Briefing

Technology and the Future of Work

Accelerating Rates of Change

The pace of technological change is accelerating, in many cases too fast for any business executive or policy-maker to comprehend or get used to. Justin Trudeau, prime minister of Canada, summed up this sentiment well in Davos this year when he said: “The pace of change has never been this fast, and yet it will never be this slow again.”

It is true that technology is unleashing a cascade of forces that are making the world more complex and volatile. Instead of revealing unseen order and predictability, it is making the future hazier and more prone to surprises. In private conversations, many business leaders and academics echo what the historian Yuval Noah Harari recently wrote in the New York Times: “It is the first time in human history that we have no idea how human society will be like in just twenty years from now. We don’t know what the labour market will be like. This is also true for the geopolitical landscape, for education and the skills we need, for family structures, for gender relations and so on and so forth”.

The public discourse on the effect that technology in general, and AI and automation in particular, will have on employment is fuelled by anxiety about what the future holds. Currently, it is divided in two camps: the techno-optimists, who claim that technology will unleash a new era of prosperity; and the techno-pessimists who think that it will create technological unemployment on a gigantic scale. Who is right? Neither. It is easy to give way to excess pessimism because it is much easier to visualize what is disappearing than what comes next, but over the coming decades, we might witness an unprecedented wave of innovation and creativity, with a global explosion of new micro industries that may employ hundreds of millions of people. What is certain, though, is that in the years to come, job losses will be consequential and the adjustment painful. Those who will benefit from automation are the owners of capital and the highly skilled / highly compensated workers. Pretty much everybody else will suffer, particularly those with low-skill occupations as their wages will either stagnate or fall.

Technological Unemployment

Keynes originally coined the expression “technological unemployment”, referring to a process in which the impact of rapid progress makes jobs disappear. It has always been the case that technological innovation destroys some jobs, which it replaces in turn with new ones, in a different activity and possibly in a different place.

The “app economy” is a case in point. It only began in 2008 when Steve Jobs, the founder of Apple, agreed to let outside developers create apps for the iPhone. Just six years later, at the end of 2014, the “app industry” was worth $100bn, more than the film industry that had been in existence for over a century. An example always offered to illustrate the point that we shouldn’t fear technological progress is agriculture: in the US, it employed 90% of the workforce at the beginning of the 19th century, and less than 2% today.  This dramatic downsizing took place smoothly – with minimal social disruption and endemic unemployment during the process. Another example is the industrial revolution in England. In the early 1800s, the Luddites were destroying the mechanical looms out of fear that they would create unemployment. They proved to be wrong: far from creating misery and poverty, the first industrial revolution in reality increased the total number of people with a job and made England much richer. If we extrapolate from the past, why should it be different this time, ask the techno-optimists? They acknowledge that technology can be disruptive but claim that it always ends up improving productivity and increasing wealth, leading in turn to more demand for goods and services and new types of jobs to satisfy it. The substance of the argument goes as follows: human needs and desires are infinite, so the process of supplying them should also be infinite, and – barring the normal recessions and occasional depressions – there’ll always be work for everybody.

Growing Evidence

With respect to the 4th Industrial Revolution, what is the evidence so far and what does it tell us about what lies ahead? The early signs point to a wave of “labour-substitutive innovation”. This is corroborated by almost every single independent academic study published on the subject. Many different categories of work, particularly those that involve mechanically repetitive and precise manual labour, have already been automated, and many others will follow. Sooner than most anticipate, the work of professions as different as lawyers, financial analysts, doctors, journalists, accountants, insurance underwriters or librarians will be partly or completely automated.

In 2014, one of the ultimate “insiders” – Eric Schmidt, then chairman of Google – warned that: “automation was a race between people and computers”. At this year’s Davos, many tech CEOs voiced their concern that job losses caused by automation will be consequential and the adjustment painful. Jack Ma, Alibaba’s CEO, even said that the angst provoked by job losses could lead to conflicts.

So far, the digital age has not created many jobs. According to an estimate from the Oxford Martin Programme on Technology and Employment, only 0.5% of the US workforce is employed in industries that did not exist at the turn of the century.  This a finding confirmed in a recent US Economic Census which shows that innovations in information and other disruptive technologies tend to raise productivity by replacing existing workers, rather than creating new products needing more labour to produce them. This may prove to be one of the most crucial conundrums of tomorrow’s world: a choice between higher productivity and lower employment, or lower productivity, higher employment but low wages, like in the UK, where the output per man-hour is now more than 20 percentage points below the G7 average.

Winners and Losers

Generally, new technologies will benefit those at the top who possess the skills and the education necessary to harness the productivity benefits that they bring.  As an example, medical doctors will not work against technology but with it, using robots and AI’s predictive analytics to better diagnose and treat patients, and increasing their revenues in the process.  Workers at the bottom of the ladder who engage in repetitive tasks that can easily be automated will lose out dramatically.  As the pace of innovation is progressing much more rapidly than previously thought, AI will exacerbate the fears of all those who risk losing their jobs.  Also, AI will increase inequality by reinforcing the ‘winner takes all’ trend that has engulfed the global labour markets over the past few years. As such, it is important to keep in mind that if technology has the potential to boost productivity and therefore economic growth, it may also constrain demand by compressing the labour’s share of income and by increasing inequality.

The potential impact of automation on employment will vary by occupation and sector. Activities most susceptible to automation include physical ones in predictable environments, such as operating machinery and preparing fast food. Collecting and processing data are two other categories of activities that increasingly can be done better and faster with machines. This could displace large amounts of labour – for instance, in mortgage origination, paralegal work, accounting, and back-office transaction processing. It is important to note, however, that even when some tasks are automated, employment in those occupations may not decline but rather workers may perform new tasks.

People Skills

Jobs in unpredictable environments – occupations such as gardeners, plumbers, or providers of care for children and the elderly – will also generally see less automation by 2030, because they are technically difficult to automate.  They also tend to command relatively lower wages, which makes automation a less attractive business proposition.

Automation will have a lesser effect on jobs that involve managing people, applying expertise, and social interactions, where machines are unable to match human performance. Workers of the future will spend more time on these activities that machines are less capable of doing. They will spend less time on predictable physical activities and on collecting and processing data, where machines already exceed human performance. The skills and capabilities required will also shift, requiring more social and emotional skills and more advanced cognitive capabilities, such as logical reasoning and creativity.

The net effect of automation on the labour markets will depend on the pace of the labour-substitution effect. If it is slow, people will have more time to adapt. If it is rapid, the scale of disruption will be greater. It is indeed essential to understand that it will be the time needed to adapt to changing economic conditions and incentives that will make the whole difference. If it takes just a few years to find a new job in a continuous and relatively smooth process of renewal, all may be (more or less) fine.

If it is a matter of an entire generation being lost to technological progress, with unemployment peaking at 20% or above and stagnating incomes, before the next generation adapts, then the world is in for trouble.

History Lessons

It is true that, in the long-term, technology has always been a net job creator from a macro perspective, but this is of little comfort to people who risk losing their jobs to machines. Back to the example of the Luddites: in the 19th century, it took six decades after the advent of weaving machines for real wages of textile workers to recover. The Luddites were not anti-tech, but the skilled, middle-class workers of their time, whose lives were upended by machines replacing them with low-skilled, low-wage labourers in dismal factories. To ease the transition, the Luddites sought to negotiate taxes to fund workers’ pensions, a minimum wage, and adherence to minimum labour standards. As these bargaining attempts were rebuffed by factory owners, they began months of “machine breaking” in 1811-1812, smashing the weaving frames, in a last-ditch effort to bring their new bosses to the table. It was then that the British Parliament declared machine breaking a capital offense. It sent 14,000 troops to the English countryside to put down the uprising. Dozens of Luddites were executed or exiled to Australia.

According to David Autor (a labour economist at MIT), we are now in the midst of another such gap when innovation and automation are creating more wealth, but are redistributing income from workers to owners.  This partly explains why American wages have been largely stagnant despite one of the tightest job markets in decades.

No one knows how long the gap will last or when workers displaced by the new automation revolution will find employment at reasonable wages. New job categories that didn’t exist 10 years ago, like social media marketing manager, or robot repair technician, are growing rapidly but are unlikely to compensate the level of job losses in the foreseeable future.

This may be different for occupations that benefit from secular trends, such as ageing populations and growing demand for care services.  It will shift spending patterns, with a sharp increase in spending on healthcare and other personal services. This, in turn, will create significant new demand for a range of occupations, including doctors, nurses, and health technicians, but also home-health aides, personal-care aides, and nursing assistants.  According to the McKinsey Global Institute, healthcare and related jobs from ageing could grow globally by 50 million to 85 million by 2030.

Thierry Malleret
01.05.2018