We live in an era of dizzying technological change, with smartphones, self-driving cars and automated stock-trading desks apparently set to transform life across the globe. What will human beings do in an automated future? Will we be able to adapt our social and political institutions to realize the dream of human freedom presaged by a new age of machines? Or will it turn into a nightmare of mass joblessness? In Part One of this essay, I identified a new automation discourse, propounded by liberal, right-wing and left analysts alike.footnote1 These automation theorists claim that mass technological unemployment will need to be managed by the provision of universal basic income (ubi), since large sections of the population will lose access to wage labour.footnote2 I argued that the resurgence of this feverish discourse is a response to a real trend unfolding across the world: a chronic under-demand for labour. However, the explanation the automation theorists offer—runaway technological change destroying jobs—is false. The real cause of the persistently low demand for labour is the progressive slowdown of economic growth since the 1970s, as industrial overcapacity spread around the world, and no alternative growth engine materialized—a development originally analysed by Robert Brenner, and belatedly and obliquely recognized by mainstream economists under the name of ‘secular stagnation’ or ‘Japanification’.footnote3 As economic growth decelerates, job creation slows, and it is this, not technology-induced job destruction, which is depressing the global demand for labour.

In Part Two, I demonstrate that employment outcomes have differed in important respects from the automation theorists’ predictions. I analyse the contemporary dynamics of the global labour market and consider the solutions automation theorists have proposed, notably ubi, before going on to consider, as a thought experiment, an alternative approach to achieving a post-scarcity future. First, however, I will argue that it is crucial that we reconceive of the present situation as marked not by the imminent arrival of mass unemployment, as automation theorists suggest, but by continuously rising under-employment. A survey of worldwide vistas of insecure work shows that this new reality has already been accepted by wealthy elites. Turning the tide towards a more humane future will therefore depend on masses of working people refusing to accept a persistent decline in the demand for their labour, and the rising economic inequality it entails. Struggles against these outcomes are already unfolding across the globe. If they fail, maybe the best we will get is a slightly higher social wage in the form of ubi. However, we should not be fighting for this goal, but rather to inaugurate a post-scarcity planet.

What lessons can be gleaned from past experiences of job losses and profit-driven technological breakthroughs? On their own, these have never overcome human drudgery altogether. Nevertheless, they do periodically result in sweeping job destruction in certain industries, particularly when they allow firms to overcome a long-standing resistance to industrial development. Agriculture, for example, was one of the first sectors to be transformed by modern production methods: in the 15th- and 16th-century English countryside, new forms of animal husbandry on enclosed farms were combined with crop rotation to raise yields. Yet farming remained difficult to mechanize, due to the uneven terrain of fields and seasonal cycles, and for centuries it continued to be a major source of employment.footnote4 In the 1940s, however, advances in synthetic fertilizers, the hybridization of crops and the mechanization of farming implements made it possible to develop industrialized forms of agricultural production, and operative logics shifted dramatically.footnote5

Labour productivity took off, as farms came to resemble open-air factories. Given the limits to the growth of the demand for agricultural outputs, the sector then shed workers at an incredible pace. As late as 1950, agriculture employed 24 per cent of the workforce in West Germany, 25 per cent in France, 42 in Japan, and 47 in Italy; by 2010, all of these shares were under 5 per cent. During the 1950s and 60s Green Revolution, methods of industrialized agriculture were adapted for tropical climates, with stunning consequences for global agricultural employment: in 1983, the majority of the world’s workers were still in agriculture; that figure has since fallen to 25 per cent.footnote6 The major global job destroyer in the 20th century was not ‘silicon capitalism’ but nitrogen capitalism. No mechanism existed within the labour market to ensure that as many new jobs were created outside of agriculture as had been lost within it.

Firms are still seeking ways to overcome obstacles to industrialization, but in the present era of slowing overall growth rates and generally slack labour markets, these innovations tend to leave working people without steady jobs. For example, on a global scale, the mechanization of electronics assembly and apparel and footwear industries would be devastating: these sectors employ large numbers of people worldwide and generate foreign exchange for otherwise cash-strapped economies. Sewing in particular has long been resistant to technological modernization: it involves detailed work with fabrics, which machines have trouble manipulating; the last major innovation in the field was the Singer sewing machine in the 1850s. Electronics assembly work, although of more recent vintage, has proven similarly resistant to labour-saving innovation, since it too requires the delicate manipulation of tiny parts. As technological laggards within larger, highly mechanized production processes, these jobs were some of the first to globalize in the 1960s. Retail, apparel and electronics firms contracted suppliers in low-wage countries to meet a growing demand.footnote7 These industries remain significant as the first links of industrial supply chains, where they are subject to fierce competition among suppliers.

Much of this kind of work has relocated to China since the 1990s. However, even as Chinese wages rise and other countries become more competitive, advances in robotics may finally be overcoming long-standing resistance to further mechanization within these fields. Foxconn is deploying ‘foxbots’ to stave off competition from electronics assemblers in lower-wage countries. In China and Bangladesh, apparel companies are using ‘sewbots’ and new knitting technologies, which may also be extended to the manufacture of footwear. These innovations are unlikely to lead to full automation in these sectors, but they could eliminate lots of jobs very quickly, and block access to the global economy for further low-wage countries, in Africa for example.footnote8 It is unclear whether these technological developments are ten or twenty years away, and they may not occur on any scale at all. Yet even without major advances in automation, ‘Industry 4.0’ and ‘smart-factory’ technologies will increase the advantages of industrial clustering in the vicinity of related services, with the result that manufacturing jobs are more likely to be globally concentrated than dispersed.footnote9

By overcoming impediments to mechanization in sectors that have hitherto acted as major labour-absorbers, new technologies may serve as a secondary cause of the under-demand for labour. However, the key to explaining this phenomenon is not the rapid pace of job destruction in specific branches, if it occurs, but the absence of a corresponding pace of job creation in the wider economy. As I argued in nlr 119, the main explanation for that is not rising productivity-growth rates, as the automation theorists claim, but inadequate output demand, due to the proliferation of industrial capacities across the world, an associated over-accumulation of capital, and a consequent downshift in rates of manufacturing expansion and economic growth overall. These remain the primary economic and social causes of the slack in the labour market that is wracking workers across the world.