The effect on a labour-surplus developing country might be devastating in the short term.
Most industrial nations and middle-income giants like India and China have seen falling GDP growth and increase in unemployment in the last 10 years. While the 2008 financial crisis and the COVID-19 pandemic have exacerbated this situation, a more insidious structural change has also been working from within economies over the last few decades to drive this. In their book, The Second Machine Age, Erik Brynjolfsson and Andrew McAfee warn us of a future where low-skilled workers would be redundant, with automation, robots and “intelligent” software replacing them all over the world.
That this would be the culmination of the profit-driven capitalist industry did not escape Karl Marx, who as early as in 1857 speculated that with the persistence of the capitalist mode of production, which create machines of increasing sophistication, an individual product would have so little labour value that it’s price would virtually be zero (in the chapter, “Fragment on Machines”, in Gundrisse). American author and satirist Kurt Vonnegut spoke of the redundancy of humans in a machine age back in the 1950s in his eerily prescient novel, Player Piano (1952), whose name evokes creepy images of playerless pianos in shopping malls belting out lifeless, machine-programmed muzak.
Bear in mind that this phenomenon of mass structural unemployment is not new as innovations such as the spinning jenny, the steam engine and computer chips resulted in thousands of workers losing their jobs in their eras. In Britain, blue-collar wages went down to less than half of what they were previously in the last decades of the 18th century, after the first Industrial Revolution.
What then is the future of data entry operators, customer service specialists, shelf stackers or the legion of office peons that still characterise Indian workplaces? A 2019 McKinsey report finds that close to half the workers in the U.S. and OECD (Organisation for Economic Cooperation and Development) countries are in jobs which are at high risk of being replaced or fundamentally transformed. That sounds alarming. Is it even possible for us to fight this future?
Well, the optimist/pragmatist would proffer the “obvious” solution: retraining, reskilling and redeployment of the same worker. This may be possible to some extent in the U.S. and western Europe, where as a result of societal affluence, individuals may increase their skills and productivity and find their way back to being gainfully employed. However, as with most optimistic theories in economics, this one too is not supported by preliminary data from even advanced countries.
According to Banerjee and Duflo (2019), the automation of simple industrial tasks has not led to an increase in the hiring of skilled personnel to supervise the automation. They argue that the disruption caused by automation increases the demand for both very skilled and very unskilled work arising from some “trickle down” of income. Software engineers and other knowledge workers get richer and now have a higher demand for dog-walkers or house helps — occupations which do not need college or even high-school education. They further say that as artificial intelligence (AI) displaces workers from clerical and administrative roles, it pushes more individuals (particularly from less-educated groups) into low-skill and physically harder jobs like food delivery, cleaning, and security, thus increasing economic and social inequality.
If the labour-displacing impact of automation has been quite significant in advanced countries from the 1990s, its effect on a labour-surplus developing country might be devastating in the short or medium term. With total factor productivity being already lower for these countries, automation or AI is often welcomed by business-owners and employers as a cost-cutting venture: it allows them to bypass the hiring of low or mid-skilled workers by getting instead a machine which is always accurate, never asks for a raise and whose health benefits can be a reasonably priced annual maintenance contract.
We are already getting a glimpse of this in India with payroll software and computerised factories using a fraction of human beings as compared to a decade ago. According to data released by NASSCOM in 2021, Indian software firms are gearing up to reduce headcount by a massive three million by the end of 2022, which will help them save approximately $100 billion annually, mostly in salaries. Of course, labour displacement will be the highest in software, IT and IT-enabled sectors.
For more traditional industries, technology adoption is slow and still relatively expensive. So, for the time being, many business-owners may stick around with good old human labour over R2-D2, especially since the latter can have installation and maintenance costs capable of burning a deep hole in corporate budgets. However, a couple of these industries, such as automobile and textiles, might get more automated in the coming decade.
What happens to the laid-off workers? In the skill-deprived Third World, who will employ these individuals for anything other than unskilled jobs, which many of them may not be willing or able to perform? By all accounts from the U.S. and western Europe regarding the first wave of computerisation in the 1990s, seamless reallocation of workers into other occupations is anything but guaranteed. It stands to reason that in the not-so-distant future, workers who are unable to find alternative occupations will have a downward impact on the generalised demand for commodities and merit goods such as education and health. Thus, over time, it is indeed possible that a significant fraction of the wealth created by the deployment of automation is eroded, with a shrinking of consumption and markets. There are other socio-economic fallouts of mass unemployment.
First, there is a rise in crime, caused by deepening inequality. This strains resources of governments who have to manage widespread unrest and restore order. Second, households with no income become a burden on government health systems, as they cannot pay if there is an illness in the family. Climate change will further deepen the miseries of the poor and the jobless by making their living and working environments more unbearable.
Given these frightening possibilities, why does this issue not get discussed more often? For one, technological progress per se has always been hailed as an end in itself for human civilisation. Add a generous dollop of positivist optimism and deep belief in the “theory” of trickle down to it and we can all blissfully relax on hammocks doing nothing while obedient robots get us our espresso.
A second reason behind the silence is that in formal economics, units of labour as a factor of production have not traditionally been treated that differently from units of physical capital. In fact, the efficiency of the capitalist system arises from the fact that units of input get matched in an unfettered way to their most productive use. Thus, different forms of capital seek out labour most appropriate to it, leaving other labour units to find other productive uses. In such a system, theoretical efficiency is not reduced if a handful of workers find high-paying jobs leaving legions of workers to near zero wages. This unfortunately is how much the marginal product of the latter group’s labour is, given the radical technological transition which has made them redundant. Of course, over time there will be adaptation and some reabsorption back into the workforce of the retrained as new jobs are created, some of them organisational roles produced by the process of automation. The question is, how long will this reallocation take? And what will be the livelihood costs of this disruption?
We are not too bothered about the way in which technologically driven inequality can dampen productive efficiency mainly because variables relating to psychological states, which crucially determine our motivation and hence productivity, have not found their way into conventional economic thinking or economic theories. All over the world, laid-off workers are prone to depression, anxiety and other mental health issues, gravitate to alcohol and drug addictions, making their seamless transition into other productive professions difficult. Work and employment also set lifestyle habits that make transitions difficult. Try hiring a laid-off junior accountant as a security guard and see how efficient he will be. Smooth retraining and reallocation in the short term is a costly affair, particularly in developing economies. Moreover, most redundant workers will not have the personal funds to affect such retraining and reskilling. The question is how to acquire the funds and mobilise the political will into paying these transitional costs.
“Robot tax” as a solution
One solution could be to make AI-enabled industrial entities pay a “robot tax” to disincentivise downsizing the labour force unless very high gains in productive efficiency can be realised. This was considered by the European Union parliament in 2017 before being disallowed on the grounds that it would stifle innovation. In the same year, the Korean government approved of a version of the tax on automation: it reduces tax subsidies for companies that automate and combines this with a tax on outsourcing so that firms which find it costly to automate because of the tax do not save on their wage bill by sending jobs to poorer countries.
A second way may be to institute some kind of universal basic income, to be given to structurally unemployed individuals, funded by a wealth tax on rich financial entities. According to Thomas Piketty and Emmanuel Saez, the idea of the high “top tax rate” or a generalised wealth tax is the only effective antidote to rising inequality. The debate on how much this tax should be and whether or not rich entities will pay it rages on.
Unlike the West, India is at the beginning of its journey towards automation. It is thus crucial for policy-makers to consider the efficiency-dampening effects of structural unemployment and introduce programmes to retrain and reallocate workers to other professions and provide them with support in the interim.
Sujoy Chakravarty is Professor of Economics, Centre for Economic Studies and Planning, at School of Social Sciences, Jawaharlal Nehru University.