Recent discussions of the future of work and employment have assumed a coming "labor shortage," as well as a "skills mismatch" between the characteristics of American workers and the needs of employers. This view has been widely propagated by the Department of Labor under Presidents Reagan and Bush. As Elizabeth Dole put it while she was Secretary of Labor, "America faces a work force crisis" because there are a diminishing number of people eligible and qualified "for the ever-increasing complexity of jobs in our economy."
This view draws heavily on the widely disseminated Workforce 2000 report, prepared by the Hudson Institute for the Department of Labor. At bottom, here is the thesis: Rapid growth of high-skilled professional and technical occupations and the service sector is said to be increasing the demand for workers of higher skill. But because of slow labor force growth and the dominance of minorities and women among labor force entrants, the pool of new workers is shrinking, and their skill and education are declining. America will thus face a mismatch between available jobs and available workers.
But this clouded picture has a convenient silver lining. With a modest amount of leadership and very little, if any, new funding from government, we are assured the economy is on a high-wage, high-skill growth path. Moreover, the analysis essentially blames students, schools, and workers for our problems and fails to examine business practices. It thus neatly minimizes the need for more active government support for upgrading jobs and pay in the economy.
Recently, the Department of Labor released another report in the same vein, titled "What Work Requires of Schools." This latest report identifies basic competencies needed for "high-performance work settings that are relentlessly committed to excellence, product quality and customer satisfaction" that draw on worker ingenuity to "think creatively and solve problems." Astonishingly, however, the same report admits that "nine out of ten employers are operating on yesterday's workplace assumptions," yet offers no guidance on how employers might restructure work to take advantage of newly equipped workers.
In fact, the skills mismatch story, in either version, is largely wrong or misleading: It is wrong in that key points are contradicted by available data, some of it in the same Workforce 2000 report. It is misleading in that key predictions are more wishful thinking than logical extrapolations of economic trends. The consequence is a set of labor market policies that ignores significant problems, such as the need to upgrade pay and skill levels of available jobs, and misdirects the focus of education and training policy.
Far from an explosive growth of job skill requirements, our analysis of BLS employment projections shows that the effect of occupational upgrading on job skills is actually projected to slow down in the future to one-third to one-fourth of its rate in the recent past. In addition, the differences in skill and pay between expanding and shrinking occupations are projected to narrow in the future, rather than widen. While projected changes should slightly increase skill requirements, they will likely decrease pay levels. And non-Hispanic whites, not minorities, will dominate entrants to the labor market in the 1990s, comprising two-thirds of the total.
Because the "labor shortage" view is factually incorrect, it misdirects policy makers away from measures that would affect the way employers structure work. Overstating the anticipated growth of professional and technical jobs also leads to an overemphasis on college education. Even an optimistic view suggests that at most 30 percent of future workers will need a college degree, up from about 25 percent in the mid-1980s. If anything, there will be a glut of college graduates. The bigger challenge is to improve the jobs, pay, and skills of workers with less than a college education. But this will not be accomplished simply by upgrading worker skills and assuming that the jobs will follow.
The labor shortage view ignores our recent problems with wage and income growth. The 1980s saw slow income growth generally and a dramatic fall in real wages for the three-fourths of the workforce that lack a college degree. A young, high-school educated man with less than five years of work experience earned 18 percent less in wages in 1987 than in 1979 and only what a comparable worker earned in 1963. Hourly compensation of nonsupervisory workers (about 80 percent of the wage and salaried workforce) fell 0.6 percent annually between 1979 and 1988. We need a new set of policies to restore wage growth.
Slowly Changing Jobs
The changing skill levels of available jobs can be analyzed in three ways: by occupation (is the economy losing janitors and gaining engineers, or vice versa?); by industry (is the computer sector growing or shrinking?); and within a particular occupation/industry category (is the work of supermarket cashiers or stockbrokers becoming more highly skilled over time?)
In the past, employment shifts toward higher paying occupations have at best barely offset the shift toward lower-paying industries. Projections by the Bureau of Labor Statistics to the year 2000 show an extremely modest, if not inconsequential, positive effect of occupation shifts on hourly compensation, raising compensation by just 0.24 to 0.37 percent over a ten-year period. This positive contribution of the expected shift to higher-paying occupations over ten years is less than half the annual decline of real hourly compensation for most workers in each year of the 1980s. Moreover, the negative consequences of the expected shift to low-paying industries, a reduction of hourly compensation by 1.2 to 1.4 percent over ten years, is three to five times larger than the positive benefits of expected occupational changes.
These results allows us to put the findings of Workforce 2000 in perspective. Their analysis of changes in the job structure focused solely on occupational shifts and concluded that "the jobs created between 1987 and 2000 will be substantially different from those in existence today. A number of jobs in the least-skilled job classes will disappear while high-skilled professions will grow rapidly. Overall, the skill mix of the economy will be moving rapidly upscale, with most new jobs demanding more education and higher levels of language, math and reasoning skills."
But this assertion is contradicted by our analysis of BLS employment projections -- the same data series used by Workforce 2000. All of the changes that Workforce 2000 describes raise overall compensation levels by less than 0.5 percent over twelve to fourteen years, hardly a "rapid upscaling." Moreover, the expected occupational changes will have a smaller effect over the next twelve to fourteen years than the effect they have had in the prior thirteen years (roughly 0.4 percent versus 2.5 percent).
Meanwhile, the Workforce 2000 analysis ignores the more consequential shift to lower-paying industries. The net result of the industry and occupational employment shifts will be to reduce pay levels. Our skepticism of a "skills explosion" is further confirmed by examining the pay differences between occupations that are expanding and occupations that are shrinking. BLS projections imply that the occupations to which the work force will be "upscaled" are, in fact, less upscale than in the past.
Do New Jobs Require More Skills?
A widely cited claim from the Workforce 2000 report projects that while only 22 percent of jobs in 1984 (the base year used in the report) required a college degree, 30 percent of the new jobs created between 1984 and 2000 will require one. Moreover, the new jobs in the fastest growing occupations have the highest skill ratings in the language, math, and reasoning areas; on the other hand, jobs in slow-growing occupations have medium skill ratings and those lost in declining occupations have quite low skill ratings. Averaging across skill areas, 41 percent of new jobs will be in the three highest skill categories, compared to only 24 percent in such categories today.
Here the claims are belied by the report's own data. A table in the report's technical appendix reveals that the average level for language skills (General Educational Development-Language or GED-L) in 1984 was 3.06, about the middle of the scale. By the year 2000, the average skill score is projected to rise to only 3.17. This is an increase of just 3.6 percent over the sixteen-year period, hardly a massive upgrading of job skills.
Why is the actual change so modest? First, much less change is occurring in the overall distribution of occupations than most people believe. The total increase in occupations with rising employment share is under six percentage points. In other words, if we think of the U.S. economy as having one hundred jobs, after the end of sixteen years six people will have moved out of the decreasing-share occupations (which are relatively low skill) into the increasing-share occupations (which are relatively high skill) -- a very slow rate of change.
Second, while the most highly skilled occupational groups are indeed generally growing the fastest, these groups are relatively small to begin with. For example, the data in the Workforce 2000 appendix show that the five most highly skilled occupational groups (which includes the three fastest-growing groups) will provide just 10.6 percent of the net new jobs added to the economy between 1984 and 2000. Even by the year 2000, these occupations, which include law, medicine, natural and social science, engineering, and architecture) constitute only 6.1 percent of the job pool.
Finally, the trend toward relatively high-skill jobs has been partially counterbalanced by a shift toward relatively low-skill service occupations, such as cooks, waiters, household workers, janitors, and security guards. This occupational group, by itself, will provide almost one-quarter (23 percent) of the net new jobs added to the economy in the sixteen-year period and constitute 16.8 percent of the overall job pool in the year 2000.
Given all these factors, we felt it was appropriate to test the Workforce 2000 assertions using the Bureau of Labor Statistics occupational projections (1986-2000 and 1988-2000) and a wide range of skill measures. We did two things the Workforce 2000 report did not do: (l)We estimated future change in the overall job structure, not just at the margin; and (2) we compared estimated future change (from 1986 to 2000 and 1988 to 2000) to change in the recent past (the 1973-79 and 1979-86 periods) so future change could be assessed against a historical yardstick. The results are startling: not only is the projected growth in skill requirements unimpressive; growth will be at just one-third to one-quarter of historical rates.
For example, skill levels in handling data went up at a ten-year rate of 4.01 percent between 1973 and 1979 and a rate of 5.07 percent between 1979 and 1986, but are projected to rise in the future at rates only about one-third the 1973-79 rate and just one-quarter the 1979-86 rate. The pattern is similar for the other skills such as General Educational Development (GED) and verbal aptitude. These trends flatly contradict the popular notion that occupational upgrading will produce an explosion in skill requirements.
Do Old Jobs Demand New Skills?
If shifts in the distribution of jobs are not likely to produce significant upgrading of employment, what about changes in tasks required in given jobs? For example, if computers are now employed extensively within an occupation (say, airline reservationists, hotel clerks, or bank tellers), while they were not used at all fifteen years ago, the skills expected in that occupation may have changed dramatically. If widespread, these changes within occupations could affect the whole economy.
One problem is that available data are weak; the last systematic survey on job content was taken in 1977. But the available data and case studies do suggest skepticism about whether there has been or will be large-scale skill upgrading simply because of technological change. This is because employer response to a given technology can vary from large increases in skill levels to none at all; new technologies may even downgrade required skills. One widely cited cross-national study of flexible manufacturing systems shows similar technologies being deployed with dissimilar effects in different countries. Countless occupations, such as bank teller or data entry clerk, work with advanced technology but in a routine way that does not necessarily lead to higher skills and wages.
We can, however, draw some limited conclusions about changes in the content of jobs from available data and case studies. First, more jobs will require at least threshold levels of literacy and numeracy. Second, some jobs in "best practice" firms within certain industries are being substantially upgraded; workers are expected independently to solve technical problems that come up in the course of their work, to learn new tasks on a fairly regular basis, and to interact extensively with fellow workers, frequently as part of a team. But finally, and most importantly, "best practice" firms that are upgrading job skills in this manner are far from typical, though their numbers are certainly increasing.
If this interpretation is accurate, much of the current talk about extensive job upgrading appears to exaggerate the extent of change in most work places. There is a problem with the "skilling" of jobs in our economy, but not the problem that people generally imagine. What we have is not too much upgrading of the job structure, but rather too little. There is great potential for upskilling, but not enough actual upskilling.
The Real Challenge
If employers are not generating a powerful demand for more highly skilled workers, does this mean that there are no problems with the quality of our work force? On the contrary, there are definite problems with the quality of the workforce, but these problems are different from what is usually supposed. The typical labor shortage viewpoint has work force quality suffering because of changing demographic patterns, patterns that are producing less educated and less productive workers.
Workforce 2000 states, "Only 15 percent of the new entrants to the labor force over the next 13 years will be native white males." While a change is indeed occurring, it is neither as drastic as this statistic suggests, nor are its implications what the report takes them to be. Workforce 2000 exaggerates the shift by focusing on net new entrants to the work force -- those who represent a new addition to the overall size of the workforce -- rather than actual entrants or the total work force, both of which are changing far more slowly. Nor should one categorize women as educationally disadvantaged, as Workforce 2000 does. For example, about one-third of actual, rather than net, entrants will be white, non-Hispanic males, while fully two-thirds will be white, non-Hispanic males and females. As a result, the overall composition of the work force will shift only modestly by the year 2000, with the share of white non-Hispanics dropping from 79 to 74 percent, while the share of blacks and Hispanics rises only for 18 to 22 percent. (Most of this increase comes from Hispanics.) Finally, the categorization of women work force entrants as educationally disadvantaged is belied by the fact that young women in the labor force are now more highly educated than men.
If the demographic composition of work force entrants is not a problem, it still would be argued that the extent and quality of schooling received by these entrants has generally declined. In fact, the schooling of our work force entrants has not deteriorated, although it has not significantly improved over the last two decades. This can be seen in trends in both the extent of schooling (years of education, high school completion rates) and quality of schooling (as measured by various test scores) of young workers. Moreover, there is a consensus among economists that the dramatic drop in the real wages of high school graduates in the 1980s was not caused by a decline in school quality since the wages of older high school graduates also declined. This can be seen especially clearly for blacks. We have never had better educated black work force entrants; years of schooling have risen, high school and college completion rates are significantly higher, and test scores -- at the bottom as well as at the top -- have dramatically improved, especially relative to whites. The wage problem facing non-college-educated black and white workers lies in access to high-paying or skilled jobs, not in their academic credentials.
To the extent that there is a work force problem we believe it lies in the inadequacy of American education and training relative to that received by students and workers in other countries. Because of this inadequacy, our economic competitors can and do pursue technological innovation and higher productivity with the benefit of workers with stronger cognitive skills. This is a competitive disadvantage that should be addressed, but it is a problem of the entire work force, not of its changing demographic composition or of the deteriorating schooling of work force entrants.
The Road Forward
Wage deterioration among the vast majority of workers, and young workers in particular, should set the context for how we evaluate labor market trends in the 1990s. The issue is not only whether incomes and wages will grow, but also whether the earnings of many workers will even recover to their levels of the late 1970s. Our own analyses, based on BLS projections, suggest that income and wage trends in the 1990s are unlikely to bring about that kind of recovery.
If we hope to see wages recover their historic growth, we need policies that put the economy on a high-skill and high-wage path. For example, employers face a choice between high-skill and low-skill approaches to work organization; policy should encourage them to select the high-skill approach. One reason to promote ideas such as works councils and worker participation is to make it more difficult for management to rely on low-skill, routinized work organization. We might also consider adopting a uniform training levy -- a 1 percent or higher payroll tax -- that employers would either have to pay to the government or use internally for upgrading work and workers. We must also advance the diffusion of "best practice" methods of production through an expanded set of industrial extension services.
Upgrading workers' living standards is also not likely to happen absent conscious intervention. The key lies in achieving greater wage growth, which is associated with higher productivity. But higher productivity may not be enough, since the slowdown in the growth of wages reflects an overall weakening of workers' bargaining power. Without reforming labor laws and raising the minimum wage back to historic levels, we are unlikely to achieve satisfactory wage growth. The optimal path therefore includes highly trained workers who are also well paid, working at jobs defined to maximize the productive use of their skills.
In sum, a substantial upgrading of skills and wages cannot be achieved only by improving the skills of the most disadvantaged and relying on market forces to do the rest. We need to upgrade the skills of the workforce as a whole, both new entrants and those already at work. And we need strategies to make that a lifelong process. The point of improving workforce skills is not to complete a rendezvous with an improbable explosion of high skill jobs. It is rather to provide a solid base of capable workers who can do the upgraded jobs that our policies should be consciously promoting.