All the same, a series of studies made it increasingly hard to dispute the fact that some workers struggled to adjust. In 1928, before the Great Depression, the Senate Committee on Education and Labor asked the Brookings Institution to inquire into “how many dispossessed laborers were being absorbed by American industry.”8 The study traced the fates of 754 workers who lost their jobs to mechanization across a variety of industries. While 11.5 percent had found a new job within a month and 5.0 percent were still looking for work after a year, the vast majority were out of work for more than three months but eventually found employment elsewhere. Other research similarly showed that transitional costs could be significant. The economist Robert Myers studied 370 displaced cutters in Chicago’s clothing industry during the period 1921–25 and found that the average duration of unemployment was 5.6 months, although after a year 12.9 percent of the former cutters were still without work.9 And workers’ ability to adjust was highly age dependent. Fully 90 percent of the displaced cutters over the age of forty-five either failed to find work or were forced into lower-paying jobs. In contrast, for the most part young adults managed to switch into better-paying jobs. Both studies found that roughly half of displaced workers found new jobs that paid as well as their former ones.10 As in Britain in the Industrial Revolution, in America older people found it especially hard to adjust to new technology. And many workers who found new employment were left worse off economically, at least in the short run.
Unsurprisingly, workers with highly specialized skills faced the most serious adjustment problems. Many musicians, for example, struggled to adapt as the motion-picture branch of the entertainment industry underwent exceptionally rapid technological change. The adoption of sound-producing machines made unnecessary the employment of living musicians in the projecting theater, so that the number of employed musicians declined. In Washington, D.C., the union of musicians and the owners of the motion-picture theaters agreed on a reduction in employment of 60 percent. Part of the decline was offset by the growing demand for musicians by local radio stations. But only a few musicians could make a living in radio broadcasting. Like other displaced worker surveys, a study of a hundred displaced musicians in Washington’s motion-picture theaters showed that the majority experienced a subsequent fall in earnings. On the bright side, the loss of the theater musician was the gain of the motion-picture machine operator. The switch from the silent to the sound movie was “accompanied by an improvement in the status of the motion-picture machine operator, both by replacing the customary boy helper with a licensed operator, and by increasing the average earnings of the projectionists.” According to representatives of five leading moving-picture theaters, more operators were added than the approximately ten thousand musicians who lost their jobs to the sound picture. But even if new types of jobs were added, that provided little relief to the musicians whose skills were not applicable to other employment.11
Displaced worker surveys, it is true, say nothing about the aggregate. The National Research Project on Reemployment Opportunities and Recent Changes in Industrial Techniques, set up to investigate the role of technology in unemployment, unfortunately failed to provide much conclusive evidence. In a 1932 article, David Weintraub, director of the project, reached an optimistic conclusion about the impacts of technical change on employment during the 1920s. However, his later analysis suggested the opposite: mechanization, he found, had been a key factor in unemployment.12 As economists today still struggle to isolate the share of nonemployment attributable to technology, it should be no surprise that research efforts of the 1930s faced similar challenges. At the time, Leo Wolman, who served in the National Recovery Administration during the Great Depression, pointed at several empirical issues limiting progress in studies of technological unemployment, of which many seemed hard to overcome.13
Despite statistical challenges, the emerging consensus among contemporary economists was that there was technological unemployment, albeit of temporary nature. The likes of Paul H. Douglas, Alvin Hansen, and Rexford G. Tugwell all argued that labor market rigidities were impeding the process by which workers would be reabsorbed into new jobs: the expense of moving between locations, the human drags of retraining, and the psychological pressures of job loss made adjustment costly and hard. To ease transitions, Douglas argued against home ownership and too narrow specialization in education and for some form of unemployment insurance and a federal employment agency. In the absence of such policies, he suggested, it will be “almost inevitable that labor will resist and oppose most attempts to raise efficiency of industry.”14
To what extent economists shaped the public discourse is hard to say. Few economists, if any, thought that policies to slow the pace of progress were a good idea. Yet displaced worker studies and the depth of the depression did prompt the adoption of unconventional policies by twentieth-century standards. The great exception of twentieth-century America was the administration of President Franklin D. Roosevelt, which actually tried to slow the pace of mechanization. Of the 280 regulations in the National Recovery Administration, thirty-six included restrictions on the installation of new machines.15 By focusing too much on worker-replacing technologies, the administration missed many of the enabling technological advances of the time. Michelle Alexopoulos and Jon Cohen write:
In this respect, it is quite possible that the Roosevelt administration’s preoccupation with the labor displacing effects of new technologies was largely a consequence of its focus on innovations in manufacturing. Had it been able to take a broader perspective that encompassed the rapid growth of new products linked to automotive and electrical advances, the administration may have been more sanguine about the employment impact of these new technologies. It may even have embraced the idea that the great wave of gadgets that swept over the U.S. from the mid-1930s actually prevented a bad situation from becoming a good deal worse.16
Instead, the debate stopped only when unemployment ceased to be a concern. As late as 1940, Roosevelt warned in his State of the Union address that America had to begin “finding jobs faster than invention can take them away.”17 It took the attack on Pearl Harbor and America’s entry into World War II for the machinery question to ebb. Beating the Axis powers required everyone to work at full capacity, and they did.
* * *
The machinery question, however, only faded temporarily. The first few introductions of electronic computers into the workplace sparked a panic about the threat of automation to jobs in the news media, and the upsurge in unemployment that came with the three post–Korean War recessions led people to connect the two. Looking back in 1965, Robert Solow noted: “Whenever there is both rapid technological change and high unemployment the two will inevitably be connected in people’s minds. So it is not surprising that technological unemployment was a live subject during the depression of the 1930s, nor that the debate has now revived.”18 As noted above, the technological unemployment debate actually predated the Great Depression. However, machinery angst in the twentieth century was clearly cyclical, and this time it followed an upswing in unemployment after the Korean War. Though it is hard to detect much progression in the nature of the debate, our vocabulary bears witness to the progression of technology: the discussions of the 1950s and 1960s centered on the new popular term “automation.”19 Just like “technological unemployment” in the 1930s, “automation” and its discontents became one of the defining themes of the postwar years.
In America, the first comprehensive inquiry into the employment effects of automation was undertaken in 1955, when twenty-six leaders of labor, industry, and government testified before a Congressional subcommittee.20 The subcommittee concluded that “all elements in the American economy accept and welcome progress, change, and increasing productivity,” but that “no one dare overlook or deny the fact that many individuals will suffer personal, mental, and physical hardships as the adjustments go forward.”21 During the hearings, nobody proposed measures to restrict the use of mac
hines or even disputed the desirability of automation. Rather, the witnesses urged that greater attention be given to the social problems emerging from dislocation and noted that there was concern that older workers in particular would face difficulty in finding new and better jobs. Union representatives voiced the traditional demands for labor to get a larger share of the nation’s increasing productivity in the form of higher wages, shorter hours, and a lower retirement age. But Secretary of Labor James P. Mitchell responded: “I repeat, there is no reason to believe that this new phase of technology will result in overwhelming problems of readjustment. Science and invention are constantly opening up new areas of industrial expansion. While older and declining industries may show reducing opportunity, new and vibrant industries are pushing out our horizons.”22
The automation debate extended far beyond America’s borders. When the International Labour Organization (ILO) convened the fortieth annual ILO Conference in 1957, automation was a topic on everyone’s mind. In connection with the conference, David A. Morse, the ILO’s director general, wrote an article on the subject for the New York Times. Just as William Green had argued that mechanization was nothing new in the 1930s but was now progressing faster than at any time in history, Morse suggested that “no one could claim that automation is new. The productivity of machines has been increasing the productivity of man for many centuries. Perhaps the newest thing about automation is its tendency to speed up the rate of technological change and thus both to multiply the opportunities for social progress and to pile up the social problems likely to go with it.”23 As in the 1930s, the 1950s was held to be different because of the growing pace of technological change. Morse also pointed to the human tragedies that emerged when automation meant displaced workers failed to find work elsewhere. But on the whole, he was optimistic about the opportunities arising for “a better living, and for a better world society.”24
In the public debate, those worker-replacing technologies that reduced the bargaining power of labor unsurprisingly received the bulk of the attention, but things were not quite as straightforward as they might seem. Take the case of the elevator operators, for instance. During a general strike on September 24, 1945, elevator operators left more than fifteen hundred office buildings across Manhattan empty. Workers swarmed in the lobbies and crowded the sidewalks, while a few brave ones tried to climb the endless stairways of the highest skyscrapers. Such disruption was immensely costly and bad for business, and the arrival of the automatic elevator seemed to be the best way to ensure that such a thing would never happen again.
However, substituting automatic elevators for manual ones required public acceptance. Many citizens were first terrified by the idea that automatic elevators might leave them hanging on a cable hundreds of meters in the air, without any operator being responsible for their safety. Such concerns seem familiar today, given the present discussions about the adoption of autonomous vehicles. But like human drivers today, elevator operators were not infallible. Injuries were frequent, and several operators in New York City are reported to have been involved in fatal accidents. One operator on Seventh Avenue was killed when the elevator “shot up and pinned him at the top of the door,” and another was found “wedged between the lift and the door” in the Bronx.25 Indeed, a report filed by the Elevator Industry Association in 1952, in response to attempts seeking to restrict the introduction of automatic elevators, came to the conclusion that automatic elevators were fully five times safer than manually operated ones.26
When the jobs of truck and taxi drivers will go remains to be seen. But people in the 1950s were clearly right in thinking that the occupation of the elevator operator would soon become a distant memory. In 1956, the New York Times predicted that “the elevator operator may be joining the coachman and trolley car motorman on the road to oblivion.”27 That year in New York City alone, 43,440 elevators (about one-fifth of those in operation in America) were estimated to have carried 17.5 million passengers a combined distance of halfway to the moon. But by 1963, it was reported that although 35,000 elevator operators were employed in New York City in 1950, only 10,000 of their jobs were left. While the Empire State Building was still one of those with manually operated elevators, one article reported that a $2 million investment would serve to cut operating costs associated with operators’ salaries, pension plans, and sick leaves. At the Chrysler Building, forty-eight of its fifty-two elevators had already been converted to automatic operation. Two-thirds of elevator operators had been reassigned as porters and handymen, while the remaining third found themselves looking for new work.28
For the most part, however, automation anxiety concerned computers. As Abraham Raskin wrote in 1961, “The ultimate horror in labor’s book was the announcement that a computer … will do the work of seventy-five of the biggest computers now in use.… When computers start creating unemployment among computers it is really time to start worrying.”29 As we will see in chapter 9, much like artificial intelligence today, the first computers did not have any meaningful impact on labor markets even in the 1960s. In fact, the effect of computers on employment was not felt before the 1980s. Even so, most commentary on the early introduction of computers focused on the fear that they would leave many Americans unemployed. Such worries reached the heart of the government, with some congressmen fearing that the displacement of government employees by electronic computers would imperil the practice of giving politicians government jobs in recognition of their service. As statistical work could be better accomplished by machines, politicians were torn between the “desire for increased efficiency and the fear that the patronage system will be wrecked,” in the words of C. P. Trussell. The matter was treated most seriously. In 1960, a subcommittee of the House of Representatives, chaired by Congressman John Leskinki of Michigan, recommended that workers who might experience displacement should be given sufficient notice to be retrained to handle new machines, equipping them with the skills to keep their jobs. In the case of a net reduction in employment, the subcommittee also recommended a hiring freeze to allow displaced workers to take vacant jobs. However, it did not advocate any policies to slow the pace of computer adoption.30
During the 1960 presidential campaign, Senator John F. Kennedy gave a spirited speech on the automation dilemma at a rally in Detroit that was quite similar to what Secretary of Labor Davis had said in 1927. The message was straightforward. The coming automation revolution, Kennedy suggested, is “a revolution bright with hope of a new prosperity for labor and abundance for America—but it is also a revolution which carries that dark menace of industrial dislocation, increasing unemployment, and deepening poverty.”31 After Kennedy became president, the first formal report by his Advisory Committee on Labor Management Policy, published in 1962, claimed that “it is clear that unemployment has resulted from displacement due to automation and technological change.” But it added that “it is impossible, with presently available data, to isolate that portion of unemployment resulting from these causes.”32 Yet such caution was not enough for Kennedy to hold back on the issue. When asked, “How urgent do you view this problem, automation?” during a news conference in 1962, he responded:
Well, it is a fact that we have to find, over a 10-year period, 25,000 new jobs every week to take care of those who are displaced by machines, and those who are coming into the labor market, so that this places a major burden upon our economy and on our society.… But if our economy is progressing as we hope it will, then we can absorb a good many of these men and women. But I regard it as the major domestic challenge, really, of the ’60s, to maintain full employment at a time when automation, of course, is replacing men.33
The tragedy of Kennedy’s assassination the following year didn’t end the automation debate. Soon after taking office, President Lyndon Johnson set up a National Commission on Technology, Automation, and Economic Progress. Like Kennedy, Johnson was not opposed to automation. When signing the bill to establish the commission, he noted that “technol
ogy is creating both new opportunities and new obligations for us.…” He saw both an opportunity for faster productivity growth and the obligation to make sure that no worker and family “must pay an unjust price for progress.” Automation, he argued, could be the “ally of our prosperity if we will just look ahead, if we will understand what is to come, and if we will set our course wisely after proper planning for the future.”34 Large parts of the commission’s report, published in 1966, was dedicated to investigating “the belief that technological change is a major source of unemployment” and the fear that technology would eventually “eliminate all but a few jobs, with the major portion of what we now call work being performed automatically by machines.”35 But unlike Kennedy’s Advisory Committee on Labor-Management Policy, the commission concluded that persistent unemployment in the period 1954–65 was not due to automation, and the conclusion was supported by some actual analysis. “The persistence of a high general level of unemployment in the years following the Korean war,” the commission asserted, “was not the result of accelerated technological progress. Its cause was interaction between rising productivity, labor force growth, and an inadequate response of aggregate demand.”36 Yet despite this conclusion, the commission considered automation sufficiently disruptive to recommend making the government the employer of last resort, expanding free education, and introducing a guaranteed minimum income.
The Technology Trap Page 21