Wednesday, 7 June 2017

The Rise and Fall of Competition

The phrase "data is the new oil" has been around for over a decade. The analogy doesn't really work. The vast majority of data (literally more than 99.99%) is close to worthless, its availability simply reflecting the negligible cost of production, distribution and storage. There is no obvious limit to data, so the idea of "reserves" is meaningless, while its intrinsic value isn't obvious. In contrast, a barrel of crude oil can be refined into a specific set of products with predictable qualities. Data is closer to poor soil than valuable carbon deposits. For this reason, "Big Data" - the promise that hitherto inaccessible potential can be extracted like nuggets of precious ore from a spoil heap - is important in preserving the illusions of the present. The inevitable result is hype, of which two choice examples would be the claims that "The value of data goes up every day AI advances" and "Data will become a currency". The first is the new alchemy, transforming base metal into gold. The sorry tale of IBM's Tay chatbot suggests that training AI with rubbish produces more rubbish. The second claim is even more incredible as data is not consistent and heterogeneous like an actual currency or a precious metal. How many Web cookies exchange for a Fitbit step?

Like Big Data, the Internet of Things (IoT) assumes that there is value in data and that this value will increase the more data we produce. But if the world's existing stores of data are largely worthless crap produced by humans, why would we expect the digital droppings of limited function machines to be any better? The monetisation of the IoT comes from commodifying our own exhaust and selling it back to us, in the form of a premium for a connected device, not from exploiting an otherwise wasted surplus. Scepticism about the IoT is good, but it also carries its own ideological load. For example, in a Guardian essay, Adam Greenfield of UCL divides his critique between "our bodies (where the effort is referred to as the 'quantified self'), our homes ('the smart home') and our public spaces ('the smart city')". Not only does this accept the functional framing of IoT evangelists, but it echoes a view of the world as discrete domains that dates back to Ancient Greece, though a view that originally emphasised the oikos and the polis (or demos) more than the self (true individualism - atomisation - is a product of capitalist modernity). Another way of presenting the tripartite division is in terms of exploitation, competition and public order. That is: the equivalence of a behavioural surplus with labour; the threat of the market intruding into the home; and the politicisation of urban space.

Put like that, you can see how the critique potentially lends itself to a reactionary as much as a progressive interpretation. The quantified self can be seen as either the exploitation of leisure time (i.e. unpaid labour) or as the market over-stepping traditional boundaries (the money-lenders have taken over the temple that is my body). The smart home can be theorised as malign surveillance by intrusive corporations or as the defence of private property against expropriation. The smart city can be imagined as a public space privatised for the benefit of those same corporations, or as a totalitarian dystopia controlled by unelected bureaucrats through unforgiving algorithms. What these contrasting visions share is a belief that data is of great value in aggregate. The algorithms may be intentionally or unintentionally biased, but there is no doubt that they are effective. But what if the data is largely rubbish and these contrasting visions are just fantasies; the one of human value, the other of social control? The popularity of the word "smart" with both liberals and conservatives (consider its regular use by both Hillary Clinton and Donald Trump) should give us pause for thought.

A central insight of Marx was that capitalism could not but help create the conditions for its own supersession, essentially by concentrating labour and obliging it to develop as a class for itself: the "revolutionary combination" of the proletariat. But this dialectic, famously outlined in The Communist Manifesto, depends as much on competition between labour as on its association: "The essential conditions for the existence and for the sway of the bourgeois class is the formation and augmentation of capital; the condition for capital is wage-labour. Wage-labour rests exclusively on competition between the labourers. The advance of industry, whose involuntary promoter is the bourgeoisie, replaces the isolation of the labourers, due to competition, by the revolutionary combination, due to association. The development of Modern Industry, therefore, cuts from under its feet the very foundation on which the bourgeoisie produces and appropriates products. What the bourgeoisie therefore produces, above all, are its own grave-diggers. Its fall and the victory of the proletariat are equally inevitable."

The standard explanation for the failure of capitalist crises to lead to immiseration and revolution, at least in the developed world (and ignoring the role of imperialism), is that technological innovation has led to an aggregate growth in prosperity that has allowed the needs and desires of the working class to be satisfied by a rising tide of material gain. In this view, liberal democracy, with its corollaries of social mobility and meritocracy, has acted both as a political pressure valve and as an amplifier of entrepreneurial and inventive talent. Central to this is the idea of competition and the marketplace of ideas. In the Early Modern era, competition was still seen as synonymous with conflict and contention (Milton's Satan is the first great competitor as much as the first great rebel), and thus the enemy of harmony and divine order. Under capitalism it was recast first as a virtue and then as the driving force that produces the harmony of equilibrium. If classical liberalism from Locke to Mill emphasised the moral benefits of competition, finding a philosophical and quasi-religious justification for what had hitherto been seen as vicious, modern strains of liberal thought, from the "big staters" of social market liberalism to the "little staters" of Hayekian neoliberalism, have sought an ethical foundation revealed through social science, which has meant an emphasis on data and calculability.

Having transformed from the profane to the sacred, the idea of competition reversed course once more as religion was supplanted by science. This latter transition was mirrored in liberal ideology by a shift from the moral impetus of producers to the amoral preferences of consumers: from Samuel Smiles Self-Help to Helen Gurley-Brown's Having It All. This isn't just a reflection of the expansion of the market economy and the profusion of commodities, but a reframing of all social relations in terms of competition and exchange. Alienation has given way to FOMO, while commodity fetishism has become an unironic way of life. This has been accentuated by information technology since the 1980s, most famously in the form of the Internet meme: the commodification of an idea as an infinitely-replicable image or slogan. If the Internet has extended the association of labour to a global scale, however loosely, what has it done to competition between labour? Objectively, information technology has heightened labour competition, through offshoring and market disruption, which can be seen in the decline in labour's bargaining power and thus wages.

As Jeremy Gilbert put it in The End of the Long 90s: "Capital always needs to find ways of ensuring that that productive potential is channelled exclusively in the direction of producing and circulating marketable commodities, and not in the rather predictable direction of producing new and democratic forms of common life. From this perspective, there’s no contradiction between a techno-cultural revolution which really does, in principle, enhance our collective creative and organisational capacities beyond the wildest dreams of earlier generations of utopians, and a culture in which [we] are constantly admonished to view each other ultimately as competitors". In other words, the "potent collectivities" of the online world, as Gilbert terms them, which have fed the accelerationist and postcapitalist optimism of many on the left, are about extending and normalising competition as much as facilitating association. Subjectively, information technology has made competition (to complete a level, to be the first to comment) a constant in our lives, and that in turn entails the constant fear of failure.

Viewed in this way, much of the liberal criticism of the tech titans, particularly in respect of the supposed subversion of democracy, is simply the age-old defence of private property rights, which means setting limits on the scope of the market. The modern clamour for we, the people to secure ownership of our data might seem an odd cause for The Economist to back, but it is simply the latest chapter in the assault on the feudal order (now reconstituted in Google and Amazon) and a continuation of the thought of John Locke. Likewise, much left criticism of digital political economy is stuck in the past, albeit a more recent twentieth century paradigm, from Evgeny Morozov's plea for a socialised data populism ("Data populists must seize our information – for the benefit of us all") to Paul Mason's autonomist imagining of a tendential fall in the price of capital goods leading to a liberating abundance. "The product is you" was adopted by many on the left as an insight into the nature of the digital economy because it echoed the established idea of the surplus value of labour and its expropriation by capital. But the phrase also obscured the competitive dynamic at the heart of the online. Perhaps the true value of Twitter and Facebook is the competition for esteem, not the supposed insights to be gleaned from their databases and sold to advertisers.

Data isn't the new oil because it isn't a naturally scarce but useful resource whose cost of production is high enough to sustain fat profit margins. The chief characteristic of data is that it can be created by anyone (a comatose patient in a hospital hooked up to a monitor, for example), with the result that it is for all practical purposes ubiquitous and free. The persistence of such a misleading analogy is surely suggestive, particularly when there are far more suitable substitutes for oil's economic and geolpolitical role, such as fresh water in an increasingly climate-challenged world. What I think it points to is the importance of maintaining the fiction that competition is necessary to the operation of the economy and thus to liberal democracy more generally. To this end we have created a social environment in the West of constant, performative competition and have insisted, against all evidence to the contrary, that our endeavours produce a stream of new value. The more our economy is defined by monopoly and the sweating of exhausted assets (hello Sergeant Pepper), the more we must project the ideal of competition onto the trivial and the marginal.


  1. The social environment of performative competition and its stream of value.
    May I ask how you assess this stream of value against the labour theory. Is it a sort of 'fictional' value, acting as the analogue of fictional capital, or does it point to how creating and harnessing Gilbert's 'potent collectivities' have always run the show. Adam Smith's Butcher et al does not act from his own self interest, but his desire to be esteemed. In this way your identification of a state of perpetual anxiety for subjects competing with one another amidst an internet of things is similar to Latour/Beck's risk society, where we all 'live dangerously'.

    1. Marx's labour theory of value addresses how actual value, i.e. a productive surplus, arises in a capitalist society. His distinction between real and fictitious capital concerns how future productive value is appropriated through monetary claims. The common point is that there is real value involved, even if it relates to production that has yet to occur.

      Polanyi makes a different distinction when he talks of fictitious commodities. In treating land and labour as market commodities that can be traded, we introduce competition into a physical and social realm where is does not naturally arise. For example, the quality of a piece of land might once have put different peoples into competition for it, but modern planning regs put different pieces of land into competition with each other independent of their ownership. This is not a theory of value but of how value forms are extended to social and material phenomena.

      The crossover between the two (and my starting point) is the creation of a labour market that pits workers into competition with each other. Historically, this concerned actual value - i.e. a real labour surplus. In contrast, non-wage labour (e.g. hobbies or recreation) produced no alienable surplus, hence competition (who could grow the biggest leek) was incidental for most people. The great claim of the digital economy is that all of this activity now produces a valuable surplus, namely our data trails, and that all must produce and compete to maximise the aggregate benefit.

      My suggestion is that this is largely not true. Most data are worthless. Most of those that are monetised are of distributive not productive value (e.g. aiding advertising spend, which merely shuffles transactions), or they simply serve monopolists and rent-extractors (e.g. the sharing economy). There is certainly data that can be used to generate productive value, but it is very little as a proportion of the whole. If no one used Facebook for a month, it wouldn't hurt the economy.

      Polanyi also makes an important point when he talks of the "embedded" nature of markets, referring to the substantivist theory of the economy - i.e. economies are plural and heavily conditioned by social norms - as distinct from the formalist theory in which the abstractions of utiliy maximisers hold equally good for all societies at all times.

      I think that the concept of embedding (which Polanyi described through the double movement of society's resistance to the market) has been recuperated by capital through the advocacy of the Internet in the language of "liberation", "sharing" and "community" (which many on the left have bought into wholesale). The result is a d├ętournement in which non-wage labour has been colonised by the idea of competition. Given the very small amount of data that produces real value, the lack of discrimination suggests that this is primarily ideological. My suggestion is that this is a compensation for a decline in actual competition in the economy.

    2. Herbie Kills Children11 June 2017 at 11:00

      “Marx's labour theory of value addresses how actual value, i.e. a productive surplus, arises in a capitalist society”

      Pedantic point but crucial as forgetting what I am about to say has led to all sorts of errors in interpreting and analysing Marx, Marx’s theory addresses how all value is produced, not just the surplus. Actually you can’t possibly explain the surplus without explaining the reproduction of existing value!

      “Most data are worthless. Most of those that are monetised are of distributive not productive value “

      A plethora of companies have been created that employ labour to extract and transform data, which is then sold in the market. In what way can it be claimed that this modern industry is not productive of value?

    3. My point is that most of this activity is concerned with getting us to buy commodity A rather than commodity B (or use one retail channel over another). In other words, it is distributing revenue between different suppliers rather than creating new products and thus additional value.

  2. Thats very clear. I asked the question because I'm trying to fit mimetic theory (Rene Girard) into the same phenomena you describe and want to understand how it intersects with other theories, such as Labour theory (Locke), marginal utility (preference), and what might be called market complexity value (mathematical real). The topic of mass data created by non waged labour and harvested for commercial sales advantage and advertising seems to draw on all of those ideas. I certainly favour your last point that the decline in competition in other areas is responsible for the lionising of market principles for the internet, and that this can work politically because it 'embeds' that market.

    Anyway, talking of leek competitions did you ever see the video of the leek competition at the bottom of this Adam Curtis blog. Pretty splendid.

  3. Herbie Kills Children8 June 2017 at 17:19

    Well data is one thing, turning data into information is another. The trick is in the latter.

    You are missing the revolutionary implications of this technology, which does indeed cheapen everything but also allows people to more easily share information, so it tips the balance away from certain interests and allows new power structures to grow. For example, people can now go online and review the holiday they have just been on, they can rate the hotel, the area, the staff etc etc. This forces the hotel and area to respond to the criticisms it receives online. In many ways this is a highly political use of the internet, albeit for frenzied consumption. And i guess this is where the limitations are, ultimately the internet intensifies the corrosive consumption culture to an even greater degree.

    So the revolutionary and liberating potential of these technologies will simply help us on our path toward death by consumption and a totalitarian all seeing state.

    I doubt even Corbyn can save us from that fate, but at least the ride will be made more palatable.