This study examines policies that can successfully address long-term unemployment. It focuses on Denmark and Sweden, where, despite sizeable job losses during the crisis, labour market indicators are at present better than in any other EU country. By looking at the interaction between labour market flexibility (especially in hiring and firing regulations) and passive and active policies, we argue that well-designed active policies matter more than labour market flexibility for employment.
After the onset of the global financial crisis, the number of unemployed increased steadily across Europe, and only in the last couple of years have there been some feeble improvements. Stubbornly high levels of jobless people are a major concern, and they hinder future recovery prospects. The longer the period that one is jobless, the more difficult it is for that person to re-enter the labour market. There are various factors that contribute to this, including the atrophy of skills during unemployment, an increase in discouragement among job seekers (which leads many unemployed to reduce their searching activities) and a stigma in the perceptions of many employers towards any long-term unemployed job seeker. When youth comprise the bulk of the long-term unemployed population, then serious concerns arise regarding the long-term loss of human capital and, consequently, future growth prospects.
Against this backdrop, what strategies can countries in Europe implement to fight long-term unemployment? What policies have proven to be effective in re-instating job seekers into the workforce?
Looking at the best performers in the EU as far as employment is concerned, this study examines how the combination of labour market policies and labour market regulations (in particular hiring and firing regulations) matters for reducing unemployment. The literature has long investigated the properties of these two factors – labour market policies (both passive and active policies) and flexibility – in enhancing employment. Yet, to date, little has been said regarding the interplay between the two.
This study focuses on this interplay, examining two EU Nordic countries, Denmark and Sweden, whose labour markets are at present performing better than those of many other European economies. With unemployment rates at about six and seven per cent respectively at the end of 2015, the two countries are well below the EU average of nine per cent. To better understand what made these two economies more resilient to the crisis, we compare and contrast labour market institutions (flexibility) as well as policies, and we discuss the role these have played for job creation. We first provide evidence on the main differences related to labour market flexibility, in particular hiring and firing regulations, and show that Denmark has a rather more flexible labour market than Sweden. Then we compare the labour market policies of these economies and highlight how they both dedicate high levels of spending to active labour market policies (ALMPs), although this spending is targeted to different categories of ALMPs. We then report evidence of job creation as well as labour market matching and show that while Denmark exhibited higher turnover and a quick job-finding rate, Sweden was better able to create and preserve jobs during the crisis. We also illustrate how ALMPs in both countries enabled better matching between job seekers and employers. In light of this evidence, we argue that well-designed active policies are what really matter for employment performance, more than the degree of labour market flexibility.
Success is based on a broad range of factors, and as such, it is not sufficient to simply look at the labour market characteristics of the Nordic model. Our conclusion thus takes a look at the broader picture and also discusses future challenges facing the Nordic labour markets.
Labour market flexibility and policies
The current consensus in labour economics posits that by making labour markets more dynamic, flexible hiring and firing practices would provide a better and more rapid allocation of resources, and this would improve growth. While the impact on growth is not questioned, and as such it has paved the way for the structural reform agenda in the context of the recent crisis and current recovery, uncertainty remains regarding the implications of labour market flexibility on employment. As argued by Pissarides,1 looser firing and hiring regulation increases labour turnover, thereby reducing employment and unemployment periods. As a consequence, the final impact of more labour market flexibility on employment (and unemployment) remains ambiguous, because it depends on whether job creation would prevail over job destruction. Empirical evidence on the matter is also mixed.2 Some studies find that looser hiring and firing regulations reduce unemployment, while others show that by increasing searching and matching activities, labour market flexibility may indeed contribute to higher frictional unemployment.3
Regarding ALMPs, most studies tend instead to find that spending on these policies, and specifically on activation, can help reduce unemployment, reversing the commonly found positive link between unemployment insurance and unemployment.4 Some of these policies have been found more effective than others in reducing unemployment, and for each policy the impact changes according to the time lag considered. A review of studies finds that job search assistance programmes have a positive impact on employment, while classroom and on-site training programmes yield positive results in the medium term but can have insignificant impacts in the short term.5 More specifically, data on EU labour market policies show that if the active labour market policy of activation takes place within the first year of unemployment instead of the second, the unemployed will find a job faster and they are more likely to find a better-paid job.6 A key element of successful ALMPs has been found to be the “coaching” or conversations between unemployed people and caseworkers. On this topic, studies by Koch et al. and Brown and Koettl show the benefits of conversations and that the more conversations with caseworkers, the better the outlook for the unemployed in the labour market.7
As highlighted in cross-country reviews of the impact of ALMPs, country-specific characteristics do matter. In particular, the type of labour market structure may have implications on the way ALMPs stimulate job creation. The importance of the interplay between labour market conditions and policies has been, for instance, advocated as the catalyst for the success of the Danish “flexicurity” model. As argued in Bjorsted and Dahl,8 flexibility in terms of hiring and firing, generous unemployment benefits and social assistance, and ALMPs are the three key ingredients in the flexicurity model. The flexibility of the labour market in terms of hiring and firing ensures that even during bad times there will be job openings. To compensate for the low protection against being fired, a strong safety net is needed to secure income during spells of unemployment. Finally, to neutralise any potential risk of moral hazard problems arising from generous unemployment benefits, ALMPs are needed to ensure that unemployed people actively search for jobs and to provide them with help and guidance. Compared with the Danish flexicurity system, the Swedish Rehn-Meidner welfare model relies more heavily on labour force mobility in exchange for income and employment security. A main condition to achieve these goals was the presence of strong trade unions (usually blue-collar unions); more recently, an emphasis has been placed on employment protection rather than job protection, to ensure that people can get some kind of job.9
Labour market trends
The 2008-09 global financial crisis hit Denmark and Sweden as severely as it did the rest of Europe. Both countries experienced significant GDP losses of more than five per cent of GDP between 2007 and 2014, more than the eurozone average. Similarly, both countries experienced rising levels of unemployment – for Denmark the drop in employment was greater than the drop in output, whereas in Sweden the employment loss was less than half the GDP loss (see Figure 1). Seven years have passed since the onset of the crisis, and the two countries are performing very well economically. Employment rates in the two countries continue to be among the highest in the EU, at 76% and 80% of the active population respectively. The unemployment rate in Denmark is only 6.0% and in Sweden it is 7.4%. Meanwhile, the corresponding figures for the EU and the eurozone are 9.4% and 11.0% respectively. Only about 20% of the unemployed in Denmark are long-term unemployed, and in Sweden the figure is 25%. This contrasts sharply with the average of approximately 50% in the EU and the eurozone. Less than ten per cent of 15-24 year-olds in Denmark and Sweden are long-term unemployed, while the share in the eurozone is around 40% (see Figure 2).
GDP and employment losses, 2007-14
Source: Own calculations based on data from Eurostat.
Employment, unemployment and long-term unemployment rates
Labour market flexibility
Denmark and Sweden feature two different degrees of labour market flexibility. As shown in data from the World Economic Forum, hiring and firing rules are much more flexible in Denmark than in Sweden, where the labour market flexibility index is at par with the EU average (see Figure 3).10 This also explains Denmark’s larger drop in employment during the crisis. In a nutshell, Sweden grants employees more protection against dismissal (both collective and individual), as it requires employers to consult with the concerned trade unions before making any decision with respect to a business. The notice period for redundancies is much longer in Sweden than in Denmark, and some retraining and reassignment of the redundant person is usually expected in Sweden. Rules to protect workers from unjust termination are much stricter in Sweden, as employees can apply for an invalidation of a termination or dismissal. Fewer obligations apply in Denmark, where, for instance, reinstatement orders are possible but rare; also, less protection applies to employment outside collective bargaining agreements. Other salient differences in the labour market regulation of the two countries include a more regulated working hour scheme in Sweden but a longer minimum contribution period for unemployment protection in Denmark.11
Hiring and firing regulations in the EU
Source: World Economic Forum: Global Competitiveness Report 2015-16.
Labour market policies
Labour market policies relate to those expenditure items that have been designed for the unemployed or under-employed. Traditionally, passive policies refer to unemployment insurance schemes and other types of welfare benefits. The term active labour market policies refers, instead, to activities that are directed at shortening unemployment spells by proactively helping jobless people to re-access the labour market. Job centres offer a variety of programmes and incentives that can be used or activated only if job seekers take specific actions. These include a wide range of activities, from recurring conversations and guidance about employment and education to public and private work experience for a limited amount of time. Looking at cross-country evidence, data show that Denmark and Sweden are at the forefront of the OECD for spending on and the use of ALMPs (see Figure 4). It is interesting to note that while unemployment benefits in Denmark remain high – higher than the OECD average – this is not the case in Sweden, where activation policies coincide with more frugal spending on benefits. Among the various programmes, a large share of spending and coverage is targeted towards training in Denmark and towards employment incentives in Sweden (see Figure 5). More precisely, the OECD classification of training refers to all measures undertaken for labour market policy reasons, other than special programmes for youth and the disabled. Expenditures include both course costs and subsistence allowances to trainees. Employment incentives instead consist of income maintenance and support payments to formerly unemployed individuals who have taken up part-time or full-time employment. Nonetheless, the use of OECD data and classification warrants some caution, as definitions do differ across countries.
Labour market policy spending and coverage
ALMPs most widely used
Whether spending on training is preferable to direct employment incentives and other measures remains an open issue. Yet, more specific information about training and employment incentive programmes both in Denmark and Sweden can shed light on what factors make these policies work.
In Denmark in 2011, the employment rate for skilled workers was 81%, while only 56% of unskilled workers were employed. This supports the claims that unemployment programmes targeted towards educating the unskilled will have a positive effect on employment. An example of such programmes is the adult apprentice programme, which ensures that unskilled workers above the age of 25 can become skilled even if they have no educational background beyond primary school. The programme gives employees who make an agreement with an institute of vocational education the right to a special salary supplement during the period of work training. One of the many requirements is that the education can lead to a job in a field with a lack of workers. The importance of training as a labour policy is also demonstrated by the fact that Denmark and Sweden have some of the highest levels of on-site training and education, at about 30% of the active population, against the EU average of only ten per cent.12
Regarding the effects of employment incentives in Sweden, studies point to the positive impact that these programmes have had on the participating individuals.13 This emphasis on employment incentives in Sweden developed in the 1980s when more traditional ALMPs were replaced by demand-side interventions, including job creation schemes and incentives. As argued in Bengtsson, this shift reflects a stronger pro-market employment orientation, which had the primary goal of actively getting people to work, without leaving the burden of searching for jobs and skill upgrading to the job seeker, as usually assumed in training schemes. More recently, Sweden adopted employment incentive schemes combined with training programmes, such as the 2007 Entry Recruitment Incentive, which combined the granting of employment with Swedish language studies for immigrants. Another rather extensive programme is the Job Development Programme, which is directed towards youth and entails in-depth assessment and job-seeking/job-coaching activities, training, and supported employment for a maximum of two years in “artificial jobs”.14
Evidence of job reallocation
What can be said regarding the implications of the labour market flexibility in these economies and their policies on jobs? The Danish labour market is known for being dynamic and for having a high rate of employee turnover compared to other countries. In 2012 there were around 714,000 job vacancies, which translates to a turnover of around 19%.15 The low protection employees have against being fired in Denmark may certainly have its downsides, but such flexibility ensures that even during a severe crisis the unemployed can still find many openings. In fact, evidence shows that, despite the crisis, there are many newly employed people in Denmark, as about 23% of the employed have been employed for less than a year (see Figure 6). Also, if one looks how long it takes the unemployed to return to the workforce, both Denmark and Sweden are performing rather well. Of those that were unemployed in the first quarter of 2015 in Denmark, almost 38.5% had found a job by the next quarter, and in Sweden this figure was almost 30%, placing both countries at the top of the EU (see Figure 7).
Newly employed among the employed
Source: Eurostat Labour Force Survey.
Share of unemployed that find work next quarter
Source: Eurostat Labour Force Survey.
In addition to this evidence, we report data on job creation, calculated following Davis and Haltiwanger as the sum of all employment gains in a given year divided by the average employment in that year and the previous year.16 We also calculate the flows into and out of unemployment for the same time period. Figures 8 and 9 show that job creation was higher and more frequent in Sweden than in Denmark, while the flows into and out of unemployment in the two countries exhibit similar patterns, although in the crisis years Denmark’s unemployment increased significantly more so.
Job creation, year over year
Changes to unemployment, year over year
Evidence of labour market matching
A major reason behind the adoption of ALMPs is that they facilitate matching. To calculate matching, we use the Diamond-Mortesen-Pissarides definition, according to which the matching process between job seekers and employers is a stable but dynamic equilibrium relationship between the search processes of unemployed workers and job vacancies.17 On these grounds, the degree of matching can be expressed by a curve whose points reflect the same combination of vacancies and unemployment, a curve more commonly defined as the Beveridge curve. Through the analysis of Beveridge curve shifts and movements, the literature has identified several factors that can reduce frictional unemployment by improving matching. Included among these are active labour market policies. Have these policies played a role in the matching process for Denmark and Sweden? Is the mismatch (or frictional unemployment) in these countries lower than in other countries? Has it decreased since the crisis?
To this end, we look at possible mismatches. As vacancies data are not available for Denmark, we use European Commission data on labour shortages in the manufacturing sector, following Bonthius et al.18 Figure 10 shows the labour shortages as obtained from the EC business survey. A value closer to zero implies fewer vacancies, as it corresponds to a lower degree of perceived labour shortages. This is evident in the crisis years, when the variable drops to zero in Denmark, Sweden and the euro area. After the crisis, the indicators for Sweden and the euro area increase, suggesting more vacancies are being posted than in Denmark, yet after 2013 both Sweden and Denmark display low levels of labour shortages. A low level of vacancies in conjunction with the relatively low level of unemployment suggests a high level of matching between job seekers and employers in the two countries. Conversely, the coexistence of higher unemployment and more vacancies in the euro area reflects a high degree of mismatch. It is interesting to note, however, that with the exception of 2006-07, the level of vacancies (or labour shortages) in Denmark was less volatile than in Sweden. Such a trend may be partly explained by the larger emphasis Denmark places on training programmes, which implies a better skilled and equipped labour force.
Labour shortages in the manufacturing sector
Source: European Commission Business Survey.
What conclusions can be drawn from the evidence?
Both countries are performing well in terms of labour market indicators. The different degrees of labour market flexibility of the Danish and Swedish models have led to slightly different reactions to the crisis. We have shown that the Danish model leads to higher employment turnover and grants more chances for a job seeker to find a job in a relatively short period of time. At the same time, the Swedish model resulted in greater resilience during the crisis, as fewer jobs were destroyed and more jobs were created. We have also shown that the large share of public spending devoted to ALMPs has led to better matching between job seekers and employers. Our comparison of training programmes in Denmark with employment incentives in Sweden suggests that this difference in focus may be behind the higher volatility of labour shortages in Sweden.
Clearly, labour market policies and institutions are only a piece of the puzzle that explains the success of these economies. The Nordic model as such has much more to offer. Other key features include centralised wage setting as opposed to market-based wage setting, high rates of labour organisations and trade unions, high collective agreement coverage, a tax and transfer policy mix that strongly reduces inequalities, high investment in education, and a political economy that puts employment first. For a comprehensive analysis of employment performance, these factors and features would also need to be taken into account. Looking at the medium-term prospects for these economies, the question emerges as to how they can preserve their fair distribution, balanced growth and full employment objectives, as well as the political support for such programmes. This question becomes even more pressing in the face of new challenges associated with increased global competition, the ecological and digital transitions, migration, and ageing populations.19
- 1 C. Pissarides: Equilibrium in the Labour Market with Search Frictions, Prize Lecture, 8 December 2010.
- 2 See A. Bassanini, R. Duval: Employment Patterns in OECD Countries: Reassessing the Role of Policies and Institutions, OECD Economics Department Working Papers, No. 486, 2009.
- 3 See L.E. Bernal-Verdugo, D. Furceri, D. Guillaume: Labor Market Flexibility and Unemployment: New Empirical Evidence of Static and Dynamic Effects, IMF Working Paper 12/64, 2012; E. Bova, J.T. Jalles, C. Kolerus: Shifting the Beveridge Curve: What Affects Labor Market Matching?, IMF Working Paper, forthcoming.
- 4 See J.P. Martin: Activation and Active Labour Market Policies in OECD Countries: Stylized Facts and Evidence on their Effectiveness, IZA Policy Paper No. 84, 2014.
- 5 See D. Card, J. Kluve, A. Weber: Active Labor Market Policy Evaluations: A Meta-Analysis, NBER Working Paper No. 16173, 2010.
- 6 Rockwool Foundation Research Unit: Aktivering kan reducere uligheden i samfundet, May 2013.
- 7 A.J.G. Brown, J. Koettl: Active Labour Market Programs: Employment Gain or Fiscal Drain?, Kiel Working Papers No. 1785, 2012; C. Koch, P.K. Madsen, V. Jensen: Veje til job – en arbejdsmarkedsindsats med mening, Danish Ministry of Employment, 2014.
- 8 See E. Bjørsted, S. Dahl: Why Are the Nordic Countries Doing so Well? The Case of Denmark and Sweden, FEPS Policy Brief, 11 January 2016.
- 9 M. Bengtsson: Transformation of Labour Market Policies in the Nordic Countries: Towards a regime shift in Sweden and Denmark?, University of Gothenburg, mimeo, 2012.
- 10 World Economic Forum: Global Competitiveness Report 2015-16.
- 11 Economic Freedom of the World data 2016, OECD and EC data.
- 12 Eurostat.
- 13 IMF, 2010: Sweden, Article IV Consultation Staff Report.
- 14 M. Bengtsson, op. cit.
- 15 C. Koch et al. op. cit.
- 16 S.J. Davis, J. Haltiwanger: Gross Job Creation, Gross Job Destruction and Employment Reallocation, mimeo, 1991; S.J. Davis, J. Haltiwanger: Gross Job Creation and Destruction: Microeconomic Evidence and Macroeconomic Implications, Bureau of Census, CES Paper 90-100, 1990.
- 17 D.T. Mortesen, C.A. Pissarides: Job Creation and Job Destruction in the Theory of Unemployment, in: The Review of Economic Studies, Vol. 61, No. 3, 2011, pp. 397-415.
- 18 B. Bonthuis, V. Jarvis, J. Vanhala: What’s Going on Behind the Euro Area Beveridge Curve?, ECB Working Paper No. 1586, September 2013.
- 19 For an analysis of the Nordic Model and future challenges, see J.E. Dølvik, T. Fløtten, J.M. Hippe, B. Jordfald: The Nordic Model towards 2030. A new Chapter?, Fafo report, 2014.