Active Labour Market Policies and Youth Employment:
A Macroeconomic Analysis of the Youth Guarantee Program in Europe
By Massaya Sirimatayanant
Economics and Governance, Public Administration (2554402)
Thesis Supervisor: Prof.dr. P.W.C. Koning Secondary Reader: Dr. J. Been
A Thesis
Presented to the Faculty of Governance and Global Affairs Leiden University
In partial fulfillment of the Requirements For the Degree of Master of Science
Table of Content
1. Introduction 1.1 Background 1.2 Research Question 1.2.1 Academic Relevance 1.2.2 Social Relevance 3 6 7 8 2. Literature Review2.1 Theoretical Overview of Youth Employment Problems
2.2 Consequences of Youth Unemployment and Justification for Action 2.3 The Role of Active Labour Market Policies
2.4 A Literature Review of Macroeconomic Studies 2.5 The EU Youth Guarantee Scheme
2.6 Hypotheses 9 10 11 17 21 24 3. Research Design
3.1 Sample Selection and Construction of Data 3.2 Method of Analysis
25 36
4. Estimation Results and Analysis 39
5. Conclusion 47
6. References 52
1. Introduction 1.1 Background
Despite the fact that young cohorts have become smaller in number and better educated than their older counterparts, the average youth unemployment rate for most OECD countries over the last decades have persistently been double the overall adult unemployment rate (Kluve, 2014a). This is because young individuals have lower labour market attachment than adults. Unlike their adult counterparts, young individuals face a higher risk of unemployment when entering the labour market and have the tendency to switch between states of employment, unemployment, and inactivity due to the nature of the school-to-work transition process (Quintini, Martin & Martin, 2007). Young individuals are considered an at-risk population because of their initial lower labour market experience, making them costly investments for firms to hire but less costly for termination. Therefore, young individuals also tend to be more sensitive to cyclical volatility in demand for labour than adults (Banerji et al., 2015; Choudhry, Marelli, & Signorelli, 2012; Verick, 2011;). This was demonstrated following the 2008 financial crisis when the average youth unemployment rate in the EU-28 countries shot up by about five percentage points between 2008 and 2009, peaking at almost 24 percent in 2013 while the prime-age adult unemployment rate experienced a
less steep incline from about six percent to 10 percent during the same time period1.
High youth unemployment is a concern because the incidence of “scarring effects” or negative impact on earnings, employment, or other labour market outcomes that are a consequence of prolonged unemployment spells (McQuaid, 2015). Although all demographics can experience scarring effects, young workers are disproportionately affected by this effect because it occurs early on in their life, causing the negative impacts to multiply over time (Doiron & Gørgens, 2008; Gregg & Tominey, 2005). Youth unemployment therefore has the potential to impact a person’s lifetime economic wellbeing, which may affect their socio-economic positions as adults and have negative consequences on their social and political participation, as well as contribute to economic inequality in the society at large in the long run as well (Scott & Acock, 1979).
1 Average youth unemployment rate is derived for the age group 15-24 years old, while the prime-age adult unemployment rate is derived for the age group 25-54 (OECD, 2019).
For these reasons, the significance of engaging youth population has become a priority on policy agendas in countries around the world. Active labour market programs (ALMPs) have become an important tool used by governments to smooth the school-to-work transition and prevent extended unemployment spells. Unlike passive labour market policies that are designed to partially compensate for loss income by providing temporary financial support, ALMPs are used to expand the workforce actively and improve employability (Nie & Struby, 2011). Activation measures can either take on a labour force attachment approach, aimed at moving unemployed workers on welfare into jobs quickly, or human capital development approach, which exhibit results in longer time horizons due to the fact that the approach seeks to improve basic and job-related skills of welfare recipients (Hotz et al., 2006). Generally, the most common ALMPs can be divided into four main categories: training and skill development programs, employment services (such as job-search assistance programs), subsidized employment programs (both private sector wage subsidy programs and public employment programs), and entrepreneurship promotion programs (Kluve et al., 2019). While evaluation of specific programs finds mixed results within the current literature, macro-econometric evidence does reveal that ALMP intensity can help curb unemployment (Martin, 2015). Unfortunately, young individuals were found to benefit less from ALMPs than their adult counterparts. In a meta-analysis conducted by Card, Kluve, and Weber (2010), only about 20 percent of the youth programs reviewed were found to have positive impacts.
Although the youth population is often cited as the most difficult demographic to assist (ibid.), empirical results from a number of systematic reviews still indicate positive ALMP treatment effects that are statistically different from zero on youth labour market outcomes (Kluve et al., 2019). Authors of meta-analysis on youth programs interpret the low proportion of programs with positive effect sizes as relating to differences in design and implementation factors of each program as well as the characteristics of the country and population of beneficiaries, rather than the effectiveness of ALMP programs themselves (Kluve et al., 2019). Furthermore, since it is found that scarring effects on young individuals are non-permanent and can be mitigated by preventing early-career unemployment (Ellwood, 1982; Möller & Umkehrer, 2015), there is empirical support suggesting that investing in young individuals through ALMPs do pay off with positive impacts on employment and earnings outcomes if programs are carefully designed and tailored to the population and the context in which it is being used.
According to the International Labour Organization (ILO), the financial crisis in 2008 had reversed the gradual declining trend in global youth unemployment rates observed between 2002 and 2007 (ILO, 2017). Recognizing the urgency in intensifying assistance to youth population who have been the hit the hardest during the crisis, European Union members made a political commitment to the Youth Guarantee (YG) to ensure that “all young people under the age of 25 years receive a good quality offer of employment, continued education, apprenticeship, or traineeship within a period of four months of becoming unemployed or leaving formal education” in 2013 (European Commission, n.d.). Through the YG scheme, the EU assisted member states in developing national YG implementation plans while the Commission monitors implementation. Implementation of the scheme took place from 2014 onwards with the European Social Fund (ESF) providing additional funding that are matched by member states to assist with the process. Policies that are supported by the EU under this scheme include a number of ALMPs that are aimed at providing opportunities for low-skilled young individuals to train, and address skill mismatches through work-based learning and apprenticeships or training vouchers. Policies extend to providing wage and recruitment subsidies to encourage employers to recruit young people in apprenticeships or job placements, as well as supporting small and medium enterprises and start-ups through entrepreneurship programs (European Commission, 2014). Additionally, the EU has established the Youth Employment Initiative (YEI) to exclusively focus on the most vulnerable groups including Youth not in employment, education or training (NEET), long-term unemployed, and those not registered as job-seekers; YEI is used as the main financial tool by qualifying member states to fund implementation of YG schemes for the aforementioned target group in regions where youth unemployment rates are higher than 25 percent. As such, only 20 member states are qualified to receive the additional YEI allocation (ibid.).
The average youth and prime-age adult unemployment rates in EU-28 countries have fallen to 14.36 percent and 5.79 percent, respectively, by 2019; an all-time low since the start of the century (OECD, 2019). While it is inviting to associate the fall in youth unemployment rates to recent changes in youth targeted ALMP schemes in the EU, improvements in the unemployment rate for both younger and older cohort is to be expected as economies recover following the financial crisis. Furthermore, the Global Employment Trends for Youth 2020 still observes concerning declines in labour force participation of youth, which reflects an increasing enrolment
in education but also the persistence of youth NEET challenges (ILO, 2020). As this pattern can also be observed in the EU-28 countries, it is beneficial to investigate the impact of the YG scheme and the incidence of increased ALMP-intensity, through the YEI, on different labour market outcome indicators for youth in order to disaggregate whether improvement in youth labour market performances can be attributable to the EU’s intensified youth targeted commitment or not.
1.2 Research Question
Due to the difficulty in isolating impacts of particular programs, and the trend to integrate ALMP service provision, the goal of this research is to investigate, from an aggregate level, the following research questions:
1. Does the use of youth targeted-ALMP have a positive impact on labour market outcomes
of young individuals?
2. Does the incidence of additional funding improve the labour market outcomes of young
individuals?
To answer the aforementioned questions, this study will conduct a cross-country level
analysis drawing on panel data for EU-282 countries between 2000 and 2019; during which the
YG scheme began implementation in 2014, and additional YEI funding is allocated for the duration of YG implementation period of 2014-2020 (intended to be frontloaded and committed between 2014 and 2015). Labour force statistics – namely, employment, labour force participation, unemployment and NEET rates of 5-year age interval categories, are drawn from the OECD Statistics and Eurostat Databases. The statistics are complemented with information on the incidence of additional YEI funding and the cut-off age for YG eligibility in each member states. Due to the fact that both the OECD and Eurostat do not observe ALMP spending stratified by age groups, the YEI as the only additional source of funding that was established in 2014 and the main
EU funding program created under the YG scheme3, will serve as the treatment within this analysis.
2 This study will include the United Kingdom as part of the EU-28 countries for the analyses as labour force statistics are derived up until 2019.
3 Although the main EU funding program for the YG scheme is the YEI, the ESF (to be succeeded by the ESF plus in 2020), a multiannual financial framework renewed every seven years, is the largest source of funding for member states to use to promote employment and social inclusion, which also includes youth employment policies
However, because the YEI is reimbursed retroactively when programs are approved, and consequently annual YEI expenditure data for each member state is unavailable, this study will be taking on an intention-to-treat analysis. A difference-in-difference design will be used to compare the labour market position of young individuals that are intended to receive YG/YEI treatment to the labour market position of their peers who are right above the cut-off age for the YG/YEI treatment. This will allow for the study of the differential effect of the treatment on a treatment group and a control group that are more similar in characteristic and age profiles.
1.2.1 Academic Relevance
The majority of studies on the relationship between ALMPs and youth have either been microeconomic evaluations focusing on determining which program or combination of programs works best, or a systemic quantitative review (meta-analysis) summarizing the multitude of impact evaluation studies produced across countries (Kluve, 2014b). However, Kluve et al. (2019) finds that contextual and design factors of ALMPs are more significant contributions to the success of ALMPs rather than the type of intervention. Therefore, although microeconomic evaluation literature is important for quantifying which individual ALMPs work, it cannot be scaled to quantify macroeconomic effects of ALMP intensity (Martin, 2015). Additionally, most macroeconomic studies on the impact of ALMPs have focused on overall unemployment rates (Hur, 2019; Boone & van Ours, 2004; Estevão, 2003; Blanchard & Wolfers, 2000; Nickell & Layard, 1999; Elmeskov et al., 1998) with few that focus on youth employment outcomes (Scarpetta, 1996; Banerji et al., 2015). In instances where determinants of youth unemployment were studied, results found were mixed. For example, in Scarpetta (1996) other variables such as employment protection rules and union density were found to exert more significant effect on increasing youth unemployment, while ALMP was found to be insignificant. This study will therefore contribute to the literature by conducting a macroeconomic analysis studying the impact of a youth targeted ALMP, using the introduction of the YG scheme as a proxy, on youth labour market performances.
allocated to promoting employment for youth specifically and because ESF allocations have been present even prior to the YG scheme, this analysis will only focus on the impact of additional funding through the YEI on youth labour market outcomes.
1.2.2 Social Relevance
Due to the COVID-19 pandemic, countries around the world are implementing the necessary protection measures that have severe impact on economic activity. According to the IMF (2020), the global economy is projected to experience a three percent contraction in 2020; a level that is more severe than the 2008 financial crisis. As it is found that young individuals experience higher unemployment probability during recessions than adults (Scarpetta, Sonnet & Manfredi, 2010), concerns regarding whether young cohorts entering the labour market under such economic crises could represent a “scarred generation” suffering a permanent decline in their financial well-being has catapulted the youth employment crisis as a priority for policy makers around the world (IZA, n.d.). An investigation into the effectiveness of the YG scheme is therefore relevant to help inform policy makers of the importance of investing in ALMPs to minimizing volatility in youth unemployment and create resilience for this upcoming recession.
The remainder of this research paper will be structured as follows: section two presents a literature review composing of a theoretical discussion of youth employment problems and ALMPs, an introduction into macroeconomic studies in the field, a short summary of the YG scheme along with the expected causal mechanism, and a statement of the hypotheses; section three presents the design of this research including a description of the data and sample selection, operationalisation of variables, and the model specifications; section four presents and discusses the estimation results from the model developed under section three; and, section five will conclude and suggest ideas for future work.
2. Literature Review
2.1 Theoretical Overview of Youth Employment Problems
Young individuals, unlike adults, experience higher rates of job turnover leading to more frequent period of unemployment, and in turn, delays their process of becoming established within the labour market. Two economic theories, which can be used to explain the challenges young individuals face in the labour market, are human capital theory (Becker, 1993) and job signaling theory (Spence, 1973).
Within labour economics, it is helpful to view the set of marketable skills a worker has as a form of capital in which workers can invest in. Becker (1993) asserts that human capital corresponds to a stock of knowledge or characteristics that a worker has, either innate or acquired, that contributes to their productivity. Increases in human capital, through investments in training, education and even work experience, therefore increase a worker’s productivity and subsequently employability. Young individuals with lower initial labour market experience than adults have lower employability, and therefore are at a disadvantage when competing for work. In order to understand how human capital impacts the mismatch between supply of and demand for youth labour, it must be complemented with insights from the job signaling theory.
According to Spence (1973), one must view “hiring as an investment under uncertainty” (p.356). He explains that in most job markets, employers are unsure of the productive capabilities of an individual being hired, both at the time of hiring and immediately after hiring. This is because it takes time and requires certain skill or job-specific training to be undertaken by the newly hired individual for employers to become aware of their full productive capability. Since these capabilities are not known beforehand, hiring decisions are made under uncertainty. Under conditions of information asymmetries that exist within the labour market, observable attributes resulting from human capital investments can act as signals for employers to make a “conditional probability assessment over (an individuals’) productive capacity” (ibid., p.357). Since previous experience in the labour market act as an important signal for productivity, young individuals with little to no previous labour market experience are seen as an expensive investment when compared to adults. This would also explain why youth unemployment rates are more sensitive to business
cycles than adult unemployment rates (Dietrich & Möller, 2016; Banerji et al., 2015; Verick, 2011). For example, in times of severe recessions, young individuals are also the first group to be laid off due to the “last in first out” principle which favours more experienced workers (ibid.).
Structural challenges relating to the performance of education and training systems, including vocation education and training (VET) also exacerbates the youth labour supply-demand mismatch. Poor education and training systems, in terms of both access and quality, relates to both human capital and signaling theories. The implication is that young individuals are not receiving the appropriate education and training that will equip them with the transferable skills required in the workplace. Stigmatism with poor quality of VET programs also sends the wrong signals to employers with regards to the quality of youth labour exiting such programs.
In addition to these structural disadvantages, young individuals as new labour market entrants also experience age-specific patterns in labor market behavior: young workers often engage in “job-shopping” behaviors to learn about their abilities and preferences. Job shopping is defined by Johnson (1978) as “the search for a suitable job when workers cannot predict perfectly either their performance in or their liking for a particular job” (p. 261). Young individuals with little to no labour market experience often engage in this period of experimentation with jobs more so than adults. Topel and Ward (1992) found that a typical worker would typically hold seven jobs, making up two thirds of their career total, in just the first ten years of entering the labour market. While job shopping plays a critical role in allowing young individuals to move upwards in the career or wage ladder, and towards building a stable employment relation typical of mature careers (ibid.), it becomes a concern when frequent job-changes continue to promote further job changes which then impacts their observable attributes and their future employability.
2.2 Consequences of Youth Unemployment and Justification for Action
Unemployment, particularly when occurring early on in an individual’s career, have been found to have negative effects on future labour market performance. “Scaring” or “state-dependence” denotes a wide range of life-long negative impact on earnings, employment, and even health and well-being that are a consequence of prolonged and repeated unemployment spells (McQuaid, 2015). Explanations for how scaring effects impact future labour market performance
also lies within both human capital and signaling theories. Unemployment spells result in forgone work experience, and even worse, skill deterioration in the instances of long spells of inactivity, while the frequency of unemployment signals low employability to future employers. Although early empirical literature on state-dependence has found mixed results, mainly due to various empirical obstacles that make it difficult to identify the causal mechanism of state dependence (Heckman & Borjas, 1980; Ellwood, 1982), more recent studies have found evidence of such effects (see recent examples: Skans, 2011;Junankar & Wood, 2016; Schmillen & Umkehrer, 2017).
Particularly, Edin and Gustavsson (2008) found statistically strong evidence of a negative relationship between work interruption and skills. This occurrence is concerning for the youth population as it implies that the value of skills acquired by young individuals during their time in education may also depreciate over time if it is not used. It is important to note however, that negative scarring effects are persistent but not permanent (Skans, 2011). A German study conducted by Möller and Umkehrer (2015) agues just that. While early-career unemployment has substantial negative effects on earnings accumulated later in life, workers with high earnings potential which can be obtained through human capital development were found to be able to offset the consequences of early career unemployment to a large extent. What can be learnt from these studies is that there is justification for policy makers to concentrate efforts to help young individuals remain attached to the labour market from a young age. This starts from smoothing out the transition from education to stable employment to activating those who are disengaged from the labour market, with the ultimate goal being: to prevent early-career unemployment spells and accumulate long-lasting beneficial effects on both employment prospect and future earnings.
2.3 The Role of Active Labour Market Policies
A variety of measures can be taken to increase both the supply of and demand for youth employment. Blanchflower and Freeman (2000) identifies three major kinds of programs that can be used to tackle youth unemployment. Supply-side programs can be disaggregated into programs that try to ease the school-to-work transition before young individuals encounter difficulties in the labour market, and “second-chance” programs that address the skill mismatch for young individuals facing difficulties in the labour market. Meanwhile, demand side programs may be
aimed at raising youth wages, such as minimum wage programs, or linking young individuals with certain employment opportunities, such as public employment programs (ibid., p. 9). ALMPs cover a range of the programs types previously mentioned.
Unlike other labour market policies, ALMPs require active participation in programs that intensifies labour market integration. The overall objective of ALMPs is to improve human capital of participants which signals their employability, while also increasing the efficiency of the job-matching process through addressing information asymmetries that exists between young workers and future employers. In terms of serving young individuals, ALMPs are therefore used to integrate unemployed youth into the labour market, stabilize their career entry, and promote vocational training or continued education offers as an intermediate step to labour market entry (Caliendo & Schmidl, 2016). The motivation behind how ALMP works is the theory of change which hypothesizes that participation in ALMPs can improve the employment and earning outcomes for young individuals by addressing the various constraints that affect their access to wage or self-employment through targeted interventions (Kluve et al., 2019). Generally, ALMPs can be systematically clustered into the following groups: training and skills development programs, employment services, subsidized employment programs, and entrepreneurship programs.
Training and skills development programs aim to increase human capital and attenuate skill mismatch, therefore aligning the skill-level of unemployed youth to labour market demands. Since education and skills are considered the core factor in determining young individuals’ opportunities in the labour market (Biavaschi et al., 2012), it is not unexpected that these programs were found to be the most commonly used ALMPs (Kluve et al., 2017; Caliendo & Schmidl, 2016). Examples of programs that fall under this category of ALMP include trade- or job-specific technical skills training, literacy programs, as well as programs that improve non-technical skills such as life skills or soft skills beneficial for young jobseekers. Consequently, training programs can be very heterogenous in nature ranging from classroom-based training to on-the-job-training programs, with varying lengths ranging from weeks to a year.
Training programs can provide young individuals with a number of benefits. Training with firms can give young individuals with some level of work experience, which sends
productivity signals to employers and increases their future employability (Caliendo & Schmidl, 2016). Furthermore, it provides them with a foot-in-the-door which is particularly relevant when young individuals lack prior labour market experience. In the short run, training and skills development programs were found to be ineffective due to “locked-in effects” as individuals engaging in training programs were found to have lower job finding rates than non-participants (Kluve, 2010; Van Ours, 2004). As such, training programs were often combined with other interventions such as general employment services to provide further activation for young individuals. However, the overall emphasis on human capital accumulation of such programs were still found to have substantial long-term positive effects (Strandh & Nordlund, 2008).
In a meta-analysis conducted by Card, Kluve & Weber (2010) which drew on 199
program impacts4 from 97 studies conducted between 1996 and 2007, classroom and on-the-job
training programs were found to be ineffective in the short run but have more positive relative impacts after two years. Unfortunately, when comparing across different participant groups, the authors found that programs targeting youth were less likely to yield positive impacts than untargeted programs (ibid., p. 453). This is, in parts, due to “crowding out” or “displacement” effects which can render targeted ALMPs to be ineffective. This can occur when young individuals who engage in training programs displace other non-participants with lower skill level. As a result, overall unemployment rate does not change while the burden on government expenditure increases (in terms of unemployed non-participants continuing to receive unemployment benefits and participants receiving assistance through government sponsored ALMPs) (OECD, 2004).
However, results from a more recent systematic review5 focusing on ALMPs impact on youth
employment and earnings outcome, found that interventions with a focus on skills training do improve employment outcomes and raised earnings, on average (Kluve et al., 2011).
4 Authors’ main analysis focuses on the sign and significance of program estimates, followed by conducting specification tests to evaluate the validity of their main analyses (Card, Kluve, & Webber, 2010, p. 462-467). 5 Note different meta-analysis and systematic reviews use different inference method to estimate program impacts. Kluve et al. (2017) draw conclusions from effect size computations, which involves a combination of two measures: a standardized mean difference (SMD) measuring effect size of treatment effects reported in each studies
incorporated in the meta-analysis, and a binary variable that turns on as one if the treatment effect is positive and significantly significant. The authors’ claim that the SMD effect sizes can capture the relative magnitude of treatment effect in a “dimensionless way”, which allows them to compare effect sizes across various outcomes and studies that are incorporated in the meta-analysis (p. 18; p.61-64).
Next, general employment services cover a range of activation measures including counselling, job search assistance, and mentoring services. Based on the acknowledgement that there are information asymmetries within the labour market, general employment services such as job search assistance aim to improve the efficiency of job-matching process by providing information and support to both sides of the labour market. Young jobseekers are informed about suitable opportunities, while employers are informed of potential unemployed youths (ibid.). The goal of PES in providing general employment services is therefore twofold: to reduce the transaction cost and uncertainty employers face when recruiting young individuals, while also providing support for young individuals to improve their job-seeking skills. These services are intended for young individuals who face difficulty in signaling their skills and credentials to employers, or lack the network or knowledge to effectively search for vacancies and connect with potential employers (ibid.). Combining job search programs with a tight monitoring scheme can help achieve fast activation of young individuals early on in the unemployment spell, which in turn can help avoid long-term unemployment spells and minimize the risk of young individuals getting discouraged, withdrawing from the labour market and entering inactivity (Caliendo & Schmidl, 2016). Fast activation through such programs also ensure that young individuals can avoid skill deterioration as a result of foregone work experience.
While these general employment services have lower implementation costs, it has been found that they tend to be less effective in the long run. Card, Kluve and Weber (2010) concluded in their meta-analysis study that job search assistance may outperform training programs in the short run, but over longer horizons, the gains of human capital development programs are larger. Furthermore, in instances where such programs exhibit positive long-run effects, it is still debatable whether these gains are transitory or displacement effects because benefits for those enrolled can come at the expense of equally eligible but untreated unemployed workers. This was shown in Crépon et al. (2013), wherein a randomized experiment was designed to evaluate the direct and indirect or displacement impacts of job placement assistance on labour market outcomes of young educated jobseekers in France. The authors found that gains for treated youth come partly at the expense of equally eligible workers, particularly in labour markets where they have to compete with other educated workers, and in weak labour markets. They concluded that such programs have very little net benefits. This conclusion was also echoed in the systematic review
conducted by Kluve et al. (2017; 2019), which found the effects on youth employment to be small and not statistically significant for employment services.
The third type of ALMP to be discussed is subsidized employment programs. These programs can be divided into either private sector incentive programs such as a wage subsidy, or labour intensive public employment programs. The goal of subsidized employment programs is to address issues of insufficient labour demand. Particularly with regards to youth employment problems, such programs stimulate youth labour demand, as well as addresses the barriers to initial labour market entry resulting from the lack of work or job-specific experience that young individuals face (Caliendo & Schmidl, 2016). Akin to public employment services, these programs are intended to fix the youth employment problem in the short run by ensuring that young individuals who are incapable of finding a job on the regular labour market remain integrated and connected to economic and social life, with the addition of working towards having a longer term labour market impact through the provision of valuable work experience (Kluve et al., 2017).
However, positive effect of subsidized employment programs in the long-run is only attainable if experience from these programs can improve longer-term employment prospects for participants. A study conducted by Strandh and Nordlund (2008) sought to compare the effects of training and subsidized employment programs, both of which represent different types of human capital investment. Using longitudinal data of Swedish workers who experienced long-term unemployment prior to participating in the respective ALMPs, the authors found that the effects of the two forms of ALMPs were differently distributed over time. Subsidized employment programs were found to have immediate positive effects that decreased relatively quickly, while training programs had a longer-term effect (Kluve et al., 2019).
Alike training programs, participants of subsidized employment also tend to experience locked-in effects, whereby engagement in ALMPs were found to reduce participant’s job-search efforts rather than increase them (Van Ours, 2004). Coupled with the fact that these programs were
expensive to implement and run the risk of inducing a deadweight loss6, their use by policy makers
is therefore only typically targeted towards activating the most disadvantaged individuals in
6 Deadweight loss resulting from the creation of more jobs than required, if implemented in a more buoyant labour market.
societies. In particular, public sector employment programs were often only implemented in times of poor macroeconomic and weak labour market conditions (Kluve, 2014). Accordingly, when implemented, subsidized employment programs were always coupled with other intervention mechanisms to ensure proper integration of workers into the labour market after exiting the program. For example, wage subsides were granted to firms that agree to provide additional training for subsidized employees, and public employment programs were paired with exit strategies such as skills training or entrepreneurship (Kluve et al., 2017).
Lastly, entrepreneurship programs have become an important ALMP found to contribute to both employment and earnings outcomes of young individuals (ibid.). Entrepreneurs are income providers and job creators that benefit both booming economies by challenging existing enterprises to innovate and compete, and economies that are experiencing slow job growth by boosting labour demand. The relevance of these program types are hence twofold: on the individual level, such programs can help vulnerable unemployed individuals obtain gainful employment and become independent of welfare benefits, while on the societal level, it can increase levels of employment through a multiplier effect, promote entrepreneurship and innovation, and create new working places (Wolff & Nivorozhkin, 2012).
Entrepreneurship promotion interventions can be disaggregated into two categories: programs that provide or facilitate access to credit and finances, and programs that offer business advisory services and mentoring, business skills training, and access to markets and value chains (Kluve et al., 2019). Entrepreneurship programs are more popular in developing countries. It has been found that increasing self-employment among the labour force is an important anti-poverty strategy in countries where there is limited scope for formal wage employment (Gindling & Newhouse, 2012). However, it has also been found to be highly effective in developed countries too. Wolff and Nivorozhkin (2012) studied the effect of participation in a new business start-up scheme for needy unemployed people in Germany. They found positive and sustained effects of the program on reducing the proportion of registered unemployed and benefits recipients among treated participants.
2.4 A Literature Review of Macroeconomic Studies
It is important to note that there are two main types of evaluation studies on ALMPs within the literature. The first type, which uses micro data to measure the impact of program participation on individual’s employment and earnings, have been discussed briefly throughout the previous section. The second type, more relevant in terms of methodology to this study, uses aggregate data to measure net effects of programs or combination of programs on aggregate employment and unemployment outcomes (Martin & Grubb, 2002). While micro studies are beneficial in terms of identifying effective programs, they face limitations in terms of selection
bias7 and provide no more than estimations of partial-equilibrium effects (Boone & Van Ours,
2004). The latter is a result of the spill-over effect of ALMP exposure. According to Kluve et al. (2017), “exposure to ALMPs is expected to create a spillover effect among non-participants, as well as, general equilibrium effects throughout the economy” (p.27). Spillovers onto non-participants can either positively or negatively affect overall employment outcomes. For example, analysis on subsidized employment programs have found large variations in their results; this is
due, in part, to the fact that wage subsidy programs induce displacement8 and windfall9 effects that
is not always incorporated or properly captured in micro-level analysis (Almeida, Orr & Robalino, 2014). As such, while microeconomic analysis may find instances where wage subsidy programs exhibit positive impact on labour market outcomes (Strandh & Nordlund, 2008), macroeconomic analyses find subsidizing employment to be costly and inefficient from the perspective of aggregate labour market outcomes (Boon & van Ours, 2004).
Macroeconomic analyses still face certain drawbacks such as the inability to disaggregate between the impact of various types of programs on labour market outcomes (leading to omitted
variable bias), and the possibility of simultaneity bias10 (ibid.). Regardless, the benefit of
7 Microeconomic evaluations are faced with selection bias because participants self-select themselves into ALMPs, and therefore, their outcome as a result of receiving ALMP treatment is being compared to a counterfactual state of non-participation. Note: Experimental evaluations have been able to address such bias to a certain extent but the majority of studies on ALMPs rely on non-experimental or observational data (Caliendo, 2006, p.3).
8 Explained in the previous section on ALMPs
9 When subsidies are paid out to workers who would have been hired regardless of whether they are being subsidized or not.
10 Or also referred to as the endogeneity problem/reverse causality, occurs when explanatory variables on the right-hand side of the causal inference model and the outcome variables on the left-right-hand side of the same model correlate with one another. For example, rather than ALMPs impacting unemployment, high unemployment rate impacts intake into ALMP programs and therefore leads to higher ALMP spending.
conducting a macroeconomic analysis is the ability to capture the net effect on labour market outcomes of ALMP exposure. Unfortunately, there are a limited number of empirical macroeconomic studies that use cross-country time series data to map out the effects of ALMPs on labour market outcomes of young individuals. A study conducted by Scarpetta (1996) is one such exception. Using annual data from 17 countries over the 1983 – 1993 period, Scarpetta uses the explanatory variable ALMPU or government expenditure on active labour market policy per unemployed person relative to GDP per capita to study the impact of the ALMP spending on unemployment rates and employment-to-population ratios. ALMPU has become a popular operationalisation of ALMP spending in macroeconomic studies because it adjusts for differences across countries in both unemployment rate and the size of the economy. Scarpetta found that ALMPs have a negative impact on overall unemployment rate and a positive impact on employment-to-population ratio; however, the small negative impact on youth unemployment rates were found to be insignificant. Instead, he found unemployment benefit, union density and
employment protection legislations11 to have a stronger effect: exerting a positive and significant
impact on youth unemployment rate.
In contrast to Scarpetta’s result, Banerji et al.’s (2015) investigation into the role of labour market institutions on youth unemployment did find that higher spending on ALMPs, especially for training, were associated with lower unemployment. In particular, the authors found that an additional thousand euros per unemployed increase in ALMP spending is associated with lower youth unemployment rates, by 0.33 percentage points (ibid., p.20). Alike Scarpetta’s study, Banerji et al.’s multivariate analysis approach also considers the effect of several labour market features simultaneously, rather than investigating the role of individual labour market institutions. Through the examination of the cyclical and structural explanatory factors behind youth unemployment dynamics during the financial crises in 22 advanced European countries between 1980 and 2012,
the authors also found other factors such as lower labour cost12 to be associated with lower youth
unemployment. Similar to Scarpetta’s findings, they also found that stronger labour market duality, of which was operationalized as the higher shares of temporary worker and lower employment
11 The latter two are determinants of labour market segmentation, or the degree to which the labour market is divided into labour market insiders and outsiders; this will be further elaborated on in the conclusion.
protection of temporary workers, tend to raise youth unemployment. In particular, the authors find that higher employment protection for temporary contracts is associated with lower unemployment rates for both young individuals and adults; but young individuals experienced stronger effects due to the fact that they make up a higher share of employment on temporary contracts.
Other macroeconomic studies that do not include a youth dimension but are worth discussing are as follows:
Elmeskov et al. (1998) analyzed annual data from 19 OECD countries over the 1983 – 1995 period to study the cross-country determinants of structural unemployment, focusing on the role of labour market policies and certain institutions. Again, ALMPU was used as the explanatory variable. The authors’ result echoed that of Scarpetta (1996), concluding that ALMPs do have a significant negative effect on the unemployment rate. Nickell and Layard (1999) conducted a similar study over a shorter time period (between 1983 and 1994) for 20 OECD countries, using ALMP spending per unemployed person as a percentage of GDP per member of the labour force as the explanatory variable instead of the ALMPU. Their results find that ALMPs do have a negative effect on long-term unemployment, but counter’s Scarpetta (1996) result by finding no significant effect on the employment-to-population ratio, possibly due to the difference in use of explanatory variable.
In contrast to the previous studies, Estevão (2003) argues that there has been a disproportionate focus on unemployment rates in ALMP studies, and therefore chose to use the “most appropriate measure for labour market performance”, that is, the employment rate (p.4). He
therefore evaluates the aggregate effect of ALMPs13 on employment rates, finding a positive
correlation between ALMP spending and employment rate in the business sector in the 1990s but not in the late 1980s. He argues that this result mimics the effect of increased intensity of ALMP spending on employment outcomes, as ALMP expenditures were still relatively small in the late 1980s compared to the 1990s.
13 As Estevão’s outcome variable is employment rates, he operationalizes ALMP as spending on ALMP as a percentage of GDP rather than the typically used ALMPU.
Lastly, in a more recent study, Hur (2019) evaluates the impact of government
expenditure on passive and active labour market policies14 on unemployment rates in the OECD
using panel-data between 2001 and 2013. In particularly, he conducts a difference-in-difference analysis to compare the effects of different policy responses to the labour market among OECD countries, and found that countries with higher ALMP spending before the 2008 financial crisis were found to be “more resilient” in terms of experiencing a smaller change in unemployment rate after the crisis when compared with countries that had a lower ALMP spending.
In summary, the studies discussed have found a negative relationship between ALMP spending and unemployment rates. ALMP impact on other outcome variables, such as employment rates still finds mixed results, depending on the operationalisation of ALMP expenditure. ALMP impact on youth labour market outcome is less studied on the macroeconomic level with those mentioned providing varying results (Scarpetta, 1996; Banerji et al., 2015). Micro data seem to suggest mixed results as well, depending on the program types; although systematic reviews have informed policy makers that ALMPs do impose a small but positive impact on youth labour market outcomes (Kluve et al., 2017; 2019).
As such, this suggests that the current study will need to utilize a variety of outcome variables that can capture the impact of ALMPs on varying dimensions of youth employment indicators. Furthermore, a particular focus on youth outcomes will allow for the investigation of macroeconomic-level impact of ALMP to corroborate with average estimates of microeconomic-level results presented in the systematic reviews. Before turning to an in-depth elaboration on the operationalisation of variables and model specification in the methodology section of this research paper, the following section will provide a brief discussion of the European experience with youth guarantee program, present the expected causal mechanisms of the EU YG scheme, and develop the hypotheses for this analysis.
2.5 The EU Youth Guarantee Scheme
“The Youth Guarantee is a commitment by all Member States to ensure that all young
people under the age of 25 years15 receive a good quality offer of employment, continued
education, apprenticeship, or traineeship, within a period of four months of becoming unemployed or leaving formal education” (European Commission, n.d.).
The notion of a youth guarantee is not new in Europe. Nordic countries having introduced such schemes since the mid-1980s, and while the programs in these countries diverged in many aspects, they all shared two common features. One, they all shared the primary goal of reducing the duration that young individuals remained unemployed or inactive, and two, to guarantee young individuals an offer of employment, academic, or vocation training opportunity. This would become the foundation for the development of an EU-level youth guarantee scheme which began since 2005 but would not become a realization that is adopted as an EU-wide commitment until
the end of 2013. While many EU countries16 have prior experience with youth initiatives that cover
a range of ALMPs, the YG scheme is distinguishable from typical ALMPs in two aspects. First, the “guarantee” concept is a commitment, which evokes the notion of a “rights-based” program. And secondly, the establishment of a limited time frame for activation, set at a maximum period of four months (Escudero & Mourelo, 2015).
Unlike preceding employment policy guidelines for young people, the YG scheme has received extensive political support by the European Commission and subsequently EU-level financial support, through both the European Social Fund (ESF) allocations and establishment of the Youth Employment Initiative (YEI). The ESF allocations are part of the standard multiannual financial framework that is renewed every seven years to member states aimed at financing the operation of all employment-related projects, including youth targeted projects. On the other hand, the YEI was established with the YG scheme in 2014 to specifically serve vulnerable young individuals (NEETs). Eligibility for additional YEI funding is dependent upon the regional youth
15 Extended to all young individuals under the age of 30 for some Member States, see table 3 in the appendix for more information on the main features of YG scheme.
16 In addition to Youth Guarantee programs already present in the Nordic countries and Austria, other notable youth initiatives include the Jugend mit Perspektive (JUMP) in Germany and the New Deal for Young People (NDYP) in the UK (Caliendo & Schmidl, 2016).
unemployment rate measured in 2012. As such the intensity of YEI funding allocated to each member state is dependent upon the number of regions within the country that experienced youth unemployment rate (ages 15-24 years old) above 25 percent in the reference year. To summarize, the policy change that all EU-28 member states faced in 2014 is the additional assistance provided by the European Commission in terms of implementation and monitoring of the scheme in national settings (including recommendations for structural changes), the commitment by each member state to allocate more national funding to youth initiatives, and the provision of additional ALMP spending through the YEI and ESF matching allocations.
The expected causal mechanism of the YG scheme follows that of a combination of ALMPs that are covered within the scheme. Employment services provide young individuals with information and networks that address information asymmetry within the labour market. Entrepreneurship programs assist young individuals who face barrier to access to credit and financial capital, as well as information gaps in terms of access to social networks and business set-up knowhows, with the added multiplier effect of creating more job opportunities for others too. Skills and training programs improve human capital and address inadequate supply of skills among young individuals, while targeted subsidized employment can be used to incentivize demand for young people’s labour and address the uncertainty that employers face when hiring young inexperienced workers. Additionally, another unique feature of the EC’s YG recommendations is that it is not a uniform one-size-fit-all scheme. Member states are given flexibility in terms of choosing the combination of interventions that are relevant within the given geo-economic context of each country (Caliendo et al., 2018). Therefore, in regions where labour demand is more flexible, YG interventions used may be geared towards training and reintegrating young individuals into the labour market (most non-YEI eligible countries), while in regions that face poor macroeconomic conditions, demand-side interventions such as wage subsidies and public sector employment may be used (multiple YEI-eligible countries).
There are however concerns with the scheme, as voiced by Escudero and Mourelo (2015), with regards to the universality of the “guarantee” notion. The authors argue that the success of such a scheme is dependent on the “capacity of labour demand to fulfill the commitments made” (ibid., p.7). Which means, if the increase in number of young people entering the labour market
competition within existing vacancies which will lead to an eventual fall in the reservation wage17 (Basu et al., 2009, as cited in Escudero & Mourelo, 2015). While theoretically a fall in reservation wage leads to lower unemployment durations which is beneficial for activating unemployed workers, the effect is determinantal if this is not met with an improvement on the demand-side of the labour market. The fall in reservation wage can disproportionately affect young individuals compared to adults in terms of forcing young workers into short-term or more precarious forms of employment. It can even be detrimental to the well-being of the most disadvantaged groups such as NEETs who have already become so detached from the labour market. As such, the EC’s idea of combining interventions with a more targeted approach (for example, the YEI funding is only aimed at assisting NEETs) will allow for the ability to service the most disadvantaged young individuals while not undermining the universality of the scheme.
The overall success of the YG scheme will therefore be dependent on the use of a full range of ALMPs that match the labour market needs of each member state, a clear identification of the target groups, coupled with reinforcement of institutional frameworks which includes the strengthening of local PES (key player in maintaining sustainable activation of needy young individuals) as well as strengthening of formal education and training structures to smooth out the school-to-work transition. If these conditions are met, the expected impact of the YG/YEI scheme on youth labour market outcome should be positive. Contrastingly, there are also issues with selection bias and the resulting displacement effects. If young individuals who are already qualified to find work on their own self-select themselves into ALMPs and displace those who are actually struggling to make the transition into the labour force, the YG/YEI treatment may have no impact or even a negative impact on aggregate youth employment outcomes.
2.6 Hypotheses
Drawing on the theory of change, hypothesizing that the implementation of YG/YEI scheme will result in a real increase in commitment to fund and implement ALMPs and other structural changes, the following hypotheses are derived:
H1 The YG scheme will have a significant positive impact on labour market outcomes of
eligible young individuals in EU-28 countries.
H2 The YEI will have a significant positive impact on labour market outcomes of eligible
young individuals in countries that receive YEI in EU-28 countries.
H0,1 The YG scheme will not have an impact on the labour market outcomes of eligible young
individuals in EU-28 countries.
H0,2 The YEI will not have an impact on the labour market outcomes of eligible young
3. Research Design
3.1 Sample Selection and Construction of Data
The EU-28 countries were selected as the sample for this study because the introduction of the YG scheme represents a policy change that all member states were committed to implementing equally from 2014 onwards. As such, analyzing the change in labour market outcomes of young individuals who are eligible to join a variety of activation programs under the YG in the EU after 2014 can act as a proxy to investigate whether the use of youth targeted ALMPs, rather than the more widely use of untargeted ALMPs, has a positive impact on labour market performances of young individuals. However, it must be noted that EU member states have varying experiences with youth ALMPs; many also have prior experience with youth guarantee schemes. The introduction of the YEI and proposed increase in both ESF and matching national allocations that accompany the YG scheme in 2014 can act as a proxy to help investigate the impact of
increased ALMP spending18 on labour market outcomes of young individuals. This will be
applicable for all countries, regardless of any prior experience with youth targeted ALMPs and guarantee schemes. This subsection will provide a description of the data sources, first by operationalizing the variables of interest, and followed by a discussion of the descriptive statistics.
Data used in this study is derived from statistical databases and European Commission publications. Cross-sectional time series data on labour market outcomes including employment-to-population ratio, labour force participation, and unemployment rates are drawn from OECD Statistics Database, while NEET rates are derived from the Eurostat Database. The panel data is
derived for every country in the EU-28 sample for the period between 2000 and 201919, with the
exception for NEET rates which are only available for the period between 2004 and 2019. The panel data is then complemented with information on the incidence of additional YEI funding and the cut-off ages for YG eligibility for each member state derived from the International Labour Organization’s systematic review of the European Youth Guarantee (Escudero & Mourelo, 2017),
18 Recall, financial contributions to the YG scheme are derived from multiple sources. While the main financial contributions are from the EU (ESF and YEI) and national budgets, other sources also include contributions from private and non-profit sectors (Caliendo et al., 2018, p. 10-11). Due to the fact that the ESF and YEI financial data are submitted through the System for Fund Management in the European Union (SFC support portal) wherein access is only granted to member state authorities, the incidence of YEI funding is therefore used as a proxy for additional ALMP funding in this analysis.
which was cross-checked with Youth Guarantee Implementation Plans submitted by member states (European Commission, n.d.b).
The YEI funding is allocated as a lump sum in 2014, provided to member states as a resource to fund specific YG projects targeting NEETs between 2014 and 2020 – as such there is no record for any annual variation across countries for YEI spending. Furthermore, it must be stressed that the lack of YEI or low YEI spending allocated to member states (see table 1 for the breakdown) does not imply low ALMP spending for NEETs or youth in a country. This funding is expected to complement national spending and ESF allocations on ALMPs targeted for young individuals, but it is not exhaustive. Therefore, although YEI represents increase in intensity of ALMP spending for some countries, for others it is just a supplementary financial instrument.
The use of multiple labour market outcomes is intended to address the various dimensions and manifestations of youth employment problems. While the unemployment rate is the most prominent indicator for labour market challenges, and often the most used outcome variable in macroeconomic studies on ALMPs, the indicator does not reflect the true proportion of those who are unemployed. Complementing this data with labour force participation rate and employment-to-population ratio can show whether a fall in unemployment rate is being absorbed into the labour force and increasing employment ratios, or resulting in a withdrawal from the labour force.
Figure 1 shows that the decline in average unemployment rate in EU-28 countries following the financial crisis is only matched with a comparatively small incline in average employment-to-population ratio. Meanwhile the decline in average labour force participation rate for young individuals still persists. Parts of the decline in young people’s engagement in the labour market, a pattern which is also observed globally, reflects the increasing enrolment in education; however, it also reflects the withdrawal from or failure to enter the labour force by an increasing proportion of young individuals leaving school (ILO, 2020).
Figure 1. Labour Market Outcomes for Youth (Age 15-24) in EU-28 countries, 2000-2019
Source: (OECD, 2020)
The analysis of labour market participation by young individuals therefore requires a different combination of indicators than the traditional analysis of adult labour markets which can be explained by employment, unemployment, and inactivity. Young individuals outside of the labour market can be divided into two groups: one, those who are still enrolled in school, college, university or other higher education establishment or training, and two, a smaller subgroup of young individuals who are neither in employment (unemployed or economically inactive), nor in education or training – abbreviated as NEETs (Eurostat, 2020b). Enrolment rates may not be useful for this analysis because it does not disaggregate between new entrants and those who are returning to education, among which, the latter may be deemed a consequence of activation measures. NEET rates, on the other hand, is a useful indicator to determine whether the most vulnerable and discourage youth are being reached and assisted by activation measures or not.
38.5 36.2 35.3 32.6 36.1 46.8 44.6 44.3 42.7 42.1 17.82 18.83 20.14 23.69 14.36 10 15 20 25 30 35 40 45 50 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Per cen ta ge Employment/population ratio
Labour force participation rate Unemployment rate
Figure 2: NEET rates for Youth (Age 15-24) in EU-28 countries, 2002-201920
Source: (Eurostat, 2020a)
An important characteristic of NEET rates is that it can be further disaggregated
depending on their activity statuses: NEETs who are unemployed but are actively seeking
employment, and those who are economically inactive or NEETs who are unemployed and are not actively seeking employment. The first group of unemployed NEETs consist of short- and long-term unemployed youths who are soon to return either to education or employment; this is, therefore, the group that is more susceptible to a country’s macroeconomic performance and business cycle (ibid.). This pattern can be observed in Figure 2, where the rise and fall of the average NEET unemployed rate among EU-28 countries following the financial crisis largely follows the same pattern over time as the average youth unemployment rate. Among those who are not engaged in the labour force or education, inactive NEETs also include other demographics such as the disabled and those engaged in household work (Elder, 2015). This is why while the NEET (unemployed) rates have risen and fallen with the business cycles, NEET (inactive) rates
have remained relatively consistent overtime. Unfortunately, due to limitations21 in NEET data
collected on Eurostat, only the aggregate data on the rate of NEETs will be used in this analysis.
20 Eurostat disaggregates NEET not employed rates into NEET unemployed (NEETs who are unemployed but are actively seeking employment) and NEET inactive (NEETs who are unemployed and are not actively seeking employment) (Eurostat, 2020b).
21 In addition to experiencing multiple consecutive breaks in the series for data in many countries, Eurostat also does not collect data for a large range of age groups for the NEET unemployed and NEET inactive indicators which
13.0 10.9 13.2 10.1 5.0 7.0 4.0 6.5 5.9 6.2 6.1 2 4 6 8 10 12 14 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Per cen ta ge
NEET (Not employed) NEET (Unemployed) NEET (Inactive)
Consequently, the outcome variables of interest in this analysis are as follows: employment-to-population ratio, labour force participation, unemployment and NEET not employed rates. These variables are measured as unweighted averages of 5-year age intervals for the 28 EU member states for the period between 2000 and 2019 (2004 and 2019 for NEET rates). Drawing on the aforementioned data, this analysis will compare the labour market performances of young individuals who are intended to receive the YG/YEI treatment with those who do not receive any treatment in the same time period. As such, the following dependent variables were constructed for each country and year:
1. A dummy variable indicating whether an age category is subject to YG treatment
or not. This dummy is equal to one for the age categories that receive YG treatment after 2014, and zero for the remaining age categories that are not eligible for the YG treatment in all countries.
2. A dummy variable indicating whether an age category is subject to YEI treatment
or not. This dummy is equal to one for age categories that receive YEI treatment after 2014, and zero for the remaining age categories in countries that are eligible for YEI treatment as well as all age categories in countries that are not eligible for YEI treatment.
It must be noted that while the YEI and YG treatment are constructed as two separate dummy variables, they are not mutually exclusive. Both variables affect the same age categories and is turned on for the same time periods (post 2014), however both dummies can appear simultaneously in the analysis due to the fact that a number of countries will experience only the YG treatment, while others experience both treatments. Table 1 summarizes the categorization of countries by treatment type and age cutoff for treatment eligibility.
Table 1. Summary of treatment type and age cutoff for treatment eligibility for EU-28 countries
Country YG_dummy ( t ≥ 2014 ) YEI_dummy ( t ≥ 2014 ) YEI amount (€ million) Age cutoff for YG/YEI
treatment Austria 1 0 0 25 Finland 1 0 0 25 Germany 1 0 0 25 Luxembourg 1 0 0 25 Malta 1 0 0 25 Netherlands 1 0 0 27 Denmark 1 0 0 30 Estonia 1 0 0 30 Slovenia 1 1 9.21 30 Cyrpus 1 1 11.57 30 Czech Republic 1 1 13.59 25 Latvia 1 1 29.01 30 Lithuania 1 1 31.78 25 Belgium 1 1 42.43 25 Sweden 1 1 44.16 25 Hungary 1 1 49.76 25 Bulgaria 1 1 55.18 30 Croatia 1 1 66.17 30 Ireland 1 1 68.14 25 Slovak Republic 1 1 72.17 30 Romania 1 1 105.99 25 Portugal 1 1 160.77 30 Greece 1 1 171.51 25 United Kingdom 1 1 206.09 25 Poland 1 1 252.43 30 France 1 1 310.16 25 Italy 1 1 567.51 30 Spain 1 1 943.49 30
Before moving onto a discussion of the specific indicators relevant to this study, figure 3 shows the average ALMP expenditure per unemployed worker expressed as a percentage of GDP per capita for four different time periods for a selected number of countries. Although ALMP expenditure is not stratified by target age groups, this indicator is adjusted for differences across countries in unemployment rates and the size of the economy and is therefore useful for illustrating the difference in level of generosity of each counties’ national ALMP programs. It is important to
note from this figure that there is a large variation among member states with regards to how generous their national ALMP programs are, even among countries that are eligible or not eligible for YEI funding.
Figure 3. Average ALMP expenditure per unemployed worker as a percentage of GDP per capita for selected countries in EU-28 sample
Note: Labour market policy expenditure (categories 2-7), derived in Euros at 2010 constant prices from the Employment, Social Affairs & Inclusion statistics database, is divided by the number of unemployed persons aged 15 to 74, in the same year and expressed as a percentage of GDP per capita. Labour market policy expenditure for Greece is missing for 2011 and 2018.
* Refers to countries that are eligible for YEI Source: (DG EMPL, 2020; Eurostat, 2020).
Figures 4 to 6 show varying dimensions of youth employment problems within the EU-28 sample. The ratio of youth-to-adult unemployment rate for each member state within the time period under investigation is presented in figure 4. This indicator provides an understanding of how young individuals fare in the labour market when compared to adults: an average ratio of about two denotes that young individuals (age 15-24 years old) are two times more than likely to be unemployed than adults (age 25-54 years old). As explained in the previous section, the discrepancy in performance is due in parts to young individuals’ limited work experience, making them less desirable for hire when competing for jobs against adults. However, it also due to the previously discussed unique characteristics of young individuals as well; unlike adults, young individuals can also rely on their families for financial support while they opt for unemployment to search for better employment options or seek further education opportunities as well (ILO, 2020).
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% Denma rk Nethe rland s Austr ia Germ any Eston ia Franc e* Irelan d* Spain * Gree ce* Roma nia* Av er age AL M P ex pen di tu re/ un em pl oy ed wo rk er , % of G DP p er c ap ita
Figure 4 is divided into four panels, split based on the average ratio of youth-to-adult unemployment rate experienced by each member state in the period right before the YG scheme was implemented (2011 – 2013).
Figure 4. Ratio of youth-to-adult unemployment rate for EU-28 countries, 2000 – 2019
Rati o of y ou th -to -ad ul t u ne m pl oy m en t r ate , %
Note: the ratio of youth-to-adult unemployment rate is derived by dividing the youth unemployment rates (age category 15 – 24) by prime-age adult unemployment rates (age category 25 – 54). This panel is split based on the average ratio of youth-to-adult unemployment rate between 2011 and 2013, as follows: low average ratio (1.60 – 2.25%), medium-low average ratio (2.29 – 2.60%), medium-high average ratio (2.71 – 3.03%), and high average ratio (3.06 – 3.95%). Unemployment rates for the age categories used in the figure for Lithuania is derived from Eurostat; while data for remaining countries are from the OECD. * Refers to countries that are eligible for YEI
1.00 2.00 3.00 4.00
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Cluster 1: Low average ratio 2011-2013
Germany Estonia
Latvia* Lithuania*
Austria Ireland*
Slovenia* European Union 28
1.00 2.00 3.00 4.00 5.00 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Cluster 2: Medium-low average ratio 2011-2013
Spain* Greece*
Denmark Netherlands
Bulgaria* Portugal*
Malta European Union 28
1.50 2.00 2.50 3.00 3.50 4.00 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Cluster 3: Medium-high average ratio 2011-2013
Slovak Republic* Hungary*
France* Cyprus*
Finland Croatia*
Belgium* European Union 28
1.50 2.50 3.50 4.50 5.50 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018
Cluster 4: High average ratio 2011-2013
Poland* Czech Republic*
United Kingdom* Luxembourg
Italy* Romania*
