Downsizing in America opens with excerpts from mid-1990s articles in The New York Times and Wall Street Journal telling of longtime workers from firms like Kodak, General Motors, and IBM who had recently lost their jobs. But, the authors ask, was this a fair account? “Did companies actually reduce the size of their workforces?” After an exhaustive analysis, supported by more than a hundred tables, William Baumol, Alan Blinder, and Edward Wolff conclude that “media attention lavished on downsizing may have left people believing that job insecurity had increased more than it actually had.” In fact, they say, “the creation of new jobs always overwhelms the destruction of old jobs by a huge margin.”

Indeed, statistics can be cited to confirm their position. Between 1980 and 2002, while the population grew by 23.9 percent, the number of employed Americans rose by 37.4 percent. In fact, as a fraction of the population, more people now hold full-time jobs than ever before. The principal reason is that millions of women have not only joined the workforce, but have helped to expand its size. The same holds for teenagers, even if we take account of the fact that most of their jobs are part-time. In view of these and similar facts, Baumol, Blinder, and Wolff propose that most large firms “may have simply altered the compositions of their workforces more than their sizes.” They describe a process of “restructuring,” in which

companies may have first reduced the number of less-educated or older workers or middle managers. Then, after a suitable interval required for retooling and reorganization, they may have replaced most of their former employees with others deemed more appropriate to the company’s current needs.

The key word here is “replaced”: Who replaces whom and what happens to those who leave?

Almost all the calculations in Downsizing in America are based on the total number of people employed by the nation’s largest corporations. (Only four tables focus on specific companies.) However, the experience of actual firms can be quite surprising. Their employment records are easily extracted from Fortune’s annual tables, which show how many people worked for each company in the previous year. The percentages cited in Table A compare the payrolls on its 1981 and 2003 lists. Between those years, for example, General Motors shed more than half its workers, going from a payroll of 746,000 to 350,000 people. Xerox dropped from 120,500 workers to 67,100, and DuPont from 135,000 to 79,000. During this period, it will be recalled, the overall population had grown by 23.9 percent.

The authors’ view is that what has been seen as a “downsizing trend is more or less restricted to manufacturing.” In part, this is the result of automation: not as many workers are needed to build a car. Some other companies, like Campbell Soups and Sears, simply haven’t kept up with the competition. Still others, such as Xerox and Goodyear, have lost ground to foreign companies. Moreover, mergers can create pink slips. The six companies on the table which folded into three together ended up with 366,929 fewer employees. At the same time, firms that manufacture computers and chips and other high-tech products have expanded their payrolls, while Procter & Gamble, founded in 1857, has shown it is possible to keep abreast of the times.

It would be well to take another look at Baumol, Blinder, and Wolff’s assertion that “the creation of new jobs always overwhelms the destruction of old jobs by a huge margin.” Yes, more new jobs have been created but most of them pay relatively low wages. The most conspicuous case is Wal-Mart, which had a payroll of 27,000 in 1981 and was employing 1.3 million people twenty-one years later. In that time, McDonald’s also added 296,000 places. As it happens, the new jobs created by these two companies alone made up for the combined losses of the dozen downsized firms listed in Table A. Yet it hardly needs saying that the typical new job at Wal-Mart and McDonald’s, even if pursued full-time, pays about a third of the wage for a lot of the jobs that no longer exist. So one effect of recent job creation is that it may take three paychecks to bring in what one did before.

Another result of downsizing is that growing numbers of young people cannot count on the kinds of careers that were once quite common. This holds not only for factory jobs, but also for those aspiring to be independent professionals such as physicians or accountants with their own practices, a point that emerges from Simon Head’s book.

Toward the end of Downsizing in America, the authors move from algebraic equations to some plainspoken comments. Too often, they say, reducing payrolls is simply a “transformation of what were once wages into profits.” Nor do they accept the arguments of executives who claim companies must become leaner to survive in modern markets. In their view


cutting the labor force is not a sure-fire way to increase the competitiveness of American enterprises, or an effective way of raising output per worker…. Downsizing is profitable at least partly because it is an effective way to hold down wages. We may be led to say that this is the dirty little secret of downsizing.

Or when not holding down wages, downsizing may be used to get more work out of the people you keep. I saw this happen not long ago, while I was visiting a brokerage house. My host had earlier called in a researcher who analyzes food companies and told her that, owing to staff cuts, her job would henceforward include beverage companies as well. Later, I asked how much more time the researcher would have to work. Perhaps six, even eight, hours each week, I was told.


Low-Wage America, another Russell Sage Foundation volume, contains twelve studies of industries that maintain their profit margins by paying their workers at or near the minimum wage. They range from telephone call centers and temporary services to expensive hotels and hosiery factories. Specifically, the book concentrates on a growing number of “Americans who do not earn enough to support themselves and their families.” As it happens, this problem is relatively new.

During the heyday of union organizing, a frequent complaint was that too few companies paid their employees a “family wage.” It was presumed that a worker was his household’s sole earner and he had to make enough to support his wife and several children. For much of the last century, one paycheck could suffice, partly because of the level of wages maintained by union pressure, and also because families got by on modest budgets. For example, most of them had only one car; the children shared rooms and fewer of them expected to go to college.

Today, most American jobs do not pay enough to support a full household, or at least to have what most people now feel are necessities. In 2001, the most recent year for earnings figures, of the roughly 100 million men and women with full-time employment, over half made less than $35,000. Indeed, as Table B shows, only 15.7 percent of all jobs paid as much as $65,000, a figure many middle-class Americans would feel is needed to maintain a family with children. Even if the threshold is lowered to $45,000, only 32.8 percent of jobs paid that much. In short, our economy is providing relatively few “family wages.” Not the least reason is that more outlays are now seen as essential even for lower-middle-class suburban life. People feel that they must have two or more cars, that each child should have his or her own room and expect to go to college. The authors also cite the decline of unions. Currently, only 13.2 percent of workers belong to a union, compared with 35.5 percent at the end of World War II. Moreover, close to half of the union members are on public payrolls, like police officers, prison guards, and teachers.

The figures in Table C show how people are coping. Among the families that include wage earners, only about a third now have just one. In some of these, one spouse has high earnings, so the other can stay home if that’s what he or she prefers. And similar choices are still being made by women in families of modest means, who know it will call for a lot of fiscal discipline. Even so, two thirds of all households with wage earners now have two or more. As the table also indicates, this ratio continues when families have children, where two thirds of the mothers work.

Low-Wage America might have said more about who precisely are the people who end up with menial pay, and how many of them should be regarded as exploited or ill-treated. Here are some of the new low-wage recruits.

Wives. Although most households now have at least two earners, fewer than half of the wives hold full-time jobs. The median paycheck for wives who worked in 2001 was under $18,000, which is close to the low-wage category. Whether by choice or circumstance, most wives’ earnings tend to be supplemental. In the typical two-job family, she contributes 36.1 percent of their joint income.

Teens. Between 1980 and 2002, teenaged employment grew by 63.2 percent, compared with 37.4 percent growth for the rolls as a whole. Most young people work part-time; yet for supermarkets, shopping malls, and fast-food franchises, they have become an essential source of labor. While their wages are low, they help pay for their cars and for independence from their parents.

Singles. Young people are postponing marriage and having children, and not all are intensely concerned with how much they make. For the 46 million Americans who in 2001 had not yet married, median earnings were $16,934. One can see some of them working at Home Depot and Barnes & Noble, or taking orders in a restaurant. Nor are all of them dropouts. A Low-Wage America study discovered that “half of the employees in the call center that services large companies have college degrees.” (In such centers employees make calls seeking orders or deal with inquiries.)


Immigrants. That more people are being paid less is in large part a consequence of the surge in immigration. Since 1970, the Hispanic portion of the population has tripled in size, while Asians have increased by a factor of five. Most recent arrivals are willing to accept low pay; indeed, many come here hoping to find those positions. The statistics on the hotel industry in Low-Wage America show that between 1979 and 2000, the proportion of Hispanics and Asians rose from one in ten to one in three of their employees. This, it was explained, helped them to keep their room rates from rising higher.

Single Mothers. For women who head households on their own, the median income was $18,472 in 2001, which means half of them made less than that. Three quarters of single mothers now work, mostly at bottom-tier jobs, often because their welfare stipends have been cut off. Most of the fathers are behind in support payments, or pay none at all. (Many can’t, since they are in prison.) One paycheck must support a family, and it is amazing that as many cope as well as they do. Currently, 22.4 percent of US children are being raised under these straitened conditions. And because of the high incidence of divorce, desertion, and out-of-wedlock births, their numbers show no sign of diminishing.1

The editors of Low-Wage America mention only in passing “the increase in inequality that has characterized the US income distribution in recent years.” And they never really link this fact to the subject of their book. Yet it seems clear that one reason for the growth in low-end jobs is that much more of the economy’s money is going to workers and families at the high end. In 1980, men who had graduated from college ended up with $1,690 for every $1,000 earned by those who had gone no further than high school. By 2001, the differential for college graduates had risen to $2,243 per $1,000 earned by high school graduates. In its reports on family incomes, the Census calculates how much of the nation’s aggregate income goes to people at different income levels. In 1980, the top 5 percent of all families received 14.6 percent of the total. By 2001, that group was getting 21.0 percent. Thanks to this upward redistribution, the already small share going to the bottom fifth of families fell by 20.8 percent. Whether or not today’s poor are materially worse off, their share of the national wealth has certainly dropped.


In his The New Ruthless Economy, Simon Head provides detailed and disturbing glimpses of how, to take only three examples, digital programs can be applied to assembling automobiles, running telephone centers, and managing medical care, with the result that wages decline and work becomes more tightly controlled.

It’s probably no surprise to learn that putting a car together is divided by management into “sub-jobs,” consisting of tasks “precisely timed to last twenty-three or seventeen seconds each.” Not to mention instructions like “extend the right hand, pick up screw, insert through seal.” But similar systems of control now prevail at call centers, where scripts dictate the exact words agents must use when they deal with statements and questions from customers. There is no room for further explanations, or for dealing with special problems. Clerks at a major health plan are warned not to exceed 225 seconds in speaking with clients. Still, there are rewards for initiative: bonuses are given to employees for dissuading callers from asking for appointments with a physician. One reason is that plan doctors often have to deal with upward of seventy patients a day, which works out to an average of eight minutes per patient.

What we are seeing, Head writes, is “the industrialization and reengineering of health care,” with the chief impact on professional judgment and discretion. Managed care plans have created “decision-making algorithms” that dictate

the proper length of a patient’s hospital stay, set out the appropriate length of time for a physician to spend with his or her patient, and rule on the treatments that patients should or should not receive.

Less well known are data-based “diagnostic systems that speed the identification of illness.” Their aim, as Head explains, is to replace old-style encounters where a doctor conducts a physical examination, asks questions, and takes specimens for tests. The problem here is that a diagnosis depends in large part on the insights that a physician has accumulated in his or her practice and on the relationship with patients whose condition may emerge clearly only after considerable inquiry. Indeed, we all know of people who had a mysterious ailment, and went to a succession of physicians before they found one who understood what their problem was. But now we have Iliad, “a medical expert software program” that offers “expert diagnostic consultations and patient simulations.” Its current edition covers some 900 specific diseases, 1,500 syndromes, and almost 14,000 “disease manifestations,” such as ringing in an ear or numbness of a toe.2 Its knowledge base has been compiled by panels of specialists, who provide diagnoses they associate with certain symptoms. A physician’s assistant can type in a version of a patient’s account and soon discover what the experts think. Of course, not all will agree all the time. So the program lists a series of possibilities, which will still help physicians to decide what is ailing their patients. As with similar “expert systems” now being used by accountants and lawyers, Iliad tries to forestall false starts and dead ends in making diagnoses; but the possibilities of ignoring the feelings and life histories of particular patients are evident.

Another “reengineering” possibility is that health plans will tell their members that before they may see a doctor, they must first go to a “diagnostic center,” where a technician will enter their medical history and symptoms into a computer. Since the physician will be given that printout, it should result in even shorter office visits. And it seems equally likely that the software will specify the treatments the plan will permit.

Or, to take another profession, professors usually lead a classroom discussion with one group of students. However, this method is very labor-intensive, and consequently expensive. So, as with medicine, technology can be used to raise productivity. Via “video conferencing,” a technique now common in the business world, one professor can teach two or more classes simultaneously in different buildings, or on different campuses. In most cases, the teacher meets personally with one set of students, who know that classmates elsewhere will be watching the professor on screens or monitors. Cameras and microphones allow for visual recognition and verbal interchanges. Here, slightly abridged, are some tips from a guide issued by Cornell University:

Wear solid colored clothing, avoid complex patterns, such as stripes or busy patterns.

Avoid distracting mannerisms, such as tapping a pen…or jingling change in pockets.

Smile when you make eye contact. Make eye contact for 2 to 3 seconds with the camera when looking at the distant class. Pretend the camera is one of your students.

Help students to be comfort-able with…speaking on-camera. Demonstrate how the technology affects interactivity.3

True, video yields reduced and flattened images of the distant students, as will be theirs of the instructor. Still, proponents of this method cite studies showing satisfaction on both sides after a period of adaptation. Indeed, the professor will probably have to do most of the adjusting, since today’s students already spend much of their lives staring at screens.

But if the professor has several classes of students, won’t there be that many more tests and papers? If that’s so, then there won’t be much of an increase in productivity, just one person doing more work. Here the teacher may eventually draw on e-rater, a software program being developed by the Educational Testing Service, which produces the Scholastic Assessment Test. Once an essay has been scanned, e-rater will proceed to

quantify such features as number of complement clauses, subordinate clauses, infinitive clauses and relative clauses and occurrences of subjunctive modal auxiliary verbs such as would, could, should, might and may.4

As for judging competence in dealing with subject matter, the ETS says that, “each argument in the essay is first evaluated separately by computing cosine distances between its weighted vector and weighted supervectors.” What this means in practice remains mysterious but the testing service claims that, thus far, the grades given by e-rater and human readers have agreed between 87 percent and 94 percent of the time, which is comparable to scoring done by two human beings.

New medical technologies that promise longer lives and improved health are now getting high praise and attention. (A pill, it should be added, is a technology in a tiny package.) Simon Head shows how managerial technologies are dictating the boundaries within which professionals must practice. He alludes to the changing ratio of medical managers to physicians in the health field. In 1980, for example, there was one such administrator for every four physicians. By 2001, the ratio had risen to one-to-one. This means that most doctors now have the equivalent of a personal health-care manager looking over their shoulders. And while managers don’t actually prescribe treatments, they set limits on what can be done or will be reimbursed. (And these figures don’t include the clerks they must deal with on the phone, or the lawyers who demand documents and depositions.) Thus the divide that Head acutely describes in industry after industry is basically one between those who impose technologies and the rest of us who will have little choice about their impact.

In the early days of American industry, Alexis de Tocqueville noted that “at the very time at which the science of manufactures lowers the class of workmen, it raises the class of masters.” He went on to predict:

The one will require nothing but physical strength without intelligence; the other stands in need of science, almost of genius, to ensure success. This man resembles more and more the administrator of a vast empire; that man is a brute.

Gallic hyperbole? Perhaps. But The New Ruthless Economy shows how American society is moving in the direction Tocqueville anticipated.

This Issue

February 12, 2004