The orthodoxy that sees economic growth as a process by which entrepreneurs orchestrate the continuous rearrangement of fungible resources rests on the assumption that industries grow and expire in a predictable sequence. It goes something like this. Step one: An individual inventor—call him Henry—has a Big Idea. Henry comes up with a turbo-charged automatic vacuum cleaner that can suck up all the dirt and dust out of a room without anyone lifting a finger. Just leave the vacuum on a shelf for five minutes and—presto!—the room is spanking clean. Step two: Henry presents his Big Idea to a group of wealthy investors who are on the lookout for hot new prospects. They agree to back him in exchange for shares in the new venture. Step three: Henry and the investors hire Ralph, who sets up and manages the factory to make the turbo-charged vacuums, and Shirley, who peddles the product to retail stores. Step four: Thousands of turbo-charged vacuums are rolling off conveyor belts manned by hundreds of drone workers. Hundreds more workers distribute, sell, and service the new machines.
Step five: As production expands, the cost of producing each vacuum declines. This is because many costs—such as machinery, factory maintenance, and advertising—are incurred regardless of how many vacuums are made, so the more production the better. (Henry Ford notably capitalized on this fundamental concept in producing his Model T.) As the per-unit cost falls, the profit on each sale rises. The company starts to make a good deal of money. Step six: At this point Henry and his backers sell the Turbo-charged Automatic Vacuum Company to a big corporation—say, Westinghouse—for a very large sum. Henry returns to his attic to come up with another Big Idea, and his backers go on the prowl for another Henry. Ralph and Shirley become middle managers in Westinghouse’s Consumer Appliance Division. Step seven: Now tens of thousands of turbo-charged vacuums are coming off the lines and the cost per unit has been pushed down to a pittance. The turbo-charged vacuum becomes the star of the Consumer Appliance Division, and Westinghouse earns a nice profit. Step eight: Westinghouse spends some of these profits acquiring another start-up enterprise, the Self-Destructing Picnic Plate Corporation. (Five seconds after it registers the disappearance of the last deviled egg, a microprocessor triggers a small explosive charge; nothing but water vapor and environmentally benign ash emerge from the fireball—saves messy cleanup.)
Step nine: Many American households now own a turbo-charged vacuum cleaner, and a competitor is about to introduce a new line of comparable machines. In response, Westinghouse drops the price of its model and embarks on a cost-cutting campaign. Ralph and Shirley, along with hundreds of workers, are laid off. Step ten: Profits have declined so far on the turbo-charged vacuum that Westinghouse decides to terminate the line. The factory and equipment that produced it are put to other uses or sold for scrap; the remaining workers, terminated. But Westinghouse does not regard the excursion as a failure. To the contrary, the turbo-charged vacuum cleaner netted a great deal of money for the firm. And anyway, the self-destructing picnic plate is just hitting its stride.
This sequence or something like it has been repeated in American industry often enough to be seen as the rule. And the rule has gone from being a description of what is to a prescription for what must be. Invention is distinct from production. Entrepreneurial efforts focus on Big Ideas—fundamentally new products or new ways of making things. The most successful of these Big Ideas will have a predictable life cycle, extending from an unprofitable period at the start, to a rewarding harvest season of expanding sales and mass production, and then to an eventual maturity and decline as the market becomes saturated and as other Big Ideas render the product obsolete.1
This suggests that a business firm, if it is to remain consistently profitable, should have a portfolio of Big Ideas at different stages of their life cycle. The typical large firm does not generate Big Ideas internally, but buys start-ups that seem likely to yield high profits when their Big Ideas reach full-blown mass production. Corporate headquarters uses the profits from its herd of winners to buy further promising start-ups. And it quickly withdraws from businesses that have reached the end of the line. Thus the ideal firm balances out rising and declining Big Ideas by diversifying its portfolio.2 In the mid-1980s the widely diversified firm—buying Big Ideas embodied in start-up firms and selling off the carcasses of Old Ideas—was more or less the rule.
From the standpoint of the lone genius like Henry, or even of the corporation like Westinghouse, this pattern makes economic sense. But from the standpoint of a national economy and the vast majority of people who depend upon it for their work and, ultimately, their standard of living, the pattern is more problematic. The only way that people like Ralph and Shirley and all the drone workers under their care can be assured of jobs and steadily rising real incomes is if entrepreneurial geniuses like Henry continue to conjure up successful Big Ideas quickly enough to replace the declines at the other end. In America, this recipe worked for a while.
But as I have sought to show, high-wage economies can no longer depend on standardized mass production. Big Ideas like Henry’s can be shipped in blueprints or electronic symbols anywhere on the globe. Workers in South Korea, Taiwan, or Mexico can churn out turbo-charged automatic vacuums just as well as American workers can, and for far lower wages. Indeed, today Henry is as likely to license a South Korean or Taiwanese company to manufacture the vacuum as he is to sell out to Westinghouse. If Westinghouse does get hold of Henry’s Big Idea, it is apt to build its own factory overseas.
In a world where routine production is footloose and billions of potential workers are ready to underbid American labor, competitive advantage lies not in onetime breakthroughs but in continual improvements. Stable technologies get away. Keeping a technology requires elaborating upon it continuously, developing variations and small improvements in it that better meet particular needs.
The Japanese are skilled at using the essential insight of a Big Idea as the starting point in such an ongoing process of elaboration and refinement. To return to Henry’s Big Idea, the Japanese knack for modification and improvement would likely foreshorten the profitable life of the turbo-charged automatic vacuum cleaner. Within one or two years after it hit the market, Hitachi would likely come up with a super turbo-charged automatic vacuum cleaner that can be programmed to suck up all the dirt and dust in your whole house in one minute flat. Anticipating this, Westinghouse is likely to sell the patent on the basic turbo-charged vacuum to Hitachi early on in return for an exclusive license to sell and distribute Hitachi’s super version in the United States. Thus Westinghouse would be able to offer American consumers a full line: cheap, basic turbo-charged automatic vacuums made in South Korea, and super turbo-charged automatics made in Japan.
The specific example may be fanciful, but the general tendency is not. Americans continue to lead the world in scientific discoveries and Nobel laureates. But we often have had difficulty elaborating upon our inventions and turning them into streams of commercial products. We get bogged down somewhere between big breakthroughs and their applications. Americans invented the solid-state transistor. Then, in 1953, Western Electric licensed the technology to Sony for $25,000, and the rest is history. A few years later RCA licensed several Japanese companies to make color TV production in America. Routine assembly of color TVs eventually was shifted to Taiwan and Mexico, but Sony and other Japanese companies continuously refined the technology into an array of continuously evolving consumer electronic products.
In 1968 Unimation licensed Kawasaki Heavy Industries to make industrial robots; the nascent American robotics industry never quite recovered, as the Japanese developed ever more elaborate versions. Americans came up with the Big Ideas for video-cassette recorders, basic oxygen furnaces and continuous casters for making steel, microwave ovens, automobile stamping machines, computerized machine tools, and integrated circuits. But these Big Ideas—and many others—quickly found their way into Japanese production, where they were steadily improved upon, step by step.
In this integrated world economy, Americans must live by their wits. If they hope to command a premium wage, their labor must generate more value. They must produce goods that continuously embody new innovations that workers in less complex and more distant economic systems cannot easily or quickly match. Such products may comprise standard components put together in unique ways, like a communications network custom-designed for a particular corporation. Or they may incorporate standard hardware but specialized software, like an “intelligent” credit card that lists credit balances, or a work station for developing and testing alternative engineering approaches. Or the products may involve particularly high standards of precision. Or they may require custom-tailoring, like made-to-order semiconductor chips, or special chemicals prepared for particular end uses. (I have elsewhere referred to these products and the means of producing them as “flexible systems.”)3
Where innovation is continuous, and products are ever more tailored to customers’ particular needs, the distinction between goods and services begins to blur. Thus when robots and computerized machine tools are linked through software that allows them to perform unique tasks, customer service becomes a part of production. When a new alloy is molded to a specified weight and tolerance, service accounts for a significant part of the value added. IBM is in the business of providing bundles of high-technology goods and services. Of its 400,000 employees, only a comparative few are engaged in “manufacturing” in the traditional sense. The rest provide services—putting together packages of hardware, software, financing, and maintenance. What IBM and other advanced companies provide are best described as “service goods.” Reports that American workers can no longer compete in manufacturing and must shift to services are only half-right. More precisely, they can keep high wages only by producing goods with a large component of specialized services or, to state the same thing differently, providing services integral to the production and use of specific goods.
The point is this: In the new global economy, nearly everyone has access to Big Ideas (and the machines and money to turn them into standardized products) at about the same time, and on roughly the same terms. The older industrial economies have two options: They can try to match the wages for which workers elsewhere are willing to labor. Or they can compete on the basis of how quickly and well they can transform ideas into incrementally better products. This second and obviously more appealing option implies a fundamentally different approach to entrepreneurialism. Instead of a handful of lone entrepreneurs producing a few industry-making Big Ideas, innovation must be more continuous and collective.
To compete on the basis of rapid improvements in product and process, rather than on the basis of the scale economies of mass production, means a new emphasis on the innovative skills of workers—the productive services they deliver—and on the organizational structure of production. Consider these evolutionary paths followed by successful firms (a few American, mostly Japanese or the Japanese parts of Japanese-American corporations): Vacuum-tube radios become transistorized radios, then stereo pocket radios audible through earphones, then compact disks and compact disk players, and then optical-disk computer memories. Color televisions evolve into digital televisions, capable of showing several pictures simultaneously; videocassette recorders into camcorders. A single strand of technological evolution is embodied in electronic sewing machines, then in electronic typewriters, and then in flexible electronic work stations. Basic steels give way to high-strength and corrosion-resistant steels, and then to new materials composed of steel mixed with silicon and custom-made polymers. Basic chemicals evolve into high-performance ceramics, to single-crystal silicon, and high-grade crystal glass. Copper wire gives way to copper cables, and then to fiberoptic cables.
These patterns reveal no clear life cycles with beginnings, middles, and ends. Instead of Big Ideas that beget standardized commodities, these series display a continuous process of incremental change and adaptation. Workers add value not solely or even mostly by tending the machines and carrying out routines, but in the analysis, experimentation, and the application of creativity. They are paid for the services they embody in goods. In this context, it makes no sense to speak of an “industry” like steel or automobiles or televisions, or even banking, because there are no clear borders around any of these service-goods. When products and processes are so protean, firms grow or decline not with the market for some specific good, but with the adaptive capacity of their work force. One thing leads to another. Producing the latest generation of automobiles involves making electronic circuits that govern fuel consumption and monitor engine performance; improvements in these devices lead to improved sensing equipment and software for monitoring heartbeats and moisture in the air. Producing cars also involves making flexible robots for assembling parts and linking them by computer; steady improvements in these technologies, in turn, lead to expert production systems that can be applied anywhere. What is considered the “automobile industry” therefore is really a variety of technologies evolving toward all sorts of applications that flow from the same strand of technological development but toward different markets.
And because there are no preordained life cycles, and machines and workers are not locked into producing long runs of any single standardized good, the firm has less need to hedge its bets. Experimentation and development are occurring constantly, so it is not necessary for the firm to leap into deliberately unrelated lines of business as insurance against declining demand in any one.
This pattern of ongoing, incremental evolution depends on the cumulative expertise of a great number of people. Workers within the firm garner experience in the process of product design, fabrication, engineering, and production. Expertise outside the firm itself is also important; suppliers of key components share intimately in the evolution. The path of change is governed by continuous feedback from consumers, sometimes spread around the globe, whose needs and priorities shift over time.
There are no sharp distinctions between goods and services, between invention and production, between entrepreneurs and drones. Because production is a continuous process of reinvention, entrepreneurial efforts must be focused on many thousands of small ideas rather than a few big ones. As the speed and precision of response become more important, so does the potential of each member of the enterprise as a source of innovation. Workers who design the software, make the hardware, or sell and service the resulting product are cultivated as fonts of valuable up-to-the-minute information about how things can be improved. Because the information and expertise are dispersed throughout the organization, top management does not solve problems nor set specific direction; it creates an environment in which people can identify and solve problems for themselves.
Collective entrepreneurialism requires close working relationships among people at all stages of the process. If customers’ needs are to be recognized and profitably met, designers and engineers must be familiar with marketing and sales. And salespeople must have an intimate understanding of the enterprise’s capacity to design and deliver specialized products. The system can adapt quickly to new opportunities only if information is widely shared, and everyone involved is actively looking for opportunities to improve, adjust, and upgrade.
Most of the training for this type of work can only occur on the job. Formal education prepares people to absorb and integrate experience; it does not substitute for experience. This is because the precise skills to be learned cannot be anticipated in advance. Opportunities cannot be foreseen. Any production process that can be taught on the basis of textbook procedures can be moved to low-wage areas or programmed into robots and computers.
Individual skills are integrated into a group whose collective capacity to innovate becomes something more than the simple sum of its parts. Over time, as group members work through various problems and approaches together, they learn about each others’ abilities. They learn how they can help one another perform better, who can contribute what to a particular project, how they can best gain experience together. Innovation is inherently collective and incremental. Each participant appreciates what the others are trying to do; he is constantly on the lookout for small adjustments that will speed and smooth the evolution of the whole. The net effect of many such small-scale adaptations, occurring throughout the organization, is to propel the enterprise forward. Because such cumulative experience and understanding is so critical, this network of people must be maintained over time. Their collective capacity, by assumption, cannot be translated into standard operating procedures and transferred to other workers. The capacity to add value resides in the whole, and this capacity is what the enterprise seeks to preserve and develop. This is in sharp contrast to the pattern in standardized production, in which drone workers are seen as interchangeable, and extruded when the industry or product approaches the end of its life cycle.
Enterprises designed primarily to reduce the cost of mass-producing Big Ideas are organized into a series of hierarchical tiers, so that each superior can ensure that subordinates are acting according to plan. But enterprises designed to continuously discover and apply incremental advances have a relatively flat structure. Here it is far less important that workers follow preordained rules than that they gain new insights into how products or processes can be improved. Coordination is achieved both through common experience—working together long enough so that signals are relatively clear—and through common understandings about what sorts of small-scale refinements are likely to improve products and processes. There are thus few middle-level managers and only modest differences in the status and income of senior managers and junior employees. Individual performance cannot be monitored and evaluated through simple accounting systems, because the quality of work is often more important than the quantity. Tasks are often so intertwined, moreover, that it becomes impossible to evaluate them separately. Since each worker necessarily relies on many others, success can be measured only in reference to collective results.
In this very different world of work, automation poses no threat to employment, as some liberals have feared. Computers are used less to reduce the cost of labor than to enhance its value. Rather than simplify and standardize jobs by preprogramming every task, computer-based information provides a means of expanding workers’ discretion. It gives workers more feedback about what they are doing and how what they do affects other aspects of the production process. Computerized data might reveal, for example, that a particular component, if slightly modified, could exactly meet the needs of a product on a different line, thereby reducing the number of parts that need to be produced overall. Or the data might show that one piece of equipment is using up a great deal of energy, and could be redesigned to be much more efficient. Or that with a slight alteration in pressure or composition, a batch of material could be put to very different end uses. These sorts of information help workers discover ways to improve product and process. They give them the means to use their imaginations—to experiment in rearranging the data to provide new insights into what is being produced and how it can be refined.4
Collective entrepreneurialism is not as unusual as it may at first seem. Rarely do even Big Ideas emerge any longer from the solitary labors of genius. Modern science and technology is too complicated for one brain. It requires groups of astronomers, physicists, and computer programmers to discover new dimensions of the universe; teams of microbiologists, oncologists, and chemists to unravel the mysteries of cancer. With ever more frequency, Nobel prizes are awarded to collections of people. Scientific papers are authored by small platoons of researchers.5
Nor is collective entrepreneurialism the sole province of Japanese enterprise. Some corners of American enterprise, too, have long been premised on the ideal of collective entrepreneurialism. This is true for professional partnerships—lawyers, doctors, accountants, management consultants, architects, investment bankers—for which the value of the enterprise rests almost entirely in the knowledge and experience of its members. Within the most successful and innovative of these enterprises there is little hierarchy and few established routines; all members have a stake in improving the performance of the entire group.
Collective entrepreneurialism also has characterized many small firms producing service-intensive goods. Coalitions of designers, engineers, fabricators, marketers, and salespeople and financial specialists race to get new product generations to market. A comparable pattern could be observed for a time in some of the notable geographic centers of technology-based production. In the late 1970s and early 1980s the areas around Route 128, which encircles Boston, and California’s Santa Clara County both represented diffused but often effective entrepreneurial networks. Technical specialists tended to be familiar with one another’s work and knew who to tap for help when a specific sort of problem arose. Firms and projects came and went, but the underlying network evolved and developed, at least for a time. Shared experiences were cultivated and preserved, and technological competence accumulated. (The divergence between private and social returns on these activities, however, eventually would threaten these learning communities, as will be recounted shortly.)
But many Americans have continued to work in enterprises that conform to the standard pattern, in which drones mass-produce Big Ideas—be they automobiles or hamburgers. Why? If collective entrepreneurialism offers higher returns in the future, why would Americans cling to the old ways?
Consider: In 1985, soon after the Reagan administration arranged for quotas on the importation of foreign steel, the U.S. Steel Corporation dropped plans for new investment in a Utah facility. Instead, it opted to import semifinished slabs from South Korea to feed its West Coast finishing mills. Soon thereafter it spent $3.6 billion to purchase Texas Oil and Gas Corporation, on top of the $6 billion it spent a few years before to buy Marathon Oil. In mid-1986 it dropped “steel” out of its name and became USX—with the last letter serving as an indelible reminder that what the corporation now stood for was unknown and unknowable. By that time energy accounted for two thirds of its revenues and all of its profits, and thousands of workers had lost their jobs.
The ensuing political debate centered, as usual, on the benefits and the pains of economic change. Unionized workers, and not a few liberals, complained that U.S. Steel was abandoning steel, and so it was. They lamented the resulting unemployment of steel workers and the decline of traditional steel towns. On the other hand, conservative disciplinarians pointed out, correctly, that there was no future in making basic steel. South Korea’s Pohang Iron and Steel Company, for instance, operated one of the most modern mills in the world, which generated over 9 million tons of steel a year; Pohang’s workers earned an average of $2.50 per hour, or about a tenth of U.S. Steel’s pay scale.
Another example from the opposite end of the industrial spectrum: In the 1970s the Zenith Corporation invested several hundred million dollars trying to implement a potentially revolutionary Big Idea: using lasers to play sounds recorded on a plastic disk. The lasers would “read” information encoded and compactly stored on the disk and reproduce sounds far more faithfully than conventional tapes or records. But by the end of the decade, Zenith had abandoned the effort. Production was simply too risky and expensive. Zenith opted to import videocassette recorders—a comparable but simpler technology—to sell under its own brand name. Sony, meanwhile, soon introduced the first successful minisized, laser-operated compact disk player, which swept the American market.6
Both U.S. Steel and Zenith made rational calculations of the cost of pursuing a market and, following the logic of standardized mass production, opted out. In principle, however, each had other options. U.S. Steel could have eased out of steel and into new alloys and plastics that combine high strength with light weight. Or it could have moved into advanced ceramics that resist corrosion and heat, or into any number of other new materials that do what steel does but better or cheaper. In most of these areas, no foreign producer was yet ready to compete. U.S. Steel could then have maintained its marketing links to its customers making cars, buildings, and appliances; American automakers, for example, were beginning to turn to Japan for ceramic engines and carbon-fiber chassis. Had U.S. Steel moved in this direction, it could have retrained many of its workers—already skilled in making one kind of durable material—to meet the same needs with new products. It would have become U.S. Advanced Materials, a robust descendant of its former self. Zenith, likewise, could have regarded the laser disk not just as one potential product but as the wellspring of a stream of potential products flowing out of the collective experience gained by making the first—items like optical computer memories, disks containing information services, videodisks that could be erased and revised. In this way, Zenith too could have evolved as its work force, and its surrounding network of suppliers and customers, also evolved.
Yet neither firm followed any such path. What are we to conclude from this? One possibility is that the notions of high-value service-goods and collective entrepreneurship are pipe-dreams, and that the only realistic options for most American workers are protection, idleness, or wages as low as their competition abroad. Another possibility is that the managements of U.S. Steel and Zenith were simply too blind to spot the sources of future profits. Neither of these explanations holds true, however, for Zenith and U.S. Steel nor for the many other American companies who cling to the logic of standardized mass production and balk at a strategy of collective entrepreneurialism. The problem is rooted in a deeper dilemma, to be taken up in the next chapter.