Data-owning democracy: Citizen empowerment through data ownership

Abstract

This article extends property-owning democracy to the digital realm and introduces “data-owning democracy,” a new political economic regime characterized by the wide distribution of data as capital among citizens. Drawing on republican theory and acknowledging data's unique role in the digital economy, it proposes a two-tier model that combines different modes of data ownership and corresponding rights. The first layer of “data-owning democracy” is characterized by a digital public infrastructure that enables citizens to collectively generate data and have a say in how their citizen data are used. In the second layer, individuals automatically receive machine-readable copies of their data whenever they are generated—a slightly more advanced form of the European Union's existing right to data portability (Art. 20). With its focus on empowerment, data-owning democracy is designed to be complementary to existing data protection regulations. It also illustrates how political theory more broadly, and republican theory specifically, can be instructive for specifying the normative components of a new political economy dealing with questions of empowerment and digital rights.

Introduction

Today, most of us are simultaneously data users and data producers. Whenever we look up the nearest coffee shop, collect some loyalty points, or pay with our credit card online, we leave a small data trail. Individually, these data points do not possess great value. Aggregated and analyzed, however, they become worth a fortune. In the Western hemisphere, this is reflected in the ever-rising value of data-driven companies such as Alphabet, Meta, or Amazon. Their scope and exact business models may vary. But what they share is a focus on extracting, analyzing, and using information about their customers to improve and expand their services. Unsurprisingly, three out of the five richest men in the United States owe their wealth in large numbers to data-driven companies and technologies (ATF and IPS, 2021; Collins, 2021). As our digital and physical worlds become increasingly intertwined, our human experiences are the newest addition to the data pool—be that datafied information about our habits, emotions, or deepest fears (Zuboff, 2019; Cohen, 2019).

Against this backdrop, calls for a better protection of individuals in the digital economy ¹ have become increasingly loud. A central aspect of such efforts are data protection regulations, such as the European Union's General Data Protection Regulation (EU 2016/679) or California's Consumer Privacy Act (CCPA). Under the liberal paradigm that continues to inform both law and economic thought (Benthall and Goldenfein, 2020), such regulations tend to operate from an understanding of freedom as the absence of unwanted interference, or more broadly, harm. Acknowledging data protection's essential role in moving us toward an equitable digital economy, this article investigates whether these efforts could benefit from complementary initiatives that go beyond questions of protection. More specifically, it discusses how personal data flows could be integrated into a political economy modeled around citizen empowerment.

The argument developed in this paper is that property-owning democracy is a useful concept for empowering citizens online and that it can benefit greatly from a republican perspective. So far, works on digital empowerment have largely been built on liberal and egalitarian grounds. Conversely, this paper shows that republican theory more broadly, and its focus on non-domination specifically, have much to contribute to the debate.

The paper is structured as follows. Section two argues that personal data qualify as capital in today's digital economy and discusses the different roles they can assume in it. ² Section three introduces the concept of property-owning democracy and extends it to the digital economy. Section four specifies data-owning democracy as a two-tier model organized around three core values: efficiency, political equality, and freedom as non-domination. In the first layer, citizens collectively generate public data and correspondingly enjoy secondary rights over them. In the second layer, individuals automatically receive copies of their data whenever they are generated, which is a slightly more advanced form of the European Union's existing right to data portability (Art. 20). Importantly, the argument presented here does not deliver a blueprint of institutional specifics. Rather, its aim is to flesh out data-owning democracy's normative underpinnings, relevant properties, and prospective structure. To increase its overall feasibility, it draws on real-world examples, such as project Decode in Barcelona. Section five discusses how the proposed structure might promote efficiency, political equality, and non-domination. Section six discusses open questions and the most pressing challenges that need to be addressed.

In linking questions of data ownership with concerns of community wealth building, political economy, and power asymmetries, this article contributes to an emergent strand of works that recognize data as a central aspect in strategies toward an equitable digital political economy (e.g. Guinan and O’Neill, 2020; Sadowski et al., 2021). For example, Hafen et al. (2014) have focused on the empowering quality of health data cooperatives. Cheneval (2021) has argued for individual property rights of personal data to promote justice and endow individuals with an additional income from their data. Others have discussed personal data platform cooperatives (PDPCs) as vehicles to promote fair equality of opportunity (Loi et al., 2020), or have proposed taxing corporations in ways that acknowledge the public's vital role in generating data streams (Feygin et al., 2019).

Discussions about new forms of data ownership rightly raise the question whether such efforts would further entrench the commodification of personal data flows in Western society, and whether that would be a desirable path to embark on. It is true that the regime proposed here does not question the commodification of people's data flows per se. Nevertheless, data-owning democracy as a new political economic regime does not further entrench or encourage said commodification processes. Rather, the question it addresses is: How, absent more substantive changes in the overall structure of the data economy, can we ensure that current data commodification processes benefit as many people as possible?

Therefore, while more substantive changes in the digital political economy remain urgent, necessary, and desirable, the rationale here is that while personal data flows continue to be commodified, they might as well be used in the most empowering way possible. That said, this article's focus on empowerment should not be understood as an alternative approach to regulatory or on-going data protection efforts. Rather, its purpose is complementary, and crucially depends on their success.

Personal data as capital in the digital economy

In a report titled “data is the new capital,” a prominent consultancy firm argues that there will be “a day in the not too distant future when data will be an asset reported and accounted for in a company's balance sheet and reflected in its market cap” (Jain, 2020). Whether that prophecy holds or not, data have already been recognized as a valuable asset in the digital economy and beyond (WEF, 2011). With an increasing part of social activities shifting online, many aspects of the world are starting to become datafied (Van Dijck et al., 2018: 33). This “datafication” process (Mayer-Schönberger and Cukier, 2013) refers to the ability to render information and activity into datapoints. In turn, datafication is becoming “a leading principle” to access, understand and monitor people's behavior (Van Dijck, 2014: 198). In other words, it is both a new paradigm as well as a new frontier of accumulation (Sadowski, 2019).

While earlier forms of capitalism were characterized by the commodification of basic inputs of production, such as land, labor, and money (Polanyi, 1944), the digital transformation has given rise to a new form of capitalism that goes a step further. Instead of merely commodifying basic inputs of production, as industrial capitalism did, this new form of informational capitalism now re-conceptualizes them as datafied inputs to new algorithmic modes of profit extraction (Cohen, 2019). There is also a new, fourth, factor of production that enters the economic system under informational capitalism (Cohen, 2019: 25, 41): data flows extracted from people. ³

Characteristics of personal data

Data assume some characteristics that distinguish them from other popular goods. First, data are essentially non-rivalrous. In most cases, they can be copied and shared without their value being depleted (Thouvenin et al., 2017: 6, 15). This is because unlike cakes, data can in principle be “consumed” more than once (Cheneval, 2021). There are exceptions of course, such as data about consumer preferences that become useless once a specific need is satisfied. But in most cases, data collection aims at deriving population-level insights from data subjects for population-level applicability (Viljoen, 2021), and their copying and sharing does not affect their value a priori.

Second, personal data are versatile. They can be applied by different actors for various purposes and in different contexts. Especially in the case of personal data, the ‘real’ market value lies in their combination with large data sets. The appliance of big data analysis, in turn, reveals connections between pieces of data that the human eye fails to detect (Andrejevic, 2014; Boyd and Crawford, 2011). To make use of this competitive advantage, one needs massive amounts of data, as well as the capacities and capabilities to make sense of them (Metzler et al., 2016). Correspondingly, even data that used to be seen as a byproduct of information flows have now become valuable inputs.

Roles of personal data: capital, commodity, or productive asset?

In an economic system engineered around information, data flows about people assume the role of capital (Sadowski, 2019). And just as capital may assume different functions in different contexts, so do data flows about people. In the digital economy, personal data can assume the role of a commodity asset. This approach “defines rights with reference not only to the particular data itself but also to the person (or company) that holds the information embodied in the data” (Thouvenin et al., 2017: 15). Correspondingly, personal data can be a commodity for a variety of actors, including corporations, non-governmental organizations, or even governments.

Secondly, data about people can also serve as a productive asset. With the appropriate technical infrastructure in place, they can be used to create valuable insights about people's behavior, as well as inferring potential personality traits and emotional states. In some cases, data flows about people are even used to create profiles of voters with the goal to infer their assumed interests and predict or manipulate their behavior (Cadwalladr and Graham-Harrison, 2018). These points illustrate that due to their versatility and non-rivalry, personal data assume different, and sometimes even multiple, roles at once. For the remnants of this paper, I focus on the broader category of personal data as capital and draw the distinctions where necessary.

Sources of personal data and questions of ownership

Most people would intuitively assert that they own data about themselves, and that they should therefore control who can access, use, aggregate, edit, and share them (WEF 2011: 16). The reality is more complex, however. People often neither own their criminal records, fitness data, or credit history, for example. This has to do with the fact that the personal data ecosystem consists of a range of actors—including identity providers, mobile operators, and service providers—which all contribute to the ecosystem that makes personal data the valuable resource they are. A more useful way to think about data ownership, then, is to detail their source, or how personal data come to be. We can differ between three sources of data: volunteered, observed, and inferred data (WEF 2011: 14). In the first case, individuals choose to explicitly share information about themselves, such as when they create a social media profile. In the second case, data are captured by recording activities of users, such as internet browsing preferences or location data. In the third case, organizations can analyze personal data to infer additional data from individuals, such as the calculation of credit scores via numerous factors deemed relevant to an individual's creditworthiness.

Today's prominent data protection laws, such as the European Union's GDPR or California's CPA, acknowledge the complexity of the personal data ecosystem, and place great emphasis on protecting individuals appropriately. While this is a central concern, there is often little focus on empowerment in these regulations. This is surprising, considering that the digital transformation was accompanied by claims to empower the individual from the very start (see: Negroponte, 1995; Matei, 2006; Rheingold, 1993). One explanation is that Western democratic legal rule and ethics have strong roots in liberalism, which emphasizes the right to be let alone (Benthall and Goldenfein, 2020; Freeden and Stears, 2013). But if we are interested in complementing data protection efforts with citizen empowerment, it is instructive to turn to a concept within political economy that's closely tied to this concern: property-owning democracy.

From property-owning democracy to data-owning democracy

In general terms, “property-owning democracy” is a political economic system characterized by the wide diffusion of capital ownership. Especially liberal egalitarians have provided important contributions to the debate in recent years, such as how the diffusion of capital is linked to distributive justice and preventing elites from controlling the economy and political life (O’Neill and Williamson, 2012b; White, 2012; Rawls, 2001: 139; Thomas, 2017). For the present purpose, I build upon their work in discussing how property-owning democracy could be extended to the digital realm, and what such a regime would look like.

The goal of a property-owning democracy is to reconcile “efficiency, equality and liberty” (Meade, 1993). This is achieved via putting “all citizens in a position to manage their own affairs on a footing of a suitable degree of social and economic equality” (Rawls, 2001: 139). Different authors suggest a variety of activities by which a property-owning democracy might be structured. These include, but not exclusively, macroeconomic planning, regulation of economic institutions, establishing market rules, implementing resource transfers, as well as providing public goods (O’Neill and Williamson, 2012a: 4; Rawls, 2001). Most importantly, they all envisage an active role of the state and the provision of a basic income or a basic capital grant to all citizens (White, 2003; Ackerman and Alstott, 1999; Hacker, 2013). In light of the limited scope of this paper, the following discussion centers on this last point: the wide diffusion of capital.

Conceptualizing personal data as capital in a data-owning democracy

So far, there have been different accounts of what a ‘republic of educated property holders’ (White, 2012: 133) would look like. They include a one-time capital grant which could be used to finance a range of activities (White, 2003: 186), creating a ‘shareholder society’ by awarding individuals with highly generalized assets (Ackerman and Alstott, 1999), or to award people with rights in less-multipurpose assets, such as new models of private and public housing (Hacker, 2013). If we acknowledge that the digital transformation has led to a new type of asset class, this presents us with a unique opportunity: to endow individuals with rights to their personal data (Cheneval, 2021; Loi et al., 2020). Data also potentially surpass traditional capital goods in terms of economic potential. Even though a person's personal data worth is hard to determine, current estimates range between a couple of hundred up to several thousand dollars per year (Bloor, 2020). This is likely to increase, considering the rise of data-intense services, goods, and the emergence of the smart home industry.

The republican contribution to data-owning democracy

When it comes to specifying data-owning democracy's normative underpinnings, republican theory appears to be a decent candidate for two reasons. First, Republican theory, as I use it here—following Pettit (2012, 1997)—conceptualizes freedom as the absence of domination. To be free, one has to be free from the arbitrary power of another (Pettit, 2012). This focus on the ability to interfere, goes considerably further than classic liberal theory. This makes republican theory particularly promising with respect to the digital economy, where corporate actors don’t need to actively interfere with individuals’ choices to dominate them. Instead, they can constitute the option sets individuals face, or structure incentives in ways to make their preferred option more likely to be chosen (Fischli, 2022).

A second reason that speaks for a republican underpinning is that the theory regards political equality, which refers to equal status freedom for all, as closely connected to economic equality. Disparities in wealth and power are only permissible as long as they do not undermine political equality (Pettit, 2012: 298). Consequently, if material domination in the economic sphere becomes a major threat to the liberty of citizens, the republican seeks to remove it by restructuring incentives so that inequalities fall only within a permissible range (Thomas, 2017: xvii–xviii). For this reason, republicans have also started to take up the idea of a political economy that is based on wide dispersal of capital, and the ability to prevent powerful elites from dominating the economy and political realm (Elkin, 2006; Dagger, 2006; White, 2012; Thomas, 2017).

Taking these points together, we can see how republican theory is useful to specify Meade's “efficiency, equality and liberty” credo for data-owning democracy. As we consider a political economy regime, we would in principle expect it to be efficient. Keeping in line with the republican emphasis on political equality and non-domination, a data-owning democracy in the republican spirit therefore strives toward efficiency, political equality, and freedom as non-domination.

Relevant properties: property rights vs. data portability

The next task is to clarify what the wide distribution of data as capital would look like. From a property rights perspective, one can in principle distinguish between rights to primary and secondary use of data. As previously pointed out, the personal data ecosystem is highly complex and consists of multiple actors with different stakes and respective contributions. This is underlined by a new generation of algorithms that are based on machine-learning, which means that in certain cases, not even their designers know how exactly algorithms make sense of data (Pasquale, 2015). If we understand personal data as data that relate to an identifiable person or “data subject,” then big data analytics and data-processing algorithms based on machine-learning render almost all data potentially meaningful (Purtova, 2017: 17). Correspondingly, it is almost impossible to try and distinguish the exact stake individuals have in this ecosystem with regards to data production, and, relatedly, at what point property rights over primary use of personal data should kick in.

Similar difficulties present themselves when dealing with ownership rights over secondary use of personal data. Secondary use of data refers to any scenario in which data are used for a different purpose than they were originally collected for. Defined this way, secondary use of data would capture large parts of current activities by social media platforms or search engines, such as selling advertisements or using personal data to improve their services (Cheneval, 2021: 254–255). For that reason, efforts to introduce property rights over secondary data use are likely to be met with paramount opposition from the corporations that currently capitalize on it. These concerns aside, it remains doubtful whether awarding individuals with secondary usage rights would be sustainable or desirable. For example, people might stop participating in data-generating activities that are important for the common good altogether (value problem). Even worse, corporate actors might also incentivize them to sign off all their rights, essentially worsening the status quo (misuse problem). These “twin problems” already exist today (Mulgan and Straub, 2019). But under a rights regime that centers around secondary data use, chances are high these problems would get worse. A further difficulty is that property rights over secondary data use could even have a disempowering effect in certain contexts. By endowing individuals with the rights to use, and potentially sell, their personal data, it's easy to imagine a scenario in which people—especially those who lack the means to generate an income otherwise—are incentivized into generating even more data about themselves, essentially trapping them in an ever-lasting data commodification cycle.

An alternative, albeit less far-reaching, approach is to refrain from property rights altogether, but instead ensuring that individuals receive copies of personal data flows already generated with their consent. The European Union's Data Protection Regulation already includes such a right to “data portability” (Art. 20). This right aims to allow data subjects to obtain and reuse their personal data for their own purposes across different services (Wong and Henderson, 2019: 173). According to the regulation, “the data subject shall have the right to receive the personal data concerning him or her, which he or she has provided to a controller, in a structured, commonly used and machine-readable format and have the right to transmit those data to another controller without hindrance from the controller to which the personal data have been provided.” For the right to data portability to work, the processing must be carried out by automated means, and be based on the individual's consent. This excludes personal data that were processed for other reasons, such as to perform a contract, public interest, or for legitimate business interests—provided these interests do not refer to direct marketing, or override individuals’ fundamental rights.

Potential right holders

Most authors raising questions around data ownership adopt individualistic reasoning when discussing data ownership and prospective rights holders (Thouvenin et al., 2017; Jones and Tonetti, 2020; Cheneval, 2021; Purtova, 2015). But such a focus neglects the capitalist logics currently shaping the digital economy, where discrimination and classification are not just undesirable externalities, but rather tightly waved into the logics of commodification, analysis, and monetization of data. The digital advertisement industry, for example, would probably pay more for the consumer data of an average U.S. citizen, as opposed to data of a refugee from the Global South. ⁴ Additionally, awarding individuals with data capital runs counter to how value creation in the data economy works. The true value of data flows lies in their aggregated form. This relates both to economic value, as well as economic power. In an economic system geared toward correlation and prediction, owning one's data streams might contribute to increased economic value for individuals, but it hardly tips the scale of power in their favor. ⁵ To the contrary, forcing individuals to sell their data in a marketplace without a major change in their a priori bargaining power stands to worsen, instead of improve, the status quo (Feygin et al., 2019: 6). In other words, absent additional action, simply awarding people with rights over the use of their data is insufficient to reduce dependencies on corporate actors. It might even exacerbate the existing inequalities in society.

Another option is to introduce collective data ownership rights. In such a set-up, a collective body—be that a community, a municipality, or a neighborhood—owns the rights to their data. There are at least three reasons that speak for such an approach. First, it is in line with the insight that today, most data extraction schemes in the digital economy aim at extracting population-level insights for population-level applicability (Viljoen, 2021). In the current economic configuration, individuals tend to be grouped together by certain attributes to predict their future behavior (Mantelero, 2016), or infer information about others with similar characteristics (Wachter, 2020). Awarding collectives with data ownership rights therefore accommodates the observation that in the networked informational architecture, individual data streams tend to affect others, too. Second, because personal data tend to be most valuable in aggregate, there is a strong case to be made for treating data as a collective good and follow the logic of data creation as collective labor in order to create funding streams for public goods (Feygin et al., 2019: 8). And third, as a collective—especially in the form of a constituency or citizenry—citizens have much more bargaining power regarding ownership claims to primary or secondary use of their data. This is especially true if they collectively own the infrastructure involved in the data generation.

Data-owning democracy: introducing the dual structure

The previous section indicated that data's intrinsic quality and the complexity of the personal data ecosystem present us with a dilemma. If efficiency is a real constraint, primary or secondary data usage rights for individuals seem out of the question. But if individuals merely own copies of their data, powerful corporate actors continue to enjoy network effects, and their ability to exercise unchecked domination continues. Drawing on Hobhouse, who differentiated between “capital for use” and “capital for power,” awarding individuals with copies of their data flows might confer “data for use,” but hardly “data for power” (Hobhouse, 1913/1994). Conversely, while a move toward collective data ownership is more in line with the realities of value creation in the data economy and might approximate “data for power,” it would do so at the cost of individual empowerment. I therefore propose a dual structure; two complementing layers that enable us to reconcile different demands, while staying true to data-owning democracy's stated purpose.

Layer one: public data and collective data ownership

The first layer of a data-owning democracy is characterized by a digital public infrastructure and corresponding collective data ownership rights. This sphere is linked to a sufficiently large association, ideally to a public body or city. Potential candidates are communities that have a sufficient degree of stability in order to have relatively stable boundaries of the data they produce to facilitate property claims and data management (Purtova, 2017: 18), such as municipalities, or smart cities. While the latter have so far received relatively poor privacy scores due to the involvement of private companies (Cecco, 2020), there are noteworthy exceptions. One of them is the City of Barcelona, which launched “project Decode” in 2017.

Project Decode was a smart city initiative with data sovereignty at its core and ran as part of the European Union's Horizon 2020 Program. The pilot in Barcelona ran from 2017-2019 and focused on the Internet of Things (IoT), as well as open democracy and data sovereignty. As one of its central features, project Decode's digital infrastructure was created to facilitate and ensure that citizens would own the data they created. To do so, the City revised the procurement deals between City hall and private sector providers and included “data sovereignty” clauses in public procurement contracts (Bass and Old, 2020: 6). As a result, any supplier that worked for the city had to return the data they gathered to deliver their services in machine readable format. That way, the data were turned into a public good and remained in the public domain, while at the same time preserving privacy, ethics, and security by design.

A core component of the digital infrastructure were data commons, platforms that enable groups of people to leverage their data's collective value (Bass and Old, 2020: 9–10). Data commons can be built on top of technologies that give people control over their data in the first place, while also providing trustworthy mechanisms to share them. They can give rise to new democratic data sharing arrangements where people collectively decide the terms by which data are accessed and used. In project Decode, the commons-based approach was characterized by using tools for decentralized identity and trusted data sharing among local communities. It also sought to shift agency and control by letting citizens decide what data they want to share, with whom and on what terms (Bass and Old, 2020: 6). This way, citizens were enabled to set the conditions of their data sharing themselves. This new digital infrastructure remained open to local companies, cooperatives and social organizations that could build data-driven services and create long-term public value this way (Bass and Old, 2020: 6).

Project Decode provides useful insights into how a digital public infrastructure can help promote citizen empowerment. First, ‘data sovereignty’ is synonymous with collective data ownership. It also acknowledges the importance of data exclusion as a core strategy to battle power differentials in the data economy (Prainsack, 2019). As a collective, citizens owned the secondary rights to the data they produced. Second, such a digital infrastructure empowers citizens with knowledge about their data and reduces their dependence on corporate actors for these services. It also addresses the value problem associated with data underuse. The current siloed nature of the data economy makes it very hard for public bodies to derive data-driven insights, because they lack both the data as well as the infrastructure to derive the relevant insights from them. Creating a digital public infrastructure that allows for large amounts of citizen data to be analyzed is a way to counter this trend. Moreover, as citizens are in charge over what data they want to share and with whom, data vetted and submitted by citizens are more likely to be accurate—as opposed to inferred information that takes place without people's awareness or consent.

A digital public infrastructure can also strengthen democratic governance by providing new tools for political participation. Examples include the ability to sign petitions, online consultations, as well as building local community fora (Bass and Old, 2020: 30). Liu (2021) has shown that complementary crowdsourcing can play a significant role in policy design, because it allows government agencies to co-produce with the public and generate a broader range of solutions than would otherwise be possible (Liu, 2021: 324; Kaminski et al., 2016). And provided the digital public infrastructure is created in a democratic state, the resulting data flows are democratically controlled. This means that citizens have, in principle, unique and safeguarded powers when it comes to collectively determining how their data are to be used.

These points notwithstanding, the first layer of data-owning democracy would have to go a step further. Most importantly, it would require the creation of a “digital wallet,” a safe and privacy-friendly storage unit for all citizens that allows easy data management. Different suggestions as to how personal data could be stored for individual use have already been taken up in discussions around data ownership (Cheneval, 2021; Bass and Old, 2020: 8; Brochot et al., 2015). In light of constantly evolving formats for safely securing personal data, such as Apple's Health Records that allows users to securely store and aggregate their health records, the creation of such a digital wallet that stores citizens’ data seems sufficiently possible. Because it would enable people to store both citizen as well as individual data, the digital wallet is the main connection between the first and the second layer of data-owning democracy.

Layer two: voluntary data and individual data flows

The second layer of data-owning democracy empowers individuals via their personal data flows. Whereas the first layer is concerned with value creation for the common good, the second layer aims at empowering individuals by letting them harness the economic potential of their existing data flows. Here, people possess copies of those data flows to whose processing they previously consented.

Apart from storing citizen data, the digital wallet also serves as the storage unit for additional personal data, such as consumer preferences or mobility data. Whenever individuals consent to creating data about themselves, for example by wearing a fitness bracelet, they automatically receive a copy of their data. As pointed out before, the GDPR already knows a right to “data portability” (Art. 20) which provides the legal basis for this mechanism. In the current form, it posits the burden of receiving data on the individual. Conversely, in the set-up envisaged here, people automatically receive copies of their data whenever they are generated and can use them for whatever purpose they want. For example, they can choose to donate, transfer, or sell data about their consumer preferences to interested parties.

They are also free to pool their data together with other individuals, for example via data cooperatives. Data cooperatives are organizations that are collectively owned and controlled by their members, and where members voluntarily pool their data to create mutual benefits (Micheli et al., 2020; Ho and Chuangt, 2019; Hafen et al., 2014). They are particularly promising in this context because they represent consumer interests on an aggregated basis and essentially work as a solicited “broker,” representing groups of users while having fiduciary responsibilities over user data, defined by a contract between user and cooperative (Feygin et al., 2019: 19). Cooperatives are commonly associated with the one-member-one-vote practice, which directly prevents elite formation and promotes collective and egalitarian modes of decision-making. Their two main pillars, democratic ownership and control, have contributed to a considerable “revival” of the cooperative in academic and activist thought, especially with regards to empowering workers and users in context of the digital economy and, more specifically, digital platforms (Sandoval, 2020; Scholz and Schneider, 2017; Borkin, 2019).

In data-owning democracy, the cooperative idea mainly refers to the pooling of individual data flows. Cooperatives may simultaneously increase the value of their member's data, create insights about them, as well as sell data or data-related services to third parties—thereby acting as “trustees” or “brokers” on behalf of their members. In principle, cooperatives could contribute to the empowering character of a data-owning democracy in various ways. In a “thin” form, they would allow individuals to pool copies of their data flows and enjoy greater bargaining power vis à vis corporate actors or other third parties. This does not necessarily imply the selling of data flows. Provided the cooperatives possess the adequate infrastructure, they can also conduct their own data analysis and sell specific services, instead of the data they possess—similarly to how big corporate actors in the advertising business operate. In their “thick” form, data cooperatives can become primary data generators themselves, where they enjoy certain primary rights over the data they produce within the cooperative infrastructure.

Such data cooperatives already exist around the world. While some are explicitly created with the goal to create profit for their members, others are not. This variety reflects the multifold possibilities data cooperatives present for a data-owning democracy. In such a regime, people can simultaneously be part of different cooperatives with different purposes. To allow data cooperatives to gain a meaningful foothold in the data economy, additional legal and regulatory interventions are required. As Feygin et al. (2019: 35) argue, a new class of regulated business entities should be established by law in order to facilitate fair and efficient bargaining over data. These entities would have special obligations to maintain independence from data-using businesses, to refrain from permanent data alienation, and to uphold strict fiduciary obligations to their members.

Discussion: approximating efficiency, political equality, and non-domination

After sketching data-owning democracy's relevant properties and prospective structure, this section discusses how the envisaged political economic regime corresponds with its republican values. It also highlights republicanism's core contribution to questions of digital empowerment: how citizens can reduce their dependence on corporations in the digital realm.

Efficiency

Despite being costly and complex in its initial set up, a data-owning democracy would likely increase in efficiency over time and create public value and impact at scale. Provided citizens express trust in the digital infrastructure and those running it, the generated public data would provide policymakers with vital insights into the public body's “inner life,” allowing them to make predictions and take precautionary action where needed. A data-owning democracy as envisaged here also addresses the problem of data underuse, by enabling public bodies to derive insights from an enormous and highly valuable dataset. Data submitted and vetted by individuals are more likely to be accurate than inferred information. Overall, data-owning democracy is therefore well positioned to embrace the potential of data-driven insights for individual and collective well-being. This expectation is in line with existing research, which indicates that a market in which people own their data might be the best option to balance efficiency and privacy concerns (Jones and Tonetti, 2020: 2820).

Political equality

Recall that from a republican standpoint, political equality does not require strict material equality (Pettit, 2012). Rather, economic inequalities are permitted as long as they don’t allow one actor to enjoy disproportionate market power over others that threatens to undermine political equality. Asking whether a data-owning democracy can approximate political equality is therefore another way of inquiring whether it can create the necessary conditions for approximating non-domination, without compromising the free status of oneself and others.

The dual structure of the data-owning democracy appears to correspond with this goal in four respects. First and foremost, citizen data are used in aggregate—and because there is no inherent profit motive in the collective data generation—each citizen's data flow is, in principle, equally valuable. Second, (collective) control over data generation and management is a core component of any strategy that aims at combatting bias, discrimination, and ultimately, inequality, through data-driven practices. This is because machine learning algorithms reproduce and reinforce the patterns that exist in the data sets used to train them (Cohen, 2022: 15). Data are never strictly “neutral,” but are already shaped by the people, infrastructures, and instruments involved in deriving or generating them (Gitelman, 2013; Kitchin, 2014). Therefore, tackling the issue of data constitution is a first attempt to address discriminating outcomes in data-driven practices. Third, the digital public infrastructure offers new means for all citizens to participate politically—provided one possesses the relevant know-how to do so. Fourth, if the digital public infrastructure is created in a democratic state, the resulting data flows and their use are democratically controlled. This means that citizens in a data-owning democracy have unique and safeguarded powers when it comes to collectively determining how their data are used.

Freedom as non-domination

Finally, and most importantly, data-owning democracy fosters non-domination through the reduction of dependencies. Even though a data-owning democracy cannot put all generated data flows in the digital economy under public control, it can help create spaces where collective data control and democratic oversight are possible. On the collective level, a citizenry lowers its dependence on corporate services significantly when it comes to insights about its members and public life. And because citizens own the secondary rights of their data in such a set-up, they challenge the current powerful standing of corporate actors in the digital economy—at least regarding insights about citizens, public life, and infrastructure. This in turn reduces corporate actors’ ability to arbitrarily exercise domination.

The observation also applies to the individual level. Especially in the context of data cooperatives, individuals are simultaneously digital wallet holders and cooperative members. As the former, they get to decide what data they want to share with whom. As the latter, they get to co-determine the digital environment in which their decisions take place, for example, one characterized by transparency, informed consent, and viable alternatives. This applies to the overall choice architecture within which data transactions occur, as well as more specific aspects—such as default settings, interfaces, or user dashboards (Loi et al., 2020: 13–14). This co-determination of the rules and standards that shape the digital environment is a central factor in addressing domination in the data economy, which often manifests precisely in the design of the options individuals get to choose from online (Fischli, 2022).

Freedom as non-domination is also fostered via resourcing. The creation of data commons empowers citizens with new knowledge and insights about their circumstances. Moreover, by strengthening citizens’ control over their data flows, they enjoy new possibilities of political participation and self-determination. Here, design is key. Interfaces, indicators, and dashboards should foster transparency and accountability, and overcome traditional failure modes, such as dark patterns designed to deceive users and interfaces that foster addictive user behavior (Cohen, 2022; Hartzog, 2018). In project Decode, for example, the corresponding governmental measure for ethical data management purported that wherever possible, projects based on data will be able to check the algorithms using simulations based on city data, and third-party technology suppliers must reveal the underlying logic behind any IT process for automated decisions pertaining to any systems used by the City Council (Barcelona Ciutat Digital 2018: 16–17).

The same holds for the second layer. Because individuals own copies of their data flows, they now have the possibility to use, transfer, and sell their data, thereby reducing their dependency on labor. Cheneval (2021) has estimated that over time, the compensation from secondary use of personal data would turn into a considerable income stream for individuals. At a time of increasing automation, such an additional income would particularly strengthen individuals whose jobs are at risk of being automated, or whose jobs are inadequately protected—such as gig workers. There is also a danger associated with this: to incentivize people toward producing more data about them, essentially worsening the status quo. While this point deserves careful attention, it should be noted that awarding individuals with copies of their data flows does not necessarily incentivize them to create additional data flows about themselves. The second layer of data-owning democracy seeks to make the status quo more inclusive, not substantively changing its underlying dynamics.

Finally, a data-owning democracy also resources individuals with new forms of power. This is especially pronounced in the case of data cooperatives, which, if more common, could build the foundation for collective action, for example by building new alliances and allowing new insights into previously undetected patterns, such as unfair working conditions. Data cooperatives, in other words, could be important drivers for change in the digital realm, moving from individual economic empowerment to collective political action.

Data-owning democracy therefore fits neatly into current community wealth building initiatives that center around the creation of collaborative, inclusive and democratically controlled local economies (Guinan and O’Neill, 2020). ⁶ It might also be seen as an instantiation of what Ferretti (2020) calls “organizational strategy”; the idea that public institutions should change the kind of organizations running club goods and platforms (Ferretti, 2020: 55). Lastly, in placing great emphasis on cooperatives, data-owing democracy also follows a well-established path within republican theory. Republican theory has a historical connection to cooperatives, precisely because they are associated with increased accountability, self-determination, and flat power hierarchies (Gourevitch, 2013).

Toward data-owning democracy: democratic control, trust, and social welfare

Instead of imagining the digital economy in a fundamentally different way, this paper has taken a less ambitious, but nevertheless important route. It started out by asking: How, absent more substantive changes in the overall structure of the data economy, can we ensure that current data commodification processes benefit as many people as possible? The resulting political economic regime, “data-owning democracy,” makes a case for strengthening citizens’ rights in the digital economy, and using (existing) data flows about people to foster both collective and individual well-being. As the discussed project Decode in Barcelona illustrated, communities have much to gain from re-asserting collective control over public data flows, for example by promoting common well-being, while simultaneously reducing their dependency on private, for-profit entities.

Further, moving toward a data-owning democracy will inevitably be accompanied by a number of challenges. First, data's unique economic status in the digital economy will likely provoke opposition from powerful technology companies to any changes in data ownership schemes. Especially in the Western world, private companies currently enjoy disproportionate power over personal data and are unlikely to consent to public procurement contracts with data sovereignty at their core without a fight. This is why broad implementations of the two-tiered system envisaged by data-owning democracy will be central to its success. Within Europe, some efforts are on the rise; After Barcelona, the City of Hamburg is also working on a similar project that is working on a data-driven strategy toward increased political participation and sustainability (The New Institute, 2021).

The second issue concerns democratic control. Recall that the republican has no problem with state interference a priori, provided citizens have appropriate means of control over how they are governed. Because a data-owning democracy relies heavily on state involvement, particularly when it comes to the creation of the digital public infrastructure, such an economic regime has to be embedded into an overall democratic structure in order to make the considerable state interference in people's lives legitimate. Because the capital assets at stake are personal data flows, the political stakes increase, too. Whether or not one ascribes to the intuition that personal data are a part of “oneself,” they do reveal a lot about their originators. Any political economic regime capitalizing on personal data flows therefore needs to devote adequate efforts toward ensuring appropriate democratic control by those subjected to it, both on an individual and collective basis.

A third challenge is the promotion and securing of trust. A data-owning democracy relies heavily on the people on whose data it draws. Absent a broad trust in the process and democratic oversight, for example by oversight boards and transparency measures, people are unlikely to participate in the data-generating process that is crucial for a data-owning democracy's ability to promote citizen empowerment. Here, transparency and useability are central factors. Increased usability is often correlated with higher trust. If one is presented with an interface that is too difficult to understand, people are much more likely to abstain from interacting with it. The authors of the project Decode report state the importance of technology to be convenient and user-friendly, for example by applying user-centered design and on-boarding (Bass and Old, 2020: 30). Furthermore, there always needs to be the option to “opt-out” of data production, and to exercise individual data control.

Fourth, we must acknowledge the dangers associated with unequal data worth and potential dependency on data flows for income. It is easy to see the idea of compensation being co-opted by certain political forces to argue for a data-generated individual income at the cost of social welfare provisions. However, if people are made dependent upon the income from their data flows, incentives will easily be structured in ways to put the burden of data production on the already worst-off, thereby worsening their situation by eroding both their privacy and autonomy. Strengthening individuals’ rights to data portability ought not to replace current welfare provisions but endow individuals with capital assets they can make use of to their liking. This is in line with the current European legal understanding that data protection is a human and fundamental right.

Finally, one could point out that the economic regime proposed here does not go far enough, or that it even legitimizes current data extraction practices by putting them into a new context. It is certainly true that data-owning democracy does not change the underlying dynamics of the digital economy, for example by introducing new property rights of personal data or getting rid of the most damaging data extraction mechanisms. Its purpose should be seen as a second-best approach: to complement existing data protection regulations and anti-trust efforts with new, albeit limited, tools to empower citizens, until more substantive reforms of the digital economy are found. Questions of how to reign in Big Data behemoths therefore continue to be urgent and relevant, and crucial to citizens’ empowerment online.

Concluding remarks

This paper extended property-owning democracy to the digital realm and introduced “data-owning democracy,” a new political economic regime characterized by the widespread diffusion of data as capital among citizens. Drawing on republican theory, it argued that the normative underpinnings of this regime could be specified around the core values of efficiency, political equality, and freedom as non-domination. Because of data's unique features and role of data in the digital economy, it proposed a regime characterized by a dual structure; two layers that combine different forms of data ownership rights and confer different forms of power.

While I refrained from proposing specific policy proposals, I modeled data-owning democracy on real-world examples as closely as possible. In so doing, I sought to define the building blocks of a data-owning democracy in ways that correspond with existing approaches or would in theory be easy to implement. That said, together with additional efforts, such as regulating tech corporations or introducing new taxes that incentivize firms to improve privacy protections, the regime proposed here could be an important step toward promoting citizen empowerment, while staying true to the republican ideals of political equality and non-domination. It would also be a first step toward including people into the value cycle of their data, a step that seems heavily overdue.