Throughout history, technological advancement has changed the structure of labor markets and societies. The Industrial Revolution in the 18th and 19th centuries both destroyed and created jobs. Machines and mass production put craftsmen out of business, but they also created a large class of factory workers. In the early to mid-20th century, the spread of powered assembly replaced these unskilled factory workers with a new class of factory line technicians. In recent years, continued computerization, advancements in international trade and the ability to outsource production have all contributed to the decreasing number of manufacturing jobs. This has led to a decline in the number of technicians but growth in the financial, logistics and computer industries, which support manufacturing.
Through all these shifts, most labor markets have rebounded by reallocating labor, avoiding chronic underemployment. Some people struggle to adapt to new jobs that require new training, leading to some negative individual outcomes. Certain local economies, too, have been decimated — the collapse of Detroit is a prime example. In most cases, however, increases in labor productivity tend to lead to more employment and positive economic gains. The regional crisis in Detroit is a problem, but does not threaten the overall U.S. economy, which is geographically and technologically diversified. This resilience depends on geopolitics — a small country like Austria would face an existential threat if there were an industrial crisis in its economic core around Vienna.
Gauging Impacts by Classifying Tasks
Those forecasting the impacts of computerization begin by looking at the tasks a certain job requires and deciding whether those tasks could be translated into computer code and performed automatically by a robot or 3-D printer, for example. This is called the Task Model, conceived in 2003 by David Autor and Frank Levy of Massachusetts Institute of Technology and Richard Murnane of Harvard. The model classifies all tasks as either routine or non-routine and either cognitive or manual. Routine tasks are repetitive and simple, such as sorting packages. Nonroutine tasks, however, do not follow standardized step-by-step sequences, like translating a foreign language. The task can then be broken down into cognitive or manual, using thinking or physical skills. Tagging each task according to these classifications leads to four categories.
The potential to computerize these tasks depends on their classification. It is easy to write computer code for easily defined, sequential work. This means that many routine/manual manufacturing tasks can and are performed by industrial robots. At the moment, computerization is breaking the most ground in replacing routine/cognitive tasks. New software and connectivity now allows websites such as TurboTax and Priceline to do work once done by accountants and travel agents. As time goes on, software will continue to replace human workers in jobs such as telemarketing or language translation.
But current limitations in artificial intelligence make tasks that require recognition, perception, social interaction or creativity difficult. These nonroutine/cognitive jobs will remain partially insulated from computerization. But advancements in computerization will over time augment the productivity of these positions. For example, the nonroutine/cognitive transportation and logistics industry could be revolutionized if driverless vehicle research by Google and others gains ground or if Amazon's delivery drone project comes to fruition. And as more consumers carry handheld devices connected to the Internet, personalized advertisements in stores could even replace salespeople.
Impacts on Income Inequality
The push to computerize routine/cognitive work will continue to impact the service sector in the developed world for the next two decades. Many of these jobs are middle income, similar to the routine/manual tasks that went offshore in the last three decades of the 20th century.
Even as the service sector loses jobs, the process will also lead to more technical positions to program, run and operate the new computerized infrastructure, including computer engineers, programmers and system testers. The spread of the Internet has already fueled the growth of high-paying financial sector jobs. These rising incomes at the top will in turn lead to new service-oriented jobs that require personal interaction and are more difficult to do offsite or through a computer. This can already be seen in the rise in the number of yoga instructors and personal fitness or wellness trainers.
The simultaneous growth of the financial and personal service industry is noteworthy because the newly created jobs are on opposite ends of the economic spectrum. Among the top 20 fastest-growing U.S. occupations between 2009 and 2013 are jobs with an average wage well below $30,000 — bartenders, personal care aides and administrative assistants. Occupations with the greatest decline, on the other hand, are largely middle class, including accounting clerks, teachers and executive assistants. The decline of the middle class is due in part to budget cuts, but even were these to end computerization will continue to erode middle income jobs without policies meant to deliberately slow the change.
The economic recession has sped up computerization and this will continue, slowing down employment recovery in Europe and North America. Jobs that involved routine tasks were among the hardest hit in 2009 and companies with slimming profit margins cut the positions that were easiest to replace through computerized processes. Even as parts of the U.S. economy recover, experience has demonstrated the financial benefits of computerizing tasks.
Bringing Back the Manufacturing Sector
Breakthroughs in robotics and additive manufacturing present the possibility of bringing manufacturing jobs back to the developed world. But this process faces constraints. The use of industrial robots to move objects is energy-intensive, and energy is expensive in the developed world, particularly in East Asia and Europe. This will limit the cost savings of bringing jobs back and also the development of artificial intelligence — computation requires a lot of energy. Future application of emissions caps and expensive alternative energy could also add to the price of manufacturing in industrialized countries.
A renaissance in robotic manufacturing would require advances in big battery technology as well as cheap and efficient large-scale storage of clean energy such as wind and solar. Widespread use of robotics for mobile applications in the industrial and service sectors would require further advances in small-scale batteries. Progress on both fronts has been relatively slow. Battery technology has historically advanced incrementally, but a breakthrough is possible over the next decade. Though even if a breakthrough were to occur, the pace of commercial adoption would depend on how much it increases battery efficiency and reduces costs compared to existing technology.
Even with the appropriate energy technology, robots will still face stiff competition from human labor in the developing world. Most manufacturing supply chains for East Asia are concentrated on the Indian Ocean and Eastern Pacific, which allows for short, cheap sea-based transportation for raw materials from China, Africa and Southeast Asia. This allows for local specialization and collaboration between countries — a television can be rapidly assembled in Japan using parts made in China, Taiwan and Thailand and then sent to the West in one long haul. Moving manufacturing to the United States or Europe would lengthen raw material supply chains and rely on comparatively expensive land routes for diversification and specialization.
North America will have a distinct advantage in overcoming the energy barrier to computerization. North America's shale gas revolution will likely provide cheap electricity for decades, balancing out slower progress in battery technology that makes renewable energy cheap. The United States' comparatively lower debt and demographic burden as well as its innovative economy will add to this advantage. Regions and nations with natural gas deficits, such as Europe, Japan and South Korea, will continue to pay more for natural gas because of physical limitations on the natural gas trade.
In spite of higher energy costs, Japan and Germany could also benefit from computerization. Neither is concerned with reintroducing manufacturing jobs as much as with maintaining historical competitiveness in high-end, export-oriented manufacturing. South Korea, Taiwan and China are now encroaching on these markets and Japan and Germany must speed up advancements in cheap robotic technology and, ideally, cheap energy. China is a formidable competitor because of its sheer size, lower labor costs and aggressive push toward industrial robots as demonstrated by Foxconn's announcement that it will deploy 1 million industrial robots in Chinese factories.
Overcoming and the Demographic Challenge
Demographic decline in the industrialized world could be the defining characteristic of the 21st century. In addition to their energy challenges and increasing competition abroad, both Germany and Japan have birth rates lower than replacement level. Japan faces particular difficulty — every future Japanese generation will be a third smaller than the one before. China, too, has its own demographic challenge.
A population drop, combined with increasing longevity, will mean that the workforce in these nations will need to support high levels of debt and large retired populations. Computerized production could ensure that productivity increases, offsetting the declining workforce. Governments could then offset the costs of their aging populations through taxes on corporate gains and distribution through a social security apparatus. Alternatively, the falling prices of goods and services would lower the cost of living for workers, who could then contribute the excess to support retired family members.
The medical and healthcare industries will become increasingly important amid demographic decline, and it will become essential to drive down the cost of this care. Few jobs in the medical profession are easily computerized, but computerization could support the efficiency of healthcare providers. Increasing robotic dexterity as an aid to surgeons and automatically administering expensive anesthesia could both drive down the cost of surgery. Computerizing nonspecialized positions in nursing homes and other care facilities could increase the ratio of patients to caregivers. All of these advances would lead to more efficiency in a growing industry.
The creep of computerization into new sectors of the economy will result in a certain level of social flux as positions are downsized. In the long term this will free up workers for new, innovative sectors of the economy. At the same time, computerization will also have a stabilizing effect by balancing out disastrous demographic trends, allowing nations such as Japan to support their growing nonworking population instead of resorting to contentious immigration reform.