A lecture at the London School of Economics, 18 November 2018
In 1991 the nobel prizewinner Herbert Simon conducted a thought experiment: what would the economy look like to Martians?
This was his answer: a bunch of organisations, represented by green blobs, a set of market transactions, represented by red lines…
and a set of blue lines — within the green blobs — representing power structures — but let’s park that issue for a moment.
He concluded that wherever they looked, from Soviet Russia to capitalist America:
“Organizations would be the dominant feature of the landscape. A message sent back home, describing the scene, would speak of “large green areas interconnected by red lines.” It would not likely speak of “a network of red lines connecting green spots.”
It was designed as an critique of the new micro-economics, where the question was being asked in all seriousness: why do firms and states exist at all, when markets could actually organise everything?
At first sight the diagram is missing a category: human beings. But Simon was actually using human beings as the unit of measurement. The blobs are big, and the lines thin, because human beings spend most of their economic time in organisations not markets.
If we bear that in mind we can tell the story of what happened between 1991 and now.
Organisations break up — privatisation, outsourcing. And people spend more of their economic time consuming, individually — both products and financial services.
A lot of tiny green dots appear. A lot more people are actually trading as tiny organisations — businesses.
And people begin to transact more with a certain kind of blob, known as a bank. They transact with Wonga, Bright House, RBS, their credit card provider.
Then the first of two wierd things happens.
Around the year 2000 the amount of red on the map begins to totally overwhelm the green. Between in the five years before the financial crisis the global money supply expands from $27 trillion to $70 trillion. And the Martian economists on board the ship say: “you know what captain, one of these blobs is gonna explode”. And boom on 15 September 2008 it does.
And to stop the system falling apart some other blobs — states — adopt the survival tactic of expanding the supply of red some more: they print $12 trillion in quantitative easing money.
And the captain of the Martian ship says: its a mess so lets go home.
But the ship’s anthropologist says: no, wait — what about the yellow lines?
What yellow lines? say the economists.
Well look, says the anthropologist: there’s millions of yellow lines beginning to appear between people and organisations. Its a new kind of transaction not involving money. People are using digital technologies that didn’t exist in 1991. They are using them to replace every other colour on the model.
So they produce stuff for free and exchange it with other people — that’s a yellow line instead of a red one. But when they do this they also interact differently — using networks: they manage themselves, horizontally — so that’s a yellow line replacing a blue, power diagram.
27,000 people are producing a thing called Wikipedia, so we don’ t know whether to put them altogether into a blob or just call them a network of little dots. We might have to have yellow blobs as welll as green ones.
How much does Wikipedia turn over a year? Says the captain. Wrong question — says the anthropologist. The right question is how much value does its existence deduct from the market economy? How much of the red stuff does it replace. And the answer is about $3 billion.
OK let’s stay, says the captain, and see what happens next.
What happens next depends on whether we understand two things.
1. What’s wrong with neoliberalism?
2. What to do about the yellow lines. Ie what to do with these non-market, non-hierachical forms of production that have emerged alongside it?
The first one’s easy.
Neoliberalism is broken.
You bring in fiat money in 1973 and that allows you to expand credit, over and over, unrelated to the amount of precious metal or indeed economic activity. Credit, derivatives and financial complexity go stratospheric.
But you simultaneously defeat and atomise organised labour, so that wages stagnate. In the USA the male median hourly wage in 2008 was in real terms what it was in 1973. Other developed countries did better for the average; but for the poorest not.
Neoliberalism’s founding promise to the developed world workforce was: we’re going to smash organised labour and make you poorer, less powerful, more dependent. But then, once that the smashing phase is over here is: the credit card, the auto loan, the payday loan, the student loan.
So then you get a boom-bust cycle: credit expands, real incomes don’t. An asset price bubble takes off. It bursts. Asian crisis, dotcom crisis, Lehman Brothers.
The neoliberal solution to these crises? Expand the money supply, cheapen credit, suppress wage bargaining power some more. Extend the implicit guarantee from the state that it will always save banks but usually let steel works go bust.
Neoliberalism feeds off, fosters but then ultimately asphyxiates information technology.
The result of weakening of labour’s bargaining is that entrepreneurs don’t have to innovate high productivity solutions: there’s a cheap labour solution to everything. When I got my first car, and it got dirty, I put it in a car wash. At that time a car wash was a machine. How did five guys with rags undercut the machine? Because the system forces and enables them to work for peanuts.
Though we are now at a point where, as Frey and Osborne tell us in their 2013 research for the Oxford Martin School, 47% of all jobs could be quickly automated, starting with office admin and retail, we don’t automate. We hold conferences about robots, AI, machine learning: but we create millions of low-skill, low pay jobs that probably don’t need to exist. Bullshit jobs, as the anthropologist David Graeber calls them.
But the boom-bust cycles are not a steady state. The first bust, after 2000, wipes out much of the private pension system. The second one, because the banks have to be bailed out, cripples many states with debt and takes a chunk out of the welfare state. The next one — which may be several years ahead of us but if we’re really unlucky not — will take a chunk out of people’s savings: debts will have to be written off, bank bail ins will happen.
My fear is that neoliberalism will end in a way that de-globalises the world. And given the choice I would rather have globalisation than neoliberalism. The Keynesian playbook says you dismantle neoliberalism by attacking financialisation, empowering workers, raising wages, growing the welfare state, re-regulating business, dis-incentivising low-wage, low social responsibility businesses.
But these are only palliative measures, and here’s why.
Once we understand the special nature of information technology we begin to understand the long-term signals the credit system is sending us.
The short term signal is: “hey, you expanded the money supply much faster than real economic activity, you deregulated complex finance, and even assets like houses or shares in Uber can’t expand faster than real incomes forever.”
The longer-term signal is: “hey, infotech is creating forms of economy where there is no monetary value being generated, where there is not enough work to go around, and where the promise implicit in a credit system — more value in future to pay the interest — might not be fulfilled”.
If I am right, creating a more regulated, neo-Keynesian and even a Green capitalism or an entreprenurial state doesn’t solve the strategic problem. You have to explore what a transition might look like beyond an economy based on states and markets.
“The information society,” writes the Oxford professor Luciano Floridi, “has been brought about by the fastest growing technology in history. No previous generation has ever been exposed to such an extraordinary acceleration of technical power over reality.”
But information technology is different. Before exploring how, I think we should reiterate something Floridi insists on. Information is physical.
Norbert Wiener, the founder of cybernetics believed otherwise: “information is information, not matter or energy”. But there is no information without representation, and that does require both matter and energy. Sure information can have dynamics, laws, that are not the laws of matter and energy — but the information layer of the world is physical.
And information is doing something radical to economics. It is corroding the price system.
It was Paul Romer in 1990 who stated the obvious. Information goods are different. If I smoke one inch of a cigarette and you smoke the next inch, (it’s one of those cigarettes you share) we cannot both smoke the same inch. Likewise we can’t both put our car in the same parking space. Physical goods are rival.
But information goods are non-rival — I can listen to an mp3 track at the same time as you. If I read the Guardian website you can also read it. I can take the same digital mockup of a Boeing 787 to two separate factories and use it to build identical planes. Information goods can be reproduced many times over with amounts of energy and matter so small that it looks like they are immaterial.
Digital information is naturally abundant in a way physical goods are not. It wants to be free.
What happens if the cost of reproducing something it close to zero? Under conditions of competition and private ownership its market price should fall close to zero. It is this zero marginal cost effect that underpins the whole information revolution.
The only way to make an mp3 track cost money is to get a good IP lawyer, extend copyright to eternity so that even when the world has disintegrated into atoms Sony Pictures will own the song Diamonds are Forever; and/or build a physical system that incentivises paying for the mp3 track rather than getting it for free.
But maintaining prices using lawyers can’t work forever, and in any case it’s not capitalism. Capitalism has to be granular. Its economic dynamics have to arise spontaneously out of the commodity itself. The price system has to operate.
Tech monopolies built on short-term technical advantage, heavy use of patents to make that which should be cheap appear expensive — will fall.
Once effectively challenged by competitors, Apple’s 95% dominance in digital music was eroded by the sharing services. The result: falling prices. The Information is Beautiful website has calculated that a signed solo music artist would have to get 1,800 plays on iTunes to make the minimum wage, while they would have to get over 1.1 million on Spotify to earn that.
And the near-zero reproduction cost effect is not just a phenomenon in online music. It is collapsing the price of physical things whose information content is high. Bandwidth, storage and processing power have all collapsed in price exponentially in the past 10 years. Deloitte tells us that because of this:
“… we are experiencing rapid advances in the innovations built on top of these core “exponential” technologies. The current pace of technological advance is unprecedented in history and shows no signs of stabilizing as other historical technological innovations, such as electricity, eventually did.”
The problem is we have no idea how to value this exponential impact. Economics are rooted in the concept of scarcity. There is nothing economic that is not scarce, Walras writes in the founding text of marginalism.
The market impact of cheapening products can be measured in the so-called “consumer surplus” — the short term advantage consumers get if their wages fall slower than the cost of basic necessities falls. But it doesn’t capture the whole effect.
As the OECD wrote, in its first major study of the economics of information,
“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 market-clearing mechanism.”
What are these expanded non-market interactions doing? I think they are reducing capitalism’s ability to make the big, 50 year adaptations long-cycle theory tells us about. On any version of long-wave theory — Kondratieff, Schumpeter, Perez — we are overdue a technologically driven takeoff.
The rapid and complete automation of everything can’t happen because there is no high-value synthesis between higher value work, higher wages and higher cost consumer goods.
If we describe industrial capitalism as a complex and adaptive system the source of its adaptablity has been the ability to create high-value industries, high wage jobs and high price consumption needs to replace all the stuff that’s being produced cheaply through automation.
Information technology makes that much harder to do.
So the most imporant impact of info-tech is to make things that were expensive cheap or free, to cascade this effect over into the physical world, and to dissolve the price mechanism, and to make accountancy guesswork.
Society’s short term response is to suppress competition and the price mechanism with gigantic tech monopolies and IP law. But that solution won’t last.
The second big thing information does, is: it allows us to blur the distinction between work-time and life; and to delink work time from wages.
The salariat? We get on a plane to Brussels at 7am, sit elbow to elbow typing into Excel. If it was a factory it would be shut down on health and safety grounds. Yet it is a factory. We are all working. But because nobody asks: when did you start work, we are not paid for the hours. We are constantly at work; and our work and leisure completely blurred. As a result the link between our wages and hours are also broken.
Towards the bottom end of the pay scale there are fast-food workers who will have been exhorted to smile, be cheerful, touch each other playfully as they work; indeed they will exhort each other to do this unprecedented emotional labour because there is a secret shopper who will come in and cancel the bonus of the entire shift if somebody doesn’t smile enough when serving his cappucino.
This blurring of work and life, and detachment of work from wages, I argue, is the result of a society where modular, target-driven work is the norm. It’s quite different from the past 200 years of industrial capitalism and in the process it is producing a new kind of human being just as quickly as the factory did. A workforce with low loyalty, weak ties, skilled at uploading knowledge into their brains and then discarding it; a population with multiple personalities, with much of the self externalised onto a device.
The third thing information does is: destroys organisational hierarchies and makes ownership less important. As soon as we could protest using netoworked, non-hierarchical technolgies we did. The first time we could build virtual utopian communities — from the basic bulletin board to Second Life — we did. Thats because the barriers to entry for collective action have fallen. We can swarm around an objective — be it a software project, a documentary, the minicab needs of a city, a street protest at COP21 — achieve it with minimal organisation commitment to each other.
So we live in an age where, after 200 years of the hierarchy being the most effective form of organisation at work and in society, instead: the network is.
Let’s consider why. First, because information work tends to be modular. On a production line for me to do my work, you next to me have to turn up. In a video design house, there’s often a whiteboard with yellow post-it notes: small teams form around one of the tasks, record their progress, leave a new post it note for another team to work on.
Once you get modular work, plus reliable digital communications, you can dis-aggregate the company.
In the physical world, standards, tooks and skills are all separate. So the 1950s pres operative enters the factory armed with his ruler, to measure fine tolerances on the metal parts he’s stamping; his lathe and the knowledge needded to work the lathe. Today the tools of engineering are virtual. Catia, the software package you use for virtual manufacturing: what is it — a tool? A standard, a skill? All we know is that, if somebody discovers and corrects a fault with it, anywhere in the world, or adds a widget of functionality, by 9am tomorrow every user can have it.
So knowledge is becoming social. Not just the free social knowledge embodied in Wikipedia and other open source projects.
Digital information means that, regardless of IP and commercial secrecy, innovations in technique tend to spill over across the whole of economic life very quickly.
Another thing information does is make physical things very efficient: so if you do virtual manufacturing, instead of over-the wall engineering where the design is passed from a design office to manufacturing plant, this is what happens. A massive reduction in defects during the prototyping process, and a massively greater level of assurance for the product. The engineer who worked on the tailfin of the Tornado told me they tested 12 diffent stress loads on it. When they designed the Typhoon, its replacement, at the other end of this guy’s career, they did 186… million.
The airliner above our heads still looks like a metal bird but it is an information product — from the virtual manufacture, the realtime engine data, to the GPS guidance system to the atomic structure of the fan blade grown in a vacuum from a single crystal of metal alloy. It is an old thing made new by information.
So let’s address the cumulative effects: networked work, horizontal work, distributed organisations, tools that act like standards and come for free; free upgrades happening every minute of every day. Massive improvements in the design process, the prediction of failure, the maintenance process. Socialised knowledge.
If you make a list of it all it does begin to look, as the Harvard law professor Yochai Benkler puts it, like a new mode of production.
Put in Marxist terms this is means of prodution straining at the boundaries of the social relations of production. It’s not a scenario Marxists of the 20th century would have recognised, because it creates the possibility of a rapid transition towards abundance in some sectors. It is granular and spontaneous in a But as I have outlined in the book, Marx himself imagined it.
In a set of notebooks not published in the West until the 1970s, Marx, writing in 1858, imagined the emergence of a “general intellect”; of knowledge becoming a factor of production and a more powerful source of productivity and profit than the simple exploitation of labour. In the Grundrisse he explores the possiblity that this will lead to a phase where machines can be made for free that last forever. These would, he said, constitute the ideal machine for capital, because they require no labour to produce and therefore make everything you spend above labour costs profit.
If you then have a society based on wages derived from work, but you reduce necessary work to a minimum and promote un-necessary work, you are creating a contradiction within capitalism that, Marx says will “blow its foundations sky high”.
I am not so keen on blowing foundations. I will settle for gradually replace.
But I do think there is a rational grounds to argue that information technology
… by corroding the price mechanism
… by blurring work and life, and delinking hours of work from wages
… by replacing hierarchies with networks
… by triggering an upsurge in collaboration, non-managed work and non-market exchange
… lays the basis for a transition to postcapitalism.
In the book I lay out some of the ways you would manage this:
· the basic income paid for out of taxation as a one-time subsidy for the rapid automation of work
· the re-regulation of finance to defuse the timebomb of a credit based system in a world where the market sector begins to disappear
· the aggressive clearance of a space for commons-based production, a solidarity economy — through regulation: we do this as aggressively as early capitalism cleared a space for the factory system, transport infrastructure and central banking
But the details are not as important as the framing. If we do free ourselves of economic necessity, as Keynes imagined in 1929, we’ll unleash a human revolution whose outcome will be unimaginable by us.
So I want to end on something more prosaic: what does it mean for economics.
It was the outraged conservative writer Tim Mongomerie who commented on my book: “this book is obsessed with obscure Soviet economists”.
Yes it is, because it was they who had the first opportunity in both theory and practice to grapple with transitional economics, where transition does not just mean towards low carbon but towards an economy where some of the basic dynamics of the market are suppressed.
One of the first things they learned of course is the market can’t be suppressed in conditions of scarcity without going backwards.
But the debates among early Soviet economists are richer than that. Let’s look at our diagram to discover why.
If you try to use money as a unit of measurement in Herbert Simon’s diagram it doesn’t work.
Think about it. The green blobs, the organisations, to be measured on the same scale as the red lines, then the red lines have to be income/expenditure, and each organisation’s size has to be its turnover.
That’s fine until you introduce yellow lines: production of free stuff using networks of decentralised people who are not paid wages. This is the problem modern accountancy has with infotech. It just looks like a giant deduction from the market economy.
What you need is a unit of measurement that can measure paid work against unpaid work, commodities in a marketplace against free stuff given away or, as in the case of creative commons, shared on a conditional basis.
This was Von Mises supposedly killer argument in the socialist calculation debate of the 1920s and 1930s. The plan can’t decide to produce a ton of wheat or a ton of iron unless there’s a common unit of measurement for both. And if one is produced in the market economy, and the other in the plan, you’re going to have chaos.
The postcapitalist project does not happen through centralised planning. Its a different route out of capitalism.
But it faces the same problem: how do we measure, predict, understand the dynamic interactions between the state, the market and the commons.
I think, in the transition to come, it is all forms of economics derived from marginalism that have the “calculation problem”: you are not calculating like with like. And this is not just a problem for the right. The left-wing economist Anne Pettifor last week, in a response to my book, argued that people who work for open source projects should be paid wages.
If you use, as Simon implicitly did, labour as a unit of value it helps solve the problem of measurement.
Mises admits in Economic Calculation and the Socialist Commonwealth that if the labour theory of value is correct, there is no calculation problem. Amounts of labour expended in the planned economy — bearing in mind that skilled labour is just a multiple of unskilled labour — can be measured against amounts of labour in the market.
At the time of the calculation debate, most of its left-wing participants opposed the labour theory of value. Their point was to prove that the perfect plan is, in theory and given adequate computing power, is just as good as the most efficient market.
That’s not my concern. My concern is to understand and manage a long transition during which more free stuff gets pumped into the economy; more work is done for free; more positive externalities are generated from the network effect.
I think the best theoretical framework for doing this is to use the labour theory of value. The labour theory is the only framework that allows you to measure a sector producting free stuff agaisnt a sector producing for the market. I’m not saying here “the labour theory is better than marginalism for understanding capitalism”. It’s not: if it was, modern econometrics and accountancy would have grown out of Marx, not Walras.
I am saying, only the labour theory allows us to understand a process whereby the price mechanism corrodes.
And like the most perceptive of Soviet economists, Evgeny Preobrazhensky, murdered in 1938 on Stalin’s orders, it will allow us to understand that transitional economies have dynamics, just as much as a stable capitalist economy has dynamics.
Of course everything I’ve said here could be wrong: capitalism might survive, but if it does so it will have to find a new kind of market within the microcosms of human interaction — just as the 1910s created a market for celluloid films and shellac records inside the leisure time and brains of the middle classes.
But it has to be a high value synthesis not a low-value one. If it survives simply by suppressing the price mechanism, reducing the reprocution mechanisms of capitalism to asymmetries of power and information, it will look more like neofeudalism.
The more you interface with the avatars of neofeudalism — the silicon valley workers trapped in their coaches as the residents of the Bay Area try to smash their windows, the Russian oligarch in his darkened limo, the vast, fake-tanned Downton Abbey that surrounds the super-rich, the more you realise how unhappy they are. They are as unhappy as the fantasy overlords in Game of Thrones, a series in which nobody to my knowledge has ever smiled, other than cruelly.
Postcapitalism will set them free.