The Washington PostDemocracy Dies in Darkness

Will robots steal our jobs? The humble loom suggests not.

Early 20th century, Lowell, Massachusetts, USA --- Mills and smokestacks line the Merrimack River in Lowell, Massachusetts. (Corbis)

Two hundred years ago the Industrial Revolution came to America on the banks of the Charles River in Waltham, Mass., where in 1814, the Boston Manufacturing Company built the first integrated textile mill.

The mill was the brainchild of Francis Cabot Lowell, a Harvard graduate in his 30s. Soon, mill towns dotted the banks of rivers across New England - a cluster of innovation that was the Silicon Valley of its day.

Realizing that smarter workers could master complex weaving equipment more quickly, mill owners sought to attract the well-educated daughters of middle-class farmers. Two centuries before Google offered gourmet meals and on-site medical care, mill owners built libraries, schools and lecture halls to keep workers intellectually engaged.

The revolution Lowell helped to launch would reshape the American economy. Steadily improving production techniques allowed the United States to surpass England and, more important, eventually to create substantial wealth for ordinary workers.

Today some people worry that this experiment has run its course. After all, despite continued technological progress, median wages haven't grown in three decades.

Lessons from the Industrial Revolution, however, suggest that today's Information Revolution could yet yield similar gains. Early soaring productivity that made factory owners and managers wealthier took decades to reach ordinary factory workers - but in time brought large wage increases.

Innovation is not the only factor influencing wages - policy matters, too - but, if history is a guide, technological progress can provide growing wealth for ordinary workers once again.

Power looms come to America

In the early 19th century, America's weaving industry was technologically backward. Cloth was woven slowly on handlooms, mostly as a part-time occupation in farm households.

While on a family vacation in England and Scotland, Lowell visited textile mills and saw cloth woven on automated power looms. He studied them, memorizing sufficient details of their construction so that, with his help, the talented American mechanic Paul Moody was able to build a working power loom.

In a traditional hand loom, a long set of threads is stretched lengthwise and alternate sets of threads are raised or lowered. The weaver propels a shuttle carrying another thread across the loom, through the opening created by the raised threads, and then lowers the raised threads so they lock the alternate thread into the cloth.

The power loom automated these basic steps, but weavers still needed to perform tasks such as filling the shuttle when it ran out of yarn, monitoring the cloth for defects and fixing breaks in threads. This partial automation meant that a single weaver could tend two looms, increasing the output per worker.

Once Moody got the power loom running, he developed a host of complementary machines and the Boston Manufacturing Company quickly became highly profitable. Lowell died in 1817, but his heirs built mills throughout New England. The first expansion took place at the company town of Lowell on the Merrimack River, named in his memory.

New skills for new technology

When Charles Dickens visited Lowell in 1842, he reported back to his English readers three facts that he thought many of them would find "preposterous:" young girls who worked in the mills played the piano, they subscribed to circulating libraries and they published their own literary magazines. To the class-bound English, such activities were "above the station" of factory workers.

British readers expected mill workers to come from the lower classes because the first British mills sought the cheapest labor. This lowest-common-denominator approach wouldn't have worked in Lowell. While much of the machinery was copied from Britain, the mills were organized differently. In Britain, specialized workshops produced a variety of cloth goods, many of fine quality. In contrast, in Lowell all of the operations involved in turning raw cotton into finished cloth were conducted in one integrated facility. That allowed the production of a highly standardized product in large quantities.

Coordinating all aspects of production under one roof required specialized technical and business skills. The Waltham mill was one of the first business organizations to use professional managers, called mill agents, who were separate from stockholders (though many mill agents also owned stock).

Waltham mills also required a different kind of worker. Mass production demanded training on a large scale and the new technology demanded new skills. In the British craft workshops, sons often learned as informal apprentices to their fathers. But apprenticeships couldn't quickly train the large numbers of workers the mills required. And technology was changing too quickly for formal classroom training to be practical. Instead, Lowell and his partners sought to recruit intelligent workers who could learn quickly from experience on the job.

This is why American mill owners encouraged the cultural enrichment activities Dickens found so "preposterous." The mill owners built the boarding houses, schools, churches, a Lyceum for lectures, the circulating library, a savings bank and a hospital.

Lowell marked a bold social experiment for a society where, not so long before, the activity of young unmarried women had been circumscribed by the Puritan establishment.

His approach to hiring workers turned out to be critical to the success of the new technology. An experienced weaver could weave over twice as much cloth in an hour as a new recruit. This bolstered the bottom line. In areas where early integrated mills did not hire a labor force who could learn quickly, the mills often failed.

Waiting for wages

At first wages were low and stagnant for decades. But by the end of the 19th century, Lowell's weavers earned more than twice what they earned per hour in 1830, after taking inflation into account, and they earned much more than workers with lesser skills. What changed? A labor market for skilled weavers developed. After the Civil War, a weaver at one mill could reliably expect to get a job using her skills at another mill. The mills mainly hired experienced weavers from a large pool of weavers who lived in or near Lowell. Because the mills had to compete to hire skilled weavers, they paid for that skill.

New technology demanded new skills, but the workers who acquired those skills were not necessarily compensated for them, at least not at first. Good pay required labor markets for the new skills. And labor markets required skills that were standardized among many competing employers.

There is good reason to think something similar is happening in the American economy today. Labor markets for some new skills are also slow to develop, although for different reasons. Sometimes regulatory changes are needed. For example, changes in medical technology allow mid-skill providers such as nurse practitioners, dental hygienists and diagnostic sonographers to deliver more health care. But in most states licensing restrictions limit these opportunities in favor of traditional care provided by doctors and dentists. Occupational licensing legislation has increased the share of workers subject to licensing restrictions from 18 percent in 1980 to 29 percent in 2008. These restrictions inhibit the development of labor markets.

Sometimes the pace of technological change itself delays robust labor markets. Consider, for example, publishing technology. In the 1970s and '80s, computerized desktop publishing replaced typesetters with graphic designers. Graphic designers learned new skills, but those skills have been constantly changing. Print designers had to acquire Web skills, and Web designers had to acquire mobile skills. Standards keep changing, too. A few years ago, designers needed to learn Flash; now they learn HTML5.

The most talented designers teach themselves and earn high pay, often as freelancers. But the average designer cannot keep up, and employers have little way to know what prospective hires have learned. While the best designers are in high demand, the pay for the average designer has not grown in decades. The median designer today earns slightly more than a typesetter earned in the 1970s.

In the 19th century, the maturation and standardization of weaving technology finally brought long-overdue raises to textile workers. Perhaps something similar will happen in the coming decades: As skills in the Internet publishing industry stabilize, it will be easier for ordinary workers to develop skills they can take to another employer. And that could give them bargaining power to seek higher wages.

The end of work?

Karl Marx saw England's impoverished factory workers as evidence that machines were replacing workers, throwing them into unemployment and poverty. For example, the automated power loom took over tasks formerly done by handloom weavers. Over the 19th century, weaver's tasks were progressively automated.

New inventions automatically straightened the edges of the cloth, detected when threads broke or replaced the yarn in the shuttle when it ran out. These changes and other incremental improvements reduced the time it took to weave a yard of cloth. As a result, weaving a yard of cloth at the end of the century took only 2 percent of the human labor it took to do so on a handloom at the start of the century; machines did the rest.

Marx observed this automation and predicted that it would result in mass unemployment. But that's not what happened. In fact, by the end of the century, there were four times as many factory weavers as in 1830. What Marx missed was that the new technology also increased demand. The greater output per weaver reduced the price of cloth. Consumers reacted by buying more cloth. Greater demand for cloth meant more jobs for weavers despite the automation.

Even so, as the market for cotton cloth matured in the mid-20th century, further automation no longer generated such increases in demand and new technology slowly eliminated jobs in the textile industry.

Today, machines are taking over tasks not only in manufacturing industries such as textiles, but also in white-collar jobs, again raising fears of unemployment.

According to 60 Minutes, "Bank tellers have given way to ATMs. Sales clerks are surrendering to e-commerce. And switchboard operators and secretaries to voice recognition technology," arguing that digital technologies are leading to persistent unemployment. But, in fact, there are more bank tellers, sales clerks and receptionists and secretaries in 2009 than in 1999, according to the Bureau of Labor Statistics. The reason: demand.

For example, it takes fewer bank tellers to operate a bank branch, thanks to the ATMs. This makes it less costly to operate a bank branch, allowing banks to open more of them. With more branches, banks can expand their markets. But more branches mean greater demand for tellers, offsetting the loss in the number of tellers per branch. Bank tellers today perform different tasks than in the past - they do fewer simple jobs like counting cash and more of the customer interaction of "relationship banking." These tasks require different skills, but ATMs have not eliminated teller jobs.

The experience of America's textile industry demonstrates that technology sometimes interacts with society in surprising and counterintuitive ways. From the beginning, social experiments have been integral to the introduction of major new technologies, which require new skills learned through experience. And learning on the job often requires new ways of organizing a workforce, new occupations and new labor markets.

Even with the right policies, these social changes can take time to work out. So while new inventions can come into use relatively quickly, it may take decades of slow learning and occupational changes before the benefits of major new technologies are shared by large numbers of ordinary workers.

James Bessen does research on technology and innovation at Boston University School of Law. He is the co-author of "Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk." He is currently writing "Learning by Doing," a new book about technology and jobs. You can follow him on Twitter.