Whilst on internship from the University of Leeds, 2nd year student George Travell explored the potential policy prescriptions for the 4th Industrial Revolution.
We’re in the midst of a technological revolution but have we been here before?
The fourth industrial revolution is well on its way to remove 20% of UK workers from the job market by 2030 - a report by McKinsey Global Institute has found. In their study of 46 countries they predicted job loss to be 800m by 2030 whilst still concluding that more jobs will be created than destroyed. Rising consumption particularly from developing economies is one source of new demand which could create approximately 280m jobs globally. Development and deployment of technology itself could create up to 50m.
One need only look back 200 years ago at the height of the US and Europe’s previous technological revolution to see that technological unemployment is nothing new. The Economist notes that in the 1960s, President John F. Kennedy declared that the major domestic challenge was “maintain full employment at a time when automation…is replacing people” highlighting how this is a persistent anxiety. Moreover, the same argument is made that with the introduction of new technology, new jobs are created in an act of creative destruction. These are arguably better jobs too - during the Industrial Revolution, the life expectancy of children increased dramatically. The percentage of the children born in London who died before the age of five decreased from 74.5% in 1730–1749 to 31.8% in 1810–1829 is one positive effect of technology on society. Though child mortality rates plummeted some estimates put the UK Gini coefficient (a measurement of inequality) change from 0.4 to 0.63 in the years between 1823 to 1871 galvanizing workers unions, socialism and ultimately very real threats of communist revolution. Governments such as Bismarck’s Germany in 1880 responded with “unemployment insurance” and Roosevelt’s America saw the introduction of the Square Deal to break up monopolies in the early 20th century. The problem resides in the pace of the displacement and government intervention being unable to keep up.
The stark similarities should be the first point of call for our governments when establishing the policy for the fourth revolution. History revealed communism erupted out of inequality, poor working conditions and no social protection. One can patently make parallels with the millennial resurgent socialist movements of today who make inequality their principle incisive critique of our times. The Gini coefficient currently stands at 73.2% in the UK.
Are we reaching the end of work?
First reflections on the effects of AI could easily lead one to question whether we are reaching the end of work. John Maynard Keynes believed that at this point we would be busier structuring our leisure time than worrying about work. Despite his immensely important contribution to economics he spectacularly incorrectly predicted that we would be now working a 15-hour working week. His rationale was reasonable enough; working hours had fallen from 53 in 1870 to 47 in 1930 when he made his prediction. Average hours today in the UK are around 40 and many economists argue it should be allowed to drop to 30 as this in fact entails greater productivity whilst also improving wellbeing.
How do we ensure equal opportunity?
Automation will increase productivity, output and national income but ultimately the greatest beneficiaries of the AI revolution will be those already working with the robots such as programmers. It is unlikely that this will be a Pareto improvement for the numerous lower manufacturing roles which will be lost and not necessarily allow for security and equal wages; indeed, the UK has seen a rise in zero hour contracts entailing minimal employee protection and uncertain hours. What’s more real wages have remained stagnant since the financial crisis, despite high income earners seeing their wages continue to rise. More relevantly all AI professionals have seen their salaries rise as much as 45% in the last five years. Just like the industrialists of the 19th century who owned the machinery, out of the fourth revolution could emerge the ‘AI owning class’. Though history has shown that technological revolutions are part of the natural evolution of employment patterns, some argue this time it is different as the change is happening at such an expeditious rate. The political fallout could be profound if workers find themselves displaced rapidly and on such a long scale.
Is there a third way between neo-luddism and automation?
Two clear goals for policy makers: allow for equal opportunity in new industries as well as ensuring welfare of the workers left behind. Education must fill the skills gap for tomorrow’s economy. Initiatives to bring coding into primary schools have now been brought into many UK primary schools but there is still a way to go. The UK and The US are also largely lagging in STEM graduates when compared to South Korea, Japan and increasingly China and India and so greater focus in these areas as well as vocational training will be vital. In addition to the school education, facilitating ‘life long learning’ will be of paramount importance to make sure age won’t be a discriminating factor.
Should we tax the robots?
South Korea recently became the first country to implement a ‘robot tax’ through reducing tax incentives for investments in automated machines as a means of filling welfare coffers for those displaced by automation. Though taxing AI may seem initially appealing to redistribute to those who have been displaced it quickly runs into problems. Firstly, just like any tax, if it is lower in another country then business will be encouraged to relocate to cheaper locations leaving any spoils of the AI revolution outside the UK. Stifling innovation which would have been used to increase competitiveness would not be wise seeing as the UK only surpassed 2008 productivity levels in 2015.
What are other policy prescriptions?
An alternative method could be to role out a universal basic income (UBI) to provide a financial floor to cover for any displacement caused in addition to maintain these individual’s consumption in the economy. However, implementing a UBI would require answering difficult questions on its goals and priorities. The most pressing issue would be how to fund such a wide role out for say $10,000 which would make a huge difference to living standards, but this could cost as much as 15% of GDP particularly if added on top of current welfare programs. The nature of UBI however would mean unconditional income with no conditions attached which could act as a disincentive to work for some. PwC suggest a more paternal form of UBI such as a ‘life long learning fund’ to allow those displaced to finance educating themselves new relevant skills. Unfortunately, outside of Finland there is little empirical evidence on the effects of UBI and so bigger experiments will be imperative.
Public and private investment in existing infrastructure will also need to play a role for ensuring sufficient work. The UK transport infrastructure is just one existing sector which requires substantial investment for modernising. Newer sectors like the digital sectors require considerably more investment – for example through place-based strategies focused on research centres, science parks and other enablers of business growth.
What is clear is that if our government is serious about a “stronger, fairer country with real opportunity” then people must have enough equal opportunity to access the spoils of the AI revolution. History revealed that it is in fact possible for governments to use the tools at their disposal to mitigate against technological displacement through welfare and tax reforms. They must act sooner rather than later to make sure policy measures can keep up with displacement.
George Travell BA Economics and Italian University of Leeds.
#AI #Automation #RobotTax #AIPolicy #Unemployment
Mabel C. Buer, Health, Wealth and Population in the Early Days of the Industrial Revolution, London: George Routledge & Sons, 1926, p. 30 ISBN 0-415-38218-1