moody 10

According to Moody, why should we be skeptical of the "rise of the robots"? Pay attention to the arguments he makes about particular industries--industrial robotics, logistics, etc.--and the larger, macro-level argument he makes about investment in technology in contemporary capitalism.

 Yet, over two decades later, there are more, not fewer, jobs in the US and the world. Work did not end, it just changed official statistical categories and got worse

Between 1990 and 2000 commercial banking did lose jobs, 111,700, about a sixth of the prediction, and this likely due to industry consolidation. By 2010, commercial banks had gained back half of those lost jobs despite consolidations and better computers, and by mid-2017 they had reached 97 per cent of the 1990 level. Computers destroyed jobs, but nowhere near on the scale imagined by Andersen Consulting or Rifkin.4

Instead, warehouse production and nonsupervisory jobs grew by 27 per cent from 1990 to 2000 and by another 83 per cent from 2000 to mid-2017, despite recession and technological advances.5

hey are much larger and hence require more rather than less labour

 The major contention here is that the susceptibility of any job to automation is secondary to the potential profitability of its actual application.

 The question is not that of the susceptibility of various jobs and occupations to automation, but the practicality of their application through actual investment.

 In conclusion, it will be argued that the dynamics of capital accumulation itself, as well as the turbulence of capitalism globally and in the US have led to a slowing-down of investment in work-related technology which remains a barrier to the sort of dramatic replacement of human labour by machines projected by the techno-futurists

‘47% of total US employment is in the high-risk category, meaning that associated occupations are potentially automatable over some unspecified number of years, perhaps a decade or two.’

hey make no effort to assess the economic feasibility or practicality of applying various forms of technology under today’s turbulent economic conditions

As might be expected, as caretakers and practitioners of capitalism, they are somewhat more circumspect in their predictions of job loss

When the WEF survey asked executives what they thought the main ‘drivers’ of change were, only 9 per cent answered, ‘Advanced robotics and autonomous transport’, while even fewer named ‘Artificial intelligence and machine learning’ or ‘3D printing’. In comparison, 44 per cent, the highest percentage, answered ‘Changing work environment and flexible working arrangements’

The only economic barrier to new technology mentioned in the survey was ‘Pressure from shareholders, short-term profitability’

They also predicted a tripling or even quadrupling of the number of robots between 2015 and 2025 

The IFR’s projections for 2015 to 2020 show only a doubling of annual robot shipments for the US from 27,504 to 55,000, while the Boston Consulting Group estimates US robot spending to increase by one-and-a-fifth-times to $24 billion from 2015 to 2025. However, their estimate for 2015 of $11 billion in robot sales equalled less than 3 per cent of GDP expenditures on ‘Machinery’ that year.12 In global terms, the US is actually behind most of the rest of the industrialised world. The IRF’s report World Robotics 2017 shows that while global shipments of industrial robots have grown significantly, those from ‘The Americas’ have never amounted to more than 18 per cent of the world total, and by 2016 were down to 14 per cent of which 20 per cent came from Canada, Mexico, Brazil, and the rest of Latin America.13 Further evidence for the relatively slow growth in robots lies in their uneven application across industries. According to a Brookings study, as of 2015 half of the nation’s 233,305 industrial robots were in auto with a huge concentration in the Midwest and upper South, the site of most car and truck-supplier and final-assembly plants. Of those 116,653 robots, 30,000 or over a quarter belonged to General Motors alone.14 Despite rapid growth in robots in a few US industries, the only industry with extensive use of robots globally as well as in the US is automobile manufacture – and that more than a half-century after their first introduction

the introduction of robots has been both highly uneven and, for the most part, slow. Even in automobile manufacture where robots have been used since the 1960s and have proliferated more than in any other industry, total employment in auto and auto parts in January 2017 was 945,000, compared to the all-time high of 1,004,900 in 1978, or 94 per cent of the industry’s highest employment level

What all this indicates is that job losses and gains do not correspond directly to the increased use of robots. Competition and the ups and downs of the car and truck market continue to be major factors in employment levels along with various methods of work intensification. More broadly, the level of output and sales, i.e. the realisation of surplus value, remains a factor in employment levels in almost any industry. That is not to say that automation and robots do not displace workers. But they have done so along with recurrent economic crises, changes in plant structure and layout, lean production, alternative shift patterns, and other forms of work reorganisation and intensification that have had a strong effect on auto and other manufacturing jobs

Robots are one factor in the loss of manufacturing jobs – though not even the main one given their limited use so far in most manufacturing industries shown in Acemoglu and Restrepo’s and the IFR’s figures. Determining with any precision just how much job loss is due to technology and how much to changes in, and the reorganisation of, work in the past thirty or so years is probably impossible.

So, although technology plays a role, the rate of investment in robotics and automation has decelerated, while that of economic turbulence and work intensification by other means, that is, essentially class war waged by capital, has increased. The major sources of job loss in manufacturing came not from robots or imports, but from the volatile course of the economy as huge numbers of manufacturing jobs were destroyed in the recessions of 1980–2, 1990–1, 2000–1, and 2008–10, and large productivity gains between recessions due mainly to the implementation of capital’s major tools of class conflict de jure: leanproduction methods beginning in the 1980s; work reorganisation; the introduction of ‘alternative work schedules’; the reduction of break time; and the accelerated monitoring, measuring and standardisation of work via computerisation and new surveillance technology

The classic case was the GM-Toyota NUMMI plant in California where, beginning in 1986, the number of seconds of actual work per minute rose from 45 to 57 seconds. While most factories are not likely to meet the 57-second standard, an increase of a few seconds per minute in a plant with a few thousand workers can create hundreds of extra hours of work at no cost to the company and without any change in technology

‘the appropriate measure of technical change is the ratio of current GDP to current-cost capital stock.’61 This, of course, is the economic measure of change, not a measure of the efficiency of the technology, but as such gives us a guide to capital’s investment behaviour. Appendix I produces this ratio for the stock of fixed capital equipment from 1980 to 2015. What this shows is that, while there has been growth in this ratio over the 35-year period from 1980 to 2015, the rate of growth has been slowing down significantly decade-by-decade, nearly grinding to a halt between 2010 and 2015 despite some growth in the economy. During the 1980s this ratio grew by an average of 1.8 per cent a year, itself not all that strong. But in the 1990s the annual rate of change slowed by half to 0.9 per cent, then dropped to 0.3 per cent from 2000 to 2009. From 2010 to 2015, during the period of recovery, the rate of growth in the ratio of technical change all but vanished at 0.08 per cent a year. Looking further into the course of capital investments, we can see that investment in new ‘information processing’ and ‘industrial equipment’, as defined by the Bureau of Economic Analysis (BEA), has not taken the course techno-futurists’ predictions would suggest. As Appendix II on New Private Investment in Equipment, compiled from BEA data, shows, investment in both information processing and industrial equipment has fallen as a proportion of total equipment investment since the early 1990s, while that in transportation equipment has risen – no doubt as a result of the expanding logistics sector. 

Roberts also notes that growth in investment specifically in new technology has decreased in recent years.63 Economic Policy Institute figures show the same trend.64 The economic factors that limit direct investment in technology have also limited spending on Research and Development (R&D),

by almost any measure the advance of new technology in economic terms does not substantiate the technofuturists’ predictions. Given all the gains in technology these futurists describe, the mystery of this poor and declining performance lies in the volatility of the US and worldcapitalist economies since the 1970s and the continuing problem of profitability. Profitability was not strong enough and could not be sustained long enough under these circumstances to justify large and continuous investments in new technology of any kind. The problem was compounded by the rapid rise in corporate debt over these years. As a result, as Roberts argues, ‘this increase in debt means that companies must raise profitability or be forced to reduce investment in productive capacity to service rising debt.’70 It appears they have done the latter. Future investment in the US and world-wide auto industry, the major user of robots, is further limited by the persistence of global overcapacity in car and light-truck production.71 Yet another indication that large-scale investment is not likely in manufacturing in the near future is the relatively low level of capacity utilisation, which has fallen from above 80 per cent in the 1990s to an average around the mid-70 per cent rate since, compared to the mid-to-high 80 per cent level of the 1960s.72 Short of an economic miracle, the pace of automation and the march of the robots in much of industry is likely to be bumpy and slow

While the past is not always a reliable guide to the future, for there to be a substantial increase in investment in automation there would have to be a prolonged period of stable economic growth and rising profit rates. That has not been the case and is not likely to improve as the very slow and drawn-out recovery of the US economy since 2009 shows.

But technology, for all the AI gains or improvements in robots, does not introduce itself to the factory, warehouse, or 18-wheeler. It has to be introduced through actual investment that promises substantial increases in profit rates to the capitalists who advanced the money, and the hope that that will materialise on a scale big enough to bring about the robot revolution in the foreseeable future seems like the biggest piece of futurism of them all.   

Moody quotes end