Demanding Activation

Abstract

The distinctive character of the unemployment policies overhaul, known as the ‘activation turn’ of the 1990s, was that OECD governments intensified coercion against unemployed individuals while relaxing their commitment to ensuring full employment. By now supply-side activation policies have been tested long enough to permit conclusive evaluations, and the article surveys the body of empirical assessments to determine the achievements of the activation turn. The implementation of welfare-to-work policies has not resulted in bringing down the rates of unemployment (independently of the business cycle), combating long-term unemployment, reducing (in-work) poverty or empowering jobseekers as consumers of public services, which were all goals of the reformed ‘activating state’. Instead, activation has been working as a mechanism of entrapment at the margins of liberalised labour markets by dint of its complementarity with employer-centred flexibility. The activation turn has thus failed to achieve its direct labour market and social objectives. However, it has produced pronounced indirect negative effects in the labour market, and its social impact has been regressive and repressive, as anticipated in normative and political-economic critiques of the rising ‘workfare state’.

Introduction

The explicit intention of the activation turn in labour market policy was to restore the proper balance between rights and duties. There had been a discontent, shared both on the right and left of centre, that certain sections within the poorer population had apparently become accustomed to relying too readily on public support without being prepared to give back to society or even recognising such an obligation. Even though requirements to look for work had always been attached to income support, policy makers of the 1990s went out of their way to bring the duty to work to prominence in public discourse and the enforcement of such requirements was made pivotal to labour market and social policy. We now have more than two decades of practical experience with welfare-to-work (or activation) policy and programmes in OECD countries. The article reviews empirical evaluations of their effectiveness by economists and questions the results from a broader perspective that directs attention to the implications of such policies for liberal-democratic citizenship. More concretely, the goals of the activation turn will be juxtaposed with its achievements to see which evaluations – ‘internal’ or ‘external’ by critics – can offer greater purchase in capturing the nature of the activation turn.

Active labour market policies include both punitive and enabling mechanisms. To clarify at the outset, the premises and benefits of enabling activation (i.e. vocational skill formation and other supports, such as childcare provision) are not disputed here. Punitive activation comprises (the threat of) sanctions for non-fulfilment of (often punishingly) extensive job-search requirements or refusing job offers, ¹ mandatory unpaid work and the psycho-compulsion ² involved in mandatory skills building and motivational training. Those who advocate and implement such strategies (‘activators’ henceforth) prefer the euphemistic term ‘demanding’ to ‘punitive’. How well has activation been working and what makes it still worthwhile in the eyes of activators? These are the initial questions to be addressed in the article. It will be argued then that reasons given in support of continued punitive activation are not good enough.

How well activation works

Active labour market programmes (ALMPs) ³ come in four main categories: (1) services and sanctions to increase job search efficiency (counselling, monitoring, job search assistance and sanctions for non-compliance); (2) training: classroom, on-the-job, and work experience. Training can provide more general education (language, basic computer and other basic skills), in which case it is called remedial, or specific vocational skills; (3) private sector incentive schemes (e.g. wage subsidies to private firms and start-up grants); (4) direct employment programmes in the public sector (usually work-for-benefits schemes). ⁴

Evaluations of ALMP effectiveness have mostly tended to focus on estimating short-term effects of the programmes (Kluve, 2010: 905). Thus in the short run, comparative evaluations of the main ALMP types find that job search assistance (with sanctions for non-compliance) programmes (Category 1) yield relatively favourable results, reducing the time in registered unemployment (Borland, 2014: 11; Caliendo and Schmidl, 2016; Card et al., 2015). Such programmes are the least expensive and their use has been prevalent in the Anglo-countries where ALMP expenditure has been much lower than in northern and central Europe.

Classroom and on-the-job training (Category 2) which has been more common in Europe shows positive effects after two or three years (Card et al., 2015). However, the overall trend in OECD countries has been to align with the liberal regimes of the English-speaking countries in giving priority to a quick return to employment (work-first) and cutting public expenditure. Therefore, at least until recently, preferences in European countries have also been shifting away from training, which only generates beneficial effects in the medium and long run, to subsidising low-wage jobs (Konle-Seidl and Eichhorst, 2008: 439).

Subsidised private sector placements (Category 3) come second best in short-term effectiveness evaluations (Kluve, 2010). ⁵ However, if one changes perspective from looking at the impact of wage subsidies at the level of unemployed individuals to the aggregated perspective of the economy – to potential indirect and general equilibrium effects of wage subsidies – the results appear to be far less encouraging. Such programmes have been found to produce very low (if any) net gains in employment due to massive deadweight, substitution, and displacement effects (Konle-Seidl and Eichhorst, 2008: 422; Martin and Grubb, 2001: 31). Lastly, studies show uniformly that work-for-benefits (workfare) (Category 4) is the least effective measure in bringing people closer to regular employment (Caliendo and Schmidl, 2016; Card et al., 2015).

To sum up the comparative evaluations of the four main categories of ALMPs, job search assistance and monitoring are cheap and fairly effective in pushing unemployed individuals off benefits. Subsidised employment in the private sector helps many individuals get into paying jobs, but does not seem to lead to the creation of additional jobs in the economy. Programmes that offer substantive vocational training, especially in the workplace, show good results in the long run, but are expensive. Finally, subsidised public sector placements almost never become pathways to regular paying jobs.

Let us now turn to aggregate assessments of ALMP effectiveness. An econometric meta-review, which claimed to ‘represent the most sound and reliable knowledge in the area to date’, revealed that they had ‘a minor positive effect’ in moving unemployed individuals on public assistance into work, increasing earnings, and lowering welfare payments in the first six years after treatment. Using comparisons between participants and control groups, the authors established that there was only a slight increase from 58 per cent to 61 per cent in the likelihood that the participants will have found work two years after participation (Smedslund et al., 2006: 424). In relation to the cost-efficiency of activation programmes, the authors estimated that it takes 33 people in activation to put one person into a job and 27 people to get one welfare recipient off benefits (Smedslund et al., 2006: 31–32).

In the first systematic review analysing the magnitude (and not merely statistical significance) of the effect of ALMP participation on unemployment insurance (UI) recipients ⁶ , the authors found ‘an overall small effect of ALMP participation on job-finding rates, and no evidence of differential effects for different programmes’ (Filges et al., 2015: 8). There was a 52 per cent chance that a treated unemployed person (i.e. one who has participated in a programme) would find a job before a non-treated person. For every 15 unemployed participants an additional person would be holding a job approximately one year after participation (Filges et al., 2015: 61). The probability of finding a job after participation increased by 9 per cent, and by 7 per cent one year after participation (Filges et al., 2015: 54).

Available research thus indicates that the aggregate effectiveness of ALMPs is only modest. The probability of finding employment is only slightly higher for participants of these programmes as compared to non-participant jobseekers. And no evidence has been found ⁷ that ALMPs are becoming more effective over time (Card et al., 2010: F463-464).

The lack of achievement in activation can also be seen from the fact that there has been no reduction in the total share of benefit claimants in the working age population (Konle-Seidl and Eichhorst, 2008: 424–425). The ‘often neglected phenomenon related to activation policy’ is that, in many European countries, falls in active benefit recipiency have been matched by rises in passive benefit recipiency (Konle-Seidl and Eichhorst, 2008: 424–425). ⁸ Countries with low unemployment rates and successful activation regimes, such as Norway, Switzerland and the UK, show high rates of long-term sickness or disability benefits recipiency (with large numbers suffering from mental ill-health). Switzerland spent 2.6 per cent of GDP in 2008 on such benefits – five times what it spent on unemployment benefits; almost one-fifth of the working age population in Norway (as of 2007) were in this category and about 5 per cent of GDP went on funding such benefits but only 0.4 per cent on unemployment benefits; almost 7 per cent of the working age population were in receipt of such benefits in the UK at the end of 2012 (Martin, 2015: 15–17).

The 1979–1997 Conservative government in the UK was accused of concealing unemployment by moving claimants from unemployment to disability benefits (McKnight, 2015: 8). It has since been recommended by welfare-to-work policy experts that these ‘escape routes’ (early retirement being the second one) from activation be closed off (e.g. Konle-Seidl and Eichhorst, 2008). The UK government in particular has pursued this objective. However, as a recent overview of the results of such efforts shows, the UK government has overestimated the share of Incapacity Benefit Claimants that would be found capable of work and underestimated the barriers they face in the labour market (McKnight, 2015: 5, 60). Nor has activation been successful in other countries in relation to long-term sickness/disability benefit recipients, especially those with mental ill-health whom employers are ‘extremely reluctant’ to engage with (Martin, 2015: 22).

Both in terms of administration and programme costs, ALMPs are expensive and the welfare state reforms have neither relieved the pressure on public finances nor brought down overall benefit dependency (Konle-Seidl and Eichhorst, 2008: 426–427). While ALMPs are expensive, they are not expensive enough to shift the hard-to-place into employment. Thus, the outcomes of the latest UK Work Programme, while generally disappointing, have been even worse for the most disadvantaged participants (McKnight, 2015: 5). The UK government has again underestimated the level of incentives that would be high enough to persuade private providers to invest in those with the greatest need for assistance. The levels of expected outcomes and therefore rates of return for disadvantaged participants are very low, and it is unlikely that anything can be achieved in moving disadvantaged individuals into work without substantial investments by the government (McKnight, 2015: 36, 40–41, 60). Yet, given the quest to cut back on social spending at the heart of welfare state reforms, this is not likely to happen. OECD countries had been cutting public spending (including the financing of ALMPs) to achieve fiscal consolidation and, even where budgets were restored to surpluses, there has been no resumption of social spending. Most governments used the surpluses to cut taxes and kept the fiscal capacity of the state restricted (Haffert and Mehrtens, 2015). Furthermore, in the context of quasi-marketised employment services, governments’ cost-cutting imperative also converges with private providers’ profit-making imperative to over-determine that disadvantaged jobseekers continue to be poorly serviced – despite all the efforts to incentivise providers to focus on such clients (Carter and Whitworth, 2015: Ch. 12; Davidson, 2014; Meager et al., 2014).

What makes activation still worthwhile in the eyes of activators?

While acknowledging both that the positive impact of ALMP is limited and that in certain respects the impact can be adverse, many of the policy analysts reviewed above continue nonetheless to support their implementation and approve in particular of strict activity requirements and the use of sanctions to enforce compliance. The purpose of this article is to question this stubborn commitment by OECD governments to expanding effective labour supply through ‘demanding’ activation while neglecting the demand-side causes of unemployment.

The main limitation of ALMPs as mentioned above is their low positive impact even in the case of the better performing programmes (Martin and Grubb, 2001), with the net macroeconomic impact (Konle-Seidl and Eichhorst, 2008: Section 1.4) and cost-effectiveness (Card et al., 2010: F476; Filges et al., 2015: 62–63) remaining undetermined. But what is beyond doubt is that supply-side intervention cannot by itself have a significant impact on the rate of unemployment. Rates of exit to jobs reflect business cycles more than anything else. ⁹ That is why critics of ALMP have dismissed it as a ‘fair weather instrument’. John P. Martin (2015: 21) however defends activation against the criticism by claiming that countries which successfully implemented the policies had experienced lower unemployment rates during the Great Recession than could have been predicted on the basis of previous recessions.

Exit to precarious employment

Yet if activation can indeed moderate rises in cyclical unemployment as well as hysteresis, it would be most likely at the expense of driving many benefit recipients into poor-quality employment. It has been widely recognised as another problem of activation that unemployed individuals are driven into low-skill, low-pay jobs, in which they continue to stay partly dependent (through in-work benefits). Such jobs are also unstable, making repeated returns to unemployment highly probable (Arni and Schiprowski, 2015; Konle-Seidl and Eichhorst, 2008: 427; Martin and Grubb, 2001: 45–46).

In combination with new modes of business organisation as discussed below, work enforcement can reduce the quality of jobs without increasing the quantity, as the latest UK labour market recovery illustrates. It was largely driven by self-employment, which expanded to 15 per cent of the workforce – the highest level in 40 years. At the same time, real average earnings of the self-employed sank by 22 per cent (McKnight, 2015: 5, 52). Many of those who had been made redundant by the cuts in public sector employment re-emerged as self-employed while still ‘selling their services back to the Government or working for private and third sector providers under contract to the Government’ (McKnight, 2015: 39). The fact that the unemployment rate has fallen considerably on account of growing numbers of self-employed, who generally work fewer hours and also earn much less than before the recession, should make the recovery less celebratory (McKnight, 2015: 59). The recovery in the US has been similarly driven by the expansion of low-skill jobs in the private sector while the public sector was retrenched (National Employment Law Project, 2014). It is likely that activation facilitated this shift because it punishes resistance to low-quality employment. Participant surveys confirm that the soul-crushing impact of the activation regime is such that it makes jobseekers willing to accept any job whatsoever just to escape it (Brady, 2014: 180; Meager et al., 2014: 185).

Meta-analysts have complained that available studies preclude a direct assessment of such outcomes as earnings, employment and hours of work (Card et al., 2010: F476; Filges et al., 2015: 55–56). One of the first econometric attempts to fill in this lacuna finds that warnings of sanctions do reduce post-unemployment earnings if not the stability of subsequent employment. However, the actual application of sanctions (i.e. benefit withdrawal) has a negative and persistent impact on both earnings and stability. The study concludes that the net effect of benefit sanctions is negative in macroeconomic terms – because equilibrium unemployment is not affected (i.e. reduced unemployment duration and reduced employment duration cancel each other out) and in microeconomic terms – because individuals are forced into unstable jobs and face a reduction in their life-time income (Arni et al., 2013). Another study yields similar findings about persistent sanction effects on subsequent job quality (Van den Berg and Vikström, 2009: 45).

It does indeed have an adverse impact on the workforce but is it coherent to regard this outcome as an ALMP implementation problem, addressing which may improve the system? Not if one takes a broader view that permits one to see how ALMP functions in relation to the contemporary modes of business operation. These modes were made possible by economic liberalisation. Businesses were allowed to pursue external quantitative flexibilisation, which refers to adjusting the number of workers to business volume fluctuations. The relaxation of labour protection laws made it easier to dismiss staff and to use a flexible workforce – part-timers, casuals and contractors. Since the purpose of labour laws was to protect workers against market insecurity and risks, their weakening meant that it was now possible for businesses to shift risks back onto workers. Flexible modes of business organisation depend essentially on precarious forms of employment, on evading the traditional employer-employee relationship. It used to be illegal to use temporary workers to carry out the regular business activities of a firm. The new mode that has now become the normal and prevalent way of operating a business – the outsourcing of core and peripheral activities – permits doing away with this ‘rigidity’. The transfer of business risk away from the parent or client company occurs in two stages: the company contracts out its activity to business services providers, and providers then hire flexible workers. While regular business activity has not changed its nature, the addition of an intermediary layer in business organisation permits the presentation of this activity as specific, discrete tasks. The traditional employment relationship that encompassed certain protected rights is now fragmented and reduced to discrete assignments set out in the commercial contracts between the client and the provider, and then reduced by providers into work tasks for workers. The provider has no obligations traditionally borne by employers because it itself is an entity that exists only temporarily with no organisational identity outside of the annually reviewed contract with the client. This transformation effectively amounts to the substitution of labour law with commercial law (Frade and Darmon, 2005).

Under New Public Management, unemployment benefit administration itself was reorganised on the same model. The government acts as a parent/client company contracting out the delivery of services to a network of providers within a quasi-market (often with ‘prime’ providers subcontracting to smaller local providers). These providers supply a subsidised contingent workforce to low-cost employers. Early research into employer engagement with ALMP, which so far has been a neglected area, indicates that it is mostly businesses that heavily rely on a large supply of low-wage, low-skill labour that are interested in engaging with ALMP providers (McGurk, 2014). It saves employers who ‘have a strategic preference for contract workers or who simply seek to fill temporary staffing shortages in low-skill, generic roles’ to utilise publicly-funded and publicly-administered sources of such contingent labour (McGurk, 2014: 12). Providers in their turn cannot hope to survive competition unless they manage to become ‘recruiters of choice’ to such large employers with regular, high volumes of vacancies and to staff these with job-ready candidates from their caseloads (Ingold and Stuart, 2015: 455–457). ‘Prime [welfare-to-work] contractors [in the UK] estimate that such job placements account for at least seventy-five per cent and up to ninety per cent of their activities’ (McGurk, 2014: 26). In any case, employers’ interest in the ALMP provider services has remained constrained by low business volumes in the depressed UK economy, that being cited as a key reason for employer disengagement (Ingold and Stuart, 2015). In theory, supply-side intervention is supposed to work because ‘when vacancies attract many high-quality applications, employers create more vacancies’ (Martin and Grubb, 2001: 52). The idea that it is possible to expand aggregate demand for labour just by expanding effective labour supply seems to descend from Say's Law (a law stating that supply creates its own demand) which, although discredited a long time ago, was revived by supply-side economics. The failure of activation proves that it should not have been.

The exit of qualified/skilled benefit claimants into low-paying, precarious jobs is uniformly considered a sub-optimal result of activation. However, considering such outcomes sub-optimal does not seem to fit with the fact that unemployment benefit recipients are induced to accept any available work. Under the activation regime, a longer job search that could lead to a better match is discouraged. After a certain period of being on benefits, jobseekers are made to understand that any jobs should be suitable for them and that they are going to lose benefits if they refuse (usually) a third job offer made by their caseworkers. For skilled/qualified jobseekers that are unable to find appropriate employment quickly enough this implies downward occupational mobility. When sub-optimal microeconomic outcomes are discussed in ALMP evaluations, these are generally losses in earnings and job security upon re-employment. Only one of the analyses that were identified and perused for this article considers downward occupational mobility as another plausible aspect of punitive activation. Analysing to what extent jobs that were accepted following the imposition of a sanction (for refusing a job offer) differed from jobseekers’ previous jobs, van den Berg and Vikström (2009) found that sanctions often led to a match in a lower occupational level. They further observe that occupational downgrading tends to be irreversible. ¹⁰ The authors note that the incidence of sanctions in the Swedish system, which was the context of their study, was the lowest among the nine OECD countries compared (Van den Berg and Vikström, 2009: 10–11). Presumably, the more punitive the activation system the stronger the impact of this kind it creates. ¹¹

The main rationale for escalating administrative coercion against benefit claimants, which continues to be evoked in defence of activation, is that the deterrent impact of compulsory policies improves overall labour market performance by moderating wage demands.

Wage bargaining models predict that supply-side policies will lower wage demands and this effect will be strongest in the lower end of the wage distribution – both a downward pressure on wage setting and an upward move in employment (Konle-Seidl and Eichhorst, 2008: 422).

If the purpose of activation, as made explicit by economists, is to undercut wage demands and increase labour supply ‘in the lower end of the wage distribution’ while employers’ obligations towards labour are minimised at the same time through deregulation, the growing numbers of the precariat should signal its success. Outsourcing is intrinsically contingent because it is based on impermanent contracts. It therefore seems unlikely that the dynamics of ALMP can be reversed without reversing the transformation that has occurred in business organisation as discussed above. Contemporary employers may not be able to provide enough secure jobs with career ladders to absorb even mainstream jobseekers, let alone the ‘unemployables’ – even if austerity governments were ready to invest extravagantly in developing the human capital of the disadvantaged unemployed.

As pointed out earlier, ALMP has been evolving away from investment into human capital development. The activation paradigm in social policy is part and parcel of the austerity regime. There seems to be little point – within the current framework – in advocating for more investment in the vulnerable groups, as Konle-Seidl and Eichhorst do (2008: 441), to enable their integration into sustainable employment and thus also to reduce the need for subsidies/in-work benefits. The whole policy tendency over the last decades has been in the opposite direction, ¹² because the activation turn was inspired by different concerns – by concerns that low-wage jobs are not taken by such groups (Mead, 1986). The Third Way manifesto of Blair and Schröder emphasised the importance of a subsidised low-wage sector. In fact, the relevant section – ‘An active labour market policy for the left’ (Blair and Schröder, 1999: Sec. VI) – only refers to creating this one kind of ‘opportunity’. The active promotion by many European governments in recent years of subsidised domestic services (Morel, 2015) clearly follows in this path.

Exit to non-employment

Exit to non-employment, on the other hand, can indeed be validly viewed as an implementation problem for ALMP. Where jobseekers respond to pressure by withdrawing from the workforce rather than intensifying their job search (Caliendo and Schmidl, 2016: 10–12; Manning, 2009), activation is counterproductive. ¹³ An Australian longitudinal study of the period between 2008 and 2010 shows that around 37 per cent of claimants exited unemployment by leaving the workforce (41.8 per cent of women versus 31.2 per cent of men), 23.3 per cent exited into full-time employment, 23.6 per cent into part-time employment and 16.5 per cent remained unemployed (Rotaru, 2014: 116–117). Arni et al. found that 12.5 per cent of the sample exited to non-employment, with ‘a remarkable rise’ in the rate of exit to non-employment – 99 per cent in response to a sanction announcement. They also found that an enforced sanction ‘results in an additional increment’ of 67 per cent to the exit to non-employment rate (Arni et al., 2013: 1165). In some cases, people become unregistered as unemployed while continuing to look for work (Arni et al., 2013: 1166). This should be mostly a welcome situation from the current official point of view: even though such unemployed people make themselves unavailable for inferior jobs, they are still participating in job search and at no cost to the ‘taxpayer’. This option however is ‘truly costly’ to the individual in monetary terms (Arni et al., 2013: 1169).

If the unemployed withdraw from the workforce, it is a mixed outcome: on the upside, the cessation of benefit payments means savings, but on the downside such withdrawals deplete the effective labour supply available to employers. When claimants disappear from welfare rolls but do not reappear as jobholders, whichever ‘escape route’ they use – private dependence and severely reduced consumption or informal work (including crime) – they effectively manage to decommodify themselves, which runs against the ALMP ambition to expand labour market participation.

Another redeeming impact that has been suggested for activation is that it spreads the burden of unemployment among more people. Despite the fact that, due to displacement and substitution, the existence of net macroeconomic gains from subsidised employment is in doubt, spending on it, the argument goes, may still be justified by its redistributive effect, i.e. ‘changing the identity of the unemployed to achieve a more equal distribution in the population of time spent in unemployment’ (Borland, 2014: 11). However, if one recognises the legitimacy of redistributive goals, why settle for such meagre positives? Why shouldn't policy seek to redistribute the burden of employment rather than the burden of unemployment? If the high cost of ALMPs can be accepted on the grounds that they help to share the burden of unemployment more widely, why not accept the costs of eliminating unemployment altogether? However, neoliberal ALMP does not seek to eliminate unemployment. From the point of view of neoclassical economics, unemployment is functional. Activators recognise as much: ‘unemployment reduces wage pressures, making business more profitable in an open economy or allowing noninflationary expansion of aggregate demand in a closed economy’ (Martin and Grubb, 2001: 52). It seems inappropriate therefore to judge the success of activation by how much it brings down unemployment rates. But it is valued for supposedly contributing to a lower natural rate of unemployment, i.e. the rate of unemployment that keeps inflation down. However, the natural rate in EU15, i.e. the core of the EU, has remained roughly stable at a high level since the early 1980s. As inflation rates had long stabilised and remained low, the EU15 actual unemployment rate now almost coincides with the natural rate (Blanchard et al., 2006: 9). In theory, strongly activating policies in the context of liberalised market institutions were supposed to speed up labour market recovery after shocks, shorten cyclical downturns, and counteract hysteresis. Yet the scale of the Great Recession and stagnation that followed hardly indicate that the economies have become more resilient as they were expected to under the impact of the labour market and welfare reforms. The missing results raise the question of how sound the premises of the activation turn were. The subject cannot be pursued here, however, due to space constraints.

Psychological activation

A survey published almost a decade ago noted that comprehensive evaluations of ALMP were hindered by the paucity of studies attempting to measure the impact of activation on ‘quality of life, social skills and self-confidence of the participants, and whether the programmes would have any adverse effects on the participants’ (Smedslund et al., 2006: 31). Nonetheless, literature of this kind seems to be emerging. Evidence from various studies shows that participation in ALMPs improves the subjective wellbeing of the unemployed even though this improvement may not be long-lasting (Sage, 2015). There is also uncertainty (at least in the British case, analysed by Sage, because the UK national surveys only furnish aggregate data) as to the impact of different types of ALMPs on wellbeing. Different groups of participants may also be affected differently. Findings on health and social benefits from ALMPs’ participation in the international literature are much more mixed, and Sage's own study finds none (Sage, 2015).

Since the focus in this article is on the programmes that involve coercion, we shall turn to studies that do differentiate between discrete types of programmes in measuring the wellbeing/life satisfaction impact on their participants. One such study finds enforced participation in job-search training, among individuals with poor labour market chances and at risk of becoming long-term unemployed, to be counterproductive: it actually decreased re-employment and had an adverse mental health impact (Malmberg-Heimonen and Vuori, 2005). Based on this, the authors advise against forcing unemployed individuals, especially the more disadvantaged ones, ‘to search for jobs that do not exist or jobs that they are not qualified for’ (Malmberg-Heimonen and Vuori, 2005: 465), because the only result of this is likely to be their compromised wellbeing. The study of participants in the German version of workfare (One-Euro-Jobs) finds that those who believed that activities matched their skills, and perceived them as increasing their chances of reemployment, felt better than non-participants. Otherwise, participation was experienced as degrading and thus life-satisfaction diminishing (Brady, 2014: 171–175; Gundert and Hohendanner, 2015; Wulfgramm, 2011). Qualitative studies of activated single parents in five English-speaking countries have registered an overwhelmingly negative impact of conditionality on mental health and wellbeing (Brady and Cook, 2015; Campbell et al., 2016). The majority of participants in the UK Work Programme believed that the regime of sanctions was unnecessary as it neither influenced their search behaviour nor increased the likelihood of getting employment (and indeed no correlation was found between the application of sanctions and exit to jobs among this group) (Meager et al., 2014: 157). ¹⁴ The awareness of sanctions generated anxiety and resentment and, when sanctions were applied, also hardship and disruption to their job-search activities (Griggs and Evans, 2010: 33–34; Meager et al., 2014: Ch. 10). These effects are not compatible with subjective wellbeing.

However, activators are either unaware of the limits to psychological resilience or unconcerned about the potential damaging psychological effects of intensifying harassment and intimidation associated with activation in job-short economies. This is not to say that there is no place for psychology in ALMP. On the contrary, intrusive psychological intervention to modify attitudes, beliefs and the personality of the jobseekers in order to create a ‘disposition of employability’ in them, occupies a central place in activation. In addition to the direct punishment by sanctions for failure to meet requirements, longer-term jobseekers are indirectly punished by mandatory work-for-benefits and compulsory ‘support’ in the form of motivational workshops, referral to psychological help or psychometric testing (Friedli and Stearn, 2015).

The benefit recipient ‘has two distinct tasks: one to find a job and the other to find a job on the terms and conditions set forth by’ activators (Brady, 2014: 97). These imperatives are not necessarily synergetic. In qualitative studies of participant experience, activation is more often than not described as a frustrating waste of time due to the arbitrary nature of requirements and the perfunctory, tokenistic nature of provided support. Participants feel ‘processed’ and ‘busied’ when they have to show up for frequent job-search monitoring appointments, take irrelevant courses, meet often unrealistic quotas of monthly applications or apply for ill-matched vacancies forwarded to them by caseworkers for no other reason than proving compliance (Brady, 2014: Chs. 6–7; Marston and Mcdonald, 2008; Meager et al., 2014: Chs. 9 & 11).

Only autonomous motivation is positively correlated with subjective wellbeing, because it fulfils two basic psychological needs: to feel that one's actions are self-determined and that one is competent (i.e. achieves results) (Van den Broeck et al., 2012: 70). Neither of these conditions is met in the context of imposed activities and absent rewards in the form of reemployment. Furthermore, attempts by ‘employability experts’ to enforce ‘positive affect’ in jobseekers (Friedli and Stearn, 2015) through controlled motivation (of the introjected type in this case) ¹⁵ appear to be essentially incoherent since positive affect is an outcome of autonomous motivation and also because the environmental antecedents of motivation are ignored.

Welfare-to-work providers consistently believe (and the terms of their remit probably condition them to) that it is poor motivation that holds their ‘clientele’ back. Yet jobseekers that are perceived to lack motivation point to factors beyond their control, such as repeated failures to find work after many months or even years of job search, inability to secure funds to pay for occupational training or professional licences, and various barriers to work (poor health, age, homelessness, lack of work experience, financial difficulty, etc.) (Meager et al., 2014: 181–182). The work-first format excludes support that jobseekers themselves believe could be really useful, such as occupational guidance, job brokering and access to further education that would lead to formal occupational qualifications (Brady, 2014: Chs. 6–7; Meager et al., 2014: Chs. 9 & 11). Qualified individuals looking for higher skilled, professional and managerial level jobs feel especially strongly that providers have little to offer and that their goals and preferences are disregarded as they are quickly pushed into any job or unpaid work placements (Meager et al., 2014: 170, 187).

Deficits that current ALMP does not address, such as job shortages and discrimination against older individuals or those without work experience or qualifications, were identified by participants as the most significant causes of their unemployment. At the same time, areas where providers do offer services, such as motivation building, were ranked as the least relevant (Meager et al., 2014: 221). If the limited availability of paid employment affects motivation, trying to extirpate possibly legitimate ‘negativity’ in jobseekers (Friedli and Stearn, 2015: 40) rather than address job shortages seems a misconceived as well as oppressive policy.

Controlled motivation (i.e. coercion) is effective in making people do something they otherwise would not and, by the same token, it reduces their subjective wellbeing. However, the fact that ‘demanding’ programmes work this way is by no means a flaw of design or implementation. To make living off benefits unbearable is their direct if not always stated or publicised function. It is known that greater employment protection and generous income replacement rates during unemployment raise the general levels of life satisfaction (Ochsen and Welsch, 2011; Wulfgramm, 2014). Yet both of these conditions (aka social rights) were deliberately undercut through neo-liberal labour market and welfare reforms on the grounds that they had purportedly exaggerated the sense of entitlement and etiolated the sense of social duty with regard to economic participation. The coerciveness of ALMP was simultaneously enhanced.

Conclusions

When faced with warnings and sanctions, jobseekers accept inferior job offers more promptly – unless they are unemployable. ‘Demanding’ activation is proven to have this kind of effectiveness. On the other hand, even activators readily recognise that activation on its own is not a solution to unemployment. Basic and generic services, which are the only profitable ‘product’ for the privatised system of service provision, are largely useless (deadweight loss) in the case of the more employable jobseekers and inadequate in the case of the harder-to-help.

One achievement of the labour market reforms is certain: the wage share in GDP across OECD countries has fallen (and to the greatest extent in countries with stiffer work-first regimes). However, the neoclassical-economic promise that this would lead to business expansion and job creation has hardly materialised (Lewis, 2009). Instead of productive investments, the greater share of accumulated capital had been mostly channelled into portfolio movements and speculation, inflating the bubble that burst in 2008 (Tridico, 2012). Downward adjustments in wages have not translated into price deflation (OECD, 2014) – the theoretical next step towards economic and jobs recovery. Unemployment remains high, long-term unemployment is becoming more intractable, and the ‘long-standing’ issue of poor job quality in the lower tier of the labour market has ‘deepened’ (OECD, 2014). Thus the conclusion (e.g. Greer, 2016; Wiggan, 2015) that the activation turn is a failure on its own programmatic terms, but a success in terms of achieving the ‘immanent’ objectives of the workfare state appears to be warranted.

Within the framework of the social investment strategy, as one review puts it, one has ‘a right and a duty to activation’ (Van Kersbergen and Hemerijck, 2012: 482, added emphasis) – not the right to work and the duty to work. The internationally codified human right to work receives no acknowledgement within the paradigm of the activating state. In line with the political change, the quantitative aspect of the human right to work (i.e. full employment) has been effectively nullified through interpretation (Raffass, 2014). Since employment is not guaranteed as of right and, statistically, there are not enough jobs to be held by every person wanting a job, what is being enforced is not even the duty to work, but merely a duty to activate. That is to supply labour to the formal market (some people do work but informally) by looking for a job, by changing occupation to possibly find a job or at least to pledge labour by demonstrating ‘work-readiness’ through participation in prescribed activities that are not genuinely expected to lead to a paid job. The abandonment of the commitment to full employment leaves a sizeable section of the population on the margins of demand-deficient economies due to various aspects of vulnerability, including scarring from unemployment itself, but also subjected to the stigma-generating compulsion of activation that becomes a recurrent and enduring state for those churned at the bottom of the labour market.