Those that advocate for Basic Income (BI) often point to the impending automated future as a reason that BI is necessary. Capitalists, they argue, will be replacing jobs with machines at such a high rate that BI will be the only method to ensure the vast amounts of displaced workers can earn an income.
There is a grain of truth to this argument, as Harry Braverman notes in his classic study of the labour process, “Marx points out that generals win their wars by recruiting armies, captains of industry win theirs by discharging armies.” But the technological determinism of those arguing that automation leads to mass unemployment, rests upon an inaccurate reading of how capitalism operates.
The labour theory of value
To make any judgement about the plausibility of full automation it is necessary to have a clear understanding of the basic inner workings of capitalism. Essentially one has to have a theory that can answer at least these two basic questions: Where do profits come from and why do capitalists utilize labour saving devices? Looking at the labour process itself helps to answer these questions.
Workers enter the labour market because they need food, shelter, clothing and other basic necessities. They do not have the ability to attain these things by any method other than selling their ability to work.
Employers hire workers for their ability to work for a given wage in a given period of time. The value of what workers produce at work is more than the value they are paid by employers. This surplus value produced by workers is the ultimate source of profits.
The competitive nature of capitalism requires businesses to vie for market share to survive. Capitalists aim to increase their market share by cutting costs and maximizing their return on investment. Inside the workplace this means capitalists will try to squeeze more productive work from workers for less.
Technology as discipline
Capitalists organize the workplace to convert workers’ time into productive energy. This is done by controlling the knowledge of work skills, organizing the flow and pace of work, and increasing oversight. The continual introduction of technology into the workplace is one of the best methods to achieve these aims.
Machines and tools enable workers to increase their output by making certain job tasks easier. It also allows management a greater ability to ratchet up the pace and oversight of work. As Braverman notes, “machinery offers management the opportunity to do by wholly mechanical means that which it had previously attempted to do by organizational and disciplinary means.”
It is important that we conceptually understand technology and labour saving devices as a form of labour discipline. Technology under capitalism is not merely just a bunch of cool stuff, like robots or computers. Technologies are things with a purpose in relation to labour. Management uses machines to dictate the pace of work and to deskill labour, separating the conception and execution of job tasks. The increased use of effective technology leads to increased output and a lowering of unit prices.
Automation and the logic of capitalism
The mechanization in a given industry means the replacement of human labour. This has a number of different consequences within capitalism. Individual businesses will employ fewer workers, have higher outputs and lower unit costs. So long as they have a technological edge, companies can increase their rate of profit even as they increase the amount of fixed capital costs relative to workers.
This technology allows them to squeeze out more value from each worker in a given period of time. But when a labour saving technology becomes generalized in a sector, and prices fall, the firm’s profit rate will depreciate. The company will be saddled with more fixed costs relative to labour, with the capacity to have higher output levels.
The reason for this is that all machines are fixed capital, the machine’s potential value to the firm is paid in advance. Highly capitalized firms are able to compete on a bigger level, but run the risk of having their capital investment superseded by competitors using a superior technology (imagine investing in a series of highly efficient pneumatic tubes just before the internet). There is also a risk that the market may experience an excess of a given product and the business owner will be faced with idle machinery that it cannot get rid off.
There are countervailing tendencies to all of this. As firms deskill labour and replace workers with machines this can create larger and larger pools of workers who are looking for work (in the context of globalized production chains, the vast pool of labour can extend beyond national borders). This puts downward pressure on the price of labour and encourages firms to hire cheap labour, and put off costly and risky investments in machines.
There is of course another added dimension to mechanization, it is not simply that workers are thrown out of work, but the automation of work creates new avenues of employment (often highly skilled work). Workers are needed to build, design, program, transport, tend to and service machines. Automation itself births a new industry that requires workers. Human labour remains at the centre of mechanization, a point many frequently forget.
Machines are the product of human labour and intelligence. When machines are introduced into a workplace, it is more accurate to conceptualize this as an outsourcing of labour than it is to see it only as a worker being replaced by a machine. Behind the vast layers of robots, machines, tools and code lies the human labour that created all of it.
Automation, even at its most advanced level, has not altered the basic premise of capitalist social relations. Humans are still the only input able to create new value. Humans have a unique and unrivalled set of physical and mental skills that are not simply better than any automated system has produced, but are of a different order.
Efforts to create Artificial Intelligence (AI) have come nowhere close to producing the complexities, nuances and imagination of the human mind. Even the most advanced machines on the planet are conceptually no different than the original Turing Machine, only able to complete a set of pre-programmed tasks (even learning machines are limited in what and how they can write new code).
Within the tech industry there are two broad categories of AI, weak and strong. Weak AI refers to computing systems, which can make decisions within a narrow pre-programmed framework. Machine learning, where computers can learn and write new code would be an example of this, but so would basic computational programs in cars, planes and cellphones. There have been notable advances in this field in recent years, but they have had surprisingly narrow application in the economy.
Strong AI, which could be something either like human intelligence or a machine that is capable of learning beyond pre-programmed limitations, remains theoretical. Computers have essential become faster at taking in and processing data (although the rate at which processing speeds are advancing is slowing significantly).
In the field of robotics, building models to do physical human tasks like walking into a room and making a bed are still along way off. But minus a breakthrough in Strong AI, robots, no matter how advanced, remain simply machines. Marx’s basic description of a machine in relation to human labour remains as relevant for a steam engine as it does for the most advanced computer:
“The machine proper is therefore a mechanism that, after being set in motion, performs with its tools the same operations that were formerly done by the workman with similar tools. Whether the motive power is derived from man, or from some other machine, makes no difference in this respect. From the moment that the tool proper is taken from man, and fitted into a mechanism, a machine takes the place of a mere implement. The difference strikes one at once, even in those cases where man himself continues to be the prime mover.”
The problem with the full automation thesis
As I noted earlier machines are, simply put, human labour crystallized as a fixed cost. They can help speed-up production, make workers work more efficiently and increase output by doing pre-programed job tasks, but they cannot create new value.
On the firm level, once technology advances are generalized in an industry, the competitive advantage a firm has disappears. The firm is left with less human labour and more fixed costs, this tends to lower the rate of profit. As Michael Roberts explains:
“Over time, a capital-bias or labour shedding means less new value is created (as labour is the only form of value) relative to the cost of invested capital. There is a tendency for profitability to fall as productivity rises. In turn, that leads eventually to a crisis in production that halts or even reverses the gain in production from the new technology. This is solely because investment and production depend on the profitability of capital in our modern mode of production.”
The flow of capital will shift from low to high areas of return on investment. Major investments will be pursed in newly developing industries and areas that often employ many workers. Just look at the phenomenal growth happening in the healthcare sector of the economy. As in all industries in capitalism, newly created ones will be pushed down the path of uneven mechanization. The capital necessary for the investments in this mechanization is ultimately derived from exploitation of workers.
The narrative that we need BI because of impending mass automation and mass unemployment, while overblown does touch on two important issues. Will the rate of new employment being created in newly developing sectors of the economy be roughly equal to the rate at which people are losing their jobs due to automation?
And is the dearth of productive investment a passing phase of global capitalism or does it point to something more structural? Will the sustained low profit rates that have lead many capitalists to simply sit on large swaths of cash or drive it into financial speculation continue, and if so for how long?
Currently, North America is experiencing record low rates of productivity growth. The period from 2005 – 2016 is by far and away the lowest it has been in the United States since 1947 (the first year of comparable data). In the manufacturing sector, which utilizes more advanced technology, the numbers are even lower.
Robert Gordon notes that the productivity bump precipitated by the Internet revolution was mostly confined to the period of 1996 to 2004. Businesses restructured their organizations and business patterns to take advantage of these newly formed web technologies (the reduced prices in semi-conductors also helped). But once businesses reorganized to adopt the new technological breakthroughs, further gains in productivity became marginal. As a 2014 study in the American Economic Review noted:
“there is also little evidence of faster productivity growth in IT-intensive industries after the late 1990s. Second and more importantly, to the extent that there is more rapid growth of labor productivity [referring to the 2008 -2010 period] this is associated with declining output and even more rapidly declining employment.”
Some like Ryan Avent, an editor at the Economist, argue that low productivity and high levels of employment does not mean we aren’t in a period of rapid technological progress. He argues that low economic growth discourages investments in labour saving technology, as labour costs are relatively low. Avent fails to see that this is precisely the point. The relationship between technology and its application is determined by the needs of capital.
The relationship between automation and profits
New technology is being generated and utilized in all businesses. But the marginal return on investment is painfully slow, which helps explain the relatively low levels of business investment. This in part reflects the excess of productive capacity that all ready exists in many sectors. Why invest in efficiencies in steel, mining or chemical sectors if you already have unused productive capacity?
A recent study published by the National Bureau of Economic Research that shows industrial robots have depressed wages and replaced workers is cited by many to argue that automation will result in mass unemployment. It concludes that every new robot added in manufacturing to a given commuting zone reduces employment by 5.6 workers. This study is treated as game changer in the debate over automation and its impact on work. However, its results essentially confirm what one would expect. That in a period of high productivity growth, the study takes place between 1990 – 2007, workers would be displaced in given geographic areas as companies reorganized and ramped up their use of labour saving devices. Productivity increases and the adoption of labour saving technology after 2007 drop off a cliff, because over-capacity in the manufacturing sector along with sluggish economic growth discouraged new investments. The authors mistakenly predict future rates of job losses and the rate of the implementation of robots using a period of high productivity growth as the baseline.
The technology that is cutting edge, such as advances in computational intelligence and robotics for the most part remain highly niche. Of course advances in areas such as driverless cars will in all likelihood shake up the transportation sector, but it is impossible to know exactly when and how that will play out. Will job loss be sudden and sweeping in the sector or will that technology be slowly integrated, like navigational systems on planes? What if any applicability does driverless technology have for other sectors of the economy? What new jobs will this technological advance create?
There has been a tendency to treat speculation about the introduction of labour saving technology into every facet of the labour market as a given. Employers, right-wing commentators, and those in the tech sector having been beating the drum about how automation will sweep through all industries, especially the service sector.
Automation in the service sector
Automation in the service sector is more complicated than most commentators are making it out to be. The narrative is that increasing wages in the sector will precipitate the rise of automation, which will displace large swaths of workers in the industry. This narrative is used by the right-wing to argue against wage increases. Parts of the Left have latched onto this as well, saying it is the reason BI is needed.
Nowhere is this argument more present than in the fast-food industry. Its highly taylorized job tasks and deskilled workforce make it the prime candidate for full automation with self-serve kiosks displacing thousands of workers. The reality of this prediction is somewhat different.
Automated kiosks in the industry remain quite expensive ($120,000 to $160,000 per McDonald’s franchise). Even if labour costs rise and the cost of the technology falls there is little evidence that kiosks will reduce total working hours. For many large fast-food chains a huge source of their business comes from drive-thru (McDonalds gets 70 percent of its business via drive-thru). The installation of expensive kiosks will not apply to the vast majority of patrons.
Where kiosks have been implemented it simply shifts labour to the back end of the restaurant. Kiosks take orders faster (people also tend to order more food via the machine), but that just puts more strain on the kitchen staff, and in the end necessitates shifting labour around to accommodate the new workload in the kitchen. Some restaurants that are using kiosks are shifting labour from front-end order takers to implementing table service. Kiosks are susceptible to vandalism and misuse and this still require workers to assist customers.
Self-checkout machines in grocery stores, which have existed for well over 15 years have not seen wide-scale implementation. The machines remain relatively costly in relation to their performance. They still require human minders to assist people. They also create the need for more oversight to discourage shoplifting. In the vast majority of stores where they exist they are part of a mixed check-out system.
Studies have shown in more complicated self-checkout processes (grocery stores as opposed to banks or buying movie tickets) that automation is not reducing wait times nor proving faster than regular checkouts. This is because complex automated self-checkout systems aren’t reducing the labour required in the process, they are simply shifting it to the customer.
Self-service machines for simple functions will see wider adoption, but where they have been implemented in fast-food and other parts of the service sector there is little evidence to suggest kiosks are leading to the doom and gloom scenario of mass unemployment. Many jobs in the service sector involve the performance of emotional labour (by making customers feeling happy, listened to, and valued) or intricate physical tasks. For these jobs, the risk of automation replacing human labour is extremely low.
There is of course the prospect that grocery and retail stores could be totally reorganized. Amazon is looking brick and mortar retail where you take items and remove them and then are automatically billed via your phone. But this requires a level of customer integration into the system that would, for now, only cater to very niche upscale markets.
The hype machine
What we can say is that current narrative around automation and the jobless economy is driven by a lot of hype. This hype is not neutral, it is the structural product of how the tech industry operates. Start-ups and mid-sized companies overinflate their technological promises and innovations in order to attract large chunks investment capital (capital which is struggling to find suitable returns on investment).
Take Uber, for example. They are seen as a cutting-edge tech company that has revolutionized the taxi industry. The reality of their business model is more crude and straightforward. They leverage billions of dollars in venture capital to run an international taxi company using sub-contracted precarious workers. They use their vast capital resources to systematically violate local taxi regulations and force their repeal through political lobbying. By ignoring regulation, downloading risk onto their employees and using low-wage subcontracted labour they effectively undermine their competitors by offering lower rates.
Some studies examining Uber’s business model suggest that fares are only covering 40 percent of Uber’s trip costs, meaning they are operating at a loss. Once the competition and regulations have been defeated they raise their rates from the position of a monopoly. They spend huge sums of money to market themselves as an innovative and revolutionary tech company in order to sustain and attract new investment. But in truth their business model barely differs from other large predatory multinationals who smash into markets by selling products and services at a loss to drive out competitors and force politicians to change regulations to sanction the new economic reality the multinational has imposed.
Automation and capitalism
The debates about automation, job loss and how best to combat them are not new. Marx observed that capitalism, “never looks upon and treats the existing form of a process as final. The technical basis of that industry is therefore revolutionary, while all earlier modes of production were essentially conservative.”
The competitive dynamism of capitalism has pushed firms to innovate to increase output. The introduction of the power loom, steam engine, the blast furnace, electrical power are in essence no different than firms employing driverless cars, machine learning, or high-end robotics. If anything the former set of technological breakthroughs increased productivity at much higher rates.
Automation in the workplace will continue to increase. What is important to understand is that the pace, extent and impact of their adoption will be shaped by the inherent contradictions within capitalism.
Automation will displace jobs and it will also create jobs. It is quite possible the rate of displacement could exceed the rate of job creation. However, if this does happen it will not only produce higher levels of unemployment (driving down labour costs), but it would also mean lower profits and growth, leading to economic turmoil. This in turn will slow the rate of investment, stall innovation, and the adoption of new technologies. As Michael Roberts explains:
“If the whole world of technology, consumer products and services could reproduce itself without living labour going to work and could do so through robots, then things and services would be produced, but the creation of value (in particular, profit or surplus value) would not. As Martin Ford puts it: the more machines begin to run themselves, the value that the average worker adds begins to decline.” So accumulation under capitalism would cease well before robots took over fully, because profitability would disappear under the weight of ‘capital-bias’.”
Capitalism will not produce a fully automated society. Instead workplaces will be automated when it is profitable to do so, not when it is socially useful or even possible. Marx noted long ago that there was a vast gulf between the potential and reality of machines under capitalism:
“The economic paradox that the most powerful instrument for reducing labour-time suffers a dialectal inversion and becomes the most unfailing means for turning the whole lifetime of the worker and his family into labour-time at capital’s disposal for its own valorization.”
Instead of producing a jobless economy automation will produce more crises, more inequality, without actually reducing total work time in society. Automation under capitalism has always been a cudgel used to threaten and exploit workers. The idea that we are headed towards a fully automated future under capitalism rests on a profound misreading about how capitalism operates. It can’t happen and it won’t happen.
This is not to say that automation will not cause problems for workers, that innovation and technological breakthroughs will cease. In fact, quite the opposite, capitalism is if anything dynamic and innovative. Workers in certain industries may indeed face serious challenges of deskilling and job loss, as has always happened under capitalism. To combat this, workers will need strategies that aim to direct the fruits of technological innovation towards the working class, not simply to accommodate themselves to the will of employers.
Those that push for BI as a solution to automation aim to unite employers and workers and move beyond Left and Right to find a policy that can fit comfortably within the existing power relations of our society. Their deterministic view of technological development leads them to see BI as the only solution to the inevitability of full automation. The fact that this viewpoint gets a wide hearing on the Left ,and even in labour, is the result of a generation of defeats and the lack of confidence that workers can organize and win a better world. The Left must break with this millenarian like thinking and begin to situate the discussion over automation within a concrete analysis of capitalism which looks to class struggle for effective alternatives.