PostCapitalism-_Paul-Mason_-_z-library.sk_-1lib.sk_-z-lib.sk_

unprecedented in history and shows no signs of stabilizing as other historical technological innovations, such as electricity, eventually did.’²⁶

It has become commonplace to think of information as ‘immaterial’. Norbert Wiener, one of the founders of information theory once claimed: ‘Information is information, not matter or energy. No materialism which does not admit this can survive at the present day.’²⁷

But this is a fallacy. In 1961, IBM physicist Rolf Landauer proved, logically, that information is physical.²⁸ He wrote: ‘Information is not a disembodied abstract entity; it is always tied to a physical representation. This ties the handling of information to all the possibilities and restrictions of our real physical world, its laws of physics and its storehouse of available parts.’²⁹

Specifically, he showed that information processing consumes energy and it should be possible to measure the amount of energy used in deleting one ‘bit’ of information. In 2012 a team of scientists built a tiny physical model demonstrating ‘Landauer’s Rule’.³⁰

So information is a product that costs energy to produce and exists as matter. Bits take up room in reality: they consume electricity, give off heat and have to be stored somewhere. Google’s famous Cloud is in fact acres of air-conditioned server farm space.

But Wiener was right to understand that the product of a computing process is qualitatively different from other physical products.

The real wonder of information is not that it is immaterial but that it eradicates the need for labour on an incalculable scale. It does all the things a machine does: it substitutes cheap labour for skilled labour; it eradicates labour altogether for some operations, and it makes new operations possible that no previous forms of labour could have achieved. The new information produced by a computer has a use value, or utility, massively in excess of its component parts.

But the amounts of labour value embodied in information products can be negligible. And once knowledge becomes truly social – as Marx imagined with the concept of the ‘general intellect’ – some of the value is contributed for free, as follows:

Information goods naturally leverage general scientific knowledge

Their users feed back, in realtime, data that allows them to be improved, for free

Any improvement in knowledge somewhere can be implemented in every machine deployed everywhere, immediately.

For example, the internet protocol, invented in 1974 and published for free, is a ‘standard’, not a product. But it is not the same as, say, the safety standard the garment factory is supposed to adhere to. It is more like the electricity grid a factory draws power from: it is materially useful. And it is free.

What happens if you insert some of this free machinery into the labour theory of value? Marx, it turns out, had actually thought this through.

In the Grundrisse, Marx says: if a machine costs 100 days’ worth of labour power to make, and wears out in 100 days, it’s not improving productivity. Much better to have a machine that costs 100 days but wears out over 1,000. The more durable the machine, the smaller the amount of its value chipped off into each product. Taking this to its logical extreme, what you ideally want is a machine that never wears out, or one that costs nothing to replace. Marx understood that, in economic terms, they are the same thing: ‘If capital could obtain the instrument of production at no cost, for 0, what would be the consequence? Surplus value [would be increased], without the slightest cost to capital.’ He lists two ways in which, even in the nineteenth century, capitalism was getting just such a free hit: from the reorganization of workflow, and through scientific advances. Marx then writes: ‘If machinery lasted for ever, if it did not itself consist of transitory material which must be reproduced … then it would most completely correspond to its concept.’³¹

We should shudder in awe at this incredible insight, written by gaslight in 1858: that the ideal form of a machine is one made of material that does not wear out, and which costs nothing. Marx is not here speaking about the immaterial but of non-transitory material: that is, something that does not degrade.

Machines where parts of the value are input for free by social knowledge and public science are not alien concepts for the labour-theory. They are

central to it. But Marx thought that if they existed in large numbers they would explode the system based on labour values – ‘blow it sky high’, as he says in the Fragment on Machines.

The worked example Marx uses in the Grundrisse makes it clear: a machine that lasts for ever, or can be made with no labour, cannot add any labour hours to the value of the products it makes. If a machine lasts for ever, it transfers a near-zero amount of labour value to the product, from here to eternity, and the value of each product is thus reduced.\*

Of course, in reality, physical machines do not yet last for ever; but what we’ve seen in the past fifteen years are machines whose utility derives from the information used to run them, design them or make them. And only the labour-theory can properly comprehend what it means economically, if the world of physical objects becomes alive with information.

WHEN MACHINES THINK

In 1981 I worked for a few months as a press operator in a small engineering factory next to the River Mersey. The stamping press worked on a mixture of electricity and compressed air: when you pulled a lever, it hammered a machine tool down on to a metal disc, bending it into shape. My job was to put the disc on the die, pull the lever and get my fingers out of the way before the guard came down. It was unskilled work, about ten repetitions per minute, and there were always a huge number of defective discs. There was no information feedback mechanism in the stamping press at all; and nothing was automated bar its single physical hammering motion.

Above me were two tool setters, semi-skilled men who fixed the tool in the machine and realigned it every few hours. In the next room were the skilled metalworkers who made the tools. They never spoke to us. However, what we all shared was this: without the skill of our fingers and a keen eye for defects, inherent danger and faulty processes, nothing in the factory worked.

Today, metal stamping is almost completely automated. The operation is first simulated on a computer, with thousands of datapoints on the metal

modelled mathematically to understand the stress placed on the metal. Then a 3D design is fed directly into a computer, which controls the machine. The die and the machine tool are often much more intricate than the one I used in 1981; and now they are positioned by laser beams, allowing far greater accuracy. If something goes wrong, the computer controlling the machine knows about it. When the part comes out of the machine, it is picked up by a robot, analysed and placed precisely where it should go next. And when the tool needs changing, a robot arm does it.

Such machines can finish in an hour what we finished in a day, free of defects and with no fingertips accidentally left on the floor – because there are no workers. What makes this possible is numerous applications of IT: computerized analysis and 3D design in the preparation; realtime feedback and analytics during the process; and data retention to aid future refinements of the process. Researchers are now focusing on ways to automate the production of the tools themselves and even de-skill their design using computer models.

So the whole machine is alive with information and so is the product: automated factories require even small parts to be identifiable individually, through tags and numbers. The press can add these as well.

We have lived through a revolution in one of the most basic operations in industrial capitalism: metal bashing. But nobody has bothered to theorize it – the academic literature on the automated metal press belongs in the engineering department, not in economics.³²

And that’s because, as we’ve already seen, nobody knows how to measure the value of information economically. You can see the impact of buying an automated press on the company’s bottom line; you can value the 3D designs and bespoke computer programs as assets, but as the SAS Institute research showed, you are basically guessing.

The labour-theory enables us to do something better than guess. It allows us to think of software as a machine; the information (3D designs, programs, monitoring reports) as finished labour in the same way the tools and metal dies are. And it allows us to trace the process by which the ‘zero marginal cost’ effect of pure information goods spills over into the world of physical products and the machines that make them.

My press shop in the early 1980s was staffed by maybe twenty-five workers. For a similar-sized operation today you would need fewer than five. The crucial difference is made by software, laser sensors and robotics.

The value of this industrial software is entirely reliant on the patent law that prevents it being used and replicated for free. Though it’s harder to pirate than, say, the DVD of a feature film, the principle remains the same: the reproduction cost of industrial software is zero; the value added is contained in the work done to attach it to specific machines and processes.

Though a machine shop smells and sounds the same as it did thirty years ago, it is as different from the one I worked in as an iTunes track is from a vinyl record.

FREE MACHINES IN A MIXED ECONOMY

We’ve seen what happens if you inject zero marginal cost products into the price model: it breaks down. We must now model what happens if you inject free machines into the cycle of capital investment.

For the sake of clarity I’m using an ultra-basic model here with all its attendant dangers of oversimplification.

Let’s say there are four lines on a spreadsheet modelling the inputs to an economy in terms of labour value. The units could be millions of hours of labour time. Let’s say the labour transferred to the final product in Period #1 looks like this:

Capital: 200

Energy: 200

Raw materials: 200

Labour: 200

The capital line of the spreadsheet is always different in the labourtheory, because machines transfer their value into the product over several years, while in the other three lines the value is consumed in the current period. So that capital line might represent machinery etc. costing 1000,

chipping off 200 units of value each year into the total output over a fiveyear lifespan.

Now let’s do something drastic to the capital line: let’s assume it represents a single machine that lasts for ever. In the labour-theory that immediately slashes the labour transferred from the capital line to zero, for ever. No matter what the initial outlay was (in terms of hours spent to make the machine), if it lasts for ever it transfers almost no value – because even a billion divided by ‘for ever’ is zero.

The total labour hours transferred by all factors of production to the final output now fall to 600 hours (keen-eyed Marxists will spot I am not including profit in this model, but see below).

Now we run the spreadsheet over time: in Period #2, the zero-effect in the capital line spills over and reduces the number of labour hours transferred to the final product – because the hours needed to reproduce labour are reduced. If you keep on running this model, without doing anything to counteract the downward pressure on labour inputs, pretty soon it is not just capital costs that are zero, but labour/raw material costs fall rapidly. Of course, in a real economy machines don’t last for ever. But insofar as they are pervaded by information, a part of the labour expended to make them ceases to circulate in the old way. The value vanishes.

Let’s run this spreadsheet down to an end-state, over several time periods where capital and labour get shrunk towards zero marginal reproduction costs. Now the labour expended is mainly focused on providing energy and physical raw materials. If this happened in real life, because the law of value operates beneath the surface, it would be possible for the price system to carry on as normal, trying to calculate the marginal utility of things. As prices fell, corporations might react by trying to impose monopoly pricing – to stop the value embodied in the machine and its product falling towards zero. But mainstream economics would be puzzled. It would seem like whole swathes of economic activity were being ‘stolen’ from the normal market framework.

And even though we are far away from the pure information economy modelled crudely here, we can already feel these effects in reality:

monopolies are arising to prevent software or information goods becoming free; accounting standards are becoming garbled as companies resort to valuation guesswork. There are attempts to stimulate wage growth, while most of the inputs to labour can now be produced with less labour.

In its first major macro-economic study of the internet, in 2013, the OECD admitted: ‘While the internet’s impact on market transactions and value added has been undoubtedly far-reaching, its effect on non-market interactions … is even more profound. Non-market interactions on the internet are broadly characterised by the absence of a price and marketclearing mechanism.’ Marginalism supplies no metric, no model to understand how a price economy becomes a substantially non-price economy. As the OECD team put it: ‘Little attention has been paid to nonmarket interactions since few, if any, well-defined and well-grounded measurements have been commonly adopted.’³³

Let’s admit, then, that only marginalism enables us to build price models in a capitalist society where everything is scarce. In return, let us insist: only the labour-theory allows us to build models whereby zero-cost effects begin to cascade over from information into the sphere of machines and products, and from there into labour costs.

Once you introduce free machines and products into a model of capitalism that runs over time, even a crude one like this, it is as electrifying as introducing the figure zero into mathematics.

The four-line spreadsheet outlined above should really have an extra row for profit and instead of simply declining, each value should grow by perhaps 3 per cent a year, representing GDP growth. But suppose you did add profit and growth? Once the zero marginal cost effect kicks in, there would have to be tremendous profits and growth to offset the eventual impact on labour costs. In other words, there would have to be new industrial revolutions every fifteen years, very rapid nominal growth and ever bigger monopoly firms.

But that can’t happen.

Capitalism worked as long as capital could move, when technological innovation brought lower costs in one sector, to sectors with higher wages, higher profits and higher-cost inputs. Capitalism does not self-reproduce in this way when the outcome is zero costs.

This simplified model also allows us to see really clearly how economics in a zero production cost society quickly comes to centre on energy and raw materials: they become the sector where scarcity still rules. Later we’ll explore how modelling the disappearance of labour value like this could translate into the actual design of strategies for transition; and how issues around energy fit in. For now, however, let’s look at how capitalism might evolve to meet these economic challenges.

WHAT WOULD INFO-CAPITALISM LOOK LIKE?

The rise of free information and free machines is new. But the cheapening of inputs through productivity is as old as capitalism itself. What stops capitalism from becoming a systemic race to the bottom is the creation of new markets, new needs, and raising the amount of socially necessary labour time used to meet these needs (fashion instead of rags, TVs instead of magazines); this in turn raises the amount of labour time embodied in each machine, product or service.

If this inbuilt reflex could work properly, faced with the information revolution, what we’d get is a fully fledged info-capitalism. But here’s how it would have to work.

It would have to stop the price of information goods falling, by using monopoly pricing: think Apple, Microsoft and Nikon/Canon on steroids. It would have to maximize the capture of externalities by corporations. Every interaction – between producer and consumer, consumer and consumer, friend and friend – would need to be mined for value. (In labour-theory terms, our non-work activity has to be turned into work contributed to the corporation for free.) A thriving info-capitalism might seek to maintain artificially high prices for energy and physical raw materials, through hoarding and other monopolistic behaviour, so their cost fed through into higher average necessary labour time to reproduce labour. Crucially, it would have to create new markets beyond production, in the field of services. The 250-year history of capitalism has been about pushing market forces into sectors where they did not exist before. Info-capitalism would have to take this to its extremes, creating new forms of person-to-person micro-services, paid for using micro-payments, and mainly in the private sector.

And finally, for info-capitalism to succeed it would have to find work for the millions of people whose jobs are automated. These could not be in the majority low-paid jobs because the traditional escape mechanism needs labour costs to rise: human life has to become more complex, needing more labour inputs, not fewer, as in the four cyclical upswings described by longcycle theory.

If all these things could happen, info-capitalism could take off. The elements of such a solution are there in modern economies: Apple is the classic price monopolist, Amazon’s business model the classic strategy for capturing externalities; commodity speculation the classic driver of energy and raw material costs above their value; while the rise of personal microservices – dog minding, nail salons, personal concierges and the like – shows capitalism commercializing activities we used to provide through friendship or informality.

But there are clear structural obstacles to making this work.

First, the normal escape route – innovation creates expensive new technologies that replace info-tech – is blocked. Information is not some random technology that just came along and can be left behind like the steam engine. It invests all future innovation with the zero-price dynamic: biotech, space travel, brain reconfiguration or nanotechnology, and things we cannot even imagine. The only way you could remove the information effect from these coming technologies would be, as in Frank Herbert’s sci-fi novel Dune, to ban computers and replace them with expensive human experts in calculation.

The second obstacle is the scale of workforce redesign. In Marx’s time, there were 82,000 clerical workers in the USA, 0.6 per cent of the workforce. By 1970, on the eve of the info-tech revolution, there were 14 million – almost one in five workers.³⁴ Today, despite the automation and disappearance of all kinds of brainwork jobs – such as bank teller,

shorthand typist, comptometer operator and the like – ‘office and admin support’ remains the biggest job category in America, with 16 per cent of the workforce.³⁵ The second category is ‘sales’, with 11 per cent.

In 2013, a study by the Oxford Martin School suggested 47 per cent of all jobs in the US were susceptible to automation. Of these, it was admin and sales that stood the highest risk. They predicted two waves of computerization over the next twenty years: ‘In the first wave, we find that most workers in transportation and logistics occupations, together with the bulk of office and administrative support workers, and labour in production occupations, are likely to be substituted by computer capital.’³⁶

In the second wave, it is everything relying on finger dexterity, observation, feedback, or working in a cramped space that gets robotized. They concluded the jobs safest from automation were service jobs where a high understanding of human interaction was needed – for example, nursing – and jobs requiring creativity.

The study provoked an outcry along familiar under-consumptionist lines: robots will kill capitalism because they will create mass underemployment and consumption will collapse. That is a real danger. To overcome it, capitalism would have to greatly expand the human services sector. We would have to turn much of what we currently do for free, socially, into paid work. Alongside sex work we might have ‘affection work’: you can see the beginnings of it now in the hired girlfriend, the commercial dogwalker, the house cleaner, the gardener, the caterer and the personal concierge. Rich people are already surrounded by such post-modern servants, but to replace 47 per cent of all jobs this way would require the mass commercialization of ordinary human life.

And here’s where you hit the third obstacle – what philosopher André Gorz called the ‘limits of economic rationality’.³⁷ At a certain level, human life and interaction resist commercialization. An economy in which large numbers of people perform micro-services for each other can exist, but as a form of capitalism it would be highly inefficient and intrinsically low-value.

You could pay wages for housework, turn all sexual relationships into paid work, mums with their toddlers in the park could charge each other a penny each time they took turns to push the swings. But it would be an economy in revolt against technological progress.