ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

History of Anxiety and Possibility

Revisiting Keynes’s Technological Unemployment

Ashish Kumar Sedai ( is a research scholar in economics at the Colorado State University, United States.

The developed world faces a new bout of anxiety given the current status of technology development. To understand the anxiety, the historical perspective on the short- and long-run effects of technological progress on labour substitution, post the Great Depression, is traced. The present technological progress stands in contrast to what Keynes called the “Economic Possibilities” of technological unemployment. The probable policies for tackling contemporary issues of technological unemployment are discussed.

The political economy of technological labour substitution is filled with insightful works, but the one that stands in the middle and sweeps across the significant part of history is John Maynard Keynes’s “Economic Possibilities for Our Grandchildren” (1930).1 The essay is futuristic and depicts Keynes’s understanding of the impending technological unemployment. A famous quote from the essay reads

As machines continue to invade society, duplicating greater and greater numbers of social tasks, it is human labor itself—at least, as we now think of “labour”—that is gradually rendered redundant. (Keynes 1930: 4)

But, to our surprise, he was not worried about it.

Keynes had claimed that a hundred years hence—with compounded economic growth and manifold increase in living standards—technological unemployment would not be a real concern, and even if it is, there would be simple solutions. He argued that with economic abundance and no worries for subsistence, human labour would willingly decline. The evolution of technologies and machines would finally allow humans to enjoy the fruits of their labour. For Keynes, technology was power—a means to overcome the imminent problem of subsistence that we faced since our inception. Mechanisation would take over long working hours, a 15-hour week would be sufficient to fulfil human appetite for work.

Technological Unemployment

For Keynes, technological progress was a manifestation of rapidly evolving capitalist framework. He expressed how technical efficiency had been taking place faster than the capacity to deal with the problem of labour absorption. But this technical efficiency would only make economic growth possible, which would envelope the short run and the long run concerns of technological unemployment. Keynes lauded the potential of compounded economic growth and technological improvements in solving the economic problem of subsistence within a century. He exerted that if capital could increase by 2% per annum, the capital equipment of the world will have increased by a 50% in 20 years, and by seven and a half times in a hundred years, which would tremendously improve living standards. To sum up Keynes’s idea of technological unemployment, one might argue that Keynes believed in the mechanisation of tedious labour processes as the source of technological unemployment. However, he did not think of technological unemployment as a manifestation of profits or revenue-based labour-saving technologies, the kind that the present-day automation portends to be.

Since Keynes’s optimistic commentary, many economists had written about technological change and its potential to change the modes of production. Solow (1957: 5), in his famous neoclassical growth model, starts off with an assumption that technological progress is exogenous to the production function and is labour augmenting. The Solow equation2 assumes that output increases when technology increases, directly pointing to the positive relation between technology and labour. Nearly 90 years since Keynes’s essay, technological unemployment has remained a contentious issue with peak periods in bouts of anxiety.

Surely, there was no lack of imagination and observation among the 19th century political economists and it is understandable that they were writing in a time when technology augmented labour and performed routine manual tasks. Machines primarily substituted for human and animal strength and gave dexterity, while capitalists got an opportunity to expand in size and scope. But these machines could neither reason or sense, nor make decisions or be responsible on their own. In effect, they acted as substitutes for physical work. Mechanisation was the key to the bright light ahead, through higher production and improved material conditions. Leisure, as many thought, would be the epitome of paradise. Thus, Keynes’s (1930: 5) idea of free time made complete sense in its literal spirit.

The Post-Keynesian Analysis

With the end of the period of capital–labour accord, coupled with sensitive and smart technologies, the question became more intriguing: Is this period of technological unemployment different from the previous one? Is it going to lead to long-term unemployment? A study by the United States (US) Congress in 1986, on the manufacturing sector from 1960–85, showed that technology replaced human labour, even in the farm sector. Although this technological progress replaced some of the most dangerous, onerous and repetitive tasks in manufacturing and services, the displacement was far from painless for some workers. The report showed that the problem of technological labour substitution was going to be bigger in the next decades.

Before World War I, it took 30 years of technological innovation to yield a commercial product, while in the 1960s it took only nine years (Picado 2017). The invention of the “Milwaukee Matic” in 1959 had revolutionised production and reduced the need for 40% of toolmakers and 50% of machine operators in Milwaukee in the US. The union office reported that without a shorter working week, 60% of their members would be out of jobs. Brynjolfsson (1993: 5) argued that the impressive advances in computer technology from improved industrial robotics, information services and automated translation is largely behind the sluggish employment growth of the blue-collar jobs between 1973 and 1993, as well as in professions such as law, finance, education and medicine.

Digital technologies change rapidly, but organisations and skills are not keeping pace with such change. As a result, millions of people are being left behind, their incomes and jobs are being destroyed, leaving them worse off than before the digital revolution (Brynjolfsson 1993: 5). Rapid technological change has been destroying jobs faster than creating them, leading to stagnant median income and growth of inequality in the US (Rotman 2013: 4). The Luddite Fallacy3—which implies that jobs lost in one area of the economy due to improved technology will be supplanted by new jobs created in other spheres—seems to have conjured up in the wake of intelligent and sensitive machines.

The new problem of technological unemployment was identified by Rifkin (1995: 1) in his book End of Work. He commented, “[S]ociety is gradually moving towards an eventual elimination of work—technology has added enormous value but has a price, the price is jobs.” According to Rifkin, the US had already experienced the third and final stage of shift in economic paradigm. While the first shift was a result of the development of steam power, the second shift, between 1860 and World War I, took place with the replacement of oil with coal, and the third or final shift, starting post World War II, actually materialised much later with the computers and numerically controlled robots. This shift has invaded the last and remaining human sphere, namely the human mind.

At present, we are on the verge of the fourth industrial revolution (Schwab and Samans 2016: 4) with 3D printing, driverless cars, artificial intelligence, drone deliveries, robots, and many such labour-saving machines that can wake us up, make our coffee and clean our closet. To take an example, Sophia, the humanoid robot created by Hanson robotics recently got her citizenship granted by the Saudi Arabian government (Future Investment Initiative 2017). Robots and humanoids have a compelling advantage over humans in terms of efficiency, accuracy and durability. This is forcing firms to rethink and retool everything that they do internally. Rethinking is bound to affect the demand of both skilled and unskilled labour. Hence, an imminent question emerges: What is to be done for the structural and involuntary unemployment created by this technological growth?

Is Shared Work the Alternative?

Marchant et al (2014) provided economic and social framework for addressing the problem of technological unemployment. An array of possible solutions was in the form of (i) slowing down innovation and change, (ii) sharing work, (iii) creating new work, (iv) redistribution, (v) education, and (vi) fostering a new social contract. Of all these solutions, some, including (i), (iii), (iv) and (vi) appear less plausible in the near future, as slowing down innovation and change counters the idea of economic growth.

For Keynes the need for reduced working hours is not only for humans to enjoy the fruit of the hard work done in the years gone by, but from the necessity and the incessant nature of financial crises, which to no one’s surprise are not by chance. Rather than laying off employees affected by technology, reducing the number of hours worked in a week is an alternative. Using a shared work policy of working four days a week instead of five days or 32 hours a week instead of 40, could allow firms to make the necessary reduction in working hours.

In 1981, the French government introduced a proposal to reduce the working hours from 40 to 35. The proposal initially manifested into a law in 1982, with 39 hours week and five paid vacations. Since then, the French government has tried and tested different approaches and procedures to increase employment by using the shared work strategy. After two decades, in 2000, the Aubry II Law4 confirmed the 35-hour reference. The law enacted permanent social tax cuts, concentrated near the minimum wage for firms applying a 35-hour agreement (Askenazy 2008: 4). This reduction in work was achieved through a partial redefinition of the work, as per former President Nicolas Sarkozy’s amendment to Aubry II 2007, which implied working more for earning more, not limiting the working hours for any employee, suppression of fiscal charges on overtimes and cuts in the employer’s social contribution.

To understand the shared work strategy, let us explain with an example. Suppose we induce the shared work alternative in a firm that is bound to cut labour employment because automation has taken over. Before the threat, an employee worked a full week earning $155 a day. But after the lay off they earn about $64 a day as unemployment benefit. In a shared work arrangement, however, if each employee is given work for two and half days each (instead of laying off some completely), not only the average daily earning of each worker will be higher than the unemployment benefit, this would also bring some relief to the problem of subsistence due to no job and no income.

Concluding Remarks

Like many caricatures, however, the above prescription has certain critical conflicting assumptions. First, the availability of firm-level data on lay offs and underemployment is hard to acquire. Sample techniques to figure out shared work polices could be counterfactual. Second, there should be a considerable number of employees in the affected unit for the switch to take place. Third, the normal 40-hour work week should be reduced by no less than 20% and no more than 40% to effectively work out the four-days’ work week. The business must provide the same fringe benefits to its employees. There should be a clear and timely communication by the business to the employees about the shift.

From the employees’ perspective, an individual must accept the contract offered by the business for the claim period. The individual must be able to work, be available full-time and be eligible for the unemployment benefits. All this, with a check on the growth of disguised unemployment of the labour force. There is a need to understand that the middle- and low-income class and imported labour does not have that luxury and the necessary wealth to choose leisure over labour. Also, their choices are constrained by the possibility of hunger if they do not work for more hours.

There is no need to take a given 40-hour work week, many European nations do not and have not had it for a long time. The concept of shared success and collective progress leads us towards the question of “Universal Basic Income.” It is an interesting policy tool for which there are many likes and dislikes. It is widely discussed as a policy option, but clearly not the only one. There could be mandatory profit sharing by firms, similar to a percentage of output, thereby collectively creating and sharing value. Minimum employment guarantee schemes for impoverished households could be another alternative. Developing economies like India and Brazil have had successes in minimum employment schemes despite various concerns. Also, ceteris paribus, it is certainly not a good idea to cut back on welfare spending in the face of technological unemployment and inadequate economic subsistence.

We live in a community where accepting the suffering of any of us can make all of us poorer and less well off. Accepting this as the default gradient should be unacceptable to us. If a cheque from the government becomes the key to a minimum living standard, it should be provided, while the issue of “how” can be a secondary matter. Apparently, that is exactly the kind of public good that government should create.


1 The final version of the essay, which appears reprinted, was included in Essays in Persuasion (Keynes 1930: 349).

2 Output Y = F(K, L) = AKα L1-α. Here K is capital, L is labour, A is technology, and α, 1-α is the intensity and the sensitivity of capital and labour respectively.

3 Luddite rebellion took place in British textiles in the early 19th century when angry rebels marched to trash the machinery.

4 Law named after Martine Aubry, the minister of labour under Lionel Jospin’s tenure.


Askenazy, P (2008): “A Primer on the 35-Hour in France, 1997–2007,” IZA, Institute for the Study of Labour, IZA DP No 3402, March, pp 1–30.

Brynjolfsson, E (1993): “The Productivity Paradox of Information Technology: Review and Assessment,” Communications of the ACM, pp 1–24.

Future Investment Initiative (2017): Sophia Robot, Hanson Robotics Ltd, Riyadh: Public Investment Fund, Saudi Arabia.

Keynes, J M (1930): “Economic Possibilities for Our Grandchildren,” Essays in Persuasion, J M Keynes (ed), New York: Harcout Bruce, pp 358–73.

Marchant, G E, Y A Stevens and J M Hennessy (2014): “Technology, Unemployment and Policy Options: Navigating the Transition to a Better World,” Journal of Evolution and Technology, Vol 24, No 1, pp 26–44.

Picado, A (2017): “The Real Reason Manufacturing Jobs Are Disappearing,” Ted Talk, February, New York, United States,

Rifkin, J (1995): The End of Work: The Decline of the Global Labour Force and the Dawn of the Post Market Era, Vol 9, Harvard Law and Technology.

Rotman, D (2013): “How Technology Is Destroying Job,” MIT Technology Review, 12 June, pp 1–12.

Schwab, K and R Samans (2016): The Future of Jobs, Employment, Skills and Workforce Strategy for the Fourth Industrial Revolution, Davos, Switzerland: World Economic Forum.

Solow, R M (1957): “Technical Change and the Aggregate Production Function,” The Review of Economics and Statistics, Vol 39, No 3, pp 312–20.

Updated On : 8th Nov, 2019


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top