A generally marginal sector, of paradoxical importance

6 Eurostat collected statistics in 2014, which were analysed in terms of purchasing power parity (PPP) between the poorest and richest countries of the European Union, above all those with the most generous benefits aside from the special case of Luxembourg. According to these statistics, social protection benefits as a whole range from 1 to more than 4.2, that is, from 2 461 euros per person to 10 644 euros [5]. This is the most prominent feature of social assistance in Europe: there is a major disparity between countries. This gap is particularly wide between the countries of northern Europe, the countries of southern Europe, and the countries of central Europe. This highlights the considerable differences in living standards within the Union. Second, the welfare sector is still a marginal part of social protection. This is apparent from an examination of table 1 in the annex, which compares the social expenditure devoted to unemployment and social exclusion as a part of overall social protection in a selection of the 28 countries. It should be noted that the content of this table illustrates the complexity of the comparisons of social assistance. This can be seen through the expenditure devoted to social exclusion in which the percentages have explanations that are primarily institutional, but also economic, and even cultural-political, e.g., the notable weaknesses in Greece and Italy and the inverse case of Bulgaria. There are also specific explanations for the relative and cyclical importance of unemployment, e.g., the case of Spain, and the inverse case of Germany. Apart from Spain, Southern and Central European countries spend less than 5% of their social spending on unemployment and “last resort” benefits, and those devoted to social exclusion are generally less than 1%. The resulting inequalities in terms of poverty are a third essential feature. They can be analysed via the so-called AROPE (At Risk of Poverty and Social Exclusion) indicator, which has now been included in the auxiliary indicators of the European semester monitoring procedure [6]. Table 2 illustrates the very clear differences over the period that includes the 2008 economic crisis between the former communist countries, the countries of the South—together with the United Kingdom—and the countries of Northern Europe, to which France is close. Changes in the various indicators should be interpreted in the light of the specific political and economic circumstances of each country and its institutional reforms. Certain results appear significant in the group of Central European countries, which does not prevent the relative level of this indicator from remaining high or very high. For the rest, we will obviously note the case of the very high poverty rates in Greece after 2012, and also the relatively high rates in the Latin countries (the value of the indicator for Portugal—which is not shown in the table—is close to that for Spain), which tend to increase after 2010. In the richest countries, Germany’s relatively high level stands out in contrast to its reputation for economic prosperity. The case of France also approaches the group of Scandinavian countries at the end of the period with an AROPE rate more than two points lower than that of Germany. This development is, moreover, within the framework of a radical failure of the Europe 2020 strategy as regards its objective of reducing the number of poor in the European Union by 20 million, an objective that will not be achieved in 2020 [7]. However, it is not possible to establish precise causal links or correlation between AROPE values and well-identified factors such as policy developments. Here we find the particular difficulty of evaluation already reported in the field of social assistance. The internal inequality of the European Union is just as marked in Table 2 as in the previous table. What dominates in the Northern countries (including France) is a certain stability of rates, including in the United Kingdom, even in the worst years of the crisis. However, even in these countries, there is no doubt that the public authorities’ response in terms of benefits is largely inadequate. The case of the Scandinavian countries is specific—given their egalitarian orientation, which, however, has tended to diminish—because they are still, in 2017, the least unequal societies in the Union. Finally, the relative inadequacy of social assistance benefits is a general feature (Table 3). Assistance benefits in the new member states in 2010 are generally equivalent to one third of the level of the monetary poverty threshold. This rate is around half in the countries of the South (excluding Greece), and between half and two thirds for the other countries, except Denmark which is the exception of generosity. These elements are not new. It is therefore not surprising that the experts brought together by the European Commission find it very difficult to identify positive aspects in their comparative assessment of social assistance systems. These experts selected five criteria in their 2016 report (Frazer and Marlier, 2016): adequacy of systems, coverage rate, the non-take-up of rights, impact on poverty reduction, and linkage with active labour market policies and other quality services. They have formulated a very global judgement on all these criteria. According to them, only two European Union countries (Cyprus and the Netherlands) are really qualified as to adequacy. As for the impact on poverty, only the Netherlands, the United Kingdom and Ireland are considered to have a system that has had a strong impact on the reduction of the indicator. Seven countries have had an impact on the intensity of poverty: Austria, Belgium, Cyprus, Ireland, Luxembourg, Portugal, and the United Kingdom (ibid., pp. 35–36). On all other criteria, the assessment is “partial” or “inadequate”. As far as an assessment at this aggregate level is possible, effective links with active policies have been noted only in two countries, Hungary and Malta. For the rest, the vast majority of countries have been classified in an “average” category, obviously very random in its determination. The Commission document suffers from approximations, taking into account its method. [8]

7 Ultimately, the overall conclusion, confirmed by other studies (Immervoll, 2009; Nelson, 2013; Pfeiffer, 2012), is that although the most vulnerable people in the majority of countries depend on these assistance delivery systems, they are generally inadequate and still cover an insufficient fraction of the needs of the poor and socially excluded. These conclusions are not recent (see Pfeiffer, 2012, which uses data from the 2000’s, prior to the crisis, and which reviews previous work for the richest countries), and do not stem from a sudden deterioration due to the economic crisis. “They are structurally and permanently necessary”. The assessment of the Commission’s experts on the link with so-called “active” policies also casts doubt on their effectiveness (Frazer and Marlier, 2016, p. 36), which we shall now turn to in more detail. It should be noted that as new studies are published (Immervoll, 2009; Pfeiffer, 2012; Wang and van Vliet, 2016), varied general analyses arrive at comparable conclusions.

The great reform of the 1980s and 1990s

8 There is now adequate perspective to report on welfare reform—which has generally taken the name of “activation policy”. Very often, this reform has been described on the basis of simplified features corresponding to political slogans in the different countries. These slogans are translated into English (or from English), which means that one could believe all reforms to be identical. Thus, the theme “rights and obligations” is often mentioned, which, for example in France, forces the beneficiary into a legal situation that obliges him to take certain steps to seek employment or activity. In Danish, the formula presented to the beneficiary is ret og pligt, but the content of the rights and obligations is extremely different from what it is in France. In Great Britain, all services are now subject to a compulsory work test, which does not exist in most other countries. As is to be expected, reforms have been undertaken according to very different timetables and methods, as well as with different ideological, moral and political justifications. Access to a solidarity benefit when resources are lacking is, in fact, typically an operation that involves solidarity within the political community (Rothstein, 1998; Barbier, 2008). We must therefore proceed in two stages: first, we must explain how the reforms have followed one another over time in Europe. France was the pioneering country here with its minimum insertion income [Revenue minimum d’insertion-RMI]—at the same time as the United States, with the Family Support Act and the associated workfare. Germany, on the other hand, was the last. Secondly, one can draw from the examination of this history the essential features that characterize the reforms, up to the point at which we can form general categories. It is by moving on to generality that we can accurately gauge what is common and what is specific to different categories and countries.

Social protection activation strategies in Europe and the United States

9 In 2017, the term “activation” is commonly used in different languages without any complement in social science research on social policies (Aktivierung; activation; attivazione; aktivering, etc.). The expression is most often unknown to people who receive assistance, except in the case where the administration addresses them utilising this word, as has been the case in Denmark since the early 1990s. This expression actually comes directly from a “political language” that really took hold in the second half of the 1990s. Few researchers have questioned the term itself, many simply taking up the notion from the mouths of politicians without trying to define it as a concept of social policy. In the early 2000s, a definition was suggested (Barbier and Ludwig-Mayerhofer, 2004) based on a paired comparison of eight countries (United Kingdom, Slovenia, Italy, France, Denmark, Norway, Germany and Austria). To summarize, reform since the 1990s has essentially consisted of two objectives: requiring people receiving benefits to change their behaviour and transforming social protection systems to make them more compatible with job creation.

10 We will avoid using the expression “activate the unemployed” or “activate the poor” (Barbier, 2011, p. 48), because it is ethically questionable. We generally “activate” mechanisms and things, but not people. It is a mark of the present period that verbs traditionally reserved for action on things are being increasingly used for people. Much confusion has crept into the initial uncertainty of defining a type of reform designed to “treat” assistance recipients. The history of these programmes make it possible to observe at least two of them: that of workfare, and that of the originality of the programmes in the form of “active employment policies” as compared to those which existed previously. The historical review involving a large number of countries makes it possible to understand why political leaders have seized on the slogan in their political action. This historical retrospective work cannot, however, be exhaustive here [9]. Studying this history is all the more important as the definitional and practical bases of the so-called “activation” policies are still essentially in force in 2017. Of course, many changes have been made, as H. Jørgensen (2013) has shown in detail concerning Denmark. Finally, while the literature on the forms of “activation of the unemployed” is very abundant, that on activation of systems is much less so. The country where the shock of change was strongest was Germany, as shown in the book coordinated by Eichhorst et al (2008). Several analyses tended to focus only on changes in the situation of the unemployed and those receiving assistance, even as broader reforms took place—for example, the powerful expansion of tax credits or, in France, the reduction of social contributions.

Swedish active policies and contemporary activation policies

11 We should first note the origin of the use of the word “active” in the field of social and employment policies in Europe. Sweden invented the concept of politiques actives du marché du travail [active labour market policies]—also known at the time as politiques actives de la main-d’œuvre [active labour policies]—(Barbier and Sylla, 2004). Debated from the 1930s, it was implemented in the 1950s. Historically, this has remained an “original concept of activation” as the cornerstone of Swedish employment policy. This implied that any unemployed person who cannot find a job should first be offered an appropriate job training or other measure. Combined with a centralised, solidarity-based wage policy and a restrictive budgetary policy [10], the active employment policies thus designed have for some years represented what has been called the Rehn-Meidner model. Indeed, it was Gösta Rehn himself, as head of the OECD Employment and Social Affairs Secretariat in Paris, who promoted the activation of labour market policies from 1962 to 1973 (Wadenjsö, 2001). A first international reference to “activation”, which, with hindsight, can be considered as “old activation”, thus took shape. As dominant economic views changed both at the head of the OECD and in the larger nations, the notion of active policies subsequently changed in nature. As employment policy—in the sense of the previously advocated “full employment policy”—disappeared from recommendations, employment policies (in the plural) became diversified interventions to assist new recommendations based on monetarist and neoliberal references. Structural reforms—including labour market reforms—were then systematically discussed. The OECD rapidly changed the Rehn and Meidner [11] conception from 1964 onwards, detaching it from its interactions with economic and income policy, to form a critique of what the organisation referred to as “passive policies” (unemployment benefits and early retirement). Hence the very important consequences of the differences between the “old conception” of policy activation and the contemporary one: active policies (social or employment policies) have become equivalent to labour supply policies. Under the guise of orthodox economic policy justifications, these measures aimed, at least partially, at delegitimising social protection (assistance and unemployment insurance) considered likely to discourage the search for employment. However, until the second half of the 1990s, the OECD made no reference to a policy or “strategy of activation”. It took other innovations for this slogan to triumph. Better still, at least at the OECD, the new vocabulary has not blurred the distinction with the classic “labour market policies” (often referred to in French as “politiques de l’emploi”, thus risking confusion with full employment policies [12]). Even today, OECD officials and its leading experts insist on this distinction (Martin, 2014).

Workfare and insertion

12 While these new employment policies based on the dominant Keynesianism were being implemented in various ways in Europe (Barbier and Gautié, 1998), a commitment to reforming welfare and unemployment coverage gradually took shape in various countries. Discussions developed rapidly in the Scandinavian countries. However, from the point of view of achievements, the latter were not the first to take the initiative on the subject of assistance recipients. In these countries (Norway, Sweden, Denmark), there were many such recipients since the 1970s who received relatively high unemployment insurance benefits, in connection with the spirit of the “general law” which inspires the Nordic regimes (Rothstein, 1998). However, the comparative literature, particularly in France, has not been sufficiently aware of the fact that the pioneering countries of what we have the retrospective right to call “activation before the letter” were the United States and France. For the former country, the Family Support Act was passed by Congress in 1988 under Ronald Reagan, generalizing the use of workfare for the implementation of the Aid for Families with Dependent Children (AFDC) program (King, 1998, p. 276–277). The term “workfare” had been invented long before, under the presidency of Richard Nixon, the word (combining “work” and “welfare”) having been formed by the journalist William Safire. However, a significant part of the literature unfortunately used this expression outside its American context (in particular Rosanvallon, 1995). Nixon utilised this expression with great success in 1969. Republican critics of assistance to poor families have used it polemically to impose programs that force recipients to work—or to accept labour market activities in exchange for receiving welfare (Morel, 2000). These programmes, in existence since the 1970s, became widely known with the 1996 welfare reform under Bill Clinton, which finally tightened eligibility requirements for social assistance for recipients, particularly for single mothers. The first “activation” of the unemployed in the contemporary United States consisted mainly in implementing punishments (through work). With Clinton, this benefited from “bi-partisan” support. Both sides of the political spectrum agreed to “End welfare as we know it”, which was Clinton’s slogan. At exactly the same time, the minimum insertion income (RMI) law was passed in France, bringing a new stage to the policy of social integration that had been developed in many forms since the 1970s (Eme, 1997). While France was also among the pioneer countries, it did not at all utilise the same logic as the American punitive approach (Morel, 2000). At that time, no one in France would have thought of French integration policy as an “activation” reform. However, the contrast between the two initiatives, American and French, would change perspective over time. What some researchers found, a posteriori, was that the RMI reform was an almost contradictory option (Barbier and Théret, 2001) as compared to the Clinton reform that was taking shape. The notion of integration (Maclouf, 1992), initially including young people followed by people with disabilities, would prove to be a sustainable French innovation in social policy. On the one hand is punitive reform, threatening to impose sanctions on those who are cheating or supposedly reluctant to work. On the other hand, we find the republican aim of bringing all citizens into their rights, even when they are poor and excluded. It is well known that these stylized forms were often different from the prosaic reality of implementation, until the new reforms that were subsequently introduced. In any case, although the word was never used, neither in the United States nor in France, the two countries at the time conducted a “form of activation of social assistance before the letter by modifying public policies in a very important way. This change in both countries focused on the poor and not on the unemployed as such.” This distinction is very important: indeed, alongside these innovations which were assuming an unprecedented political importance [13] on both sides of the Atlantic, the traditional forms of unemployment compensation in both countries were of course still in force, as were the forms inherited from the past from labour market policies (again with great differences between the two pioneer countries).

Danish aktivering

13 Political analysis, in conjunction with a debate on the place of work in society, continued in the Scandinavian countries. In Sweden and Norway, a form of “activation” of assistance which at first did not take that name, was introduced in the first half of the 1990s under the label of the “line of work” (in Norway, “arbejdslinjen” [Dahl and Dropping, 2001, pp. 270–77]; in Sweden [Hort, 2001, pp. 252 ff.], “arbetslinje”). Moreover, researchers in Norway, who were familiar with American developments, began to speak of reforms as “workfare” (Lødemel and Trickey, 2000), using the American word with their American colleagues. Already at that time, the justifications that would later be generalized elsewhere as to the preference of work (or “work first”) over assistance, were present. However, it was Denmark that would become famous with its introduction of the aktivering policy. The word hit the nail on the head: “activation” in the strict sense used in connivance between the ruling Social Democratic Party and the trade unions that were closely involved in the implementation of these programmes. Denmark, at the time, was entering a phase of activism on the European scene, putting forward its “flexicurity model”, which it would later promote. Aktivering and “flexicurity” were closely linked, and both reforms had a clear (classic) rivalry dimension with Sweden. The policy of activating assistance and the unemployed was deployed in Denmark starting in 1992, under the socialist government of P. Nyrup Rasmussen, in three stages: the first reform concerned young people and obliged them to accept either a job or training if the administration offered them it to them. The introduction of this first reform was a success and young people stopped registering for assistance services. They were “activated” by the threat of the imposition of a job, and, the authorities said, by the temptation to train themselves. The measure of “activation” of persons was then generalised to all the unemployed. Then, in 1998, it was extended to the beneficiaries of welfare assistance. Of course, people eligible for assistance, which in Denmark is the responsibility of municipalities, were not all immediately considered “available for the labour market”, and the history of aktivering in Denmark illustrates a long series of reforms that followed those introduced in the 1990s. Each time, it was a question of being more and more demanding with the unemployed and the beneficiaries of assistance (Jørgensen, 2013).

Great Britain and Germany

14 The United Kingdom has been among the most fervent supporters of “activation” in Europe, starting in 1997 and with the arrival of New Labour in power. Contrary to what is often believed, however, this was not a simple application of the old principles of British social protection. The United Kingdom, in adopting its own policies for activating social protection, was not a pioneer, since it came after the United States, France and the Scandinavian countries. Under the influence of a sociologist (A. Giddens) and an economist (R. Layard), the Labour Party made the decisive turning point while the party was preparing for its accession to power. Labour was guided in this by both the US Earned Income Tax Credit (EITC) policies and the reform of requirements for recipients of Aid to Families of Dependent Children (AFDC). Critics spoke of the “Wisconsinization” of British social protection, because the pioneering experiments came from the state of Wisconsin. The influence of the EITC was more discreet. The British, moreover, did not readily recognize the influence of workfare, particularly because these programs had a bad reputation. British unions did not want the American example, and, when they supported the first Labour reforms, they were careful to explain that what was being done in Britain was not workfare but welfare-to-work. The Labour Left resumed its criticism of workfare only much later, under the Cameron government, when Minister Iain Duncan Smith introduced his radical reforms. In the early 1990s, the Conservatives, led by John Major, embraced the idea of tightening the already restrictive conditions for access to unemployment benefits. However, they would not discuss employment assistance initiatives, fearing that the state would become an “employer of last resort”. They did not really pursue an activation strategy. Labour’s activation logic was therefore a real innovation. Politics took the form of the New Deal, one of the pillars of contemporary activation. From 1998, it aimed its first and main programme at young people aged 18 to 24 who had been unemployed for more than six months. Subsequently, they gradually extended these programmes to other categories and were part of a general logic of extending tax credits, social contributions having been systematically reduced for all employees. The New Deal aimed at rapid professional integration into the labour market from the outset. In quantitative terms, options other than immediate placement on the market remained marginal. This was particularly true for the use of employment subsidies initially reserved for the private sector, which were then extended to the public sector. The state has remained constant in its rejection of an employer of last resort function. Intermediate labour markets (ILM) and supported employment programmes have always remained marginal. However, the New Labour activation project was not limited to the implementation of this strategy of transition from assistance to ordinary employment, through the implementation of programmes to boost job search and the provision of improved services by the public employment service. It used a highly punitive logic, marked by a very rigorous provisional sanctions program, which would become even more rigorous during the years 1997–2017. However, beyond the simple adoption of an English-style workfare, Labour undertook an overall reform of social protection starting from the core of assistance, making specific demands for the quality of the public employment service and an installation of relatively high minimum wages while simultaneously generalising the tax credits supposed to be incentives.

15 Germany is undoubtedly the last country to have reformed the activation of social protection. In its Agenda 2010, the Schröder government wanted a radical reform of social protection. On employment issues, a timely scandal was discovered concerning the false statistics of the former Bundesanstalt für Arbeit (Public Employment Service, now Bundesagentur). This triggered a chain of reforms. Germany had already adopted an important law, called JobAqtiv Gesetz, at the end of 2001, but—as in France—was rather reluctant to accept a logic of activation, which ran counter to the logic of unemployment insurance. Numerous reforms followed, but the most significant, painfully decided and applied on 1 January 2005, marked a profound turning point in German social protection based on the insurance principle. By that time, and since the law was adopted in 1961, an overwhelming majority of welfare recipients should in theory have already been “activated”. However, this had not yet happened. They were then classified as employable (persons of working age and employable for at least three hours a day) and had to look for work, while being eligible for a new benefit called Arbeitslosensgeld II (Alg II). The same category also includes young people entering the labour market and former unemployed persons who had exhausted their entitlement to insurance benefits, the majority of whom are insured persons over 45 or 50 years of age. This is a “paradigm shift” (Knuth, 2009), because previously, insured persons who had exhausted their rights continued to receive an assistance benefit proportional to their previous income (Arbeitslosenhilfe). At the same time, since the mid-1990s, a logic of flexibility in employment contracts has been coupled with reform. This has considerably increased the role of jobs previously called “marginal” (geringfügige Arbeit) and later called “minijobs”. In theory, all job-ready people are “activated” with the reform. This includes jobs which in derision public opinion has called “one euro” jobs in the “second market”. Germany had always refused such jobs, contrary to the French assisted contracts introduced in the 1980s. At the same time, many measures have been adopted reforming social security contributions. As in the French system, this makes the German system less Bismarckian, and leads it to playing jointly on the incentive to work and on the demand for work by companies.

The activation of social protection: essential features of restructuring

16 The preceding review of the stages in the reform of social protection activation illustrates that the process, which varied from country to country, took place over a long period from the late 1980s to the early 2000s. Since then, these strategies for activating social protection have become commonplace and part of assistance policies in all the rich countries and, by contagion, in all the Member States of the Union. These reforms were often limited in the latter countries, by the simple fact that the elements of social protection that could be “activated” were very limited (see Table 3). However, the OECD and the services of the European Union have disseminated this political watchword in English. The first task of the various countries therefore has consisted of complying with this orientation in the action plans that they regularly draw up. International organizations have been spreading the political word consistently since the second half of the 1990s. In its 1995 issue of Employment Outlook, the OECD praised the method invented by Denmark, which introduced an individual handlingsplan (individual action plan) for recipients of assistance and unemployment benefits. At the same time, the OECD also promoted tax and social reforms that increased the pecuniary interest in work, under the slogan “Make work pay”. Political and administrative leaders in all countries have started talking about “making work pay”. However, this has not yet led to policies actually converging. The declarations of international organizations have undoubtedly contributed to the discursive and symbolic existence of these policies (Barbier, 2006, p. 13-14). The 2008 crisis barely interrupted the use of the slogan of applying active strategies to people on welfare and the unemployed, as well as reforming social systems to make them more employment-friendly. The political orientation in favour of “activation” is therefore still present in 2017, now constituting a constant element in assistance policies (Frazer and Marlier, 2016). A general feature characterizes them: assistance policies have been profoundly reshaped everywhere by the fact that they no longer make a distinction between poor people, who were the object of special treatment as poor, and the unemployed, who traditionally came under employment policies. From now on, all these people were potential workers, people likely to be inserted into the labour market. “People of working age” is their new label. Even more, the best assistance policy everywhere is supposed to be the one that provides work and employment. [14]

17 The first to publish an international comparative analysis of the reform of assistance to the poor and unemployed in terms of activation were I. Lødemel, a Norwegian, and his British colleague H. Trickey. Influenced by their work in the United States, they took the term “workfare” in American political jargon and proposed a minimal definition that it was programmes and schemes that require people to work in return for social assistance (2000, p. 6). Not only was the term “workfare” replaced by the notion of activation as a result of this pioneering research, but, above all, its meaning was extended beyond the most apparent empirical effects, namely the obligations of the unemployed and those who received assistance. In fact, activation exists in a reform as soon as an explicit link (regulatory or legal) is introduced or reinforced—even reactivated when it already exists at a more informal level—between social protection and professional activity. Whether the systems are Beveridgian or a fortiori Bismarckian, G. Esping-Andersen has shown that this link is historically constitutive of their structure, whatever their degree of “de-commodification”. [15] In this sense, an earlier form of activation has existed since the beginning of social protection: whether it is the obligations imposed on the poor in the British workhouse system, or the obligations to seek employment in unemployed people’s mutual organisations in Germany, etc. (Barbier, 2004). What is new in contemporary activation is that such a link—or reinforcing an existing link—gives rise to a critical redefinition of social programmes, under the dual justification of efficiency and equity. This is done in the sense of a systematic preference given to the engagement of beneficiaries in activity in the labour market. This increasingly takes the form of a condition of activity that is more or less restrictive in terms of “rights and obligations” and that reinforces the conditions for eligibility for benefits. [16]

18 In other words, there is both “activation of persons” and “activation of systems”. In this respect, the reform of social protection systems has taken some time to recognise the importance of introducing tax credits (the various British Working Tax Credits, inspired by the example of American Earned Income Tax Credits—EITC—used both at federal and state level). In France, a functional equivalent is the reduction in social contributions, the amounts of which are considerable, to the point where one could speak of a “hidden welfare state” (Zemmour, 2014). However, the majority of researchers have focused on what they call “activation of the unemployed”, “activation of the poor”, or often just “activation”. In reality, the segments of social protection still undergoing this transformation in 2017 have not been limited to social assistance. These principally include unemployment compensation, (active) employment policies, and assistance and solidarity benefits. The field of retirement and early retirement is now fully concerned. Policies that can be considered “functional equivalents” of French family policies are also “activated” in several countries—for example, in the United Kingdom using tax credits. However, sick leave benefits have not been concerned because they are increasingly becoming a universal form of coverage, although at the same time the reforms of daily sickness benefits may also be interpreted as part of a general logic of activation, such as is the case in Sweden. The reform of the financing of social protection and the tax-contribution linkage are also part of the same systemic dynamic. The official objective is encouraging professional activity. This may be at the global level of labour demand—for example, the reduction in employers’ social contributions in France, the introduction and extension of the generalised social contribution (CSG), and the taxation of social security. [17] Alternatively, it may be at the level of individual incentives such as British or American-style tax credits or so-called “profit-sharing” forms associated with unemployment benefits in Germany, France, etc. These strategies should not be confused with the “flexibilisation” of legal systems, particularly of labour law, which continue to have their own logic (Barbier and Fargion, 2004), even if they are also coordinated.

19 The notion of activation is more neutral as compared to the notion of workfare. It limits the excesses of an analysis that, in one of its caricatured normative versions, sees the transformation everywhere of the so-called “welfare state” into a “penal” system. [18] It allows the analysis of empirical programs as they actually work. [19] The advantage is that with a broad definition of activation, there is an appropriately extended viewing angle covering the overall system. Indeed, it is the reform of social protection as a whole, and not simply the assistance sector alone, that is at stake. Activation, in this sense, is best understood as a form of what Pierson (2001) called the “restructuring” of “welfare states”. It is inextricably linked to the activation of people, services and systems. Like restructuring, it is likely to diversify in several dimensions, even in several forms, depending on the social protection family in question (see Barbier, 2002; 2004). In the end, apart from the particular cases of the United States and France mentioned above, the aid reforms took place initially in Beveridgian systems (Scandinavian systems classified as “social-democratic” and generous and British system based on residual tendencies). Table 4 outlines the political problem to which activation reform is supposed to respond, as well as the collective norms in force in this type. Identified in the early 2000s, these two types have persisted with the evolution of policies in different countries, while a type more in line with the Bismarckian ideal has never really found its place (Barbier and Knuth, 2010).

20 We should add a clarification to these main features of the welfare reforms of the last twenty years. While the representation of policy beneficiaries—the poor and the unemployed—has tended to merge them into a single category of “unemployed”, the distinction has persisted in public policies between activation policies and active employment policies (inherited from the Swedish tradition), as has often been pointed out by OECD leaders (Martin, 2014; see also Weishaupt, 2011). This distinction is important at a time when some analyses have tended to confuse “activation strategies” and “active employment policies”. Unlike Bonoli (2013), Martin notes that the OECD has changed its vision. At first, he said, it was to promote active policies rather than “passive policies”. However, for him this was too “naïve” because it neglected the fact that a country could spend little on “active” labour market policies while having a low unemployment rate. “Economic theory has indeed highlighted the important interactions that exist between the level of unemployment compensation systems, the variety of programmes, and the conditions for seeking employment in compensation programmes” (ibid., pp. 10-15). In the two typical cases in Table 4, the question of the cost of assistance mechanisms remains an essential motivation for the implementation of the reform, whether in Great Britain or in countries such as Denmark. While the Scandinavian case is relatively balanced between rights and duties, implementation is rigorously more demanding vis-à-vis beneficiaries in the British case. In both cases, however, the restructuring of systems included a requirement to contain the increasing costs of this part of social protection. The demand was at the heart of American reform, as the “bipartisan” electorate of the Clinton reform in 1996 was mobilised against welfare recipients represented as parasites. New Labour in Britain, following the Conservatives government, made a similar presentation. On the other hand, in the Scandinavian countries, this argument was never used at the beginning of the reform, which was supposed to correspond to the egalitarian and universal spirit of the Scandinavian social protection systems. [20]