1 Introduction

A key feature of the present global economy is the belief that data resources are the engine of economic growth, leading to a pervasive urge to capture, store, and process data (Fourcade and Healy 2017). The global battle for control over data resources—including their movement across borders—is fierce. The US, home of the leading global Internet corporations and working in their interests, has resorted to bilateral and plurilateral trade agreements to secure the corporations’ transnational control over data resources and “free trade in data” (save for national security reasons). Any regulation putting restraints on the free transborder flow of data is cast by the US as illicit “digital protectionism” obstructing economic growth and technological innovation.

Julie Cohen suggests that US domestic data law and policy has engendered a “biopolitical public domain”. The expression public domain is drawn from US intellectual property law and signals that a domain is characterised by “abundance”, that prior claims to resources in the domain are absent and that the domain may be approached as “a source of raw materials that are there for the taking and that are framed as inputs to particular types of productive activity” (Cohen 2019, pp. 48–49). The public domain does not allow for exclusive claims to resources, but the resources may be appropriated freely as inputs to profitable activity (Cohen 2019, p. 50). Cohen (2019, p. 67) characterises the public domain of personal data as biopolitical, because the data extracted from the domain are refined to identify patterns and produce predictions with a view “to make behaviours and revealed preferences calculable, predictable, and profitable in aggregate”. The US international law project to secure the free flow of data across the globe, could in line with Cohen’s conceptualisation, be described as an effort to construct a global biopolitical public domain.

This effort, however, predates both the framing of transnational data movements as a trade issue and the ascendancy of the biopolitical use of data as the dominant form of data use. When computer technology was connected with telecommunications in the late 1960s, making possible the movement of considerable quantities of data across borders in no time, it was not self-evident how the movements of data across borders should be cast juridically. Through acts of legal imagination, which Koskenniemi (2021) describes as the act of employing old legal vocabularies in new ways in new circumstances to achieve desired outcomes, states wishing to contest the de facto mobility of data across borders, presented data as a resource over which permanent sovereignty could be claimed. In response, the US, home to the corporations in charge of the bulk of the transborder data transfer, suggested that data were the kind of information (speech), the free circulation of which existing human rights instruments endorsed. The project of ensuring the global free flow of data (hereafter “free-flow project”) was thus born.

Over the course of the more than 50 years that the free-flow project has existed, there have been two major shifts in how the project has been legally articulated. The free speech argument gave way to a free-trade framing in the early 1980s when the terms of free trade were being re-negotiated during the WTO Uruguay round, while transborder business-to-business digital transactions of goods and services were expanding. The WTO agreement entered at the same time as the breakthrough of the Internet as a transnational marketplace in the mid-1990s. The limited and complicated ways in which certain kinds of cross-border data mobility are protected from certain kinds of restrictions in the WTO framework were unsatisfactory for the US-side from the beginning, but the securing of transnational data mobility became even more important with the rise of the data-driven economy in the early 2000s and the biopolitical turn in data usage. At this point, the US turned to concluding bilateral and plurilateral free trade agreements with a view to create a global public domain of personal data and suggested that cross-border data transfers are a matter of international trade even when unrelated to transactions in goods or services.

In this article, we give context and detail to the ways in which the free-flow project was legally articulated prior to its current framing as a matter of free trade, even without any link between data transfers and trading goods or services. To do this, we rely on existing accounts, from different time periods, in which the free flow of data has been discussed. We will particularly consider how existing legal tools and resources have been mobilised for new purposes to highlight how legal imagination was used in the service of, and against, the free-flow project. After presenting the two legal framings of the free-flow project in chronological order, the article ends with reflections on the workings of legal imagination in the context of informational capitalism. Koskenniemi clearly links legal imagination to the possibility to exercise agency and creativity, but also suggests that its use is about finding a powerful justification for acting or taking decisions in some particular way; decisions that have been arrived at on the basis of considerations other than existing law. What were those primary concerns that legal imagination served in the struggles for and against the free-flow project? How and why did these concerns change? And to what extent, and in what sense, did the exercises of legal imagination in and against the free-flow project, display the exercise of agency and creativity? As these questions show, we assume a constitutive, but not commanding, role for law in informational capitalism.

2 The 1970s and 80s: the rise of transnational data flows and US efforts to fend off regulation

The convergence of digitalised telecommunications and computers in the 1960s and 70s produced integrated systems for collecting, processing, transferring, and accessing data that came to span much of the globe. Vast amounts of digital information could be transmitted over long distances in no time and at low cost through a “complex, multilayered fabric of transnational interconnection among heterogeneous private and public systems” (Drake 1993, p. 270). The impetus behind the development of these systems was the needs of transnational business (Hamelink 1994, p. 30), and by the end of the decade “a significant portion of the global economy had migrated into the transnational cyberspace” (Drake 1993, p. 271). Transborder data flows had become “the backbone of international business and banking” (Hamelink 1979, p. 147). The globally interconnected computer/communication systems were of particular importance to large transnational corporations (TNCs) who depended on them to coordinate and optimise their globally dispersed production and distribution processes.

The US was not only home to the TNCs responsible for the absolute majority of the transnational transmission and processing of data (Mowlana 1984, p. 45) but also to the electronics corporations controlling the necessary infrastructure. Reflecting these “interests”, the US position was that data should be free to move across the globe and that the calls for regulation of data movements across borders (through taxation, localisation requirements, etc.) at national or international level had to be pushed back. US policy on international communications included assuring “US multinationals and others non-discriminatory access to low-cost, efficient information systems” and “non-discriminatory commercial opportunity for U.S. firms that are marketing international data processing and data bank services” (US Department of State, quoted in Pipe 1979, p. 117). Tagging on to the established notion that the public has a right to communicate across borders, as articulated in article 19 UDHR and elsewhere, the US presented the free-flow project as a human rights issue relating to “freedom of information”, which would supposedly include a right to “corporate speech” (Drake 1993, pp. 279 and 294). “Freedom of information” was thus not only the banner under which “Voice of America” did its famous radio broadcasts in the Soviet Union at the time but also provided a human rights-framing for the right of US TNCs to move data across borders.Footnote 1

The flows of data across borders had in the 1970s become “an essential ingredient in the wider globalisation process that [was] integrating national economies more deeply than before” (Drake 1993, p. 288) and many states feared that private capacity to transfer, for example, electronic money across borders in no time would undermine the efficiency of national legislation and debilitate jurisdictional authority. Realising that the world had entered the age of informational capitalism, competing industrial states became concerned about US dominance and control over the information intensive industries (Drake 1993, pp. 273–274). The fact that “valuable information [was] being deposited in the United States without an equal flow in the reverse direction” (Novotny 1980) was feared, because it would lead to economic dependence and reliance on US private firms.

In a similar register, Third World states regarded the unbalanced North–South flow of data together with the unequal distribution of databanks, critical data, and data processing capacity as a contributing factor to the continued dependency of the Third World on industrial countries. They argued that free flow would undermine their development process and, in particular, stifle the development of domestic information technology and services (Cong 2022, p. 54; Drake 1993, pp. 275–276). Against the US’s free-flow project, they mobilised “sovereigntist positions”: advocating national restrictions on data outflow, industrial policies, and transfer of technology programs to boost domestic capacities and the introduction of an international regulatory regime on transnational data flows (Drake 1992, p. 275).

The international alliance that came to articulate the most systematic position against the free-flow project was the Non-Aligned Movement, the members of which were mostly Third World states. To strengthen their own communication capacity and to rebalance the international flow of information, an agenda for a New World Information and Communication Order (NWICO) was launched in the mid-1970sFootnote 2 (Cong 2022, p. 49). Here transnational data flows were placed in the much broader context of asymmetries in flows of information, including flows of mass communication, between developed and developing countries that, so the NWICO-protagonists argued, undermined their political and ideological self-determination and would lead to the spread of capitalist cultural values and thought patterns to the rest of the world (Cong 2022; Drake 1993).

The proponents of NWICO denaturalised the “American” version of the freedom of information by pointing out its intimate ties to the freedom of commerce and to the global expansion of the US economy (Cong 2022, p. 67). The new order envisioned would be built on a conception of information as a public good, not a commodity, while freedom of information would no longer be the proprietary rights of “those already controlling the media to control the information flow and to disseminate information” but instead be given a collective meaning of control over information infrastructures (Cong 2022, pp. 52–53).

UNESCO picked up the cause of NWICO and became the principal forum for discussion on the terms of international information flows. The move to UNESCO, which is mandated by its constitution to “promote the free flow of ideas by word and image”, became a move towards pragmatism as UNESCO focused on building and financing communication infrastructures (Cong 2022, p. 73) and yet the association of UNESCO with NWICO and the New International Economic Order (NIEO) to which it was strongly linked, led to the withdrawal of the US and UK from the organisation in 1984 and 1985, respectively. In the words of Wanshu Cong, “by the end of the 1980s, NWICO and its radical normative challenge on the liberal-capitalist notion of freedom of information became largely irrelevant to the discussion and policymaking on international communication” (Cong 2022, 66). With this, sovereigntist positions regarding transnational data flows also disappeared as a relevant position in the international law contexts. Inger Österdahl has suggested that the conflict over the meaning of the right to freedom of information, which included the issue of transnational data flows, ended simply because of the end of the Cold War and the Eastern Bloc states’ embrace of a liberal understanding of human rights (Österdahl 1992). William J. Drake has offered a number of more specific reasons for the “premature resolution” of the debate on transnational data flows.

Drake emphasises the significance of corporate mobilisation. In the early 1980s, American TNCs aimed to overcome “the US versus the rest”-framing of the problems of transborder data flows by forging new alliances of business associations supposedly representing the perspective of corporate users everywhere. The understanding was promoted that restrictions on data movements would decrease efficiency of production and suppress wealth creation in general by denying domestic corporations everywhere access to vital information and to new and innovative services (Drake 1993, p. 295). This proved to be a successful strategy at a point in time when the notion was taking hold that regulating corporate activity would be bad for everyone. Drake also points out internal splits in the pro-regulation camp as a relevant factor. Many industrialised states shared the concerns of Third World states about losing economic independence, and there was also tangible anxiety in a few, such as France and Canada, about the wider effects of networked information on culture and national identity (Drake 1992, p. 277). Nonetheless, these states agreed to “quickly work through the issues and preemptively establish the international consensus” (Drake 1993, 276) because of worries that a North–South redistributive agenda might emerge proposing the transfer of wealth and strong controls. Such a development would negatively impact their own TNCs, which were struggling to catch up with the Americans.

The emerging international “consensus” was that regulation of transnational data should only be allowed to protect personal data. Privacy had been a major concern in several European countries ever since the 1960s and the emergence of computerised data banks. The merger of telecommunications and computers only exacerbated those. In the 1970s, multilateral discussions had been initiated in both the OECD and Council of Europe (COE) concerning what information about people would be permissible to gather and process in computerised databases (Palka 2020). Soon transnational data flows emerged as an issue in these discussions, because the fact that massive amounts of data about people could easily be transferred beyond borders undermined national regulatory efforts. COE adopted a convention on transnational data transfers and privacy in 1981 which allows governments to prevent transmissions destined for non-signatory states. OECD adopted rules with similar content that same year. Due to US pressure, the rules were non-binding guidelines. By the early 1980s, many industrialised states had adopted data protection laws in line with the COE convention, but corporate opposition had blocked the adoption of a coherent law in the US. However, data protection rules were on the whole quite ineffective, because they relied “on TNCs willingness to comply” (Drake 1993, p. 298). Personal data protection frameworks continue to be inefficient today and ultimately rely on corporate goodwill (see e.g. Brännström 2023).

Before continuing to the next stage of the free-flow project, it may be worth recapitulating how legal imagination was put to use during this initial period. The efforts to secure the growth, mobility, and effectiveness of US TNCs hinged on imagining data as information, and its exchange as a matter of freedom guaranteed by international human rights law. Thus, “freedom of information”, including a right to “corporate speech”, became the main justification for free flow. The privatisation and commodification of the “freedom of information” argument was contested through a counterimaginative move, most clearly articulated in the NWICOs “sovereignty over information”.

3 The 1980s and 90s: the Internet as a “global marketplace” and free flow as trade

The fact that the International Chamber of Commerce (ICC) adopted its “Uniform Rules of Conduct for Interchange of Trade Data by Teletransmission” already in 1987 testifies that business-to-business international e-commerce was well established already in the late 1980s. International e-commerce was conducted through what at the time was called Electronic Data Interchange (EDI), that is to say “computer based systems that allow companies to order, invoice, and bill their products and services electronically” (Garcia 1995).Footnote 3 However, transnational data flows were also used to serve hosts of customers distributed around the world in industries, such as advertising, insurance, accounting and finance, legal services, computer and data processing services, and so forth (Fiekele 1995). Already in the late 1970s, trade analysts and policy makers in the industrialised world suggested that these kinds of international, electronically mediated, transactions should be conceptually treated as trade and brought under the General Agreement on Tariffs and Trade (GATT) regime (Drake 1993, pp. 287–288). In the 1980s, the US officially presented transnational data flows as a trade issue related to trade in services (United Nations 1982: p. 6; Fascell and Schlundt 1983). Thus, even during the era which we described in the previous section, transnational data transfers were not only facilitating the internal coordination of corporations but also creating a marketplace in cyberspace, to which existing services could be redirected and within which wholly new services could be created.

By the early 1990s, nominal growth in world trade in services substantially outpaced that in merchandise and amounted to one-fourth the size of merchandise trade (Fiekele 1995). This happened at the same time as the closed US ARPANET infrastructure was opened up for private telecommunication actors, which eventually morphed into the Internet by political fiat. This changed the preconditions of international trade (Packard 2020). With the emergence of platforms that served consumers, international e-commerce was no longer limited to business-to-business transactions. Also, a range of products that traditionally had been delivered as goods, such as media products, film, printed material, and games, could now be delivered through the simple movement of data (Mattoo and Schuknecht 2000, pp. 4–5).

President Clinton was convinced of the potential of the Internet becoming the “most active trade vehicle [for the US] within a decade, creating millions of high paying jobs” (White House 1997). His senior advisor for Policy Development, Ira C. Magaziner argued at the time that there were two different models through which global Internet commerce could develop. One was the telecommunications and broadcast model, where the governments of the world would own, operate, and regulate industries. The other model was one “where buyers and sellers come together freely and do business, and the role of governments is simply to set a predictable legal environment for commerce” (Magaziner et al. 1998, p. 528). Choosing the latter model, the US policies aimed to institutionalise the internet as a “global marketplace”, resting on the principles of private sector leadership and a limited role for government intervention (White House 1997). This desired state of affairs required international regulation of “positive action” to make sure that the legal framework supporting commercial transactions was consistent and predictable regardless of where the buyer and the seller doing business were located. Regulation of “negative action” was also necessary to make sure that states did not introduce new and unnecessary regulations and procedures or new taxes and tariffs on commercial activities that take place via the Internet. “The internet”, a policy document from 1997 stated, is a truly global medium, and all nations will benefit from barrier-free trade across it.” (White House 1997).

Concluded by December 1993 by 117 nations, the Uruguay Round of multilateral trade negotiations generated the first comprehensive agreement on trade in services covering all the different means through which cross-border service transactions can occur. However, the introduction of barrier-free e-trade into the WTO framework would not come easy. E-commerce was first introduced at the WTO in 1998 at a Ministerial Conference with a Declaration of a Work Program on Electronic Commerce (Abendin and Duan 2021, p. 710). With the General Council in an oversight role, all major bodies within the WTO were tasked to take charge of discussions of trade-related issues of e-commerce: the Council for Trade in Services, the Council for Trade in Goods, the Council for Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), and the Committee for Trade and Development. The only outcome was that a 1998 2-year moratorium to the effect that “members would maintain their existing practice of not enforcing customs duty on electronic transmissions” (Abendin and Duan 2021, p. 709) has received regular renewals every second year. According to Abendin and Duan, three critical issues that have been discussed since then are: (1) the classification of electronic transmissions as goods, services, or intellectual property, (2) the effect of the moratorium on revenue-making, and (3) the technological feasibility of customs duties on electronic transmissions (Abendin and Duan 2021). Whereas the US and most other developed economies have sought trade liberalisation, developing economies wanted to decide on the above-mentioned preliminary issues or further investigate the implications for their economies of e-commerce and digital trade (Abendin and Duan 2021).

When the process of reaching new agreements on e-commerce within the WTO framework came to a halt in the early 2000s, the US spearheaded a move to bilateral and plurilateral free trade agreements which would include, inter alia, uninhibited transnational data transfer provisions (see, e.g., Wunsch-Vincent 2003; Streinz 2019). As prefigured in the introduction, this happened at a time in which the economy of the global flow of data was undergoing significant change related to emergence of digital commercial platforms and the data-driven economy more generally. Data-resources were no longer a means for providing digital products and services but an asset in themselves, because they could be used as an input to a range of profitable activities. As mentioned earlier with reference to Cohen, the commercial data use became biopolitical, because data are now refined to identify patterns and produce predictions with a view “to make behaviours and revealed preferences calculable, predictable, and profitable in aggregate” (Cohen 2019, p. 67). The appropriation of data resources takes place without money exchange and its casting as trade, stretches the meaning of the latter notion.

The rationale driving the US free-flow project has, however, changed after the US embarked on its bilateral and plurilateral trade policy. The project is no longer solely motivated by helping domestic data-driven companies to keep their dominant position globally, but also to counterweight the technological and economic rise of China, which transformed itself from an exporter of low-tech, labour-intensive products into a leading manufacturer of technological products and provider of digital services within a generation. In the words of Alvaro Santos et al., “China’s rise to economic predominance, and the West’s reaction to it, arguably constitutes the most important development affecting trade and investment law today” (Santos et al. 2019, p. 12).

China in its turn stresses that absent international legal obligations, each sovereign state is at liberty to decide if and how data resources may cross its border. Thus, the sovereigntist position in relation to transnational data movements has re-emerged. Beside China, measures restricting data flows have been considered in for example India in the last few years with a view to grow a domestic AI economy fuelled by “Indian data” (see Vila Seoane 2021). The United Nations Conference on Trade and Development (UNCTAD) suggests that such policy might be “the only way for developing countries to exercise effective economic ‘ownership’ of and control over the data generated in their territories” (UNCTAD 2019).

To sum up, during the second stage of the free-flow project, the primary use and value of transnational data transfers was no longer in facilitating the internal coordination of corporations but in sustaining a transnational marketplace in cyberspace, to which transactions could be redirected and in which new services could be created. During the period, legal imagination was invested in incorporating transnational data transfers which were necessary for digital cross-border transactions to take place, in the multilateral legal regime of international trade law that was in a process of consolidation and expansion. During this time, the opposition to the free-flow project was no longer concerned with the effects of the free flow of data on communities and legal counterimagination narrowed to matters relating to individual privacy. Starting in the early 2000s, the US free-flow project underwent major changes which led up to the current legal articulation of the transnational data transfers as a free trade issue, autonomous, and unrelated to trade in goods or services. Thus, the US turned away from multilateral to bi- and trilateral trade agreements to legally institutionalise this understanding at the international level. In parallel, however, sovereigntist positions reappeared which went beyond the privacy concerns that dominated the period from the 1980s until the early 2000s. The contemporary international legal discussion on transnational data movements is nevertheless still narrow in comparison with the 1970s, focused as it is on matters of economy and security.

4 Legal imagination in the age of informational capitalism

The free-flow project emerged to secure the growth, mobility, and effectiveness of US transnational corporations. Its aim was to prohibit international and domestic regulation of data movements across borders, and the choice of legal justification hinged on imagining data as information, and its exchange as a practice of liberty guaranteed by international human rights law. That the US would turn to human rights to support the global free-flow project was very much aligned with Cold War ideological battle lines.

The most radical challenge to the free-flow project came in the context of NWICO. The free-flow project was in this context seen as ultimately part of a grab of economic power by transnational corporations which reduced information to a privately held commodity—subject to a paradigm of liberal freedoms—and made the Third World into mere consumers (Cong 2022, p. 7). The legal counterimaginative move to US “freedom of information” became “sovereignty over information”, where “information” was cast a collective “social good” and a “cultural product”. Notably, the agenda of the NWICO saw the issue of the movement of data across borders in a very encompassing way, covering news, science (such as remote sensing of the earth's surface), and other forms of data too. This reminds us that the idea of sovereignty evoked in the context of NWICO was formed out of Third World concerns about colonialism, self-determination, and development and meant something more than independence pure and simple (see Anghie 2017).

The US turn to the rules and principles governing global markets and trade would soon change the parameters of the debate. What was achieved in this move was that a “broad and complex issue area” was “significantly narrowed” (Drake 1993, p. 260).

The opposition of “freedom of information” and “sovereignty over information” entailed a clash of liberal rights and public goods, giving way for an anti-imperial, anti-capitalist critique, centering wider questions of social and cultural independence. The shift to a trade framework marked a limiting of the discussion to a more technical debate within a largely neoliberal consensus.

The shift in how the free flow of data was justified occurred in parallel with the Cold-War bipolar dynamics giving away and being replaced by US unipolarity and reflected at once new ways of making profit off of data and the spread and gradual domination of a neoliberal consensus. As we have described, in the years to come, the question of privacy, raised already at the outset of the debate on transnational data transfers, would be the only check left, marking the outer limit of the freedom of trade.

In the early to mid-2000s, there was a shift in how profit could be extracted from data resources: whereas in e-commerce data resources function as a means for the provision of goods and services, it is the analysis of data resources which opens opportunities for profit-making in the data-driven economy. With the advent of the latter, the US free-flow project transforms into constructing a global biopolitical public domain and to secure the conditions for transnational data appropriation. Legal imagination is put to work with the effect that transnational data movements (involving no exchange of money) are now casted as an autonomous trade issue unrelated to trade in goods or services. Furthermore, as Streinz has argued, cross-border data flows do not conform to the prevailing understanding of trade law as concerning the exception to the default economic interchange within integrated national economies. Cross-border data flows are the norm in today’s digital economy and intra-country data flows are the exception (Streinz 2019). This is due to developments in the Internet’s digital infrastructure’s non-conformance with territorial limits (Streinz 2019). Involving no money exchange and reversing a foundational distinction between internal and external interchange, regulating the data flows of the data-driven economy by means of bilateral and plurilateral trade agreements is a rather creative act of legal imagination. This is true even if it builds on what was inherited from the e-commerce era. To some extent, trade agreements seem like an attractive avenue to regulate cross-border information flows for a different reason. They are both binding and enforceable (Burri 2017). The shift to bilateral and plurilateral free trade agreements can partly be explained by how the balance in the global trading system has made it increasingly difficult to conclude multilateral agreements within the WTO (Subramanian 2015). Free trade agreements have been a solution to large actors from the developing world, notably China, having an increasing impact, with some added advantages as well, notably, that it allows the US, as a large economy, to deal with smaller economies, thus retaining an overwhelming balance of bargaining power (Subramanian 2015).

In this project, we have tried to highlight the role of international law and of legal imagination in justifying and contesting the global free flow of data. We have done so assuming a constitutive, but not commanding, role for law. We have done so, because we do not believe that international law has been a reason for adopting a particular position with regard to free flow. At the same time, it cannot be said to have been entirely at the disposal of other forces, in particular the geo-economic interests of the US. The “entirely” is emphasised here, because, as Susan Marks (2009) has shown, and as later articulated by Robert Knox, “a vitally important aspect of international law as an ideological form is its tendency towards ‘false contingency’” (Knox 2014). In our case, the sense that the continuity of the free flow of data, and the profits made off it, is a random outcome. There is a persistence but also flexibility to the project, constantly adjusting to new means for profit-making, that should be acknowledged. This is not to say that the outcome was inevitable (that would amount to the opposite fallacy of “false necessity”). It is merely a way of pointing to the “close link between international law and processes of capital accumulation” (Knox 2014).