m_topn picture
Atlantic Monthly Sidebar

July 1983

The Declining Middle

With most jobs being created at the top and the bottom of the ladder, America may have difficulty remaining a middle-class society

by Bob Kuttner

There is a good deal of evidence that job opportunities in the United States are polarizing, and that, as a result, the country's future as a middle-class society is in jeopardy. What the decline of the middle class would mean to the country can only be guessed at, but it presumably would be unwelcome to the millions of parents who hope that their children can move up the economic ladder; to American business, which needs a middle class to consume its products; and to everyone who is concerned about fairness and social harmony.

As the economy shifts away from its traditional manufacturing base to high-technology and service industries, the share of jobs providing a middle-class standard of living is shrinking. An industrial economy employs large numbers of relatively well-paid production workers. A service economy, however, employs legions of keypunchers, salesclerks, waiters, secretaries, and cashiers, and the wages for these jobs tend to be comparatively low.

This increasing concentration of jobs at the extremes of the country's range of earnings is an issue quite distinct from high unemployment--the most visible threat facing the work force today. It is also an issue distinct from the country's poor rate of economic growth. Restoring growth may make "everyone" better off, statistically, but growth by itself has little bearing on the distribution of wages. Moreover, the government's existing income-support programs simply are not designed to address this new problem. The fact is that statistically, most dollars of government transfer income go to the elderly, through Social Security and Medicare. Thus, even counting all the welfare checks and food stamps for the poor, and all the dividend coupons, interest payments, and capital gains for the rich, the working-age population still depends on jobs for fully 85 percent of its income.

During the three decades after World War II, real income grew steadily, along with the number of good jobs. Between 1958 and 1968, manufacturing added 4 million jobs; state and local government added another 3.5 million. The child of a working-class family who graduated from high school in the 1950s and 1960s could choose among a number of relatively well-paid jobs--precisely the jobs that are disappearing now. Manufacturing, which pays on average almost three times the minimum wage, added fewer than a million jobs between 1968 and 1978, and has lost nearly 3 million jobs over the past four years. Construction and production work, taken together, account for about one job in eight today, compared with one in four in 1950. According to Barry Bluestone, a professor of economics at Boston College and co-author of the book The Deindustrialization of America, "The pattern of wages in the old, mill based economy looked just like a normal bell curve. It had a few highly paid jobs at the top, a few low-wage jobs at the bottom, and plenty of jobs in the middle. But in the new services economy, the middle is missing."

The erosion of the middle of the labor market is easy to misinterpret, because its roots are multiple. During the 1970s, the entry into the work force of an unprecedented number of women and of young adults born during the baby boom resulted in too many workers for the jobs available, and depressed wages. The decline of the middle also has something to do with the explosive growth in world trade since 1960. As manufacturing technologies have become more mobile, and multinational firms more footloose, production jobs have migrated from the U.S. to countries where wages are low. In addition, technology itself has helped to provoke the shifts in the job market. For example, fewer American workers would have been needed to make steel in 1980 than in 1960 even if the pressures of global competition had not been a factor, because new machines have made many of their tasks redundant. Finally, the high rate of unemployment caused by these trends has tended to drive wages down further, especially at the low end, since it forces unskilled workers to compete for their jobs with unemployed people who are willing to do the work for less.

Although demographic shifts, stepped-up world trade, unemployment, and especially the advance of technology all have had an effect on the shape of the job market, middle-level jobs have been disappearing ultimately as a result of the ways in which technological gains are being distributed. When a machine replaces a production worker, both the firm and consumers as a group benefit. The loss falls mainly on the worker who is displaced. If that loss is generalized to millions of high-paid workers, they suffer as a group, and the economy as a whole suffers a loss of worker purchasing power. Thus the lack of a mechanism to distribute some of the financial gains from technology to the work force comes back to haunt the entire economy.

As it happens, production work is becoming heavily automated at a time when the most powerful agents of distribution--the trade-union movement and the public sector--are in decline. There is nothing intrinsic in assembling cars, mining coal, or pouring molten steel that requires high wages. These jobs do pay well mainly because of the efforts of strong unions. Just as high wages are not a necessary characteristic of factory work, low wages are not a necessary characteristic of a service economy. But because unions have organized few fast-food workers, typists, bank clerks, or even computer assemblers, most of the new service jobs that are replacing factory work pay far lower wages. And the chief realm of a service economy in which wages are good--public service (teaching, social work, nursing, and so forth)--can no longer be relied on to take up the slack. Although the number of jobs funded by state and local government grew substantially throughout the 1960s and 1970s, it peaked in 1980, and had fallen by 200,000 at the end of 1981--the most recent year for which figures are available. Whether or not one favors strong unions or the pervasive presence of government in society, unions and government undeniably have operated as instruments of a relatively egalitarian distribution of wages. As their influence declines, so does their impact on the national pattern of earnings.


Labor statistics show that wage and salary income is becoming more unequally distributed in the economy as a whole, and in service industries in particular. It also turns out that the fastest-growing industries are those with the most extreme wage spread.

The Bureau of Labor Statistics ranks earnings in ascending order by percentile. Thus, a worker at the eightieth percentile makes more money than 79.9 percent of the rest of the population. A study conducted in 1980 by two economists at the department of Labor, Peter Henle and Paul Ryscavage, found that between 1958 and 1977, the earnings of men at the twentieth percentile increased by 130.6 percent (with no adjustment for inflation), but the earnings of men at the eightieth percentile increased by 206.7 percent. The book Services: The New Economy, by Thomas M. Stanback, Jr., and three colleagues, published in 1981, concludes, "For services as a whole, the most important observation is that there tends to be a heavy concentration of employment in better-than-average and in poorer-than-average jobs. In contrast, in manufacturing and construction, the distributions are more heavily weighted toward medium and above average jobs."

Statisticians at the Department of Labor have been comparing the salaries paid in the twenty fastest-growing occupations with those in the twenty that are in the sharpest decline. Although their results will not be released in full until later this year, Ronald Kutscher, an assistant commissioner of the Bureau of Labor Statistics, says that workers in the fastest-growing industries are earning, on average, $5,000 a year LESS than workers in industries that are declining or that are growing most slowly.

There is, of course, independent evidence of widening inequality in America, apart from changes in the distribution of jobs. The number of Americans living in poverty, which dropped to about 23 million in 1974, is greater than 30 million today. The 20 percent of households in the country's highest tax brackets take 41.9 percent of all income paid in the U.S.--the highest share in two decades. Paradoxically, the entry of more women into the labor force has exaggerated the inequality of family incomes. On average, the working wife of a man who makes $50,000 earns $5,000 more than the working wife of a man who makes $10,000. Homeownership, for years the badge of membership in the middle class, has become a nearly impossible feat for those who do not have a house already. A new house of average price is beyond the budgets of more than 75 percent of all households.

One might fault the Reagan Administration for cutting taxes at the top and social programs at the bottom, but the widening polarization of the labor force remains a phenomenon distinct from these policies and largely ignored in the national debate. Long after the Reagan Administration ends, this problem will plague the American economy.

In 1982, the Department of Labor ranked the number of actual job openings in 1980 by category. These openings reflect turnover as well as net job growth. The top fifteen job categories, with a single exception, are ones that middle-class parents hope their children will avoid:

JOB OPENINGS, 1980
Retail salesclerks757,750
Managers and administrators (not elsewhere classified) 711,793
Cashiers 617,973
Secretaries (not elsewhere classified) 599,216
Waiters and waitresses 465,628
Cooks (except private household) 437,341
Stockhandlers 358,393
Janitors and sextons 333,309
Bookkeepers 304,789
Miscellaneous clerical workers 299,940
Nursing aides and orderlies 284,332
Child-care workers, private household 277,525
Building-interior cleaners 259,528
Typists 250,276
Truck drivers 245,377

Of the top fifteen, only one, the grab-bag category "Managers and administrators (not elsewhere classified)," suggests a well-paid line of work. The first profession to appear on the department's roster is job number thirty-six, accountants, with 140,108 openings. Far down on the list come computer programmers (30,963) and electrical engineers (23,089). According to the most recent comprehensive forecast of the Bureau of Labor Statistics, released in 1982, the economy will generate some 19 million new jobs between 1980 and 1990, about 3.5 million of which will be "professional and technical." Low-wage service and clerical work will account for almost 7 million new jobs.

Within the manufacturing sector, the job structure of the one major growth industry--microelectronics--is highly stratified. Computer scientists and engineers typically have more than a college degree and earn $30,000 a year and more. The workers who assemble computers, however, earn about $12,000 a year. The number of skilled technicians whose wages fall between these two extremes is surprisingly small. The rate of productivity of the electronics industry is accelerating so fast, in fact, that electronics is not expected to be an important source of new production jobs, good or bad. The more significant effect of the computer on work is being realized not in the computer's manufacture but in its application. The Bureau of Labor Statistics projects that the U.S. will require about 120,000 more computer programmers and 125,000 more electrical engineers during the next decade, but it will require some 3 million more secretaries and office clerks.

By 1990, 600,000 more janitors, 500,000 more salesclerks, and 400,000 more fast-food workers will also be needed. To be sure, the economy has always included low-wage service jobs, but in the past it also included a larger middle range of relatively well-paid production jobs. These were the jobs that created the postwar, blue-collar middle class. Today, a welding job in an automobile plant is performed more cheaply by a robot or a foreigner; that job once paid somebody $12 an hour, and there are few openings at $12 an hour in restaurants, hotels, or computer-assembly plants.

Given the 10 percent unemployment rate, it is tempting to conclude that lots of low-wage, "entry-level" service jobs are just what the economy needs. Unfortunately, although these jobs do provide an entry into the work force, few offer opportunities for advancement. The fast-food industry, for example, employs a small number of executives, and hundreds of thousands of cashiers and kitchen help who make about $3.50 an hour. With some variation, keypunchers, chambermaids, and retail sales personnel confront the same short job ladder.


Statistics do not do justice to the story. A look at actual offices, stores, and factories is even more revealing. The Boston area, with its heavy concentration of financial services (banking, credit reporting, real estate, and so forth) and high-technology firms, can serve as a model of where the economy as a whole is headed. To get a practical glimpse of the emerging labor market, I interviewed personnel executives at several companies there.

The work force of financial-service industries grew 32 percent, to 5.4 million, between 1973 and 1982--a period when durable manufacturing lost workers. At Blue Cross Blue Shield, which falls within this sector of the economy, the work force grew 46 percent, to 88,000, between 1970 and 1982. As in most insurance companies, the main activity of Blue Cross Blue Shield is collecting premiums and paying claims. Computerization and remote data transmission have made these tasks more routine and have allowed them to be performed almost anywhere.

The company's offices in Massachusetts employ about 4,000 people, of whom some 400 are rated managers, another 1,000 are rated professionals (attorneys, actuaries, programmers, doctors, nurses), and some 2,600 are rated clericals. According to Theron R. Bradley, who oversees employee training for Blue Cross Blue Shield of Massachusetts, computerization has not had the same effect on everyone who works for the company. For professionals like him, it has provided new possibilities. Bradley says, "As the manager of training, if I wonder who in the company has taken which courses, I can have the computer research that at the touch of a button. If I need to analyze the results of testing, I can have the computer produce research that I could not have imagined allocating the resources for a short time ago. Our accountants can now spend their time doing what they were trained to do--setting up accounting systems and interpreting the results. In the past, a lot of their work was glorified bookkeeping." At the same time, however, the computer has had a different effect on Blue Cross Blue Shield's large clerical force. Bradley says, "In the past, the lowest six levels of our clerical structure looked like an inverted pyramid. At the very bottom were jobs like filing, sorting, mail delivery, and basic data-entry jobs. You'd almost have to try not to be promoted in order to avoid rising through the clerical ranks. Most people did get promoted. With good oral and written communication skills, you could move fairly easily into a supervisory or junior management job. Those senior clerical jobs are disappearing, because the computer is taking over so much of the decision-making. We have a lot more routine jobs at the bottom, and a few more complex jobs at the top, where you need someone to analyze hundreds of pages of data."

Blue Cross Blue Shield's largest job category is that of claims examiner. The company recently opened "outreach" offices in the towns of Plymouth and Rockland, outside Boston, to house the newly organized Blue Shield claims department. The work is part-time, in order to attract mothers of young children, students, and others who will accept reduced hours. As it turns out, all of the claims examiners on the day shift at Plymouth are women. The day shift runs from eight until two. "You can drop Junior off at school, and drop down to our office at the local shopping mall," Bradley explains. Each clerk sits at a computer screen with a pile of claims, and types in the subscriber's name or number, the code number of the medical procedure claimed, the fee, and the identification number of the doctor. The computer verifies that the subscriber is covered. A different department handles more complicated claims, and still another mails out reimbursement checks.

A claim takes about a minute to process, and a good clerk can dispatch 300 a day; the computer keeps track of productivity. On average, the clerks are paid between $4.50 and $5.75 an hour, but wages are adjusted every two months to reflect each clerk's output. Blue Shield provides a fifteen-minute paid lunch break and a week of vacation after one year. It provides no health insurance.

When the company announced its plans to move to Plymouth, a Blue Shield Day was held at a local vocational high school to recruit applicants. Bradley says. "We locate these offices in areas of high labor demand so that we don't have to outbid everybody else's wages. We have no trouble finding people. Every time we open one, we are swamped with applications."

According to one of the clerks at Plymouth, who requested anonymity, few of the women who began work with her at the office when it opened, in March of 1982, are there today. The most demoralizing thing about the job, she says, is the isolation. Virtually all communication is between the clerk and the machine. Since pay is based on output and the job itself requires no human interaction, there is every incentive to keep social conversation to a minimum. "Last Christmas," my informant says, "they organized a little office party. They had a 'Christmas Grab,' where everybody picks a name out of a hat. You bring in a little gift for somebody and they bring one in for you. We realized that nobody knew anyone else's name."

The new system has indeed improved the productivity of the company's claims department, but more intensive use of technology and higher output have not translated into better jobs. Despite the space-age technology, each clerk can be trained in only four weeks. The work is more routine, and the company's labor costs are lower.


It is intriguing to consider whether the loss of middle-income jobs reflects technological imperatives or merely management preferences. Much has been written on the tendency of technology to degrade work, but there are enough instances to the contrary to suggest that managers have substantial latitude in how they choose to reorganize work.

Banks provide examples of technology both making work more tedious and making work more challenging. Like armies, banks divide their troops into officers and enlisted personnel. Behind every officer is a cadre of back room clerks. Although in theory the lowliest teller can aspire to be an officer, in practice there is little opportunity to move up from the clerical force into the officer ranks. For example, First National Bank of Boston, the nation's seventeenth largest bank, hires about a hundred people a year for its loan-officer-development program. As many as ten may be promoted from within. Most new loan officers, however, are recent graduates of colleges or business schools. The competition is keen, and the bank will pay more than $30,000 to attract a promising MBA holder. Clerical workers at First National outnumber officers by more than two to one. The 621 tellers earn an average salary of $6.69 an hour, or about $263 a week, and below that rank there are thousands of other, lower-paid clerical jobs.

One job that has become more routine as a result of technology is that of money-transfer clerk. Like most banks, First National has automated its money-transfer operation, plugging into a computer-connected network. When a customer requests that money be transferred from one bank to another, the clerk punches up a standard format on the computer terminal, checks whether funds are available, enters the number of the bank to receive the funds, and electronically consummates the transaction. According to Steven Bavaria, the bank's vice president in charge of personnel, "We are able to use more junior people for transfers since that was automated. We even use some part-time high school students. Because we build more safeguards into the system, the people operating it can concentrate all their attention on five or six variables: the account number, the amount of money being transferred, the name of the bank, and so on, rather than having to type and proofread an entire Telex message."

In contrast, the computerization of First National's letter-of-credit department has upgraded the status of several dozen employees. A letter of credit is a document from the bank to a foreign supplier's bank, guaranteeing payment for goods that are about to be shipped. Formerly, processing a letter of credit involved assembling pieces of paper, with a chain of clerks verifying shipping and credit documents and typing forms. Now the bank's letter-of-credit department is mostly what Bavaria terms "junior professionals," each with a computer terminal. "The same person can review the documents. create the file, send the letter, and, when we have verification that the order has been shipped, authorize payment." Bavaria says. Most letter-of-credit employees earn between $300 and $400 a week, and the position is sometimes a way station in the officer training program.

Gross statistics on the earnings of clerical workers nationwide suggest that this kind of upgrading is unusual. The vast majority of clerks, men and women, earn less than $300 a week. For full-time female clerical workers, the average weekly salary in 1980 was about $220. Of the total 17 million female clerical workers, fewer than 500,000 earned as much as $400 a week. Fewer than a million earned more than $335. The study by Henle and Ryscavage indicates that clerical wages are becoming increasingly skewed to the low end of pay scales.

Retail sales is another growth industry in which wages are clustering more and more at the extremes. The number of people working in the Department of Labor's "wholesale and retail trade" category has grown from 8.7 million in 1950 to more than 18 million today. In 1981, nonsupervisory workers in trade and retailing earned an average of about $150 a week. The 2.4 million salesclerks earned an average of $178 a week, and the 410,000 manufacturers' sales representatives earned an average of $434.

In theory, the shift to an economy based on services and information ought to be compatible with the equitable distribution of income and jobs. Substituting technology for human labor makes the country as a whole richer. Some economists argue that as the economy grows more productive, technology will gradually eliminate many routine, dead-end jobs, replacing them with high-skill positions and more leisure time for all. The market does not necessarily convert rising productivity into an acceptable distribution of income and leisure, however. The pattern of development in much of the Third World demonstrates that it is quite possible for an economy to attain a high rate of growth based on rapid technological progress, and still function as a society in which one half of the work force waits on the other half, and makes a good deal less. Low wages tend to foster more low wages. If you have to pay a welder $12 an hour, it may make economic sense to substitute capital for labor and build an automated welding machine. But if fast-food workers cost only $3.50 an hour, automation may be a long time coming. It is often the high-paying jobs that are replaced by machines, not the low-paying ones.


Few economists expect the booming high-technology field to be an important source of new jobs: its productivity is too high. According to projections by Data Resources, Inc., a corporate consulting firm, the entire high-tech industry, broadly defined, will provide fewer than a million new jobs over the next decade. The American Electronics Association projects that the number of jobs for skilled electronics technicians will grow by only about 15,000 a year. Moreover, the relatively low-skill assembly jobs in microelectronics plants, which do not pay well to begin with, are rapidly being automated or shifted to countries in which wage scales are even lower.

The Hewlett-Packard Company--a California-based manufacturer of computers, computer software, scientific instruments, and testing devices, with assets totaling about $4.3 billion--offers a good example of technology's effect on the high-tech labor force. The company prides itself on "the Hewlett-Packard way"--an informal style of management with few barriers between the executive suite and the shop floor.

The medical-products group, one of five major corporate divisions, makes an array of sophisticated electronic monitoring devices used by hospitals to analyze blood chemistry, to monitor fetal functions, and to check heartbeat, blood pressure, and oxygen intake. It employs some 2,700 people in two plants near Boston. During the past two years, the number of engineers and computer scientists at the company has grown by about 1,250 nationwide, but few technicians or assembly workers have been hired. Despite the group's steady growth in sales, averaging between 15 and 20 percent a year, productivity has also been rising so rapidly that the company has barely avoided layoffs.

Hewlett-Packard's large plant in the suburban town of Waltham occupies a long, low, brick building. The shop floor is quiet, clean, and well lit. There is no assembly line. In the main section of the floor, employees sit at benches attaching the final few components to printed circuit boards that have been produced mainly by machines elsewhere in the plant. James Phelps, the group personnel manager, says, "For years, loading circuit boards was like building the pyramids. You needed armies of human labor. All the components were hand-inserted. Today you can design a circuit board by computer, and the same computer program will load the machine that produces the board, put the components into the right sequence, and produce most of the board."

At the end of the long shop floor is the sequencing machine--about fifteen feet long and five feet high. In bins above the machine are supplies of dozens of different components. Following a computer tape, the machine loads the components in proper sequence onto what look like ammunition belts for a Gatling gun. The belts then feed the circuit-board-inserting machine, and out come nearly finished circuit boards, with capacitors, resistors, and diodes all in place. Only a few parts remain to be inserted by hand. One worker tends the sequencing machine. She was trained for the job in a few days. A second semi-skilled worker tends the inserter.

Automation is evident, and equally impressive, elsewhere in the plant. A printed circuit board begins with a layer of inert fiberglass sandwiched between two sheets of copper plate. Holes are punched into the blank board to receive the components. In the "printing" of circuits onto the board, an ink screening process is used to indicate the tracks that will eventually be etched into the board. In the past few years, machines have been introduced that plate the copper, punch the holes, and even do the screening. In every case, a few machine tenders now do the work of several production workers, and they need not be highly skilled technicians. Workers were trained to use the automatic screening machine in a single day; there was no upgrading of status, job description, or pay.

According to James Phelps, "We've peaked in terms of our need for the classical technician. The machinery is more and more self-diagnosing. There's less of a judgment factor. The computer spits out an instruction to change such-and-such a component, and you do it. We can upgrade a $5-an hour assembler to a $6- or $7-an-hour junior technician with little advanced training. A few years ago, we were recruiting skilled technicians as aggressively as we're recruiting engineers today. We're not doing it now. Organizations that train technicians should realize that they're preparing people for a smaller and smaller marketplace."


A few miles away from Hewlett-Packard, the Gen-Rad Corporation makes electronic testing instruments, which are used in quality-control operations throughout the electronics industry. According to Richard Cambria, Gen-Rad's vice president for personnel, the instruments are having the same effect on Gen-Rad's production force that they are having on Gen-Rad's customers. There is a reduced need for the traditional job of "test technician," because most testing has become integral to production. Cambria says, "It used to be that the assemblers would assemble, and the inspectors would occasionally throw the work back at them. With the quality control built right into the manufacturing process, we've been able to get rid of an entire layer of inspectors,"

In principle, the proliferation of computers will require skilled repair people to maintain them. The Bureau of Labor Statistics projects about 130,000 new repair jobs for all types of office machinery over the course of this decade. But according to Michael Odom, a human-resources executive with Digital Equipment Corporation, the nation's second-largest computer maker, computer repair is likely to be a field of short-lived growth. Many of the new minicomputers on the market are self-diagnosing. If a circuit malfunctions, the screen will literally tell the user which panel to pull out. The circuit boards slide out like trays in a baker's rack. Digital gives its customers a supply of yellow mailing envelopes, with instructions to send the defective board back to the company for a replacement.

"Things that break are usually things that move," Odom says. "The circuit board is being condensed into a chip. You don't repair a chip. The old personal computers had 100 chips; the new ones have nine. Pretty soon, computers will have no moving parts other than an on-off button. When the whole operation is on a few chips, the greatest likelihood of malfunction is probably random failure cause when a circuit is struck by an extraterrestrial gamma ray.

"When the systems were more complex, we needed more sophisticated technicians to maintain them," Odom adds. "As the computer becomes a mass-market commodity--like a stereo--sales, distribution, and repair are all becoming routinized and simplified."

Odom believes passionately in educating people to work with computers. He is currently on loan from Digital to the Boston city school system to help develop a computer curriculum. But he scoffs at those who expect high technology to fill the gaps in the middle of the labor market. "It's just wrong to hold out hope for high-tech as a mass provider of jobs," he says. "Increasingly, it's an elite industry of professionals. The job growth is in the industries that use our technology, and I worry about what kind of jobs those will be and who will get them."

Within manufacturing generally, automation will affect the middle of the labor market in ways that are not always easy to predict. One important innovation just beginning to be applied is "computer aided design/computer aided manufacturing," or CAD/CAM. Using a computer to make and modify designs, based on an engineer's specifications, increases the productivity of a draftsman many times. The same computer that produces the blueprint can instruct the machinery to build the part. A number of junior colleges have introduced CAD/CAM programs to retrain and upgrade a whole generation of draftsmen. But, according to Lon Bonczek, a corporate vice president of Computervision, a leading CAD/CAM manufacturer, just over the horizon is something called CIM, or "computer-integrated manufacturing." The job of CAD/CAM technician, which is regarded as a hot new opportunity, will diminish, because the engineer himself will sit at the computer terminal and transform his specifications directly into a product.

The economy, of course, will still require technicians to perform traditional jobs. During the 1980s, about 200,000 more auto mechanics will be needed. But contrary to expectations of a smaller but more highly skilled blue-collar work force serving high technology, there is little evidence on the shop floors or in federal statistics of a boom in demand for skilled technicians. The Bureau of Labor Statistics projects that by 1990 about 1.5 million new technical jobs will be created, accounting for 8 percent of the total job growth. But these "new technical jobs" include dental hygienists and surveyors as well as electronics technicians.


If the changing shape of the nation's economy indeed signals a change for the worse in the distribution of jobs, there has been surprisingly little fruitful discussion of what is to be done. Public-policy debate focuses primarily on the gross unemployment rate, and on the need for government to create jobs or to subsidize job training. The quality of work, aside from industrial health-and-safety hazards, and the distribution of job opportunities are not visible public issues.

Some economists believe that demographic shifts are responsible for the polarization of the work force, and that demographic shifts will eventually restore equilibrium. In this view, the explanation for both the depression of average wages and the widening gap between high and low wages is the disproportionate number of young people and of women who entered the work force during the 1970s. Young workers characteristically earn less than older ones, because of their inexperience, and women earn less than men, because of discrimination, and because of the poor salary histories that past discrimination has engendered. So it is often argued that as the work force ages and discrimination recedes, wages will fall closer to the center. Also, some economists say that as the baby-boom generation is absorbed into the work force, labor supply and demand will again balance out and employers will have to pay more to attract workers.

The trouble with this theory is that it fails to acknowledge the changing range of occupations, as well as the opportunity to rise within them. The demand for workers is increasingly skewed to the extremes as a result of technology--a pressure independent of supply. And as the youthful work force ages and more women seek to rise from clerical status, the competition will become even more intense for the scarce good jobs. Demographic shifts are likely to aggravate the competition for at least a generation. In 1975, about 16.7 million workers were between thirty-five and forty-five years old--the prime earning period. In 1995, almost 34 million people will fall into this age group. Twice as many middle-aged people will be competing in 1995 for the limited supply of bread winning jobs. A recent study by the economists Martin Dooley and Peter Gottschalk, who teach at McMaster University and Bowdoin College, respectively, considered whether or not demographics alone could account for the inequality in wages during the 1970s. Even after correcting for the age, sex, size, and education of the work force, Dooley and Gottschalk found that wages remained unaccountably skewed.

Others argue that as the productivity of the economy continues to rise, wages must logically rise as well. Isabel Sawhill, the former director of the National Commission for Employment Policy, says, "In the long run, the cure for low wages is increasing productivity."

This cure does not treat the distributional malady, however. It is entirely possible for the economy as a whole to grow more productive while wages and occupations continue to concentrate at the extremes. At the Gen-Rad Corporation, company revenues per worker rose from $59,000 in 1981 to $72,000 in 1982 to an estimated $83,000 in 1983. Nevertheless, the wages of assembly workers have not grown nearly that fast, and the company is having no trouble filling those jobs.

The fragmentation of the economy into dissimilar realms has been debated for years by economists. John Stuart Mill, in his work Principles of Political Economy, published in 1848, noted the fact that a common laborer could not reasonably aspire to be an artisan, and analyzed the barriers to free competition among workers for available jobs. Several generations of American students of trade unionism, beginning with the institutional economist John R. Commons in the 1920s, have pursued this line of inquiry. In the early 1970s, the labor economists Peter Doeringer and Michael Piore began to write about dual labor markets. The "primary" labor market, they found, is typically union-organized or professional, with relatively good wages and benefits, fairly predictable careers, and job security. The "secondary" labor market, in contrast, is made up of less desirable, low-wage jobs. Its defining characteristic is instability, with high turnover and a large proportion of part-time or seasonal work. Piore, Doeringer, and other labor-market theorists also point out the predominance of white males in "primary" jobs, and argue that the dual system both reflects and reinforces discrimination. A surprisingly large number of primary jobs, they note, do not have high formal entry requirements; skills are learned instead on the job. Through informal referral networks, white men are steered into the better jobs while blacks and women end up with the inferior ones.

Economists of the rival "human-capital" school have responded to dual labor-market theorists by arguing that if such a schism exists at all, it reflects mainly differences of skill and education. They say that undoubtedly some racial and sexual discrimination persists, and is to be deplored, but if workers are only marginally literate, with little training and sloppy work habits, it is hardly surprising that they should find only marginal jobs. Indeed, they say, it is fortunate that society provides jobs for such workers at all. According to the human-capital theorists, the obvious way to upgrade jobs is to upgrade the qualifications of the work force. This approach has informed most federal job-training efforts, including the Manpower Development and Training Act of 1962, CETA, and, more recently, President Reagan's retraining initiative.

A favorite counter-example by the dual-labor school is the mobilization of factories during World War II. All kinds of people who had never seen the inside of a plant learned a variety of skills on the job. But jobs were available then. If a demand for workers today is increasingly shifting toward professional jobs at one extreme, and low-wage, routine jobs at the other, it is no solution to upgrade the qualifications of workers. Although nobody should oppose the movement of basic education for its own sake, education will have little impact on the kinds of jobs available.

Ronald Kutscher, of the Bureau of Labor Statistics, reports that true shortages of qualified workers exist only in a few engineering specialties and in the highest-skilled blue-collar occupations, such as journeyman tool-and-die maker. Since the real problem is the supply of good jobs rather than the supply of good workers, an emphasis on education and training will make the work force even more frustrated than it is now. According to Isabel Sawhill, during the 1970s one college graduate in five was forced to take a job that did not require a college degree.


If retraining and better basic education are not by themselves solutions to the polarization of the supply of jobs, society is not entirely without remedies. It can intervene in several ways to influence the mix of jobs, to maximize career possibilities within the workplace, and, if nothing else, to assure that routine jobs at least pay decently. Most such measures, however, are either unpopular or relatively untested.

The simplest way to promote relatively equal wage scales is to encourage trade unionism. According to James Medoff, a professor of economics at Harvard who specials in labor markets, unionism by itself tends to increase wages by between 10 and 20 percent. In retail trade, for example, workers in unionized stores earn an average of $179 a week, according to the Bureau of Labor Statistics. Non-union retail workers earn only $111 a week. Unionism is a force for wage equality for several distinct reasons. First, unions understandably seek to organize workers from the bottom up, both for reasons of solidarity and to assure that workers whose wages are low do not underbid on members. Second, unions tend to negotiate standard wage rates across firms and regions. Unions are also the principal lobby for a minimum wage, which tends to improve the distribution of wages generally. It is interesting to note that the minimum wage, which was once 54 percent of the average wage, has fallen to a record low of 39 percent--another reflection of widening inequality, and of the declining power of unions.

When unions represent most workers, they undoubtedly improve the overall distribution of wages. When they represent only about one worker in five--as in the United States today, where union membership is at a forty-seven-year low--they tend to reinforce the trend toward a dual labor market, because their members enjoy a degree of market power to bargain for cost-of-living and productivity adjustments that other workers lack. In countries with strong unions the labor movement acts as a prime force for what is sometimes called a solidarity wage policy--the deliberate narrowing of the range in wage and salary income. In Sweden, where both the trade-union federation and the Social Democratic Party have recognized the limits of taxation and income transfer as instruments of redistribution, wage equalization is seen as a more efficient version of the welfare state. If wages are more nearly equal to begin with, there is less need for social programs and high taxes.

Unions can be a force for egalitarian labor markets insofar as they are the principal constituency for clear career ladders, apprenticeship programs, and on-the-job retraining. Although some unions, notably those for the building trades, have acted to restrict opportunities for apprenticeship, others have sought to expand them. In the tool-and-die industry, for example, where it takes as long as four years to train a skilled journeyman, many small shops are reluctant to invest that much effort in a worker who is free to leave for a better-paid job. The International Association of Machinists has long lobbied--unsuccessfully--for publicly subsidized apprenticeship training. The United Auto Workers recently bargained for a joint program with Ford, which allows both active and laid-off workers to train for new jobs, partly at company expense. In health care, which is fragmenting into a number of paraprofessional specialties, the Service Employees International Union has fought a rearguard action to allow orderlies and practical nurses to train on the job for better positions, rather than letting hospitals recruit all their technicians from technical schools and their registered nurses from four year colleges.

Unions can also influence the way technology is applied. Officials of the United Food and Commercial Workers (UFCW) report that the new optical scanners are popular among unionized check-out clerks. A clerk has more time to chat with customers, while the scanner rings up the groceries. The scanners obviously help management improve inventory control. In bargaining over the introduction of the new technology, however, the union made sure that it would not be used to tie wages or promotions to output. This contrasts sharply with the policy of Massachusetts Blue Cross Blue Shield, whose claims examiners are not represented by a union and whose automated equipment accomplishes precisely what the UFCW sought to avoid.

Although American unions are in decline, and therefore lack a pervasive influence on how management makes use of technology, management's own interests are sometimes served by the upgrading of jobs as a result of technology. While many Marxist economists, notably the late Harry Braverman, have contended that the "de-skilling" of the labor force is an inevitable consequence of advanced capitalism, the contrast between the status of the Blue Cross Blue Shield claims clerks and that of First National Bank's letter-of-credit clerks suggests that management has some latitude in the way it applies technology to the tasks at hand. William Spring, a White House adviser to President Jimmy Carter on employment policy, believes that few managers pay close attention to internal labor markets. While most corporations are nominally committed to a policy of promotion from within, maintaining a broad middle sector of the work force is rarely a corporate goal for its own sake. And the wish to cut short-run costs seems to encourage just the sort of de-skilling that the Marxists denounce.


The other unfashionable remedy for the polarization of work is the public sector itself. Tom Wolfe, in his essay "Mau-Mauing the Flak-Catchers," which described an anti-poverty office in San Francisco, noted a truth about public-service agencies that poor people themselves had perceived immediately. The real value of the war on poverty lay not in the programs offered by the poverty agencies but in the opportunity of holding jobs in the poverty agencies. The public sector has been the main avenue of upward mobility for blacks and women. In 1970, according to the economist Martin Carnoy, of Stanford University, 57 percent of all black male college graduates, 79 percent of all black female college graduates, and 62 percent of all white female college graduates held public-sector jobs.

As more and more of society's material goods are produced by fewer and fewer workers, the kind of society that results is ultimately a political choice. We can permit the owners of wealth to spend most of their earnings privately--and the service jobs will include salesclerks, waiters, and open-heart surgeons--or we can tax away some of those earnings and let the service jobs include more teachers and social workers. The choice exists within a realm of public policy that is the least charted of all: the means for redistributing the fruits of productivity and automation.

Harry Bridges, who was the president of the International Longshoremen's and Warehousemen's Union from 1937 to 1977, was among the first labor leaders to bargain explicitly with management on the consequences of automation. Beginning in 1960, as the West Coast docks were automated, Bridges negotiated the following arrangement: The shipping companies would be free to automate, but his members would not be laid off. The wages of longshoremen holding "A" union cards would rise with productivity, and the number of longshoremen would gradually be reduced by attrition. Bridges liked to say, "At this rate, by the year 2000 there will be one longshoreman left on the docks. But he's going to be the best paid son of a bitch in the United States." As Bridges recognized, the deal was fine for union members holding "A" cards, but not so good for the work force as a whole. To generalize from Bridges's quip to an only slightly more absurd extreme, what would happen if all physical goods were produced by robots? The one worker who flipped the switch would boast astronomically high productivity. What should he be paid? And what would everyone else do for a living?

Twenty years ago, many futurists predicted that society would soon become so productive that the main problem would be what to do with leisure time. It hasn't quite worked out that way. The productivity of manufacturing is rising, but since 1973, average household income, adjusted for inflation, has declined about 11 percent, and average wages have dropped 16 percent. In more than 50 percent of families with children, both parents work. Almost half of all mothers of children under six years of age work. Far from the ideal of part-time jobs, shared child rearing, and increased leisure, most parents in two-income families are instead working full-time, juggling babysitters, and wondering why they have no time even to watch television.

Among the various possible remedies, the Luddite answer is surely the wrong one. It makes no sense to preserve manual jobs doing hot, dirty work, when machines would liberate workers to do something else. But, still, the question is what? Other countries, notably in Western Europe, have made an effort to redistribute productivity by lowering the retirement age, shortening the work week, increasing vacation time, and promoting part-time work. Although none of these experiments has yet had impressive results--the European economy continues to generate more products with fewer workers, and to create more low- than high-wage jobs--the obstacles to a radical redistribution of technology's benefits seem to be as much political as they are economic.

The industrial West as a whole is coming under heavy pressure, both from technology and from trade, to lower its wages. If machines are available to replace most human production jobs, and if unemployed women in the Orient are eager to take the remaining jobs producing for Western markets at ninety cents an hour, then the postwar social contract that developed in Western Europe and North America is in severe jeopardy. That social contract included high wages, job security, and a costly welfare state. It enabled the West to defy Karl Marx's prediction that under capitalism the labor force could only become worse off. Now, paradoxically, one hears that in order to maintain our standard of living we must first lower it: to remain competitive, the West must jettison the very social baggage that has assured a comfortable passage. This conclusion somehow defies common sense. From the viewpoint of the entrepreneur, wages are an unfortunate cost of production. From the perspective of the entire economy, however, they are the stuff of social stability.

A distinguished company of essentially conservative political economists--Charles E. Lindblom, Karl Polanyi, Joseph Schumpeter, and even Adam Smith--have recognized what Lindblom was eventually to call "the limited competence of markets." Free markets are superb at energizing human creativity and innovation. They incur painful costs of transition, however, which are often imposed unfairly. Through no fault of their own, textile workers in North Carolina bear the costs of a new technology invented in Tokyo. The market is a marvelous engine of growth, but it is not very good at providing a socially defensible distribution of income. Over the course of this century, governments big and small have intervened to smooth out business cycles and distribute income more equitably; in this country, the devices have included minimum-wage laws, legal recognition of trade unions, Social Security, and a progressive income tax, among others. Perhaps we should add to the list of things that the market does imperfectly the distribution of work opportunity, and find ways to intervene here, as well.

Wassily Leontief, the Nobel laureate in economics, argues that the loss of purchasing power from the decline in wages brought about by automation could doom the economy to a paradoxical state of rising productivity and rising destitution. Leontief proposes an ever-shorter work week, in order to assure a fair distribution of jobs and leisure, and a greater reliance on government transfer payments to distribute income and maintain consumer purchasing power. Although millions of workers are unemployed, millions more are working longer at their jobs than they might wish: the two obvious categories are people near retirement age and working parents. Following Leontief's approach, the government might supplement the pay of workers who voluntarily cut back to a three-day work week at the age of sixty. It could also supplement the income of working parents so that they could work half-time for, say, five years at a full-time rate of pay. Millions of new jobs would thereby be opened up, and at the same time, those who chose to work fewer hours for a while would suffer no loss of purchasing power. This system would be expensive in tax dollars over the short run, but as the unemployment rate dropped, it would also generate economic growth, and hence more tax revenue. Unions could then welcome automation without fear that society as a whole would benefit at the expense of the livelihood of production workers.

Automation may be the ultimate solution to the inroads of foreign competition on American markets, and to the loss of American plants to poorer countries. As more efficient machines replace human workers, smarter capital equipment matters more than cheaper labor. But for automation truly to serve as a remedy of this country's economic ailments, society must become far more explicit about how the benefits of automation are distributed. Leontief's idea--automation combined with redistribution--would require an even greater role for trade unions and for the state than they now play. Although the government as the redistributor of income and trade unions as the defenders of wages have been declared all but extinct lately, they may be the best instruments we have to resolve an explosive dilemma. The debate about the distribution of wages and the content of work has scarcely started, but it is certain to become more urgent.


Bob Kuttner is co-editor of The American Prospect and frequently writes for The Washington Post, The Boston Globe, and Business Week.

Copyright © 1983 by Bob Kuttner. All rights reserved.
m_nv_cv picture m_nv_un picture m_nv_am picture m_nv_pr picture m_nv_as picture m_nv_se picture