The past 30 years have seen amazing progress in computer technology, and that progress has transformed a wide variety of American industries. Yet that same 30-year period has been a disappointment when it comes to the incomes of ordinary Americans.
A new book by economist James Bessen argues that it's not surprising for new technologies to be associated with stagnant wages. Indeed, something similar happened in America's first high-tech boom: the textile industry of the mid-1800s. From 1830 to 1860, cloth production skyrocketed; wages barely budged.
But then weavers' wages started rising. By 1900, they were more than twice their level from 40 years earlier. Bessen argues that this and other historical examples offer important lessons about the state of the labor market today.
Some people, such as economists Erik Brynjolfsson and Andrew McAfee, have portrayed a future in which computers destroy more and more jobs, leaving millions out of work. Bessen is skeptical. Computers obviously do eliminate some jobs, but they almost always create other jobs in the process. The tricky thing is that these new jobs often demand different skills, and workers are struggling to keep up. Still, Bessen paints a basically optimistic view of a future in which technologies mature and create new, higher-paying jobs for ordinary workers.
Computers aren't destroying jobs (on net)
In recent years, there's been a lot of worry that increasingly sophisticated computers would replace more and more jobs, which could lead to a future where many low-skilled workers are unable to find a job at all.
But Bessen argues that this is a misunderstanding of recent economic trends. "I just don't see evidence of it," he says. "There are very few occupations where everything has been automated."
Bessen points to bank tellers as an example. During the 1990s and 2000s, banks installed thousands of automated teller machines. Yet surprisingly, the number of human tellers actually grew slightly during the same period, as this chart from the Atlantic shows:
Bessen argues that this is not an isolated example. "Voicemail systems and automated telephone systems dramatically reduced the jobs for telephone operators, but the number of receptionists increased," he says. Government data shows that from 1999 to 2009, the economy lost 162,000 typists and 102,000 switchboard operators. But during that same period, there was an increase of 215,000 secretaries and 64,000 receptionists and information clerks. That adds up to a net gain of about 16,000 jobs.
In part, these workers handle tasks that are still too complex for automation. But these jobs have also persisted because people find it useful and pleasant to talk to other human beings. A human teller can answer basic questions about a bank's services better than any computer program.
This isn't just an abstract issue for Bessen, an economist whose empirical research has had a big impact on the patent reform debate (and who created a card stack for Vox last year). In the 1980s, before embarking on a second career as an economist, Bessen founded a company that built one of the first desktop publishing systems. As he watched his software make some jobs in the publishing industry obsolete even as it created many others, Bessen became fascinated by the question of how technology transforms the labor market.
How technology creates new jobs
Bessen says today's technological changes aren't so different from those that happened in the past. "Over the course of the 19th century, technology took over 98 percent of the tasks required of a weaver," he says. "But that 2 percent became more valuable."
Bessen sees similar trends occurring all around him. For example, in 2013, 60 Minutes did a feature on a Massachusetts company that uses robots rather than humans to move products around its warehouse.
"They figured out that each robot was doing the work of one and a half workers. And 60 Minutes concluded that this means we're just headed for technological unemployment."
But Bessen doesn't buy it. "If you talk to people in the industry, the managers of these warehouses, they're saying they're having difficulty hiring enough workers who have the appropriate skills"
There are two big reasons for this. The obvious one is that introducing robots into warehouses creates new, high-skilled jobs building, programming, and repairing robots.
But there's also a more subtle effect that parallels the experience of the textile industry 150 years ago, where cheaper cloth induced consumers to buy a lot more clothing. As technology has made warehouses more efficient, companies have demanded a lot more from them.
For example, Bessen says, "there's been a huge increase in the variety of items stocked in a store. Today's supermarkets carry over 50 times as many items as the grocery stores of 80 years ago. That's been enabled by all these various systems that can keep track of things and ship them and restock them when inventories are getting low."
The astonishing variety of a modern grocery store would have been impossible using the information technologies available in the 1960s. That complexity means there are a lot more warehouse jobs than there would have been if companies had simply computerized the simpler supply chains they had in past decades.
Bessen argues that the same dynamic was at work with growing bank teller employment during the 2000s. ATMs made it cheaper to open a bank branch, so banks opened more branches, which created additional jobs for tellers.
Standardization gives workers leverage
If technology is constantly creating new jobs, why aren't we seeing more employment and wage growth? Bessen argues that many workers today face the same problem textile workers faced in 1845: until technology is standardized, it's difficult to profit from investments in new skills.
Early textile companies built their looms in slightly different ways, and they were constantly tinkering with them. These differences made job-hopping difficult. Someone who had mastered one company's equipment wouldn't necessarily be more productive at other mills in town. Each employer knew that other mills wouldn't pay a big premium to hire an experienced worker because they'd still have to spend months training her to use their equipment. And that gave workers little leverage in wage negotiations.
The strong bargaining power of mill owners provided healthy profits in the industry's early years. But the situation wasn't great for owners, either. Weavers' low wages and lack of bargaining power meant that few treated it as a profession. Employers faced high costs every time they hired a new worker. A persistent shortage of skilled workers also limited opportunities to expand production.
Everything changed after the Civil War, as mill owners began to standardize their equipment and production processes. Bessen writes that the "New England Association of Cotton Manufacturers was formed in 1865 to exchange technical and management knowledge. The first technical school for textile managers and workers opened in Philadelphia in 1884." As weaving technology became standardized, workers could move from one job to another and take their skills with them. That gave them leverage to demand wages that reflected the value of their work.
This isn't an isolated case. Bessen points out that the market for typists took off around 1900, when the QWERTY standard became the dominant keyboard layout. Once all typewriters used the same keyboard layout, women (they were almost all women) became a lot more willing to pay for typist training knowing that they'd be able to choose from a wide variety of employers.
Bessen argues that many industries today suffer from a similar problem. Rapid technological change means that workers' skill investments become obsolete quickly. Bessen points to graphic designers as an example. "Desktop publishing replaced typographers in the 1980s. Graphic designers had to learn new skills. Then they had to learn web publishing skills. Then they had to learn mobile publishing skills."
When will wages start rising again?
The most talented graphic designers have thrived in the digital age. Employers are willing to pay a premium for people who are able to master cutting-edge technologies. But ordinary designers have fallen behind.
And the rapid pace of change discourages workers from investing in skills. "The graphic design schools have a hard time keeping up," Bessen says. Why spend thousands of dollars learning Flash design skills that might be totally obsolete five years later?
So when will wages start to rise again for average workers? Bessen's theory suggests that it depends on how long it takes for new technologies — like online publishing and supply-chain management — to mature and standardize. Once that happens, it will become easier for ordinary workers to gain skills, for schools to teach them, and for workers to earn a living from them over long periods.
But that might take a while. In past periods of history, new technologies tended to affect relatively narrow slices of the economy. By contrast, computers are changing almost every industry, and there's reason to expect more dramatic changes ahead.