Taxing data as an instrument of economic digital constitutionalism: elements for a normative agenda

Abstract

Digital constitutionalism rarely focuses on value creation, extraction, and distribution. This Article introduces a symposium that contributes to filling this gap, using data taxation as an entry point and sketching the elements of a normative agenda. The contributions advance different proposals, but they share the view that the externalities of informational capitalism have constitutional significance. Based on this, this introduction keeps four issues together: (1) the impact of excessive datafication on contemporary societies; (2) the role of data in contemporary economy; (3) concrete tax design; (4) the interaction of data taxation with other legal regimes and social justice issues, also at the global level. The first goal is to increase the dialogue among strands of legal scholarship that do not necessarily speak the same language. The second goal is to expand the analytical and normative scope of digital constitutionalism, which cannot address such issues as accidental elements but needs to be (also) an economic constitutionalism. The Article proceeds as follows. Section 2 focuses on the link between the digital revolution and constitutional states, especially on their role in value creation, extraction, and distribution. Section 3 identifies such an issue as a gap in digital constitutionalism and opens the way to the following sections. Section 4 is divided into four subsections. Section 4.A stresses the need for critical approaches to datafication, which needs to be seen as an autonomous object of regulation. Section 4.B highlights the role of data within contemporary economy and offers normative justifications for its taxation. Section 4.C highlights the need to include Pigouvian, progressive, and rent-targeting elements into data tax design. Section 4.D puts these issues within the context of economic governance, highlighting the role of (global) institutions in creating, extracting, and distributing value as well as the political nature of the underlying policy choices. Section 5 concludes.

1. Introduction

By and large, digital constitutionalism – roughly understood as the constellation of scholarly and policy discourses exploring the relationship between constitutionalism and the socio-legal challenges of the digital revolution¹ – has not focused on value creation, extraction, and distribution of data/informational capitalism.² To be sure, legal scholarship has long investigated the relationship between the digital revolution – with its underlying political economy and legal infrastructure – and social justice.³

However, authors rarely thematise these issues within explicitly constitutionalist frameworks.⁴ Scholars using such frameworks, on the other hand, mainly focus on questions such as access to the Internet; free speech and disinformation; privacy and data protection; procedural guarantees (transparency, participation, fairness);⁵ digital administration⁶ and justice;⁷ applicability of constitutional law standards to private actors via (some variation of) the ‘horizontal effect’ of fundamental rights.⁸ Moreover, most constitutional analyses take the regulatory capacity of modern states somehow for granted and do not account for tendencies towards the re-feudalisation of socio-political relationships deriving from the business models of data economy.⁹ Put differently, the impact of digital and algorithmic innovation on societal processes and on (the legitimation of) both national and non-national political institutions¹⁰ is hardly linked to the latter’s capacity to effectively pursue policies such as social justice and egalitarian objectives.¹¹ At an even deeper level, constitutional lawyers do not thematise the dangers coming from the excessive datafication of society¹² as a self-standing problem.

Against this background, this Article introduces a symposium that contributes to filling this gap, using data taxation as an entry point. More specifically, it sketches the elements of a research and policy agenda concerning data taxation within a (digital) constitutionalist framework. While the single contributions adopt distinct approaches and advance different proposals, they share the view that the externalities of informational capitalism are issues of constitutional significance.

Starting from this view, this Article and the symposium (try to) keep four macro-issues together: (1) the impact of excessive datafication on contemporary societies; (2) the role of data in contemporary economy and the justifications to its taxation; (3) concrete tax design, coherent with the regulatory purposes of a (digital) constitutionalist agenda; (4) the interaction of data taxation regimes with other legal regimes and social justice issues, also at the global level. The interlinked questions underlying these issues constitute the building blocks of an expanded digital constitutionalism, which includes the socioeconomic dimension not exceptionally but structurally.¹³

Put differently, this Article – and the entire symposium – takes on the issue of data taxation to explore the interconnection among such questions, both analytically and normatively.¹⁴ The first-order goal is to increase the dialogue among different strands of scholarship – constitutional law, law and technology studies, ‘law & political economy’ (LPE), and tax law, among others – that do not necessarily speak the same language. The second-order goal is to expand and strengthen the analytical and normative scope of digital constitutionalism.

These goals are based on a specific view of the normative purposes of digital constitutionalism: if the latter aims to rise at the level of complexity required by the challenges posed by the digital revolution and informational capitalism, it needs to address issues related to value creation, extraction, and distribution, also through interactions with legal fields that do not speak the language of constitutional law. If digital constitutionalism aspires to be authentically normative, it cannot address such issues as contingent, accidental elements.¹⁵ To be fully normative, digital constitutionalism needs to be (also) an economic constitutionalism.¹⁶

After this introduction, the Article proceeds as follows. Section 2 focuses on the link between the digital revolution and (the legitimacy of) constitutional states, notably when it comes to the latter’s role in value creation, extraction, and distribution. Section 3 identifies such an issue as a gap in digital constitutionalism and opens the way to the following sections. Section 4 is divided into four subsections. Section 4.A highlights that, to address the issues mentioned coherently, one has to start from a critical approach to datafication. This means that excessive datafication needs to be seen as a source of problems in itself, to be addressed as an autonomous object of regulation. Section 4.B highlights the role of data as an economic factor within the data economy, beyond its monetary value and the specific conceptualisations in different legal fields. Analysing the role of data in the contemporary economy, one can individuate normative justifications for its taxation, even in the light of constitutional limits to states’ taxing power. Section 4.C moves to tax design issues, highlighting the need to include Pigouvian, progressive, and rent-targeting elements. Finally, Section 4.D puts these issues within the context of economic governance, highlighting the role of (global) institutions in creating, extracting, and distributing economic value, and the political nature of the underlying policy choices. Section 5 concludes.

2. Böckenförde’s dilemma, governability and the (digital) constitutional state

‘The liberal secularised state lives by prerequisites which it cannot guarantee itself.’¹⁷ Böckenförde’s (in)famous line looks at any lawyer standing at the gates of constitutional theory. At its core, it is a sobering reminder of the insufficiency of modern (constitutional) law in generating the preconditions for its own normativity. When god(s) cannot be invoked anymore – not directly, at least – and nationalism has generated monsters, when justice and values mean something different to each social group, the legitimacy of political institutions in post-war constitutional states has come to rely on their capacity to adopt collectively binding decisions¹⁸ while simultaneously preserving room for conflicts to emerge and mediating them.¹⁹ Post-war liberal, constitutional states with capitalist modes of production are thus characterised by an intrinsically precarious balance which puts their own ‘governability’ into question.²⁰

Indeed, the capacity of political institutions – both national and supranational – to generate, extract, and distribute value, while also striking balances among societal actors, is crucial to their legitimation and, ultimately, to the performance of their societal functions.²¹ Effective, socially legitimated systems of value creation, extraction, and distribution are not only instruments in the hands of political apparatuses aimed at preserving social peace. At a deeper level, they are essential to the capacity of political institutions to adopt consensus-based, collectively binding decisions, that is, to the performance of their societal functions.²² As instruments of social justice, tax systems are crucial to both socioeconomic governance²³ and the legitimation of modern political authority.²⁴

Such elements are relevant to digital constitutionalism. The digital revolution, the platformisation of socioeconomic relations, and data- and automation-driven business models have profoundly impacted the political economy and the public sphere(s) underlying contemporary societies.²⁵ The COVID-19 pandemic has accelerated pre-existing dynamics, touching upon the relationship between authority and freedom, state and society, politics and economy, collective and individual actors.²⁶ The public sphere(s) where individuals and groups generate debate, contestation, and conflict have become extremely different from those presupposed at the foundational time of modern constitutionalism²⁷ and often go beyond the territorial borders of nation-states. These developments, in turn, take place alongside processes of political-economic globalisation, dispersion, and fragmentation triggered or dominated by neoliberal policies. In this context, the value of traditional procedures legitimising constitutional states – especially elections and other institutions of representative democracy – is more and more eroded.²⁸

These developments are further linked to the crisis of economic governance models presupposed by post-war constitutional states. In addition to the race to the bottom triggered by the competitive alignment of regulatory, fiscal, welfare, and labour protection systems,²⁹ and by so-called austerity policies, the business model of the data economy impacts socio-political integration by also affecting the capacity of value creation and distribution – that is, taxation and welfare systems – presupposed by modern constitutional states.³⁰

Models of value creation based on digital services and finance, personalised advertisement, big data analytics, cryptocurrencies, smart contracts, metaverse(s) – in one phrase, informational capitalism³¹ – accelerated and exacerbated pre-existing social dynamics of late-20^th-century economic globalisation, weakening the capacity of political institutions to govern social processes also through economic distribution. Following processes of globalisation, dispersion, and intangible-isation,³² the data economy makes traditional taxation systems – especially the income tax – much less effective³³ and aggravates the fiscal crisis and ‘base erosion’ that emerged even before the global neoliberal turn of the 1980s.³⁴ Informational capitalism does not influence only individual behaviour and how societies collectively reach (presumptive) consensus on specific issues and accept political authority as legitimate. It also affects political institutions’ capacity to govern the economy via collectively binding decisions and politically legitimated law-making. The externalities of profit-driven datafication affect the ability of politics to produce socially legitimised decisions;³⁵ of science to produce socially shared truth;³⁶ of medicine to improve collective health;³⁷ of economy to create value for the whole society.³⁸

3. A gap in digital constitutionalism

By now, there is a vast literature on how law ‘codes’ data, making it one of the factors of production within the data economy/informational capitalism³⁹ and contributing to ‘datafying’ other factors of production.⁴⁰ Social justice issues have been explored especially by labour law scholars, for example in the context of platform work,⁴¹ workplace democracy,⁴² or the impact of digital technologies on processes of value-creation and -distribution at the global level.⁴³ Tax law scholars, for their part, investigate how informational capitalism affects states’ fiscal capacities, potentially contributes to tax avoidance,⁴⁴ or how Artificial Intelligence (hereafter AI) and digital technologies may optimise tax systems, making them more efficient.⁴⁵

However, after a false start in the 1990s,⁴⁶ only recently have tax law scholars begun conceptualising data as autonomously taxable wealth.⁴⁷ More generally, authors do not thematise these challenges within constitutional-theoretical frameworks. However, the relationship between digital technologies and constitutional law goes beyond the protection of the integrity of the ‘free marketplace of ideas’, the guarantee of ‘digital rights’, or more broadly, the political public sphere and market competition. Instead, it touches upon the very sources of integration/legitimation of constitutional states as they emerged following the secularisation of modern societies.

Digital constitutionalism, for its part, is by now a relatively settled strand of constitutional scholarship.⁴⁸ Recent contributions have highlighted how digital constitutionalism is intrinsically characterised by different perspectives and, in this sense, represents a discursive field whose elements are both complementary and contradictory.⁴⁹ However, if it aims at overcoming some limits of the liberal political theory underlying modern constitutional law and avoiding overlooking issues of societal power, a normative digital constitutionalism is called to frame excessive datafication and social justice as parts of a single project concerning the relationship between the digital revolution, the economic system, and constitutional law.

In both analytical and normative terms, social justice and value distribution are pieces of a broader puzzle within any digital constitutionalism aiming to stand as a counter-power to different kinds of authoritarianism. In this sense, digital constitutionalism needs to be an economic constitutionalism.⁵⁰ Looking at the trajectory of modern constitutional state, this means considering and revitalising the normative elements of economic and industrial democracy,⁵¹ currently largely missing in digital constitutionalism discourses. Indeed, while traditionally overlooked or abandoned by constitutional theorists of the past century, such elements have historically played a significant role in the reality of 19^th-century constitutional states.

Taking into consideration the different perspectives shaping its identity, digital constitutionalism is called to address at least four issues as part of a comprehensive, coherent research and policy agenda: (1) the negative impact of excessive datafication on contemporary societies (critical approach to datafication); (2) the legal conceptualisation of data for the extraction and distribution of its value (data as an economic factor); (3) the design of data taxes (tax design); (4) the interaction of data taxation with other legal regimes and issues of social justice at both national and global levels (economic digital constitutionalism).

4. From a critical approach to datafication to economic digital constitutionalism

A. Critical approach to datafication, regulatory goals, and data taxation

Any digital constitutionalism that takes the ‘datafication question’ seriously requires a critical approach and an awareness of the externalities deriving from datafication itself. Such an approach⁵² is necessary to expand the regulatory goals of data governance and entails several consequences.

First, it forces new problems to appear or reframes existing ones.⁵³ ‘Big data’ and digital technologies are not only a means for received models of governance of populations and subjects. The amassing, analysis, and mobilisation of hybrid data repositories and real-time data flows – primarily driven by the profit-maximisation compulsion of (informational) capitalism⁵⁴ and the power-maximisation compulsion of political systems⁵⁵ – open to new and potentially dangerous forms of governance, a new ‘digital political economy’.⁵⁶ The combination and mutual reinforcement of private/commercial and public/political surveillance, powered by digital and algorithmic technologies, end up affecting or manipulating individuals, groups, and social systems alike.⁵⁷

In Habermasian language, the excessive datafication of society narrows the ‘life-world’ spaces within societies.⁵⁸ At the same time, it allows for their colonisation/corruption by rationalities – economic, political, but also scientific, medical, mass-mediatic, etc. – whose inner expansive tendencies are less and less constrained.⁵⁹ Datafication is not a problem ‘only’ to the extent it reinforces the power- and profit-accumulation imperatives of political and economic actors. It re-frames, constrains, and potentially corrupts the social processes which allow for the protection and free development of individuals, collective groups, and functional systems within society.⁶⁰

Second, a critical approach highlights other – actual or potential – harms, eg, the energy consumption and environmental degradation linked to the data economy;⁶¹ the exploitation and invisibilisation of old and new forms of labour;⁶² and the socioeconomic costs deriving from data- and automation-driven business models.⁶³ Such issues are hardly taken into consideration in data governance strategies. The latter – typically focusing on harms to individuals and issues related to data protection, hate speech, discrimination, misinformation,⁶⁴ market competition,⁶⁵ and, more recently, (urban) safety and sustainability – do not include the reduction of the compulsion to datafication among their goals. Instead, they aim to increase the availability, quality, sharing of data⁶⁶ or, at best, the participation of data subjects in its control/management.⁶⁷ Despite its potential,⁶⁸ the principle of data minimization – one of the cornerstones of data protection law⁶⁹ – remains underenforced⁷⁰ and, in any case, limited to the relatively narrow realm of personal data.⁷¹ What was designed as one of its primary safeguards – the possibility to refuse or withdraw individual consent to data treatment⁷² – has long shown its limits, especially in the age of so-called digital resignation.⁷³ Moreover, AI systems and intensive data processing put into question the very dualism between personal data/non-personal data on which data protection law has been built.⁷⁴

More generally, by creating ‘data markets’ controlled by data intermediaries;⁷⁵ by circumventing or softening fundamental right guarantees concerning mass surveillance, often in the name of child protection;⁷⁶ by encouraging ‘smart’ urban planning;⁷⁷ and by relying on the mirage of data anonymisation,⁷⁸ more recent regulatory instruments incentivise data (over)production, commercial surveillance, and stabilisation of dominant market positions of ‘data producers’.⁷⁹

The challenges underlying data governance do not derive only from the role of business actors in highly concentrated markets, actors that can often escape ‘hard’ regulation and create their own regulatory standards.⁸⁰ They also come from the fact that regulatory approaches based on data as economic good are ‘hardwired to only produce governance strategies that will facilitate the provision of more or better-quality data’,⁸¹ thus overlooking other societal goals beyond data provision. Even risk-based regulatory approaches⁸² and ‘hard’ prohibitions of more recent instruments⁸³ – pursuing goals such as protection of the democratic process, innovation, privacy, or minors’ wellbeing – have only limited effect if the internal incentives to excessive datafication are unaffected.

The reduction of surveillance-based, profit-driven, attention-maximising datafication can be pursued without foreclosing the prospects for authentic inclusion, participation,⁸⁴ solidarity, and emancipation brought by digital innovation, as well as the economic opportunities brought by safe digital technology.⁸⁵ Certainly, datafication of contemporary societies, on the one hand, and the emergence of informational capitalism, on the other, have been deeply entangled. But such entanglement remains contingent: technological progress does not necessarily need to consolidate the institutions and actors of (informational) capitalism and may be directed towards authentic socio-political change, whatever the latter might be. A digital constitutionalism that (re-)asks the ‘social question’ is not about fighting rear-guard battles. Rather, it is about reducing the incentives to the excessive datafication of society and the closely related pressure to societal manipulation/colonisation, which, under current modes of production, is in turn linked to other perverse dynamics of global neoliberal capitalism (especially the free flow of capital and the financialisation of the economy).⁸⁶ Here, tax policies represent an overlooked instrument.⁸⁷

Historically, taxes – especially progressive ones⁸⁸ – have served multiple goals, well beyond the ‘mere’ funding of governments.⁸⁹ Taxes may be imposed to reduce risks associated with lawful but potentially harmful activities, to tackle the societal power coming from the accumulation of profit by economic actors, to enhance social mobility, inclusion and purchasing power of low- and middle-income population.

However, the normative justifications of proposals for ‘digital services taxes’ – in the EU or elsewhere – normally focus on the need to target new forms of intangible wealth that escape tax systems (so-called base erosion) but rarely mention disincentivising excessive datafication as such.⁹⁰ Once again, this is a regulatory blind spot, showing a persisting market-driven imprint.⁹¹ As long as the contrast to excessive datafication does not become a goal as such, any regulatory strategy is doomed to overlook the related risks or, worse, contribute to their invisibility.

Against this background, tax law is only one instrument of a necessarily diverse regulatory mix that also needs strategically to include both state law – notably corporate, labour, administrative, tort law⁹² – and non-state normative systems (international law,⁹³ but also social and private norms of different kind⁹⁴). As long as it makes the data economy – which, as we will see below, is primarily a rentier economy⁹⁵ – less profitable, reduces the compulsion to excessive datafication, and slows down automation,⁹⁶ taxation is a useful and, so far, relatively underexplored tool to tackle the negative externalities of data capitalism.⁹⁷

To summarise, a critical approach helps identify datafication as a self-standing problem and orient the related regulatory strategies. Here, one needs to consider the potential contradiction between conflicting regulatory goals of data taxation: budget-funding/distribution of value, on the one hand; reducing/disincentivising excessive datafication, on the other. Indeed, any data tax effectively reducing the datafication compulsion might shrink the tax base⁹⁸ of data companies and therefore the targeted revenue.⁹⁹ Certainly, which goal should prevail is a policy issue that needs to be addressed through politically legitimated decision-making processes. However, thematising the reduction of datafication as a regulatory goal is worth in itself, as it makes the associated dangers more visible – it forces them to appear as a problem, opens to more informed debates, and re-politicises data governance.¹⁰⁰ Further, the automatic link between data taxes effectively reducing datafication and the shrinking of the tax base is not to be taken for granted and is so far supported by little empirical evidence. There is indeed the possibility – and initial evidence¹⁰¹ – that raising the marginal costs of data collection might push companies to pass the burden to users by charging them for using digital services that are currently ‘free’. This dynamic could bring people to use less of it or, more realistically, re-expand the base of traditional income tax – the one most affected by the ‘base erosion’ linked to the data economy¹⁰² – thus potentially increasing tax revenues.¹⁰³

Be as it may, before elaborating concrete tax designs, scholars and policymakers need to explore understandings of data providing normative justifications to data taxation also in light of constitutional limits related to taxpayers’ ability to pay. In other words, any effective data taxation policy needs to frame data as a factor contributing to data-collectors’ tax base.¹⁰⁴ In this way, a critical approach to datafication represents the point of departure towards topics investigated by LPE scholarship and institutional economics.¹⁰⁵

B. Data as a legitimate tax base within a rentier economy

Today, it is almost a cliché to observe that data is the new oil¹⁰⁶ or, as the EU Commission put it, ‘the lifeblood of economic development’.¹⁰⁷ Besides more or less questionable metaphors, data has undoubtedly become a major economic factor in contemporary economies. Profits of business actors – both dominant and non-dominant – heavily rely on their capacity to access, extract, process, and monetise data.¹⁰⁸ Some scholars even argue that data might progressively replace money or at least some of its functions.¹⁰⁹

From a legal point of view, data is not a ‘thing’, as it is normally defined as a ‘digital representation of information’.¹¹⁰ Law ‘codes’ data as a ‘representation’, constantly replicated for a potentially infinite number of times. The term ‘data’ clusters a constellation of activities, notably inputting, processing, organisation, abstraction, units, aggregation, and resourcing of information through digital technologies.¹¹¹ But the law does not only define data. It also contributes to making it an economic factor.

Taking a step back, one can observe that, from an economic point of view, data becomes a valuable good through the aggregation and analysis of pre-existing information in huge amounts, via technologies created by software engineers and data scientists.¹¹² This constellation of activities, however, is difficult to evaluate economically, as it is contingent on specific technoscientific, legal, and institutional ecosystems.¹¹³

Indeed, legal scholars have proposed numerous conceptualisations which, according to different policy goals,¹¹⁴ variably frame data as object of property rights, privacy rights, or as commons.¹¹⁵ However, it is crucial to stress the role of law in shaping the capacity of business actors to produce, commodify, and monetise data. The law ‘codes’ data as capital by making priority of use, durability, convertibility, and universality possible.¹¹⁶ Law – or rather, a combination of (intellectual) property, contract, collateral, trust, corporate, bankruptcy, procurement law¹¹⁷ – contributes to this result in three ways.

First, the law keeps the ‘raw’ information potentially representable or processable via digital technologies a free, abundant resource, open to use, ‘mining’, even ‘scraping’.¹¹⁸ Large networks of users ‘hooked’¹¹⁹ to digital platforms and vast amounts of publicly available, ‘free’ information build the datasets necessary for the business model of data companies,¹²⁰ both those providing social media services and those focusing on AI.¹²¹ Information potentially representable and/or processable via digital technologies is treated as res nullius rather than res communis.¹²² But data is an economically relevant factor only above certain thresholds that can be reached only by (few) governments¹²³ and companies through their control of digital infrastructures.¹²⁴ In this sense, the data economy is an economy of scale.¹²⁵

Second, the law protects and stabilizes – notably via proprietary technologies¹²⁶ – the unequal control over digital and computational infrastructures¹²⁷ and, therefore, business actors’ capacity to appropriate and process information, excluding others from the use of data,¹²⁸ primarily through trade secrecy protections.¹²⁹ In theory, data could be an (impure) public, universally accessible good only with the combined presence of some elements: no or minimal constraints to computing and processing power; digital infrastructure built with interoperable standards; open-source (or unlicensed and in the public domain) software.¹³⁰ Absent such elements, data becomes de facto privatised.

Third, the law ensures the possibility of exchanging data for some direct or indirect advantage: direct monetisation, improvement/optimisation of one company’s own services, and influence on the social and political environment.¹³¹

To sum up, data’s economic value varies depending on who uses it and how it is being used.¹³² One feature – possibly the feature – of informational capitalism is that data is an economic good but its contribution to taxpayers’ wealth oscillates and is extremely difficult to pre-determine in monetary terms.¹³³ This makes it difficult to individuate its role in data-producers’ ability to pay¹³⁴ and therefore a pre-determined tax base to target.¹³⁵ Also for this reason, data does not appear in corporate balance sheets and legal systems do not ‘code’ it as an economic good that can be taxed as such. This could be seen as an obstacle to taxing data in the light of the constitutional guarantees requiring taxes to be limited by taxpayers’ ability to pay and tied to some pre-determined tax base.¹³⁶

However, this obstacle also derives from the almost exclusive focus on the exchange value of data, that is, the possibility to convert it into other economic goods, notably money. Relatedly, in most advanced economies, the bulk of taxes presently falls on production, income, and consumption rather than rents. Therefore, tax policies are normally evaluated through the relatively narrow lens of the income tax as a default taxation model.

In contrast, recent scholarship has highlighted the use value of data, that is, the ‘value of being able to infer or predict likely future actions or effects’.¹³⁷ The use value of data, in other words, consists in exponentially enhancing data companies’ predictive power.¹³⁸ Predictive power and algorithmic control over user attention and digital infrastructures enable data companies to extract a rent ¹³⁹ from the actors in their ecosystems (users, suppliers, and advertisers).¹⁴⁰ Even without an immediate monetary benefit – and often engaging in non-profitable or loss-taking investments¹⁴¹ – acquiring predictive power provides structural economic advantages, further incentivising datafication.¹⁴² The data economy, in other words, is primarily – although not exclusively¹⁴³ – a rentier economy.¹⁴⁴

The individuation of data and the predictive power it brings as an autonomously relevant economic factor – more specifically, as a source of rent – offers a normative justification for data taxation even in the light of constitutional norms imposing the pre-determination of a tax base as a limit to governments’ taxation power.¹⁴⁵

Further justifications may be figured out by looking at the process of data production. As seen above, this process involves the ‘mining’ of publicly available information and its subsequent transformation into effectively enclosed data. In this process, the intellectual labour and the technology made possible by data science are crucial.¹⁴⁶ Based on these observations and the scholarship supporting the existence of a right to science in international human rights law,¹⁴⁷ one could resort to the following argument: to the extent it dispossesses different types of workers of (the value of) their labour, without necessarily compensating such loss with a commensurate increase of aggregated demand, investment, and job gains;¹⁴⁸ and to the extent it is based on a common good – science – whose benefits are the object of an international right, profit-driven datafication provides data companies with a rent.¹⁴⁹

Put differently, data contributes to the tax base – and taxing data can be considered legitimate under traditional constitutional guarantees – not only because, above certain thresholds, it gives companies a crucial economic advantage through the use value/predictive power but also because it structurally deprives other actors – notably workers, households, economic actors in subaltern market position, and humankind more generally – of economic opportunities and benefits to which they are also legally entitled. Either way, the data economy emerges as a typical rentier economy, encouraging further rent-seeking or ‘conspicuous consumption’ rather than productive investment.¹⁵⁰

These considerations relate to the objection that, because of its elusive monetary/exchange value, data does not per se contribute to the data collector’s ability to pay and cannot be used to determine a tax base. To be sure, several issues are left unaddressed, although, as we will see below,¹⁵¹ such considerations have an impact on tax design.

Here, one needs to stress that grounding data taxation on data’s use value is compatible with different (legal) notions of data. In other words, as far as taxation goes, qualifying data as an economic factor is compatible with different legal qualifications (personal and non-personal, sensitive and non-sensitive, ‘simple’ data and metadata, ‘synthetic’ and ‘authentic’, etc.) and regimes (ownership, commons, etc.) inspired by distinct regulatory goals (protection of privacy and individual self-determination, market competition, well-functioning political sphere, intellectual property, etc).

C. A constitutionalist data taxation: Pigouvian, progressive, rent-targeting

As seen above, the regulatory goals of a ‘constitutionalist’ data taxation (may) include the reduction of excessive datafication, as well as issues of social justice, distribution, and legitimation of constitutional states. The ultimate choices concerning tax design are political and should be open to debate, struggle, conflict, and collective decision-making. Digital constitutionalism does not and should not offer a pre-made, ready-to-go, depoliticised policy agenda, be it economic or not.¹⁵² Its normative nature, if any, does not consist in providing a pre-set project of a ‘good society’ but rather in highlighting the crucial role of some aspects in policy fields where ultimate decisions will still be based on the interaction and conflict among politically engaged, strategic societal actors. In this sense, what a ‘good’ digital revolution is – or will be – can only result from this ‘blind’ collective dynamic. However, the arguments concerning the ability-to-pay objection are not neutral in terms of tax design.

Firstly, and more generally, any tax design coherent with the regulatory goals of digital constitutionalism needs to incorporate Pigouvian elements. This means that data taxes should target data companies and other actors of the data economy as actors who produce externalities and be designed to change their behaviour.¹⁵³

Secondly, data taxation informed by digital constitutionalism includes elements of progressivity.¹⁵⁴ In data capitalism, any form of data taxation has structurally progressive effects on the overall tax system, regardless of whether the specific tax imposition is designed as actually progressive or not, with increasingly higher rates¹⁵⁵. While ‘data-rich’ taxpayers – basically, big data companies – easily avoid traditional income taxes,¹⁵⁶ low- and middle-income (‘data-poor’) taxpayers (eg, the individual Internet user) cannot. Therefore, via an exemption threshold calculated based on the volume of data traffic,¹⁵⁷ a fiscal imposition targeting (the collection and transmission of) data could easily capture only high-volume users that otherwise avoid income tax. Creating a ‘no-data tax’ area below specific data traffic volumes is not only sound in policy terms.¹⁵⁸ It is also a way to make data taxation compatible with relevant constitutional principles, given the specific features of the data economy. If the ability-to-pay principle can be respected only based on the use value/predictive power which comes from the amassing of huge amounts of data – typically those reached by data companies in an economy of scale – then a ‘no-data tax’ area for small-time data producers is a constitutional requisite for any legitimate form of data taxation.¹⁵⁹

Besides that, specific fiscal impositions could easily be made progressive. Once certain thresholds are crossed, the marginal tax rate could be designed as a logarithmic function to streamline data tax and prevent cliff effects.¹⁶⁰ The more data is produced, the (gradually) higher the tax rates and brackets would be. Once again, making data taxation progressive is not related only to the options concerning the distribution of the value extracted by oligopolistic economic actors. It also relates to underlying policy goals deriving from a digital constitutionalist agenda aware of the effects of excessive datafication as such and, at the same time, of the potential for inclusive economic growth coming from digital innovation.¹⁶¹ Put otherwise, disincentivising excessive datafication means designing data taxes not as ‘flat taxes’ but rather as progressive ones.

Thirdly, a constitutionalist purview provides reasons to design data taxes as direct taxes targeting rent-providing positions.¹⁶² For example, data taxes could be designed as royalties rather than digital services taxes (DSTs).¹⁶³ Royalties are fiscal impositions related to the extraction of (often) public resources. Those wishing to extract the resource must pay a ‘concession’ to the public authority for the right to do so. While not a tax in the technical sense, a royalty is commonly used by governments to raise revenue from extractive industries. Designing data taxes as royalties fits the reality of the data economy – as described above – taking into consideration the critical role of governments in developing digital technologies and infrastructure¹⁶⁴ as well as their continuing control over most computational and data infrastructures. Besides that, designing a data tax as a royalty has further advantages.

First, royalties do not require the activity to be productive and therefore are well suited to target sources of wealth that provide only potential rents.¹⁶⁵ As they do not target income deriving from data but data use as such, they bypass the problem of the delayed income realisation typical of the data economy.¹⁶⁶ Second, contrary to excises,¹⁶⁷ royalties can easily be made progressive – thus increasing based on some production measurement (eg, data traffic volumes) – and the related burden cannot easily be transferred to consumers. Third, contrary to DSTs, legal bases for adopting royalties can be found in (an evolutive reading of) existing international tax and trade treaties, which allow withholding taxes at source such as are applied to royalties. Therefore, they might forestall retaliatory measures, notably by the US, and reduce the need for international cooperation.¹⁶⁸

Even the criticism normally raised against royalties in the industries where they are most common – that they deter investment – does not necessarily apply to the data economy. Progressive royalties on data would only deter a specific business model. Moreover, the political economy surrounding data ensures that the ‘raw’ information potentially representable or processable via digital technologies, contrary to oil, remains a free, abundant resource, open to use and ‘mining’ even by non-profit actors. As highlighted above, the revenues of a few oligopolistic data companies are more akin to a rent and hardly create new value as such.

Fourthly and finally, a digital constitutionalist approach does not only aim at reducing excessive datafication and distributing wealth but also at countering the ‘governability’ and legitimacy problems of contemporary constitutional states accelerated by the digital revolution.¹⁶⁹ To that purpose, the related economic benefits should not necessarily be directed to individuals whose information was used to accumulate data (so-called data dividend taxes).¹⁷⁰ Indeed, data dividends feature some of the same issues as income taxation, that is, the problem of assessing the value of data collected from a specific territory, with the valuation and localisation problems coming with it. Further, the data economy is not based only on personal data but also – and, after the AI revolution, mainly – on non-personal, synthetic, and meta-data. Most importantly, data dividends de-responsibilise governments and de-politicise data governance.¹⁷¹

Put otherwise, deciding how to specifically distribute the value produced in the data economy through the institutions of the political system potentially re-legitimises constitutional states and strengthens their role within (global) economic governance. Contrary to data dividends, direct taxes and royalties on data give governments the opportunity to regain legitimacy precisely because they do not take away from them the burden of adopting collectively binding decisions over different distributive options.

However, such decisions must consider that the data economy is intrinsically transnational. An authentically normative digital constitutionalism needs to thematise distributive issues through the lens of states’ legitimacy and within the framework of global (economic) governance.

D. Economic digital constitutionalism within global governance

In the globally dispersed data economy, location of income and ownership of data mean little to tax purposes. Relatedly, the externalities of datafication are also highly dispersed. Decisions concerning data governance, the contrast to excessive datafication, and distributive justice need to take these elements into account,¹⁷² also from the perspective of the North/South divide. Such issues have already been debated for a long time¹⁷³ and go well beyond the scope of this Article. This section highlights their analytical and normative interconnectedness from the perspective of digital constitutionalism, especially if the latter aims to hold onto its ‘global’ nature¹⁷⁴ and does not want to reinforce hegemonic positions within world society.¹⁷⁵

Taking a step back to tax design, one can observe that data taxes designed as excises, direct taxes, or royalties can be purely domestic. They do not target income and, therefore, bypass the problems of income realisation/localisation with which tax cooperation initiatives have been struggling¹⁷⁶ for over a decade.¹⁷⁷ For the same reasons, such taxes are less likely to trigger trade conflicts.¹⁷⁸

This notwithstanding, global tax cooperation still needs to be thematised within a digital constitutionalist framework, especially in light of trends towards the liberalisation of data trade in international trade law¹⁷⁹ and the persisting unequal control over digital infrastructures.¹⁸⁰ Overlooking such elements by designing purely domestic taxes targeting data traffic individuated through digital infrastructure may be convenient to Global North countries and the main competitor of the US for global hegemony – China.¹⁸¹ However, data taxation policies overlooking issues of international cooperation and the North/South divide risk becoming a redistributive endeavour involving already-hegemonic actors that does not tackle and might actually reinforce digital colonialism.¹⁸²

At the same time, when it comes to distributive obligations with respect to taxing rights, it is crucial to differentiate between the collective and the individual level, as ‘the reallocation of taxing rights from state to state does not necessarily help when it comes to fulfilling duties of justice towards individuals’.¹⁸³ Put otherwise, the normative goals of an economic digital constitutionalism are not necessarily met by a value distribution from state to state but require a consideration of the actual political decisions over the distributive choices towards both individuals and communities.

Such issues are linked to the role of global institutions, notably international organisations and transnational networks.¹⁸⁴ This consideration opens to issues of institutional design and management of conflicts across different governance regimes.¹⁸⁵ Put differently, an effective constitutionalisation of the digital sphere can – has to – involve institutions that compensate for the absence of a world state (and tax authority)¹⁸⁶ by collectively fulfilling the functions performed by political systems at the global/transnational level.¹⁸⁷

As recalled, institutions such as the OECD have become incubators of global tax cooperation – especially in the field of digital economy – and should be considered at least as important as those traditionally at the forefront of digital constitutionalism’s reflections (eg, the ICANN). Whether such efforts will be successful/effective or not is an open question. At the moment, the multilateral anti-BEPS initiative¹⁸⁸ seems to have stalled, potentially giving rise to new trade conflicts.¹⁸⁹ However, such fora allow for the thematisation and, potentially, contestation of key issues of economic global governance.¹⁹⁰ From this perspective, the recent emergence of a competing initiative within global tax governance¹⁹¹ might be seen as a positive development not only from an experimentalist perspective but also because it contributes to the further politicisation of such regime. The purposes of re-legitimation of political institutions and global justice impose a political reflection on how to direct the related revenue.

Indeed, besides the more general goal of tackling the fiscal crisis of states,¹⁹² it is crucial to strengthen the social fields – eg, medicine, information/press, science – most affected by the commodification trends and the reduction of aggregated demand deriving from the business models currently dominating the data economy.¹⁹³ If digital constitutionalism wants to live up to the aspirations of any authentic constitutionalism and constrain the expansive tendencies of economic systems ‘freed’ by the digital revolution, it has to embed forms of social protection,¹⁹⁴ especially for the processes instrumental to political participation,¹⁹⁵ socioeconomic inclusion, and generation of socially shared consensus, knowledge, truth. Moreover, focusing on the protection of social processes as such might help address two admittedly difficult regulatory questions: first, the choice between the collective and the individual levels of revenue reallocation/distribution; second, the individuation of the point where datafication becomes ‘excessive’.¹⁹⁶

In this regard, scholars have formulated proposals concerning, for example, funding some form of universal basic income (UBI).¹⁹⁷ Other contributions explore the possibility of using the data tax revenue to fund open medical research, supranational organisations such as the EU, especially considering their role in tackling other pressing global governance issues such as the climate emergency.¹⁹⁸ But many other solutions and proposals might be advanced, based for example on the environmental, social, and governance (ESG) agenda promoted by the United Nations Environment Programme Finance Initiative;¹⁹⁹ or – even more interesting and disruptive – the Program of Action on the Construction of a New International Economic Order elaborated by the Progressive International,²⁰⁰ published on the occasion of the 50^th anniversary of the UN General Assembly resolutions adopted in 1974 that aimed at establishing a New International Economic Order (NIEO).

Before concluding this section, one needs to address another question: Why ‘economic digital constitutionalism’ instead of the less conceptually demanding ‘digital constitutionalism for the economy’ or ‘constitutionalism for the digital economy’? And why is taxation important?

By using that phrase, the purpose is to link the analytical and normative frameworks of digital constitutionalism to the older conceptual tradition of the economic constitution, which, in its different understandings, goes beyond the ‘mere’ constitutional governance or regulation of the economy.²⁰¹ Importantly, that phrase brings back an entire set of normative elements related to forms of democratic legitimacy and decision-making that, instead of exclusively focusing on the traditional schemes of representative democracy turning around formal political institutions and procedures, take the condition of subaltern groups – especially within economic relationships – as the starting point to build any effective form of constitutional legitimacy and normativity even in the digital age.²⁰²

Relatedly, arguing that data taxation is a crucial instrument of a properly normative and therefore economic digital constitutionalism is meant to highlight two elements. First, data taxation aims – should aim – at constraining the expansive/colonising tendencies of both economic and political systems ‘freed’ by datafication processes. Second, value distribution through taxation may ultimately (re)legitimise the institutions of the political system, ‘exhausted’ by the persisting necessity to produce collectively binding decisions even when social consensus declines or is absent.

In other words, and without necessarily subscribing to the famous dictum of the US Supreme Court ‘the power to tax involves the power to destroy’,²⁰³ data taxation could be a crucial element to re-assert the centrality of political institutions in global governance, especially in the light of the historically genetic link between taxation, democracy, (popular) sovereignty, and modern constitutionalism.²⁰⁴

Along the same lines, advancing tax-based policies and proposals does not necessarily mean validating or naturalising governance approaches treating systemic problems in global capitalism as externalities and the results of failures of otherwise ‘perfect’ systems. The proposals mentioned above are rough and need to be discussed, analysed, and tested, also based on empirical elements provided by other disciplines. In no way are they supposed to be exhaustive, and, most importantly, they do not exclude regulatory measures of different kinds. In particular, they do not rule more radical options out, for example those questioning the current governance of digital infrastructures, one of the key structures underlying the data economy itself.²⁰⁵

5. Conclusion: a call for normative and institutional imagination

An authentically normative digital constitutionalism may point at some common principles²⁰⁶ and regulatory directions but, as such, is not and should not be a post-political discourse offering a pre-made agenda.²⁰⁷ Neither constitutional law (theory) nor law in general have the communicative/symbolic potential to solve, by themselves, the material challenges of the digital revolution and global governance. However, recognising the role of the law in constituting the institutions of (digital) capitalism – like the recent LPE scholarship does – is a fundamental step to address the related challenges and contribute to their transformation.²⁰⁸ At the same time, the transformative possibilities of the law should not be overestimated. This holds particularly true when it comes to more radical forms of contestation of global economic (re)production.

Digital constitutionalism can and must provide a discursive field to thematise constitutionally significant economic issues: a ground for conflict and collisions to arise and produce constituting and limiting norms.²⁰⁹ Opening the gates to conflict – forcing the problems to appear as problems – strengthens the analytical and normative potential of digital constitutionalism. It also establishes fruitful conversations with other political–economic discourses and practices concerning the material conditions of the digital revolution at the global level, which, in turn, might be enriched by more direct interactions with explicitly constitutionalist frameworks.

To be sure, the cluster of informational capitalism,²¹⁰ ‘technofeudalism’,²¹¹ data colonialism,²¹² and different forms of both public and private digital authoritarianism²¹³ will hardly be overcome only through litigation before (constitutional) courts. Nevertheless, collectively rethinking the meaning of ‘(un)constitutional’ in the digital era is a necessary step towards radical transformation. Digital constitutionalism can emerge as a counter-discourse and be authentically normative if and to the extent it (1) breaks the conflation between a ‘public sector’ driven by digitalised power and a ‘private sector’ driven by digitalised profit;²¹⁴ and (2) contributes to the re-legitimisation of both national and international political institutions, reducing their tendency to decomplexify their social environments through techno-authoritarian solutions.²¹⁵

In both cases, data taxation should be considered as a piece of a complex puzzle whose contours (should) remain open to institutional and legal imagination.²¹⁶