by Gordon Hull
Cloud computing – where users keep their data (and often their applications) online - poses significant theoretical and regulatory problems. Many of these concern jurisdiction: it’s very hard to even know at a given moment where data is kept, and it’s often unclear (in the case of privacy, for example), which jurisdiction’s privacy and data protection rules should apply (the one for the data subject? the company that collected the data? the companies processing it? etc.). Not only that, U.S. and EU law are wildly inconsistent on the point, even though any large big data company has to serve multiple jurisdictions.
A recent piece by Paul M. Schwartz does some valuable work disentangling these issues; here, I want to focus on one moment. Schwartz notes that cloud computing will likely induce significant changes in how firms are structured, and how they structure their data handling. Back in 1937, Ronald Coase proposed that companies will decide between doing something in house and outsourcing it based on a comparison of the costs of each. If it’s more efficient to do something in-house, using the hierarchical control structure of the firm and avoiding the complexities of dealing with markets, that’s what we can expect. If, on the other hand, it turns out that it’s more efficient to hire somebody else to do the job, we can expect companies to do that. Companies have to balance the difficulties of managing a project in-house versus the costs of negotiating contracts with independent vendors.
There’s nothing surprising here, and Coase’s paper is fundamental, but Schwartz’s suggestion as to how cloud computing will likely significantly revise corporate decision-making is worth a close look:
“Instead of reliance on supervised relationships within firms, the cloud makes possible a new use of the price system. In the context of the cloud, Coase’s 1937 insights point to the conditions under which companies would shift from “make” to “buy” for networked computing services. Coase’s “Nature of the Firm” predicts this result when the transaction costs of purchase, including the negotiation of the necessary contracts, are less than the management costs of computing operations within the firm” (1658).
Coase thought that the IT of his time – specifically telephony – would tend toward “Buy,” as companies went in-house. Schwartz, by comparison, suggests that firms will increasingly be able to treat data storage and processing like electricity, buying the capacity they need on “spot markets” (the comparison of server space to the electrical infrastructure was made a while ago by Nicholas Carr). This saves significant resources on computing and other IT, as well as dramatically reducing the need for technical staff to support those computing resources. Although Schwartz doesn’t make the point, I think one can add that centralizing applications to corporate servers can make them more secure, by reducing the need to persuade millions of users to install security patches (see Jonathan Zittrain on this development, which he thinks threatens the open architecture of the Internet). Of course the privacy and individual security implications of all this are worrisome, as deeply personal information travels rapidly all over the place, and is often for sale at very low prices, as many have noted.
Why do I think this is significant outside of the realm of privacy and security? The idea that the rise of neoliberalism over the last few decades has significantly extended the reach of markets into new spheres of life is very common (I think it’s bascially correct, too). This is one of Foucault’s main points in his discussion of Becker, and it’s also a point explicitly made by Italian autonomists in discussions of the “complete subsumption” of society by capital. What’s interesting here is that we don’t have the inclusion of something new in market relations, but rather an increase in the spread of the price mechanism within market relations. In short, we have the increasing prevalence of a key neoliberal concept within capital, i.e., that capital relations are increasingly oriented around the price mechanism and associated techniques.
For Hayek in particular (here’s I'm outlining the view, not evaluating it), the price mechanism is fundamental to both the epistemic and normative defense of markets. It’s normatively preferable, he thinks, because it involves minimal coercion in comparison with command economies: nobody is required to produce or not produce anything; instead, they can tailor their economic activities however they see fit, and with an eye toward their own gain. Epistemically, the price mechanism offers a decentralized way to signal the social value of a given good. This has substantial advantages of Soviet-style planning (for Hayek, there really are only two Manichean alternatives: very free markets, and command economies), which had to master a lot of data in a hurry. The other feature of the price mechanism (one sees this emphasized in Foucault’s account) is that it enables more areas of “competition” and entrepreneurialism insfoar as corporations rely on competitive market structures rather than internal hierarchies.
What Schwartz’s analysis underscores, I think, is how technological developments like cloud computing are deeply intertwined with the increasing diffusion of (and enthusiasm for) these neoliberal economic structures, and how its technical affordances tend to structure the kind of capitalism we have. It’s not that technological developments are somehow exogenous to the economic system in some transcendental Heideggerian sense; rather, it’s that they are part of the story that needs to be told as we develop a more precise picture of how the move to “cognitive capital” and the decline of Taylorism is paralleled with the emergence of neoliberal theory.