Difference between revisions of "Left and Big Data"
(Created page with " =Discussion= James Meadway: "The case for democratic planning has recently been restated by Leigh Philips and Michal Rozworski in their very readable The People’s Republi...")
Latest revision as of 06:46, 23 February 2021
"The case for democratic planning has recently been restated by Leigh Philips and Michal Rozworski in their very readable The People’s Republic of Walmart. Their argument hinges on the presence of ‘islands of planning’ in a sea of market competition in the form of the vast operations of corporations. The possibilities of using the technologies and organisation mobilised by the likes of Walmart for more obviously human ends than is the animating thread of their book – just as was observing the prodigious efforts of capitalist states planning production for total war for earlier generations of socialists.
In particular, fast computers and cheap digital sensors open up the prospect of continuous, real-time planning of the economy in ways that have never been feasible before. The ‘socialist calculation problem’ of old – that society was too complex to plan everything – appears solved. Yet this may not be an entirely desirable solution. In the World Review of Political Economy, Bibin Yang and Xiaoyan Li have recently proposed the implementation of a planned economy using real-time Big Data to allow immediate adjustments to plans, creating a hybrid ‘market-based, plan-driven’ system. Like Philips and Rozworksi, they suggest that major corporations – particularly the platform companies like Alibaba and Amazon – already operate on such a model. All that is needed, they suggest, is for the state to replicate such functions and expand its sphere of planning. Jack Ma, founder of Alibaba, has suggested that his company is a stepping stone to a planned economy: ‘with access to all kinds of data, we may be able to find the invisible hand of the market.’
There are two important points here. The first is that ‘finding the invisible hand’, whilst a striking metaphor, is very much in line with conventional economics thinking on the relationship between planning and the market. The ‘ideal’, welfare-maximising distribution of societies’ resources, as selected by a perfectly benevolent planner, is identical to the ‘ideal’, welfare-maximising distribution of societies’ resources as selected by an entirely free market, in the standard proof of the desirability of a market society. The presence of Big Data turns this hypothetical social planner into a real possibility, overcoming the standard free market criticism of planning, exemplified in the work of Von Hayek – that the information needed to perform the task, in any reasonably complex society, is too great to be performed. We have not stepped outside the boundaries that neoclassical economics usually establishes for the problems it seeks to address.
Second, if we do look beyond those boundaries, we are back in the world of a state-capital fusion. ‘All kinds of data’ has deep implications: it could include, for instance, social credit scoring, as applied in China (and of which the recently-departed Dominic Cummings was something of an enthusiast). It is not obvious that this should be acceptable. Francis Spufford’s semi-fictional account of the development of Soviet planning, Red Plenty, makes exactly this point: that even with faster computers and more data, the demands of a (presumed) benevolent planner might well be rejected by those expected to follow its edicts. (In the book, changes in the prices of food, intended to produce a fairer distribution according to the plan, provoke riots.) We would have to exhibit a very high degree of trust in the planner to simply allow for its diktat to set the course of our lives.
Arguably, via the mass data-collection empires of Big Tech, we do display that sort of trust, happily handing over to them unimaginable volumes of information about ourselves and our relationships. We might read this as a challenge to Graeber’s vision of reciprocity: that a version of free exchange – even a gift relationship – can be established with an entirely faceless bureaucracy subject to no obvious legal or democratic or even, if we are not careful, human restraints. The latter point is particularly important with AI that can teach itself and devise new rules for us, as is already occurring. We are implicitly offering our consent to a process that is running ahead of us – and one which, under Covid-19, has accelerated markedly.
There are two profound side-effects to the presence of mass data that the Left has not yet got its collective head around. The first is the gradual disappearance of consumer-facing markets from our lives. Facebook is, as it says on the homepage, ‘Free, and it always will be.’ No exchange of money takes place between the Facebook consumer and the Facebook company, as in a more conventional market operation. Instead, Facebook monetises the data it gathers and sells it elsewhere. But this is a challenge to how we conventionally think about capitalism, and certainly how we conventionally try to register its activities. Constructing GDP becomes harder when there is less being obviously traded, and the location of where ‘value’ is realised becomes harder to pin down. This is a process of capital-led decommodification, at least for its end-products; but the corollary recommodification of data is also still poorly understood. Katharine Pistor has recently explored some of the implications. In ‘Rule by Data: The End of Markets?’, she speculates persuasively that the data will likely displace markets and law as the dominant organising principle of human societies. We are already seeing the displacement of markets; law is further down the line. This process runs counter to the neoclassical expectation that the improvement of markets would lessen the dependence on law and reduce the role of firms in the economy. In fact, Pistor suggests, we have seen the reduction of transactions costs to bare minima via digitisation, but, alongside and as a direct consequence of this, the massive growth of ‘corporate behemoths’ in the form of the data giants. Law is being bent towards their will, directly in the sense that they are very powerful lobbyists for specific legal structures, and indirectly (and ultimately more consequential) in that their actions create new structures enjoying a form of legality in advance of any formal lawmaking.
There are three sets of responses to this. One is to attempt to reassert market rule, as with the various anti-trust efforts against Big Data. A second is to reassert law, as in the various attempts to regulate and control data. Both of these look doomed to fail, at least on their initial terms: the former looks set to, at best, reproduce multiple versions of problematic institutions; the second merely hands over more control to a central authority. Both ‘market sovereignty’ and ‘state sovereignty’ fail when set against the need to bring data under human control. This will not stop their political protagonists arguing for either: a reassertion of neoliberal politics lies in one direction, and potentially ugly nationalism in the other. An anti-egalitarian political right can exist on the basis of either (or, quite possibly, some combination of both at once).
Pistor favours a third option: a reassertion of the fundamental reciprocity – the ‘primitive communism’ – in allowing smaller collectivities of data users and producers to form, then seeking their protection through the use of ‘data trusts’. These would be a mixed form of institution – potentially (if the trustees wished it) operating in a market, enjoying a legal protection, but constituted on the basis of a collective interest. (I have recently discussed such trusts in Creating a Digital Commons.)
The pandemic has shown us how a collective interest can run far ahead of the state: judging by transport usage statistics, the ‘lockdown from below’ was beginning well ahead of the laggardly government response. The range of new governmental powers have not in general been directly used, because the wider social determination of a shared interest in maintaining social distancing, wearing a mask, and so on, has been very solid. The moves towards authoritarian control under Covid-19 – despite the hysterics from the pro-virus right – have occurred out of sight, in the collection, management and analysis of data." (https://salvage.zone/articles/acting-as-if-one-is-already-free-david-graebers-political-economy-and-the-strategic-impasse-of-the-left/?)