Are machines stealing our jobs? Automated teller machines handle transactions formerly done by bank tellers; accounting software does tasks that bookkeepers used to do; and e-commerce cuts out sales clerks. According to MIT professors Erik Brynjolfsson and Andrew McAfee, technology is causing persistent unemployment and a slow recovery from the Great Recession.
But according to the Occupational Employment Survey conducted by the Bureau of Labor Statistics, there were more bank tellers, more bookkeepers, and more sales clerks in 2009 than there were in 1999—three-quarters of a million more despite the recession. How can this be? The answer is important because it determines what kind of policies will hasten the economic recovery.
At least since Karl Marx, people have been predicting that technology would create mass unemployment. However, these predictions were consistently wrong because they ignored the offsetting benefits of automation. For example, during the 19th century, machines took over tasks performed by weavers, eliminating 98 percent of the labor needed to weave a yard of cloth. But this mechanization also brought a benefit: It sharply reduced the price of cloth, so people consumed much more. Greater demand for cloth meant that the number of textile jobs quadrupled despite the automation.
Something similar is happening in quite a few occupations today. Because ATMs perform many teller transactions, fewer tellers are needed to operate a bank branch. But because it costs less to operate a branch office, banks dramatically increased the number of branches in order to reach a bigger market. More bank branches means more tellers, despite fewer tellers per branch.
Although ATM technology did not eliminate bank tellers, it did change the way tellers work and the skills they require. They now perform relatively fewer simple transactions in favor of complex ones, and they provide the personal service that is an important part of “relationship banking.” When technology eliminates some tasks involved in a job, it makes remaining, related tasks more valuable. Sometimes this greater value can create job growth.
In other cases, technology creates offsetting job growth in different occupations or industry segments. For example, word processors and voice mail systems reduced the numbers of typists and switchboard operators, but these technologies also increased the number of more highly skilled secretaries and receptionists, offsetting the job losses. Similarly, Amazon may have eliminated jobs at Borders and other national book chains that relied on bestsellers, but the number of independent booksellers has been growing and with it, more jobs for sales clerks who can provide selections and advice that Amazon cannot easily match.
But in some cases, the benefits of technology do not offset job losses from automation. This is especially true in mature manufacturing technologies. So even though textile technology created jobs during the 19th century, recent advances in weaving have eliminated jobs. Technology continues to drive down the price of cloth, but lower prices no longer motivate consumers to buy much more clothing. Demand does too little to offset the labor-eliminating effect of automation.
So technology will only generate persistent unemployment when it no longer brings major new goods and services that consumers want. And there is no evidence that this is the case today any more than it was the case 200 years ago.
So don’t blame technology for persistent unemployment. A more likely culprit is the move to austerity in economic policy. The recovery from the Great Recession has been accompanied by government spending cuts and the largest downsizing of government employment in modern history. Moreover, manufacturing—where technology has its greatest negative impact on jobs—has been hurt by offshoring. Technology does place a burden on displaced workers: They often suffer lower wages until they learn the skills to use new technology. The need for new skills might explain why wages are stagnant. But technology is not the cause of the slow economic recovery.