The US economy is about to enter its sixth year of recovery from the recession of 1990–1991. Even though the economy’s many “restructurings” and “downsizings” have been a cause of great anxiety for Americans, the economy’s leading indicators between 1992 and 1995 showed strength rarely seen since the 1950s. Investment has been growing swiftly, productivity has been rising at twice the rate of the 1970s and 1980s, and inflation has remained low. At the beginning of 1996, despite the economy’s slow start, the consensus of economists polled by The Wall Street Journal is for steady growth without inflation during the coming year.1

The growth of investment has been particularly impressive, rising at its steepest rate since World War II.2 Spending by private business on information technology, mainly computers and software, has been leading this investment boom. In 1993, spending on computers alone grew at the phenomenal rate of 52 percent; it grew by 27 percent in 1994, and was projected to grow by 23 percent in 1995.3 This very rapid growth of investment during the 1990s has also boosted the growth of productivity—the output per head of private-sector employees.

Productivity is a basic measure of economic efficiency, and its growth rate reveals how effectively such components of investment as new plants and equipment are actually being used. In the view of many economists, including those of the Clinton administration, the rate of productivity growth also largely determines the living standard of most Americans. According to the Administration, productivity is “the primary source of income growth” and its increase “is the answer to stagnant real wages.”4 It ought to be a positive sign that the growth of productivity continued to accelerate in 1995 even though the economy’s rate of growth was slowing down.5

But amid this news of productivity growth, one set of statistics has not recovered since the recession that ended in 1991: the living standard of most Americans. According to the Census Bureau, median family income in 1994, adjusted for inflation, was $38,782, one percent below the level of 1991. Preliminary data suggests that there was no significant increase in median family income during 1995.6 Moreover, this recent stagnation is part of a larger downward trend that has been taking place since the early 1970s.

Government statisticians lump the 80 percent of working Americans whose jobs fall below the higher executive, managerial, and technical levels under the heading “production and non-supervisory workers.” The average weekly earnings of these mostly rank-and-file workers, again adjusted for inflation, fell by 18 percent between 1973 and 1995, from $315 per week to $258 per week. By contrast, between 1979 and 1989 the real annual pay of corporate chief executives increased by 19 percent, and by 66 percent after taxes.7 For the other eight out of ten workers in America, the renewed growth of productivity during the 1990s has not brought a renewed growth of real wages.

During the 1970s and 1980s US productivity grew at less than half its pace in the 1950s. Nevertheless, between 1973 and 1995 output per person of all non-farm workers in the private sector still rose by 25 percent, while the real hourly earnings of production and other non-supervisory workers fell by 12 percent.8 During the present recovery, increases in real wages have not matched increases in productivity. Between 1990 and 1995 the productivity of all non-farm private sector employees increased by 10.3 percent, while during the same period the real hourly wages of rank and file workers were unchanged.9 This is the first time in American postwar history that the real wages of most workers have failed to increase during a recovery.

Alan Greenspan warned Congress in July 1995 that the growing inequality of income in the United States could become a “major threat to our society” (though, characteristically, he went on to say that solving the problem was “beyond the power of the Fed”).10 Commenting on the same trends, Robert Solow of MIT, a winner of the Nobel Prize for Economics, has warned of a society “which might turn mean and crabbed, limited in what it can do, worried about the future. 11 Felix Rohatyn, senior partner of the Wall Street investment bankers Lazard Frères, has spoken of an “advanced capitalism” whose “harsh and cruel climate” imposes “stringent discipline on its participants”:

What is occurring is a huge transfer of wealth from lower skilled, middle-class American workers to the owners of capital assets and to a new technological aristocracy with a large element of compensation tied to stock values.12


From the time of the industrial revolution onward, one kind of middle-income worker, the skilled craftsman, has been consistently under threat from the mass-production economy, with its automated production and standardized products. But today a much broader variety of middle-income workers finds itself threatened by a new mass-production economy, one shaped by the application of information technology both to manufacturing and to service industries.

In manufacturing there has been the advent of “lean production,” the techniques of mass production originating mostly in Japan and now diffused throughout the industrialized world. In industries such as automobiles, electronics, and machine tools, lean production has three main requirements: products must be easy to assemble (“manufacturability”); workers must be less specialized in their skills (“flexibility of labor”); and stocks of inventory must be less costly to maintain (components arrive at the assembly plant “just in time,” and so save on both warehousing and financing costs).


The second big change has been in service industries such as banking, communications, and insurance, where “reengineering” has transformed the work of many employees, costing large numbers of them their jobs. Just as Henry Ford once found a substitute for skilled craftsmen in rows of machines arranged along an assembly line, so the experts called reengineers have combined the skills of specialist clerks and middle managers into software packages that are attached to desktop computers.

We have no exact measure of the percentage of the US Gross Domestic Product (GDP), or of the total US work force, that has been subject to reengineering or to lean production. But there is no doubt that these two innovations have already radically changed the main manufacturing industries such as automobiles and electronics and the largest service industries such as banking and insurance.

Martin Kenney’s and Richard Florida’s Beyond Mass Production: The Japanese System and Its Transfer to the US (1993) describes how lean production methods have been adopted by virtually every leading industry: autos, auto parts, steel, computers, consumer electronics, and machine tools among them. As for reengineering, in his book The Reengineering Revolution (1995)13 Michael Hammer quotes two separate 1994 surveys, conducted by “Big Six” accounting firms, which showed that between 75 and 80 percent of America’s largest companies “had already begun reengineering and would be increasing their commitment to it over the next few years.”

The McKinsey Global Institute’s exhaustive study, Manufacturing Productivity (1993), gives perhaps the best recent account of lean production and its effects throughout the world.14 The report assesses US, Japanese, and German industrial performance both in the big unfashionable industries like food processing as well as in the more sophisticated industries examined by Kenney and Florida—autos, computers, and machine tools. The report’s most startling findings concern the relationship between productivity and skills in Japanese manufacturing. In industries such as automobiles and machine tools, McKinsey reached the familiar conclusion that Japanese productivity was well ahead of that achieved by US and German competitors such as GM and Volkswagen.

But McKinsey also found that when Japanese corporations achieved higher rates of productivity, their need for skilled labor declined, particularly on the shop floor. In industry after industry, according to McKinsey, “ease of production” was a prime factor in explaining the differences in labor productivity between the Japanese and their foreign competitors—differences which persist despite the current problems of the Japanese banks and other financial institutions.15 In designing automobiles, for example, Japanese manufacturers “remain very conscious of how such designs affect the manufacturing function, and attempt to specify them so that they are easier to produce and assemble.’16

In the machine-tool industry, McKinsey found that by streamlining production Japanese companies used one tenth the number of parts and tools—ball bearings, high speed drills—required in German craft production.17 McKinsey, therefore, repeatedly criticizes German industry and its arduously trained engineering Meisters and Technikers, since such reliance on craft labor can foster rigidity and “compartmentalization” on the line and so undermine productivity.18

Instead McKinsey praises what it calls the “integrated worker” of Japanese lean production, where a worker fits easily into a work “team.” In fact the “integrated worker” employed by Japanese industrial corporations need have few recognizable skills at all, as I found during my own visits to the European plants of Nissan and Honda. I asked personnel managers at both plants what importance they gave to the educational and vocational qualifications of most of their prospective shop floor employees. In both cases the answer was “very little.” What they were looking for, they explained, were dexterity, enthusiasm, and an ability to “fit into the team.” In one job test, candidates were shown piles of nuts and screws of various sizes and were asked to match up the fitting pairs as quickly as they could.

In a well-known account of lean production in the world auto industry, the economists James Womack, David Jones, and David Roos argued that increased reliance on less-skilled and lower-paid workers was one of the chief selling points of the new system. One advantage of lean production, they write, is that it

dramatically lowers the amount of high-wage effort needed to produce a product of a given description, and it keeps reducing it through continuous incremental improvement.19

These aspects of lean production help to explain why the rising productivity of blue-collar workers is not being matched by the rising value of their paychecks: corporations that can achieve higher productivity while dispensing with entire categories of skilled labor are unlikely to pass those savings on to their remaining, and less-skilled, workers.


But the disparity between productivity and real wages has other, deeper, causes as well. The large industrial corporation can cut its costs even further by contracting out (or “outsourcing”) as much production as possible to smaller, independent producers, whose wage costs will be lower than its own. McKinsey found that such outsourcing was taking place in the manufacture not only of automobiles but of machine tools, computers, and consumer electronics as well. And here again the Detroit Big Three have been following the example of their Japanese competitors. Toyota, a pioneer of lean production, farms out no less than 70 percent of its components manufacture. They can do so because, as production becomes simplified, a contract manufacturer can perform most tasks as efficiently as the parent company itself.

In the US, the workers who are hired as a result of outsourcing are usually non-union and among the lower paid. When the parent company awards contracts to various satellite manufacturers, it does so as part of a strategy to keep costs down. Moreover, US companies who do this save more than do companies in Japan, where a good many workers in the satellite companies receive traditional benefits of Japanese industry. The employees of companies such as Nippondenso, Toyota’s main supplier of electrical components and engine computers, often belong to unions and are guaranteed “employment for life.” It is only further down the supply chain, among the small shops that make more basic components such as screws and fasteners, that one can see the full effects of outsourcing on working conditions. In the US, however, there is no well-developed layer of privileged suppliers that can cushion the work force against the effects of outsourcing, particularly the lower wages paid by suppliers competing to charge less.

McKinsey found that contracts between the Detroit Big Three and their suppliers were “traditionally short term in nature” and that the Big Three often had several suppliers for the same items. Always implicit in such contracts was the threat that one of the big companies would “switch suppliers” once the contracts expired. Such arrangements help to explain why “suppliers seem to have given away.” much of their profit “in order to retain [the corporation’s] business.”20 In its own survey of outsourcing, Fortune claimed that Detroit’s “stunning revival” has been a “salute to contract manufacturing.” 21

Indeed, with 70 percent of its components made by outside companies, Chrysler now relies on contract manufacturing as much as Toyota does. According to Fortune, GM, too, is “closing plants and outsourcing production, showering new business on contract manufacturers.” This rush to farm out automotive production has crippled the United Auto Workers and been ruinous for many blue-collar employees. Last June The New York Times pointed out that unionizing workers in the components industry was the most urgent task facing Stephen Yokich, the new president of the UAW.22

During the late 1970s, when most components manufacturers were wholly owned subsidiaries of the Big Three, two thirds of the hourly work force in the industry belonged to the UAW. But today, with outsourcing, only one quarter of the components work force are union members. This growth of the non-union work force has also had a devastating effect on the earnings of workers throughout the auto industry. Between 1975 and 1990 the percentage of low-wage employees in its total work force grew by 142 percent, from 17 percent to 40 percent.23—proof, if nothing else, of the claim that lean production lowers the number of high-wage employees needed to produce cars.24


While lean production can keep down wages in manufacturing, over 80 percent of Americans now work in service industries, the vast, amorphous world which includes, among others, banks, insurance companies, S & Ls, retail and wholesale outlets, restaurants, hotels, and barber shops. Many of these industries have their own rough equivalent of lean production—reengineering—whose chief proponents and theoreticians are Michael Hammer, a business consultant, and James Champy, a former professor of business studies at MIT, and now also a consultant. Between them Hammer and Champy have now written three books about reengineering.25 Their first book, Reengineering the Corporation (1993), had an international sale of nearly two million copies, and by now their ideas are familiar to most high corporate executives.

Reengineering is having an enormous effect on the economy, but it is a less grandiose activity than it sounds. The main task of the reengineer is to use information technology to streamline many of the more routine activities of business life. In a service industry such as insurance, reengineering in some form has been going on at least since the late 1960s.26 The reengineering boom came only in the 1990s when computer software became increasingly able to do the jobs of, and often replace, large numbers of workers in hundreds of companies. With the help of some adroit marketing and public relations, Hammer and Champy have tried to give the impression not only that reengineering is their invention but also that reengineering cannot be successfully carried through without the help of high-priced consultants such as themselves.

According to Fortune, however, “mass production has come to the [reengineering] industry” itself: “Once you’ve streamlined one account payable process, you’ve streamlined them all.” With mass production “those high flying armies of MBA’s have given way to in-house task forces.” Among reengineering consultants such as Gemini, and Champy’s own CSC Index, there are now signs of “cost cutting and price pressure.”27 Nevertheless the three books of Hammer and Champy, with their many case histories, provide the most detailed account of reengineering we have.

One case perhaps best epitomizes how reengineering works.28 With its headquarters at Stamford, Connecticut, the IBM Credit Corporation, a subsidiary of IBM, provides credit for clients who want to purchase IBM products. Five years ago specialists were responsible for each stage of the business. The low-level managers and skilled clerical workers in one department logged in credit applications. The employees in another department inserted special conditions for particular clients. Another department fixed the appropriate rate of interest. Yet another gathered all the relevant information into a “quote letter” to be sent to the IBM salesman and his customer. The entire system depended on vast quantities of paper creeping slowly from one specialist department to another.

Reengineering teams descended on these clerical assembly lines and swept them away. They installed in their place employees using software that could accomplish the tasks hitherto carried out by the specialists. Except for a few exceptional cases, which are still handled by a team of specialists, the functions once divided among the various departments have now become the responsibility of a single employee who is called the “deal structurer.” Sitting at his computer terminal, the deal structurer handles the entire credit-granting process, from the moment that IBM receives the customer’s credit application to the moment that the company’s “quote letter” is sent off by Federal Express. A software program is able to handle all but a few of the questions that arise at each stage. The increases in productivity achieved can be prodigious. Reengineering at IBM increased the productivity of its credit business “not by 100 percent, but one hundred times,” or 10,000 percent.29 The “deal structurer,” moreover, may displace not only many clerical workers but more highly qualified employees who held what they thought were secure “middle-level management” positions.

Hammer and Champy praise their clients—and, by implication, themselves—for carrying out the new and simpler divisions of the work force which reengineering makes possible. But the real stars of reengineering are the software and hardware packages that form the core of virtually every reengineering project. Without these packages there could be no consolidation of many tasks into a single task performed with a speed unimaginable in the pre-computer age, and no spectacular leaps of productivity of the magnitude achieved at IBM Credit.

If one reads Hammer’s and Champy’s case studies carefully, one finds that such packages are the real forces making reengineering possible, whether the process involves, to cite only a few examples, the fulfillment of orders at Texas Instruments, auto sales at Chrysler Financial, insurance claims at Aetna, CIGNA, Capitol Holdings, Liberty Mutual, and Progressive Insurance, credit card management at AT&T, or customer complaints at Brooklyn Union Gas. Reengineering places great power in the hands of those in charge of the relevant technology—those who understand the software and hardware packages (and may have created them), and who know how to use them to change a corporation’s entire way of operating.

These high-level technocrats make up a large part of Felix Rohatyn’s “technological aristocracy,” and they are uniquely well placed to influence the speed, direction, and scope of reengineering. But within the corporation they are also closely linked to another small group, the handful of top executives, led by someone whom Hammer and Champy call the reengineering “czar,” who have the final say on investment decisions and who also have the power to hire and fire.

Outside the two groups that direct the reengineering effort, there is unlikely to be any competing center of power that might represent the interests of most of a company’s employees. The people in charge of reengineering seldom have to deal with unions, since today there are hardly any unions in the service industries in which most of the reengineering described by Hammer and Champy has taken place. When the two authors mention unions, they usually dismiss them as antiquated relics of the mass-production age. Hammer and Champy take centralized power for granted.

Michael Hammer in particular celebrates the schemes of reengineers, likening their work to that of real engineers. “Engineers are creative people,” he writes, “and they often have great emotional investments in their designs.” Reengineers “have similar feelings,” and they naively expect that “the elegance of a new process design will automatically cause everyone to wholeheartedly embrace it.” But “elegant, efficient and original” though the process design may be, there will always be those employees who ask such tiresome questions as “What’s in it for me?”

Hammer and Champy recommend various public relations techniques to win skeptical employees over to reengineering (“share their fear and pain”); but Hammer also warns that tougher tactics may have to be used. According to him, “slapping people’s wrists instead of breaking their legs” is “another sign of weakness”; and again, “making it clear that termination is the consequence of their behavior is a very valid technique.”

Hammer also writes with considerable frankness about the climate of fear that reengineering can create. He mentions the “trepidation and anxiety” of the work force, and the “abject terror” and “total inner panic” of those who are told that they are “going to have to change what they do.”30 Nowhere are Hammer and Champy prepared to admit that employees have interests which ought to be the subject of negotiation and compromise. Nor do they have anything to say about arguments over pay, because wage bargaining rarely takes place in the reengineered corporation. Salaries are usually a matter of private discussion between the employer and the individual employee. In these isolated, bilateral encounters, the employer generally has the upper hand.

Hammer and Champy have much to say about the techniques of reengineering. But in their infatuation with the reengineer as technician, they seem to have forgotten about the reengineer as businessman, someone striving constantly to cut costs, increase profits, push up the share price, and so add to his own net worth as well as the shareholder’s. As Felix Rohatyn has pointed out, the compensation of the CEO and the top reengineers is more and more linked to the corporate share price. Far from being a neutral force of modernization, the power accumulating in their hands determines how the rewards of greater productivity are to be redistributed among shareholders, top managers, and employees.

One of the most authoritative accounts of how this new distribution is taking place in the economy of lean production and reengineering has been provided by Lawrence B. Lindsay, a governor of the Federal Reserve Bank who sometimes speaks more forth-rightly than most other central bankers. In March 1994, speaking before an audience of Baltimore bankers,31 Lindsay discussed who was winning and who was losing both among those who worked for wages and salaries and among those who owned capital and benefited from interest and dividends.

Among employees, Lindsay includes all wage earners, the 20 percent who are senior executives, managers, and technicians, as well as the 80 percent who are “production and nonsupervisory workers.” But even with the incomes of the high earners included, Lindsay found that the employees’ share of increased income had fallen dramatically during the past ten years. When he compared the distribution of income gains in 1993 to the pattern prevailing between 1981 and 1989, the share going to wage earners fell from 52 percent to 38 percent; the share paid out as capital income remained constant at 28 percent. Lindsay pointed out that the share going to wage earners was the lowest in recent times.32


At the middle and lower echelons, more and more workers rely on skills that are easier to learn than those of their predecessors. The reengineered “generalists,” the “deal structurers,” and others who use desktop computers to handle the flow of corporate work are often less skilled than the specialists they have replaced. They often have less education. In fact the computer skills of generalists can easily be supplied by large temporary-help agencies such as Manpower.33 In January 1995 Mitchell Fromstein, the CEO of Manpower, boasted to Fortune that “we can manufacture a skilled worker if we start with a person who has earned a high school degree.” It takes only two months to train workers in basic computer skills, one of his executives told me.

The process described by Felix Rohatyn—“the huge transfer of wealth from lower skilled middle-class workers to the owners of capital assets and to a new technological aristocracy”—is largely a response to market forces and will not disappear at the next turn of the business cycle. The skills of the technical aristocracy who design software and reengineer work forces will remain in strong demand among corporations that cannot afford to lag behind the competition in technological innovation. But since such skills are difficult to acquire, they will remain in short supply, and their price will continue to rise. There will also be a continuing demand for the kind of middle-level skills which, a generation ago, provided so many Americans with a comfortable, living wage, whether machinists to build aircraft and turbines, or plumbers and electricians to construct and maintain homes and office buildings. But the relative demand for such skilled workers is rising at a slow rate, if at all, and it seems unlikely that their wages will do much more than match inflation.

There will also be a demand for workers with the amplified skills of the new mass-production economy. But such workers can more and more be hired from what has come to be known as the “contingent work force.” Last November, for example, the United Auto Workers’ long and bitter strike against Caterpillar Inc. ended in defeat in part because the company found that it could rely on temporary-help agencies to replace virtually all its striking shop floor employees.34 The claims of Labor Secretary Reich and President Clinton that job training is a sufficient solution to the problem of static or falling real wages therefore seem exaggerated.35

Retraining programs are certainly necessary to help discharged workers acquire new skills, often skills involving some form of computer processing. But the requirements of the new mass-production economy make themselves felt most strongly among the younger people entering the labor market. The economy of lean production and reengineering has no need for a large and growing supply of younger workers who have had the kind of German-style training for technically advanced production work that so attracts Clinton and Reich. We constantly hear from Democratic politicians and from corporate executives like Bill Gates that “education” must be improved if it is to counter the effects of corporate downsizing. In his State of the Union speech, President Clinton proposed a $2,600 voucher for “unemployed or under-employed” workers “to use as they please for community college tuition or other training.” Clinton’s program might well help some young unemployed workers acquire the skills of a “deal structurer” or even of an old-fashioned engineering craftsman. But it is an illusion to believe that such job programs can by themselves recreate in significant numbers the secure, well-paid, and relatively high-skilled jobs that members of the middle class have traditionally held. Already in the 1990s the US economy has achieved a high rate of productivity growth, an even higher rate of growth of investment in information technology, and the creation of over eight million new jobs. But these successes have been achieved without any noticeable improvement in the nation’s systems of vocational and high school education.

As more and more Americans are drawn into the new economy of mass production, will they accept the harsh verdict of the market and live with its social consequences—the advent of the “mean and crabbed society” predicted three years ago by Robert Solow? The political parties and their leaders have not been of much help in confronting the problems of corporate downsizing and declining incomes that now affect tens of millions of workers. Newt Gingrich’s Contract With America says in passing that “the American Dream today exceeds the grasp of too many Americans,”36 but to revive it the Republicans rely on their all-purpose solution of balancing the federal budget within seven years, which they believe will lower interest rates and so add to the growth of the GDP. Even if the budget were balanced and the anticipated growth were to occur, nothing in the record of the past twelve years suggests that growth alone will be sufficient to reverse the growth of inequality in the US. During the present recovery, growth has been strong enough to push stock prices, corporate profits, and the pay of CEOs to record levels while the pay and living standard of most Americans has continued to stagnate. The central problem of the 1990s has been more the maldistribution of wealth than failure to increase overall wealth.

In his presidential campaign four years ago Bill Clinton briefly revived the idea that new federal programs and spending could be relied on to create a “high wage, high skill economy.” But Clinton has more and more embraced the strategy of his opponents, relying mainly on cuts in the deficit to promote growth and reverse the decline in real wages. Largely overlooked by Democratic politicians is the tradition of New Deal progressivism by which Franklin Roosevelt used legislation to curb excessive concentrations of power in the hands of corporate interests. The enduring, if much diminished, monument of this philosophy is the National Labor Relations Act, which gave unions the right to organize and protected the principle of collective bargaining at the workplace.

One of the few leading Democrats to recognize the contemporary relevance of this approach is Senator Bill Bradley. In his new book Bradley describes constituents who have suffered from corporate policies and argues that employers have obligations to their longer-serving ex-employees which are not fulfilled when, to quote one of his constituents, “after twenty years they throw you out with six months severance and no medical.”37 Bradley proposes that employers should be obliged by law not only to cover the medical insurance of longer-serving former employees for up to a year but also to contribute substantially to an employee’s retraining costs. In addition he urges that employee pensions be fully transferable from one job to another.

Measures such as these would help the victims of reengineering and down-sizing. But what can be done to protect those who are still employed and who suffer from the present imbalance of power within the workplace? It is not difficult to conceive of laws requiring that, when reengineering and restructuring take place, the views and interests of employees will be taken into account and not trampled underfoot. Where unions are still a considerable if waning force—for example, in the steel, auto, and aerospace industries—much could be achieved through strengthening the NLRA itself. The rights of unions to organize would have to be better protected by new laws so that, for instance, the Detroit Big Three would not be able to drive down wages in the auto industry by shifting production to non-union plants.

The right to strike needs to be strengthened against employers like Caterpillar who use “permanent replacement workers” to wear down and defeat their employees. But employees also need some alternative form of representation in those many sectors of the economy, notably in services, where unions have virtually ceased to exist. A leading labor economist, Richard Freeman of Harvard, cites the European-style works council as a model that can give white-collar employees as well as blue-collar workers extensive rights of “co-determination.”38 In countries where works councils are strongest, such as Germany and Holland, management must obtain the agreement of the employees’ council on such matters as reengineering and training. The most advanced forms of co-determination are often to be found among employee-owned businesses whose organization and growth in the US could be encouraged by new federal laws.39

These approaches to the problems of the work force are based on the principle that new laws can protect citizens from the less acceptable traits of the market—its greed and its harshness. Today, as in the 1930s, the market is creating more losers than winners. But sixty years ago the divisions among workers, managers, and owners were much clearer, and entire categories of workers could not be displaced so quickly by new technology. Today the distinction between winners and losers and its causes are less well understood. Until political leaders can analyze and confront the realities of the new mass-production economy and convince large numbers of workers that new laws can protect them from its worst abuses, the current trend toward a more and more unequal society will continue.

This Issue

February 29, 1996