Learning by Doing

The Real Connection between Innovation, Wages, and Wealth

James Bessen

• Publisher: Yale University Press
• Year: 2015
• ISBN:
  9780300213645
• Topic: Volkswirtschaftslehre
• Subtopic: Arbeitsmarktforschung

"Bessen sets out to refute the arguments of... techno-pessimists, relying on economic analysis and on a fresh reading of history" ( The Wall Street Journal ). Technology is constantly changing our world, leading to more efficient production. But where once technological advancements dramatically increased wages, the median wage has remained stagnant over the past three decades. Many of today's machines have taken over the work of humans, destroying old jobs while increasing profits for business owners and raising the possibility of ever-widening economic inequality. Here, economist and software company founder James Bessen discusses why these remarkable advances have, so far, benefited only a select few. He argues the need for unique policies to develop the knowledge and skills necessary to implement rapidly evolving technologies. Currently, this technical knowledge is mostly unstandardized and difficult to acquire, learned through job experience rather than in classrooms, but labor markets rarely provide strong incentives for learning on the job. Basing his analysis on intensive research into economic history as well as today's labor markets, Bessen explores why the benefits of technology can take decades to emerge. Although the right policies can hasten the process, policy has moved in the wrong direction, protecting politically influential interests to the detriment of emerging technologies and broadly shared prosperity. This is a thoughtful look at what leaders need to do to ensure success not only for the next quarter, but for society in the long term. "Everyone agrees that education is the key to wage growth. But what kind of education?... This enlightening and insightful book... shows that economic history can provide some useful and surprising answers." —Hal Varian, chief economist at Google

PART I

TECHNOLOGY

WHAT DETERMINES WHETHER AND HOW TECHNOLOGY will benefit large numbers of ordinary people? Social institutions and investments influence the pace of technological change, but they alone are not sufficient for understanding the effects of technology on wages and wealth. Chapters 1 through 4 present a view of technology as knowledge, arguing that the dynamics of knowledge are key to understanding the interaction between technology and society. Many people need to acquire technical knowledge in order to implement major new technologies. Developing new technical knowledge on a large scale is often a difficult social problem. The way societies solve this problem changes over time, and these changes shape how technology is adopted, how fast that change takes place, how new wealth is generated, and how that wealth is shared.

CHAPTER 1

More Than Inventions

IN THE EARLY 1840S, LUCY LARCOM BLOSSOMED as a poet. She took courses in German and botany; attended lectures by such leading thinkers as John Quincy Adams, Edward Everett, John Pierpont, and Ralph Waldo Emerson; and avidly read books from the circulating library on a variety of subjects. Still, it was as a poet—reading, writing, and discussing—that she excelled. She published dozens of poems in the Lowell Offering, one of several girls’ literary magazines. Some of them were reprinted in national periodicals and in collections edited by Henry Wadsworth Longfellow and other leading American poets. She attracted the attention and support of John Greenleaf Whittier, who became her mentor and friend. These years launched her career as a writer and teacher.

Lucy Larcom was not a college student. She was a factory worker in Lowell, Massachusetts, the home of some of the first integrated textile mills in the United States. The Merrimack River Valley, where Lowell is located, was a center of the U.S. Industrial Revolution, a nineteenth-century Silicon Valley.

Most mill girls came from farming families from across northern New England, and they stayed in boardinghouses in Lowell during their time in the mills, which often lasted only a year or so. Lucy’s family was different in that they lived in Lowell. Lucy was born in nearby Beverly, Massachusetts, the daughter of a sea captain and one of nine siblings. Her father died when Lucy was eight, and her mother moved the family to Lowell to run one of those boardinghouses. At eleven, Lucy left school and went to work in a mill to help out her family.

But she did not want to abandon her education. She resolved to “learn all I could, so that I should be fit to teach or to write, as the way opened. And it turned out that fifteen or twenty of my best years were given to teaching.”¹ She taught in Illinois and at Wheaton Seminary in Massachusetts, now known as Wheaton College, where a dormitory is named after her.

Yet, although Lucy was an outstanding poet and writer, her efforts to learn were not exceptional. As she wrote in 1889,

For twenty years or so, Lowell might have been looked upon as a rather select industrial school for young people. The girls there were just such girls as are knocking at the doors of young women’s colleges to-day. They had come to work with their hands, but they could not hinder the working of their minds also. Their mental activity was overflowing at every possible outlet.… They were improving themselves and preparing for their future in every possible way, by purchasing and reading standard books, by attending lectures, and evening classes of their own getting up, and by meeting each other for reading and conversation.

This was not an accident. The mill owners had expressly designed the city of Lowell to be the kind of place where, in the words of one investor, “the daughters of respectable farmers were readily induced to come into these mills for a temporary period” through the lure of a rich educational, cultural, and religious environment.²

In 1816, Francis Cabot Lowell established a very successful cotton mill in Waltham, Massachusetts, using Yankee farm girls to operate power looms, an invention he had copied from British models. Following his death the next year, his investors sought to build a new city, named in Lowell’s honor, on the site of a small farm village that had an excellent source of water power to run new mills. But the mill owners built more than factories. They also built boardinghouses for the mill girls, to be run by moral women. They made sure that every major Protestant denomination had provisions for worship. And they encouraged and supported a variety of other institutions, including twenty-three schools, the Lowell Institute, a Lyceum for lectures, a circulating library, a savings bank, and a hospital.

The owners did all this because they wanted intelligent and morally disciplined workers to run the looms. The girls they hired were not poor women desperate for work; they were the daughters of “respectable” farmers, most of whom had modest wealth or better.³ Although Irish immigrants built the mills and canals and Irish women did domestic service in Lowell’s better-off homes, the girls hired to work the looms during the 1820s and 1830s were almost exclusively of Yankee origin. They were almost all literate, at least to the level of being able to sign their names.

The mill owners’ motives were not purely philanthropic. They needed bright, able workers who could learn how to use this strange new technology efficiently. Mills in other towns had failed because they had not recruited enough high-quality hands. By the 1840s, the weavers in Lowell’s mills were more productive than those in English mills, which hired relatively fewer literate workers, or those in American mills that did not use such a select labor supply.⁴ It would appear that the skills, knowledge, and intelligence of ordinary production workers were critical to the adoption of this new technology.

Implementation

The behavior of the mill owners seems surprising, but it should not be unfamiliar. For similar reasons today, Google offers employees gourmet meals, on-site medical care, and a whole variety of amenities in order to attract talented people and keep them intellectually engaged. Yet the mill owners’ behavior seems surprising because we often forget how difficult it was to implement technologies in the past; we forget that implementation required new skills and knowledge that took time to develop and time to learn, even for mechanical inventions. We tend to focus instead on the original act of invention.

The distinction between invention and implementation is critical and too often ignored. The key invention central to the dramatic rise of textile mills in Lowell and other towns was the power loom, a machine that partially automated the work of weaving, greatly reducing the labor required to produce a yard of cloth. The first commercially successful power looms were operated in the United Kingdom. American inventors began developing power looms around 1810, influenced by British designs. The design that eventually dominated the U.S. textile industry was developed by William Gilmour, an immigrant mechanic who had experience with textile equipment in Scotland. After Gilmour built his first loom in 1817 for Judge Daniel Lyman in Rhode Island, he shared his drawings with other mechanics, and this design was quickly adopted throughout New England.

But the widespread replication of this invention was hardly sufficient to guarantee its efficient use. Economic historian Robert Zevin notes that for at least two decades there was a shortage of people who could build, install, operate, and maintain the new machinery.⁵ As late as 1845, machine shops building textile equipment suffered delays due to shortages of skilled mechanics. Because machine shops making looms and other textile equipment possessed mechanics with specialized knowledge, they were able to earn high profit margins for twenty years. The textile mills earned even higher profits. They too often could not find enough workers who knew how to use the technology efficiently, organize a factory, and train a workforce. Zevin found that mills that lacked a close relationship with a machine shop typically failed. Much of the knowledge needed by mechanics and overseers and loom fixers and weavers was highly mundane, yet for two or three decades, the difficulty of acquiring it limited the ability to operate the new technology on a sufficient scale to meet demand. Given these difficulties, it should be no surprise that the mills went to extraordinary lengths to attract and keep workers who could learn the needed skills. And the most talented mechanics and managers, called mill agents, sometimes received equity in the new companies, as is common practice among tech firms today.

Effective use of the power loom was delayed because the whole of the technology involved much more than just the original invention of the power loom. Implementation on a large scale involved several developments. First, large numbers of people in various occupations had to acquire new, specialized knowledge, skills, and know-how in order to use the technology effectively. Second, the technology itself often needed to be adapted and improved for different applications. These improvements included many secondary inventions, many of these invented by mechanics who learned of new needs and possibilities from practical experience. Third, businesses had to figure out how to best use the new technology, whom to hire, what division of labor to employ, how to organize the workplace, and how to market. Finally, because so many different people in diverse occupations needed to learn new skills and knowledge, implementation required new training institutions and new labor markets that provided incentives to learn these new, specialized capabilities.

All of this took time, yet it was quite important. The invention of the power loom was vital, but so, too, was the implementation. Indeed, the implementation was responsible for most of the economic benefit. Weavers on the first power looms could produce 2.5 times as much coarse cloth per hour as a weaver on a handloom. But over the next eighty years, improvements in the looms and in the knowledge and skills of workers generated a further twentyfold increase in output per hour.

Economic historians have long understood that most of the economic benefit from many major new technologies does not come from the initial commercialization of the original invention but from the eventual implementation.⁶ New knowledge and a long string of improvements follow for decades after the invention, continually altering the technology and the skills needed. In petroleum refining, over a forty-year period, the cost savings achieved during implementation were three times the cost saving realized on the initial installation of new technologies.⁷ Improved efficiency during implementation was responsible for an eightfold increase in the amount of electric power generated per ton of coal from 1900 to 1960. If these examples are typical, then 75 to 95 percent of the productivity gains from many major new technologies were realized only after decades of improvements in the implementation. Incremental improvements and new knowledge also brought major productivity gains in metal cutting, railroad transportation, the steam engine, and the production of rayon. Even the Industrial Revolution itself, despite introducing many important innovations, showed surprisingly low productivity growth for many decades, prompting some economic historians to question whether this period should be termed a “revolution” at all.

None of these gains would have been possible without the initial inventions, but implementation was clearly important, too. Moreover, implementation is slow and difficult for reasons I elaborate over the next several chapters: much new knowledge must be acquired through experimentation or learning by doing (Chapter 2); implementation often involves a long feedback loop where each incremental improvement in the technology requires new skills, which then make further improvements feasible (Chapter 3); and the large-scale acquisition of new skills often requires new training institutions, new standards, and new labor markets (Chapter 4).

Four Distinctions

Much thinking, both popular and scholarly, tends to ignore or downplay the significance and difficulty of implementation, making it hard to grasp how technology is affecting society today. Since implementation is the focus of this book, it is helpful to begin by outlining how consideration of implementation changes things. Four conceptual distinctions are key.

Technology vs. Original Inventions

While scholars realize that implementing a technology involves much more than invention, others are in the grip of the popular heroic view of technology. Museums and textbooks too often highlight a misleading history of Great Inventors who brought wondrous inventions and wealth to the ignorant masses. According to this narrative, the inventions were “designed by geniuses to be run by idiots.”⁸ While the Great Inventors did make important contributions, it took much more than idiots to realize the benefits, including the development of skills and knowledge by large numbers of ordinary working people.

Technology is more than just inventions because implementation involves a lot of mundane technical knowledge and specialized skills, often among diverse people. Some of this knowledge is based on science; for example, semiconductors are based on quantum physics. Engineering knowledge may be required to design a large, efficient chemical plant. The design of an invention itself represents an idea, another kind of knowledge. But not all technical knowledge is so special. Joel Mokyr categorizes useful knowledge into two general types: knowledge of “what,” which is descriptive knowledge of natural phenomena and regularities; and knowledge of “how,” practical knowledge of techniques.⁹ The latter can be very detailed and technology-specific. Much of it is also tacit—that is, not written but learned through experience or by watching others.

Some new inventions really are deployed quickly and easily, but that is not how things seem to work with many of them. Economic historians have highlighted the central role that skilled workers played during the Industrial Revolution.¹⁰ According to Mokyr, they were

an army of mostly anonymous artisans and mechanics, the unsung foot soldiers of the Industrial Revolution whose names do not normally appear in biographical dictionaries but who supplied that indispensable workmanship on which technological progress depended. These were craftsmen blessed by a natural dexterity, who possessed a technical savoir-faire taught in no school, but whose experience, skills, and practical knowledge of energy and materials constituted the difference between an idea and a product. They were mechanics, highly skilled clock and instrument makers, metalworkers, woodworkers, toymakers, glasscutters, and similar specialists, who could accurately produce parts of the precisely correct dimensions and materials, who could read blueprints and compute velocities, and who understood tolerance, resistance, friction, lubrication, and the interdependence of mechanical parts. These were the applied chemists who could manipulate laboratory equipment and acids, the doctors whose advice sometimes saved lives even if nobody yet quite understood why, the expert farmers who experimented with new breeds of animals, fertilizers, drainage systems, and fodder crops.¹¹

Joel Mokyr and Ralf Meisenzahl attribute Britain’s early lead during the Industrial Revolution to these craftsmen and mechanics rather than to her inventors per se.¹² A common observation during the early years of the Industrial Revolution was that the best technology was “invented in France and worked out in England.”¹³

This distinction is important because technologies can take a very long time to perfect. Also, it implies a different view of the role of technology in history. In the Great Inventor acc...