How the Responsibility of Digital Support for Older People is Allocated? The Swedish Welfare System at the Crossroads

A great welfare challenge today is to promote opportunities for greater digitalization, while limiting social inequalities from digital divides, especially for older people. While the digital divide is a dynamic problem, shifting from physical access to skills and usage, public policies to close the divide do not necessarily follow. This study explores who is providing digital support in Sweden by looking at three institutions: (1) the municipal eldercare system, (2) popular education institutions, and (3) the family. The results show that the Swedish policy relies heavily on popular education and family arrangements, leaving many young-old Swedes in need of digital support without public support, while the opposite occurs for very old Swedes who are mostly consumers of welfare technologies. Issues of dependency or the other way around arise. Given this, the role of the Swedish welfare state, which sets the tone of the Swedish welfare regime, needs to be re-evaluated, especially in light of the demographic challenge (a growing number of older people).

RASP -Research on Ageing and Social Policy 10(1) 50 opulation aging is a problem that affects all welfare regimes and a way to solve it is through technology. Accordingly, in 2015, the Swedish government launched the "Digital First". Digital First marked a radical shift in the view of service and information exchange in the state and municipal sectors. One risk though is that many welfare service users are members of communities most likely to be digitally excluded, such as the old people.
Despite increasing digitalization across European welfare regimes, few scholars have explored how digitalization affects welfare institutions and their relations with citizens (Hansen et al., 2018, p.67). All too often both scholars and policymakers have depoliticized digitalization, treating it as a merely technical issue and downplaying its political content and consequences (Hall, 2008;Löfgren & Sörensen, 2011). This neglect is especially pronounced in the context of social exclusion and marginalization. Although researchers have explored the digital divide since the 1990s (Norris, 2001), very few have analyzed its intersections with welfare institutions. Digital divide research and digital government research have evolved without much interaction (Helbig et al., 2009).
While country-specific digital policies are developed in political systems that belong to distinctive types of welfare regimes, the digital policy is not usually analyzed from a welfare perspective (Alexopoulou, 2020). This study addresses this gap by analyzing the case of Sweden. The Swedish government aims to make Sweden "the best country in the world when it comes to taking advantage of the opportunities of digitalization" (Regeringskansliet, 2017, p.6), and simultaneously claims to be "best in the world when it comes to digital inclusion" (SOU, 2014, pp.13-11). Hence, Sweden is an influential case that provides ample opportunities for understanding the welfare challenge of combining these intertwined goals. This paper will address the following research question: How is the responsibility for digital support for older people allocated in the Swedish welfare regime? For the design of this study, it was very influential another academic paper under the title "Welfare states do not crowd out the family: evidence for mixed responsibility from comparative analyses" (Motel-Klingebiel et al., 2005). The key difference is that we conducted a case study, which explores the various institutions, which are supposedly responsible for solving the digital divide problem in Sweden or investigate the phenomenon P 51 The Swedish Welfare System at the Crossroads -Alexopoulou & Åström at hand and how successful they are in their task. In our view, the Swedish welfare structure has an impact on the digitalization process and the kind of digital support offered to older people, something that hasn't been explored before.

Understanding the Grey Digital Divide in Sweden
The question of whether increased digitalization will increase inequality over time remains important and deserves the attention of every welfare regime (Esping-Andersen, 1990. The persistence of inequalities treats democracy as a whole and citizenship including the digital aspect of it (Mossberger et al., 2014). However, the digital divide, unlike more inequalities in housing, welfare, health, or educational attainment, has remained invisible in terms of social justice (Bustillos, 2017).
At first glance, the digital divide might seem to be closing in Sweden. In 2020, up to 96 percent of the Swedish population stated that they used the Internet and 98 percent similarly said that they had an Internet connection at home (The Swedish Internet Foundation, 2020). Also, the oldest age group is the one that is increasing its internet use most: nearly three out of four (73 percent) Swedes 76 years or older use the Internet today; ten years ago, it was just under one in four (23 percent). However, the digital divide is changing rather than closing. The divide is no longer between those who do and do not have access. It is not between Internet users and those who never go online. Davidsson et al. (2018) reported that those who do not or rarely use the Internet in Sweden more often live in rural communities, have a lower education level, have lower income, and are female. Rather, the divide is between frequent users with great skills and vast opportunities and infrequent users with significantly more limited skills and opportunities (The Swedish Internet Foundation, 2020).
Older people in Sweden are at risk. According to the Internet Foundation, many older individuals in Sweden are still on the wrong side of the second (skills) and third (opportunities) divides, and therefore risk "digital exclusion" (The Swedish Internet Foundation 2014. Helsper and Reisdorf (2017) noted the risk of a "digital underclass" developing, partly RASP -Research on Ageing and Social Policy 10(1) 52 because the relatively few Swedes who are not digitally involved are concentrated in already vulnerable groups and the strong digitalization norms in society make non-use more excluding.
With digitalization as a means of welfare state restructuring, access to physical welfare services is deteriorating. Contacts between citizens and the public sector, which previously relied on paper forms or interpersonal communication, are moving to online platforms (Schou, 2018). Digitally disengaged citizens risk not receiving welfare services to which they are entitled. New technology is adding a new layer of prerequisites for achieving functional ability and autonomous living, which are key elements of wellbeing and quality of life, especially for older citizens (Siren & Knudsen, 2017).
At the individual level, digitalized services of any sort could support older adults' social participation through activities that they think relevant in their lives, and perhaps facilitate them to live longer at home (Fischl et al., 2020), but this remains a 'far-reaching dream' for the 'digital laggards'. The only way for older people to find access to digital means is "borrowed access" (Reneland-Forsman, 2018). Otherwise, they are completely deprived of innovative solutions and technologies that can support their well-being, independence, and health (Fang et al., 2019).
In 2020, the Swedish Internet Foundation asked a random sample of the Swedish population: If they were asked to conduct various activities on the Internet, would they do these themselves, or ask for help? According to the results, age is the factor that stands out regarding the need for digital support.
As Graph 1 shows, the need for support is substantial and increases gradually with age. For instance, nearly six out of ten in the 76+ year group stated that they would need help creating a mobile banking ID and opening an e-mail account. Between 37 and 46 percent would need help buying an item online, paying bills, booking and paying for tickets, and booking a doctor's appointment online. Overall, the results clearly show that access is not the same as skills or opportunities and that many older people risk digital exclusion.

Digital Policy and the Swedish Welfare System
This section explains the connection that exists between a given welfare regime and the produced digital public policies from an institutional perspective. To understand the public policy regarding the grey digital divide, we turn to institutional theory (e.g. Powell & DiMaggio, 1991;Fountain, 2001;Agre, 2002;Yang, 2003). While digital policies are defined as "anything a government chooses to do or not to do" (Dye, 1972) in this field, the institutional perspective emphasizes how the embeddedness of government actors in cognitive, cultural, social, and institutional structures influences policy.
On one hand, this means that institutionalists often question the celebratory digital discourse on participation, inclusion, and usage, keeping open "the possibility that the world could be different than it is" (Calhoun, 1995, p.290). On the other hand, as Barber (2001, p.43) suggested, institutionalists often recognize that "new technologies tend to reflect rather RASP -Research on Ageing and Social Policy 10(1) 54 than to alter the culture that produces them … technology cannot save us from ourselves, it can only reflect all too candidly who we are". The institutional approach has the potential to add new knowledge to research on the digital divide, which usually works closely with users to learn about their preferences, capabilities, and interests, but tends to ignore that policy and programs change and evolve within bureaucratic environments (Helbig et al., 2009).
Country-specific digital policies are developed in political systems that belong to distinctive types of welfare regimes. Sweden has one of the world's most generous social-democratic welfare regimes (Esping-Andersen, 1990. Eldercare services in Sweden are universal: relatively comprehensive, mainly publicly financed, and available to all citizens according to need rather than their ability to pay (Szebehel and Trydegård, 2012). However, the number and quality of these services differ considerably at the municipal level. Elderly care is provided in Sweden by local governments and is supplemented by the family and other institutions, such as the third sector (Montin, 2015).
The main responsibility of county councils is healthcare and hospital care, while municipalities provide almost all home help services, institutional care, transportation services, security alarms, meals-on-wheels, and other services for older people and the handicapped (Sundström et al., 2002). Many changes occurred since the 1990s. As a result, uniform public services are no longer the case; instead, the new services are based on a mix of private providers and service user choices, while the relations between the various actors follow market-like dynamics (Blomqvist, 2004). According to Blomqvist (2004, p.148), the introduction of NPM ideas modified: the culture of the Swedish elder-care sector, which is evident not least in the new "economistic" language used in this sector even by the authorities, where care is referred to as "products" and the elderly as "consumers".
Another side-effect of the marketization/privatization is that fewer people receive public care, with the focus being directed to those most in need of support (Blomberg et al., 2010) and not to the whole population. Additionally, older individuals should first get support at home, and only when it is necessary to move into care facilities (Sobis, 2013) and always along the lines of stricter selection criteria (Grassman, 2014). But what has happened to the 55 The Swedish Welfare System at the Crossroads -Alexopoulou & Åström institution of the family and what is the role of technology today in front of growing demands for older people?
In Sweden, it is rare for older parents to live with their adult children, as occurs in other countries with more familialistic welfare regimes. Adult children take care of their parents, but parents live in their own homes or, less commonly, in nursing homes (Grassman, 2014). Official Swedish social policy anticipated that the institution of the family would respond by filling the gaps in formal support provision fueled by the service's cutbacks in the 1980s and 1990s (Motel-Klingebiel et al., 2005). This assumption was not entirely well-grounded, because some older people do not have families, or their families have abandoned them.
To complicate the issue further, shifts also occur due to demographic changes, which take place in Sweden. A potential way to resolve this intriguing "welfare Gordian knot" is by using the possibilities offered by technology, including "welfare technologies". These technologies are "associated with the promise of an improved public sector with services to support patients and independence as well as cost-effectiveness and improved working environments for healthcare and social care professionals" (Frennert & Baudin, 2019, p.1).
Older people and their helpers in Sweden (e.g. nursing personnel or family members) are expected to use digital tools. In the health sector, a new kind of citizen has emerged: "the digitally engaged patient" (Lupton, 2013). This sort of mentality goes hand in hand with the active aging concept that proposes a shift to more active citizenship in which individuals must remain self-reliant, independent, and flexible for their entire lives (Jensen & Principi, 2014) or, more concretely, must keep their aging bodies busy for as long as possible (Katz, 2000).
Active aging ideas seem to have a close affinity to the neoliberal notion that recalls that the state should interfere to a minimum and individuals should take care of themselves. Active aging policies are considered to be largely shaped by the welfare context in which they are implemented (Ney, 2005) and we suppose that the same happens with the digital policies having as a target group older people in Sweden.

Material and Methods
The empirical inquiry is designed as a case study of digital policy concerning the grey digital divide in Sweden. As noted by Yin (1981Yin ( , 1998, the case study is particularly suitable for empirical inquiries into contemporary phenomena in their real-life contexts, where the boundaries between phenomenon and context are not obvious. Our case study is mostly descriptive and exploratory. The fragmented nature of the digital divide in the Swedish context (many actors in comparison to other countries) made inevitable the use of different empirical sources for this paper. Therefore, for the first institution, eldercare, we primarily considered national policy documents such as the Swedish Digital Strategy and eldercare policies, reports, and statements. For the second institution, popular education, we mainly relied on evaluations, reports, and interview data. For the third institution, the family, we based our analysis primarily on two surveys of the SPF Seniorerna (2019) and Internet Foundation in Sweden (2020), complemented with interviews.
Fifteen interviews, conducted from May 2019 to March 2020 with various authorities from the public and third sector in Sweden (see table 1, In addition to the interviews, we also conducted participatory observation at the public library in Örebro responsible for the First Help program and at an elder's house in which two members of the IT-Guide organization helped seniors use new technologies to understand how IT assistance is provided to older seniors. After the interviews were completed, they were immediately transcribed in order not to lose the 'general feeling' that the interview transferred to the researcher.

Participants' Characteristics
The participants usually held an upper position and acted as representatives of their institutions/organizations to the external environment. The participants' sociodemographic characteristics were not taken into account, but only their views. The same applies to the participatory observation at the public library in Örebro where the main researcher had a discussion with a librarian and at an elder's house, close to Örebro University, where the IT-Guide staff (young immigrants) offered digital help to older people.

Data Analysis
The interviews and the observations were complemented with analysis of written documents and survey data. The gathering of multiple forms of data including interviews, observations, and examination of documents rather than relying on a single source is common in qualitative research (Creswell, 2012). All the different sources were analyzed with the use of qualitative content analysis in an inductive way. When applying qualitative content analysis "data are presented in words and themes, which make it possible to draw some interpretation of the results." (Bengtsson, 2016, p.10). The categories and consequently the themes were constructed following their commonalities. "Focused coding" (Charmaz, 2006) was employed for selecting the most appropriate categories. Morse (2008, p.727) defined a category as the "collection of similar data...important for determining what is in the data (the 'what')" and a theme as "a meaningful 'essence' that runs through the data… It is the basic topic that the narrative is about, overall." What was analyzed in the Swedish case was the views of the representatives of certain institutions regarding the digital divide in Sweden via common themes, which frame the problem at hand. At the same time, existing documents and written sources were used, while in some instances interviews' quotes were also included for supporting the argumentation. We did not care about word frequencies. We reached data saturation when the identified themes began to be reported repetitively in the interviews

First institution: Digital Divide and Eldercare Services
Facing an aging population and increasing digitalization, the universal model of welfare is challenged without though entering into a process of lamenting for the Golden period of the Swedish welfare regime. As noted by Eichhorst and Rinne (2017), the key welfare challenge is to master a balancing act between both trends: to promote the opportunities of digitalization while limiting new social inequalities and bridging divides. Despite the rhetoric of being the best in the world in both respects, Swedish policy documents clearly emphasize the first much more than the second.

The Swedish Welfare System at the Crossroads -Alexopoulou & Åström
In digitalization policy as well as in documents outlining future eldercare, government policy is preoccupied with making the most of digitalization. The government set the direction of its digital policy in its Digitization Strategy (Government Offices of Sweden, 2017). The overall goal is for Sweden to be the best in the world at exploiting the possibilities afforded by digitalization, and an important prerequisite for this is digital competence: that everyone is familiar with digital tools and services and can follow and participate in digital developments. However, the strategy does not include discussions of measures for developing digital competence among people, such as older people, not in the education system or the labor market, nor does the Swedish Municipalities and County Councils' action plan for digital development (Sveriges kommuner och landsting, 2017).
In a similar vein, in social policy, the focus is likewise on promoting digitalization, primarily by way of welfare technologies in eldercare (Government Offices of Sweden 2016; SOU 2020, p.14), which usually involves very old Swedes with various health problems. According to a recent report by Nordens Välfärdscenter (2020), welfare technology is the area with the most potential to promote active and healthy aging. Still, the main problem seems to be that the development is too slow and that there are still very few older people with access to welfare technology services (SOU 2020, p.14; National Board of Health and Welfare, 2020a). Welfare technologies, according to the Ministry of Social Affairs, can help improve welfare services, working environments for welfare professionals, and resource utilization. In line with the active aging framework, welfare technologies are considered to contribute to increased safety, participation, and autonomy (The Ministry of Health and Social Affairs, 2018, p.18).
While the social policy is closely intertwined with digital policy, digital divides are paid little attention. In the 600-page report Technology of the Future in the Service of Care (SOU 2020, p.14), for instance, there is almost no discussion of how to address the social implications of the digital divide among older people. Just as the active aging framework has been criticized for neglecting dependent seniors, Swedish policy documents are not engaging with the digital divide and don't recognize it as a problem that affects the majority i.e. the digital participation is very high. The question is how a very developed welfare regime as the Swedish one can offer equal opportunities even if a small portion of older people cannot be digital and socially active? RASP -Research on Ageing and Social Policy 10 (1) 60 To answer this question, it must be stressed that in Sweden elderly people's right to care is based on the Social Services Act, Chapter 4, Section 1, which states that anyone who cannot meet their own needs, or have them met in other ways, is entitled to care from the municipality. By law, digital help can be applied for, although medical and physical needs are prioritized in needs assessments (Katzin, 2014). In recent years, an opportunity has been given for individuals to receive certain kinds of help without first being subject to investigation and assessment (SOU 2020, p.47). A few pioneering municipalities have used this opening to bridge the grey digital divide.
In the municipality of Mölndal, for instance, the Health and Care Committee decided in April 2019 to introduce a new initiative in the home care service to provide digital support in the home. The purpose of this effort is to maintain or increase independence and reduce the digital exclusion of seniors. If one has reached the age of 70 years, one is offered this service without a needs assessment, so no investigation is required to receive digital support at home. Elderly people are granted this digital support for a maximum of 10 hours per year (Mölndal Municipality, 2020). However, municipal eldercare more generally is not given the responsibility for providing digital help and these two aspects usually are treated are two different things.
To sum up, whereas digitalization is a national public concern and the government reviews its governance regarding the basic conditions for using welfare technology (SOU 2020, p.14), the digital divide seems to be viewed as more of a private than a public matter. National policy documents do recognize that older people risk early exclusion and need help with 'simple things'such as shopping, paying bills, ordering train tickets, or taking a bus (National Board of Health and Welfare, 2020b)but this problem is presented in terms of digital skills and the measures discussed merely focus on individuals and educational institutions (Government Offices of Sweden 2017, p.13).
As the public sector increasingly requires citizens to use digital services instead of, for example, getting help from an administrator in an office, society should primarily ensure that there is relevant education. (SOU 2020, p.14, 535).
A final comment is that older people in policy documents are treated as a homogenous group, while in reality, the digital divide varies extensively between young-old individuals (65-74 years old) who do not face serious health problems in comparison to very old Swedes (75 years and older) who primarily come in contact with technology through welfare technologies to keep a level of independence.

Second Institution: Digital Divide and Popular Education
The first steps to using popular education as a way to bridge the digital divide were taken very early in Sweden. Popular education is offered for free outside the traditional school system. Its purpose is to offer citizens an open and democratic way of absorbing new knowledge and values through, for example, study circles, libraries, conferences, and lectures. As such, popular education holds a strong and cherished position in Swedish society. For the interest organization of Swedish study associations, popular education is a cornerstone of Swedish democracy. Its importance is also reflected in Olof Palme's often quoted description of Swedish democracy as "study circle democracy" (Larsson, 1999).
In the 1980s, study circles were already being used to increase computer literacy and awareness of the role of computers in society. According to Emanuel (2009), the motive at the time was to counteract increasing knowledge gaps regarding what was considered a key future technology. Around the turn of the millennium, that future was approaching: about 50 percent of the general population and 12 percent of the 65-84-year age group had Internet access at home. Yet, many Swedes still did not see how they could benefit from the Internet. The IT Commission, therefore, suggested that efforts to bridge the digital divide should be made, primarily by introducing more people to IT at libraries and via other forms of popular education. The Commission also proposed that issues related to the use of IT in the home should be addressed by the Ministry of Culture (IT-kommissionen, 2002).
When many countries rolled out digital skills trainingin schools, telecenters, libraries, and community centerspopular education was the default and path-dependent choice in Sweden. At that time, the digital divide was not yet a matter of life and death, as the UN declared this year. Digitalization was not yet all-encompassing, it was not the norm. On the contrary, the challenge was, as the IT Commission put it, "to find the best way of reaching out to those who never before have been in contact with the RASP -Research on Ageing and Social Policy 10(1) 62 medium, not in school or work, nor at relatives, friends or neighbors" (ITkommissionen, 2002, p.12). In this effort, important help, apart from the popular educational institutions, is offered by the pensioners' organizations that organized IT courses only for their members.
The nature of the digital divide has changed, but libraries and study associations are still at the very heart of relevant Swedish policy. As Marking (2010) concluded, few other municipal initiatives target digital inclusion, even though the Internet is increasingly important for public institutions' interactions with citizens (Marking 2010; see also Nordqvist & Wihlborg, 2019). If there has been a change in policy over the years, it has merely been along the lines of "the same, only more so". In 2009, the government has commissioned Folkbildningsrådet to promote digital inclusion (Folkbildningsrådet, 2015). To reinforce the role of libraries, the Library Act (SFS, 2013, p.801, § 7) was amended to state that libraries must help citizens manage digital services. As the Swedish Library Association's chair Johanna Hansson (Svensk biblioteksförening, 2018) stated, "working for digital inclusion has become one of the public libraries' core tasks".
According to a survey providing a recent update of the activities carried out in municipal libraries, about 80 percent of libraries organize activities to promote digital inclusion and participation (Norberg, 2017). About half of libraries use campaigns such as the "Get Online Week" to organize activities, and just as many organize ongoing activities over the year. Most representatives of the libraries not doing anything say that the main reason is lack of time and resources. Taking a closer look at what libraries do, the survey shows that the most common activity is providing information and support about the libraries' services. While almost all libraries said to be promoting digital inclusion offer such information and support, many fewer organize information and support concerning government e-services more generally: only 24 libraries (of 264 respondents) reportedly do this regularly (i.e. at least once a month), and 88 reportedly do this once or twice a year, for instance, during a campaign week (Norberg, 2017).
In addition, most libraries offer public computers to their visitors and provide digital help when they are used. However, there are often ambiguities about the limits: how far should this help go, and how quickly. While some visitors ask help here and now with many different questions, library staff do 63 The Swedish Welfare System at the Crossroads -Alexopoulou & Åström not have the specialist skills that are sometimes expected of them. As one interviewed librarian at Örebro Library explained: First of all, we would like to help with the library's digital services that we offer, like e-books and such, but we also help people having problems with their mobile phones or computers. We don't help them with overly personal matters such as bank ID.
Digital help is widely discussed in the library sector, with some believing that too much librarian time is spent helping people with digital community services at the expense of disseminating literature and media (Dahlquist, 2019). In addition, it is recognized that problems often arise with, for example, privacy issues when librarians help individual users manage services such as bank ID or various authorities' e-services, something that the previous quotation confirms. As Nordqvist and Wihlborg (2019) argued, library staff largely consider themselves as lacking the training, mandate, and support required to handle many of the issues they are being asked to address.
The study associations operate under the motto "free and voluntary" (Statskontoret, 2018). The government issues only general guidelines for what grants to these associations are to be used for, but does not otherwise control their activities. How popular education is designed is therefore based on each organization's unique profile. Another hallmark of popular education is that the participants are given great influence over the structure and content of each study circle. Their activities are based on the participants' needs and experiences and all people's lifelong right to freely search for knowledge. Accordingly, most people in study associations do not feel that there is a national plan or that they have a specific national mandate regarding digitization (Folkbildningsrådet, 2019b). Digitization and digital inclusion are only a small part of the matters that study associations address. According to the annual reports of Folkbildningsförbundet (2017, 2018, 2019a), most people participate in music, theater, and dance activities. Although some interviewees said that "there is no lack of opportunities" and that "in every bush there is a course" (IP: Äldrecentrum), only about one to two percent of all study circles fall into the computer literacy category. Today, about 25,000 Swedes per year participate in these courses, though at times there have been more. The Digidel campaign between 2011 and 2013, for instance, aimed at getting 500,000 Swedes online (Digidel, 2013).

RASP -Research on Ageing and Social Policy 10(1) 64
Regardless of the scope of these efforts, one may ask whether free and voluntary popular education is appropriate for dealing with the third level of the digital divide. One case in point concerns the digital payment system. In the last decade, Swedish payment services have changed dramatically. Cash is being used less and less and digital payment services are increasing in scope (Post-och telestyrelsen, 2017). In 2020-2021, the Covid-19 pandemic accelerated the reduction in the use of cash in society, making it even more difficult to pay with cash than in previous years. According to several interviewees (IP: County Administrative Board; IP: Test Bed Elderly and Disabled People in Örebro Municipality; IP: Internetstiftelsen), society will face a serious divergence in outcomes when the payment system, a key societal function, becomes inaccessible to those who are not comfortable online. Support is not always available to those people: We have mapped non-profit and public actors who offer computer courses and IT support in Uppsala County. We have not found any specific initiatives or courses for digital payment services (Länsstyrelsen & Uppsala län, 2019).
Without clear responsibilities, key tasks end up "falling into the cracks". Several interviewees (IP: IT-Guide; IP: Seniornet; IP: Health Department) demanded more active support from the state and municipalities. While others indicated the previous standpoint, an interviewee (IP: PTS 2017) underlined that "It is a difficult question. I think you can always do more. It is a kind of priority discussion, you know".
Another case in point concerns the target group. When the target group was a large and diverse group of non-users, free and voluntary measures could be employed across social contexts. Today, non-use is concentrated among the elderly, and policies and interventions need to focus on the hardest to reach (cf. Helsper & Reisdorf, 2017). As concluded in the Digidel project (2013), many of those who remain digitally excluded today do not want to use the Internet and may even be opposed to the technology. As these excluded people cite lack of interest as their main reason for not using the Internet, they are unlikely to volunteer for relevant training. It was argued in Digidel (2013) that municipalities' eldercare services have a key role in reaching many of those who do not use the Internet or do so very rarely: If care staff could help those who receive care, also teaching them to use the Internet, municipalities' contact costs would decrease and the well-being of those people would increase (Digidel, 2013).

Third Institution: Digital divide and the Family
If digital help is not adequately provided by either the formal eldercare system or popular education institutions, perhaps the family has a role to play? Defamilization has been a key principle of the Nordic welfare regime: the individual should be made autonomous from his/her family. The basic idea is that public care should free us from obligations to relatives so that all adults can work. Both social policy reforms and tax reforms have sought to free individuals from their families, with the help of society's support and common resources (Katzin, 2014). However, as public care for the seniors has decreased in scope, the care efforts of relatives have increased (Szebehely & Trydegård 2007;Szebehely & Ulmanen, 2008). Katzin (2014) argued that starting from the 1970s goal of making informal care by relatives redundant, change has proceeded via the 1990s perception that care by relatives is sometimes an economic necessity, to the 21st-century emphasis on the positive qualities and desirability of family care. Besides the economic considerations, there has been a discursive change in the role of families and in what is regarded as belonging to the public and private spheres.
The family seems to be the main provider of digital support, not only in the familialist welfare regime of Greece (Alexopoulou, 2020) but also in the social-democratic welfare regime of Sweden. This was a recurrent theme in several interviews: I think that many people, and many of the elderly, have children and grandchildren they get a lot of support from when it comes to technology. (IP: Project Leader, Örebro municipality, Municipal Board Administration). I don't have any statistics about it, but I know that there is a discussion about the elderly often getting help from their children or grandchildren … So, there is a discussion about how the public sector should be able to provide help for people who are alone. (IP: Ministry of Enterprises and Innovation).
The informal and familialist character of digital support in Sweden has been confirmed by two recent surveys. A survey by SPF Seniorerna (2019) examined the extent to which seniors can obtain support with digital services and tools. First, the respondents were asked whether they have anyone to ask for help with problems in using digital services, to which 83 percent answered in the affirmative. These people were then asked a follow-up question about who they usually ask (Table 2), and nine out of ten replied that they get help from family and friends. Eight percent ask for help from ISP customer service, IT support from a private company, or a membership organization. Strikingly, only one percent said that they received help from publicly-funded welfare services such as a library, IT café, or IT course. Further information on this astonishing number (90% percent) is not provided by the above survey, for instance, what kind of relatives usually help older people within the family, i.e. children or grandchildren. The Swedish Internet Foundation's (2020) survey of a representative sample of the entire population also confirmed this picture. In total, 82 percent of the population turns to family members for digital help. Among seldom users and non-users, largely from the older age group, the corresponding figure is 97 percent. Here, too, it is striking how little public-funded services are used, in this survey represented by libraries: despite all the efforts made in libraries to provide IT guidance, no respondents said that they would turn to libraries for digital help.
Another key finding in the report is that the people who need support the most, the seniors, have much less access to support than do younger citizens (Internet Foundation, 2020). As Helsper and van Deursen (2017) showed, small social support networks correlate with digital exclusion: individuals with few socioeconomic resources more often seek informal support from family and friends; people with greater socioeconomic resources often turn to formal sources instead, such as IT support in the workplace or from other experts. As this may increase the gap between those who do and do not need support, Helsper and van Deursen (2017) suggested that the seniors should be offered alternative support, so as not to have to rely on family members and friends. Learning from family, they argued, might not always be "evident, efficient, or preferential since family members might be unavailable, be reluctant and impatient to help, or even refuse support" (Helsper & van Deursen, 2017, p.702).

Discussion and Conclusions
The way digitalization was handled in Sweden was started as a promising path (popular education) but now seems to be a rather inefficient arrangement: many older people and some of them young-old seniors (65-74 years old), who are still active, apparently need digital help, but they are not provided public support. Eldercare institutions promote welfare services (see welfare technologies) for a few and very old Swedes, while young-old Swedes are out of their scope. As a result, the latter face the problem of the grey digital divide. Popular education either provided by the municipal libraries or the third sector (NGO's), does not cover always in a satisfactory and permanent way the needs of older people in Sweden, while pensioners' institutions make considerable attempts to close the digital divide by organizing a number of IT courses which are only targeted at their members Moreover, the offered services cover mostly young-old Swedes without health problems, who can visit these places.
Family, in the Swedish context, is often neglected as an official institution that can offer digital assistance to older individuals because the policy response is more or less taken for granted by policymakers, who believe that the municipalities or the popular education system will address the grey digital divide. Even more worrisome is that in many cases policymakers in Sweden RASP -Research on Ageing and Social Policy 10(1) 68 consider that there is no digital divide issue, given that it affects only a small portion of the total population. This leads us to ask why.
In public policy research, as in comparative welfare research, the nature of public policymaking is often viewed as a matter of path dependence as current policy choices are constrained by choices made in the past. In the words of Jakob Torfing (2001, p.71), the possibility of reforming public policy depends on "institutionalized legacies that structure our perceptions of problems and goals, define the range of appropriate and feasible options, and determine the costs and benefits of policy changes". This case study revealed that much has changed in Sweden as regards the digital divide problem, but the institutions that are responsible for providing digital support remain the same, while family assumes increasingly much of this burden.
Recently, many digital divide scholars have suggested that a shift is needed from a focus on access and skills to a third-level digital divide highlighting the tangible outcomes of Internet use (Scheerder et al., 2017). Accordingly, many interviewees in this study argued that digitalization has important welfare implications: that the digital divide makes many seniors increasingly dependent, and that the state/municipalities can do more at many levels to bridge the divide. The changing nature of the digital divide speaks in favor of a more active state role, which will include the provision of digital help. In her thesis on Swedish eldercare, Helene Brodin (2005) concluded that the single most significant factor in transforming eldercare has been the changed view of the elderly from dependent to healthy. In particular: [T]he single most important reason for the changing view on the public responsibility for eldercare services is however the shifting perceptions of the elderly from a dependent to an independent group. This shift in perception of the elderly has since the 1980s forcefully been summarized under the metaphor of the elderly as healthy, which in turn has legitimized a more passive role for the state in organizing and financing social services to the elderly (Brodin, 2005, pp.206-207).
The present results indicate that this view of the seniors as independent is reinforced by the discourse of digitalization, according to which citizens are increasingly expected to be active and self-provisioning individuals serving themselves across a range of welfare domains (see also Schou, 2018, p.2). Accordingly, in digital policy, active aging is understood to mean that older 69 The Swedish Welfare System at the Crossroads -Alexopoulou & Åström people should themselves be adaptable and learn how to use technology. Very old Swedes (when applicable) are benefited from the welfare technologies with the assistance of the municipalities and family but they are not always independent, while young old Swedes who are not familiar with digital technologies have to find by themselves ways to confront their lack of skills (second-level of digital divide) and to increase their actual opportunities (third-level of digital divide), for instance by visiting a public library or a third sector organization. The key institutions, which provide this sort of assistance are popular education and family. The oxymoron is that young-old Swedes who lack digital competencies are not independent and digitally active, but they confront serious problems with the ongoing digital transformation of the Swedish welfare regime.
Indisputably in both previous cases (young-old and very old Swedes), the family institution appears to be the 'alpha and omega' for their well-being status. This study goes in line with the analysis of Motel-Klingebiel et al., (2005), according to which the total quantity of help received by older people is greater in welfare states with a strong infrastructure of formal services. The key difference is that in the Swedish case, despite the existence of the system of 'mixed responsibility' (families and welfare state services act accumulatively), the burden falls greater on the family institution. This leads the Swedish welfare regime to a crossroads: If we follow the current wellworn path, we must accept that the changing nature of the digital divide will culminate in an increasingly "familialist" model of welfare, while this comes in opposition with the Swedish welfare mentality. If we choose to take a new path, the traditional Swedish model must be adapted to the changing nature of the digital divide with new more effective institutions. If a more familialist path is taken, then the family members in Sweden could be reinforced economically (financial benefit) by the Swedish welfare state to handle the digital affairs of their older relatives.
To finish, the social contribution of this qualitative case study is to raise awareness about the shifts that take place in Swedish society and to show the need for more suitable policy responses that will allow older individuals not to fall through the digital and social net. It is also offered a potential solution (economic benefit) for addressing the greater involvement of the family institution in supporting older adults in Sweden.