Facebook and Finance: On the Social Logic of the Derivative

Derivatives

What is the social logic of the derivative? Louise Amoore speaks of the data derivative as a distinct way of processing digital data for surveillance and security purposes. She distinguishes the data derivative from what she calls ‘modernist disciplinary data’ (Amoore, 2011: 27). Drawing on her account and similar ones we can identify two dimensions to this distinction that might help us to articulate a better understanding of the concept.

First, derivatives operate with derived qualities: qualities that have been derived from an underlying entity, or simply an ‘underlying’ to use financial terminology. This in itself is nothing new, all kinds of measurements deal with derived qualities. However, contrary to what Amoore calls ‘modernist disciplinary data’, derivatives deal with derived qualities while disregarding the underlying. In the social sciences, modernist disciplinary data used derived data to discuss the nature of an underlying social reality that was conceived to have a positive existence. Derivatives instead operate by constructing a ‘virtual’ reality composed of relations between qualities that need not correspond to the ways in which those qualities are related in the lived practice of their underlyings. The result might be used to act on and govern an underlying reality, but in order to do so derivatives create a separate virtual reality that becomes the actual referent for the decisions that they support or enact. In Bryan and Rafferty’s words:

Stripped of mathematical formalism, the idea of derivatives is quite simple. They involve deconstructing a ‘thing’ (and we use a bland term intentionally) into a set of constituent elements or attributes, and configuring those attributes in a way consistent with quantification. (Bryan and Rafferty, 2014: 892)

On financial markets arbitrage is practiced and value generated by finding unknown correlations between the movement of certain derived qualities, say the volatility of a commodity index, and the movement of other such derived qualities, say the price of interest rate futures, without those qualities necessarily being related to each other as aspects of any actually exiting underlying asset (Stark, 2011). In Big Data analytics, security threats or new market niches can be discovered by finding correlations across data sets, which themselves might have no basis in lived social processes but that nevertheless allow for something to be said and done about such lived social processes. Derivatives operate by shedding causal connections to the world of underlyings – assets or people – and by creating a new topological space (Lury et al., 2012) where derived qualities can be related to each other without the world of underlyings disturbing the process.

A second fundamental characteristic of derivatives is their orientation to the future. A derivative is based on a derived direction – a future extrapolated from the past. The original Black-Scholes equation for option pricing, the simple origins of modern derivative financial instruments, solved the problem of option pricing extracting the volatility of an asset from its price history. The riskiness and hence price of the asset could be calculated by means of a derivative – in the strictly mathematical sense of the term – a path projected into the future. However, the future that derivatives extrapolate from the past is a particular kind of future. First, it is not about the future of a particular underlying asset. In Louise Amoore’s words: ‘it is not about who we are’ but about ‘what can be imagined and inferred’ – not necessarily about us – ‘based on our past proclivities’ (Amoore, 2011: 28).

Together these two dimensions point to a fundamental aspect of derivatives. Derivatives operate by creating a virtual reality of relations and connections that need not have any ontological grounding. This imagined future – a future without a subject – simply constitutes a common space where qualities and aspects that might be incompatible in the life of underlyings can be related to each other. Different kinds of assets – the value of a brand and the riskiness of a corporate bond – that might not be comparable in terms of any common denominator can be put in relation to each other by comparing proclivities implicit in their implied future trajectories. This reality is virtual in the sense of ‘existing in practice if not in reality in a strict sense’ (Arnoldi, 2004: 24). However, this virtual reality comes to affect real life, as decisions that derivatives enabled are performed. As derivatives are used to price an asset, this virtual reality affects tangible aspects of lived social reality.

The virtual reality that derivatives create is oriented towards the present: The extrapolation of the future serves only to make a decision in the present, and then its function has been exhausted. Derivatives do not so much predict the future as render a posited future actionable as a parameter for decisions in the present (Esposito, 2011; Amoore, 2011: 4). Consequently these futures evaporate once they have been acted on, once the derivatives contract has expired, once an arbitrage has been exploited, or once a particular data correlation has been identified and acted on. Whether they are deployed in finance or in other areas, derivatives create, as a material effect, a series of presents, a series of singularities that remain unrelated to each other in ‘the real world’: In Fredric Jameson’s words:

This is the ultimate logical conclusion of the paradox of the derivative: not that each derivative is a new beginning, but that each derivative is a new present of time. It produces no future out of itself, only another and a different present. The world of finance capital is that perpetual present – but it is not a continuity; it is a series of singularity-events. (Jameson, 2015: 123)

The Social Graph

How does this this apply to Facebook? At the time of writing Facebook’s greatest revenue comes from the sale of advertising space. However, the logic behind Facebook’s sale of advertising space is different from that which operated in the modern, Fordist advertising market.

The market in advertising is a market in potential attention. A medium is valuable as a vehicle for advertising if it is likely that its audience will pay attention to a certain advertising message (Smythe, 1981). In the Fordist advertising market that likelihood was inferred from the assumed socio-economic characteristics of the subjects making up the audience: Were they working class? Urban or rural? And so on (Arvidsson and Bonini, 2015). Facebook also values its advertising space based on the probability that certain users will accept certain messages. However, such probabilities are based on assemblages of abstract qualities that are derived from the lived practice of actual users but, significantly, recomposed in ways that are independent of how they are related in the lived practice of their underlyings. Facebook assembles micro-publics, not necessarily of individuals, but of interests, affinity and other expressions of affect that are tailored to a specific advertising message.

This leads to a first basic insight: the ways in which Facebook assembles qualities and people into social forms, like advertising targets or ‘micro-publics’ of users, is not based on the ways in which such qualities and people are related in actual life processes, but on the ways in which they are related in the abstract data space that Facebook creates, its ‘social graph’. The social graph constructs a virtual sociality: a set of relations between ‘doings’ that have no ontological referent. It is used to make decisions about the reality of users’ lives on the platform: about the relevance of their profiles for a particular message, or about their value as advertising targets. This way the modus operandi of Facebook in creating social assemblages mirrors the basic logic of derivative financial instruments.

The social graph was the technology that enabled the launch of Facebook ads and pages in 2007, opening up the platform for commercial revenue. Essentially it works by transforming the lived social processes that unfold on Facebook, and since the launch of the Open Graph protocol million of sites and apps outside of the domain of Facebook as well, into abstract informational entities called ‘objects’ and ‘edges’ (Kaldrack and Röhle, 2014 ). The category ‘objects’ comprises the various informational objects that can be derived form the lived practices of Facebook users: content such as postings, as well as users themselves. ‘Edges’ represent interactions between users – a user who messages another user constitutes an ‘edge’ – as well as interactions between users and content: a user who likes a post also constitutes an edge.

Objects and edges are represented as a network of informational objects and their interactions: together they make up the ‘social graph’, a topological representation of the life processes that unfold on Facebook. The topological map of objects and edges that results from the social graph is used to structure the flow of information that appears on the pages of individual users. The probability that a posting will appear on the page of a user depends not on some socially anchored common characteristic (like, for example, common demographics), but on the place of the objects that represent, respectively, that posting and the user, within the topological map of the Facebook social graph. For example, a posting that originates from a user whose homepage I often visit and that has been liked by several people whose postings I too have liked is likely to appear on my page. It is important to note that the inference here is not primarily about the content of the posting, nor the nature of the relationship in real life. Rather it is an inference that uses past patterns to make an argument about the probability of future directions. Because I have taken an interest in this user’s postings in the past, it is probable that I will do so also in the future. And the intensity of my past interest, or more correctly the affinity of past interactions between relevant objects and edges, is used to calculate the probability of my future interest. On the basis of an inference of such future interest, Facebook makes a decision regarding the present allocation of informational objects.

The social graph is also used to place advertising messages on relevant user pages. Facebook’s advantage in the advertising market is not the effectiveness of its advertisements per se (indeed Facebook has relatively low click-through rates – few people who see an ad actually click on it – relative to, for example, Google; Smith, 2013). Instead, Facebook’s selling proposition is its ability to place advertisements on user pages in such ways that it is probable that they will be compatible with a user’s interests and passions. To achieve this advertisers are asked to indicate a number of segmentation preferences. Let us quote Facebook’s own pedagogic example: ‘People who live in the US, in Nashville (TN), who are interested in cooking and cookbooks’. The first two parameters, just like age and gender, are derived from users’ own profiles. They are ‘real’, so to say. The second two parameters are slightly more complicated. To be included in a target potentially interested in cooking or cookbooks does not necessarily mean to have expressed an interest in these things. It means having a pattern of doings that is statistically similar to clusters of doings where ‘cooking’ and ‘cookbooks’ are central: in other words, interest in cooking and cookbooks is a virtual construct. The values inserted by an advertiser define a public as an assemblage of informational objects that are located at a certain distance from each other in terms of degrees of statistical similarity. In order to be included in this public I might never have directly expressed interest in a cookbook, but simply ‘liked’ a restaurant, shared an article on food trucks or even interacted with someone who is interested in these things.

The social graph is a topological space where relations between derived qualities can be created and calculated in ways that pay no attention to other aspects of the life of the underlyings – the users. Instead what counts is the affective distance between objects (me, pizza, cookbooks, food delivery services) and the direction in which the affective distance is likely to evolve in the future (I have liked pizza in the past, I am likely to like pizza in the present. I have never liked cookbooks in the past but my friends have liked cookbooks and I have liked pizza, so…). These affective derivatives are calculated by a proprietary algorithm, the precise makings of which are a trade secret. However, its most important dimensions are well known. Facebook calls them Type, Recency, Interest, Post and Creator.

Type and Recency are exogenous variables that give additional weight to certain types of doings (a comment is worth more than a like, for example) and to recent doings over less recent ones. The remaining three principle variables, Interest, Post and Creator, are developments of what the old Edgerank algorithm used to call Affinity, i.e. the structural similarity between the network of affective expressions that surrounds one user or posting and that which surrounds another in the social graph. The differences in such structures of affinity are used to make probabilistic inferences about future interests or affinities. Based on the past relations between the network of doings surrounding ‘pizza’ and the network of doings surrounding user A, we can infer that user A is likely to take an interest in a posting about pizza. On the basis of such inferences the platform establishes the value of users as advertising targets or as potential sources of attention for postings. Such valuations are always topological: it is a matter of understanding the potential and proclivities of one user in relation to the potential of a post or an advertisement.

But the relation is not a matter of finding a common external denominator. In Celia Lury’s words: ‘The measure of value does not have an ontological referent in a metric outside the space of the assemblage itself’ (Lury, 2009: 79). Rather, the procedure reassembles the ways in which arbitrage is practiced on financial markets. The inferred proclivity of one object – an asset or, in this case, an informational object – is put in relation to that of another object, and the probability of their correlation is inferred from the ways in which the two objects are represented in an abstract space of derived qualities (volatility, beta gamma and all the Greeks in the case of financial markets), or likes, messages, page visits in the case of Facebook’s social graph. The platform performs valuation by comparing such implied proclivities and potentials, just like the Black-Scholes equation enabled pricing decisions by making the inferred volatility of different asset prices comparable. These are both comparisons that are acted out in a separate topological space, that of the social graph or that of a historical time series of market data.

Brands and the Derivative: A Common Genealogy

Facebook’s modus operandi does not come out of nowhere. Once the platform grew too large to be manually managed by Mark Zuckerberg and his friends, it started deploying algorithms that built on four decades of market segmentation: the same algorithms that had made brands in their contemporary form possible. Hence, this way of ordering the social by means of the derivative is not unique to Facebook; it has developed as a core principle of contemporary branding.

Along with other contemporary scholars, Celia Lury (2004) suggests that brands have evolved from the mere symbols of products or producers into what she calls assemblages. Contemporary brands are particular social forms in themselves. They are constructed as aspects of social life – desires, interests, actions, symbolic expressions – that are taken out of context and reassembled into a coherent whole that follows its own logic. Branding entails the selection of aspects of the environment and their re-composition into elements of a system that follows its own ‘code’. For example, in the 1990s elements of inner city ‘ghetto’ culture, the sports star Michael Jordan, the culture of fitness and a distinctively consumerist version of female empowerment were reassembled into a coherent Nike brand organized around the abstract value or principle expressed in Nike’s device, ‘Just do it’ (Lury, 1999; Goldman and Papson, 1998). The reassembling also entails rendering the disparate qualities of the brand compatible, which they might not be in their ‘original’ social form (Lury, 2009: 75).

A brand is a mechanism that allows for the control and prediction of the life of consumers, and crucially the direction of consumer demand (although other kinds of life processes, like a particular attitude towards life – for example, ‘creative’ city – might also be the object of branding; Pike, 2011). This control is exercised by positing a particular future direction that can be acted on today. Facebook continuously performs such brand-like assemblages: it puts users in relation to postings and ads on the basis of affinity between data structures in its social graph. In this sense Facebook is a machine for the reassembling of the social into brand-like micro-publics without the need for human ‘creative’ intervention. These publics are endowed with direction and, by implication, predictability: Facebook is a machine for the transformation of the lived social practice into a branded sociality, composed of temporary assemblages that allow for the control and predictability of life processes.

This structural similarity between the operations of Facebook, of contemporary branding, and of derivative financial assets is no coincidence. Rather, derivative financial instruments, and contemporary branding have a common genealogy. The development of contemporary derivative finance occurred at the same time as market research went through a series of transformations that would lead up to the development of contemporary brands. These transformations built on advances in probability theory and the diffusion of computers (MacKenzie, 2006). They were also inspired by the same paradigm, what Mirowski (2002) calls the cyber-sciences – the development of probabilistic principles for the control of complex systems – that had developed during the war effort. This ‘cyber’ paradigm built on two fundamental novelties.

First, a new perspective on reality. The disciplinary paradigm had built on an intervention in reality, and an idea that reality could be reshaped, re-educated and controlled through such intervention. In the new cyber paradigm, the ambition of shaping reality was abandoned. Reality was instead conceptualized as environmental noise: the market as a series of ‘random walks’, consumer culture as a chaotic mass of desires or, in the later systems theory of Niklas Luhmann (1995) – and its perhaps unconscious materialization in Facebook’s social graph – the social as meaningless ‘complexity’.

Second, and as a consequence, there was a re-formulation of the problem of order. The notion of order no longer needed a normative anchoring: contrary to the paradigm of disciplinary governance, order was not a matter of maintaining certain pre-established principles. Rather, order became defined as the problem of finding temporary direction and coherence in an environment perceived to be chaotic. The cyber-sciences built on an acceptance of the inevitable nature of complexity, and on a more modest idea of order as the possibility to act and make decisions in the midst of such complexity: to find ‘arbitrary enclaves of order and system’, such as the ‘statistical difference’ between ‘signal’ and ‘noise’ that constituted the basis of Shannon and Weaver’s mathematical theory of information.

In the emerging field of financial economics this gave rise to a reconceptualization of the market as a random series of prices, and a turn towards calculating the value of assets in terms of their risk, conceived without regard to their specific underlying characteristics. Instead, the calculation of risk looked at the volatility of asset prices. This way the definition of risk became a self-referential – Lury would say topological – probabilistic construct that referred only to the overall system of financial price movements and their fluctuations. During the 1960s this approach gained in influence with the new discipline of financial economics – based initially in business schools rather than in economics departments in major research universities – as computers now enabled the processing of large data sets (Arnoldi, 2015; Merton, 1995). In the 1970s the development of the Black-Scholes model coincided with the expansion of currency markets and the market for currency-based derivatives caused by the collapse of the Bretton Woods system in 1971. The approach pioneered by Black, Schooles and Merton became the underlying driver, the ‘engine’ of the development of an ever more complex derivatives markets (MacKenzie, 2006).

This driving function did not emerge because the Black-Scholes model actually represented the empirical reality of asset markets. Rather the contrary: the Black-Scholes model created a reference for the rational price of an option that was independent of the dynamic of actual prices and to which these could be compared in order for traders to identify possibilities for arbitrage. As MacKenzie points out, new probabilistic financial instruments were performative; not only did they create a market for the instruments that they pioneered, but they created a market that operated according to the logic of the instruments that they had pioneered. The development of modern mathematical finance – from the CAPS model of the early 1960s, via Black-Scholes to the introduction of highly advanced probability theory in the late 1970s – managed to separate the pricing of assets from information about the performance of underlying assets. It managed to introduce and institutionalize a new way of determining the value of assets.

Something similar happened in market research. Up until the 1950s the main paradigm of marketing had been the standardization of consumer demand. Similar to the paradigm of disciplinary governance, consumers were to be re-educated to abandon their traditional desires and to adopt ‘modernity’ (Marchand, 1985). In the 1950s this began to change as the ‘marketing concept’ began to make its way in the profession. The marketing concept introduced a subtle difference: marketing was no longer about standardizing consumer demands across markets and adapting them to the productive capacity of industry. Rather, it was about extrapolating the potential of a brand, conceived of as an assemblage of consumer needs, material products, distribution channels and other potentials, and orienting production and consumption to the potential of that assemblage. Rather than referring to actual embodied life processes, the central concept of ‘the consumer’ came to stand for disembodied wants and needs that could be reassembled around products and, importantly, brands. As Robert J. Keith, the Pillsbury manager who defined the ‘marketing concept’ in a classic Journal of Marketing article from 1960, put it:

Marketing begins and ends with the consumer. New product ideas are conceived after careful study of her wants and needs, her likes and dislikes. Then marketing takes the idea and marshals all the forces of the corporation to translate the idea into product and the product into sales. (Keith, 1960: 38)

At the same time the arrival of computers and the impact of new probability theory on market research began to provide marketers with the necessary instruments to perform such reassembly of derived qualities. Psychographics, or lifestyle segmentation, was introduced in the 1960s and would radically change the ways in which market researchers conceptualized consumer demand. Up until then categories of market research had been rooted in the lived practice of individuals. Market research had mainly been about counting the number of individuals that belonged to a particular category, usually a demographic category reflecting social class, or prejudices thereof. With psychographics, market researchers instead administered questionnaires to consumers irrespective of social position or demographic allocation. These questionnaires were very long, some including up to 300 questions, that covered a variety of areas, like social class, consumer preferences, lifestyle issues, moral values, as comprised in the name of one such approach – VALS (Values and Lifestyles) – that became popular in the 1970s (Sheth, 1970; Mitchell, 1983).

Answers to these questions were subsequently treated as independent variables that were related to each other in a distinct data space, without the relations in the life of underlyings being taken into account. Their relations in this separate topological space constituted clusters that were sold to advertisers as lifestyle. Essentially a ‘lifestyle’ was a derived asset, a cluster of correlations between derived data with a certain affinity that could be used for extrapolating a certain affective proclivity (e.g. people interested in BMWs, French wine and Playboy magazine are likely to be interested in advertisements for hi-fi stereos), without this being necessarily grounded in the lived reality of any individual. Conversely, lifestyle clusters were used as the building blocks for the assembly of brands. A brand becomes a reification of a lifestyle, a rendering explicit and material of its implicit proclivities: its transformation into an intangible asset.

Psychographics, as well as modern financial economics, were part of an overall transformation of managerial practice (in the wide sense of the term) that occurred in the post-war years. This new approach was grounded in cybernetics and general systems science that had developed during the war effort to become what was known as operations research, a series of mathematical models oriented towards the optimization of performance in an environment understood as complex and unpredictable. In the post-war years operations research was promoted by government agencies, and actors like the Ford and Carnegie foundations, as a key to the modernization of managerial science and practice, and to strengthening the scientific credentials of business schools (Cochoy, 1999: 167ff.). In marketing, operations research influenced the twin development of marketing management and the marketing concept, whereby a previous descriptive approach was gradually replaced by a prescriptive approach that developed models for establishing a predictability and managing of demand in a complex environment.

While the influence of mathematical modeling in marketing practice was negligible (as opposed to the emerging academic field of marketing science, which got its own journal, Marketing Science, in 1969), this development led to a number of influential models, like the ‘marketing mix’ model developed by Neil Borden in 1964 and its sequel, the 4P (price, product, place and promotion) proposed by Jerome McCarthy and Philip Kotler in the 1970s, which continued the spirit of this new paradigm, without the mathematics. In this new paradigm of corporate governance, which Nigel Thrift (2005) has identified as the origins of what he calls ‘knowing capitalism’, the managerial sciences began to adopt a worldview similar to that of Philip Mirowski’s ‘cyborg sciences’, where the main preoccupation was the establishment and continuation of order in a world of complexity.

This orientation was common to both financial derivatives and the new marketing practices that stood behind the transformation of branding. Just like psychographics represented consumer desires as probable directions in a chaotic world, so the models of modern mathematical finance calculated price in relation to probable directions in a world of chaotic randomness. Just like psychographics was not directed towards an empirical representation of consumers as actually living human beings, so the models of modern mathematical finance were not primarily interested in adequate representation of the value of underlying assets. Rather the knowledge interest was concerned with finding new ways of establishing order in a world perceived to be increasingly complex.

Indeed, according to Robert C. Merton (1995), the development of the Black-Scholes equation responded to a situation where financial markets had become so complex that traditional ‘rule of thumb’ methods no longer were sufficient. Similarly the ‘father’ of lifestyle segmentation, William D. Wells, recognized in his foreword to the American Marketing Association’s 1974 volume on ‘Lifestyle and Psychographics’ that this new approach had developed as a response to a social environment that was perceived to be increasingly dynamic and marked by ‘rapidly changing values’ (Wells, 1974: v). By this opening up of a different way of acting on reality the models of modern mathematical finance, just like psychographics, could construct tools of predictability and order necessary for capital accumulation in a world where older forms of disciplinary governmentality were losing their grip. As these models were invested with the power of corporate capital – ‘marshaling all the forces of the corporation’ – they became performative: derivatives enabled the enormous financial expansion that has marked the world since the 1980s and onwards; the marketing of brands enabled a similar expansion of consumer culture and came to significantly affect the underlying reality of consumer desires (Yuran, 2014). Derivatives and contemporary brands together developed the foundations for a new financial law of value.

Facebook and the Law of Value?

In the post-war years consumption and finance have become central to capital accumulation. It is quite natural that these ‘levels’ would serve as laboratories for the development of new forms of capitalist governmentality, just like prisons and factories did in the 19th century. Both levels have pioneered probabilistic control in response to a world that was becoming too complex for the traditional power/knowledge nexus of disciplinary government. While the world can no longer be known, at least digital media can abstract the kinds of data that allow for probabilistic inferences. When it can no longer be transformed according to an ideological program, at least its insecurities can be transformed into risks that can be acted on and, in so doing, money can be made. Order and direction can be maintained, even in the absence of an underlying unity. As Randy Martin describes this new approach to governance:

We can say that the derivative logic speaks to what is otherwise balefully named as fragmentation, dispersion, isolation, by allowing us to recognize ways in which the concrete particularities the specific engagements, commitments or interventions that we tender and expend might be interconnected without first or ultimately needing to appear as a single whole or unity of practice or perspective. (Martin, 2013: 87)

At present Facebook mostly exploits its ability to operate the social logic of the derivative via the advertising market. But the company and its investors are imagining a much more extended application of this approach. Currently Facebook does not make enough money out of advertising to remain sustainable in the long run. Like most social media companies Facebook is strongly overvalued by traditional standards (Arvidsson and Colleoni, 2012).

What would seem like a consistent overvaluation of social media companies might to some extent be the result of a speculative bubble, a Web 2.0 boom that never ended but simply ‘re-routed’ around the financial crash in 2007. However, if there is any substance to Facebook’s valuations, they rest with the expectation that that platform should be able to significantly extend its business in the future:

Indeed the exorbitant $19 billion that Facebook paid for WhatsApp was justified by its potential to operate as a ‘lite’ mobile-based extension to Facebook, particularly in ‘the Global South’ where it can radically extend the grasp and inclusivity of the Facebook platform. Facebook has made similar acquisition of start-ups that are developing technologies to directly beam internet access, via satellite or drones, to areas with bad connectivity. This points to a desire to extend the reach and inclusivity of the Facebook platform. Acquisitions of virtual reality technology companies like Oculus and inroads into the micropayments market indicate a drive to also enable a wider range of diverse social relations and transactions, like immersive gaming universes or remittances from migrants, to unfold on the platform. The significant expansion of Facebook’s presence on mobile platforms is also taken as a promising development towards a deeper immersion of the platform in the everyday life of its users.

Overall it is significant that Facebook seems to have suspended the question of the calculable profitability of these investments, simply suggesting that their promise and potential to aggregate diverse social relations onto its platform will, some day, pay off. As Mark Zuckerberg commented on the purchase of WhatsApp with Wall Street analysts: ‘We believe that once we get to being a service that has 1 billion, 2 billion, maybe even 3 billion people one day, that there are many clear ways that we can monetize’ (Sikka, 2014). With the data at Facebook’s disposal, and with additions to that data following from its expansion into less connected markets, the platform has the potential to expand the social logic of the derivative into the minute details of everyday life, enabling new and more detailed forms of credit ratings, the pricing of health insurance, and the valuation of freelance labor based on reputational measurements of risk (Neff, 2012). As Evgeny Morozov (2014) argues, Facebook, along with the diffusion of an infinity of data gathering devices – from ‘smart’ fridges to wearable sensors – is opening up a potential for social media companies to become a sort of clearinghouse for extending financial securitization into the minute details of everyday life, at a global level. This would significantly strengthen the present trend towards integrating social media data in financial value decisions (Karppi and Crawford, 2015), while at the same time expanding the logic of financial valuation to ordinary life processes.

Like in the case of derivatives, the market valuation of Facebook is not based on its ability to capture things like ‘audience labor’ (Fuchs, 2010) but on ‘the commodified role of commensuration’ (Bryan and Rafferty, 2006: 154) that the platform performs: its ability to establish the value of diverse forms of life that can no longer be reduced to a single common denominator. This makes the strategy of Facebook consistent with present tendencies towards the ‘financialization of everyday life’ (Martin, 2002). And Facebook’s strategy would extend such financialization in ways that are compatible with the overall logic of the derivative. This way, Facebook might be evolving into becoming a platform for the establishment of the social logic of the derivative as the basis for a new law of value. Let me spend some concluding thought on how to understand the place of such a new law of value within the overall context of contemporary ‘cognitive capitalism’.

Conclusion: Cognitive Capitalism, Social Media and Lived Intangibles

Cognitive capitalism is premised on a shift in the logic of capital accumulation from tangible to intangible assets. Intangible assets, like brand, innovation and flexibility, are defined by their impossibility to be measured according to traditional accounting standards. (This is indeed the only solid definition of an ‘intangible’: ‘an asset that we suspect adds value but that we cannot measure with any certainty’; Arvidsson and Peitersen, 2013). This immeasurability results from the inability to reduce intangible assets to any common denominator like capital expenditure or labor time. The value of a brand or the capacity of a company to be innovative are not related to investments in labor time in any linear way. Rather these assets result from what Paolo Virno calls the virtuosity of cooperation (Virno, 2004). This leads to a first insight: intangible assets do not derive from measurable investments of identifiable productive factors but from the more or less virtuous forms of cooperation that emerge in a complex and networked global economy based on general intellect. Within contemporary cognitive capitalism derivatives have evolved to measure and price the riskiness of such cooperation. In global supply chains the insecurity of inter-firm relations is commodified as tradable risks. The ability of companies to extract and commodify virtuous forms of cooperation from their environment is capitalized on as ‘brand value’, which, as Lury has shown, is estimated in ‘topological’ ways that follow the logic of the derivative (Lury and Moor, 2010).

On financial markets intangibles in excess of the book value of companies are priced through arbitrage trading and derivative instruments. Overall, as Bryan and Rafferty (2006: 131) argue, derivatives are premised on the fact that there are no underlying fundamental values that make different forms of capital comparable. Rather, the work of the derivative is to offer a solution to the absence of fundamentals. The incompatibility of different forms of capital is ignored; instead these are treated as ‘bits’ of abstract general capital that can be evaluated in terms of their relative risk. Within the capitalist economy the rise of derivatives has been a response to the ‘explosion’ of the traditional capitalist law of value, due to the production process becoming complex and globalized (Negri, 1999). Derivatives have emerged and spread as a new way of valuing such otherwise immeasurable assets, as a new solution to the problem of value.

Facebook extends this valuation to what we can call ‘lived intangibles’. Once again this has a history. The becoming complex, mediated and networked of consumer culture in the post-war years created a demand for ways of valuing and commodifying consumer tastes and desires that exceeded the rigorous definitions of the Fordist ‘consumption norm’ (Agliettta, 1978). Psychographics and similar derivatives-like instruments were deployed for these purposes, leading up the development of contemporary brands. Contemporary brands commodify such excess and unruliness as concrete intangible assets, thus establishing a direct link between the ability of life, and particularly life empowered by mediation, to generate an excess, on the one hand, and capital accumulation, on the other. Facebook, and similar social media platforms, extend this integration yet one more step, offering tools that enable the valuation of such lived intangibles in all walks of life, and not just in consumer practice. Lived intangibles like the virtue or trustworthiness of potential employees or the reliability or value compatibility of freelance workers can find a value on Facebook.

Facebook’s expansion promises to extend such valuation practices beyond the market for knowledge work, where it currently operates, to, for example, the market for low-value freelance labor, gathered worldwide through Amazon Turk and similar platforms. This way Facebook comes to concretely perform the function that Randy Martin ascribes to the social logic of the derivative: the ability to ensure productive cooperation without the need for normative integration, but based instead on the pricing of probabilistically derived risks. In so doing Facebook also extends the grasp of financial capital, transforming a range of hitherto ‘pristine’ life processes into the basis for the abstraction of intangible assets. Facebook comes to perform the direct biopolitical function of financial capital, integrating ordinary life into its processes in a calculable way. It provides a way of giving universal value to the lived excess of the global multitude while abstracting from its lived practice. Facebook provides an alternative to ethics in valuing bios in excess of nomos by extending and materializing contemporary capitalism’s new ‘spirit of calculation’ (Appadurai, 2012).