The Digital Welfare State: Conceptual limits and possibilities

Abstract

The concept of the digital welfare state (DWS) has gained popularity in research on welfare system digitalization, but its meaning is under-theorized and largely unproblematized. Invoked without clear criteria, the concept leaves open fundamental questions about what qualifies as a DWS and why. In this article, we propose three key pillars as foundational criteria for analyses of the DWS: (1) the presence of a welfare state, (2) processes of digitalization – encompassing older digital technologies and more advanced technologies and their interplay with enduring manual processes, and (3) the degree of digitalization and whether this warrants a particular welfare state's classification as digital. We argue that these pillars, although seemingly self-evident, raise under-explored questions about the theoretical foundations and boundaries of the DWS concept. In highlighting these questions, we reveal that much of the literature is characterized by single country case studies, assumptions of linear change, and a degree of Western-centrism, which leaves broader patterns of digital welfare reform, and the role of non-state actors therein, under-theorized if not overlooked. By foregrounding these limitations, we demonstrate how greater theoretical rigor and reflexivity can sharpen the concept's analytical and practical value. Properly specified, the DWS provides a useful heuristic for researching broad patterns of digitalization without flattening national or local specificities, and for equipping both scholars and policy makers with a clearer lens for critical engagement and political contestation.

Introduction

Welfare regimes worldwide are rapidly adopting a range of information and communication technologies (ICTs) to digitize, automate and streamline welfare state functions. This includes the automation of ‘back end’ welfare operations, such as the digitalization and automation of benefit calculation and payment mechanisms, entitlement checks and fraud detection (Alston, 2019; Eubanks, 2018), as well as the introduction of digital identification systems, online platforms and apps to augment or replace in-person service delivery (see, e.g., Bhandari, 2020; Meers et al., 2024; Raso, 2023). The ‘digital welfare state’ (DWS) has become a common shorthand to describe this diverse set of changes to social protection systems wrought by ICTs (see, e.g., Alston, 2019). To date, however, the concept of the DWS has not been adequately defined or theorized. The term often functions as more of a broad descriptor than as a robust analytical construct. Without explicit examination of its theoretical foundations, it becomes difficult to assess what exactly constitutes a DWS and the extent to which welfare systems have fundamentally changed as a result of digitalization (van Gerven, 2022).

This article interrogates the DWS with a view to contributing to the development of a more theoretically robust engagement with the concept. We introduce three key pillars as foundational criteria for applying the DWS label with greater rigor and reflexivity. First, the concept of the DWS assumes the existence of a welfare state as a locus of transformation. Second, the concept encompasses welfare state digitalization, including the adoption of ICT technologies – whether advanced or simple – and attendant socio-technical shifts in personnel, expertise, governance and administration. Finally, welfare states must reach a threshold of digitalization substantial enough to justify their classification as DWSs. Seemingly self-evident and embedded in the DWS label itself, each pillar nonetheless raises hitherto unexamined questions about the concept's theoretical foundations and limits. Foregrounding these questions, our approach positions the DWS as a heuristic device for cross-disciplinary analysis of how welfare states co-evolve with, and transform through, digitalization. We challenge the view of digitalization as a predominantly technical or operational process, attending also to its entanglement with deeper socio-political transformations in welfare states in the Global North and South. Although not an exhaustive empirical analysis, our study establishes some common entry points for research, including comparative research, seeking to advance a richer theoretical understanding of digitalization's role in reshaping welfare states globally.

Our approach to this task recognizes that scholarly vocabularies carry conceptual baggage. As such we seek to examine the underlying assumptions and conceptual limitations present in applications of the DWS concept across disciplines. Whilst the concept of the DWS is undoubtedly less value-laden compared to pejorative labels such as ‘welfare dependency’ (see Fraser and Gordon, 1994) or the ‘welfare queen’ (see McCormack, 2004), its use nonetheless has intellectual and political consequences that warrant critical reflection. Ways of describing social phenomena can, for example, tacitly allocate blame or present particular outcomes as natural or inevitable while foreclosing others (Foucault, 1969[2002]; Fraser and Gordon, 1994). Similarly, scholarly labels like the DWS can silently prioritize particular forms of knowledge or action and shape how digital welfare futures are envisioned. We aim to make some of these tacit assumptions explicit, fostering a more thoughtful and reflexive approach to studying digital welfare.

We begin by reviewing the literature on digital welfare, then charting the emergence and use of the concept of the DWS in scholarly literature and public commentary. We then consider the definitional features and boundaries of the concept, proposing three key pillars, before reflecting on some conceptual limitations and embedded assumptions. We conclude by reiterating our proposed approach and outlining its intellectual and practical utility.

Literature on digitalization of the welfare state

The welfare state has a long history of technological innovation (see, e.g., Harlow, 2003; Henman, 2010). While the evolution of welfare regimes has long been intertwined with technological processes including digitalization (see Galis and Vlassis, 2025; Kaun and Masso, 2025), the pace of change has accelerated in recent years across many sites globally, albeit with distinct local patterns and variations (Alexopoulou, 2025; Raso, 2023). Scholars have, for example, traced the digitalization and automation of eligibility tests, including the roll out of biometric and digital identity systems (see, eg, Bhandari, 2020; Well, Currie and Stewart, 2023); the adoption of algorithmic personalization of services (Allhutter et al., 2020), and the expansion of data-driven debt recovery and fraud detection tools (Alston, 2019; Eubanks, 2018). This growing body of work highlights not only technical transformations but also changes to the way citizens interact with welfare systems. As Meers et al. (2024: 119) observe from the UK context, ‘[t]he front-line of the welfare state is increasingly not a letter, phone call or face-to-face visit, but an online interface’.

Research across disciplines has explored the significance and consequences of the rise of the DWS, particularly in advanced capitalist countries (Alexopoulou, 2025). Sociologists and anthropologists, for instance, have analyzed digitalization's transformation of both policy processes and public sector governance (Henman, 2019; Schou and Hjelholt, 2018), as well as the institutions and logics of welfare provision more specifically (Huws, 2020). These shifts have sparked debate over whether digitalization is driving a reconfiguration of welfare state governance (van Toorn et al., 2024) along increasingly neoliberal lines (see, e.g., Mann, 2020), or potentially in ways that depart from neoliberal orthodoxy (Dammann et al., 2022). Sociological research emphasizes that digitalization is not a neutral technical process but one that transforms the epistemological and political foundations of governance itself (Henman, 2019). Technologies increasingly shape how welfare problems are represented; often privileging datafied, machine-readable formats that determine what counts as legitimate knowledge for state intervention. Whilst being careful not to frame technology as an autonomous force, this work argues that technologies possess agency through their material structuring of the world, enabling and legimitizing certain actions while constraining others (Henman, 2022).

Research in human geography conceptualizes digitalization as a reconfiguration of the spatial contours of state–citizen relations (Schou and Hjelholt, 2019). Maguire and Ross Winthereik (2021) explore digitalization as a process of “territorializing” state land and resources with new forms of data infrastructure and exchange. Political economy analyses also raise questions about whether digitalization is eroding state capacity (Collington, 2022), shifting state powers and functions to private actors (van Toorn, 2024a), and deepening the public sector's operational integration with and reliance on private technology companies (Dencik, 2022; Valtysson and Jørgensen, 2025).

Since the 1990s, a substantial body of work has examined digitalization through the lenses of digital government, e-government, and Internet governance (see, e.g., Dunleavy et al., 2006; Grönlund and Horan, 2005). Situated primarily within public administration and related disciplines, this literature reveals the ways technological innovations have reshaped government information practices and transformed frontline administrative processes. There is, however, a tendency in this literature to overlook the deeper structural transformations of the state itself. That is, this literature tends to foreground technology and organizational dynamics while neglecting the distinct normative and political changes that accompany digitalization. Even so, this research yields valuable insights with wider relevance. Dunleavy et al.'s (2006) seminal work was among the first to highlight the growing influence of IT corporations under New Public Management (NPM), followed by (a) the first wave of “digital era governance” characterized by post-NPM reintegration and digitalization of services, (b) a second wave associated with big data analytics, government cloud infrastructure, new data retention and open data initiatives, and (c) a third wave involving applications of data science and artificial intelligence (Margetts and Dunleavy, 2013).

While much of this work focuses on the back-end operations of government, other scholars have turned their attention to the front lines of service delivery. Public administration scholars, for instance, have explored how traditional “street-level bureaucracy” (Lipsky, 1980) is increasingly being replaced by “screen-level” or “system-level” bureaucracy (Bovens and Zouridis, 2002), raising important questions about shifts in decision-making processes and the exercise of discretion (see, e.g., Allhutter et al., 2020; Breit et al., 2021; Ranerup and Svensson, 2023). Overall, the public administration literature usefully maps the mechanics of change but without sustained attention to welfare-specific socio-economic and normative shifts.

A growing body of critical scholarship has taken up some of these concerns, examining how digital welfare administration can propel inequality, hardship and social harm. For example, Eubanks’ (2017) landmark ethnographic study of automated welfare programs in the United States illustrates how digital technologies can disempower and exclude low-income groups, effectively “automating inequality.” Other studies similarly show that digital welfare systems and interfaces can entrench digital exclusion and discrimination (Buchert et al., 2023; Jaspars and Sathyamala, 2021; Larasati et al., 2023; Zheng and Walsham, 2021), while also expanding the surveillance and control of welfare recipients (see, e.g., Well et al., 2023). While such outcomes are not inevitable – indeed, many scholars point to the potential for inclusive and empowering digital services (Chohan and Hu, 2022; Kim and Lee, 2024; van Toorn, 2024b) – the literature nonetheless reveals a recurring pattern in which digitalization is embedded in broader reform agendas aimed at containing welfare expenditures by limiting access and at intensifying oversight of claimants (Alston, 2019; Coles-Kemp et al., 2020; Eubanks, 2018). While this critical work often addresses concerns central to social welfare, the state itself is rarely foregrounded as an object of analysis, and its institutional and normative transformations are usually implicit rather than explicitly theorized.

Broadly speaking, studies of digitalization in Western welfare states, ¹ and particularly ‘strong’ welfare states within Nordic nations and Western Europe, continue to dominate the field. There is, however, a rapidly growing body of literature that explores digital welfare development in the Global South. Explored further below, this literature examines, for example, how welfare digitalization can simultaneously enlarge state and corporate power (see, e.g., Chaudhuri and König, 2018; Khan, 2021; Singh, 2024), serve authoritarian ends (see, e.g., Pan, 2020), and entrench exclusion in already fragmented welfare states (Palacio Ludeña, 2021). This diverse interdisciplinary literature highlights common themes related to exclusion, surveillance and corporate influence while also showing how these manifest differently in countries marked by, for example, diverse histories of colonization, relationships of dependency and local traditions of welfare provision.

Each of the literatures reviewed here offer important insights on digital welfare, which we draw on and extend in our conceptualization of the DWS. Before doing so, we trace the emergence of this precise term across disciplines to consider how it is currently understood and to highlight the imprecision with which it is often deployed.

Tracing the emergence and development of the Digital Welfare State label

Arguably the most prominent use of the DWS label is by United Nations Special Rapporteur of Extreme Poverty and Human Rights, Philip Alston (2019). In his landmark report, Alston (2019: 2) charts a worldwide trend in which ‘social protection and assistance are increasingly driven by digital data and technologies that are used to automate, predict, identify, surveil, detect, target and punish’. Alston's report powerfully details the risks of digital welfare “reform”, punctuated by vivid examples of digital welfare harm and failure. But the report does little to, nor is it aimed at, contributing to a deep exploration of the conceptual origins and underpinnings of the concept of the DWS.

Whilst Alston did not coin the term – earlier references to the DWS can be found in, for example, the work of Pors (2015) and Watling and Crawford (2010: 212) – his report nonetheless sparked its broader use. The DWS label has since been deployed by scholars working in a range of disciplines, often with explicit reference to Alston's report (see, e.g., Choroszewicz and Mäihäniem, 2020; Kiely and Swirak, 2025; Madianou, 2020; van Zoonen, 2020). However, in many of these works, the concept is invoked but not developed or explained in detail. Often, the DWS is mentioned in article titles only or in passing as if its content and meaning are well understood (see, e.g., Ala-Fossi et al., 2019; Safarov, 2023; van Zoonen, 2020; Vonk and Ranchordás, 2025). In other settings, scholars have described the concept with reference to a list of indicative developments only (see, e.g., Pedersen, 2019). Certainly, these scholars carefully explain and explore digital transformation within specific welfare states and domains. But there has been little attention to theorizing or critically engaging with the concept of the DWS itself.

Where an explicit definition is offered, the term is most often used in a fairly general way to capture the integration of ICTs into social protection systems, from relatively basic forms of digitalization through to more advanced forms of automation capturing changes to both ‘back-end’ and ‘client-facing’ systems (see, e.g., Jørgensen, 2023; Kaun and Liminga, 2023; Meers et al., 2024; van Zoonen, 2020). Nguyen et al. (2024: 475), for example, adopt an explicitly ‘broad’ definition of the DWS in their study of an online welfare-to-work system in Indonesia, defining it as where ‘governments embrace digital technologies for the purpose of improving welfare service provision’. Similarly, Halliday et al. (2024) use the DWS as an umbrella term capturing various technological developments in modern welfare states. For these authors, the rise of the DWS is one of the ‘most profound shifts’ in recent history, which ‘has seen the welfare state increasingly dependent on digitalized, automated, and data-driven forms of public administration, which are altering the nature of welfare provision itself’ (Halliday et al., 2024: 64), including the rise of an ‘interface-first bureaucracy’ (Meers et al., 2024).

van Toorn et al. (2024) offer one of the more precise and meaningful elaborations of the concept. Here, the DWS ‘refers to a particular state formation in which wide-ranging digital technologies, including automated decision-making systems, algorithms, big data analytics and artificial intelligence (AI), are integrated into the administration and provision of welfare services by government agencies, which, in some cases, are then delivered through a complex set of contracted providers, which social policy scholars refer to as the mixed economy of welfare’ (van Toorn et al., 2024: 208). For Bennett (2025: 2; citing Zouridis et al., 2020) the DWS is ‘a term that encapsulates the growing use of technology in welfare design and administration’. Whilst Bennett begins by listing indicative technologies, she also considers the political and organizational consequences of this ‘digital takeover’, which includes changes to organizational values, processes and governance. As Bennett (2025: 2) stresses, ‘Digital welfare is not simply a technical and apolitical development’.

Similarly, van Gerven (2022) calls for a more rigorous conceptualization of the DWS, grounded in welfare state and social policy debates. She criticizes current literatures for over-emphasizing changes to welfare state tasks and tools. In van Gerven's words (2022: 258), ‘it is not just the presence of technologies that makes up a DWS’. Without proposing a precise definition of the DWS, van Gerven (2022) urges scholars to adopt a more theoretically robust approach to the DWS concept tied to welfare state literatures and theory. This would entail thoughtfully considering the relationship between technology and the institutions, governance and outcomes of the welfare state. Together, these studies offer the strongest theoretical engagement with the DWS, while stopping short of explicitly defining its core features.

Whilst the DWS label has recently grown in popularity, it is by no means ubiquitous in the literature. Dencik and Kaun (2020: 2) prefer the term ‘datafied welfare state’ to highlight how the welfare state is ‘intricately linked to digital infrastructures that result in new forms of control and support’ (see also Bagger et al., 2023; Dencik, 2022). The authors historicize processes of datafication, recognizing the longstanding entanglement of technology and welfare reform as well as pre-digital forms of data generation about the welfare state and its recipients (see also Kaun and Masso, 2025). Elsewhere Kaun et al. (2023: 880; see also Kaun and Masso, 2025) use the concept to the, ‘data welfare state’ to refer to the ‘increased reliance on digital data for decision-making and welfare provision’. These scholars do not explicitly reject the concept of the DWS and some, including Dencik (2022), use this label almost synonymously. The choice of ‘datafied’ or ‘data’ as the primary descriptor nonetheless foregrounds facets of change in the welfare state most closely associated with the generation, circulation and governance of data (Kitchin, 2014). Whereas the idea of the DWS is arguably more encompassing, capturing processes of digitalization, datafication and automation as well as attendant changes to digital and welfare infrastructures (van Gerven, 2022).

Like the broader scholarship on digitalization and welfare regimes, the ‘DWS’ label is more frequently deployed in studies of Western welfare states (but, see, e.g., Gupta, 2023; Nguyen et al., 2024). This may indicate that the DWS label has less purchase or is otherwise less relevant to digital welfare agendas in the Global South, a point to which we return below. There is also a trend in the research literature to use the term ‘digital welfare’ without necessarily tying this concept to the ‘state’ (see, e.g., Coles-Kemp et al., 2020; Hjelholt, 2024). This may indicate an attempt to disentangle the phenomenon of digital welfare development from the institutions of the nation state – a choice whose consequences we also return to later in this article.

Taken together, this literature illustrates the growing popularity of the DWS label particularly since Alston's (2019) report and within industrialized, ‘Western’ nations with ‘strong’ welfare states (see Busacca, 2025). Across this scholarship, the DWS is commonly used as an umbrella term encompassing a suite of changes to welfare programs, services and systems involving a wide range of technologies, including advanced technologies and simpler forms of automation and digitalization. Yet, few scholars have elaborated on the concept in detail, nor have we identified any examples from the literature where scholars interrogate the conceptual strengths and limitations of the term itself.

Defining the digital welfare state

What, then, is the DWS? In this section, we critically examine the concept, unpacking its meaning and the range of technologies and processes it encompasses. We refrain from proposing a singular, exhaustive definition. Instead, we propose an approach anchored in three key pillars: first, the existence of a welfare state; second, processes of digitalization, whether occurring through novel and sophisticated technologies, older forms of digitization or, more commonly, a complex interplay of both alongside enduring manual processes; and, third, the degree of digitalization warranting the DWS label. Instead of fixed meanings, these pillars provide conceptual touchstones for investigating neglected dimensions of welfare statehood in DWS research.

The existence of a welfare state

The concept of the DWS presupposes not only the existence of a state but specifically a welfare state as the institutional context in which digitalization occurs. Yet, the concept of the welfare state itself is a deeply contested concept, open to diverse interpretations and theories (see Offe, 1984; Pierson and Leimgruber, 2021). Some view it as a conscious effort by governments to enhance social solidarity and human flourishing. For example, in his book, Beyond the Welfare State, Myrdal (1960: 45) argues that “in all the rich countries in the Western world, [the State] has become a democratic “Welfare State”, with fairly explicit commitments to the broad goals of economic development, full employment, equality of opportunity for the young, social security, and protected minimum standards as regards not only income, but nutrition, housing, health and education, for people of all regions and social groups.’ More critical perspectives highlight its political function under capitalism, helping to stabilize societies defined by unequal class power (Gough, 1979; Offe, 1984; Soss et al., 2011). In short, there is no universal standpoint from which to evaluate the impact of digitalization, since the very object of its influence – the welfare state – is contested.

Given this conceptual ambiguity, we urge scholars using the DWS concept to be explicit about how they define the welfare state, and to specify the theoretical assumptions guiding their analysis of its digitalization. A rich body of scholarship provides a strong basis for unpacking this concept. Classic analyses, such as Esping-Andersen's (1990) typology of capitalist welfare regimes, suggest that states cluster according to shared structures of social stratification and degrees of commodification. More recent work on welfare states in the Global South further demonstrates that “actually existing welfare states” vary significantly, with studies identifying populist, neoliberal, and residual models (Niño-Zarazúa et al., 2012; Yörük and Gençer, 2022).

This variation raises further questions about how the DWS is conceptualized and whether its application presumes a shared normative foundation and/or historical trajectory that may not, in fact, be universally present. Historically, the welfare state has been tied to a specific path of Western state formation linked to industrialization and the expansion of government responsibility for social welfare in the twentieth century (Øverbye, 2024). Yet, as we will discuss below, many contemporary states have developed extensive digital welfare infrastructures without having followed the institutional pathways typically associated with mature Western welfare states. Further research is needed to explore whether the concept of the DWS adequately accounts for these differences or, alternatively, whether it risks imposing a homogenizing framework that flattens digitalization paths that are actually quite distinct. Empirical DWS research should be both guided by and contribute to theoretically robust frameworks capable of making sense of the complex interactions between digitalization and the evolving functions of welfare systems (if not states).

At minimum, we argue, scholars should clarify what justifies the designation of a DWS with attention to the specific welfare state functions and attributes digitalization is said to transform and situate it within broader welfare state theory. In practice, this requires identifying key welfare-state features and linking them to relevant scholarship to assess how digitalization has reshaped them. Is it, for example, a welfare regime based on low state expenditure, a system of residual state-administered benefits and bolstered by extremely high employment rates, ‘occupational welfare’ and familial support, such as the Japanese welfare state (Esping-Andersen, 1997)? Or, is it a highly centralized and bureaucratic welfare state with social welfare benefits primarily funded from general taxation, such as Australia's ‘wage earners welfare state’? (Castles, 1985). Clarifying the ‘welfare’ dimension of the DWS is intended to support a more rigorous and thoughtful approach to its study, ensuring careful attention to its conceptual content and boundaries. Attending to these underlying features ensures that analyses of the DWS are not abstracted from the institutional, historical, and political specificities of welfare regimes, allowing scholars to better understand not only how digitalization is reshaping welfare provision but also how it may be transforming intrinsic qualities of (welfare) statehood itself.

Defining the ‘digital’ in the DWS

At the most basic level, the concept of the ‘digital’ refers to data that has been converted into a digital form (binary code) that can be ‘read, processed, transmitted, and stored by computational technologies’, whilst digitization is the process of converting analog data into this digital form (Leonardi and Treem, 2020: 1602). The concept of digitalization is more expansive. It captures the ‘socio-technical process of leveraging digital products or systems to develop new organizational procedures, business models, or commercial offerings’ (Saarikko et al., 2020: 828). Or, as Leonardi and Treem (2020: 1602) explain, digitalization refers to the ‘ways in which social life is organized through and around digital technologies’.

In the DWS literature reviewed above, the ‘digital’ refers to a wide variety of technological tools, systems and resources as well as the attendant socio-technical processes of digitalization. This encompasses an extremely broad range of technologies and processes, from relatively simple computerization (e.g., electronic forms and databases) through to novel and sophisticated systems (e.g., AI-driven chatbots, predictive tools and machine learning algorithms), as well as related changes to personnel, policies and organizational processes. As Ball et al. (2023: 1168) point out, this ‘nebulous’ approach to the concept can ‘conceal distinct “varieties of digitalization”, which deserve specific and thoughtful theorization and analysis’.

For others, the ‘digital’ connotes an earlier form of technological innovation associated with the rise of ‘e-government’ in the mid- to late 1990s, coinciding with the rise of the Internet (see, e.g., Alexopoulou, 2025, esp. Figure 1). This included the introduction of government websites to disseminate information, limited online service delivery and the integration of lCTs to support ‘back end’ welfare functions. Many DWSs have since embraced more sophisticated and novel technologies that rely on fast processing of ‘big data’ to automate decision-making and predict or recommend interventions. Recent advancements in large language models, for instance, are being deployed as online chatbots or ‘virtual agents’ to facilitate communication and ‘complement’ in-person services (Germundsson, 2022; Henman, 2019; Meers, 2024).

Hence we have seen the emergence of alternative terms such as the ‘automated welfare state’ (Carney, 2023) or even ‘Welfare 4.0’ (Buhr, 2017), which highlight more recent technological developments but also risk downplaying the continuing importance of ‘simpler’ or older technologies of digital welfare. As Raso (2021: 4) notes, ‘novel tools tend to be layered on top of older information technology infrastructures, such as aged hardware, legacy data, and decades-old programming languages’. Some of the most notorious examples of digital welfare reform have involved relatively simple technologies. One such example is Australia's ‘robodebt’ program, which, while introducing new automated processes, relied on a relatively simple form of data matching that had been in use for decades (Carney, 2018). Rather than sidelining earlier digital tools as relics of a pre-AI era, research on digitalization needs to account for the socio-technical assemblages that emerge from these layered technologies. The concept of the DWS is helpful here in so far as it captures both the variety of technologies used and their non-linear adoption.

A handful of DWS scholars provide more detailed explanations of the ‘digital’ component of the DWS. Meers et al. (2024: 2), for example, consider digitalization and automation as two distinct, albeit overlapping, aspects of digital welfare development. Where the former refers to ‘front-end’ client-facing systems, automation captures the back-end ‘algorithmic factory floor’ (Meers et al., 2024: 2). van Gerven (2022, 252) similarly separates the concept of digitalization (how technology organizes social life) from automation (‘the replacement of human action with a technological process’). To this van Gerven (2022: 252) adds ‘datafication’ as a distinct concept to capture the transformation of social processes into meaningful data.

What is instructive about these approaches is how the DWS term is deployed less as a simplistic catch-all and more as a conceptual umbrella-term describing distinct but connected processes of digitalization. Used in this way, the DWS label facilitates analysis of various components and processes of welfare state digitalization as a whole-of-state, cross-sector and indeed cross-national transformation, while also allowing for nuanced analyses of its distinct forms and local manifestations. The problem so far has been the tendency to use the term either as a broad, all-encompassing descriptor or otherwise as a narrow analytical tool focused on specific, localized instances. But what we are urging is a more integrative approach, one that leverages the term's capacity to bridge macro-level structural changes with the everyday realities of digitalization as experienced in particular settings. Such an approach would link research across different sites without assuming a linear or convergent trajectory. It would also be better positioned to encompass both new technological developments as well as earlier forms of digital innovation, which are often overlaid and cannot always be easily distinguished (Raso, 2021).

When does the welfare state become digital?

At the outset of his report, Alston (2019) declares that the ‘DWS is either already a reality or is emerging in many countries across the globe’. He goes on to discuss examples of digital welfare reform in tens of countries, including Kenya, Australia, the UK and India, each reflecting varying degrees of digital transformation. In highlighting this shift, Alston raises an important, hitherto unanswered, question: when does a particular welfare state become digital? Or, put differently, at what point does the level or type of digitalization justify labeling a particular welfare state as ‘digital’? It seems self-evident that not all welfare states can be given this label. But if we accept that digitalization is a complex, uneven and fragmented process, then the challenge lies in determining the criteria by which we can classify a welfare state as ‘digital’. This calls for careful consideration, as the very process of classifying countries tends to recognize and privilege some features while ignoring others. Absent more robust theorizing and definitional work, as van Gerven (2022: 253) notes, it is ‘difficult for social policy researchers to evaluate what constitutes a DWS and to what extent welfare states have fundamentally shifted into a new state of being’.

In any case, scholars have confidently described particular welfare states as digital. For example, the Danish welfare state is often highlighted as especially ‘digitalized’ and ‘datafied’ (e.g., Bagger et al., 2023; Collington, 2022), with digital strategies emerging as early as the 1990s and significant reforms like the Digital Post mandating that all citizens communicate with the public sector through an online platform (Hjelholt and Schou, 2017). In other settings, ranks and measures of digitalization have become commonplace, from measures of digital inclusion to the UN's E-Government Index (United Nations, 2025), and the European Commission's Digital Economy and Society Index, which includes a ‘Digital Public Services’ indicator (European Commission, 2022).

As is clear from these various indices, digitalization is already being quantified and analyzed through different measures. The question of degree is an important one, as it challenges the blanket categorization of all welfare states as ‘digital’ and instead calls for more precise articulation of the features and processes – both social and technical – that set highly digitalized welfare states apart from those that are less digital. That is, it urges scholars to be explicit about their understanding of what it means for a welfare state to be considered ‘digital’, and thus to more fully elaborate the nature and consequences of this transformation. Since what counts as meaningful measurement varies across cultural and institutional contexts, scholars need to determine appropriate criteria in relation to the settings they study.

In making these judgements, we caution against treating digitalization as something that can be captured solely through quantifying technical developments. Doing so risks privileging technocratic changes over more fundamental transformations in the cultural, institutional and political organization of the (welfare) state. Digitalization is not simply a technological process but also a fundamentally political one. It reconfigures, for example, the mechanisms through which decisions are made and the spaces in which citizens encounter the state. Yet the literature to date has not fully addressed these deeper transformations, with a few exceptions. Dencik and Kaun (2020: 5), for instance, highlight the changing political economy of welfare brought about by digitalization, particularly the ways in which ‘public services are organizing themselves around the disruptive activities of technology companies that over time come to dictate societal arrangements’ (See also Dencik, 2022, discussed above). Schou and Hjelholt (2019: 44) in a similarly critical register explore ‘how contemporary processes of state digitalization are creating new forms of spatialized power and governance’ and elsewhere how these same processes are ‘re-engag[ing] citizens with state mechanisms, all while moderating opposition’ and dissent (Hjelholt, 2024: 653). As Kaun and Masso (2025: 16) contend, the process of automating welfare is itself ‘an ongoing process of social change’, encompassing interconnected shifts to infrastructure, norms and values, entwined with structural and cultural processes of change. What these scholars share, and what we are advocating here, is a commitment to DWS research grounded in robust social analysis, including a deeper engagement with welfare state literature as a means of clarifying the dynamic relationship between technical systems and welfare state institutions in context.

Our core claims in relation to this third pillar can be summarized as follows. First, we make the modest yet often overlooked point that not all welfare states can be considered digital. Even where digital or automated programs exist, if they remain isolated and do not fundamentally reshape core institutional processes, norms or values, the DWS label may not be appropriate or analytically helpful. A second and related point is that we should attend to patterns of state transformation, not just the extent of technological uptake or isolated technologies. As noted above, this calls for approaches that look beyond measuring technological change and instead grapple with the complex and context-specific socio-technical dynamics transforming welfare state institutions and practices.

In this section, we presented three pillars or prompts intended to sharpen how the DWS is defined and applied. Used in this way, the concept represents a useful and adaptable heuristic device (see Swedberg, 2016) with the potential to facilitate richer connections and shared language and foci in what is currently a fragmented field. We now turn to consider the assumptions and potential limitations of the DWS concept, with a view to fostering more reflexive engagement in future scholarship.

Critical reflections on the limitations and assumptions in DWS research

Notwithstanding the flexible and adaptable approach that we advocate above, we recognize that labels carry conceptual baggage, with significant intellectual and political consequences. Even the most inclusive and flexible terms prioritize particular lines of inquiry or modes of understanding whilst sidelining alternative approaches and visions. As such it is important that scholars engage with the DWS concept with an awareness of its implicit assumptions and the ways they might shape and limit interpretation. Here, we discuss three potential limitations embedded in the DWS as a starting point for this reflexive engagement.

The assumption of completeness and linear progress

It is now a well-rehearsed point that processes of digitalization, including those associated with welfare delivery, are neither neat nor linear. Instead, digitalization unfolds unevenly across and between different sectors, countries and scales of state activity (Ingold et al. 2025). Even so, accounts of welfare digitalization often imply, perhaps unconsciously, a continuous, inevitable and unidirectional process. Whilst not wholly unfounded, this assumption risks skewing analysis toward ‘new’ and accelerating areas of digitalization whilst neglecting older technologies, manual processes or even processes of de-digitalization (see, e.g., Redden et al., 2022 on “cancelled systems”). The DWS frame may reproduce similar analytic assumptions. To describe a welfare state as ‘digital’ might imply that its technological integration is locked-in, lending the DWS an air of inevitability.

Centering the (western) welfare state

Second, while the DWS carries the conceptual baggage of the traditional welfare state, it also potentially inherits the limitations of that field of scholarship. Historically, the welfare state has primarily been considered a Western construct arising alongside the development of modern European capitalism, industrialization, and the rise of a self-conscious working class (Kuhnle and Sander, 2010). Global scholarship on the welfare state has generally reflected this Western centrism. The classic and most highly cited studies are largely focused on Western European and North American contexts where welfare states are most entrenched. Dominant frameworks, such as Esping-Andersen's (1990) typology of capitalist welfare states, for example, fail to capture the diverse forms of social protection found in the Global South. Still today, the overwhelming majority of social science research on welfare regimes centers advanced, industrialized nations (Cammett and MacLean, 2011: 2). As a result, research on the DWS risks perpetuating similar hegemonies of Western-centric scholarship.

An important body of scholarship on digital welfare in the Global South is foregrounding alternative trajectories of digitalization (see, e.g., Gupta, 2023; Jaspars and Sathyamala, 2021; Raj and Juned, 2022; Sivaramakrishnan and Pellissery, 2023). This work prompts useful questions about what a welfare state is and how digitalization may transform its nature, particularly in contexts where formal state capacity is limited, and social protection is mediated through informal systems of familial, religious or community support. In these contexts, corporate actors have been shown to perform substantive roles, not merely as contractors or service providers, but as central architects of digital systems (Jain and Gabor, 2020). Their involvement extends beyond implementation to shaping the very structure and logic of welfare provision. As a result, a system may have the appearance of an advanced digital welfare regime – streamlined, data-driven, and digitally integrated – yet beneath this digital façade, it may lack some or all of the features typically associated with (Western) welfare statehood, such as democratic accountability, public oversight, redistributive commitments, and a rights-based orientation to social protection (Dahdah and Alam, 2024; Raj and Juned, 2022). The Indian experience exemplifies this hollowed statehood, revealing a digital welfare system that, while technologically sophisticated, relies on private banks and digital platforms to connect citizens with support (Jain and Gabor, 2020). Its DWS is a hybrid formation that expands the state's reach, but through private intermediaries.

In tracing the diverse trajectories of DWS development, this literature shows that digitalization can generate new and hybrid forms of welfare statehood with characteristics that depart from extant categories in the welfare state canon. Indeed, these studies suggest that the very processes through which digitalization is layered into welfare state architectures and logics may be its most revealing feature, offering important practical and analytical insights. Remaining mindful of the Western bias in conventional welfare state scholarship, we contend that revisiting its core themes provides an essential foundation for empirical, comparative and normative analysis. A deeper engagement with perspectives beyond the Western canon can deepen understanding of how digitalization is transforming not only the technical infrastructures of welfare provision but also the variegated institutional and normative features of welfare statehood.

Centering the nation state as the locus of transformation

Another potential drawback is that the DWS implicitly centers the nation state, redirecting attention from non-state institutions, actors and processes. Certainly, digital welfare reforms have profoundly impacted public sector agencies and the nature of work for state employees at all levels of government (Bovens and Zouridis, 2002). A focus on these formal spaces of state power is therefore justified, especially where digitalization is seen to threaten public sector values and the professional identity of civil servants (Dencik and Kaun, 2020). Moreover, the DWS concept lends itself to consideration of developments in specific national sites. The Danish DWS, for instance, has attracted attention for its distinctive national characteristics, notably, its high levels of digitalization coupled with a social democratic approach to welfare, offering valuable insights into Denmark's model and similar Nordic digital welfare regimes (Collington, 2022; Jørgensen, 2023).

In centering the nation state, however, the DWS concept risks overlooking the power and role played by private actors. This is important because of how digitalization is redrawing state-market boundaries, with private actors increasingly exercising ‘functional sovereignty’ in areas that were once the exclusive domain of the state (Dencik and Kaun, 2020; Pasquale, 2018). van Toorn's (2024a) study, for example, shows how ICT outsourcing arrangements grant private tech companies de-facto decision-making powers in social security and disability benefit determinations. In the Danish context, Collington (2022: 2) similarly shows how ‘the digitalization of public sector administration and services delivery has entailed the transfer of responsibility for critical infrastructures to private actors’. This is not a simple case of welfare state retrenchment. Rather, Denmark's digital welfare strategy aims to repurpose state-based digital capabilities, assets, and datasets created through public welfare programs as resources for private sector profiteering (Collington, 2022; Maguire and Ross Winthereik, 2021). Rather than a hollowed out or privatized state, the DWS is invariably a complex and always-shifting amalgamation of public sector responsibility and private sector profit-making, where state-developed digital infrastructure and data assets are leveraged for corporate gain, often shifting the balance of control and responsibility for welfare.

A further complication with state-centric analyses is that in poorer states with limited public infrastructure, social protection is provided through a diverse array of institutions not formally designated as state or public, such as religious organizations, community groups, international development agencies and NGOs, and informal networks of mutual aid, which often fill critical gaps left by the state. Even in nations in the Global South with advanced digital welfare systems such as India's Aadhaar biometric identification program, non-state actors are vital intermediaries linking citizens to essential services (Singh, 2024). Although they hold no formal state authority, these actors are essential for ‘sustaining the data infrastructure that underlies the organization of India's DWS’ (Singh, 2024: 579), underscoring the limitations of applying state-centric models to such contexts. A related concern is that the DWS concept may center developments at the national scale at the expense of cross-jurisdictional patterns or change at the local level. This may be particularly problematic in places where the digitalization of welfare programs is often localized, such as in Sweden or India.

Reflections and conclusions

In recent years, the concept of the DWS has become a popular shorthand for describing diverse forms of digitalization in welfare state regimes across the globe. In much of the literature, the term has been applied rhetorically without much precision or theory to anchor analysis. This article has sought to advance an approach that facilitates more rigorous and reflexive conceptual engagement both with specific DWS systems and with the term itself. In this final section, we reflect on the utility of the concept both within and beyond the academy.

We suggest that the concept of the DWS can be understood with reference to three key pillars outlined above, which serve as both threshold criteria for the application of the term and as conceptual prompts for the kind of deeper, theory-informed analysis we suggest is needed. Used in this way, the DWS serves as a flexible heuristic that scholars can use to investigate and contextualize particular national and subnational digital welfare systems, employing different theories and methodological orientations. The hope is that these contextualized deployments of the term will themselves deepen, develop and perhaps also complicate the concept. We also see this approach, and specifically our calls for explicit, theoretically rigorous engagement with the concept of, and literatures on, the welfare state, as means to position scholars to offer deeper insights into the under-theorized question of how welfare states have changed in the digital era. It may also help address the tendency in strands of the literature to focus narrowly on technological transformations or the reorganization of welfare state ‘tasks’ (van Gerven, 2022) in isolation from the broader dynamics of welfare state change. The intellectual value of this approach lies in its capacity for enriching our understanding of these broader patterns and transnational parallels between welfare states, without conflating or erasing distinct local experiences.

We also note the pragmatic appeal of the DWS label due to its accessibility and relative normative neutrality. Compared to its conceptual neighbors, including the datafied and automated welfare states, the DWS label is arguably more immediately intelligible, lending it purchase both within and beyond the academy. This could explain the term's broader uptake, which extends across scholarly fields and into public arenas including the media and human rights discourse (see, e.g., Danish Institute of Human Rights, 2023; McCully, 2020; Pilkington, 2019). The DWS represents a useful shorthand for naming a wide-ranging suite of socio-technical systems and developments, old and new, while remaining open to detailed and sophisticated applications of the concept.

For this reason, the DWS label holds promise for human rights and social justice advocates seeking to scrutinize and contest how digital technologies are deployed within welfare systems. Whilst its recent popularization has often been tied to critique, most notably the United Nations’ stark warning of an impending ‘digital welfare dystopia’ (Alston, 2019), the concept itself remains relatively neutral. It does not reflect or prescribe any particular stance on welfare. Nor does it imply any inherent value or harm in digital technologies. Rather, these questions are left open to empirical and normative analysis. Even in Alston's (2019) scathing critique, his diagnosis of the problem lies primarily in the purpose behind digital technologies, not in the fact of a digitalizing welfare state, per se. Alston takes aim at governments’ ‘obsession’ with ‘fraud, cost savings, sanctions and market-driven definitions of efficiency’ against the prioritization of human rights and social justice (Alston, 2019: 20). Implicit in this critique is the possibility of a just (or at least benevolent) DWS of some kind. From this perspective, the DWS label allows scholars, commentators and advocates to assess the merits of specific digital initiatives without committing to a wholesale endorsement or rejection of digitalization. This is not to celebrate political neutrality, but to underscore the value of a concept that can frame inquiry while remaining analytically open.

In this article, we have proposed an understanding of the DWS as a heuristic tool for cross-disciplinary and cross-national analysis. The approach we advocate involves more care in applying this label to specific sites and urges deeper engagement with especially the literatures and theories of the welfare state. This approach to the concept, we argue, better positions scholars to interrogate the changing architecture and norms of the welfare state under digitalization. We also call for greater reflexivity in the use of this term, which includes being mindful of its hidden assumptions and conceptual limitations. By unpacking the DWS concept, including its analytical affordances and limitations, we provide modest starting points for this more reflexive mode of scholarship, in the hope that they will advance richer discussion of both specific DWS systems and broader patterns of digital welfare transformation.