing has developed especially rapidly with respect to Hong Kong, Taiwan, Korea, and Singapore what we designate collectively here as the NICs [newly industrializing countries]. U.S. firms are sourcing subassemblies from low-labor-cost countries where usually the exchange rate has moved favorably that is, where the dollar has remained relatively strong. As a result of U.S. firms' sourcing decisions, the trade balance with the NICs has deteriorated significantly. (p. 31) Sanderson (1987) and others (e.g., Cyert, 1985), however, argue that widespread adoption of new computer-based manufacturing technologies within U.S. manufacturing, as well as the increasing competitive impor- tance of shorter product development cycles, may reduce the attractive- ness of offshore manufacturing for many U.S. firms in the future. The benefits of offshore manufacturing also should be reduced by declines in the foreign exchange value of the U.S. dollar from the levels it achieved in 1984 and 1985.26 Unit Labor Costs in U.S. and Foreign Manufacturing, 1950-1985 Technological change and productivity growth can accelerate output growth by enhancing the competitiveness of U.S. industry. As stated in Chapter 1, because productivity and output growth are linked in an open economy, growth in productivity within U.S. manufacturing, which reduces the labor costs of U.S. products, can reduce the price of U.S. manufactured goods in overseas markets. Reduced prices in turn lead to Caterpillar, Inc., which manufactures earthmoving equipment, increased offshore production from 19 percent of total sales in 1982 to 25 percent in 1986; it also increased its use of foreign sources of parts and components by a factor of four. The declining foreign exchange rate of the U.S. dollar sharply reduced the profitability of this strategy and contributed to the firm's loss of $148 million in the fourth quarter of 1986 (The Economist, April 4, 1987).
82 TECHNOLOGY AND EMPLOYMENT expanded U.S. exports of manufactured goods and reductions in U.S. imports. Conversely, if technological change and productivity growth in U.S. manufacturing industry fall sufficiently behind those of our trading partners, markets for U.S. products will shrink. Nevertheless, the higher productivity growth in U.S. manufacturing after 1979 largely failed to improve, and occurred simultaneously with dramatic declines in, U.S. trade performance. This section examines one explanation for the dis- junction of U.S. productivity and trade performance after 1980. Table 3-8 summarizes the price dimensions of international competi- tiveness and the contributions of productivity growth to the price com- petitiveness of U.S. manufacturing. The table shows changes in unit labor costs for the manufacturing sectors of other industrial nations relative to those of the United States. Unit labor costs measure the labor cost per unit of output of manufacturing industry; they grow with increases in the nominal wage of manufacturing labor. Productivity growth offsets the effect of wage increases on unit labor costs as unit labor requirements decline, so will unit labor costs. If wage increases are comparable to growth in labor productivity and if exchange rates remain constant, unit labor costs will be unchanged. The top panel of Table 3-8 shows the changes in foreign unit labor costs (measured in U.S. dollars) relative to U.S. unit labor costs for five intervals during 1950-1985; entries in the top panel are the sum of the entries in the three lower panels of the table. Negative entries indicate reductions in foreign unit labor costs relative to those of the United States. Because these costs are measured in U. S . dollars, they are affected by exchange rate movements as well as by movements in wages and labor productivity. Since 1979 unit labor costs in all of these foreign industrial nations except Canada have declined relative to those of the United States, the first period during which this has occurred since 1950-1957. As a result, the ability of U.S. manufacturing to compete in world markets declined significantly during 1979-1985, despite significant productivity growth in U.S. industry. Technological change, which affects labor productivity growth, played a major role in the behavior of unit labor costs during this period. The second panel from the top in Table 3-8 shows foreign labor productivity growth rates relative to those of the United States. Negative entries indicate more rapid productivity growth in other nations' manufacturing industry; the table clearly shows that U.S. productivity growth has lagged behind that of other industrial nations throughout the postwar period. The 1979-1985 data, however, suggest that during this most recent period, U.S. labor productivity growth rates approached those of most industrial nations (with the exception of Japan). The relative productivity perform- ance of U.S. manufacturing obviously improved during 1979-1985 to an
LABOR SUPPLY AND DEMAND 83 TABLE 3-8 International Comparisons of Unit Labor Costs, Productivity, and Compensation in Manufacturing in Selected Industrial Nations Annual Growth Rate of Foreign Unit Labor Costs (in U.S. dollars), Relative to the United States Country1950-19571957-19661966-19731973-19791979-1985 . West Germany-2.03.06.93.3-8.9 France-0.60.02.13.2-8.2 Italy-2.81.95.01.8-6.1 · Japan-22.214.171.124.5-6.1 ·U.K.2.01.9-0.27.0-5.4 Canada1.2- 1.80.2- 1.00.2 Annual Growth Rate of U.S. Productivity (output per hour), Relative to Foreign Manufacturing West Germany-4.8-3.3-3.0-2.8-0.2 France-2.2-3.2-3.6-3.5-0.9 Italy-3.6-3.6-4.1- 1.9-0.8 Japan-7.5-5.2-8.3-4.0-2.5 U.K 0.8 -0.5 -2.2 -0.2 -0.9 Canada - 1.4 - 1.3 -2.2 -0.8 1.4 . Annual Growth Rate of Foreign Hourly Compensation (in domestic currency), Relative to the United States Lest Germany 2.7 5.8 4.1 0.0 -0.8 France 4.3 5.0 4.2 6.0 5.2 . Italy -0.7 5.6 8.1 9.7 8.5 Japan 3.8 7.0 9.5 3.0 -2.2 U.K. 1.2 2.5 3.9 9.7 3.7 Canada 0.8 0.8 1.3 2.5 1.4 Annual Growth Rate of the Foreign Exchange Rate, Relative to the U.S. Dollar West Germany 0.0 0.5 5.9 6.2 -7.9 France -2.6 - 1.7 1.5 0.7 - 12.5 Italy 0.0 0.0 1.0 -6.0 - 13.9 Japan 0.0 -0.1 4.2 3.6 - 1.5 U.K. 0.0 0.0 - 1.9 -2.4 -8.2 (Canada 1.9 - 1.3 1.1 -2.6 -2.6 SOURCE: U.S. Bureau of Labor Statistics (1986c).
84 TECHNOLOGY AND EMPLOYMENT unprecedented extent. Moreover, measured in terms of their domestic currencies, hourly compensation for foreign manufacturing workers (the third panel from the top of Table 3-8) consistently has grown more rapidly than compensation for U.S. manufacturing workers. Prior to 1979, the impact of more rapid growth in labor costs on the competitiveness of these nations' manufactured exports was offset by productivity growth rates that also exceeded those of U.S. industry. During 1979-1985, however, increases in foreign worker compensation continued to exceed those of U.S. manufacturing workers, while the productivity gap between U.S. and foreign manufacturing narrowed. Yet, the growth of U.S. unit labor costs remained well above that of other industrial nations. Why? The answer to this question is contained in the bottom panel of Table 3-8. Upward movement in the foreign exchange rate of the U.S. dollar during 1979-1985 more than offset declines in U.S. labor costs. The bottom panel of Table 3-8 shows that foreign unit labor costs (expressed in U.S. dollars) have declined, and this decline offset the effects of low growth in U.S. manufacturing compensation and high growth in U.S. manufacturing productivity. Had the dollar not appreciated against foreign currencies during 1979-1985, U.S. unit labor costs would have declined relative to those of all of the industrial nations in Table 3-8 with the exceptions of West Germany and Japan. Even relative to these nations with higher produc- tivity growth rates, the increase in U.S. unit labor costs would have been far smaller, thus making U.S. exports more competitive in world markets, reducing import penetration of U.S. markets, and reducing the incentives for U.S. firms to locate their production facilities onshore (Finan et al., 1986; Kravis and Lipsey, 19861. Indeed, in the absence of the surge in the foreign exchange rate of the U.S. dollar, U.S. employment growth during the past 7 years might have exhibited a rather different pattern, as productivity gains in U.S. manufacturing supported increases in exports and higher output growth, both of which could have led to growth or slower declines in manufacturing employment.27 2'Data on changes in foreign hourly compensation and labor productivity trends through 1986 are not yet available to bring this comparative analysis of unit labor costs up to the end of 1986. Neef (1986) noted in his study that the decline in the foreign exchange value of the dollar from its peak in 1985 through October 1986 had not yet brought the dollar to its 1980 value vis-a-vis the currencies of Western European nations, although substantial deprecia- tion had occurred against the Japanese yen. Moreover, the dollar's depreciation against many Latin American and East Asian currencies was minimal. The U.S. dollar also had not depreciated against the Canadian dollar, which accounted for 25 percent of U.S. manufac- tured exports. The declines in the foreign exchange value of the dollar since October 1986 are likely to reduce some but not all of the disparities that developed during 1980-1985 in U.S. and foreign nations' unit labor costs.
LABOR SUPPL Y AND DEMAND 85 SUMMARY This chapter has examined the determinants of labor supply and demand by focusing on the role technology plays in influencing aggregate employment and unemployment levels. The increase in aggregate unem- ployment since 1973, as well as the large number of experienced workers suffering permanent job losses in recent years, is disturbing. The direct contribution of technological change to these trends appears to be minor. As the growth rate of the labor supply declines during the next decade, at least one source of upward pressure on aggregate unemployment should diminish. Differential rates of productivity growth in manufacturing and nonmanufacturing industry, combined with low rates of growth in the output of manufacturing industry, have contributed to higher rates of decline in manufacturing's share of total. employment during 1979-1985. The decline in manufacturing's share of nonfarm employment does not represent a departure from longstanding patterns of economic growth and development in the United States, but the rate of decline has accelerated during the past 15 years. . . Changes in the international economic environment during 1980-1985 have exacerbated and accelerated the reductions in manufacturing's share of total employment. The declines in the foreign exchange value of the dollar since late 1985 should improve the competitiveness of U.S. industry vis-a-vis a number of foreign competitors. Nonetheless, relying solely on this policy option to restore U.S. competitiveness will require severe (and in the view of this panel, unacceptable) declines in the purchasing power of U.S. workers and consumers. Technological change and productivity growth remain indispensable to the improvement of U.S. industrial competitiveness and real wages. In view of the fact that U.S. involvement in the international economy is likely to expand still further in the near future, the rapid generation and adoption of new technologies are essential to preserving and expanding U.S. employment and living standards during the next two decades.