Paper for FISS Conference, Sigtuna, Sweden 2-4 June 2014

1 Paper for FISS Conference, Sigtuna, Sweden 2-4 June 2014 Stream: ‘Social protection and labour market change (including active labour market policie...
8 downloads 0 Views 812KB Size
1 Paper for FISS Conference, Sigtuna, Sweden 2-4 June 2014 Stream: ‘Social protection and labour market change (including active labour market policies)’ Contact: Peter Davidson1 Title: Comparing activation policies in Denmark and the United Kingdom: testing the convergence thesis Abstract This paper reports on in-depth interviews with policy makers and administrators and independent experts to compare activation policies for long term unemployed people in the United Kingdom and Denmark from 1990 to the mid 2000s – countries with different welfare regimes that were regarded as policy ‘leaders’ in this field. It tests the ‘convergence thesis’ which holds that previous national policy differences in this area – including the design of unemployment payments and investment in different active labour market programs, diminish as activation policies mature and are widely adopted2. Activation policies are compared according to their relative emphasis on three goals: income protection, improving employment capacity, and improving employment incentives. Simple statistical comparisons indicate that substantial differences persisted between the two countries in both the generosity of unemployment insurance payments and the overall level of investment in labour market programs. In support of the convergence thesis, evidence from the interviews indicates that from the late 1980s a new ‘activation paradigm’ supplanted the previous emphasis on income protection (improvement and then retrenchment) in both countries, as activity requirements were formalised and intensified for unemployed people and extended to benefit recipients previously outside the paid workforce. However, the form of activation differed in significant ways, consistent with pre-reform institutional arrangements. In the UK, activation policies originated with ‘Restart’ interviews and programs of work experience and training were subsidiary to job search requirements and assistance. For example, in the ‘New Deals’, access to work experience and training programs was limited to those long term unemployed people who passed a ‘Gateway’ of intensive job search assistance. In Denmark, activation was, at least at first, synonymous with participation in work experience and training programs that were originally designed as income protection measures but later adapted as employability programs. There was also consistent variation in the form and purpose of employment assistance offered to insured and uninsured income support recipients in Denmark, while in the UK distinctions 1

PhD student, Social Policy Research Centre, UNSW, Australia [[email protected]]. I have also analysed, and hopefully influenced, activation policies in Australia over the past 20 years in my role as Senior Adviser at the Australian Council of Social Service, the peak body of not-for-profit community services, though of course this paper is not written in that capacity. I would like to thank Professors Dan Finn, Thomas Bredgaard and Flemming Larsen for their insights into activation policies in the UK and Denmark, and help in connecting me with potential interviewees for my research, and my supervisors, Professors Peter Saunders and Peter Whiteford for their insights into welfare states and conducting a PhD thesis. 2 Eichhorst&KonleSeidl 2008, ‘Contingent convergence, a comparative analysis of activation policies’, IZA Discussion Paper 3905, Institute for the Study of Labor, Bonn.

2 between these two categories were virtually eliminated. Evidence from the interviews suggests that enduring differences in activation arrangements between the two countries may have a basis in the institutional arrangements for the payment of unemployment benefits, including the persistence of corporatist arrangements in Denmark and their disappearance in the UK. Further, higher benefit levels rendered an incentives-based approach (‘making work pay’ policies) less feasible in Denmark. Introduction This paper compares activation policies for long term unemployed people in two countries, the United Kingdom and Denmark, which are regarded as archetyal ‘liberal’ and ;social democratic’ welfare regimes. It is derived from research for a PhD thesis comparing these policies in four countries (the others being Australia and the Netherlands) from the late 1980s to mid 2000s. The research focusses on employment assistance programs for long term unemployed people receiving social insurance or social assistance payments. It identifies the extent of convergence or divergence in the character of, and level of investment in, these programs in the four countries as activation policies developed. Following a period of retrenchment of social security payments for unemployed people in the 1980s, the United Kingdom and Denmark were among the pioneers of a new set of ‘activation’ policies to resolve the tension between poverty prevention and fiscal austerity3. After the recession of the early 1990s, unemployment was still above 10% in most OECD countries and long term unemployment was a growing problem. The concern among policy makers was that if long term unemployment became entrenched, the level of structural unemployment would remain high, imposing fiscal and social costs similar to those experienced after the recession of the early 1980s. The central idea of the new ‘activation’ strategies was that welfare costs could be curbed, not by reducing payments, but by keeping people consistently active in the labour market, improving work incentives and replenishing their skills. ‘Work tests’ were not new, but they were now applied more intensively to recipients of unemployment payments, and extended to groups whose labour market participation was previously neglected – at first long term unemployed people and later to income support recipients with disabilities and caring responsibilities. Activation policies typically require employment services to closely monitor job search and work preparation and develop individual employment plans supported by investment in work experience and training for those disadvantaged in the labour market. Reform of employment services was often complemented by a suite of policies to ‘make work pay’ by reducing effective tax rates for those entering employment.4. The spread of activation policies challenged prevailing comparative theories of the welfare state. The dominant theory of this period, Esping Andersen’s ‘three worlds’ of welfare capitalism, posited enduring differences between three regimes of welfare provision: ‘social democratic’ (characterised as offering generous and universal income support, coupled with greater 3

Barbier J & Ludwig Mayerhofer W 2004, the many worlds of activation, European Societies 6:4, pp 423-436; Castles F & Obinger H 2008, Worlds, families regimes - clusters in European and OECD area public policy , West European politics 31:1, pp321-344. 4 Finn & Schulte 2008,Employment first, activating the British welfare state, in Eichhorst W, Kaufmann O, Kohnle Seidl R, Bringing the jobless into work? Berlin. Springer. CarcilloS & Grubb D 2006, From Inactivity to work: the Role of Active labour market policies.OECD Social employment and Migration Working Paper No 36.

3 regulation of the labour market and high labour market participation including among women), ‘liberal’ (with more parsimonious payments targeted towards poverty alleviation, less regulated labour markets, and high labour market participation though less supportive of female workforce participation), and ‘conservative’ (with payments segmented by occupational status, a high degree of labour market regulation, low workforce participation, and a pronounced ‘male breadwinner’ family model)5. These regimes were associated with the Nordic countries, the Anglophile countries, and central and southern Europe respectively. Esping Andersen argued that welfare regimes could be distinguished by comparing the extent of ‘de-commodification’ (reduced reliance on market income), univeralism (scope and consistency of coverage and payment rates across the population) and defamilisation (reduced reliance on income and services exchanged within the family). The ‘three worlds’ thesis was complemented by a strand of political theory, ‘historical institutionalism’. This emphasised the ‘path dependency’ of institutions such as welfare states. That is, once the basic institutional form of welfare systems was set in place, including the level and structure of payments and the stakeholder interests that surrounded them, it was difficult for policy makers to fundamentally alter it. The path dependency thesis was used to explain the resistance of welfare states to retrenchment in the 1980s6. Instead of abolishing or drastically reducing payments, policy makers made a series of incremental adjustments within the system, though over time those adjustments could destabilise the old regime and open the path to structural change. Five mechanisms by which this might occur were identified by Streek and Thelen: ‘displacement’ (where a new program progressively takes over the space occupied by an existing one), ‘layering’ (where a new program at first supplements an existing one, but gradually becomes dominant), ‘drift’ (where a program becomes redundant through policy inaction), ‘conversion’ (where an existing program is adjusted to meet a qualitatively different objective), and ‘exhaustion’ (where an institution collapses under the weight of its own internal tensions)7. Cox and Weishaupt point to the role of ‘paradigm change’ in the dominant set of ideas according to which public policy problems are defined and solutions are framed, in bringing about structural rather than merely incremental change. Activation itself is represented in the literature as a new and increasingly dominant paradigm in labour market policy, replacing previous ones such as poverty alleviation through social security and ‘manpower policy’8. Both the ‘three worlds’ thesis and much of the research informed by historical institutionalism emphasise structural differences among ideal-typical national welfare systems. Yet with the spread of activation policies, the liberal and social democratic welfare regimes appeared to be converging towards a new system with ‘re-commodification’ at its core. While activation was at one level perfectly consistent with the Nordic pre-occupation with active labour market participation, at a deeper level it potentially undermined the foundations of the Nordic unemployment payment structure. This was traditionally dominated by generous, prolonged, and widely available social insurance payments supplemented by a residual and less generous social assistance or ‘safety net’ payments administered by a lower level of Government. Yet 5

Esping Andersen G 1998, Social Foundations of post industrial economies, Oxford, Oxford University Press. Pierson P 2000, ‘Increasing returns, path dependence and the study of politics’ American Political Science Review 94:2, pp251267. 7 Streek W and Thelen K, 2005, Beyond continuity, institutional change in advanced political economies, Oxford. Oxford University Press. 8 Cox 2001, ‘The social construction of an imperative, why welfare reform happened in Denmark and the Netherlands but not in Germany’, World Politics, V53:3, pp 463-498; Weishaupt J 2011, ‘From the manpower revolution to the activation paradigm, Amsterdam. Amsterdam University Press. 6

4 Clasen and Clegg argue that the systematic extension of activation to new groups of social security recipients requires a degree of ‘homogenisation’ of payments and employment services among different beneficiary populations9. As we shall see in the Danish case, activation reforms are associated with shorter durations of unemployment insurance and a weakening of the traditional role of ‘social partners’ (employers and unions) in the administration of social security and employment services, a key feature of social democratic’ welfare regimes. Turning to the liberal welfare regimes such as the United Kingdom, Weishaupt argues that the progressive activation of client populations more distant from the labour market requires those countries to raise their relatively low levels of investment in skills development10. The picture is further clouded by profound changes at the level of welfare administration, where activation policies were twinned with prevailing ‘New Public Management’ ideas which called for the separation of policy and implementation, replacement of direct public control of services with network or market based systems of governance, and the funding of employment services to employment outcomes rather than program inputs11. Significantly, this trend was also associated with a reduced role for the social partners in social security and employment services administration. In some countries such as the Netherlands, structural change in the administration of benefits and labour market assistance was so rapid and sweeping that the original institutions were hardly recognisable.12 Academic debate ensued over whether the new activation policies and institutional changes associated with New Public Management marked a new convergence, at least between the social democratic and liberal regimes that were the first to adopt these policies, or whether they were simply a new set of administrative arrangements ‘layered’ on top of distinct welfare regimes13. To make sense of these changes, and to understand their implications for comparative welfare state research, it is necessary to measure and compare the ‘stance’ of activation policies over a sufficiently long period to separate structural change from the effects of the business and political cycles. Since the earliest adopters of activation policies were liberal or social democratic welfare regimes, this paper compares activation policies in one liberal (United Kingdom) and one social democratic (Denmark) regime. One strand of quantitative comparative welfare research compares the generosity and scope of social security payments. Given the close connection between income support and labour market assistance in activation policies, another common comparative measure is the level of investment in labour market programs. High payments and investment in labour market programs are associated in this literature with social democratic regimes, and low payments and investment with liberal regimes14. 9

Clasen J & Clegg D 2006, ‘Beyond activation, reforming European unemployment protection systems in post industriallabour markets.’ European Societies, 8:4, pp 527-553. 10 Weishaupt 2011, From the Manpower revolution to the activation paradigm, Amsterdam.Amsterdam University Press. 11 Considine M 2001, Enterprising states, Cambridge, Cambridge University Press; Bredgaard T & Larsen F 2001, Redesigning the governance of employment policies. CARMA 25th anniversary conference, Aalborg, University of Aalborg. 12 Sol E 2008, Activation as a socioeconomic and legal concept, laboritorium the Netherlands, in Eichhorst W, Kaufmann O, KohnleSeidl R, Bringing the jobless into work? Berlin. Springer; Finn & Schulte, op cit 13 Lodemel I &Trickey H 2001, An offer you can't refuse: workfare in international perspective, London. Policy Press; Serrano Pascual A 2007, Reshaping welfare states, London, Edward Elgar. 14 Starke P, Obinger H, Castles F, 2009, Convergence towards where? Journal of European Social Policy, 15:7, p975; Rueda D 2006, Social democracy and active labour market policies: insiders, outsiders and the politics of employment promotion, British Journal of Political Science, 36:3, pp 385-406.

5 However, the overall level of investment in labour market assistance does not tell us much. Comparative welfare state researchers have developed more sophisticated measures that distinguish between different types of labour market programs. A common thread running through these comparative frameworks is a dichotomy between ‘human capital development’ which emphasises skills development, and ‘work first’ approaches which prioritise a rapid transition to (mainly low skilled) employment15. This literature usually associates the former with social democratic regimes and the latter with liberal regimes, This dichotomy is consistent with the labour market structures commonly associated with these two welfare regimes. Social democratic regimes are associated with relatively high minimum wages which raise the formal qualifications required for workforce entry while liberal regimes are associated with lower minimum wages and a longer ‘tail’ of low skilled employment to which income support recipients might be referred without a great deal of preparation16. One drawback of this distinction is that it is easy to confuse activation policies in general with one side of this dichotomy – the ‘work first’ approach - although it is also possible to ‘activate’ people through training in vocational and other skills 17. In order to clearly separate activation policies generally from a ‘work first’ approach, and to bring social security payments into the comparative framework alongside employment programs, this paper distinguishes between activation policies which strengthen ‘work capacity’ (whether through preparatory training or participation in work experience programs that directly strengthen employability skills) and those which strengthen ‘work incentives’ (whether through changes to payment rates and income tests, in-work payments, monitoring of job search, or labour market programs such as those work-for-benefit schemes designed to make reliance on income support less attractive)18. This dichotomy encompasses both ‘sides’ of activation policies: employment assistance to improve skills and ‘make work pay’ policies. It facilitates a deeper and more comprehensive comparisons between activation ‘systems’ in different countries and welfare regimes than simple comparisons of benefit replacement rates and overall investment in labour market programs, or the strictness of activity testing. It enables us to test two hypotheses regarding convergence and path dependency in policies to reduce structural unemployment in different welfare regimes: 

That both liberal and social democratic welfare regimes were predisposed towards the adoption and progressive extension of activation policies because they already privileged high levels of labour force participation (so activation was not entirely a ‘new idea’), and it enabled Governments of both the Left and Right (and the social partners where they have a role in policy making) to reduce the cost of social security and poverty at the same time, without retrenching payments. Therefore, policies will converge towards ‘more activation’.

15

Aurich, P. (2009). Activating labour supply, European strategies from 1990-2005. Conference Paper, Activation and security, Brno, ASPEN. 16 17

OECD 2006, Employment Outlook.

Lodemel I &Trickey H 2001,op cit. Castonguay J, and Sol E 2007, Is work working? Social security in the labour market. Warsaw, International Social Security Association; Johansson, H and B. Hvinden B 2007, Reactivating the nordic welfare states. International Journal of Sociology 27:7/8, pp 334-346. Bonoli G 2010, The political economy of active labour market programs, Politics and Society, 38:4, 435457. 18

6 

That social democratic welfare regimes such as Denmark, whose unemployment payments are typically higher relative to minimum wages, are less likely to lean towards incentive-strengthening approaches to activation since there is less scope for them to do so without cutting benefits. On the other hand, if qualification barriers to entry level jobs in these countries are higher (because wages and skills and more equally distributed) they are likely to rely more on capacity building approaches (since the alternative is to cut minimum wages which is likely to be politically resisted). In liberal welfare regimes such as the United Kingdom, the opposite applies: minimum wages and skills barriers to labour force entry are lower but benefits are much lower, creating more space for work incentive strengthening policies to succeed without the need for substantial investment in re-skilling unemployed people19. Therefore, activation policies will follow different tracks in the two welfare regimes (towards the left or right in figure 1 below).

Analytical framework and methodology For the purposes of this paper, ‘activation policies’ are defined as a set of policies that systematically privilege employment as an objective for social security systems and a pathway out of poverty, through activity requirements, financial incentives, and work capacity-building. The analytical framework used to compare activation policies is illustrated in Figure 1.

Figure 1: The activation diamond

Income protection

Work capacity

Work incentives

Employment

In this figure, activation policies are represented by the area bounded by the ‘activation diamond’. They aim to transition recipients of unemployment payments (work-tested social insurance and social assistance payments) from ‘income protection’ (or de-commodification) at the top to ‘employment’ at the bottom of the diamond20. Turning to the horizontal axis, this 19

This dichotomy between ‘Anglophile’ and ‘northern European’ (i.e. ‘Nordic’) approaches to labour market policy is also hypothesized in a comparative analysis of different ideal-typical strategies to reduce structural unemployment undertaken by the OECD. See OECD 2006, op cit. 20

Employment programs as well as income support can have an ‘income protection’ character. Bonoli 2010, (op cit) classifies traditional ‘job creation’ programs as ‘occupational’ since they keep people occupied rather than

7 process can follow two pathways: towards the left or right, according to the degree of emphasis on work capacity building and work incentives, respectively. This research has two inputs. First, a brief statistical analysis of benefits and investment in labour market assistance was conducted, comparing the generosity of social insurance and social assistance payments in the two countries, and also the level of investment in labour market assistance. Data from the OECD are used for these comparisons. The second input is a series of 12 in-depth interviews (six in each country) with equal numbers of Government officials, employment service providers (or peak organisations representing them), and academics with expertise in employment assistance for unemployed people. The interview schedule divided the study period (late 1980s to mid 2000s) into two periods, corresponding roughly with change of national Government in each country (1997 in the United Kingdom, 2001 in Denmark). Interviewees were asked about the main purpose and character of employment assistance programs for long term unemployed people in each of these periods (where the programs were located in the activation diamond above), and in the case of Denmark any significant differences between programs for insured and uninsured unemployed people. They were asked about the key ideas and political stakeholders influencing major policy changes, including the impact and use of official program evaluations and the role of two international institutions: the OECD and European Union. In coding their responses, and mindful of the theoretical literature discussed above, the following factors influencing policy development were identified:  Changes in the characteristics of the client group (long term unemployed people);  Benefit levels , duration and structure (social insurance and social assistance)  Labour market structure (including the skills required for entry level jobs, minimum wage levels)  Governance arrangements for activation policies (centralised, decentralised, corporatist, or market oriented)  Policy ideas (activation, rights, duties, structural unemployment, New Public Management, and labour market disadvantage).  Official evaluations of labour market programs and international policy exchange, including through the OECD and European Union. Statistical comparison (1) United Kingdom The principal unemployment payment in the United Kingdom from 1996 was ‘Job Search Allowance’ (JSA). This has a social insurance component available to former employees who meet the qualifying conditions and have made sufficient contributions, and a social assistance component for those who do not meet those qualifying conditions. The only substantive active in the labour market. To the extent that they operate in this way, they are a form of de-commodifying income protection, an alternative to securing income from open employment.

8 difference between these payments is that the social assistance payment is means test to family income, while a working hours test applies to continued receipt of the insurance component. The maximum rate of JSA is not tied to previous wages, and it is generally much lower than its Danish counterparts. Prior to 1996 there were two payments for unemployed people: Unemployment Benefit (a flat-rate insurance based payment) and Income Support (a flat-rate, means tested social assistance payment). Figure 2 shows the incidence of overall and longterm (defined as 12 continuous months) reliance on JSA and its predecessors, in proportion to the overall population of working age (not the labour force). Figure 2:

Sources: United Kingdom: Office for National Statistics NOMIS data base; OECD labour market data base. ‘Unemployment Assistance’ refers to Unemployment Benefit and Income Support for unemployed people prior to 1996, and to Jobseekers Allowance (both contributory and income based) subsequently. Note that these statistics are expressed as a proportion of the population of working age, not the labour force.

The impact of the recession of the early 1990s is clear. Afterwards, there was a sharp decline in recipients, including those in receipt of JSA long-term. This suggests, on the face of it, that British policies to reduce reliance on unemployment payments were successful, at least from the mid 1990s to the early 2000s. After this, the decline in recipients was much slower. Note that the recession of 2009 is beyond the scope of this study. (2) Denmark Income support for unemployed people in Denmark is divided into a social insurance payment called ‘Dagpenge’ which is administered by various social insurance funds largely run by the trade unions, and a social assistance payment called ‘Kontanthjaelp’ administered by the Muncipalities. Dagpenge is paid to qualifying social insurance fund members, generally at 90% of the previous wage up to a low ceiling (so that it is close to a flat rate payment for fulltime workers, but at a much higher level than the British JSA). Kontanthjaelp is paid at a lower flat rate (60% of the maximum rate of Dagpenge for a single adult) and is means tested on family income. Figure 3 shows trends in overall and long-term reliance on these payments.

9 Figure 3

10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

% of popln of working age

Reliance on unemployment insurance and social assistance (% of popln. 16-64 yrs) Denmark

Unemployment insurance (UI)

unemployment

Social assistance (SA)

Long term UA & SA

Source: Ministry of Employment (2005); Danmarks Statistik, Jobindsats series; OECD labour market database. Note that separate data were not available for long term unemployment insurance and social assistance clients, respectively, though the majority of long-term recipients were social assistance clients. Note that these statistics are expressed as a proportion of the population of working age, not the labour force.

The impact of the early 1990s recession, and a mild downturn in the early 2000s, can also be seen in this graph. From 1993, reliance on both unemployment insurance and social assistance payments fell substantially, suggesting that policies to reduce them were effective. However, reliance on social assistance payments was ‘stickier’ after the early 2000s. Table 1 compares key features of labour market and social security policies in the two countries, with the average for the OECD member countries, at the beginning and end of the study period. Table 1: Comparative labour market data Denmark

United Kingdom

Proportion of people of working age employed (%)

OECD average (a)

1999

Male: 80% Female: 71%

Male: 82% Female: 63%

Male: 78% Female: 54%

2006

Male: 81% Female: 73%

Male: 78% Female: 66%

Male: 76% Female: 57%

Proportion of people of working age lacking Year 12 or equivalent qualifications (%)

(a)

1999

20%

38%

38%

2006

18%

31%

32%

10 Denmark

United Kingdom

OECD average

Proportion of fulltime employees receiving less than 67% of median wage (% in 2000) Women Men All

(b)

18%

32%

25%

7%

14%

13%

11%

20%

17%

Gross (before-tax) benefit replacement rate (OECD summary measure, after-tax replacement rates in (c) brackets) 1990

52%

18%

28%

2007

48% (80%)

12% (57%)

29% (72%)

(d)

Expenditure on labour market programs (% of GDP) (with % of this spent on job search assistance and PES administration in brackets) 1990

1.06%

0.42%

0.5%

2000

1.89%

0.24%

0.68%

2005

1.58% (30%)

0.43% (80%)

0.56%

Denmark

United Kingdom

Index of strictness of activity requirements

OECD average

(e)

1995-97

3.0%

2.6%

2.9%*

2004

3.2%

2.4%

2.9%*

(a) OECD 2001 and 2007, Education at a glance, Paris, Organisation for Economic Cooperation and Development. (b) OECD labour market statistics; OECD average is for the 15 out of 30 OECD nations for which data are available. (c) The gross replacement rate is the income of unemployed people on maximum rates of social insurance or social assistance payments (before tax) as a proportion of wages (before tax). For 2007, the net replacement rate (after tax) is also reported in brackets (not available before 2011). The OECD summary measure is defined as the average of the unemployment benefit replacement rates for two earnings levels, three family situations and three durations of unemployment. These estimates exclude housing benefits, which significantly increase replacement rates in the UK. ‘OECD average’ is the average for the 21 countries for whom relevant trend data are available. OECD tax benefit models data base (for further details of methodology, see OECD 1994, The OECD Jobs Study. Paris, Organisation for Economic Cooperation and Development. (d) Includes employment services, training and wage subsidy schemes and disability employment programs. OECD social expenditure database; OECD 2007, Employment Outlook Paris, Organisation for Economic Cooperation and Development. (e) Based on job search requirements, requirements to participate in employment programs, and levels of sanctions. The higher the score from 1 to 5, the stricter the requirements. Danish Ministry of Finance 1999,The Danish economy - medium term economic strategy; Danish Ministry of Finance 2007, Availability criteria in 25 countries, Working Paper no.: 12/2007. * average of 16 OECD countries in the mid 1990s and 25 countries in 2005, surveyed by the Danish Finance Ministry.

11 The first two indicators in table 1 suggest that the distribution of skills and wages at the bottom of the Danish labour market is much flatter (more equal) than in the United Kingdom. Although Denmark does not have a statutory minimum wage, the workforce is highly unionised and minimum wages in industrial agreements are well above those in the United Kingdom. Further, the education level of ‘low skilled’ workers is relatively high in Denmark, consistent with higher minimum wages21. Unemployment benefit payment levels are much higher in Denmark than the United Kingdom and OECD average levels. British unemployment payments, which are only indexed to inflation declined substantially relative to low and mid-level wages, whereas Danish benefits broadly retained their relativity to wages. When housing benefits are included, payments to unemployed people in the United Kingdom are significantly higher than indicated, but still well below Danish levels. These data suggest that financial incentives for Danes to move from unemployment to a full-time low paid job are very weak. Conversely, poverty rates among working-age households have been consistently about half those in the United Kingdom (around 4% in Denmark and 8% in the United Kingdom)22. From this simple statistical standpoint, there are enduring differences between the two countries (along with others usually classified as social democratic and liberal regimes23). A key area of convergence is the steady reduction in the maximum duration of the more generous unemployment insurance payments in Denmark, which was effectively halved from nine to four years from 1994 to 2000, and now sits at two years. In the United Kingdom the duration of unemployment insurance was reduced from one year to six months in 1996. Danish Government expenditure on labour market assistance for unemployed people is among the highest in the OECD, though their spending levels are inflated in the OECD data by inclusion of educational allowances paid to participants in training programs. Consistent with its low benefit levels, the United Kingdom spends relatively little on these programs. On closer examination, there was a modest convergence in investment in employment assistance from the early to late 2000s as Denmark reduced its expenditures from a high base (especially on training and direct job creation programs in the public sector) and the United Kingdom modestly raised its spending levels from a low base (especially in job search assistance)24 However, the structure of spending on these programs was markedly different, with the United Kingdom devoting around 80% of spending to employment service administration and job search assistance compared with just 30% in Denmark: ‘Actually we spend virtually the lowest in the OECD on active labour market policies because essentially our active labour market policy is to call people in and nag them. Then if we want to spend more we call them in more often.’ (Senior official, Employment Ministry, United Kingdom)

Denmark is relatively strict in its activity requirements imposed on unemployed people (for example, to search for an accept jobs) compared with the United Kingdom, though this measure, developed by the Danish Finance Ministry, takes relatively little account of the monitoring of job search efforts and may therefore exaggerate the relative strictness of the Danish system. 21

Unlike the UK, there is no statutory minimum wage in Denmark. Minimum wages are set in each industrial agreement. The poverty line is 50% of median household income, using modified OECD equivalence scale; Data are for households with a working age head, rather than individuals. See OECD 2008, Growing unequal, Paris. Organisation for Economic Cooperation and Development. 23 OECD 2006, Employment Outlook, Paris. Organisation for Economic Cooperation and Development. 24 Weishaupt T (2011), op cit. 22

12 To summarise, these indicators suggest that the United Kingdom and Denmark continued to walk separate paths in their labour market and social security and employment assistance policies, though there is evidence of some convergence in two areas: the maximum duration of unemployment insurance payments, and overall levels of investment in labour market programs. A comparison of activation policies and influences We need to move beyond quantitative measures of the stance of activation policies to uncover their policy intent and the political forces that shape it25. For example, among work experience programs, there is a world of difference between the use of paid public sector employment to requalify unemployed people for social insurance (an income protection objective), wage subsidy schemes whose purpose is to help unemployed people learn skills on the job and prove their value to the employer (a work capacity objective), and work-for-benefit schemes designed to discourage reliance on income support (a work incentive objective). The following comparison of the development of activation policies in the two countries draws on the in-depth interviews, especially views that were common to interviewees from at least two of the three target groups (policy makers, employment service providers and academics). It also draws upon literature outlining the history of labour market policy development during the study period, including official documents such as Government policy statements and reviews, and academic analyses. The study period is divided into two parts: broadly speaking the emergence of activation policies in the late 1980s and mid 1990s, and their evolution following changes of Government in the late 1990s to early 2000s.This enables us to test the extent to which the character of activation policies was a function of the political ‘colour’ of the Government or the business cycle, and whether policy development exhibited path dependency regardless of those factors. The focus here is on the purpose, design and level of public investment in employment programs for long term unemployed people, especially the balance between income protection, work incentives and capacity building elements, and the policy influences that led to those outcomes. Discussion of each of the two periods begins with a description of relevant institutional arrangements in the labour market and social security systems at the time (the building blocks of activation systems), and identifies other key influences on the emerging activation policies including new policy ideas and evaluations(the mortar that held them together). The major shifts in activation policy, both in their initial development and evolution, are then outlined followed by discussion of the character of the resulting employment assistance programs, the extent of convergence and enduring differences between them in the two countries, and the key policy influences leading to these outcomes. The construction of activation policy in Denmark and the UK in the 1980s and early 1990s (1) Denmark As noted in the statistical comparison, the key institutional building blocks for activation policy in Denmark in the early 1990s included a labour market with a high level of workforce participation, especially among women, high wages for low-skilled workers, and a highly-qualified workforce. In concert with their fellow Nordic countries, the Danes had long favoured high levels of participation in paid work including among women. Child care was practically free and widely 25

Bonoli G(2013), The origins of active social policy, Oxford. Oxford University Press.

13 available, as was paid parental leave, though older workers had access to generous preretirement benefits which discouraged workforce participation. ‘So you have not had this redundant unskilled population which you had to put to work somewhere in services, on one hand. And on the other hand, you have had a lot of unskilled or low skilled service work in the public sector at a proper wage and so on and proper working conditions which means, also, that you have seen an increase in precarious jobs in nearly all European countries but in Scandinavia, they actually declined.’ (Academic, Denmark)

The governance system for labour market policies was ‘dense’, with multiple ‘veto points’ a strong ‘corporatist’ flavour: that is, representation of employer and employee interests was embedded in the policy making and administrative processes26. The unions essentially ran the social insurance funds and the public employment service had a tripartite governing structure comprising union, employer and Government representatives. At the broader level, the main national political parties, the Social Democrats on the Left and Liberals and Conservatives on the Right, were generally unable to govern in their own right and had to assemble coalitions with minor parties. Central Government had to deal with another player: the Municipalities, which ran a parallel social assistance payment system for those who did not qualify for unemployment insurance. Further, national labour market and social security policies were developed in consultation with the social partners, with advice from a tripartite ‘National Economic Council’. Other key building blocks for activation policy were the social security system and public employment service. For insured workers, the social security system provided a high and prolonged level of income protection. As indicated above, unemployment insurance was set at 90% of previous wages up to a ceiling. The system maintained a high level of coverage among unemployed workers. This was buttressed by regular offers of six months’ work experience and/or training in the Municipal sector at standard rates of pay, whose main purpose was to requalify unemployed workers for unemployment benefits. An insured unemployed worker could, in this way, continue to receive either unemployment insurance or a subsidised wage for up to nine years after losing their job. Thus, within our ‘activation diamond’, the main labour market program for unemployed workers at this time had an explicit income protection role. The public employment service offered job matching and vocational training services, training programs were voluntary and job search requirements were not strictly enforced. Unemployed people not eligible for insurance, including new migrants, young school leavers, and single parents, entered the separate Municipal social assistance system. Although this was predominantly funded by central Government, and there was a national legislated framework of payment rates and eligibility rules, Municipalities enjoyed a high degree of organisational autonomy. Moreover, social security policies were divided into ‘labour policies’ for insured workers and ‘social policies’ for the uninsured, administered by separate Government Ministries. Although the social security legislation required social assistance recipients to register with the public employment services and keep themselves available for employment, many were exempted from these requirements on the grounds of social disadvantages such as health problems. The main form of (voluntary) employment assistance they received was unpaid work and informal training in a separate system of Municipal ‘projects’. This form of work experience also had a strong income protection function, since it operated as an alternative to open employment rather than a path towards it. ‘Often it ended up as a project with very, very low expectations and a lot of people not showing up, and it was not really training to a relevant job, it was more wasting your time in an unpleasant way. Sometimes it wasn’t very 26

Kvist J, Pedersen L, Kohler P, Making all persons work, modern Danish labour market policies. In Eichhorst W, Kaufmann O, Konle-Seidl H 2008, Bringing the jobless into work. Berlin. Springer; Kaspersen L & Schmidt-Hansen U 2006, Corporatism in Denmark. International Centre for Business and Politics, Copenhagen Business School, Working Paper 32.

14 pleasant doing nothing, but it didn’t help people to realise how to act on the labour market.’ (Senior employment policy official, Denmark)

The Conservatives ruled in coalition through most of the 1980s, until 1993. By the late 1980s there was agreement across the political divide between the Social Democrats, Liberals and Conservatives, and among the social partners, that existing policies were not effective in reducing unemployment, which reached double digits in the recession of the early 1990s. This coalesced in the work of a Government Commission, the ‘Zeuthen Commission’ whose core idea was ‘structural unemployment’: that the main cause of high unemployment was not the business cycle but the structure of labour market institutions – wages, benefits and skills. It followed a Government Employment White Paper a few years before which called for slower wage growth and lower benefits, both of which were opposed by the unions and Social Democrats. The Zeuthen Commission, a tripartite body including the social partners, took a more consensual approach, recommending a shortening of the nine year maximum duration of unemployment insurance benefits, tighter activity requirements for unemployed people, and investment in training for long-term unemployed people to raise their skill levels27. Thus, the ideas and policies underpinning a new activation paradigm were hammered out in a structured corporatist dialogue, informed by experts. Employers regarded these policies as labour supplyimproving, unions regarded them as a more palatable alternative to cuts in wages and benefits, and the major political parties hoped that they would reduce the very high public expenditures on unemployment benefits. ‘It changed the way we looked at labour market policy, and it was quite dramatic, but a very important change in mindset from many actors. On this Committee we have both the social partners and different Ministers, and it did a lot of analysis and it did a lot of politics actually, and they found a solution.’ (Senior Employment Ministry official, Denmark)

Given the suspicion of the unions and Social Democrats that the Conservative-led Government was pursuing a policy of welfare and wage retrenchment, it took a change of Government to a coalition led by the Social Democrats in 1993 for this ideational consensus to bear fruit. The activation ‘turn’ in Denmark is generally associated with that Government’s reforms of 1994. This reform converted the previous work experience and training schemes for insured long term unemployed workers from an income protection device to a capacity building tool, by removing the ability for workers to use work experience to requalify for benefits and making work experience and training programs compulsory after four years on unemployment insurance. At the same time, the maximum duration of unemployment insurance was lengthened from two to seven years, though given the ‘carousel’ of benefits and work experience this was, in effect, a reduction from nine years to seven! These reforms were bolstered by the previous tripartite consensus over structural unemployment, and a new idea: that the rights of unemployed people to income support and employment assistance should be balanced by duties to participate in activation programs28. Consistent with prevailing New Public Management ideas, these rights and duties were embodied in a formal activity agreement between the jobseeker and the public employment 27

Goul Andersen J & Pedersen 2011, Continuity and change in Danish active labour market policy: 1990-2007.CCWS Working Paper 54, Centre for Comparative Welfare Studies, Department of Economics, Politics and Public Administration, Aalborg University. 28 The language of ‘rights and duties’ can be traced to another Government review carried out concurrently with the Zeuthen Review but with a focus on social assistance, referred to as the ‘Social Commission’.

15 service, which was intended to be tailored to individual circumstances and local labour market conditions. To facilitate this personalisation of employment services, the employment service was restructured to devolve decision-making to the regional level, at which the unions and employer organisations were represented on new management structures: ‘There was nearly a cultural revolution because it was a very standardised system because they had this offer you had to work for a period and then, after some years again, if you had no educational background, you were going into some type of education and then you had your second offer for work. …… suddenly they were supposed to sit in front of these unemployed and have some kind of sense about what is your need. And they even had to make a contract with the unemployed.’ (Academic, Denmark) But what was happening at the same time was that the social partner has huge, huge, huge influence when they made their reform in 1994. They were actually the ones in charge of prioritising what types of instruments to use, what types of target groups to give services. And they had a huge, huge budget, these regional boards. (Academic, Denmark)

The activation policies of the mid 1990s were a grand bargain between unions, employers and Government in which workers relinquished a degree of income protection in return for a guarantee of public investment in employment capacity-building programs to help reintegrate them with the labour market. This was the template for reform of activation policies for insured unemployed workers in Denmark. ‘Well you could also say that the activation for the insured was part of an agreement between the government and mostly the trade unions that okay, we’re lowering the time you can get the uninsured benefits but on the other hand we from the government would do more for these people. We will give them job training, we will give them education, etcetera, etcetera.’ (Municipal service provider, Denmark)

The resulting employment programs for insured workers had a strong work capacity-building character. In 2001, 86% of activated long-term unemployed insured workers received around six months of vocational training or work experience in the public sector paid at regular wage rates combined with training. Expenditure on employment assistance for insured workers totalled DKR 2.4 billion in that year. Given the long duration of unemployment insurance and high levels of unemployment, there was no rush to move people into open employment. The programs had a heavy emphasis on work preparation through training and paid work experience: ‘You had in the beginning much more focus on training, educational training, and you had focus on individual clients. So you did have this conversation with the unemployed and together you should find out what would be the possible and the best way back to the labour market.’ (Labour market authority official, Denmark)

Municipal social assistance clients registered with the public employment service also faced new activity requirements, but these were introduced more slowly, and many were still exempted at the discretion of Municipal social workers. Of those who were required to participate in employment programs, only 26% joined paid work experience or training programs, while 71% participated in ‘Individual Job Training’ (a period of unpaid work experience and training in a low-skilled job) or ‘Municipal projects’ (work for benefit schemes with an emphasis on social support rather than formal vocational training). Although social assistance clients were widely considered to be more disadvantaged in the labour market, and the number of these clients ‘activation’ exceeded insured workers by around 50%, investment in programs for social assistance clients totalled DKR 1.4 billion 2001, around half the amount spent for insured clients29. The reasons for the widespread exemption of social assistance clients from activation, and the different forms of employment assistance offered to them, were hotly debated. The Municipalities argued that the main reason was their clients’ greater social disadvantage, and they pointed to the classification of many social assistance clients as 29

Ministry of Employment, 2002, National Action Plan for Employment, Copenhagen.

16 ‘unemployable’ according to the prevailing profiling system for unemployed people. However, researchers from the Government’s social policy research centre found that there was a high degree of discretion and inconsistency in the profiling system30. Since only half the cost of activation was met by Central Government, the Municipalities had an incentive to exempt clients31. Meanwhile, the institutional separation of labour and social policy continued, with the Municipalities failing to gain seats on the local labour market boards. Employment programs for Municipal social assistance clients retained their essential ‘income protection’ character. This view was supported by official program evaluations in the early 2000s, which found that participants in Municipal projects rarely moved into open employment32. The same evaluations found that the most effective programs in assisting long term unemployed people into open employment were wage subsidies in the private sector, and that training programs and public wage subsidies (the main programs for insured workers) were relatively ineffective in increasing transitions to employment in the short term and delayed re-entry to the labour market (the ‘lock in effect’). They also found that job search assistance combined with the ‘threat effect’ of compulsory referral to programs had a major impact on exits from benefits to employment among less disadvantaged clients, but that their impact was weak for more disadvantaged clients33. Significantly, while the Danish Economic Council drew the conclusion that although most of the longer work experience and training programs were not cost effective, its solution was to re-focus them on transitions to employment rather than abandon them34. ‘That opened our eyes to the problems with this lock in effect – if you looked at [these evaluations] today, you might say that they were a little too tough on the outcome of education, because the problem with studying is, as you know, its effects take too long.’ (Senior Employment Ministry official, Denmark)

(2) United Kingdom British activation policies were made with very different institutional building blocks. As the statistical comparison indicates, the British labour market was characterised by a long ‘tail’ of low skilled workers and low minimum wages. Employment participation was lower than in Denmark, especially among women, and female workforce participation was not well-supported by child care and parental leave arrangements35. The system of Government was also very different. For the most part, the British ‘Westminster model’ was a ‘winner takes all’ system in which either the Labour or Conservative parties ruled with little need to consult with minor parties the social partners, or Local Government (which by the late 1980s had only a minor role in the social security system). The benefit structure was also markedly different. While there was a system of unemployment insurance, by the late 1980s the maximum payments were set at the same flat rate as social assistance, and both were in effect run by central Government. During a period of high unemployment and fiscal austerity during the Conservative Government of the 1980s, the 30

Stigaard M 2006, Municipal activation. Copenhagen, Danish national social research institute (SFI). Beer F et al 2008, State and local employment services: implementation of 'More People in Work' before structural reform. Copenhagen, Danish Institute for Social Research [SFI]. 32 Danish Ministry of Employment 2002, Effects of labour market programs. Copenhagen, Ministry of Employment; Graversen B and Weise H 2001, Effects of activation measures for social assistance recipients. Copenhagen. Danish national social research institute (SFI) Working Paper. 33 Graversen, B. 2004, Employment effects of active labour market programs: do the programs help welfare benefit recipients to find jobs? Aarhus, Department of Economics, University of Aarhus. 34 Danish Economic Council 2002, Danish Economy. Copenhagen, Danish Economic Council. Autumn 2002. 35 Crouch C 1999, Employment industrial relations and social policy, new life in an old connection. Social policy and administration 33:4 pp437-457. 31

17 insurance based elements (including earnings related supplements and access to non meanstested payments for up to two years) were progressively stripped out of the system, an instance of displacement. This process culminated in the homogenization of unemployment insurance and social assistance payments for unemployed workers into a single flat-rate Job Seekers’ Allowance in 1996. The main remaining difference between the insurance and social assistance components of this payment was the non-means testing of insured workers, but even this only continued for the first six months of unemployment. Moreover, as our statistical comparison suggests, in the early 1990s the maximum rate of benefits for unemployed people was low by OECD standards. Unemployment payments were only indexed to inflation, not movements in wages so that they fell behind community living standards over time, an instance of drift. The Jobseeker’s Allowance was widely available without time limits to unemployed workers in low income households, but it had the character of a residual poverty-alleviation program for those who were unable to support themselves in the private market. Compared with Denmark and most other OECD countries at the time, the British policy response to unemployment relied heavily on the incentive effects of low benefits and ‘self help’ rather than investment in training and other capacity building programs. For the first half of the 1980s, recipients of unemployment benefits were not even required to register with the public employment service, and the level of investment in labour market programs for unemployed people was towards the bottom of the OECD league table36. Despite the ‘winner takes all’ character of the political system, in the 1970s and 1980s labour market policies had important corporatist elements. A tripartite statutory ‘Manpower Services Commission’ (MSC) controlled the public employment service and also administered vocational training programs. In seeming contradiction with the incentives-based approach described above, the Conservative Government’s response to high levels of long-term unemployment in the mid 1980s was a large expansion of paid work experience in the public and community sectors. On average, in each of the five years before it was closed in 1988, this ‘Community Program’ provided 100,000 to200,000 term unemployed workers participants with six to 12 months of part-time work experience at prevailing wage rates37 (King, 1993). This program, along with vocational training provided by the MSC was voluntary. Since the Community Program was designed more to ‘soak up’ unemployment (at a time when it was very high) rather than prepare people for open employment, it had an income protection character. Official evaluations concluded that it was not cost effective in reducing unemployment since few participants progressed to other jobs38. Unlike the Danish work experience programs, the Community Program was not converted into the centrepiece of the new activation regime established in the late 1980s. By this stage the Government was under less political pressure over unemployment, and tensions between the Conservative Government’s market-oriented labour market policies and the corporatist system embodied by the MSC came to a head. The Employment Department and the Government – appointed head of the MSC (Lord Young) became concerned that widespread reliance on unemployment payments and the Community Program promoted ‘welfare dependency’ and the Government decided to re-orient the system away from income protection: ‘The emphasis has now to be switched from programs which provide temporary work to programs which can retrain unemployed people to compete successfully for the permanent jobs becoming available. At the same time we must motivate those who have given up looking for employment or feel that they have no incentive to return Wells J 2001, From Restart to the New Deal in the UK. Labour market policies and the public employment service, Paris. Organisation for Economic Cooperation and Development. 37 King D 1993, The conservatives and training policy 1979-92. Political Studies 41pp 214-235. 38 Price D 2000, Office of hope: a history of the employment service, London. Policy Studies Institute. 36

18 to work....In addition we must take further steps to ensure that benefits are paid only to those who are genuinely unemployed (Ministry for Employment 1988, p5) ‘One of our two secretaries of state at that point, Lord Young, was a seriously innovative thinker I think. …. One of the things he said was, "Well, if there are so many people long term unemployed and we're worried about them being out of touch, why don't we just talk to them? Why don't we call them in?" "Well, they won't come in." He said, "Well, surely they should lose their benefit, then." (Government service provider, United Kingdom)

The MSC’s control of the employment service was removed and its role was limited into a tripartite Training and Enterprise Council (TEC) administering vocational training programs for both employed and unemployed workers. The institutional link between employment services and the training system was removed, along with union and employer representation on the governing body for the employment service. In 1986 the Government conducted an experiment called ‘Restart’, comprising compulsory interviews for recipients of unemployment payments who were unemployed for six months. This, and not the large scale work experience and training programs used to activate unemployed people in Denmark, was the template used for subsequent activation policies in the United Kingdom, an instance of displacement. At its core, Restart emphasised work incentives rather than employment capacity building. Long term unemployed people were interviewed a number of times by the employment service, asked about their efforts to find work, and offered job search assistance such as coaching and help with resumes. The interviews were compulsory, the idea being to ‘shake out’ of the system those who were either in undeclared work or not actively seeking employment. If after a short period of time an individual was unable to find employment they would be offered job search training or referral to six to 12 months of training or work experience. ‘All of that led to the creation of the Restart regime which actually brought everything together for the first time. Suddenly you had something which is kind of still in place to this day, which is an advisory regime which gives you a tough advisory gateway at the very start of the process that sort of sets the scene, explains what things are like, draws up a job seeker's agreement and then puts in place an arrangement to see people roughly six monthly thereafter. Now, it has been tweaked in many ways since then; managed in many ways, but that's the fundamental.’ (Government service provider, United Kingdom)

The main vehicle for work experience and vocational training was a new program called Employment Training, run by the TEC. Fearing that the Government planned to introduce a compulsory ‘work for benefits’ scheme as part of its new ‘incentives based’ approach, the unions withdrew from the TEC and the British experiment with tripartite management of labour market assistance came to an end39. An official evaluation of Restart found that the interviews themselves had a major impact on transitions to employment, not a surprising result given the laxity of the previous activity requirements40. This, together with the fall in unemployment which relieved political pressure on the Government, convinced it that the monitoring, compliance and job search assistance elements of the Restart strategy rather than large investments in lengthy work experience and training programs should form the centrepiece of its activation policies. The Employment Training program was not introduced on the scale originally planned, and it was replaced in the

39

40

Price 2000, op cit. Dolton P and O’Neill D 1996, Unemployment duration and the restart effect, Economic Journal No 106.

19 early 2000s by work experience and training schemes of shorter duration such as ‘Training for Work’. ‘Of all the schemes introduced by the Government, Restart had the most marked effect on the unemployment 41 figures’ ‘We needed to retain a community work, work experience, social enterprise type of program for the real hard cases where there isn't much going on. That meant that the Community Action Program was born out of the ashes. But, actually, what we really needed to do was to prod, provoke, enlist, enrol people back into the labour force through a combination of sanctions through encouraging self-employment with various smaller scale but successful enterprise programs, and to actually - once people got into the labour force they would then acquire the skills that would enable them to progress. It was called a work first strategy.’ (Government service provider, United Kingdom)

From 1990 to 1993, a period of sharply increasing unemployment, spending on the Employment Service rose by almost half while spending on training programs for unemployed people fell by a third42. By the time the Conservative Government fell in 1997, unemployed people were required to participate in Restart interviews as well as employment programs, and the focus of both was ‘work first’ – to get long term unemployed people quickly into the first available job rather than to skill them up for the labour market. Unlike in Denmark, the new ‘duties’ of long term unemployed people were not balanced by a corresponding ‘right’ to substantial ‘capacity building’ programs. Rather, employment programs were developed as an adjunct to the administration of unemployment benefit activity requirements, with a clear focus on self help and work incentives. As in Denmark, official evaluations found that lengthy periods of training or work experience that lacked a clear focus on transitions to open employment had ‘lock-in effects’ that often exceeded their positive impacts on future employment, at least in the short term. This was consistent with the emerging views of the OECD43. However, the lesson drawn by the British Government from these evaluations was different to that of the Danes. The Conservative Government concluded that major investment in capacity building was not worth the cost, since the same outcomes could be achieved much more cheaply through Restart interviews and job search assistance44. The evolution of activation policy in Denmark and the UK in the late 1990s and early 2000s The Conservatives lost power to the Labour Party in the United Kingdom in 1997, and Danish politics took a turn in the opposite direction in 2001 when the Social Democrat-led coalition Government was replaced by one led by the Liberals. This study extends to the mid 2000s, just before the international recession, which enables us to assess the impact of these changes of Government in activation policies. By this stage, unemployment had fallen substantially in both countries (to 4% in Denmark and 6% in the United Kingdom) and this positively reinforced Government commitment to the new activation strategies. As the political pressure on Governments from high unemployment eased, they found themselves dealing with labour market shortages at a time when the profile of the remaining population on unemployment payments became more disadvantaged as their number 41

Chancellor Nigel Lawson, cited in Price 2000, p256. Robinson P 2000, Active labour market policies, a case of evidence based policymaking? Oxford review of Economic Policy. 16(1):13-26, p18. 43 Martin & Grubb, 2000. What works among active labour market policies. OECD Economic Studies No 230, Paris. Organisation for Economic Cooperation and Development. 44 Price 2000, op cit pp291, 292. 42

20 diminished 45. The logic of the activation paradigm was to deal with these problems by extending activity requirements and labour market supports to groups of income support recipients previously exempted from activation, especially Municipal clients and older workers on ‘pre-retirement pensions’ in Denmark, and sole parents and people with disabilities in the United Kingdom46.The challenge here was to provide the support and incentives required by these groups to (re)enter the labour market, often after a long period out of work. In both countries, there was evidence to suggest that an incentives based approach on its own was not effective in assisting long term unemployed people into jobs47. (1) Denmark The policy response of the new Liberal-led Government in Denmark to these challenges was to shift the focus of employment programs from meeting participation targets in activation programs towards transitioning people more quickly into open employment, and also to shift labour market program investment towards the ‘less activated’ Municipal social assistance clients: ‘Compared to today, efforts must be made much more flexible and must focus on the direct route to employment. 48 Most people know the stories of unemployed people who have ended up in mindless activation’ . ‘We moved in this period from a heavy weight on education and qualification to more discussion on availability as we called it – a solution that people should seek these jobs.’ (Senior Ministry of Employment official)

In addition to the tightening labour market, an important influence on this policy shift was a steady reduction in the maximum period of unemployment insurance from seven to four years introduced by the Social Democrat-led Government, combined with a tightening labour market. This in itself re-focussed employment assistance towards more rapid transitions to employment: ‘The more the demand increased the more the policies were drawn in the direction of the, what do you call it, subsidised jobs and more away from education and qualifications because we had to get people out on the labour market, and fast. Whereas in the 90s we had seven years to qualify people for the job and of course it was a lot more long term job plans were made and there was room for it as well.’ (Municipal service provider, Denmark)

There was a tension between this new emphasis on faster labour market transitions and the more disadvantaged profile of the remaining unemployed clients. So capacity-building programs such as wage subsidies and education and training were maintained for long term unemployed people and others with barriers to employment: ‘The unemployed who can take a job tomorrow just have to be conveyed to a job or helped to look for work. Those who lack qualifications will be offered training or an internship at a company. Those who lack work experience, an internship or work experience in a private or public business. Those who have problems in addition to unemployment will have support that enables them to take a job’49.

The Liberal-led Government took a number of steps towards work incentives-based strategies which in some ways brought Danish policies closer to the United Kingdom activation model. This included six monthly interviews to assist with job search and check compliance with activity 45

Cappelari L & Jenkins S 2008, The dynamics of social assistance receipt. OECD Social, employment and migration working papers No 67. Paris. OECD; Rosdahl A, and Petersen K 2006. Recipients of cash assistance, a literature review.

Copenhagen, Danish Social Research Institute (SFI). 46

Finn & Schulte 2008; Goul Andersen and Pedersen 2007, op cit Graversen, B. 2004, op cit; Hasluck C, and Green A 2007, What works for whom? A review of evidence and metaanalysis for the Department for Work and Pensions. Research Report 407. London, Department for Work and Pensions. 48 Ministry of Employment 2002, Action Plan, 'More People at Work.' p8 49 Ministry of Employment 2002, Action Plan, 'More People at Work.' p8 47

21 requirements (on the surface similar to the Restart approach), and reduced payments for some groups. However, payment retrenchment was selective. An attempt in 2003 to reduce unemployment insurance payments for workers who were previously in higher-paid jobs was repulsed by the Opposition and overall benefit replacement rates for both social insurance and social assistance payments were hardly altered50. Those whose payments were actually cut were social assistance clients, especially new migrants from non-western backgrounds who formed around 30% of social assistance clients by this stage and had become political ‘outsiders’51. In 2002 a ‘start benefit’ at a 30% to 50% lower rate was substituted for standard social assistance payments for new migrants, and in 2006 couples were denied the married rate of social assistance after two years on income support unless the ‘second earner’ had recently participated in paid work. The latter policy targeted Arabic-speaking migrant communities in which female workforce participation was low. Social assistance payments and supplements were also capped so they could not exceed the relevant unemployment insurance payments. The Government abandoned a previous administrative target that 70% of all long term unemployed people should be in an activation program (work experience and training) and used the savings to introduce a modest universal 2.5% tax deduction for employment income52. Yet the limits to policies to improve financial incentives for unemployed Danes were spelt out in an OECD report published at around this time: ‘Although [earned income tax credits] is suggested as a way of improving incentives for low paid workers, closer analysis indicates that in the Danish case – characterised by a strong compression of wages and generous 53 unemployment benefits – such schemes do not add much to labour market performance .’

The above changes to payments were either evolutionary, or at the margin. Significantly, apart from the payment cuts they were for the most part supported by the Social Democrats in Opposition. The policy changes that had the greatest potential to undermine the foundations of the existing system of labour market assistance were in the governance of employment assistance54. A key reform, the ‘Municipalisation’ of the governance of employment assistance for both insured and uninsured workers, was strongly opposed by the Social Democrats. The Government progressively transferred control of the public employment service (including employment services for insured workers) to the Municipalities. This move was partly aimed at the Social Democrat’s political support base in the unions, whose influence was much reduced in the process though unions and employers were still represented on the central and regional oversight boards for the new system. It was also intended to homogenise employment assistance for insured and uninsured workers in a way that paid more attention to activation of the latter group. Ironically, the decentralisation of employment assistance to Municipalities was also intended to tighten central Government control over employment programs in order to re-orient them towards quicker transitions to employment - by reducing the influence of the social partners. The Government gradually tightened central control over employment assistance through its funding arrangements with Municipalities. 50

Jørgensen H & Madsen P 2007, Flexicurity and Beyond, Copenhagen. DJOF Publishing. Rosdahl A, Petersen K 2006, Recipients of cash assistance, a literature review, Copenhagen.SFI. 52 Goul Andersen J & Pedersen J 2011, op cit. 53 OECD 1999, Economic Report: Denmark, p84. Paris. Organisation for Economic Cooperation and Development, p84. 54 Bredgaard T & Larsen F 2008, op cit. 51

22 ‘I think that if you look at the period from early 90s to 2007, the political intentions were to sort of make it equal, the services you could get should not depend on whether you get your income from social security or unemployment security. If you’re insured or uninsured you would, if you were able to take a job you should have the same opportunities.’ (Municipal service provider, Denmark) Some of these governance changes was quite successful in standardising and controlling the municipalities. (Academic, Denmark)

These changes to governance of employment assistance had the greatest potential to alter its character in a more fundamental way towards a work incentives-oriented model in which substantial paid work experience and training programs played little or no part. This was underscored by the Government’s decision in 2003 that Municipalities should contract out employment assistance for at least 10% of their clients to private providers on a ‘payment by results’ basis, according to short term employment outcomes55. When this strategy was pursued in Australia and the Netherlands, the result was a re-focussing of assistance away from work preparation and substantial work experience programs towards low cost job search assistance and compliance-oriented interventions56. Also, in the latter half of the 2000s the funding of Municipal employment assistance was altered to encourage regular and timely activation, and the use of employment programs with the strongest connections to open employment, especially wage subsidies and internships in the private sector. Up until 2010, the public employment service continued to assist social insurance clients and its programs were still 100% funded by central Government. However, controversially, ceilings were placed on its funding of longer training programs.57 58

‘I think was around 2007, 2006, 2007, they said it was important that as many as possible got job training . So they changed the funding for the municipalities so they got, I don't remember the percentages but if you used courses as a training program your funding was reduced to 30%, whereas the funding for job training and especially private job training was perhaps 50%.’ (Labour market authority official, Denmark)

In 2005, the Government policy statement ‘A new chance for all’ announced a concerted push to remove the remaining inconsistencies in activation for insured and uninsured unemployed workers. A new profiling system was intended to reduce exemptions from activation to a small minority of social assistance clients, whether or not they had ‘problems other than unemployment’ such as a mental illness. Municipalities were financially penalised for delays in activation, and the use of Municipal ‘projects’ was discouraged, in favour of subsidised jobs and internships in the private sector. ‘As you say the more skilled got jobs, those who did not have any social problems got jobs and what was left, those with a lot of problems. So the government also launched a special program called ‘A New Chance for All’, for the uninsured to get jobs. And they had more focus on activation on a job and not these [Municipal projects]. I think that was the main change.’ (Municipal service provider, Denmark)

By 2006, a higher proportion of unemployed workers were in activation programs, but the profile of employment assistance for insured and uninsured workers had not changed dramatically. Among insured unemployed workers who were ‘activated’, 78% participated in vocational 55

Bredgaard T,.& Larsen F, et al 2005, Contracting out the public employment service, a quasi market analysis. Transitions and risk, new directios in social policy. Centre for Public Policy, University of Melbourne. 56

Considine M, Lewis J, Sullivan, S 2009, Activating States: transforming the delivery of welfare to work services in Australia, the UK and the Netherlands. Australian Report back to Industry Partners, School of Social and Political Sciences, University of Melbourne; Sol E 2007, The individual jobseeker in the sphere of contractualism, International Journal of Sociology and Social Policy 27:7-8, pp 301-310. 57 Beer F, et al 2008, State and local employment services: implementation of the 'More People in Work' before structural reform. Copenhagen. Danish Institute for Social Research (SFI). 58

Job Training comprised paid work experience together with related training, a form of short apprenticeship.

23 training or public work experience programs (previously 86%)59. The major change was to the duration of these programs, which were mostly limited to six months or less, and a modest shift from public to private wage subsidies. Among social assistance clients; 55% participated in special activation projects or individual job training (previously 71%). The major change here was an increase in use of special activation projects and a reduction in individual job training, though these shifts in the data may be due to reclassifications in employment assistance data. This suggests a high degree of path dependency in the Danish activation system. Despite a stronger policy emphasis on ‘work first’ approaches, backed by financial incentives in central Government funding arrangements, and the findings of evaluations on their cost effectiveness, most activated social insurance clients continued to participate in training and public wage subsidy schemes60. Despite the negative evaluation findings for ‘special projects’ for Municipal clients, their use increased. It is possible that the latter trend was due in part to the activation of group of social assistance clients whom Municipal social workers regarded as highly disadvantaged. However given previous inconsistencies in assessments of labour market disadvantage, the role of institutional inertia cannot be discounted. Interviewee2: So I think okay it’s changed but it’s not changed that much. I think it’s almost the same picture. And there are some kind of historical traditions. Interviewee 1: Yes but also because a lot of the persons in the uninsured, they have other problems than unemployment and you have to work with them, you can’t just send them out into a company. You have to do something else with them. (Municipal service providers, Denmark)

Reinforcing this picture was the decline in contracting of employment services to private providers after 2005, when the Government removed the requirement for Municipalities and the public employment service to refer a minimum proportion of clients to them. The widespread privatisation of employment assistance, which had a large impact on the character of assistance offered to long term unemployed people in Australia (and more recently the United Kingdom), did not take root in Denmark. ‘The municipalities, they have a low tradition for contracting out the services, they have a tradition of keeping things inside. And all the contracted market was for the insured unemployed. So the government wanted to encourage them to still contract out services so they made some kind of additional funding if they would contract their services out. But [later] it completely disappeared.’ (Academic, Denmark)

On the whole, the Danish activation story following the change of Government in 2001 is one of path dependency and evolutionary change rather than a rupture with the previous system. There was a shift in emphasis away from capacity building among insured unemployed workers (towards shorter paid work experience and training programs) and towards a strengthening of work incentives (mainly by intensification activation requirements and their enforcement rather than changes in benefit replacement rates). There was a decline in expenditure on labour market programs driven largely by reduced spending on training for insured unemployed workers, and their re-orientation towards faster transitions to employment in a tight labour market. However, despite the attempt by central Government to homogenise employment assistance for insured and uninsured workers, most social assistance clients received assistance that was qualitatively different to that for insured unemployed –in effect, alternative forms of income protection through programs such as the special projects. It is not clear to what extent this was due to the different profile of these clients or institutional inertia. 59

Ministry of Employment (2004), More People in Work, Analysis Paper No 5; Danish Employment Council 2006, Social policy report. 60 Note that negative evaluation findings of training programs focussed on short term employment outcomes. Interviewees emphasised that their longer term impacts on employment were likely to be stronger.

24 Importantly, there was also a degree of political consensus over the broad direction of reform in the early 2000s. Apart from the lower payments for migrants and the Municipalisation of employment services, the main policy changes were supported by the Social Democrats61. Further, to a significant degree these changes were set in motion by policies of the previous Government, especially the reduction in the maximum duration of unemployment insurance payments. Despite a significant ideological shift under the Liberal-led Government towards the ‘duties’ of unemployed people, the idea that they had a ‘right’ to employment assistance as well as benefits remained: ‘There was very much an emphasis on right and less on duty and the beginning…..And then, of course, as unemployment declined, there was more and more emphasis on duties and less on right but still a fair balance, you could say, with much emphasis on the right, the so-called right to activation.’ (Academic, Denmark)

The enduring differences between employment assistance programs for insured and uninsured workers suggests that social insurance systems and their associated corporatist governance arrangements were significant influences shaping the capacity-building character of employment assistance for this group in Denmark. For insured unemployed workers, a ‘right’ to work experience and training replaced the previous ‘right’ to extended income protection. Uninsured unemployed workers, and especially recent migrants, did not have the same claim on these ‘rights’. (2) United Kingdom The election of a Labour administration in the United Kingdom was also a test of path dependency in British activation policy. Key features of the existing system were retained: the 1980s corporatist administrative system was not restored, unemployment benefits were not increased above inflation, and the recent merger of unemployment insurance and social assistance embodied in the Job Seeker’s Allowance was retained. Yet from the outset, a new approach to unemployment was a central feature of the political discourse of the ‘New Labour’ administration. The new Government had promised to move 250,000 unemployed young people off benefits and into jobs. To achieve this goal, it committed itself to strengthen and renew the activation paradigm, through a fresh combination of incentives and investment in capacity building. The Government’s language echoed the ‘rights and duties’ discourse of Danish activation reforms: ‘Our ambition is nothing less than a change of culture among benefit claimants, employers and public 62 servants – with rights and responsibilities on all sides

The core elements of the new strategy were the ‘New Deal’ employment assistance programs for unemployed people and a set of income support and tax policies designed to ‘make work pay’63. The centrepiece of the renewal of employment assistance was the ‘New Deal for Young People’, which was supplemented by a series of ‘New Deals’ for other target groups including long term unemployed adults. In one sense, this was layered on top of the previous Restart interviews. After six months’ unemployment, a young person would receive up to three months 61

Goul Andersen and Pedersen 2011, op cit. Secretary of State for Social Security 1998, New ambitions for our country: A new contract for welfare. Department of Social Security. London. Cm 3805. 63 Finn D, and Schulte B 2008, Employment first, activating the British welfare state. Bringing the jobless into work? in W. Eichhorst, O. Kaufmann and R. Kohnle Seidl. Berlin, Springer 62

25 of intensive job search assistance, the equivalent of the Restart interviews and job search courses in the previous system, but delivered in more intensive and sustained way by a personal adviser in the employment service. This was called the ‘gateway’. ‘I think you can regard the gateway almost as being an extended Restart period.’ (Official, Employment Ministry, United Kingdom)

The New Deal also converted the Restart arrangements and previous work experience and training programs into a single integrated scheme. If after three months of interviews and job search assistance a jobseeker had not secured employment they would be given four ‘Options’: work experience or training programs generally lasting up to six months that either had a work capacity emphasis (vocational training or subsidised employment), or an incentivestrengthening emphasis (work for benefits on environmental projects), depending on the adviser’s assessment of individual circumstances and needs. Participation in the New Deal was compulsory: there was ‘no fifth option’ of inactivity on benefits: ‘It was almost as though the work experience model was turned off, activation was turned on, that was progressively expanded and ramped up, and then there was an attempt to try and see whether you could get more out of combining the two than if you do the two separately. That was certainly my reading of the literature at the time and it was what I took into government as being the logic of New Deal.’ (Academic, United Kingdom)

The two elements of previous labour market programs: interviews to help with job search and work experience and training programs, were thus combined into a single ‘New Deal’ program. The key innovations were more intensive job search assistance and a new sorting mechanism: only around one third of young jobseekers with substantial barriers to employment would be referred to Options, and not all Options had a work-capacity focus. This was considered more cost effective than referring all long term unemployed people to capacity building programs. This was a hybrid activation system sitting in between the existing interview-based Restart strategy and the Danish model of activation of long term unemployment insurance recipients through work experience and training programs. In converting the Government’s manifesto commitments into a new program, the Employment Ministry emphasised its job search assistance and work incentive-strengthening aspects: ‘If you look at what they specified in the manifesto and what the New Deal looked like you could see similarities but they are not exactly the same. You say that there was a degree of continuity between what was there before and the New Deal, I'm quite pleased about that because I designed the New Deal….Actually the purpose of the ‘no fifth options’ was very much Employment Department. They were not employability measures. They were a ‘no fifth option’. They were there to say that you've got to do something active and that the focus of that activity should be that you will always be to have an eye on getting a job in the future basically.’ (Senior official, Employment Ministry, United Kingdom)

The average unit cost of the New Deal for Young People in 2000 was 1,500 pounds, which sat about half way between the cheapest elements of the Conservative Government’s activation programs (such as job search training courses and job clubs, which cost around 100-200 pounds in 2000 values) and the more expensive vocational training and work experience programs such as Training for Work (whose average unit cost was 3,000 pounds)64. To further strengthen work incentives and keep unemployed people engaged with the labour market, the public employment service and benefits office were merged into a new Government agency called Job Centres Plus:

64

Finn D, and Schulte B 2008, op cit.

26 ‘The more I've been around the more I think the financial incentives are not the big drivers in all of this. That's why you need other incentives and disincentives which are the administrative incentives.’ (Senior Official, Employment Ministry, United Kingdom)

Within three years, with the aid of a tightening labour market, youth unemployment fell substantially and the political target was met. The Government turned to other groups to draw them into the new activation regime: long term unemployed adults, sole parents and people with disability. Since long term unemployed adults were a much more disadvantaged in the labour market than young people, this was a test of the Government’s commitment to investment in capacity building. Here, the Government faltered, at first not including ‘Options’ within the New Deal for Long Term Unemployed adults and then implementing a ‘cut-down version’ of the New Deal for Young People in which paid work experience or vocational training were either not offered, or were of very short duration. The voluntary New Deal for Lone Parents essentially offered career counselling and job search assistance. Compared to a unit cost of 1,500 pounds for the New Deal for Young People, the New Deals for long term unemployed adults and lone parents cost an average of 400-500 pounds per client. By 2003, one third of ‘New Deal for Young People’ participants had already been through the program previously, and this ‘churn’, especially of lower skilled and disadvantaged jobseekers, was the subject of policy concern and debate65. Advocates for greater investment in capacity building argued that long term unemployed people, especially adults, were not gaining the skills they needed to move from insecure low skilled jobs into sustained employment. In response to these concerns, the Government introduced on an experimental basis the StepUp program for people who had already been through a New Deal program and were still unemployed. Modelled on ‘transitional labour market programs’ run on a small scale by local community organisations, StepUp provided a combination of paid work experience in a regular job, relevant training and mentoring. Despite the findings of its evaluation that StepUp had a significant impact on employability among long term unemployed adults, the program was discontinued after a few years: ‘Yeah exactly, the group who have proved chronically hard to place. So the options were you repeat the [New Deal] program ad nauseum, right until such time as they die or something. You put in place some kind of alternative more intensive program, which is what Step Up was, or you move towards some kind of punishment model or ‘work for the dole’ kind of model. So the Labor government tried the ILM model, such as Step Up, and there were various experiments, small scale things dotted around which looked moderately encouraging given the severely deprived nature of the group that you’re talking about but it was expensive. It never got past that – the evaluation was not strong enough in my view rather than it was not good... But [the Employment Ministry] felt that it was too expensive for the kit that it was getting. So they never pursued that extensively.’ (Academic, United Kingdom)

Instead, The Government shifted the New Deals in a new direction. The ‘Flexible New Deal’ (FND) was introduced in 2007 to streamline all of the population-based New Deals into a single, more flexible program. The logic of this move, which was supported by evaluation evidence, was that the New Deals would be more cost effective if assistance was personalised rather than tied to each benefit or population category66. Rather than paying work experience and training providers to deliver fixed programs, they would be paid according to short-term employment 65

Bivand P 2006, Evaluation of StepUp pilot, final report. London. Department for Work and Pensions.; Finn& Schulte 2008, op cit. 66 Department for Work and Pensions 2007, Flexible New Deal evidence paper, London. DWP.

27 outcomes. This would give the model the flexibility required to scale it up and extend it on a compulsory basis to new groups including people with disability and sole parents who up until then were not required to participate in their respective ‘New Deals’. However, the actual content of the FND was a lower cost, abbreviated version of the original. ‘Well we’re driven by the sort of per unit costs and you know, there has been a progressive reduction over the period. New Deals early on, it was probably still paying out at around about 2,000 pounds, maybe two and a half thousand pounds and FND was considerably less and Work Program was less again. Now, the issue there is that, you know, that’s feasible within a program design that does have the Gateway upfront and is high volume, low cost, moves people onto other destinations. (Official, Employment Ministry, United Kingdom)

The design of the FND was influenced by the New Public Management idea of contracting out employment services and paying providers according to employment outcomes. When the New Deals were first introduced in the late 1990s, a parallel experiment called ‘Employment Zones’ was undertaken in which long term unemployed people were referred to private providers who were paid according to their effectiveness in reducing reliance on unemployment benefits. This funding model sharpened the focus on quick employment outcomes: ‘This particular funding device helps explain why job search, job preparation and short work-focussed training courses tend to be used in preference to longer and more structured programs of training and assistance....The work first approach is thus a direct consequence of the way in which Employment Zones are funded and 67 incentivised ’.

In the FND, the four New Deal ‘Options’ (which were only fully in place for young people) were reduced to a four week period of compulsory work-for-benefits intended to reduce incentives to remain on payments. The FND looked forwards towards outcomes-based funding of employment services, but it also looked backwards to the Restart model, with its emphasis on interviews, job search assistance and incentive effects: “Well, the Flexible New Deal was the first one where they got rid of the personal advisor and they got rid of the follow through - it became much more harsh. It wasn't about what is the best? It was more conditionality than: ‘we can help you but you've got to take up the help’.” (Senior official, Employment Ministry, United Kingdom)

Another elements of the Labour Government’s employment policy at this time carried the potential of stronger investment in capacity building for long term unemployed people. An official review of workforce skills policies recognised that the long ‘tail’ of low skilled workers was an underlying cause of structural unemployment and recommended a guarantee of access to training for at least semi-skilled qualifications for all school leavers and employees: ‘The short-term focus [on employment outcomes within 13 weeks] moves attention away from pre-work interventions that might improve the sustainability of employment, as well as creating little incentive to build links 68 with in-work support that might improve retention. Two-thirds of Jobseekers’ Allowance claims are repeat ones’ .

However, in its response to the Leitch review the Government did not change the delivery of employment and training services for unemployed people much on the ground. Beyond ‘skills health checks’ and advice from a new ‘Careers Service’, activation policies for unemployed people continued to follow a ‘work first’ approach. Skills training was mostly reserved for those who had already secured employment.: ‘The idea is how can we join up work-first employment interventions which get people through the door, with then training programs which will progress them, advance them, help them stay in work?’ (Academic, United Kingdom)

67

Griffiths R, and Durkin S 2007, Synthesising the evidence of Employment Zones. London, p17. Department for Work and Pensions. 68 Leitch S 2006, The Leitch Review of Skills: prosperity for all in the global economy - world class skills. Final report. Norwich, UK Treasury, p13.

28 Despite the prominence of the New Deals in the Government’s employment policy discourse, and the significant boost to investment in capacity building for long term unemployed young people through the New Deal for Young People, its fiscal commitment to them was overshadowed by the other key element of the new activation system: ‘make work pay’ policies. In 2000, 910 million pounds was spent on labour market programs for unemployed people (of which one third was spent on the New Deal for Young People), while the cost of the centrepiece of the Government’s ‘make work pay’ policies, the Working Families Tax Credit, was 6,000 million pounds69. Together with a new national minimum wage, the Working Families Tax Credit was intended to improve financial incentives for families to move from benefits to at least 16 hours a week of low paid employment. Increases in unemployment benefits or investment in capacity building programs for unemployed people to bring them closer to Danish levels was not part of the agenda. ‘The ‘better off in work’ line of thinking strand ran through the Blair Government just as vigorously. The only difference was that their proposed solution was not lower benefits or a tougher benefit regime, because they thought all that had been done. The only other thing you could do was affect the other side of the equation; affect the demand side by saying … we'll pay you in-work benefits.’ (Government service provider, United Kingdom)

These policies also had an income protection purpose for low paid workers, a key Labour Party and trade union constituency. The Government committed itself to halve child poverty in a decade and set about achieving this through a combination of improvements in family payments, the new tax credit, and ‘work first’-oriented activation policies: ‘Then the final bit is the adaptation from the US looking at the experience of work supplements alongside work requirements, and the decision to first of all introduce working tax credits, or family tax credit it was called, but then crucially in the dynamics of the New Labour politics and progressive politics, how that agenda then becomes absorbed into the child poverty agenda, and we see tax credits used as one of the primary tools for reducing child poverty, a small part of which is about Making Work Pay. Mostly it's about reducing child poverty.’ (Academic, United Kingdom)

The main area of convergence with Danish employment policies in the 2000s was the extension and intensification of personalised activation. The work capacity-building elements of British activation policies were strengthened through the New Deals, and towards the end of the Labour Government’s tenure a number of seeds were planted that might have taken capacity building further through programs such as Step Up and the skills strategy. However, as in Denmark, there was a strong element of path dependency in British labour market policy. The ‘Restart’ activation ‘template’ which prioritised regular interviews and intensive job search assistance was stamped upon the New Deals, especially their final iteration, the ‘Flexible New Deal’: The basic institutional underpinnings of the British activation model –including tight central Government control with minimal social partner involvement, low unemployment benefits, a ‘work first’ approach, and a large low-paid workforce, remained in place despite the change of Government. Capacity building programs were strengthened, but harnessed to an activation strategy that had job search assistance and work incentives at its core: In contrast, activation of insured unemployed workers in Denmark essentially meant enrolment in capacity building programs that were converted from their original income protection purpose,

69

Finn and Schulte 2008, Table 31.

29 albeit with a stronger emphasis on faster transitions to open employment after the Government changed and the labour market tightened: ‘The options in the New Deals in the UK were sort of part of an activation system and had a particular purpose as you explain. My sense is that [what] they were doing - was different in those other countries [Denmark and Holland]. They were seen to have their own intrinsic purpose which, in addition, was activating.’ (Senior official, Employment Ministry, United Kingdom)

Conclusions The evidence from the interviews and official literature on activation policies in Denmark and the United Kingdom supports our first hypothesis: that both countries introduced and then progressively extended activation policies because they offered Governments of both Social Democratic and Liberal or Conservative persuasions an alternative to the more politically challenging course of payment retrenchment. The activation paradigm was accepted in both countries because their labour market policies and stakeholders historically favoured high workforce participation. In that sense, there was policy convergence away from income protection at the expense of labour market participation towards policies that encouraged, required and supported transitions to employment. This was supported at the level of ideas by a combination of elite concern about structural unemployment and the cost of social security and a ‘rights and duties’ agenda which Governments projected into the public discourse. The evidence from both the statistical comparison and the interviews supports the second hypothesis, that Denmark would rely more on work capacity-building activation strategies while the United Kingdom would rely more on incentive-strengthening approaches. This difference persisted despite a policy re-alignment in Denmark in the 2000s towards the British emphasis on quick employment outcomes, and growing concern in the United Kingdom that the low skills of unemployed people might be responsible for the ‘churn’ of unemployed people between New Deal programs, employment, and back to benefits. Employment assistance policies for long term unemployed people exhibited a high degree of path dependency in both countries, after the initial ‘activation turn’ in the late 1980s and early 1990s. In Denmark the template for activation of insured long term unemployed workers was participation in intensive work experience and training programs that were adapted from their original income protection purpose to fulfil work capacity building objectives. Activation of insured workers was for many years defined as participation in these programs. In contrast, the benefits and employment assistance arrangements for Municipal social assistance clients continued to emphasise income protection. Those who were not exempted from activity requirements on the grounds of social disadvantage were usually placed in ‘make work’ programs with little connection to the labour market. In the United Kingdom, the template for activation was job search assistance and monitoring of work test compliance through Restart interviews, which was later intensified through the ‘gateway’ phase of the New Deal. Work experience and training programs were treated as ‘add-ons’ and targeted towards the most disadvantaged and the least engaged. More broadly, the United Kingdom invested substantially more in ‘make work pay’ policies specifically targeted towards unemployed people and substantially less than capacity building programs for unemployed people than in Denmark. These differences in the treatment of insured and uninsured workers in Denmark, and between long term unemployed people in Denmark and the United Kingdom, persisted despite changes on Government in both countries in the middle of the study period.

30 The interviews clearly point to one source of this path dependency: the institutions delivering labour market policies – the benefit system and employment assistance and the stakeholders surrounding those systems. In both Denmark and the United Kingdom, unemployment insurance, especially when embedded in corporatist forms of labour market regulation, was associated with an emphasis on work capacity building policies70. Insured workers had rights to employment assistance (in addition to benefits), which did not apply in the same way to recipients of social assistance payments. These rights were embedded in the trade-off between longer income protection and capacity building for insured unemployed workers in the 1994 activation reform in Denmark, but they were not yet cemented in place in the United Kingdom before its corporatist phase of labour market policy development, and indeed unemployment insurance itself, were effectively terminated. This suggests that, as Clasen and Clegg and Palier argue, the structure of the benefit system and the interests represented in its administration generates path dependency in labour market policies affecting unemployed people – either facilitating or blocking activation programs and also shaping the character of those programs71. Once a benefit system is set in place it is politically difficult to fundamentally change it, and labour market assistance for unemployed people is largely an outgrowth of the benefit system. Over time, in each country, activation policies have ‘congealed’ around the underlying benefit structure, as senior officials within their Employment Ministries imply: ‘In the old days it was totally different systems in social welfare, and it was also part of the period in the new government from the start of 2001, that this system should be more and more integrated and the same. You might say from a scientific point of view that it was a bit silly because there’s always a fundamental difference,… and the main difference between these systems, as you know, in the unemployment system there’s no means testing of the family income. It’s a minor difference you could say, but in practice, it’s a totally different world.’ (Senior official, Employment Ministry, Denmark) ‘So I do sort of, I feel that the fact that we don't have a social insurance system anymore, does give – the social insurance makes the benefits feel like rights and entitlements to the recipients, that it’s their money. Which means that the activational stuff is more naturally applied to social assistance, and that’s kind of why that model was not that advanced in the UK and it was scrapped a long time ago. What was interesting about Denmark is that as you say they had a strong social insurance model but felt a need to keep people in the system. So they were creating the time limit and then they had to requalify people.’ (Academic, United Kingdom)

Further research could usefully explore other explanations for persistent differences in activation strategies in the two countries. One possibility is that employment programs were developed and altered pragmatically in response to changes in the labour market (including labour market tightening in the early 2000s), and the profile of unemployed people. This includes the greater labour market disadvantage of social assistance clients in Denmark, though it is not straightforward to identify the equivalent population within the British benefits system. This study has not examined gender differences in the treatment of benefit recipients or the ways in which different ‘care regimes’ shape activation strategies, through this could be incorporated into both sides of the activation model used in this study: work capacity building and incentives72. 70

Larsen F 2006, The importance of institutional regimes for active labour market policies, the case of Denmark, Transitional Labour Markets Workshop, Rotterdam, CARMA, University of Aalborg. 71 Clasen J & Clegg D 2006, Beyond activation, reforming European unemployment protection systems in post-industrial labour markets, European Societies, 8:4, pp527-553. Palier B , Beyond retrenchment, Center for European Studies, Working Paper Series 77, June 2001 , in Gilbert, N van Voorhis, R, Activating the unemployed, New Jersey, International Social Security Association.

72

Daly M & Lewis J 2000, The concept of social care and the analysis of contemporary welfare states, British Journal of Sociology 51:2.

31

With these caveats, this study supports, but does not necessarily confirm, the underlying premise of the second hypothesis: that there was less scope for incentive-strengthening policies in Denmark (beyond compulsory activity requirements, which were intensified in both countries) due to its higher benefits and the need for unemployed workers to improve their skills in order to compete for entry level jobs while in the United Kingdom, the labour market was structured differently in ways that made a ‘work first’ approach the path of least resistance. In particular, benefits and minimum wages were lower and there was easier access to low-skilled jobs: ‘I haven't really thought about this but I think it may have something to do that the entry points in both - to work, in both the Netherlands and Demark to some extent are more formal than they are in the UK….There's much more unionisation. There's much more tripartism. There is a segmented labour market. So if you want to undertake better jobs and do training for better jobs those better jobs are all in …the sort of standard labour market. … That sometimes means that they need to get to a higher level possibly in order to get into this more formal element. …Whereas in the UK… there's big advantages of having lots of opportunities but it isn't half wasteful in terms of information. You just need (a) to know what the opportunities are and (b) you just need someone kicking you to keep on looking for them.’ (Senior official, Employment Ministry, United Kingdom) ‘And I think it was also a period where it was sort of more a modernisation of the Danish economy from the type of jobs where you could just accept people from the street making simple work, to a more advanced economy where you just couldn’t use people without some sort of more engagement. So this old type of public employment service were gradually brought down’ (Senior Employment Ministry official, Denmark)

This brings us back to welfare regime theory. After comparing investment in human capital strengthening and incentive-building labour market programs in a number of countries, Bonoli concluded that the former were favoured by Social Democratic Governments and the latter by Liberal and Conservative Governments73. This study supports that view, but goes further: it appears that that these differences are embedded in welfare regimes and persist despite changes of Government. Activation was pursued in both countries because it met a need for Governments of both Left and Right to reduce social security expenditures without retrenching benefits. But the same activation ‘wave’ took on a different shape and course in each country, depending on the structure of the underlying benefit systems and labour market institutions. To fundamentally alter the course of its activation policies, a country would have to change its labour market institutions.

73

Bonoli G 2010, op cit.

32 References: Aurich P 2009, Activating labour supply, European strategies from 1990-2005. Conference Paper, Activation and security, Brno, ASPEN. Barbier J & Ludwig Mayerhofer W 2004, the many worlds of activation, European Societies 6:4, pp 423-436. Castles F & Obinger H 2008, Worlds, families regimes - clusters in European and OECD area public policy, West European politics 31:1, pp321-344. Beer F et al 2008, State and local employment services: implementation of 'More People in Work' before structural reform. Copenhagen, Danish Institute for Social Research [SFI]. Bivand P 2006, Evaluation of StepUp pilot, final report. London. Department for Work and Pensions. Bonoli G 2010, The political economy of active labour market programs, Politics and Society, 38:4, 435-457. Bonoli G 2013, The origins of active social policy, Oxford. Oxford University Press. Bredgaard T,.& Larsen F et al 2005, Contracting out the public employment service, a quasi market analysis. Transitions and risk, new directios in social policy. Centre for Public Policy, University of Melbourne. Cappelari L & Jenkins S 2008, The dynamics of social assistance receipt. OECD Social, employment and migration working papers No 67. Paris. OECD Carcillo S & Grubb D 2006, From Inactivity to work: the Role of active labour market policies. OECD Social employment and Migration Working Paper No 36. Castonguay J, and Sol E 2007, Is work working? Social security in the labour market. Warsaw, International Social Security Association. Clasen J & Clegg D 2006, Beyond activation, reforming European unemployment protection systems in postindustrial labour markets, European Societies, 8:4, pp527-553. ClasenJ & Clegg D (2011), ‘Unemployment protection and labour market change in Europe: towards ‘triple integration?’, in Clasen J & Clegg D, ‘Regulating the risk of unemployment: national adaptations to post-Industrial labour markets in Europe.’ Oxford, Oxford University Press. Considine M, Lewis J, Sullivan, S 2009, Activating States: transforming the delivery of welfare to work services in Australia, the UK and the Netherlands. Australian Report back to Industry Partners, School of Social and Political Sciences, University of Melbourne. Crouch C 1999, Employment industrial relations and social policy, new life in an old connection. Social policy and administration 33:4 pp437-457. Daly M & Lewis J 2000, The concept of social care and the analysis of contemporary welfare states, British Journal of Sociology 51:2 Danish Economic Council 2002, Danish Economy. Copenhagen. Danish Employment Council 2006, Social policy report. Danish Ministry of Employment 2005, Analysis of long term unemployment in Denmark. Copenhagen Danish Ministry of Employment 2002, Action Plan, 'More People at Work.' Danish Ministry of Employment 2002, Effects of labour market programs. Copenhagen, Ministry of Employment; Graversen B and Weise H 2001, Effects of activation measures for social assistance recipients. Copenhagen. Danish national social research institute (SFI) Working Paper. Danish Ministry of Employment, 2002, National Action Plan for Employment, Copenhagen. Dolton P and O’Neill D 1996, Unemployment duration and the restart effect, Economic Journal No 106. Eichhorst W &KonleSeidl R (2008), Contingent convergence, a comparative analysis of activation policies, Bonn. IZA Discussion Paper No. 3905. Esping Andersen G 1998, Social Foundations of post industrial economies, Oxford, Oxford University Press. Finn D, and Schulte B 2008, Employment first, activating the British welfare state. Bringing the jobless into work? in W. Eichhorst, O. Kaufmann and R. Kohnle Seidl. Berlin, Springer. Goul Andersen J & Pedersen 2011, Continuity and change in Danish active labour market policy: 1990-2007.CCWS Working Paper 54, Centre for Comparative Welfare Studies, Department of Economics, Politics and Public Administration, Aalborg University. Graversen, B. 2004, Employment effects of active labour market programs: do the programs help welfare benefit recipients to find jobs? Aarhus, Department of Economics, University of Aarhus. Griffiths R, and Durkin S 2007, Synthesising the evidence of Employment Zones. London. Department for Work and Pensions

33 Hasluck, C. and A. Green 2007, What works for whom? A review of evidence and meta-analysis Hasluck C, and Green A 2007, What works for whom? A review of evidence and metaanalysis for the Department for Work and Pensions. Research Report 407. London, Department for Work and Pensions. Johansson, H and Hvinden B 2007, Reactivating the nordic welfare states. International Journal of Sociology 27:7/8, pp 334-346. Jørgensen H & Madsen P 2007, Flexicurity and Beyond, Copenhagen. DJOF Publishing. Kaspersen L & Schmidt-Hansen U 2006, Corporatism in Denmark. International Centre for Business and Politics, Copenhagen Business School, Working Paper 32. King D 1993, The conservatives and training policy 1979-92. Political Studies 41pp 214-235. Kvist J, Pedersen L, Kohler P, Making all persons work, modern Danish labour market policies. In Eichhorst W, Kaufmann O, Konle-Seidl H 2008, Bringing the jobless into work. Berlin. Springer. Larsen F 2006, The importance of institutional regimes for active labour market policies, the case of Denmark, Transitional Labour Markets Workshop, Rotterdam, CARMA, University of Aalborg. Leitch S 2006, The Leitch Review of Skills: prosperity for all in the global economy - world class skills. Final report. Norwich, United Kingdom H.M. Treasury. Martin J & Grubb D 2000, What works among active labour market policies. OECD Economic Studies No 230, Paris. Organisation for Economic Cooperation and Development OECD 1994, The OECD Jobs Study, Paris, Organisation for Economic Cooperation and Development. OECD 1999, Economic Report: Denmark. Paris. Organisation for Economic Cooperation and Development. OECD 2006, Employment Outlook, Paris. Organisation for Economic Cooperation and Development. Palier B, Beyond retrenchment, Center for European Studies, Working Paper Series 77, June 2001. Pierson P 2000, Increasing returns, path dependence and the study of politics American Political Science Review 94:2, pp251-267. Price D 2000, Office of hope: a history of the employment service, London. Policy Studies Institute. Robinson P 2000, Active labour market policies, a case of evidence based policymaking? Oxford review of Economic Policy. 16(1):13-26, p18. Rosdahl A and Petersen K 2006. Recipients of cash assistance, a literature review. Copenhagen, Danish Social Research Institute (SFI). Sol E 2007, The individual jobseeker in the sphere of contractualism, International Journal of Sociology and Social Policy 27:7-8, pp 301-310. Starke P, Obinger H, Castles F, 2009, Convergence towards where? Journal of European Social Policy, 15:7, p975; Rueda D 2006, Social democracy and active labour market policies: insiders, outsiders and the politics of employment promotion, British Journal of Political Science, 36:3, pp 385-406. Stigaard M 2006, Municipal activation. Copenhagen, Danish national social research institute (SFI). Streek W and Thelen K 2005, Beyond continuity, institutional change in advanced political economies, Oxford. Oxford University Press. United Kingdom Department for Work and Pensions 2007, Flexible New Deal evidence paper, London. United Kingdom Ministry of Employment 2004, More People in Work, Analysis Paper No 5 United Kingdom Secretary of State for Social Security 1998, New ambitions for our country: A new contract for welfare. Department of Social Security. London. Cm 3805