An alien who visited Earth in the mid-1980s and came back today would see one clear difference: far more people in public spaces ignoring the world around them and staring at a small rectangular device, what she would come to learn is called a “smartphone.” Soon, the smartphone may be replaced by a device implanted in our body that connects with our mind and provides instant access to both computing power and enormous databases. Computer-enhanced humans are no longer the realm of science fiction. The information and communications technology (ICT) revolution has fundamentally changed what we spend time on, how we interact with one another, what work we do and where we do it, and even how people commit crime. Most importantly, it has upset the balance between the three pillars—the state, markets, and the community.
The ICT revolution has not just followed the course of previous revolutions by displacing jobs through automation; it has also made it possible to produce anywhere and sell anywhere to a greater degree than ever before. By unifying markets further, it has increased the degree of cross-border competition, first in manufacturing and now in services. Successful producers have been able to grow much larger by making where it is most efficient. This has created spectacular winners, but also many losers.
The technology-assisted market has had widely varying effects across productive sectors in a country. Some of the effects stem naturally from technological change, and some stem from the reactions of people and companies to it. Indisputably, it has raised the premium on human capabilities. As a result, some well-educated communities in big cities have prospered, while communities with moderately (typically high-school-) educated workers in semirural areas dominated by manufacturing often have not. More generally, as with past technological revolutions, the need for people to adapt has come rapidly, before the benefits have spread widely. Indeed, the communities that are required to adapt the most, as always, are the communities that have been experiencing the greatest adversity, and have the least resources to cope.
The antistate ideology that gained momentum beginning in the late 1970s has not been inconsequential. While the withdrawal of the state induced more competition in markets that were deregulated, it also allowed the acquisition of rents by the few and the relative shrinkage of opportunity for the many. Technology-induced inequality and human-induced inequality have built on each other.
The shift in attitudes was best epitomized in a new paradigm to guide corporate behavior, the principle of shareholder value maximization, which focused the corporation’s energy on enhancing value for a narrow class of investors. By and large, this has moved corporations toward greater efficiency and away from vague notions of doing social good, yet it has also undermined their public support by legitimizing actions the community believes are grossly unfair. Corporations have compounded their political vulnerability by attempting to enhance profits, not just by building a better mousetrap, but by influencing rules and regulations in their favor. As a result, not only is the private sector more dependent today on state benevolence to sustain such anticompetitive barriers, which make it a less effective counter to state power; it is also less likely to enjoy broad public support if the state moves against it because it is seen as part of the crony swamp.
Another important consequence of the ICT revolution is that, by enhancing the wage premium that goes to those with strong capabilities, it has strained community cohesion. To build the capabilities in their children that the market demands, people who have the income and the choice are tempted to move out of mixed or declining communities into communities of people like themselves. This is a phenomenon that can be seen in the United States, but it is also happening elsewhere. While the truly rich have always lived apart, the upper middle class has also been pushed to secede into its own enclaves. Even as job opportunities become more unequal, economic diversity within communities has fallen while diversity between them has increased. This sorting of human capital across communities has increased the inequality in access to the capabilities necessary to compete in the market.
To the extent that progress makes products cheaper, there could be more demand for them, and the overall quantum of human work can even increase. The new work will, however, be different.
Inequality, not just in economic outcomes but in opportunities, has therefore become an enormous problem. In the US, it shows up between residents of big cities or rich suburbs and small, economically devastated rural towns; between workers in big, young, successful service companies and small, older, struggling manufacturing ones; and between the top earners and the rest. The roots of this inequality lie not only in technological change, but also in the failure of the community and the state to balance and modulate markets.
These inequalities are also present, if not to the same degree, in continental Europe. Moreover, the path of integration that continental Europe has chosen has highlighted new inequalities, between the protected jobs of the older generation and the poor-paying jobs available to the youth or immigrants, between the political power of the large European countries and the weakness of the smaller ones, and between the economic well-being of disciplined Northern Europe and the relative backwardness of the unreformed southern periphery. Through integration, all these inequalities have come into the European fold.
The effects on jobs
The ICT revolution has eliminated certain categories of jobs, while enhancing the importance and reach of others. It also has had an indirect effect via trade, allowing certain tasks to be outsourced while increasing the insourcing of others.
As a number of researchers have pointed out, in recent years new technologies have eliminated jobs that involved well-specified routines or simple, predictable tasks. For example, the Amazon Go store (opened first in Seattle) tries to create a shopping experience with no lines and no checkout counters. As you walk in, you use the app on your phone to register your presence, pick up what you need, and walk out. Later, your Amazon account is billed. Computer vision and machine-learning algorithms, similar to the ones used in driverless cars, help identify what you pick up, and tote up your bill. Not only does this do away with checkout clerks; the underlying software has also reduced the need for someone to monitor stock levels, order new inventory, or reconcile the store’s books at the end of the day. The automated system does it all.
Of course, it has not done away entirely with the need for humans. There are still shop assistants to help guide people to the products they are looking for, stock shelves as they run out, and prepare some of the fresh meals that Amazon sells. The point is that humans have moved to handling exceptions, and to intermediating as experts between ordinary people and the system. So long as stores structure all this well, they improve the overall buying experience, even while cutting down on costs.
Routine jobs have been automated out of existence for decades now, regardless of whether the jobs required skills or not. Banks had hundreds of thousands of cashiers taking in and paying out cash, as well as counting it at the end of the day—a routine job that required integrity but no higher skills than basic numeracy. The job paid decent wages in order to attract honest people, and keep them that way. Automated teller machines and cash-counting machines displaced them, and now electronic payment systems such as Alipay and Apple Pay, which bypass cash entirely, are rendering physical cash, and the security apparatus that services it, redundant. Sweden has many bank branches that now refuse to take cash. Churches flash their bank account numbers on a screen so that parishioners can contribute their weekly offering using their phone. No doubt payments will get easier still in years to come. Yet, if anything, employment in banking has gone up as new cheaper bank branches are opened, and tellers morph into relationship managers advising retail customers on their loan options and their investment portfolios. According to the US Bureau of Labor Statistics, jobs in commercial banking and related areas have gone up from 2.4 million in 1990 to 2.7 million in 2017, despite widespread automation and an intervening banking crisis.
The data certainly are consistent with an increase in the number of jobs at both ends of the skill spectrum, and a decline in the middle.
New jobs are being created even as old jobs are lost. Consider, for example, skilled tax accountants, whose specialty was to know every arcane element of the tax code. Such jobs have also been displaced, in this case by tax software available for a few dollars. This leaves the highly trained tax lawyer, whose work is to erect customized international tax shelters for her high-net-worth clients, unscathed. Her work is not routine, since each shelter has to be crafted for the client’s specific situation, where her knowledge of the tax code and prior cases, as well as her creativity, are essential. The ICT revolution helps her do her job—she can access prior rulings or the relevant tax code much more easily—but it has not displaced her, at least not yet. Indeed, because she can create shelters faster using more readily accessible information, and because she becomes known internationally, both the supply of her services as well as the demand for them increase, enhancing her income significantly.
Importantly, tax software also creates new jobs for people with moderate skills. A high-school graduate with some training and familiarity with computers, employed by a tax-preparation agency, can assist ordinary people with their taxes—people who do not want to spend the time doing it on their own, or who are unfamiliar with computers. Earlier, they could not afford an accountant. Now, they can afford the assistant.
Let us focus on this last example more carefully. Historically, the complaint about machines and automation was that they rendered the craftsman redundant. For example, it is well known that Henry Ford added tremendously to his workers’ productivity by manufacturing cars in a moving assembly line. The line broke car assembly into multiple sequential tasks, allowing each worker to specialize in only one of many tasks. Equally important, and less well known, is that Ford insisted parts be honed to high tolerances so that they were interchangeable, so that each part did not have to be specially machined to fit the car. Interchangeability, coupled with the breaking down of tasks, allowed Ford to dispense with craftsmen and hire modestly skilled workers for his assembly lines, thus creating the mass-market car. We see similar de-skilling with tax software, with the middle-class tax accountant replaced by a lower-paid computer-literate assistant with only a few weeks’ training. The assistant, aided by software, is probably more competent than most accountants but less creative. Most people don’t want their tax accounting to be creative. De-skilling makes ordinary craftsmen or accountants largely redundant, but making the car or the tax service cheaper increases demand and may increase jobs overall.
When we think about technological change, we often assume the aggregate amount of work is fixed, and therefore consider that what is displaced by automation will increase unemployment. Economists sometimes refer to this as the lump of labor fallacy: that there is only so much work to go around. To the extent that progress makes products cheaper, there could be more demand for them, and the overall quantum of human work can even increase. The new work will, however, be different.
Of course, this means some kinds of workers will no longer be needed, at least in their old jobs—the aforementioned accountant, for example. Even so, as routine work gets automated, there is more demand for skilled people who can handle the nonroutine exception that is thrown up. In Ford’s time, mechanics who truly understood cars could set up repair shops, where they diagnosed and fixed the unique problems that each mass-produced Model T developed through wear and tear. Similarly, the accountant who can go beyond the routine can find employment at the tax-preparation agency or the tax-software company to handle special-situation queries—for an additional fee. Since such accountants don’t really do routine work anymore, they need more capabilities and enthusiasm than the ordinary accountant, but may be better rewarded for it.
Thus the direct effect of technological change, at least for the foreseeable future, may not be so much on the aggregate quantum of human work—unemployment in most developed countries is at historic lows at the time of writing—but on its redistribution. The rich, skilled tax lawyer earns significantly more, and has more work than she can handle; the middle-class tax accountant is typically worse off; and there are entry-level jobs for the computer-literate assistant, without much hope for additional skilling or career progression embedded in them.
The data certainly are consistent with an increase in the number of jobs at both ends of the skill spectrum, and a decline in the middle. Better-paid managerial, professional, and technical jobs and lower-paid service jobs have all increased significantly in the US as a share of jobs in the past three decades, even while middle-wage jobs have fallen. This polarization of jobs, with low-pay/low-skill occupations and high-pay/high-skill occupations gaining at the expense of jobs in the middle, is not just a US phenomenon. Studies find that in 15 of 16 European countries for which data are available, high-paying occupations expanded relative to middle-wage occupations in the 1990s and ’00s, and in all 16 countries, low-paying occupations expanded relative to middle-wage occupations.