2. Making sense of global institutional change in the context of digital transformation

Until recently, digital technologies were not part of mainstream social policy debates. Expectations about the current wave of innovations led observers, however, to suggest that institutions should be adapted to technological change: it was understood that “digitalisation and platformisation require profound rethinking of 21st century welfare provision” (Eichhorst et al., Reference Eichhorst, Hemerijck and Scalise2020: 28). Recent publications in this journal indicated a growing interest in the potential of digital innovations to deliver welfare policies (Henman, Reference Henman2022), and to coordinate decision-making in state bureaucracies (Considine et al., Reference Considine, McGann and Ball2022). Scholars also paid attention to policy responses on the effect of digitalisation on jobs markets (van Doorn and van Vliet, Reference van Doorn and van Vliet2022), and the implications it might have for welfare attitudes (Busemeyer and Sahm, Reference Busemeyer and Sahm2021). Beyond the value of these and other studies, most of the literature has had a national or local focus, leaving aside the question about the role of international actors in digital transformation. This is particularly remarkable given that this is a structural process that, as such, should involve global responses and initiatives. This paper aims to fill this gap in the social policy literature, adding the global level to institutional analyses of this process of transformation.

The conceptual approach I propose here to make sense of this agenda and its rationale consists in a sociological understanding of global institutional change, inspired by Sarah Babb and Alexander Kentikelenis’ (Reference Babb and Kentikelenis2021) study on the Washington Consensus. Even though this variant of institutionalism has primarily focused on stability, I propose that it can be utilised to make sense of change and adaptations in international institutions. Babb and Kentikelenis have stressed the influence of three interrelated dimensions in the making of a new global consensus, which I use as a basis for the analysis. First, there is an organisational aspect, referring to the agendas of international institutions and their diffusion of a given set of policies. These agencies tend to follow particular interests and mandates (Kaasch and Martens, Reference Kaasch, Martens, Kaasch and Martens2015), while their strategies influence national states in various degrees (Gilardi, Reference Gilardi, Carlsnaes, Risse and Simmons2013). Second, there is an ideational dimension, alluding to the predominance of theories and evidence that legitimate institutional reforms, often based on specific epistemic communities (Haas, Reference Haas, Bodansky, Brunnée and Hey2008) and knowledge networks (Stone, Reference Stone2008). The third factor that shapes global institutional change is the state of the international political economy (IPE), i.e. the reordering of political and economic forces on a global scale, as well as the irruption of wide-reaching events that put pressure on institutions to react (Mahrenbach and Shaw, Reference Mahrenbach, Shaw and Shaw2019). This framework, in my understanding, rather than indicating explanatory factors, shows conditions of possibility for a consensus to emerge. To that degree, this study has a primarily interpretative character.

The interaction of these three dimensions can catalyse institutional transformations in international actors and national states alike. Such change can be interpreted in two lines, drawing upon previous works from sociological institutionalism. One is the creation of a new transnational policy paradigm, meaning a set of goals and instruments that are shared by a substantive number of countries and international actors (Babb, Reference Babb2013). Paradigms constitute “cognitive assumptions that typically reside in the background of debates about change” (Campbell, Reference Campbell2004: 156), which are supported by elites until they lose legitimacy before mounting anomalies and inadequacies (Hall, Reference Hall1993). Another result of global institutional change refers to the expansion of institutional models of actors (Meyer, Reference Meyer2010) – for example, the emergence of a certain ‘model’ of state that countries are urged to emulate. In the sections below, I will apply these sociological concepts to make sense of global institutional change in the context of digitalisation, with a focus on social policy and welfare states. The argument is that in the previous decade a convergence of organisational agendas in multilateral agencies, the salience of inclusive growth as a concept in the international arena, and the threat of far-right populism created the conditions for a new consensus, based on a Schumpeterian paradigm and a model of the state oriented to promoting innovations through social policy.

The methodological strategy utilised focuses on three institutions that have been particularly influential in social policy debates in past decades (Kaasch and Martens, Reference Kaasch, Martens, Kaasch and Martens2015): the International Labour Organization (ILO), the Organisation for Economic Co-Operation and Development (OECD), and the World Bank (WB). Even though practically all major international institutions have developed initiatives on digitalisation, it seems more empirically fruitful and plausible to reduce the scope to a limited number of cases. These three have been chosen to provide a picture of the diverse ideological and geographical characteristics of international organisations. In that line, the ILO is an agency with global presence and based on a social-democratic imaginary; the OECD comprises rich democracies and promotes inclusive liberalism, while the WB works in developing nations advocating for liberal values and strategies. I have studied the digital transformation agendas of these respective institutions through document analysis of flagship reports and policy documents published between 2015 and 2020 on digital technologies. The documents were scrutinised through a content analysis strategy based on a coding matrix focused on three categories with sub-dimensions: problem definitions, policy solutions and conceptual foundations presented in the reports (based on Mehta, Reference Mehta, Belánd and Cox2011). A list of the analysed documents can be found in the appendix to this paper, as well as the matrix of categories used to code them.

3. The IPE dimension: the erosion of the divide between ‘economic’ and ‘social’ institutions in global governance

In order to understand the appearance of a new consensus in the international arena, it seems necessary to make reference to the history of antagonism that characterises global social governance. This field has commonly been characterised by scholars as a ‘war of positions’ between ‘economic’ institutions that focus on growth and competitiveness, on one side, and ‘social’ institutions that have a humanitarian and rather normative orientation, on the other (Deacon, Reference Deacon2007). In the 1980s, the consequences of structural adjustment made international financial institutions face opposition from United Nations (UN) agencies (Jolly, Reference Jolly1991); in the following decades these actors would constantly clash over the management of economic globalisation. For instance, in the 1990s, the World Bank was a strict defender of market openness, highlighting the benefits of foreign investment for job creation (World Bank, 1995). The ILO became an advocate of core labour standards to enforce minimum norms for an increasingly integrated world economy (Alston, Reference Alston2004). Even in the context of the Great Recession after 2008, these agencies had contrasting positions on the role of the state and the extent of social protection in the crisis (McBride and Merolli, Reference McBride and Merolli2013). The literature likewise used this binary division to make sense of global institutions’ positioning on emerging events and agendas.

At the beginning of the 2010s, inter-organisational relations in global governance started showing signs of a different logic. The general adherence on the idea of social protection floors in Bretton Woods institutions and the UN system indicated that the contradiction between these two positions was giving way to a relationship of ‘accommodation’ (O’Brien, Reference O’Brien, Kaasch and Stubbs2014). I further propose that there were two contextual factors in the international political economy that led, by the mid-2010s, to a dissolution of the divide between ‘economic’ and ‘social’ institutions in global governance, thus making possible the Schumpeterian consensus I will describe below. The first process referred to the expansion of inclusive growth as a dominant paradigm in the international community, including the economics profession and multilateral agencies. As Rianne Mahon explained (Reference Mahon2019), in the past decade the WB and the OECD respectively re-elaborated their approach to social policy by emphasising the importance of inclusiveness in promoting economic growth, instead of opposing efficiency and equity as contradictory aims. Issues of sustainability also became part of their institutional agendas, beyond the orthodox perspectives they had endorsed in recent decades (Deeming and Smyth, Reference Deeming, Smyth, Deeming and Smyth2018).

A second factor, now related to the balance of forces in international relations, was the irruption of far-right populism in the 2010s. The latter posed a serious challenge to the liberal international order and multilateralism (Mearsheimer, Reference Mearsheimer2019), provoking a global wave of conservatism and exclusionary measures (Fischer, Reference Fischer2020). In terms of social policy, its effect was to force a social-democratisation of agencies conventionally positioned in the neoliberal spectrum, and contrarywise a liberalisation of institutions that would represent social democracy in the international arena. More specifically, the OECD understood populism as a by-product of widening socioeconomic inequalities in advanced political economies (Gurria, Reference Gurria2017), thus justifying a reorientation of its institutional discourse towards a social-democratic imaginary concerned about redistribution and social dialogue (e.g. OECD, 2018b). Social policymaking at the ILO followed the opposite direction. The confrontational strategies of business at the ILO and their strategic alignment with far-right populist governments led to a substantive ideational change in the organisation at the end of the decade (Silva, Reference Silva2021). Despite having represented social democratic principles internationally –tripartism and social justice-, the ILO has recently adopted a more pro-business position that sees the private sector as the engine of economic development (e.g. ILO, 2019a). These two institutions – the OECD and the ILO – established a common agenda on inclusive growth and decent work, embodied by the ‘Global Deal’ (ILO and OECD, 2018), demonstrating that the existence of a clear divide between a liberal and a social-democratic camp was not so clear anymore.

4. The organisational dimension: trajectories of global institutions in the digitalisation agenda

Global institutions in the past decade, especially those concerned about social affairs, explicitly engaged with the digitalisation agenda. Each organisation ‘responded’ from their respective perspectives and political processes, but they shared the preoccupation for ‘investing in people’ for the future of work, for preparing society for the new digital era. A brief description of their trajectories will demonstrate the previous. For one, in 2013 the ILO turned its attention from the social consequences of globalisation towards the impact of new intelligent technologies. This theme was chosen by Director General Guy Ryder to become the main focus of the organisation’s centenary for 2019 (International Labour Office, 2013). The ILO Future of Work Initiative launched in 2015 included the formation of a Global Commission that in 2019, in the context of its centenary, published a report that proposed a ‘human-centred approach’ for the future (ILO, 2019b). The framework of the Commission aimed to embed technological change in capabilities’ development, promoting innovation that would increase workers’ resources and autonomy. Such approach suffered straight opposition from employers and right-wing governments in the making of the ILO Centenary Declaration (Thomas and Turnbull, Reference Thomas and Turnbull2020), who re-oriented the institution’s agenda towards skills development and adaptability for job creation.

The goal of ‘humanising’ digital technologies in the UN system was also assumed by the United Nations Educational, Scientific and Cultural Organization (UNESCO) and its Recommendation on the Ethics of Artificial Intelligence (AI) from 2021. It encouraged member states to create institutions that preserve “human rights and fundamental freedoms”, as well as “human oversight” over AI systems (UNESCO, 2021). The OECD had pioneered in that line some years before with the making of a Recommendation on AI, for which “AI should respect the rule of law, human rights and democratic values” (OECD, 2019d). These cooperative efforts between multilateral agencies and countries found ample support in instances such as the G20, the group of the twenty largest economies, agreeing on a framework on ‘AI principles’ in 2019 largely based on previous work from the ILO and the OECD (L20, 2019). The WB has also thematised the ethics of new technologies, supporting the safeguard of privacy rights to address “the power asymmetries between (individual) data subjects and data processors and collectors” (World Bank, 2021: 194). In other words, institutions with largely different aims and interests seemed to recognise the necessity to set institutional and normative boundaries to technological developments in the age of AI.

Apart from the human-centred approaches, there were two additional trajectories that must be examined here. The OECD experienced in the past decade a substantive ideational turn in the context of its inclusive growth initiative, distancing this organisation from the orthodox position it defended in the 1990s (McBride et al., Reference McBride, McNutt and Williams2007) and its focus on activation from the 2000s when it embraced ‘inclusive liberalism’ (Mahon, Reference Mahon2011). In the following decade, after the Euro crisis, the OECD focused their social policymaking on digitalisation. On this process, in 2016 labour ministers and the secretariat decided to make a new Jobs Strategy – ‘Good jobs for all in a changing world of work’- which was published in 2018 (OECD, 2018a). In a scenario where digital ‘superstar firms’ and platform labour were notoriously expanding, the OECD proposed to strengthen the institutions of work –centralised collective bargaining, minimum wages, and inequality reduction – in a clear turn to a social-democratic imaginary (see OECD, 2019b). At the same time, the OECD committed to renew its social policy framework for the challenges of the digital era and re-elaborated its Skills Strategy in 2019, encouraging member countries to invest in training over the life course, foreseeing that in the future jobs and occupations will disappear due to automation (OECD, 2019c). As in the Jobs Strategy, the one on skills supported stakeholder engagement and social dialogue in the governance of skills systems. The re-making of social policy frameworks at the OECD played a key role in the institution’s projects related to technological change: the most important of them, the ‘Going Digital’ initiative (OECD, 2019a), presented a roadmap in which well-being, good jobs and social inclusion were considered central to build a sustainable digital transition.

As the OECD, the WB experienced a similar pattern of institutional change in the context of the digitalisation agenda, as it explicitly embraced the ‘flexicurity’ model in social policy provision as a response to such transformation. In a number of World Development Reports, the Bank exposed the economic gains and opportunities brought by new technologies – for example, by creating new industries and jobs, or making organisations more efficient (World Bank, 2016, 2021). Despite the optimism shown in the Bank’s approach, the institution repeatedly showed concerns about the appearance of new social problems in such context. The World Development Report of 2019 – ‘The changing nature of work’ – emphasised that the demand for skills was going to change due to technological change; the risk of polarisation and the already decreasing labour share in the product were worrying trends to consider (World Bank, 2019). The two institutional initiatives developed as a response to these analyses were the creation of the ‘Human Capital Project’ in 2018 (World Bank, 2018b), which provided technical support for countries to invest in their people, and the ‘Protecting all’ social policy framework from 2019. The latter positioned flexible security as the key goal for the future, giving companies freedom to adopt innovations while guaranteeing that all workers have universal access to social protection – e.g. via a universal basic income (Gentilini et al., Reference Gentilini, Grosh and Rigolini2020) or a ‘pentapartite’ model that would include the self-employed and non-standard workers in social dialogue (Packard et al., Reference Packard, Gentilini and Grosh2019).

The previous indicates that major players in global governance put digitalisation and technological change at the centre of their respective organisational agendas. They also remade their social policy approaches, with different emphasis to be sure, but with the common idea that states should ‘invest in people’ to face digital transformation (see Table 1 below). Their organisational trajectories did not seem to be structured by an opposition between economic and social dimensions, or by an ideological dispute between neoliberal and social-democratic positions. Even though they developed their own frameworks and supported specific instruments, the next section shows that they were actually operating within the same policy paradigm, especially in terms of problem definitions and conceptual foundations.

Table 1. Trajectories of global institutions in the context of the digitalisation agenda (2015-2020)

Source: Author

5. The ideational dimension: the Schumpeterian consensus and the new welfare state for the digital era

The diverse trajectories found in the organisational dimension should not obscure the fact that, by the mid-2010s, the mentioned actors had adopted a single understanding of economic development and, perhaps more important, of the role of the state in welfare provision. The concept of ‘consensus’ here is presented as it has been utilised in international political economy to denote the convergence of actors around certain policy paradigms, as in the ‘Washington consensus’ (Williamson, Reference Williamson, Serra and Stiglitz2008) or the ‘Wall Street consensus’ (Gabor, Reference Gabor2021). However, in this case I am not referring to industrial and macroeconomic policy, but to social and labour market regulations at a global level. The new model that became a transnational policy paradigm in the 2010s can be called the ‘Schumpeterian consensus’, which was fundamentally inspired by ideas, not by geopolitical/financial interests like the ones just mentioned. The essence of this paradigm relies on two notions. The first is a conception of economic and technological development grounded on Schumpeter’s idea of ‘creative destruction’. A second aspect is the promotion of ‘investing in people’ to prepare societies for digital transformation. The social investment approach and the flexicurity model, as an extension of the Schumpeterian consensus, became part of this transnational policy paradigm, being endorsed by the various actors in global social governance: multilateral agencies, international experts, and civil society organisations.

The idea of creative destruction was popularised by Joseph Schumpeter in his Capitalism, socialism and democracy (1994), representing capitalist development as a force that continuously grows at the expense of already existing economic and institutional forms. In this view, innovation –expressed in the creation of new markets, organisational methods or productive techniques- is constantly disrupting the economy, altering the general equilibrium taken for granted by neoclassical economics (Screpanti and Zamagni, Reference Screpanti and Zamagni2005). The logic of creative destruction is considered to be the core feature of capitalism, unravelling as an ‘organic process’ (Schumpeter, Reference Schumpeter1994: 84), thus inciting a succession of economic cycles linked to technological innovations. Even though Schumpeter did not advocate for institutional interventions in this ‘organic’ process, it was Neo-Schumpeterian economists who argued that state institutions and regulations were necessary to keep the creative potential of technological change, but within a “narrow corridor” where growth and innovation were not curtailed (Hanusch and Pyka, Reference Hanusch and Pyka2007: 287). In terms of the digitalisation agenda in global governance, in the 2010s there were explicit signs of a common narrative on such process:

Global institutions remained somewhere in the middle between a strict economic conception of the process that understands new risks as a ‘necessary evil’, and an approach concerned about its social implications, willing to restrain the spread of certain disruptive technologies. In other words, neither Hayek and unbounded markets, nor Polanyi and embedded economies took the lead. On the opposite, as expressed by influential Schumpeterian economist Phillipe Aghion, there had to be an investor state that could focus public initiatives on innovation and industrial policy, and an insurer state that could “protect individuals against the risks induced by innovation and creative destruction” (Aghion et al., Reference Aghion, Antonin and Bunel2021: 272). In the 2010s this dual view on the role of the state became paradigmatic in the international community, at least in relation to social policy. The state had to invest in people to prepare them for the future, adapting social and labour policy to let out the creative side of digitalisation –generating jobs and new tools of work and management, while creating new job opportunities through digital platforms. At the same time, the state was expected to invest in mitigating the new set of risks attached to this wave of innovations – namely, technological unemployment due to automation, the polarisation of the jobs market, the rise in non-standard forms of employment and the spread of monitoring mechanisms and data analytics in organisational settings (Figure 1).

Figure 1. The Schumpeterian consensus in global governance: creative destruction and social investment.

Source: Author

I call this new model of welfare the ‘Schumpeterian social investment state’, which must mediate between creative and destructive forces, channelling them towards inclusive growth and human-centred development. Apart from feeding from Schumpeterianism, it borrows elements from the social investment paradigm.

Social investment as a concept rests on the assumption that there is productive value in social policy, against views that oppose efficiency and equity, particularly dominant before the 2000s (Morel et al., Reference Morel, Palier, Palme, Morel, Palier and Palme2011). I can give a general description of its rationale. First, the ‘investment’ dimension does not necessarily expect social returns in the sense of neoclassical economics, as measurable outcomes (Nolan, Reference Nolan2013). In a broad sense, investment is focused on raising “the quality of the ‘stock’ of human capital and capabilities”, as well as on providing “universal safety nets ‘buffers’ for micro-level income protection and macro-economic stabilization” (Hemerijck, Reference Hemerijck2018: 816–7). Second, the result of investing is understood not only in relation to growth but also in a broader sense, emphasising aspects such as well-being, equity, or social cohesion. A third element is a focus on enabling transitions, both in relation to life events (e.g. school-work) and various domains of social life (e.g. work-family), mainly under the assumption that economic productivity and opportunities are determined over the life course (Esping-Andersen, Reference Esping-Andersen, Sassen and Esping-Andersen2005). Fourth, derived from the previous point, the investment paradigm is associated the flexicurity model in social policy, i.e. the promotion of flexible arrangements in the labour market with social protection and activation measures for individual workers (Auer, Reference Auer2010). This strategy would, in principle, reconcile the imperative of adaptation from business and the need for support in transitions and new risks (Greve, Reference Greve2014).

The question about how these policy orientations were addressed in the digitalisation agenda in global governance will be covered in the next section. Before that, one final conceptual reflection is offered. Until the mid-2010s, the social investment perspective had been endorsed by global institutions in relation to different strategies: some of which were associated to what Bob Jessop (Reference Jessop and Cahill2018) called the Schumpeterian ‘workfare’ state. For this model, in a globalised economy, countries had to develop ‘defensive’ activation measures and disciplinary approaches to social policy to remain competitive (Torfing, Reference Torfing1999). As Jane Jenson suggested (2017), the OECD recommended a turn towards active labour market policies in the Global North and the WB backed conditional cash transfers for poverty reduction for the South; they shared, nonetheless, the same emphasis on the disciplinary role of the state and the importance of investing in children development.

There are three key differences, however, between the Schumpeterian workfare state and the Schumpeterian investment state that I am delineating (see Table 2). First, the latter presents itself as a truly globalised state model, an institutional template to emulate in a context of worldwide industrial transformation. Even if in the 2010s countries in the Global South were still focusing their attention on all-too-common problems like informality or low internet access (see ILO, 2017), the discourse of international institutions was promoting this state model regardless of countries’ development situation. Second, in the Schumpeterian investment state the disciplinary element is virtually absent, while the context that gives meaning to its principles is digital transformation, not economic globalisation, which was displaced in the 2010s as a centre of gravity in global social governance. Lastly, a third element that separates these models has to do with policy targets. Workfare had a strong emphasis on childcare and children-related investment; on the contrary, the expansion of automated technologies in the past decade forced a reorientation towards adult workers that could be displaced by labour-saving digital innovations.

Table 2. Models of Schumpeterian state in global social policy: two meanings of ‘investing in people’

Source: Author, inspired by Jenson (Reference Jenson and Hemerijck2017) and Jessop (Reference Jessop and Cahill2018)

6. Policy expressions of the Schumpeterian consensus: a new global social model?

The Schumpeterian consensus rested on a shared conception of development and the role of the state in welfare provision. Global institutions, I have mentioned already, supported different policy instruments and goals according to their own organisational agendas, but they also showed noticeable similarities in the problem-definition aspect of digital transformation. They identified issue areas related to the ‘social model’ (Dølvik and Martin, Reference Dølvik and Martin2014) that had to be addressed – insufficiencies in employment relations, social policy and skills formation systems that had been exposed by technological change and that, consequently, had to be faced by states. In terms of employment relations, even though in previous agendas on the globalisation of production the standardisation of work had been a primary concern (Bernards, Reference Bernards2017), this was no longer the case in the digitalisation agenda. Instead of countering precariousness and informality via formalisation, multilateral agencies were now announcing the death of the standard model of employment as an empirical reality and as a normative horizon:

These observations sustained the idea that future labour markets were going to be flexible, partly due to digital innovations that were disrupting the standard model (e.g. the gig economy), partly due to an already ongoing process of casualisation related to a global decrease in labour power in employment relations (OECD, 2019b). In this context, and considering that states had to mitigate risks beyond adaptability, welfare in the new social model supported by international institutions was fundamentally based on the flexicurity approach, now detached from its Nordic origins and presented as a global institutional template:

The renovation of social policy systems implied, then, letting employers displace workers through automation while keeping universal social provision detached from employment status. Even though universalism was already on the table when the digitalisation agenda emerged, especially in the widely supported concept of social protection floors (Zelenev, Reference Zelenev2015), in this case there was the idea that all workers could be exposed to technological replacement, and therefore governments had to ‘protect all’ (Packard et al., Reference Packard, Gentilini and Grosh2019). The appropriate instruments to achieve that objective varied: for instance, the WB recommended a universal basic income (Gentilini et al., Reference Gentilini, Grosh and Rigolini2020), but the ILO opposed this policy, given its possible regressive effects (Ortiz et al., Reference Ortiz, Behrendt and Acuña-Ulate2018).

In relation to skills formation, the third element of the Schumpeterian consensus’ social model was the imperative to reform educational systems to prepare youths and adults for rapidly changing labour markets in the digital era. Investment in a lifelong conception of education, particularly a change from knowledge-centred learning towards the transmission of adaptable soft skills, became the mainstream position in international debates in the 2010s (Penpraser, Reference Penpraser and Gleason2018). In parallel, in global governance there was increasing support for subordinating educational institutions to labour markets:

The need to reskill workers to cope with technological change was also noticeably advocated by business consulting groups (Schlogl et al., Reference Schlogl, Weiss and Prainsack2021). Global think tanks, echoing such views, called for a ‘reskilling revolution’ (World Economic Forum, 2018), arguing that governments should “help workers develop skills best suited for the automation age”, focusing on “developing agility, resilience, and flexibility”(McKinsey Global Institute, 2017: 18–9). The idea that disruptions caused by technological innovations could be mitigated primarily by adapting skills formation to the needs of the private sector constituted an ‘educational fix’ (Peters et al., Reference Peters, Jandrić and Hayes2019) that was an unequivocal component of the Schumpeterian consensus.