Digital welfare: designing for more nuanced forms of access

Abstract The aim of many forms of digitalized welfare is to offer a personalized, holistic service that is affordable, sustainable, efficient, encouraging and leaves room for voluntary action. We argue that for these goals to be achieved, consideration has to be given both to the design of the system delivered by the welfare provider and to the ecosystem that further shapes the experience of the system. In such an ecosystem not only should state-provided welfare be considered but so too should community support, as well as alternative methods of accounting for societal contribution. In this paper, we use theoretical perspectives on access and security to ideate sketches that invoke new user experiences of welfare. These sketches reflect the importance of both designing for and understanding the ecosystem in which welfare systems are accessed, in order to articulate a different welfare ethos that can encompass both complementary and conflicting perspectives. Using the ideas of Buchanan (1992, 2001a) our synthesis of theories related to security with the practical implementation of digital welfare aims to shape the placement (Buchanan 1992) of digital welfare by embedding access points of different types further into the welfare ecosystem.


Introduction
The early 21 st century has seen digitalized welfare (welfare policy delivered through technological means) implemented in a number of countries, most notably in the UK and the USA, with other countries such as Australia duplicating their initiatives. Digitalized welfare is typically part of a wider program of welfare reform. In the UK, welfare reform was set out in the Welfare Reform Act (DWP 2012) and digitalization formed part of a program to simplify welfare and make the process of claiming welfare more transparent. A key critique of this program relates to the lack of access opportunities and the difficulty of access to the digital welfare system: both in terms of the methods of access and the conditions on which access is granted. The issue of access was highlighted when the UN Special Rapporteur for Extreme Poverty and Human Rights, Philip Alston (2019a), led a study in the UK that sought evidence from civil society, individuals, researchers and government. The digitalization of welfare systems puts technology in the spotlight and shines a light on the ways in which technology is used to regulate access. In this paper, we examine ways in which the notion of access might be re-designed when delivering digital welfare to meet both the security needs of the welfare provider and the individual.

Universal credit: an overview
The focus of the UK's welfare reform program is to, "make the benefits and tax credits system fairer and simpler by creating the right incentives to get more people into work" (Department for Work and Pensions [DWP] 2010). The program can be traced back to a series of reviews and proposals, in particular the green paper "A New Deal for Welfare: Empowering People to Work" (DWP 2006) and the green paper, "No-one Written Off: Reforming Welfare to Reward Responsibility" (DWP 2008).
Not only were the reforms intended to simplify and make the system more efficient but also to change the conditions upon which access to welfare was granted. The DWP (2010) published a white paper which laid out information about the upcoming plans for the reformed system of welfare. In this white paper, the Secretary of State for Work and Pensions, Iain Duncan Smith stated, "too much of our current system is geared toward maintaining people on benefits rather than helping them to flourish in work," citing the need for, "reform that tackles the underlying problem of welfare dependency." These proposals, "firmly placed conditionality and responsibility at the heart of welfare policy" (Reeve 2017).
Universal Credit was the name given to the new payment system for benefits. Not only was the welfare system reformed but its method of delivery was also re-designed to achieve the twin goals of simplicity and conditionality. Accordingly, claimants of Universal Credit will typically encounter the following process: Claimants apply for Universal Credit on-line in all but a very few exempted situations. In order to apply, claimants must access the government website and input their bank details, an email address, information about their living situation, details of their income, details of savings and any details about childcare responsibilities. Once the application has been made, there is a claim interview where it is determined whether the claimant is required to look for work and the support that the claimant will be given to become work ready. The claimant will typically be asked to sign a Claimant Commitment, where the conditions for receipt of welfare and the claimant's responsibilities are specified. The conditions and responsibilities are a personalized commitment calibrated to the claimant's situation, taking account of family responsibilities, health and potential earnings.
Evidence of compliance with the conditions is subsequently submitted by the claimant primarily using on-line means.
There are different categories of requirements or "conditionality" that claimants are expected to meet. As the Public Law Project (2020) describes in their leaflet for welfare claimants the main conditions are as follows: Some people will have no work-related requirements. This could be because they are already earning enough money from work, or because of an illness, disability or impairment, or because they have caring responsibilities for a disabled person or a child under 1 year old. Others will only have interview requirements, or this may be combined with work preparation requirements. Which one applies to you will depend on your circumstances, such as childcare responsibilities or whether you have limited capability for work. Some people will have all work-related requirements as a condition of claim. This means that as a general rule, the DWP will impose a work search requirement and a work availability requirement on all claimants in this group, and may also impose interview and/or preparation requirements too.
If claimants do not meet their work-related commitments, their Universal Credit payments may be reduced for a set period of time, this is known as a sanction. Sanctions are escalated if the system detects what it understands to be repeated or extended noncompliance. Sanctions are a principle means of controlling access to welfare. The "Review of the DWP Conditionality and Sanctions" by a task force led by Salford Council for Voluntary Services (2014), states that sanctions, "ensure the system is fairer for the taxpayer" and that "seventy-two per cent of claimants say they are more likely to follow the rules due to the presence of sanctions." 3. Universal credit: Access and security DWP has been clear that its motivation for a digital-by-default system is to, "improve outcomes for society, make DWP more efficient and effective, re-imagine customer experiences" (Prakash 2017). Reliance on digital delivery, however, raises fundamental questions about the accessibility and availability of Universal Credit to those who need it most. As Alston (2019b) points out, digitalized welfare assumes that individuals will not only have access to documentation but will also be able to upload it digitally; that they will not only have a credit history but a "broader digital financial footprint" to accompany it.
Access to digital welfare is often supported by the wider social and relational networks in which a claimant sits. For example, The Huffington Post (UK) reported that 462,000 people "required help from friends, family, the Job Center or a charity to apply" (Youle 2019) a problem, that as Murugesu (2019) noted in the New Statesman, often results in front-line library staff assisting claimants. However, access is not only controlled by the technical access to the service but also by a risk assessment of the likelihood that an individual will defraud the welfare system. The DWP (2011) piloted Risk Based Verification (RBV) with a number of local authorities, and as Torbay Council explains on their website: A risk profile will be given to each customer, determined by proprietary software using statistical information and risk propensity data gathered over many years about what type of claim represents what type of risk. The higher the risk, the greater the checks used to establish that the claim is genuine. (Torbay Council 2011) The RBV system assigns claims based on low, medium or high risk, and, "low risk claims are processed more efficiently, with greater attention being given to the medium and high-risk cases" (Chichester District Council 2017). A Local Authority Insight Survey states that increased verification for high-risk claims include: credit reference agency checks, visits to the claimant's home, increased document checks, meeting claimants and phone-calls, as well as follow-up telephone calls (DWP 2017). As RBV shows, the basis of access control decisions rests on the interpretation and implementation of the principles of conditionality and responsibility.
Conceptualizing the system as solely technological and focusing the understanding of access on the control of access, gives rise to many of the critiques of Universal Credit (Dwyer and Wright 2014; Morris 2020; Pantazis 2016; Reeve 2017) One means of responding to these critiques might be to design for a broader notion of access where access control can be maintained but attention can also be given to the wider access goals of welfare. Ribot and Peluso (2003) theory of access offers a framework through which different aspects of access can be brought together. Their theory conceptualizes access as the ability to derive benefit from something and as, "bundles and webs of powers that enable actors to gain, control and maintain access" (Ribot and Peluso 2003). This push toward a broader conceptualization of access brings traditional notions of technological system security through access control into dialog with the broader conceptual framing of security that considers security both in terms of enablement and emancipation of people as well as protection from harms (Roe, 2008). In the case of welfare, it is important that not only the system and the financial aspects of welfare are protected but also that the claimants are enabled to achieve employment and make themselves financially secure.
To both secure the welfare system and to enable the claimant to achieve a state of financial security, multiple logics of security must be brought together. Doty (1998) identifies three main security logics or means of framing security: national security logic, societal security logic, and human security logic. National security logic is a traditional, state-centred conceptualization of security where security is framed as the absence of external threats to the national or state territory. Societal security logic foregrounds identity politics and society as a site of security but it also looks to the state as a site of security because society resides within the protection of the state. Human security logic is a human and individual-centred conceptualization of security. It is a form of security perceived as a desired good which enables the pursuit and enjoyment of a good life. It is a logic of inclusion, which transcends state boundaries and binary conceptions of identity.
We argue that the lived experience of digital welfare might improve if the welfare system could be designed to articulate these multiple logics of security. In the following sections we sketch how digital welfare could be reconceptualised as a system that exists within a wider welfare ecosystem. In such an ecosystem, not only is state-provided welfare considered but so too is community support, as well as alternative methods of accounting for societal contribution. As the sketches show, the benefits of access can be many and varied if the wider ecosystem is considered and these benefits can be realized in different parts of the system. The consideration of an ecosystem of welfare enables policy change and welfare re-focus. At the time of writing, during the COVID-19 crisis, questions of welfare focus and the benefits that are needed are the subject of much debate across many countries (Rayner 2020). Conceptualizing welfare as an ecosystem and considering access to welfare through several points in that ecosystem offers flexibility for policy change as well as enabling a closer alignment with the claimants' lived experiences.

Designing for the lived experience
The UK's focus on digitalized welfare is not unique. As has been noted by Millar and Whiteford (2020) while there are differences, there are also many similarities in the way Western nations are implementing new welfare systems. In particular the UK, Australia and the US each have a focus on digitalized welfare. The welfare models are similar too and both the UK and Australia aim to provide support for greater self-sufficiency, limiting government expenditure and encouraging independence (Hinton 2018). The Australian Anglicare Report (Hinton 2018) on the Australian welfare system is interesting in its attempt to document the lived experience of claimants in Southern Queensland, Tasmania and Western Australia, as well as community support staff who often act as the bridge between claimants and the welfare process. The similarities between the Australian and UK systems mean that we can assume perhaps some similarity in the lived experiences of claimants.
In Australia, the Welfare Payment Infrastructure Transformation Program (WPIT) is part of an initiative to deliver better management of the social welfare system. A particularly contentious issue in the program is the Online Compliance Intervention (OCI). OCI is referred to colloquially as Robo-Debt because of the way it uses automated data matching to determine if claimants have been overpaid and issues automated initiation letters to alert claimants and to start the claim-back process. The system matches income declared to the Australian Tax Office (ATO) against records of income held by Centrelink (the Australian government department responsible for issuing benefits). Problems arise because Centrelink records are based on fortnightly income and the ATO record a lump sum for each financial year which is then averaged for comparison with the Centrelink records. While the need to claim back overpayments is not disputed the process of data matching causes a high number of discrepancies which can then only be challenged through an online portal or a phone service.
An Australian report produced by Anglicare about the Australian digital welfare system suggests an alternative, more inclusive welfare could be achieved through, "a system that could acknowledge and work with the reality of people's lives, that was respectful and that 'cared' about its customers" (Hinton 2018). This report further highlights that the current Australian system presents economic barriers resulting in a lack of access to a computer or smartphone, lower levels of digital literacy and lower quality broadband access (Hinton 2018). Economic barriers also mean less access to transport necessary to meet the demands of conditionality or mutual obligations, in turn this can lead to claimants driving, "uninsured, unregistered and unlicensed" (Hinton 2018). The move away from face-to-face interaction and a system founded on human relationships toward a system of digital interactions and tasks means that an important element of identifying acute needs of claimants disappears (Hinton 2018). It is clear that similar arguments may be made about the delivery of Universal Credit.
In the following section, we explore how it might be possible to change these narratives within the lived experience of the present system. We do this by re-framing the meanings of access and security in a way that enables different logics of access and security to coexist within the welfare ecosystem and promoting different articulations of welfare.

Possible futures in dialogue with the lived present
UK government's Open Policy Unit has developed an approach and toolkit for collaborative and participatory processes for co-designing public policy (HMG 2016). Design thinking is an important means of both developing an understanding of the lived experience of communities and of producing responses that support and promote positive versions of that lived experience. A fundamental element of design thinking is the use of abduction to synthesize ideas from disparate sources and disciplines (Kolko 2010). Innovation occurs through, "inference or intuition" (Kolko 2010) as a result of synthesis. Using the ideas of Buchanan (1992Buchanan ( , 2001a our synthesis of theory with the practical implementation of digital welfare aims to shape the placement (Buchanan 1992) of digital welfare. In Buchanan's (1992) definition placements have boundaries that, "shape and constrain" mental models but they are also fluid and can offer a way of reframing people's mental models. Using Doty's (1998) theory of security logics and Ribot and Peluso (2003) theory of access we use theory as a provocation that allows the ideation of new placements that enable digital welfare to be reframed and the lived experience improved.
Design says something that goes beyond language (Buchanan 2001a) and a product or service says something to those who understand it so that, "all productsdigital, analog, tangible, intangibleare vivid arguments about how we should live our lives" (Buchanan 2001a). We offer three sketches as a way of, "shifting semantic perspectives" (Kolko 2010) to create new placements for notions of access and security and by doing so, we speculate how digital welfare may invoke more positive lived experiences for claimants. Buchanan (2001a) goes on to distinguish between the classical themes of rhetoric linking them to aspects of design. Logos is the rational argument underpinning the design, pathos is about the affordances of the design and ethos is the, "voice, character or personality of the product" (Buchanan 2001a). Using Buchanan's categorization, the sketches address the ethos of digitalized welfare. The personality of digitalized welfare prioritizes the values of efficiency, employment and self-efficacy, and technological security controls enforces such prioritization using access control to protect the system and the integrity of the welfare state. Challenging these values by engaging with the complexity and the contradictions of lived experiences demonstrates the opportunity to reshape the ethos of digitalized welfare.
Each sketch suggests a design that enables the voices of marginalized and underserved communities to grow stronger, a call that is often made through participatory design, as exemplified in the work by Hillgren, Seravalli, and Agger Eriksen (2016).
Re-casting the values of the welfare system so that efficiency, employment and self-efficacy are matched with values of solidarity, reciprocity and care creates new placements of access and security that are located within the wider welfare ecosystem. Framing welfare as a form of care acknowledges the potential for conflict but also provides capacity for working with, rather than against, the breakdowns of everyday patterns and interactions (Hall, Coles-Kemp, and Heath 2018).
The three sketches encourage us to anticipate how welfare and reflect on the design of welfare policies and systems of the present.

Feminist conceptions of labor
To reimagine a caring system of welfare, society needs to be redefined to recognize the unpaid labor that props up society, predominantly carried out by women (Andersen 2019). As has been discussed, currently Universal Credit requires work preparation and job-search requirements of lone parents. This erasure of care work is deeply gendered and, in particular, encodes single mothers as a, "moral and financial risk" and reinforces visions of the, "private, self-sufficient family unit" (Cain 2016). An alternative system would, firstly and most importantly, reconfigure the digital design of welfare to include caring labor in the calculation of hours worked by claimants and include pay for said labor in the benefits package received.
Online community hubs could be used to support this process where a particular emphasis could be placed upon supporting parents (especially lone parents). One way this may work is if claimants can use the hub collectively organize skills-swapping for caring and domestic labor. For instance, one individual may have work and be unable to pick their child up from school, and another may be unable to go food shopping due to an appointment at the welfare office. The former claimant could take a cooked meal to the latter claimant's house, in exchange for their child being picked up from school. Additionally, the hub could be used by claimants to make extra money in their benefits package, if they perform caring labor for recipients unable to access technology (for example, taking food shopping to an elderly service user) or for users who do not wish to skills swap. Again, this could be overseen by frontline staff at the welfare office and be subject to checks, but the goal would ultimately be to shift the rhetoric on care from individualized responsibility to valued community work, that benefits both state and society.

Digital communities
The current system of digitalized welfare is cold and faceless, and creates immense frustration and desperation amongst claimants (Hinton 2018). However, welfare technology could be reimagined from a point of compassion and care, with the central goal to support vulnerable members of society. This draws on Doty's (1998) interpretation of security as "security through development, not through arms, security of all people everywherein their homes, in their jobs, in their streets, in their communities, in their environment." For instance, welfare centers could have online communities which function as social platforms for claimants to meet and support one another, as well as receive guidance from frontline staff. Websites and apps may act as tools for skillssharing and swapping, as well as posting guidance and tips for navigating the system. This may be led by frontline staff who could perform tasks ranging from responding to queries to organizing group activities for service users in the local area, in order to combat isolation (for example, art or cooking classes). Additionally, welfare centers could run paid courses to train claimants in using the digitalized service, and once qualified, they might be employed to make home visits for claimants who have little to no access, and assist them in navigating the online welfare service and engaging with the community hub. Such a reframing of welfare digitalization repositions claimants as part of a creative web of social support, rather than an atomized individual being punished for their "work shy" behavior. Doty (1998) notes that definitions of security can be expanded to focus on the welfare of individuals in all areas of their lives, which will benefit state and society. At the heart of many digitalized welfare systems are a deep-rooted mistrust of claimants, meaning that welfare technologies have led to a culture of cruelty that leaves vulnerable individuals disenfranchized, isolated and excluded. We draw from Bales (2013) to suggest that digital welfare services can be reframed around the conception of formal solidarity, which centers on the, "collective contributions of individuals" to improve communities and the, "living conditions of the poorer members" of society. This involves a dramatic shift away from individualized responsibility and suspicion toward a system built upon empathy and care. One way this can be exercised is through wider participation in the design process of digitalized welfare, where government and local communities work together to develop inclusive systems either before they are implemented or as a response to unintended consequences after implementation. Ehn (2008) examines these approaches in depth, drawing on Redstr€ om's (2008) definitions of participatory design as "use-before-use" and meta-design "design-after-design." For example, government officials, digital designers, activists, charities, academics and community members could organize workshops for welfare claimants to re-imagine digital welfare. This may involve community learning about the algorithmic make-up used for Universal Credit, wherein recipients and service designers co-construct digital welfare as a mutually beneficial service to the state and society, potentially through creative visualization methods (Heath, Coles-Kemp, and Hall 2014). Not only would such an approach assist in the articulation of empathetic digital welfare, wherein the thoughts and experiences of individuals impacted by Universal Credit are valued, it would enable claimants and community members to re-assert their identities, and deconstruct the violent boundaries of the, "deserving" and "undeserving" poor that are central to current welfare practices (Doty 1998).

Reflections
The expansion of digital welfare is partly driven by the desire of governments to rationalize, simplify and deliver policy through the use of technology but there are unforeseen consequences of this turn to technology and data-driven decision-making. In our three sketches we consider how digital welfare can be designed around human relationships facilitated by civic technology to augment and, in some cases replace, a centralized digital welfare system. Human relationships become a key site of security and the security principles are of care, solidarity and community. These sketches do not replace the conditions of welfare access and the methods of delivery but augment and filter these conditions with community-supported forms of access that enable additional forms of benefit and recognize a wider range of labor. In each sketch, the proposed ethos of the welfare system re-frames the system by embedding the access points within the community. Furthermore access is framed as forms of benefit that meet the needs of the vulnerable in society, thereby clearly addressing issues of human and societal security without undermining the need to digitally protect the welfare system.

Reflections for design
Design should be a negotiation between the, "intent of the designer and manufacturer and the expectations of the communities of use" (Buchanan 2001b). Our sketches aim to give voice to a response to the efficiency ethos of the existing digitalized welfare system. The first and second sketches speak across the bureaucratic divide, with the first using the rhetoric of labor to construct a broader placement than paid labor to encompass the value of unpaid labor within the home and communities. The second uses the rhetoric of access to construct a broader placement than technological access controls to demonstrate how claimants can support each other. The third sketch is located in the rhetoric of participatory and meta-design described by Ehn (2008) making an argument for involving the claimants in the design of the system. Through this use of theory as provocation we have aimed to move some way to closing the gap that often exists between theory and practice and to start to develop a, "pathway for bringing theory into a closer relationship with practical action" (Buchanan 1992) not solving the problem but at least starting to, "constructively deal with disagreements" (Ehn 2008). By using such methods, we hope to develop constructive and practical ways to, "bring the structuring of inequalities into view" (Julier and Kimbell 2019).

Reflections for policy practice
What can policy makers take from this use of theory as provocation for design thinking in digitalized welfare? In making the suggestions below we take inspiration from the recently produced "Government as System" toolkit (HMG 2020) and Kimbell (2015).
We demonstrate the potential value in understanding the lived experience of welfare claimants beyond their use of technology. We've aimed to do this through imaginative sketches that facilitate abductive reasoning that can lead to insight and creative problem solving. Another way of doing this would be to capture stories from claimants that span both their use of the digital welfare process, the technology and their wider lived experience.
The sketches offer an expansive view of the possibilities of digital welfare rather than a reductive view focused on the scope of the technological system. They offer a way of imaginatively prototyping ideas that policy makers could then use to form the basis for more structured modeling or experimentation. Our sketches suggest the benefits that could be realized on both sides of the bureaucratic divide through the use of participatory and co-design so that both policy makers and claimants gain. For a digital welfare system that already exists the use of theory as provocation opens a space for policy intervention that could underpin the redesign of the current system as the sketches that emerge give a voice to claimants. This space is an opportunity for policy makers to pause and refocus to ensure that digitalized welfare services are human-centred and to engage in actively and imaginatively addressing flaws and limitations.

Conclusion
These sketches offer a view of welfare that counters a vision of digitalized welfare that positions claimants as potential miscreants and subjects them to technological security controls. They instead articulate a system that is situated in the everyday lives of claimants and that connects the state to the lived experience of claimants so that claimants feel both seen, understood and cared for.
The technology present in these sketches is used to support a welfare ecosystem that further includes welfare claimants in society and fosters empowering spaces. Sketching alternative futures of welfare in this way encourages us to question whether the current security logics of digitalized welfare are inevitable, necessary and appropriate. Crucially, creating such sketches offers not only questions but also alternative responses and thereby responds to aspects of the negative lived experiences articulated in the critiques of digitalized welfare.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by Engineering and Physical Sciences Research Council (EPSRC) [Grant EP/R033382/1].