1. Introduction

Our understanding of western societies is heavily influenced by research on the unequal distribution of wealth and income. Stiglitz (Reference Stiglitz2015), Atkinson (Reference Atkinson2015), and Piketty (Reference Piketty2014) note the dramatic increase in income disparity between ‘the richest 1% and the rest’ and the growing concentration of wealth and income in the upper classes. Stiglitz (Reference Stiglitz2016, p. 169) and Hacker and Pierson (Reference Hacker and Pierson2010) note that the rise in the wealth and income of the 1% directly correlates to the growing financial difficulties of the middle classes and inequality of opportunity and outcome. It is only in the last decades that social inequalities have become central to any understanding of the middle class and of social stratification (Reference PaynePayne, 2013). In the 1990s, when inequality began to rise in many OECD countries, research focused on the increase of poverty or on the consequences of deregulation, rather than on the decline of the middle classes and how to tackle this (OECD, 2011; Reference KenworthyKenworthy, 2007; Reference DallingerDallinger, 2013; Reference Giesecke and GroßGiesecke & Groß, 2003).

In my paper, I argue that a large part of the research focused on the relation between extreme inequality and the decline of the middle class does not capture the extent of the changes in its living conditions, the level of hardship and the impact on civil coexistence and the functioning of economic and political institutions. Many OECD countries are still ‘middle-class nations’, moreover, the middle class is no longer a locus of social and systemic integration, but instead includes both people who occupy social positions that provide them with material and cultural advantages and people whose income is low and status precarious. The latter group constitutes the vast majority. In many OECD countries, there has been a growth of an extended social stratum formed by the majority of individuals, which is separated not only from the ‘super rich’, but also from the established middle class and the most disadvantaged households. I believe that the condition of this stratum constitutes a significant and critical ‘juncture’ for any welfare strategy aimed at combatting social and economic processes that are increasing the vulnerability of many households and further degrading social life. Furthermore, the relevance of these changes for the organisation of welfare services is not adequately emphasised. Social policy responses to middle class decline have been weak. They appear less and less capable of protecting the middle class from the growing risk of instability, and do little to foster social mobility or improve its social and economic condition.

In the first part of my paper, I present my theoretical framework, placing it within the broader debate on inequality and its influence on middle class decline. In the second, I seek to interpret some key dimensions of the changes in the condition of the middle classes. In the third part, I argue that many western European societies not only present marked inequalities but also the merging of a considerable part of the middle and working classes into an extended social stratum, characterised by serious economic difficulties. In addition, there is another issue that should be highlighted: the attitudes emerging among these groups go against the functioning of institutions and weaken civic values, creating more political and social instability. The paper concludes with some comments on possible future developments in social policy that could protect this social group against the growing risk of instability and activation policies that support their autonomy across different spheres of life.

2. A framework for detecting less visible trends

Over the last few decades a large body of theoretical and empirical research has documented the effects of the unequal distribution of wealth and income, the loss of economic security, opportunity and upward mobility (Reference LindLind, 2004; Reference KrugmanKrugman, 2009; Reference PressmanPressman, 2007), and the declining living standards of middle class households (Reference ParkerParker, 2013). It shows that all the core aspects of the middle class are under siege (Reference Hacker and ParkerHacker, 2013; Reference Hacker2010). Moreover, economic and technological changes are altering the relations of the middle class with other classes and increasing inequality within the middle class itself. One noticeable change is the fact that inequality is leading to the review of the central position of the divide between the middle and working class (Reference GruskyGrusky, 2014). This is because the divide that underpinned class analysis since its foundation in the thinking of Marx and Weber is becoming less and less clear (Savage et al., Reference Savage, Devine, Cunningham, Taylor, Li, Hjellbrekke, Le Roux, Friedman and Miles2015; Reference SavageSavage, 2015).

However, according to Sassen (Reference Sassen2014, Reference Sassen2015, Reference Sassen2007), the usual tools used to interpret today's range of transformations – such as rising inequality, rising poverty, rising government and household debt – do not help us to detect ‘conceptually subterranean trends’ nor to capture the larger reality we must face. In many societies other dynamics are at work as well; dynamics that cut across familiar and well-established conceptual/historical boundaries and are for the most part invisible to current analytical tools (Reference SassenSassen, 2014, p. 8).

Sassen writes of a systemic logic that mostly produces ‘expulsions’. Her point of inquiry is the proliferation of ‘systemic edges’ where a condition takes on a format so extreme that it cannot be easily captured by the standard measures of governments and experts and becomes invisible, ungraspable (Reference SassenSassen, 2015; Reference Sassen2014). The dynamics highlighted by Sassen notably alter the relationship of the middle class with the top and the bottom of the social stratification, as well as the relationship that these classes have with welfare, public services and the right to support and defend them.

Sassen's analysis is useful in understanding that today's transformations cannot be conceptualised only in terms of extreme inequality and a higher incidence of severe poverty: the relevance of these changes and their effects is undeniable, but in many societies diverse ‘subterranean’ and less visible ‘trends’ are growing. These tools only partially help us to understand the growing social instability and the unpredictability of majority attitudes and behaviour in both the public and private spheres of life.

In my paper, I try to understand the complex nature of the changes affecting the majority using Sassen's approach as a basis for my theoretical framework. However, I also take into account that her approach does not pay enough attention to the interplay between structure and agency. By mainly focusing on the importance of systemic logic, it leads to an overemphasis on the power of footloose capital and macrostructures, and encourages a view of the global urban labour force as playing a passive role in globalisation and of cities as containers instead of active agents (Reference RobinsonRobinson, 2009; Reference RushingRushing, 2004). In my opinion, to go beyond these limitations, we must refer to key concepts of different versions of relational sociology (Reference ArcherArcher, 2015; Reference CrossleyCrossley, 2011; Reference Donati and ArcherDonati & Archer, 2015) and to individualisation and reflexivity theories as put forward mainly by Beck (Reference BeckBeck, 1992; Beck & Beck-Gernsheim, Reference Beck and Beck-Gernsheim2001). With these references, we could provide a more comprehensive understanding of the current hardship and welfare needs of the middle class, and of the impact of emerging middle class attitudes and social identities on civil coexistence.

3. The great division

In their article, Alvaredo, Atkinson, Piketty, and Saez (Reference Alvaredo, Atkinson, Piketty and Saez2013) note that the rise in the wealth of the very rich – sometimes described as the 1% – has had a noticeable effect on overall income inequality in the United States and other countries, particularly in continental Europe. The authors argue that for three decades, the debate about rising income inequality in the United States has centred on the dispersion of wages and the increased premium for skilled/educated workers, attributed in varying proportions to skill-biased technological change and to globalisation. In recent years, however, there has been a growing realisation that this growth in inequality is associated with a large increase in the share of income going to those at the very top of the distribution (p. 3). The great divide between the 1% and the 99% is recognised as the biggest social threat of our time and the one issue which a majority of the global protests calling for economic justice have centred on. The vast majority of the 99% are becoming both more equal and often poorer than they were in 2008 and the gap with the 1% is increasing, while inequality within has fallen (Reference DorlingDorling, 2015a, Reference Dorling2015b).

The slogan of the ‘99 per cent’, whose well-being had been sacrificed to the interests of the 1%, became a mainstream topic in American and in many other countries’ political life (Reference CastellsCastells, 2012). The Occupy Wall Street movement with the slogan ‘we are the 99 percent’ thrust the issue of inequality into the centre of the nation's political conversation (Reference MilkmanMilkman, 2017). On the web site ‘wearethe99percent’ personal experiences are described: ‘we are the ones who are getting kicked out of our homes, we are forced to choose between groceries and rent, we are denied quality medical care, we are working long hours for little pay and no rights’.

In contrast, it is possible to argue that there are many social and economic divides within this ‘99 percent’ that risk remaining hidden from the analyses on extreme inequality. The OECD (2015) indicates that while the flashy lifestyles and incomes of the top 1% are certainly eye-catching, exclusively focusing on them risks obscuring another area of growing concern in inequality – namely the declining situation of low-income households, which is by no means a small group. In recent decades, as much as 40% of the population at the lower end of the distribution has benefited little from economic growth in many countries’ (pp. 21–27).

The concept of the separation of the top 1% and the rest of the population, as argued by Piketty (Reference Piketty2014), Atkinson (Reference Atkinson2015), Stiglitz (Reference Stiglitz2015, Reference Stiglitz2016) and Dorling (Reference Dorling2015a, Reference Dorling2015b), partially highlights some tendencies and dynamics that are changing structural aspects of many OECD countries.

Within the ‘99 percent’ we can distinguish actual social groupings that share common living conditions – in particular in terms of income, work and relation to institutions – which are, however, very different between them. More emphasis should be placed on how the growing concentration of wealth and income in the upper classes is remaking the social stratification and the different groupings that compose the working and middle classes (Reference CromptonCrompton, 2008; Reference Butler and SavageButler & Savage, 1995). The concept of social stratification consists of more than simply inequalities in life chances and refers to the idea that individuals are distributed among social strata that are real social groupings, forged together through both their economic relations and their associated social relations and interactions. A social stratum connects individuals together and helps build up boundaries that enclose groups and divide them from others (Reference Scott and PayneScott, 2013, p. 29).

4. The income divisions within the 99%

To grasp the differences within the 99% Piketty (Reference Piketty2014) suggests observing the specific dynamics that distinguish each decile of income. The income-inequality trajectories of the various deciles are significantly different. The global top 10% income share grew until 2005 and then decreased significantly in 2016, while the income share of the bottom five deciles at the same time grew at a steady rate. Referring to the four intermediate deciles, the income share decreased significantly until 2006 to then increase in the following years (World Inequality Database; Alvadero, Chancel, Piketty, Saez, & Zucman, Reference Alvadero, Chancel, Piketty, Saez and Zucman2018). Moreover, Piketty (Reference Piketty2014) argues that the top decile encompasses two very different worlds. There is ‘the 9 percent’ in which income from work clearly predominates, and ‘the 1 percent’ in which income from capital becomes more important: ‘the frontiers are of course porous, but the differences are nevertheless clear and systematic’ (pp. 280–281). The global top 1% income share increased until 2007 and declined slightly in 2016; the income share of the remaining 9% increased until 2005, and later progressively dropped significantly to 2016 (World Inequality Database; Alvadero et al., Reference Alvadero, Chancel, Piketty, Saez and Zucman2018).

The 1% is not a unified segment of the population: ‘top executives’ (the ‘working rich’) and top capital owners (the ‘rentiers’) are two distinct groups of this segment. Top executives do not replace capital owners but rather ‘co-habitate’ reflecting the changes happened in professional and managerial occupations (Reference Atkinson, Piketty and SaezAtkinson, Piketty, & Saez, 2011). Refering to key concepts of Sassen, the income and the wealth of the 1% takes on a format so extreme that the language of inequality is not enough to seize the transformation that has been taking place. The segment of the population involved considered as a whole forms a ‘social world’ clearly distinct from the rest of society, from the sociality, social responsibilities, values and rules of coexistence shared by the majority, beyond the ‘systemic edges’ of the society.

In the other deciles, divisions are growing; within them, and especially in the middle classes, there is a broad spectrum of people with different life chances, qualifications, remuneration, and social identity (Reference CromptonCrompton, 2008). In much of the recent comparative research, these distinctions rely on income-based definitions: being middle class is defined as having an income within a scale constructed around the median and which has typically been symmetric. The lower and upper cut-off points have been set in diverse and ad hoc ways (Reference RavallionRavallion, 2010). Lester Thurow, the first author who adopted this approach (Reference Thurow1984), defined the middle class as including households with an income between 75 and 125% of median household income.

The lower thresholds of the middle class are naturally linked to the poverty threshold, which is set at 60% (EU) or 50% (OECD) of the national median equivalised disposable income. Ravallion (Reference Ravallion2010) argues that middle class living standards in many developing countries begin where poverty ends. Horrigan and Haugen suggest fixing the lower endpoint of the middle class at an income level significantly above the poverty level. In the EU, Atkinson and Brandolini (Reference Atkinson and Brandolini2011) note that the former criterion would bring us to identify the lower bound of the middle class with the at-risk-of-poverty threshold, set at 60% of the median, whereas the second criterion would rationalise the 75% cut-off as defining the ‘margins’ of poverty plus a quarter of the at-risk-of-poverty threshold (p. 8).

On the ‘margins’ of poverty, Whelan, Nolan, and Maítre (Reference Whelan, Nolan and Maítre2016) place those between 60% and 75% of the median. The middle class can then be said to be those not in poverty or in the margins of poverty, between 75% (lower thresholds) and 166% of the median (upper thresholds). Within this they distinguish a ‘lower middle class’ between 75–125% of the median and an ‘upper middle class’ between 125% and 166% of the median. Those whose incomes are at least 167% of the median are considered as the affluent class (pp. 7–8). Each group tends to have a distinct social identity, live in different conditions and face diverse social risks; however, they no longer form an interrelated network of social relationships and values that might previously have characterised all the groups of the middle classes.

In the margins of poverty a different social condition starts, involving households able to cope with daily expenses only with increasing difficulty, despite having levels of income above the conventional poverty line. The lower bound of this group is just above the poverty threshold, while the upper bound is at 75% of the median. This very small part of the population that sits between the poor and the middle class can be defined as ‘the missing class’ or ‘near poor’, households that earn enough to not be considered poor, but not enough to share in the living conditions of the middle class (Newman & Tan Chen, Reference Newman and Tan Chen2008). From a longitudinal perspective, a large part of the ‘near poor’ is made up of downwardly mobile middle class households, and by the movement of poor households towards a marginally better position.

5. The emerging welfare needs of the fragmented majority

In the last two decades only a small section of the upper middle class has registered an improvement in their condition entering into the established middle class (Savage et al., Reference Savage, Devine, Cunningham, Taylor, Li, Hjellbrekke, Le Roux, Friedman and Miles2015) and a segment became the global professional class (Reference SassenSassen, 2007). On the other hand, the vast majority of middle class families has worsened its living condition and has lost a variety of certainties related to work stability and income. Between the latter and the former the social and economic divides are growing. Moreover, the distinction between this part of the middle class and the working class in terms of income, consumption and quality of work is getting smaller.

For these groups in the middle of the social stratification it is difficult to draw clear boundaries compared to the specific economic, social and cultural capital of each class (Reference CromptonCrompton, 2008; Savage et al., Reference Savage, Devine, Cunningham, Taylor, Li, Hjellbrekke, Le Roux, Friedman and Miles2015). Many authors show that a consistent part of the middle and working classes is at the same growing risk of exposure to short-term and occasional poverty (Reference JenkinsJenkins, 2011; Reference Nolan and WhelanNolan & Whelan, 2011; Whelan & Maître, Reference Whelan and Maître2008; Whelan et al., Reference Whelan, Nolan and Maítre2016). Furthermore, an OECD analysis (2015) indicates that the biggest impact of inequality is the growing gap between lower income households and the rest of the population. This is true not just for the very lowest earners – the bottom 10% – but for a much broader swathe of low earners – the bottom 40%. The analysis suggests that countering the negative effect of inequality on growth is thus not just about tackling poverty but about addressing low incomes more broadly; especially since the dominant mechanism through which inequality seems to affect growth is by curbing opportunities for the poor, the working class and the lower middle class (pp. 21–27). A report by Eurofound (2015) shows that the lower quintiles of the working population (up to 60%) receive around 35% of the total wage mass, while the share of wages in the 60–80% quintile is around 23% and in the remaining upper quintile around 43%. The widening of the wage gap is the result of both high growing earnings at the top and declining earnings at the bottom. Over the last 10 years, in countries as United States and Canada, the size of middle class has decreased by 1.2 percentage points (Reference PressmannPressmann, 2017), instead the share of the population that considers itself as belonging to the middle class has fallen significantly and, in the same countries, it has fallen from two-thirds to one half. The drop in middle class perceptions may reflect rising economic insecurity, accompanied by a lack of confidence in future labour market and economic prospects. Higher job insecurity and income volatility may have pushed middle-class self-identification down as people feel less certain about maintaining their socioeconomic status in the future (OECD, 2016). There is substantial risk for middle-income households to fall into low income and poverty over their life course: one-in-seven households, and one-in-five of those living closer to those with lower income slide into the bottom 20% over a four-year period. There are also signs that these risks have increased over the past two decades. In particular, a further divide among the middle classes appeared in several countries: for those closer to lower income, the risk to further slide down has increased. At the same time, those closer to the middle and the more affluent members of the middle class today have lower risks to fall into low income. There are thus signs that the middle class may be fracturing (OECD, 2018, pp. 15–16).

Taken together, these findings indicate changes in the middle of the social stratification that have not been adequately emphasised, even though they are extremely relevant to how variations in the values and norms of civil coexistence occur and in the organisation of welfare services.

In the middle of the social stratification, between the established middle class and households in persistent poverty, a social stratum is consolidating, a sort of what Harold L. Wilensky (Reference Wilensky1975, pp. 118–119; Reference Wilensky2002) called ‘middle mass’ (the lower middle class and the upper-working class). Like Wilensky's middle mass, this extended social stratum comprises of a large part of the middle and working classes. Unlike the ‘middle mass’, this group is a sort of multitude with a scarcely defined social identity, which however shares conditions of economic difficulty and feels unable to deal with the new risks of a globalised society. This social stratum is afraid that its current condition of economic difficulty will become stable, with little hope of increasing its income and opportunities for upward mobility in the near future. It is a social stratum that is particularly faced with processes of individualisation that further fragment group homogeneity and weaken traditional and modern forms of belonging and social ties. People in this group appear to be less and less able to build their own life projects that would allow them to survive amidst growing insecurity. The condition of this social stratum constitutes one of the most significant critical issues of strategies aimed at contrast social instability and the degrading of social life. The attitudes emerging among these groups strike at the functioning of institutions and weaken civic values.

In his seminal study (Reference Lipset1960), Lipset claimed that political radicalism in the form of Fascism and National Socialism, constituted an ‘extremism of the center’ or ‘extremism of the middle class’. Betz (Reference Betz1994) argued that the rise and success of the ‘extremism of the center’ and middle class disenchantment with and profound distrust of major social and political institutions can be interpreted as the result of an increasing social and cultural fragmentation, relative deprivation and loss of status (or fear of the loss of status). Empirical research clearly shows that workers and the old middle classes are indeed overrepresented among ‘radical’ voters (Reference Lubbers, Gijsberts and ScheepersLubbers, Gijsberts, & Scheepers, 2002; Reference NorrisNorris, 2005; Reference RydgrenRydgren, 2007). The rising inequality in advanced economies, in conjunction with job insecurity and stagnating real incomes for a segment of the population, has led to growing public backlash against globalisation (International Monetary Fund, 2017). Among the middle and working classes confidence in institutions, civic values and trust in other people, emancipative ideals and emancipative forms of social capital have been declining and represent a threat for stable democracy (Reference Inglehart and WelzelInglehart & Welzel, 2005).

In recent years, the middle and working classes are not only disenchanted with major social and political institutions. In many developed countries, rather than passive attitudes and behaviours what prevails is an active individualism which is less and less integrated. This active individualism is mainly created by the radical processes of individualisation of the second modernity (Reference BeckBeck, 1992) that has weakened the ties between individuals and their families, has liberated people from traditional roles and constraints and undermined ‘intermediates institutions’ as trade unions, political parties and ethnic groups. At the same time, these processes are not often creating new forms of reintegration and control: in many cases, individualisation means ‘disembedding’ without ‘re-embedding’ (Reference BeckBeck, 2016; Beck & Beck-Gernsheim, Reference Beck and Beck-Gernsheim2001, pp. 202–203).

The enabling state is less and less able to manage and include this multitude of active and reflexive individuals with weak collective ties. Economic and social changes are weakening a founding principle of competitive societies that want to valorise individual effort: individuals are builders of their biographies, but they should pursue their life projects in disciplined ways, respecting the rules and social institutions which they themselves contribute to through their active and dynamic participation in society. First Parsons (Reference Parsons1937, Reference Parsons1978) and then Beck (Reference Beck1992) and Beck and Beck-Gernsheim (Reference Beck and Beck-Gernsheim1995, Reference Beck and Beck-Gernsheim2001) called this kind of individualism ‘institutionalized individualism’ oriented by norms and values which form the functioning of the central institutions of society. The expression ‘institutionalized individualism’ sums up and reconciles the two aspects of Parson's theorization: one emphasising the objectivity of the social order, the other the activity of individuals. The individual moves within an action space, wherein they find both support for their action and impediments to it, both resources and constraints (Reference BourricaudBourricaud, 1981, pp. 14–15).

The middle classes historically personified, in many respects, the part of society which intended to actualise this kind of individualism with the calculated aspiration to climb the social ladder and improve their living conditions through education and hard work (Bellah, Madsen, Sullivan, Swidler, & Tipton, Reference Bellah, Madsen, Sullivan, Swidler and Tipton1996, pp. 142–165). In the recent decades the ‘institutionalized individualism’ is no longer able to promote the integration in an active society of relevant social groups. The extended social stratum that comprises a large part of the middle and working classes has become a sort of multitude which believes largely to have lost the economic and social conditions needed to build his life project. In a social context marked by heightened risk and offering few opportunities for social mobility other forms of individualism are growing between these social groups which are less and less functional and beyond the control of social institutions. To the ‘institutionalised individualism’ of Parsons and Beck may be juxtaposed ‘survival strategies’, in which individuals put together very different values and ways of life, which are neither regulated nor easily manageable and are characterised by a confusing array of aspirations, resentments, and intolerance towards rules and institutions. This non-apathetic discontent for what one is surrounded by extends to every aspect of social bonds, rules and institutions. The condition of this social stratum constitutes one of the most significant critical issues of strategies aimed at contrast social instability and the degrading of social life.

The attitudes emerging among these groups strike institutions’ ability to function and weaken civic values. This type of individualism is much harder to govern, also because of the social divisions between the 1% and the ruling classes in their entirety and the rest. A large part of the ruling class no longer intends to pursue a proactive, mediating role in supporting institutional rules and functions. In many OECD countries, those at the top feel outside of the resources that the system normally offers (state schools, public healthcare, social services) but also outside the duties and rules of the social life of the majority (Reference SizaSiza, 2018). The upper classes decline to take social responsibility, to represent collective interests, to act as a role model and to create limits and criteria that help distinguish between morally acceptable and morally reprehensible behaviour (Reference SassenSassen, 2014, p. 15).

6. Subordinating the ‘new welfare state’ to prevailing economic logics

In his analysis on the declining middle class based on microdata of the Luxembourg Income Study, Pressman found considerable empirical support for a Keynesian hypothesis: government spending and fiscal policy largely determine the reduction of the size of the middle class and the balance between downward and upward mobility (Reference Pressman2007, Reference Pressman2010, Reference Pressmann2017). The impact has been quite remarkable in a number of countries. The more progressive the national tax system, and the more generous government spending programmes, the larger is the proportion of middle class households (Reference PressmanPressman, 2007, Reference Pressman2010, Reference Pressmann2017). A longitudinal research, based on the same microdata, analysis the changes of middle classes in 19 countries between 1985 and 2005 to understand to what extent public policies mitigate the market position of different income groups. Thanks to the welfare state, the percentage of the population classified as middle class rises, though not uniformly across nations. In the US and Switzerland, taxes and transfers enlarge the middle class only by a few percentage points. The largest middle classes can be found in Sweden, Norway and Germany, to a large extent an achievement of the welfare state. Based on market income alone, the middle class in both Sweden and Germany is rather modest (15%), but with taxes and transfers, it increases to more than 40% of the population (Reference DallingerDallinger, 2013). Other authors (Reference Hacker and PiersonHacker & Pierson, 2010; Reference Garfinkel, Rainwater and SmeedingGarfinkel, Rainwater, & Smeeding, 2005) argue that the middle class has been eroded in countries with low social expenditures and a marginal welfare state.

As Hacker and Pierson (Reference Hacker and Pierson2010) have noted, late capitalism ‘has turned its back on the middle class’, job security has been eroded, and pillars of security and opportunity, such as education, pensions and social insurance and other welfare spheres, have come under strain (Reference ParkerParker, 2013). Hacker and Pierson mainly refer to the US, but the situation in Europe is similar. In many European countries, significant changes have reduced the capacity of welfare to address the needs not only of households in a condition of poverty, but also those of the central classes of the social stratification. New forms of dualization, in terms of access to benefits, efficiency and quality of services (Reference Palier and KathleenPalier & Kathleen, 2010; Seeleib-Kaiser, Saunders, & Naczyk, Reference Seeleib-Kaiser, Saunders, Naczyk, Emmenegger, Häusermann, Palier and Seeleib-Kaiser2011) have impacts on the majority the middle class that are underestimated by policymakers and in academic and political debates. Dualization is described by Emmenegger et al. (Reference Emmenegger, Häusermann, Palier and Seeleib-Kaiser2012, p. 10) as a ‘differentiation of right, entitlement and services provided to different categories of recipient’. It may take on three different forms: a deepening of existing institutional dualism, that is, the differential treatment of insiders and outsiders becoming more pronounced; a widening of existing institutional dualisms, that is, groups that have been previously treated like insiders increasingly treated like outsiders; or finally, it may be the product of a new institutional dualism (Emmenegger, Häusermann, Palier, & Seeleib-Kaiser, Reference Emmenegger, Häusermann, Palier and Seeleib-Kaiser2012, p. 10).

The last two decades have witnessed a significant decrease in the numbers of the most privileged categories of worker – those with the most protection – as well as a further erosion of job protection, pension benefits, and social security benefits for outsiders. The degree of dualization increased further: being an outsider does not have the same consequences in all countries and welfare state regimes but depends on the particular social policy and labour market institutions of each country (Häusermann & Scwander, Reference Häusermann and Scwander2009). Pension reforms have reduced public pension benefits, gradually extended the official retirement age, and fostered privately funded pensions (Reference EbbinghausEbbinghaus, 2011). New forms of dualization are also emerging in other realms. For example, the protection offered by the public health system is no longer sufficient and the population is increasingly divided between those who enjoy adequate private protection and those who are covered by very limited private insurance and occupational schemes. In terms of access to health care, the support offered by the public health system is no longer sufficient. Many countries have achieved universal (or near-universal) coverage of health care costs for a core set of services, but since the onset of the economic crisis the share of the population covered has decreased in many countries (OECD, 2015).

The politics of retrenchment has worsened the living conditions of the majority in the last 20–30 years. Welfare retrenchment involves more than spending cuts alone. It presupposes a limited area of intervention for public welfare and is aimed at a reduction of the number of beneficiaries: those deemed to be in conditions of severe deprivation and who cannot gain the material resources to live independently. It has been thought that social democratic or corporatist welfare-state regime-types that tend to be universalistic and which were founded by a class coalition would be less likely to suffer the politics of welfare retrenchment (Reference Esping-AndersenEsping-Andersen, 1990, p. 33). Le Grand's analysis of expenditure under the Conservative Government in the UK (Reference Le Grand and Winter1986) showed that those services most used by the middle-classes were protected from the expenditure cuts implemented as part of its austerity programme.

Actually, the middle classes do not represent an obstacle to such policies anymore. They are encouraged to opt into the private welfare through tax incentives (Reference Horsfal and HudsonHorsfal & Hudson, 2017).

New instruments for welfare retrenchment are being used in many countries almost independently from welfare-state regime-types and the involvement of the middle class: deep economic crises or the need to reduce budget deficits and restore economic growth and employment have become the principal circumstances that justify the politics of welfare retrenchment (Reference Levy, Castles, Leibfried, Lewis, Obinger and PiersonLevy, 2012).

The living conditions created by economic and technological changes and by weak or misjudged responses of the policymakers place the central classes of the social stratification at risk of not having sufficient public services, at the same time as having only a limited ability to access complementary welfare. Public services contribute less and less to the improvement of living conditions of this social stratum. Moreover, private welfare provision plays a compensatory role almost exclusively only for social groups with consolidated incomes and employment, and is not accessible to a large part of the population.

7. Rethinking activation policy

The social investment perspective aims to promote an active society and to better address the new social risks, to encourage active participation in the mainstream of social and economic life (Esping-Andersen, Gallie, Hemerijck, & Myles, Reference Esping-Andersen, Gallie, Hemerijck and Myles2002; Reference Morel, Palier and PalmeMorel, Palier, & Palme, 2012). The relation between social investment-oriented policies and inequality variations has been the focus of a large part of the systematic and comparative empirical studies. Some research has claimed that there is no evidence that focusing on active social welfare and new social risks disturbs the egalitarian ambitions of the welfare state (Reference VaalavuoVaalavuo, 2013). In contrast, other authors note that there is some evidence of a correlation between poverty and inequality variations and shifts in expenditure for social investment programmes (Reference CantillonCantillon, 2011; Ghysels & Van Launcher, Reference Ghysels and Van Launcher2011; Van Vliet & Wang, Reference Van Vliet and Wang2015). In many cases, activation programmes have created new forms of exclusion, new social divides and new stratifications rather than reduce inequality (Pintelon, Cantillon, Van den Bosch, & Whelan, Reference Pintelon, Cantillon, Van den Bosch and Whelan2013). While the impact of the transformation of the welfare state based on the social investment logic should not be underestimated, Van Kersbergen and Hemerijck (Reference Van Kersbergen and Hemerijck2012) argue that rising inequality has probably not been directly caused by this and that ‘further empirical analysis is therefore needed’ (p. 489).

In the academic and public debate on social investment other criticisms of this perspective are not adequately explored, although referring to central features and crucial dimensions of its conceptualisation. In particular, I am referring to policies oriented towards individual autonomy and individual responsibilities. The central purpose of the social investment approach is to promote a renewed integration of the social and economic spheres. In fact, governments’ capacity to improve both economic and social outcomes was weak in the Nineties (Reference Taylor-GoobyTaylor-Gooby, 2004) and has further decreased in the last decade. Contemporary capitalism appears a society whose system integration is critically and irremediably weakened (Reference StreeckStreeck, 2016, pp. 13–14). Many societies are unable to reconcile the needs of economic growth with the quality of work, remuneration levels and the quality of human relations and to manage the new economic and social divisions that are creating.

The active individual advocated by the social investment perspective founded his life project on fragile and very contradictory conditions. The economic crisis has further dissolved securities and life projects, and has emphasised the limits and ambiguities of the processes and forms of individualisation that have prevailed in these years. For many social groups, the economic and care resources needed to face old and new risks have been decreasing significantly and the existing solutions proposed by social investment-oriented policies appear less and less accessible and effective. The welfare needs of the middle classes are changing. Existing research shows that the more educated segment of the middle class remains a strong supporter of the social investment development and its emphasis on future distributive effects of activation, human capital formation and early childcare education. A large part of the middle and the working class demand instead for programmes that secure minimum protection and immediate welfare benefits for addressing their economic insecurity (Beramendi, Häusermann, Kitschelt, & Kriesi, Reference Beramendi, Häusermann, Kitschelt and Kriesi2015; Gingrich & Silja Häusermann, Reference Gingrich and Silja Häusermann2015; Häusermann & Palier, Reference Häusermann, Palier and Hemerijck2017). Nevertheless, in many countries there is a tendency to believe that activation may substitute conventional income maintenance guarantees. The focus increasingly placed on individual responsibility and labour market participation is restricting the rights of social beneficiaries (Cantillon and Vandenbroucke, Reference Cantillon and Vandenbroucke2014).

What we need is a new focus for social policy: the old social risks have not disappeared and the minimisation of poverty and income insecurity is a precondition for an effective social investment strategy (Esping-Anderson et al. Reference Esping-Andersen, Gallie, Hemerijck and Myles2002, p. 5; Reference HemerijckHemerijck, 2017). Autonomous life projects require a basic security net that is independent from paid work (Reference Beck and Beck-GernsheimBeck & Beck-Gernsheim, 2001) and which is constructed by uniting formal and informal relationships of social support and care. A new focus does not place into question the logic of activation, inasmuch as people's autonomy and capacity to take the initiative are social resources that should also be reinforced by social policy. In highly individualised societies, in which individuals are not passive reflections of circumstances and where individual agency assumes a central place (Reference BeckBeck, 1992), the autonomy and desire to construct one's own life project should not be seen as an obstacle, or something to be constrained within standardised schemes, but could be recognised as a real resource for activation programmes. Therefore, the aims of non-standardised activation measures could be to reinforce social networks, giving them the opportunity to realise their projects even only in part and to influence the choices and the direction of policy. Middle class disenchantment and its profound distrust of major social and political institutions could be dealt with through more personalised interventions, recognising care work in the family as equally important to paid work (Reference LarsenLarsen, 2004), promoting activation programmes in which economic support is part of a broader aim to strengthen nets of social support and to instill more confidence in institutions.

In their books Wilkinson and Pickett provide compelling evidence that changes in inequality and trust always go together: inequality corrodes trust, divides people and deteriorates the quality of social relations (Reference Wilkinson and Pickett2009, pp. 49–56) reduces social cohesion and participation in community life; more unequal societies become more fragmented as social distance increases (Reference Wilkinson and Pickett2018, pp. 56–58). People in more economically unequal countries trust each other less and think of many others around them as less deserving. Lower trust and greater inequality are highly correlated (Reference DorlingDorling, 2016, p. 29).

Active social policy may reasonably contribute to overcoming many social divisions and to rebuilding a sense of belonging and strengthening civic values and trust in other people. This would help in the middle class feeling part of the decision-making process of public projects. This could become part of a social policy project to increase individual autonomy and at the same time to create opportunities for families and individuals to take part in collective activities which are meaningful and useful to the community. Active social policy can contribute to promoting new forms of social interaction in many spheres of life and new care relationships compatible with the requirements of an individualised existence and individual autonomy (Reference Beck and Beck-GernsheimBeck & Beck-Gernsheim, 2001). It is in the realm of social relations that a solution to the widespread degradation of social life is to be sought. Society is promoted not merely by a relationship of rights and duties, but to an ever greater and more fundamental extent by relationships of reciprocity, solidarity, cooperation, which is, a new meaning of what is considered ‘civil’ (Donati Reference Donati2011; Reference Donati and ArcherDonati & Archer, 2015, p. 258).

8. Concluding remarks

In most wealthy countries the division separating the very richest from the rest is undeniable, as are the decline in the living standards of middle class households, the loss of economic security and opportunities for upward mobility. By applying Castel's concept of the three ‘zones’ (Reference Castel2003) it can be seen that the middle class is no longer the zone of integration, but is becoming more a zone of vulnerability and of precarity.

In my paper, I have argued that in the academic and political debate many other types of economic and social divisions and declines have not been adequately emphasised. In many OECD countries, there has been a growth of an extended social stratum formed by the majority of individuals. This extended social stratum comprises of a large part of the middle and working classes, who feel, to various degrees, that they are in a shared condition of economic difficulty, lacking security, status and trust in the future. Not only have confidence in institutions, civic values and trust in other people been declining among this group, but also there has been an undermining of ‘institutionalized individualism’ as a founding social principle. Moreover, the upper classes and the established middle class do not feel the social responsibility to mediate between different interests and promote values, resources and conditions for social integration.

Often, social policies cannot conceptually grasp and tackle these divides with adequate policy actions and therefore are unable to cope with their social effects. They appear less and less capable of protecting the majority from the growing risk of instability, and do little to foster social mobility or improve their social and economic condition. In many countries ‘the new welfare state’ has created new forms of exclusion, social divides and stratifications rather than reduce inequality. Reductions in the welfare state are worsening the living conditions of families who are at risk of not having sufficient public services, and which have limited possibilities to access a supplementary welfare.

In my article, I have observed that a much broader concept and design of activation is needed, founded on a combination of income support and social promotion. Starting with the person and their life project rather than the service, active social policy could be able to mobilise the productive potential of citizens and capable of contributing to the rebuilding of a sense of belonging and identity. A new welfare configuration does not place into question the logic of activation, inasmuch as people's autonomy and capacity to take the initiative are social resources that should also be reinforced by social policy.

The financial and economic crisis has created a strong disorientation especially among the middle classes, yet in the long term it could allow for the consolidation of new collaborative relationships, new opportunities for social growth and different ways of looking at the world. A civil coexistence differently organised, which has as its base a greater reciprocal trust, could be capable of giving vitality and sustenance to individuals who are currently isolated and lacking trust in traditional collective bonds. As Beck (Reference Beck2016) reminded us, it is possible to increase the degree of social integration only if we are able to mobilise and motivate people to meet the challenges that lie at the centre of their lives.