9.1 The Global Scope of the Tax Problem

This chapter examines the global regulatory issues concerning international taxation with particular attention to recent developments, in order to generate proposals for means of addressing platform content moderation issues at both the rule-making stage and the rule-application stage. We therefore provisionally assume that a potentially correct analogy between international tax regulation and platforms content regulation can be based on the fact that policy issues in areas are engendered by the prominent role that platforms – intended as “a foundation technology or set of components used beyond a single firm and that brings multiple parties together for a common purpose or to solve a common problem”Footnote ¹ – have assumed in the new models for conducting business. We will subject this assumption to critical analysis.

We therefore surmise that there is a similarity between the “opportunistic” platform-generated profits diversion that will be described in this chapter and “immoderate” platform content use and develop an analogical argument based on this similarity: since the two situations are similar, if it is true that the international community should establish regulatory structures that “moderate” the opportunistic use of new technological platforms that artificially divert profits away from taxation, then it is also probably true that the same international community should establish regulatory structures that “moderate” the use of new technological platforms that artificially distort information.

It is essential to remark here the first phenomenological resemblance between the tax regulatory problem and the regulatory problem of content moderation: non-state actors through platforms access, manage, and regulate flows of capital and similarly manage flows of information, designing new imaginary gateways for both capital and information that exploit regulatory differentials. So, in both cases there is the need for a regulatory response both at international and domestic level.

We can therefore argue that there is a striking “homology” between capital and information. Homology is not just a phenotypic similarity in the current aspects but a more profound similarity of two phenomena over time because they have a common origin.Footnote ² In social sciences, homology is a situation in which different human beliefs, practices, or artifacts share similarities due to genetic or historical connections. Here, in the discussion of regulatory options about profit diversion (based on capital) and content moderation (based on information), the homology is the underlying similarity between capital and information that in platform economy tend to be merged, becoming a hybrid constituted by intangible information, defined here as “computational capital.”

The upshot is that policies that counteract aggressive tax strategies that rely on fully mobile capital may be considered, in a broad sense, a potential model for policies for platform content moderation that rely on fully mobile computational capital. Therefore, one can hypothesize that tax regulatory approaches may be of potential interest for platform content moderation in spite of the fact that, as we will see, they have not so far proved entirely successful.

The chapter will begin by explaining the scope of the global tax problem (Section 9.1) and then address the related regulatory process at two levels. At a first descriptive level, the chapter will detail the process and institutional structures by which policies have been actually developed as well as their shortcomings (Section 9.2). At a second deeper explicatory level, the chapter will attempt to flesh out the phenomenology of actual platform operation and its global tax implications (Section 9.3). The chapter will conclude suggesting possible lessons that can be learned from tax regulation for platform responsibility rules (Section 9.4).

Let’s start with explaining the scope of the global regulatory tax problem. The international tax framework based on treaties established in the twentieth century aimed at preventing double taxation in case of reciprocal investments among countries. This framework appeared to be relatively stable until a novel issue emerged: countries unilaterally reduce their effective tax rates to attract capital, thereby affecting the mobile capital base determining what has been defined as “tax competition,” that is a race-to-the-bottom of tax rates, a competition among States to attract investments.Footnote ³ Under these competitive constraints, capital is invested in jurisdictions where the tax rates are lower.

Bilateral tax treaties do not address tax competition because they do not include tools to limit the decisions of key players about the location of investment and are just aimed at preserving the Westphalian sovereignty of the Contracting States. The consequence is that tax treaties do not provide tools for multilateral cooperation and governance in respect to tax competition so that there is the need to devise new regulatory tools.Footnote ⁴

The tax predicament, however, not only rests on a competition among States but also on the fact that multinational enterprises (MNEs) pursue action plans that thoroughly exploit tax differentials. We will see in detail what is the actual operation of platforms currently used by MNEs (Section 9.3), but one initially should be aware of the fact that, in a “classic” sense based on corporate legal concepts, MNEs respond to central management and control, relying on foreign controlled companies and chains of companies (“units”) that are coordinated under global tax strategies, even if they might be decentralized in terms of markets. By contrast, individual states only rely on local control of a territory for tax purposes. MNEs can thus be fittingly denominated “Global Actors” because they enter an asymmetric game in which the other actors are states, that is, local actors.

So, there is an additional dimension of tax competition that goes beyond a conflict-game among states leading to race-to the-bottom of tax rates: private actors such as MNEs intermediate between the state where they are based (the home state) and the states where they market their products or services (the host states) to reduce or eliminate taxation.Footnote ⁵ This goal is pursued through top-down “aggressive tax strategies.”

These strategies deployed by Global Actors are difficult to be defined in simple terms because they have reached a level of staggering complexity that cannot be recounted here but are in essence based on the “separate unity principle” accepted by virtually all corporate tax systems. This principle is derived from corporate limited liability and recognizes the separate nature of the units belonging to a Global Actor.

Under this principle, a Global Actor that is truly an integrated economic concern on a planetary scale is instead viewed tax-wise as a cluster of separate units, each unit being subject to taxation in the jurisdiction where it operates. Thus, it is possible for Global Actors under a centralized strategy to attribute profits or losses to those separate units of the appropriate jurisdictions: profits will flow to low-tax jurisdictions, costs will be deducted in high-tax jurisdictions, and losses will flow to profitable units to offset their profits.

The result is that Global Actors tend to be subject to a global effective rate that is lower than the weighted algebraic sum of the nominal rates of the different countries where they operate, and this leads to a situation where they operate in a sort of “meta-nation” ideally located in between the home state and the host state, not defined by traditional state boundaries.

This engenders a regulatory-tax arbitrage in which Global Actors are asymmetrically confronted by the limited reach of territorial states. This is a new brand of a geo-political architecture in which states and non-state organizations such as Global Actors access, manage, and regulate the intersection of territory and flows of capital, and, in so doing, design new imaginary gateways for a fully mobile capital that segregates profits in jurisdictions which offer lower rates under the constraint of tax competition.

In the tax world this regulatory-tax arbitrage has been defined as “base erosion and profit shifting” (BEPS). The base erosion component of BEPS is that deductions are made in high-tax countries, so that the national tax base of these high-tax countries is reduced by the subtraction of tax-deductible costs. The profit shifting component of BEPS is that these sums that are deducted are paid to entities of the same group located in low-tax countries, triggering a transfer of gross revenues to those countries.

The structural feature of BEPS comes from these two complementary components: there is a de-coupling of tax-deductible costs (artificially located in high-tax jurisdictions) and gross revenues usually of a mere financial nature and detached from actual operation (artificially located in low-tax countries, often defined as “tax havens” or “offshore financial centers”). In the context of content moderation, there is a similar arbitrage since platforms can be established in low-regulation jurisdiction, and yet offer services in high-regulation jurisdictions.

A burgeoning literatureFootnote ⁶ attempts to showcase the basic techniques of BEPS aggressive strategiesFootnote ⁷ that rely on the manipulation of the remuneration of debt (interest) or intangibles (royalties or other fees) within the different units of Global Actors.Footnote ⁸ Manipulation of intra-group transfer prices is also an important profit-shifting channel for Global Actors through the tax-induced management within MNEs of debt, intangibles, or intra-group services.Footnote ⁹

Another group of techniques that involve BEPS are those in which Global Actors are capable of achieving “double exemption” as a result of regulatory gaps or through complex techniques denominated “hybrid mismatches,” which take advantage of the differences of tax systems. Profits may even escape all forms of taxation: this situation has been defined as “stateless income/profits”Footnote ¹⁰ and can be restated as “stateless capital” in light of the fact that income/profits are generated from a fully mobile capital. Stateless capital challenges the foundation of the taxing powers of states anchored to a territory.

It is essential to remark here the second phenomenological resemblance between the tax regulatory problem and the regulatory problem of content moderation: if there is stateless capital, then equally there can be “stateless information” in the sense that both terms – stateless capital and stateless information – directly point to the fact that both capital and information escape the regulatory reach of states. Of course, this resemblance does not imply that content moderation should be proprietary to one or more individual states to the detriment of others, but some form of standard must be achieved.

9.2 The International Policy Response

These situations of reduced or nil taxation of MNEsFootnote ¹¹ epitomized in the BEPS acronym have been addressed by the OECD/G20 Action Plan (the BEPS Project), which unfolded in 2013–17, leading to the release of fifteen reports in 2014–15Footnote ¹² and to the modification of the OECD Commentary and Model in 2017,Footnote ¹³ followed in 2018–22 by the Inclusive Framework (the Inclusive Framework) initially with an Interim Report,Footnote ¹⁴ then by proposing revised “nexus rules” (Pillar One)Footnote ¹⁵ and different types of “profit allocation rules” (Pillar Two),Footnote ¹⁶ and finally by advancing a comprehensive proposal together with a program of work.Footnote ¹⁷

This section will detail at a first descriptive level the process and institutional structures by which policies have been actually developed as well as their shortcomings.

The fifteen Actions of the initial 2013–17 BEPS Project can be divided into three main areas: global disclosure rules, guidelines for national tax policies, and modifications of tax treaties. These BEPS measures were expected to become applicable via changes to bilateral tax treaties or through changes in domestic laws, so the purpose of the BEPS Project was to facilitate through soft law a complex implementation process both at an international and domestic level. The problem is that these OECD BEPS measures were not introduced through a multilateral legal instrument and therefore were not binding (except for those eventually included in BEPS-modified tax treaties).

In addition, the 2013–17 BEPS Project was affected by other structural flaws. First, the OECD was trying to forge consensus about complex rules to be implemented by states based on the idea that profits should be taxed where economic activities are performed and value is created.

The consensus should instead have been about direct constraints on the behavior of Global Actors; without such consensus, the BEPS situation was basically tolerated in the name of Westphalian tax sovereignty. Second, the 2013–17 BEPS Project failed to focus on the fact that aggressive tax strategies of MNEs exacerbate the negative effects of tax competition impinging on the powers of states. Third, the 2013–17 BEPS project preserved the single entity principle and thus did not hinder the capability of MNEs to shift around deductions, profits, and losses to the units located in appropriate jurisdictions.

In conclusion, in the initial 2013–17 BEPS Project, there was a certain amount of international cooperation in discussing policies, but there was no plan of an institutional architecture that could ensure adequate multilateral governance.Footnote ¹⁸

The BEPS Project at OECD level was continued by the 2018–23 Inclusive Framework, which focused on “nexus rules” (Pillar One) and profit allocation rules (Pillar Two).Footnote ¹⁹ After initial hesitation, the policy process of the Inclusive Framework was kickstarted in 2021 by the proposal by the United States of a “global minimum tax” in which individual home countries undertake the ethical and political obligation to tax their own multinationals on global profits in a kind of defensive alliance.Footnote ²⁰ In addition to the global minimum tax administered by home countries of MNEs based on OECD Pillar Two, market jurisdiction can of course adopt rules that protect their territorial tax prerogatives on the basis of OECD Pillar One.

If a global minimum tax proposed by the United States and other countries is agreed at the international level, each individual state will participate in a cooperative framework by unilaterally subjecting multinationals based in that home state to a minimum rate of tax (15 percent or more) on global profits. In this sense, the global minimum tax is a country-by-country extraterritorial tax.

Thus, the global minimum tax is not a systemic international mechanism to allocate profits of multinationals among countries but simply the alignment of home states unilaterally taxing their own multinationals. Each home state undertakes the responsibility to tax its own multinationals on the basis of extraterritorial legislative jurisdiction while host states remain free to exercise their tax power on profits sourced in their territory.

This is a promising basis for an analogy between international taxation and content moderation in terms of potential policy strategies. If regulation by the home state appears to be a “dominant strategy” in international taxation in the specific sense of game theory, that is, a strategy that is superior no matter what the other players do, then the same could be true in respect to content moderation. In that context, therefore, the home states of the main platforms should, in theory, adopt a dominant strategy by undertaking the obligation to regulate these platforms converging to basic principle through multilateral agreements. Once platforms are subject to common rules agreed by home states of the Global Actors owning such platforms, the policies of host states will be, to a certain extent, irrelevant insofar as they impose lower limitations.

The salient point here is that in international taxation, home state regulation is a dominant strategy because it ensures the collection of taxes irrespective of what other states do; while in content moderation, home state dominant strategies appear more problematic because the policy issue is not just in terms of collection, but it involves complex cultural and semantic problems of what content moderation should be.

A global minimum tax does not require fully inclusive multilateralism but can be pursued through so called minilateralism, a word and a concept suggested by Moises Naim in 2009:Footnote ²¹ when it is difficult to reach broad consensus, the most effective move is the inclusion of the smallest possible number of countries needed to have the largest possible impact on solving the problem of profit shifting. The same principle could apply in content moderation.

It should be clear from the outset that the OECD is not currently endorsing a global minimum tax. The term “global minimum tax” is often used in a generic way to denote the efforts of the international community to tax multinationals, but this term has an exact meaning that specifically comes from the US proposal of a “country-by-country minimum tax.” So a “global minimum tax” is a tax unilaterally imposed by each home state through extraterritorial legislative jurisdiction on the multinationals that are based in that state (i.e., when the ultimate parent company is resident there). To consider a simple example: the United States already imposes a kind of global minimum tax on the worldwide income of its citizens; this tax could be extended to multinationals based in the United States, and the same would respectively occur in other home states.

The OECD is not currently endorsing such a global minimum tax but rather is pursuing something different: an Inclusive Framework based on a pervasive consensus of more than 130 countries on a very detailed agenda that establishes a mechanism to allocate profits of multinationals among countries. This is an ambitious idea of establishing thoroughly common criteria to allocate profits but is not a country-by-country minimum tax “US style.” The OECD/G20 Inclusive Framework has in fact continued in its search of a global consensus about common rules trough a “two-pillar solution” based on Pillar One and Pillar Two.Footnote ²²

With regard to the reallocation of taxing rights to market jurisdictions (Pillar One), the OECD has defined as “in-scope companies” affected by the changes that are now negotiated the multinationals with global turnover above 20 billion euros and profitability above 10 percent (i.e., profit before tax/revenue). Extractives and regulated financial services are excluded.

There will be a new special purpose nexus rule permitting allocation of so called “Amount A” to a market jurisdiction when the in-scope multinational derives at least 1 million euros in revenue from that jurisdiction. Revenue will be “sourced” to the end market jurisdictions where goods or services are used or consumed. To facilitate the application of this principle, detailed source rules for specific categories of transactions will be developed.Footnote ²³

“Source” is a term of art in international taxation since its inception and indicates the establishment of criteria that indicate where specific items of income are actually generated, created, or located. This idea implies that it is possible to identify the structural features of economic activities on the basis of which allocating rules are capable of attributing taxing powers to the so called source state. Although commentators have criticized the idea of being able to pinpoint the actual source of income in the context of a globalized/digitized economy, the OECD is still advocating this idea, with the added complication that the proposed criteria are overly complex, presenting critical problems in terms of actual implementation.

There is also a so-called Amount B to allocate profits to market countries that reflects in-country baseline marketing and distribution activities in those countries that will be applied on the basis of the arm’s length principle, which will be simplified and streamlined.Footnote ²⁴ In-country baseline marketing and distribution activities are those typically developed by local units (such as agents, distributors, and affiliated companies) in the relevant markets to carry out the core minimally required business presence of multinationals and therefore a commensurate amount of income must be attributed to such units to reflect those baseline marketing and distribution activities.

By contrast, Pillar Two consists of two interlocking domestic rules (together the Global anti-Base Erosion Rules [GloBE] rules): (i) an Income Inclusion Rule (IIR), which imposes a top-up tax on a parent entity in respect of the low taxed income of an entity of the group; and (ii) an Undertaxed Payment Rule (UTPR), which denies deductions or requires an equivalent adjustment to the extent the low tax income of an entity of the group is not subject to tax under an IIR.

In addition, a treaty-based rule (the Subject to Tax Rule, STTR) allows source jurisdictions to impose limited source taxation on certain related party payments subject to tax below a minimum rate. The STTR will be creditable as a covered tax under the GloBE rules. The GloBE rules will apply to multinationals that meet the 750 million euros threshold as determined under BEPS Action 13 (country by country reporting). OECD guidelines for a minimum effective tax on multinationals (Pillar Two) have been thereafter released at the end of 2021.Footnote ²⁵

These OECD guidelines of 2021 were followed closely by the proposal of an EU Directive, which was finally adopted at the end of 2022,Footnote ²⁶ which introduces a minimum EU 15 percent tax rate that will apply to any large group, both domestic and international, including in the financial sector, with combined financial revenues of more than €750 million a year and with either a parent company or a subsidiary situated in an EU Member State.

The effective tax rate is established in each Member State by dividing taxes paid by the entities in that Member State by their income. If the effective tax rate for the entities in a particular jurisdiction is below the 15 percent minimum, then the Pillar Two rules are triggered and the group must pay a top-up tax to bring its rate up to 15 percent, which applies irrespective of whether the subsidiary is located in a country that has signed up to the international OECD/G20 agreement. The calculations will be made by the ultimate parent entity of the group unless the group assigns another entity.

If the global minimum rate is not imposed by a non-EU country where a group entity is based, Member States will apply the Undertaxed Payments Rule, which is a backstop rule to the primary Income Inclusion Rule. It means that a Member State will effectively collect part of the top-up tax due at the level of the entire group if some jurisdictions where group entities are based tax below the minimum level and do not impose any top-up tax. The amount of top-up tax that a Member State will collect from the entities of the group in its territory is determined via a formula based on employees and assets.

The last phase of the process is currently under way and reveals major criticalities of the OECD approach: the OECD in fact, at the World Economic Forum in Davos in May 2022, has admitted that its two-pillar project will not be implemented in 2023, as announced. One of the reasons for the impasse is certainly the sheer complexity of the guidelines proposed by the OECD, but there is also a structural reason, that is, the fact that the OECD is trying to reach an over-inclusive consensus (more than 130 countries) on a hyper-detailed agenda and at the same time is lacking a strategic vision: the attention is focused on technical details but there is a lack of institutional design. This reflects the depth versus breadth debate about policies and governance in international relations literature, also forming the basis for the concept of minilateralism.

In light of this impasse, the global minimum tax initially proposed by the United States in 2021 and then endorsed by other countries actually appears to be the only effective way to address the root of the problem: if the home states adopt a dominant strategy and tax the global profits (i.e., domestic + foreign) of MNEs based in those home states irrespective of what other host states do (tax or exempt), there will be no tax arbitrage because there will at least be a minimum tax rate (15 percent or more) in the home states, so that aggressive tax strategies that shift profits to lower tax jurisdictions will be futile.

So, to develop the potentially correct analogy that compares platforms and taxation, “global minimum moderation” should be pursued through a kind of “essential harmonization” with mutual recognition of minimum standards or criteria. This could be the platform analog of “global minimum taxation.”

As mentioned, the OECD has not fully adopted the proposal of a global minimum tax but has introduced very complex arrangements that jeopardize the establishment of an international tax regime that would trigger a paradigm shift toward cooperative multilateralism. Pillar One advances profit allocation rules to protect market jurisdictions and Pillar Two only proposes optional recapture rules (the so called top-up tax) for taxing foreign profits if subject to lower rates, but this is just a fallback option, and, in addition, countries are not obligated to adopt it.

The lack of an explicit will of the OECD to push for a real global minimum tax is a failure of its initial “inclusive” proposal. It is, however, still possible that countries freely adopting the global taxation minimum standard will form “defensive alliances” against tax competition and profit shifting through multilateralist geo-political compacts inspired by minilateralism,

As a matter of fact, in 2022–23, several countries autonomously began to implement Pillar Two through different types of country-by-country minimum taxes. At this stage, there is no explicit minilateralist pact among those countries, but their moves are an indication that minilateralism is beginning to work in the tax arena.

Thus, minilateralism could also work in the platform context of content moderation on the basis of its potential capabilities that have been described in international relations literature. These minilateral compacts in fact are a well-known form of institutional design because they aim at putting together a sufficient number of players to address a policy issue, without the need to strive in the direction of unattainable forms of unanimity or full coordination, often prone to free ride problems.

For example, even in the context of global warming – a typically global problem that requires a global coordinated response – minilateralist solutions have been propounded. Minilateralist solutions do not present the problems of the current over-inclusive proposals to introduce a mechanism of binding rules because they are based on the incentives of each actor to participate. Minilateral compacts in fact initially operate in a framework that has been defined by international relations literature as an “international regime.” International regimes are identified in the literature as “sets of implicit or explicit principles, norms, rules, and decision-making procedures around which actors’ expectations converge in a given area of international relations.”Footnote ²⁷

These principles, norms, rules, and procedures prescribe certain actions and proscribe others, so they imply obligations, even though these obligations are not enforceable through a hierarchical legal system. Thus, this definition adopts a conception of regimes as social institutions rather then sets of binding legal prescriptions.

Minilateral compacts have several main advantages. First, they do not imply the detailed rules through formal international agreements as they only require the participating home states to extend a minimum to Global Actors based in those states on their global operation. Second, cooperation is pursued among countries genuinely sharing the common policy subjecting multinationals to regulation. Third, it is more likely that a compact is formed for effective change on clearly identified measures when the number of participating actors is not over-inclusive. Finally, minilateral compacts initially operating as international regimes facilitate the eventual achievement of more inclusive cooperation and thus are the basis for subsequent binding international agreements. These advantages of minilateralism could potentially apply to international tax regulation and content moderation.

Specifically, in tax minilateralism, the states that are the most capable of taxing multinationals on global profits act together with the states that are most vulnerable. It is also possible that states cluster in different compacts on the basis of geopolitical criteria in a kind of “polycentric cooperation.” In a scaling down of the current inclusive approach, the OECD could, acting as an orchestrator, thus become the venue of one or more minilateral compacts. A similar institutional dynamic could also be activated in content moderation, possibly through the orchestration of an international organization.

Minilateralism could also be seen as the “club of the powerful,” for example, in tax matters developing countries often balk at the strategy of a core of home states taxing their MNEs globally, claiming that it would only be in the interest of OECD capital-exporting countries. It should, however, be clear that such a strategy would not limit the tax prerogatives of other countries, such as BRICS or developing countries: these countries obviously retain their right to tax multinationals under OECD Pillar One or otherwise, irrespective of the fact that multinationals will be taxed in their home states.

Similarly, in content moderation, regulation by home states simply ensures application of a minimum standard but does not exclude regulation by host states. It is worth mentioning that in the context of content moderation platforms often have limited moderating resources and so focus those resources in highly regulated jurisdictions such as the US or EU, as opposed to low regulated jurisdictions; but when a minimum standard is ensured in home states, moderating resources can be internally distributed within the Global Actor on the basis of opportunity and efficiency, rather than on the basis of level of regulation.

At the end of the analysis of the global policy process developed in this section, it is essential to remark a third phenomenological resemblance between the regulatory problem of content moderation and the tax regulatory problem: if it is true that we conceive of a minimum standard in a minilateral setting – at least initially – through which individual home countries undertake the ethical and political obligation to tax their own multinationals on global profits in one or more defensive alliances, we could conceive of a minimum standard through which individual home countries undertake the ethical and political obligation to moderate platform content.

The alternative regulatory model – both for global taxation and content moderation – is to pursue an inclusive consensus on a very detailed agenda to be enforced through formal multilateral treaties: in that context, if consensus on detailed rules cannot be achieved and pinned down in binding rules, resort must be made to guidelines and soft law, which can prove to be ineffective.

9.3 Tax Implications of Global Platforms

After having described in Section 9.2 the process and institutional structures by which policies have been actually developed as well as their shortcomings, this section will attempt, at a deeper level, to explicate how and why actual platforms’ operation have relevant global tax implications and possibly what inferences or comparisons may be drawn from the tax regulatory world to content moderation.

I anticipate that these structural implications point to a potential problem of “intractability” for tax regulation of Global Actors. I define as “intractability” the conundrum or cul-de-sac in which state actors would find themselves if they exclusively rely on their territorial prerogatives to tax profits that are deemed to be located “somewhere,” while they are essentially located “anywhere.”

So, the point I make here is that to overcome the problem of intractability the only feasible way appears to be a multilateral approach based on the individual extraterritorial jurisdictions of state actors willing to participate in an alliance that sets certain minimum standards on the taxation of global profits.

Aggressive tax strategies described in Section 9.2, when considered at the deeper explicatory level of this section, should be viewed as being foundationally based on technological constructs that rely on the “extraction” and use of data through platforms. This “data extraction” is explained as follows: platforms rely on raw data to generate knowledge and to achieve that goal, raw data is recorded and organized to be usable by the corporate legal owners of platforms, the Global Actors. In the extraction process raw data consists of information about facts of the world, while knowledge consists of explanation and use of such facts.

Therefore, for Global Actors the raw material that must be extracted is the data, the source of this raw material are the activities of users on platforms, and the organization of such data is the knowledge accumulated by Global Actors. Platforms extract, analyze, and use the increasingly large amounts of data that are recorded as a form of computational capital to make profits.

Once the new business models became wholly based on such technology aimed at the extraction and use of data through platforms, it came as a natural result that aggressive tax strategies were molded by their technological base of data extraction and thus morphed into brand new structures significantly different from the previous ones based on legal arrangements. At the same time, however, the traditional corporate form of Global Actors essentially remained unchanged and based on the separate unity principle described in Section 9.1, which allows base erosion and profit shifting. So Global Actors “own” platforms in the traditional corporate sense but through them pursue aggressive tax strategies that are different from preceding ones.

In a very general sense, these new aggressive tax strategies exploit to the maximum extent the potential of data to generate through platforms new types of profits that are not taxable. We generally denominate these new business models as the “digital economy,” alluding to the fact that the traditional industrial economy based on production and distribution of goods and services now relies on new information technology that improves that production and distribution. This is only a partial dimension of the new reality: what in effect happens in such a “digital economy” is structural reliance on global platforms to expand the reach of traditional profitability in an “extractive mode” that leads to the accumulation of so called big data. Massive new expanses of potential data are in fact now accessible so that data itself has become a fundamental resource to be extracted, as much as oil and other natural resources in the past.

For example, Google initially limited itself to use data to pivot advertising revenues away from traditional media, but then platform data acquired primary importance for many other Global Actors as data analysis is itself generative of data in a continuous feedback loop. The result is that data serves a whole range of functions across different industries ranging from technology (Google, Facebook, and Amazon), dynamic start-ups (Uber, Airbnb), manufacturing (GE, Siemens), agribusiness (John Deere, Monsanto), and so on. The common denominator is that data mainly generates new markets and products and cost reductions of revenue-making activities, therefore leading to more revenues and higher profitability.

Aggressive tax strategies exploit “platforms” to extract data, but what exactly are platforms in this context and how are they used taxwise? Platforms are IT infrastructures that enable two or more groups to interact. Therefore, platforms are intermediary technological frameworks that not only bring together different users (customers, advertisers, service providers, producers, suppliers, even physical objects)Footnote ²⁸ but also provide tools that enable these users to build their own products, services, and marketplaces. For example, Apple and Microsoft enable developers to build and sell new apps to users, while Google provides a platform for advertisers and content providers to target people searching for information.

So corporate owners of platforms became intermediaries vis-a-vis users to generate new products and profits,which could not be generated through traditional methods. There are several characteristics of platform content that have a connection with profit-generating activities and aggressive tax strategies aimed at diverting such new profits away from taxes. These brand-new characteristics of platforms are the following, each of them generating an asymmetry between Global Actors and taxing States: (1) intermediation and positional asymmetry, (2) network effects and growth asymmetry, (3) complementary strategies and per-capita asymmetry, and (4) neutrality and strategic asymmetry:

• First, platforms provide the infrastructure to intermediate between different groups and therefore Global Actors who own platforms do not have to build a marketplace nor even necessarily maintain a physical presence in certain market jurisdictions. For example, Google relies on online search activity, Facebook on social networking, Uber shifted urban private transportation into an online platform. This is the key to the advantage of platforms over traditional business models, which require physical presence in market jurisdictions.

• The disappearance of physical presence is also a competitive advantage of platforms owned by Global Actors over taxing States and results in a positional asymmetry: while Global Actors owning platforms do not need to position themselves on a geographical space, tax regulators traditionally need geographical space to detect profit-generating activities and assert jurisdiction.

• Second, platforms are based on network effects: the more numerous the users who use a platform, the more valuable that platform becomes for all other users and for the owner of the platform. For example, Facebook has constantly increased its social networking platform and Google improves search algorithms through the increased use online. This implies that platforms are capable of rapidly scaling up their operation by relying on preexisting infrastructure and incurring low marginal costs of production, so that there are few natural limits to the growth of platforms. The result is another type of asymmetry, namely a growth asymmetry: while Global Actors relying on platforms are innately bound to expand on a global scale, tax regulators are limited by their local geographical space.

• Third, platforms use complementary strategies: for example, one division of the Global Actor reduces the price of a service or good (even providing it for free) and generates losses, while another division increases prices to offset these losses generating profits. To pursue these strategies, platforms increase the number of their users at a global scale by offering services that often are not directly priced. For example, Google notably started as a completely free service but then began to make profits through advertising and then through a host of services.Footnote ²⁹

• The result is that there is another type of asymmetry, a per-capita asymmetry: while Global Actors relying on platforms do not cater to a population that is identified by the allegiance to state territory, tax regulators inherently need to interface directly with a locally settled population that exhibits an ability to pay taxes that are linked to the use of local indivisible services and public goods.

• Finally, platforms present themselves as neutral spaces for users to interact but in fact pursue deliberate strategies in terms of product and service development, marketplace interactions, and so on. The typical example is Facebook, which epitomizes the evolution of online social networks presenting its the platform as a free forum when in reality it is a tool to collect valuable data. This strategy is also used by many industrial players as a complement to their core business.

The result is that there is another type of asymmetry, a strategic asymmetry: Global Actors portray themselves as neutral and transparent platforms but deploy strategies that are not visible to users and to tax regulators; by contrast, tax regulators necessarily use and announce policies and regulations that are thoroughly visible in the public-civic space to Global Actors.

In addition to these four basic characteristics/asymmetries of platforms, there are also different types of platforms depending on the specific activities that Global Actors perform: advertising platforms, cloud platforms, industrial platforms, product platforms, and lean platforms.

Advertising platforms (e.g., Google, Facebook), extract information on users and then use the products of that process to sell advertising and for other uses. Cloud platforms (e.g., AWS, Salesforce), which own the hardware and software of digital-dependent businesses and are renting them out as needed. Industrial platforms (e.g., GE, Siemens), build the hardware and software necessary to transform traditional manufacturing into internet-connected processes that lower the costs of production and transform goods into services. Product platforms (e.g., Rolls Royce, Spotify), generate revenue by using other platforms to transform a traditional good into a service and by collecting rent or subscription fees on them. Finally, lean platforms (e.g., Uber, Airbnb), which reduce their ownership of assets to a minimum and profit by reducing costs.Footnote ³⁰

The four characteristics – asymmetries of platforms (intermediation and positional asymmetry, network effects and growth asymmetry, complementary strategies and per-capita asymmetry, neutrality and strategic asymmetry – are thus combined within different types of platforms (advertising, industrial, product, and lean platforms) generating different and specific tax strategies that cannot be recounted here in detail.

What are then the tax implications of global platforms? In a general sense, intermediation and positional asymmetry significantly limit host-States’ tax enforcement power for the lack of physical presence, typically a fixed place of business in the form of a permanent establishment (PE) or through a dependent agent, so called agency PE. Network effects and growth asymmetry attribute to Global Actors flexibility and resilience unattainable by state actors, while complementary strategies and per-capita asymmetry create vast opportunities for regulatory tax arbitrage. Finally, neutrality and strategic asymmetry create a gap between the opacity of strategies of Global Actors and the forced transparency of tax regulators.

The upshot is that Global Actors using platforms have the capability of escaping local state tax enforcement, are much more flexible than States in implementing advantageous strategies, rely on regulatory arbitrage, and are capable of pursuing aggressive tax strategies that are opaque to tax regulators. The practical result is that for each individual Global Actor, there is a combination of platform-based features that is the signature of that player. In addition, these features are entirely proprietary to that player and very opaque to the outside, with the consequence that the “formula” of the aggressive tax strategy of each individual Global Actor is fiercely guarded against intrusions by tax regulators.

The capability of tax regulators to “pierce through” the unique “formula” of “opportunistic” platform use would be tantamount to the dismantling of aggressive tax strategies, but this achievement is, at the current stage, highly unlikely. This deficiency points to what I consider a problem of intractability, that is, the impossibility of a territorial approach, which poses the need to strive for a multilateral (initially minilateral) approach based on the global reach of regulatory states aggregating in minimum standards compacts.

It is essential to highlight at this point that there is a fourth resemblance between the tax and content moderation regulatory problem as they are both affected by this intractability problem: if it is true that from a territorial perspective, BEPS involving the strategies described above is not readily tractable, then it is probably true that content moderation is not tractable exclusively from a territorial perspective as it relies on similar strategies. So, in both cases, minilateral minimum standards should be pursued. Of course, the intractability of content moderation is, in addition, caused by various factors that are not relevant for tax policy.

9.4 Possible Lessons for Platform Responsibility

As tax regulatory approaches can potentially be of interest for platform content moderation, we draw in this section a few conclusions to suggest possible lessons for platform responsibility rules. In doing that, I would like to make three clarifications about differences between the tax regulatory problem and the content moderation problem.

First, while the tax regulatory problem, by definition, involves a top-down exercise of sovereign powers to limit the opportunistic use of global platforms that leads to regulatory arbitrage by corporate actors, the content moderation problem involves a delicate balance between the exercise of sovereign powers and the protection of freedom of speech in many cases of individual citizens that access and use global platforms.

Second, in the international tax arena, a balance is pursued to ensure that each item of income is taxed at least once but no more than once, while in content moderation, a balance could be pursued to ensure that common principles are established once and for all, but this goal is much more difficult. In international taxation, a “follow the money” principle decides the game: once the home state has ensured collection of global profits of multinationals that are subject to its responsibility, the game is over and taxation is secured. This is not necessarily true in content moderation, where complex issues are at stake and national policies may overlap.

Third, from a strictly semantic perspective, the two problems appear different: tax regulation does not look at the meaning of information but strives to tax profits that arise from the use of data/information by platforms, while content moderation inevitably looks at the meaning of information to moderate it.

In spite of these three differences, our analysis has demonstrated that in both tax regulation and content moderation platforms managed by Global Actors have an impact on the perception of reality by users and influence their behavior. A platform – which is a technology that brings multiple parties together for a common purpose or to solve a common problem – is in fact a standards-based system that at the same time (i) distributes interfaces among users through remote coordination and (ii) centralizes the integrated control of interfaces and users through that same coordination. As a result, platforms distribute some form of autonomy to the edges of their networks while standardizing the conditions of communications.

Interfaces appear obvious and transparent to users, but the internal architecture of the platforms remains unknown to them and constitutes the medium through which strategies are developed by Global Actors. As Bratton notes:

Interfaces are representations, in a broad sense “images”: Jean Baudrillard, a postmodern French philosopher, had anticipated during the 1980s this idea of representations or images as a means of concealing this absence of reality that characterizes this current cognitive fundamentalism, calling such representations “simulacra.”Footnote ³² He envisaged successive phases of the modes of representation in which signs become increasingly empty of representational meaning and indeed the apex of this process is reached when users access certain existing platforms through their interfaces.

Baudrillard so described the “image” in a process that starts from faithful representation of external reality to the dissolution of it:

So essentially the “immoderation” of content generated within platforms is in most cases a lack of representational meaning for the users tethered to a precise platform’s strategy, what Bratton defines as “forms of cognitive fundamentalism.” In addition, “immoderation” of content is also the result of deliberate and controlled distortion of reality when strategies are willfully pursued through “echo-chambers,” “trolls,” and other techniques to influence users in the most diversified contexts.

These innate platform constraints on the representational function of content point to the need of a regulatory regime that relies on rules on procedural accountability that include, for example, common principles for the “notice-and-takedown” procedure and the strengthening of the cooperation with public authorities.Footnote ³⁴ Such a regime of course would not exclude traditional types of moderation that depend on some sort of semantic control, such as automated moderation, pre-moderation or post-moderation, reactive moderation, and distributed moderation.Footnote ³⁵

These rules on procedural accountability for content moderation are somehow similar to the basic obligation of Global Actors owning platforms to report all their global profits at least in one state, their home state, and therefore to be accountable/pay taxes on those profits. So, in both tax regulation (dealing with capital) and content moderation (dealing with information), the common regulatory problem is how to subject to procedural accountability inappropriate outcomes of platforms that strategically control computational capital.

At the beginning of the chapter, I argued that there is a homology between capital and information that in a platform economy tend to be merged, becoming a hybrid computational capital; and the chapter has foregrounded three forms of resemblance between the tax regulatory problem and the regulatory problem of content moderation. First, non-state actors access, managem and regulate through platforms flows of capital and similarly flows of information exploiting regulatory differentials, so that there is the need of regulatory alignment in both cases. For example, the host state may be lenient on content moderation in comparison to the home state of the Global Actor owning the platform (or the other way around), thus there can be conflicting/overlapping policies.

Second, since both capital and information escape the regulatory reach of States, some form of common standard must be achieved in both cases. For example, a platform may escape the regulation on content moderation by both the host state and the home state of the Global Actor owning the platform, so that a common regulative approach is required so that it establishes at least a minimum standard of procedural accountability to minimize regulatory arbitrage.

Third, such common standards can be achieved only if individual home States of Global Actors owning platforms assume together the ethical and political obligation to moderate profit diversion as well as immoderate use of platform content through procedural accountability, in light of the fact that the alternative regulatory model would be to pursue an inclusive consensus on a very detailed agenda, a task that is quite hard in the context of content moderation and that is proving unsuccessful in the tax world.

This homology between capital (leading to “opportunistic” platform-generated profit diversion) and information (leading to “immoderate” platform content use) points to the fact that information extracted by platforms becomes a hybrid computational capital and therefore to the need of regulatory structures that reduce excessive opportunism and immoderation in the use of this computational capital by platforms.

This leads to a final conclusion about the regulation of computational capital: as we have assessed that there is an intractability problem of both profit diversion and content moderation when addressed exclusively from a territorial state perspective, in both cases multilateral minimum standards should be pursued. These standards should rest on a country-by-country assumption of responsibility by States to regulate Global Actors based in those countries, initially in a minilateral forum creating an international regime that does not necessarily require formal international agreements, to be possibly extended in a more inclusive format that may contemplate formal agreements.