The loss of manufacturing jobs in advanced economies is due not only to the offshoring of production to emerging markets, but also due to the encroachment of robotics, automation and artificial intelligence on these processes.
Robots are becoming more reliable, proficient and cheap. The combination of machine learning, through studying large data sets, and artificial intelligence is expanding the number of functions that robots can take over from humans. In addition, the incorporation of advanced information technology systems in production has provided an abundance of relevant data for both sides of the supply chain, increasing production capacity by around 20%. According to the consultancy McKinsey, around 50% of existing jobs could feasibly be taken over by automation right now. By 2030, almost one-fifth of all jobs will be automated.
Companies that deploy automation can realise substantial performance gains and develop competitive advantages in their industries, their efforts contribute to the aggregate level in productivity. This incentivises companies to automate processes, despite the risk of major disruption to traditional labour markets. Such structural shifts are accentuated in many developed countries by aging workforces and the struggle to replace skilled retiring employees.
The pace at which full automation can be implemented will depend on technical feasibility, which requires consistent innovation and falling installation costs. Industry will have to consider, too, the availability of low cost labour and whether adopting full automation will provide greater economic benefits than retaining human workers.
Advanced economies’ manufacturing sectors will be the first to be displaced by automation, though services are liable to be impacted as well. Highly predictable work done in structured environments, such as data collection and processing, can be automated using algorithms. Banking jobs will also be affected. Deutsche Bank has said that almost half of its staff can be replaced by AI and automation, while Citi plans to shed half of its 20,000 technology and operations staff in the next five years.
Emerging markets will also feel the effects. As soon as wages and technology adoption in these markets begin to reach parity with those of advanced economies, automation will prevail. China has been able to adopt a commanding position in global manufacturing thanks to cost advantages in labour. But Chinese hourly wages are increasing, up 64% in 2016 from 2011 at $3.60 per hour – higher than average wages in India and Indonesia, among others – and will continue to rise as the expanding Chinese services sector puts additional pressure on wages. Eventually, wages in India and Indonesia will rise, causing similar displacement in these major markets.
Some argue that entirely novel jobs will materialise naturally as others are displaced by automation, as has happened throughout history; the introduction of the tractor and shift away from labour intensive farming did not lead to long-term mass unemployment. Automotive industries in the US faced increasing automation between 1950-70, but wages and employment continued to rise steadily over the same period as jobs were redistributed to the emerging services sector.
However, today’s pace of innovation and technological development is becoming exponential. Labour markets will find it increasingly difficult to identify new jobs and reduce slack over the medium term. Technological progress may make a society more prosperous in aggregate, but not everyone will benefit. Inequality will increase in the absence of political and regulatory intervention, as both blue- and white-collar workers are made to find new jobs that are likely to be less well paid, while business owners profit from automation.
Universal basic income and taxes on robots have been suggested as methods of redistributing income and wealth more evenly. Others propose that a shorter working week will help redistribute time if some jobs are partially automated. Improving education in science, technology engineering and mathematics can mitigate poor labour market outcomes. Equally, bolstering industries that do not rely on robotics and are instead tied to human interaction, such as care-giving and teaching, will be essential.
For policy-makers, striking an appropriate balance between economic growth through automation and the defence of workers will be a significant challenge in both advanced and emerging market economies.
Bhavin Patel is Economist at OMFIF.