NEET CONCEPT, NEET

Transcription

1 Number 3 December 2015 Mireille Zanardelli SOCIAL SECURITY STATISTICS BRIEF NEETs in a dynamic perspective Analysis using administrative data CONTENTS 1. The NEET concept 2. Measuring the NEET CONCEPT, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people 3. NEET trajectories: analysis of persistence and recurrence

2 Author: Mireille Zanardelli

3 NOTICE This publication is the first document drafted as part of a project initiated by the National Youth Service (SNJ) and conducted in collaboration with the Inspection générale de la sécurité sociale (IGSS) and the Luxembourg Institute of Socio-Economic Research (LISER). One of the SNJ's objectives is to support young people who have left the school system and are unable to join the labour market. To achieve this objective, SNJ wanted to better understand the population of young people in difficulty in Luxembourg, so as to improve its action and propose measures that are more suited to the characteristics of these young people. Because of its scope and originality, this project exceeds the needs of the SNJ and falls within the broader perspective of public policies designed to support young people in difficulty. For the part carried out by LISER, the project benefited from financial support from the European Social Fund (ESF), whose national counterpart was financed by the SNJ. A collaboration agreement with the SNJ was drafted for the part carried out by IGSS.

4 Contents 1 INTRODUCTION THE NEET CONCEPT Definition and interpretation Limitations of the NEET concept and the tools proposed to overcome them Heterogeneity of the concept Lack of dynamic dimension in the NEET concept MEASURING THE NEET CONCEPT, INCORPORATING A DYNAMIC DIMENSION: USING ADMINISTRATIVE DATA TO CONSTRUCT INDIVIDUAL TRAJECTORIES OF YOUNG PEOPLE TRAJECTORIES OF NEETS: ANALYSIS OF PERSISTENCE AND RECURRENCE Results Discussion CONCLUSION BIBLIOGRAPHY... 24

5 INTRODUCTION 1 INTRODUCTION In recent years, the employment situation of young people has become a major political issue, both internationally and in Luxemburg. One of the reasons for the growing concern is the economic crisis. While in relative terms the latter has not affected youth unemployment in Europe as much as that of their elders, the level of youth unemployment in Europe is now a matter of concern and has reached such values that it has given rise to the newly termed sacrificed generation 1. In this context, international bodies have introduced, as part of their long-term strategy, specific targets for young people to promote national policies that can improve their situation on the job market. Increasingly, these objectives include a specific indicator known as NEET: Not in Education, Employment or Training. The NEET rate, which measures the proportion of young people aged between 16 and 24 and not in employment, or continuing across the entire age group, was selected by the United Nations as part of the Post-2015 Sustainable Development Goals ; objective 8 aims to substantially reduce the proportion of NEETs. It has also been one of the European Union s official indicators since In addition, the need to focus more on NEETs has also been placed at the heart of the integrated guidelines proposed by the European Commission for employment policies. The NEET concept was proposed in the late 1980s in the United Kingdom in order to overcome the unemployment rate, which was considered insufficient to understand the precariousness and vulnerability of young people on the job market. The unemployment rate because it only refers to the active youth population (including those in employment and those seeking employment) does not take into account young people who are not employed, not in training and not seeking employment. It is nevertheless essential to take an interest in this category of young people, since this inactivity must be overcome as soon as it appears to reduce future exclusion and precariousness. Thus, unlike the unemployment rate, the NEET rate takes into account the entire age group and provides a more comprehensive overview of the situation of young people, taking account of both exclusion from the job market and exclusion from the training system. However, the NEET rate has a number of limitations. Indeed, while it offers a more complete overview of the situation of young people, it is true that its interpretation is difficult, precisely because it combines several dimensions relating to the situation of young people (Elder, 2015). As a result, the NEET rate is often presented as a complex concept summarising potentially very heterogeneous situations that may reflect different levels of vulnerability and precariousness. This heterogeneity should be taken into account in order to put the phenomenon into perspective and to provide multiple responses that can be used to target the diversity of the situations 2. Another limitation of the NEET rate, which is not referred to as often in the literature, seems to be more fundamental to us. Almost all work on NEETs has a static dimension and studies the situation of young people at a given moment, without ever accounting for the dynamic dimension of the phenomenon throughout the trajectories of young people. Yet it can be assumed that being NEET at one given point in time is not a problem; the problem lies in either the recurrence of the situation in the trajectory of young people, or the status' persistence in the medium/long term. 1 Since the start of the economic crisis in 2008, the youth unemployment rate (i.e years of age) in Europe increased from 15.9% to 22.2% between 2007 and 2014 (an increase of 40%). The unemployment rate among older citizens also increased by more than 40% over the same period, but at much lower levels, rising from 6.1% to 8.9%. 2 This critique, however, is not exclusive to NEETs; many phenomena combine heterogeneous situations that need to be accounted for in order to conduct correct analyses. 5

6 INTRODUCTION In an attempt to provide answers to the two limitations above, the Inspection générale de la sécurité sociale (IGSS), in collaboration with the National Youth Service (SNJ) and the Luxembourg Institute of Socio-Economic Research (LISER) has developed a project with two objectives: 1/ To better understand the characteristics of young people and be able to measure the heterogeneity of situations that they face, by conducting a survey of a sample of NEETs. It is clear that the available surveys do not achieve this objective; they rarely cover all areas that need to be documented and never at the same time in order to provide a comprehensive and appropriate picture of the situation of young people. For example, the use of drugs or alcohol, which is central to the issues surrounding young people, is rarely measured and even more rarely combined with a measure of how marginalised a young person is or with a measure of their cognitive skills; this last point is essential for identifying chances of emerging from NEET status. 2/ To construct and analyse the trajectories of young people in general and NEETs in particular, using the administrative data available in Luxembourg 3. There are three reasons why administrative data was used as a complement to the survey: i) it is always difficult to construct trajectories of individuals using surveys because it is a retrospective exercise that can be affected by memory bias; ii) because the plan is not to replicate the survey, it cannot enable a permanent monitoring tool to be set up for the situation of young people; the implementation of such a tool seems fundamental to us in the current context of increasing difficulties for young people; the administrative database can fulfil this objective since it can be updated regularly, for example, to measure changes in NEET rates, especially to estimate the effects of public policies implemented in favour of young people; iii) in addition, administrative data provide a complete census. 3 From an operational point of view, the survey was implemented and run by LISER, while the administrative database was created by IGSS. 6

7 The NEET concept 2 THE NEET CONCEPT In recent years, the NEET concept has become very popular. It is an official indicator for both the European Commission and the United Nations, and is the subject of in-depth studies by the Dublin Foundation (EUROFOUND, 2012) and the OECD (Carcillo. et al., 2015). However, this concept is not necessarily uncontroversial, firstly because of mistakes in its interpretation and secondly because of its limitations. Therefore, we will firstly provide a precise definition of the NEET concept and relate it to the other concepts used to describe the situation of young people in order to understand the interpretation problems noted in the literature (Elder, 2015). Secondly, we will highlight the limitations of the NEET concept, as outlined in the introduction, which will enable us to justify the complementary developments proposed in this document. 2.1 DEFINITION AND INTERPRETATION The acronym NEET stands for Not in Education, Employment or Training. A NEET is therefore a young person who, at any given time, is not in employment and not part of a training system either initial or continuing. In most cases, the studied population concerns young people aged 16 to 24 years. In order to understand the outline of the NEET concept, we must list the various statuses that young people may experience. These are as follows: 1. in employment 2. looking for a job while simultaneously undertaking continuing 3. looking for a job, without continuing 4. inactive and enrolled in an initial training system 5. inactive and in continuing 6. inactive and not in any training system (initial or continuing) Based on this division, NEETs correspond to cases 3 and 6. Young people looking for a job while simultaneously undertaking continuing (case 2) are not considered NEETs because the training takes precedence. This is also the case for young students reporting that they are looking for a job, for which student status takes precedence (they are included in case 4). The rate of NEETs is calculated as follows: rate neets = (3) + (6) [16 24 year olds] It can also be written in the following way: rate neets = [16 24 year olds] (1) (2) (4) (5) [16 24 year olds] rate neets = (1) [16 24 year olds] (4) [16 24 year olds] (2) + (5) [16 24 year olds] The term (1) [16 24 year olds] corresponds to the employment rate of the age group since it reports the number of young people in employment out of the total number of year olds; the term (4) [16 24 year olds] corresponds to the enrolment rate of the age group since it reports the number of young people in school or in out of the total number of year olds; the term (2)+(5) [16 24 year olds] corresponds to the number of year olds in continuing. The latter rate includes two different situations: on the one hand, that of inactive young people deciding 7

8 The NEET concept to undertake training on their own initiative and, on the other hand, that of inactive young people on training courses proposed and financed by the Public Employment Service (i.e. ADEM in Luxembourg). The first case is relatively rare insofar as life-long learning is most often practised within companies and, as such, it is mostly reserved for those in employment (since in this case, the company is responsible for the costs of continuing ). As a result, the term (2)+(4) [16 24 year olds] measures the participation rate of year olds in active employment policies (AP), and mainly accounts for the intensity of these policies. The more a country supports job seekers by providing training to improve employability, the higher the and the lower the NEET rate. (2)+(4) [16 24 year olds] rate will be Thus, as illustrated in figure 1, the following relationship can be deduced from the foregoing developments: rate neets 1 employment rate enrolment rate AP participate rate The NEET rate is therefore a rate combining three dimensions: the first (1-employment rate) relates to the problem of unemployment and exclusion from the job market, the second (1- rate) relates to the exclusion of initial training systems and the third (1-AP participation rate) relates to the exclusion of active employment policies. This combination of three elements within the same indicator explains the critiques of the NEET concept with regard to difficulties in interpreting this indicator. Elder et al. (2015) present a detailed analysis of NEETs by breaking them down according to these dimensions. 4 For example, they show that the contribution of each of these dimensions to the NEET rate is not equal for men and women. This critique seems well-founded and leads us to the following conclusion: if we want to understand the evolution of the NEET rate, it should be disaggregated to identify which components are involved in the observed evolution. However, this critique does question the value of this synthetic indicator, which has the advantage of providing an overall measure of young people who are potentially in precarious situations due to their exclusion from both the job market and and training systems (including AP). As such, the NEET indicator is, in our opinion, an effective alert that summarises the risks faced by young people and measures the way forward if the objective is not to leave any young people on the wayside. 4 Elder et al. (2015) group the second and third dimensions into one, conjoining those in both initial and continuous training. 8

9 The NEET concept Diagram 1 Definition and breakdown of NEETs NEET rate = 1 - Employment rate - School enrolment - AP participation rate in employment in employment in employment in employment Looking for a job and in continuing Looking for a job and in continuing Looking for a job and in continuing Looking for a job and in continuing 3 NEETS 3 NEETS 3 NEETS 3 NEETS All young residents [16-24 years of age] Looking for a job and not in continuing Looking for a job and not in continuing Looking for a job and not in continuing Looking for a job and not in continuing 4 Inactive at school or in (initial training) 4 Inactive at school or in (initial training) 4 Inactive at school or in (initial training) 4 Inactive at school or in (initial training) Inactive, in continuing Inactive, in continuing Inactive, in continuing Inactive, in continuing 6 NEETS 6 NEETS 6 NEETS 6 NEETS Inactive, not in school, or training Inactive, not in school, or training Inactive, not in school, or training Inactive, not in school, or training 9

10 THE NEET CONCEPT 2.2 LIMITATIONS OF THE NEET CONCEPT AND THE TOOLS PROPOSED TO OVERCOME THEM The NEET concept has two important limitations: the first is due to the heterogeneity of the situations to which it corresponds, while the second is due to the absence of a dynamic dimension in measuring the concept Heterogeneity of the concept The heterogeneity of NEETs stems from two elements. The first is the direct consequence of what we described above; since the NEET concept includes job market alienation, as well as alienation from initial training systems and APs, NEETs now face a vast array of situations. Therefore, a NEET may well be a young person who has left the school system with a certain level of and who is unable to find a job, or a school dropout with no qualifications. The second element of heterogeneity is the diversity of levels of vulnerability and precariousness to which the NEET status may correspond. The work of the Dublin Foundation (2012) has, for example, conceptualised five NEET subgroups: the conventionally unemployed, the unavailable, the disengaged, the opportunity-seekers and the voluntary NEETs. These categories, although they can be discussed, nevertheless reflect very different situations in terms of alienation from the job market and ability to integrate. It would therefore be advisable to perceive this heterogeneity in such a way as to put the phenomenon into perspective and to provide multiple answers that target the array of situations. It is in this perspective that the decision was made to conduct a survey of a sample of young NEETs with the aim of gathering the necessary information to identify the diversity of NEET situations. A literature review on the, very rare, surveys dedicated to NEET issues and a reflection on the fundamental elements which we believe need to be understood, led us to propose a comprehensive and innovative questionnaire that explores the following dimensions: i) measuring the level of marginalisation of the young person as well as their alienation from the job market and public institutions responsible for supporting NEETs, ii) iii) identifying addictive behaviour in relation to drugs or alcohol, measuring the young person's cognitive skills through a series of face-to-face tests. This part of the survey, which is unprecedented to the best of our knowledge, aims to distinguish those NEETs for whom it is reasonable to expect integration into the primary market and those whose capacities might be insufficient. In terms of public policy, this distinction seems fundamental, since it refers to various measures. In the first case, it involves, for example, providing adequate training, which will improve the employability of young people, whether it concerns the form (CV format, behaviour throughout the application process, etc.) or substance, through training the young person for specific jobs where opportunities could be identified beforehand. In the second case, subsidised employment, specifically for young people in difficult circumstances, must be created in sectors of the economy that have been identified as being capable of absorbing this type of workforce. The survey referred to here was carried out by LISER between October 2014 and June It will give rise to publications, the purpose of which will be to account for the heterogeneity of NEETs and to identify factors related to this status Lack of dynamic dimension in the NEET concept As explained in the previous discussion, the NEET concept is measured in a static way. Indeed, a NEET is a young person who is not in employment, nor in a training system (initial 10

11 THE NEET CONCEPT or continuous); not for a certain duration but at a given moment. The lack of dynamic analyses of the NEET problem is probably due, in part, to the lack of data available. On a European level, the NEET rate is calculated based on Labour Force Surveys (LFS), which do not always have a longitudinal dimension and which, where relevant, often do not contain sufficient numbers to analyse the trajectories of NEETs in a satisfactory manner. However, a persistent or recurrent NEET status, which can only be perceived using a dynamic approach, seems to introduce another source of heterogeneity. Two young people, both of whom are NEET at a given point and with a certain skill level, can nevertheless face very different situations; while the first remains in NEET status for many months, unable to find a job, the second only remains NEET for a few weeks. Similarly, two young people will be in different situations if, after they have both found employment, one becomes NEET within a few months and the other remains in employment. Failing to account for the persistence or recurrence of NEET status would probably lead to the number of NEETs that genuinely need support being overestimated and to amalgamation between heterogeneous situations which undoubtedly require different political responses. All these arguments led us to try to look at NEET issues in a dynamic perspective by aiming to reconstruct the trajectories of young people. As mentioned earlier, administrative data was used in such a way that we could regularly update the database and therefore have a permanent monitoring tool on the situation of young people. In the next section, we will detail the procedure used to create the administrative database (choice of data sources, interconnection procedures for administrative files, etc.). We will also discuss the methodological issues raised during this exercise, the limitations of these data and the procedure used to ensure the relevance of administrative data for NEET issues. Readers who are not interested in data-related methodological questions may find it more useful to refer directly to Section III, where results are discussed. 11

12 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people 3 MEASURING THE NEET CONCEPT, INCORPORATING A DYNAMIC DIMENSION: USING ADMINISTRATIVE DATA TO CONSTRUCT INDIVIDUAL TRAJECTORIES OF YOUNG PEOPLE When this project began, it was not certain that administrative data could be used to construct trajectories of young people. Insofar as the administrative data is not collected for statistical purposes, it is not always possible to use it to adequately measure the phenomena being studied. We therefore needed to conduct exploratory analyses to validate the relevance of administrative data for the NEET project. Ultimately, this was validated at the expense of certain limits which we will present later in this document. To create a database to identify the trajectories of young people, we took the following approach: i) We first created a file containing, for different dates, all young people aged 16 to 24 who were resident in Luxembourg (from September to the latest data available at the time of writing this document, on a quarterly basis 12/2010, 03/2011, 06/2011, 09/2011, etc.); ii) Then, for each young person in this first file, we added their status at each of the chosen dates; the various statuses are as follows: 1) in employment, 2) looking for a job while simultaneously undertaking continuing, 3) looking for a job, without continuing, 4) inactive and enrolled in an initial training system, 5) inactive and in continuing, 6) inactive and not in any training system (initial or continuing); this succession of quarterly statuses was used to construct and analyse the trajectories; iii) for each young person, we also added all socio-demographic characteristics that were available in the administrative data, so we could delve further beyond the NEET count alone and characterise the young people concerned in an effort to identify risk factors. To build this database, we interconnected the different administrative files managed by various public institutions, each covering part of our needs. In this section, we will give a detailed account of each of the sources used, which form the official base for calculating this rate. We will describe their use, describe their limitations and explain the consequences of said limitations when estimating the number and rate of NEETs compared to that which is obtained from Labour Force Surveys (LFS). i) The first step in our approach was to identify all young people aged 16 to 24 residing in Luxembourg. For this purpose, we used files from the Inspection générale de la sécurité sociale (IGSS) relating to persons covered by the Luxembourg social security system. These files identify two categories of young people: 5 This date was chosen as, prior to September 2010, data that could identify students was unavailable. 12

13 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people - The first category corresponds to young people with personal insurance; they are mainly young people who work in Luxembourg, but also young people who receive replacement income, subject to contributions, for example unemployment benefits, or guaranteed minimum income. It is important to note that among young active residents, two groups are not included in the administrative data: firstly, young people residing in Luxembourg and working abroad who do not pay contributions in Luxembourg (the number of young people concerned is very low 6 ); secondly, young people who work in Luxembourg but who are linked to a social security system outside of Luxembourg. This second group, which mainly concerns international civil servants, is also very small 7. Excluding these two groups leads to a marginal underestimation of the total number of young people aged residing in Luxembourg and thus to an overestimation, which is also marginal, of the NEET rate in relation to LFS estimates. - The second category of young people covered by the social security system corresponds to those who are co-insured, i.e. they are affiliated to the Luxembourg system, not in a personal capacity, but as a beneficiary of one of their parents. Again, co-insured young people who have international civil servants as parents and are therefore affiliated to a social security system outside of Luxembourg do not appear in these files, which also leads to an underestimation of the denominator of the NEET rate. Moreover, assuming that the risk of being NEET is lower than average among children of international civil servants, their absence from administrative records may lead to an overestimation of NEETs (which is the case for the NEET rate calculated using the LFS surveys). It should also be noted that the IGSS files contain (unlike LFS surveys) young people living in collective housing (if they are in some way affiliated to the social security system). It can be assumed that these young people, estimated at some 900 individuals for the age group 8, are more frequently in a precarious situation; including all or part of this category of young people leads to an increase in the denominator, but also to a (proportionately higher) increase in the numerator of the NEET rate, which should lead, all else being equal, to a higher NEET rate than that calculated using LFS data. ii) Files from the Administration for Employment Development (ADEM) allowed us to identify unemployed young people seeking employment, by distinguishing between those who are benefiting, at different moments in time, from an active policy in the form of training proposed by ADEM (case 2 of diagram 1) and those who do not benefit from the same (case 3 of diagram 1). It is well known that some young people looking for employment do not register with the public employment service, the main reason being that they are not entitled to unemployment benefits due an insufficient contribution period 9 (mostly young people who have recently left school). But this situation does not pose a problem when identifying NEETs. Indeed, even if these young people do not appear in the ADEM files, one can assume that they appear in the IGSS files, since the young people concerned are probably co-insured, being linked to one of their parents. Therefore, they will be counted in the group of NEETs. iii) The Ministère de l Education Nationale, de l Enfance et de la Jeunesse (MENJE) enabled us to identify young people enrolled in public secondary in Luxembourg. As a result, two categories of pupils are not included in these files: pupils enrolled in schools abroad and those in schools established in Luxembourg, but not under the management of MENJE 10. In Luxembourg, in view of certain national specifications, around 1000 young people are enrolled in secondary abroad and the number of pupils enrolled in institutions not under the aegis of MENJE is far from negligible. From the administrative 6 See the Situation du marché de l emploi dans la Grande Région - Mobilité des frontaliers [Labour market situation in the Grande Région - cross-border worker mobility] report by the Observatoire Interrégional du marché de l'emploi (November 2014, p.5). 7 About 1% of young people aged 15 to 24 residing in Luxembourg are international civil servants (source: STATEC Census, 2011). 8 Source: STATEC, Census, ADEM manages unemployment benefits therefore, anyone who wants to receive unemployment benefit must be registered with ADEM. 10 Private schools and European schools are not under the aegis of the Ministry of Education, meaning their students are not registered in the Ministry files. 13

14 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people data at our disposal, we looked for other ways to identify young students who slipped through the MENJE files. After investigation, it appears that before a young person turns 18, we cannot identify them using administrative information. Nonetheless, after they turn 18, we can identify them using information from the National Family Allowance Office concerning the payment of family allowances. Since 2010, family allowances, which are paid for all children up to the age of 18, are discontinued beyond that age unless the child is enrolled in secondary school (regardless of the school). Therefore, due to the impossibility of identifying 16- and 17-year-old students enrolled abroad or in schools not under the aegis of MENJE, we have limited the scope of our study to young people aged 18 or over. This restriction is clearly detrimental, since year olds are a key target group in public policies aimed at young people, especially in relation to dropping out of school. Studying the 18-plus group is however interesting, meaning the administrative database created as part of this project has some use. iv) In Luxembourg, no single file centralises all young residents enrolled in higher. To identify young students, we used files from the Centre de Documentation et d Information sur l Enseignement Supérieur (CEDIES - Centre for Documentation and Information on Higher Education). Indeed, CEDIES is the body that manages the allocation and payment of scholarships paid by the Luxembourg Government. Until 2010, these bursaries were in addition to family allowances that received by families and were means-tested; young people listed at CEDIES therefore constitute a part of all students (those with the lowest incomes). A new law, passed in 2010, removed child benefits from the age of 18 and replaced them with bursaries. Therefore, all young people pursuing higher after secondary must, if they wish to obtain a scholarship, register with CEDIES. In our study, the question that arises following this legislative change is in relation to the comprehensiveness of the CEDIES files. Do they include all young students? In which case is it reasonable to rely on these files to identify students, or, are the files fragmented? In which case, will using them cause us to overestimate the number of NEETs and the NEET rate? Three factors may affect the completeness of CEDIES files. - The first concerns the possibility of not making use of the bursary. However, we believe that this can in no way be significant enough to justify not using the CEDIES files. Indeed, the assumption that households do not have the right information, which is the main reason for not making use of the bursary (students not using their rights do not submit an application) does not seem likely to us. Besides the fact that the legislative debate on bursaries has been widely publicised, their introduction in 2010 replaced family allowances, which, overnight, were stopped for those aged 18 or over; families had been accustomed to keeping them in the event of the child undertaking higher. Families who may not have been aware of the legislative reform to introduce bursaries, most likely turned to the Caisse Nationale des Prestations Familiales [National Family Allowance Office] to find out why their right to family allowances was suspended and would then have been referred to CEDIES to receive a bursary if their child were pursuing higher. This sudden switch from family allowances to bursaries seems to be a sufficient safeguard to rule out the hypothesis of not making use of the bursary due to a lack of information. - The second element that could affect the completeness of the CEDIES file is linked to cases of multiple resits. CEDIES rules indicate that students have n+1 years to complete a course of n years. In other words, only one resit is granted for a given course. Therefore, if young people are resitting a year for the first time, they will receive funding the following year. However, if the student resits again, they will no longer receive a grant. The consequence of this rule in the event of resitting depends on the student s attitude. If they are aware that the second resit means they are ineligible for a bursary and decide not to submit an application, then they will not appear in the CEDIES files. In this case, they will not be considered as a student and will be wrongly considered as NEET. If, however, the student is unaware that they are ineligible and submits an application, they will be counted among the students because the CEDIES files not only provide information on bursaries paid, but also 14

15 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people those that have been filed and rejected. To ensure completeness of the CEDIES files, we sought to compare the number of students estimated from CEDIES with those provided by some identifiable schools in the CEDIES files. For example, we asked some of the largest high schools offering BTS (technician certification) training to provide the number of students enrolled in their school for the academic years and so that we could compare the number of students listed in the CEDIES files. The same comparison was made for students enrolled at the University of Luxembourg. The figures from both sources were very close on all occasions. This enables us to conclude that the comprehensiveness of the CEDIES files is more than satisfactory for the most quantitatively significant channels The third element that may lead to the completeness of the CEDIES files being questioned is the existence of training courses that are not recognised as higher and are not relevant to secondary, which means they are ineligible for scholarships. I.e. accessible training after high school but not degrees. The data from CEDIES highlights that very few applications from young residents are refused based on this criterion. However, it is possible that some young people enrolled in such courses eluded us: those who are aware of the bursary award criteria and do not file an application with CEDIES. v) The SNJ files allowed us to identify the young people who used the voluntary career guidance service (SVO), coordinated by the SNJ. These young people were considered active. 11 Unlike files from the Ministry of Education, which contain a variable that indicates the start and end of enrolment in the school, the CEDIES data does not show whether the young person interrupted their ongoing studies during the year. Once a request was submitted to the CEDIES, they were considered a student for the entire school year. It is therefore likely that we underestimated time spend in NEET status for a student whose studies were interrupted during the year. 15

16 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people Given the above, the NEETs comprise the following two categories: i) young people registered with ADEM (paid or not) receiving no training measure ii) and co-insured young people 12 who are not registered with ADEM and are not pupils (i.e., not in the MENJE files), not students (i.e., not in the CEDIES files) and not using the voluntary career guidance orientation (SVO). This second group of NEETs is defined by default, i.e. through a lack of any other status. To make the connection with the figure 1, we summarised, in figure 2, data sources that were used to identify each category of young people and the inherent limitations of the data 13. Based on this information, the status of young people between 18 and 24 years of age has been identified at different points in time, on a quarterly basis from September 2010 (NB: those under 18 years of age were removed from the scope of our study, because it was not possible to identify those who are studying abroad or in private schools in Luxembourg). It is from the status of each young person at different points in time that we were able to build their trajectory and measure the persistence and recurrence of NEET status. The NEETs at time t were defined as young people who are not in employment, or training at ADEM or beneficiaries of an SNJ measure. Based on all the elements presented in this second section, using the identification of a cohort of young people who became NEET at the same time, we will analyse their trajectories so as to better understand the diversity of profiles and different levels of precariousness that may lead to this concept. 12 i.e. those who do not work or do not receive unemployment benefits. 13 Given the various limitations of the administrative data and given the area covered by the same, it is not surprising that estimated NEET rates obtained from administrative data vary from those obtained using Labour Force Survey data. According to administrative data, the NEET rate for years is approximately 14% versus 9% according to the LFS for the same age group (confidence interval: [7.9%; 9.7%]). This difference is explained by: i) LFS including international civil servants, while administrative data does not; this difference results in a higher NEET rate when administrative data is used compared to LFS data, since it can be assumed that children of international civil servants are less frequently NEET than other young people; ii) LFS not including young people living in collective housing as is partly the case with administrative data; this difference also results in a higher NEET rate when measured using administrative data than LFS data, since one can again assume that young people in collective housing are more frequently NEET than the others. In addition, the comparison between the estimated NEET rate from EFT and from administrative data is difficult insofar as LFS uses a sample spread over the entire year; this means any seasonal variations can be removed, while the rate estimated using administrative data is calculated for a specific time. Depending on the month, the NEET rate increases or decreases by 4 percentage points (from August until the end of the year, the NEET rate is much higher than other months due to all young people finishing school and seeking employment). The rate of 14% is an average of the monthly rates. 16

17 Measuring the NEET concept, incorporating a dynamic dimension: using administrative data to construct individual trajectories of young people Diagram 2 Use of administrative sources for measuring NEETs Administrative sources Limitations of administrative sources 1 In employment IGSS / files related to affiliation with the social security for employed persons Does not include young people who work outside of Luxembourg or those who do work in Luxembourg but are affiliated to another social security scheme (international staff) 2 Looking for a job and in continuing ADEM / job seekers registered with ADEM and receiving an active policy measure All young residents [18-24 years of age] 3 NEETS Looking for a job and not in continuing ADEM / job seekers registered with ADEM and not receiving an active policy measure 4 Inactive, in continuing No existing source but numbers are considered low 5 Inactive at school or in (initial training) MENJE / students enrolled in a school in Luxembourg National Fund for Family Allowances / students over 18 in secondary CEDIES / students receiving a scholarship Does not include students attending school abroad or private schools in Luxembourg Does not include students who repeat at least two classes and did not apply for a grant Does not include students enrolled in Postgraduate training that does not lead to a recognized diploma field restricted to young people who are 18 and older 6 NEETS Inactive, not in school, or training Obtained by deduction, combining all data These are coinsured young people, in their parents name (identified from IGSS data) and are neither in school nor in or training at ADEM 17

18 Trajectories of NEETs: analysis of persistence and recurrence 4 TRAJECTORIES OF NEETS: ANALYSIS OF PERSISTENCE AND RECURRENCE Firstly (3.1), the key results will be presented first in order to secondly (3.2) provide a basis for a discussion of the fundamental lessons to which they lead. 4.1 RESULTS To study the trajectories of NEETs, we established a cohort of young people that became NEET during the same period. For this, we selected all young people who were NEET in March 2012, but who were not six months ago, i.e. in September In other words, we selected a group of young people who were NEET in March 2012 for a period varying from 1 month to 6 months. For young people from this cohort, we made half-year observations on the chance of leaving NEET status and also how this was achieved, where applicable. We identified three possible wasy in which people leave NEET status. The first two were expected: leaving NEET status for employment and leaving NEET status for school, or continuing. The final way in which people leave NEET status was less expected it is directly linked to Luxembourg s significant migration flow. Indeed, given the latter, a significant number of young people immigrate to Luxembourg and then return to their country of origin. This specificity must absolutely be taken into account to the extent that, as we will see, it makes meaningful changes to the appearance of the NEET issue compared to that of a country in which migration flows are much lower. We will also show how the number of young people who return to their country is estimated using administrative data. Based on the individual trajectories, we analysed persistence and recurrence in NEET status. Figure 3 shows the timing of emerging from NEET status and shows the number of young people experiencing recurrence after going on to employment, or training. 18

19 Trajectories of NEETs: analysis of persistence and recurrence Diagram 3 The trajectories of the cohort of young people that became NEET between September 2011 and March 2012 March September 2012 March 2013 September 2013 March ,708 young people became NEETs between Septemb er 2011 and March 2012 In employment: 549 (32%) In : 119 (7%) Training with ADEM: 51 (3%) Still NEETs: 803 (47%) Recurrence of NEET status between September 2012 and March 2014: 261 (15%) In employment: 160 (9%) In or training: 11/17 Still NEETs: 525 (31%) Recurrence between March 2013 and March In employment: 83 (5%) In or training 20/4 (1%) Still NEETs: 380 (22%) Recurrence between In employment: 63 (4%) In or training: 5/4 (-1%) Still NEETs: 276 (16%) Left Luxembourg: 186 (11%) Left Luxembourg: 90 (5%) Left Luxembourg: 38 (2%) Left Luxembourg: 32 (2%) Sources: IGSS/ADEM/CEDIES/MENJE / Calculations: IGSS Note: the % in brackets are calculated using all 1,780 young people in the cohort 19

20 Trajectories of NEETs: analysis of persistence and recurrence This figure highlights the following results: i) Within the study cohort (made up of young people who became NEET between September 2011 and March 2012), 64% leave NEET status in the two years that follow. Of these, 50% go into employment (32% within the first six months) and 14% into or training (9% within the first six months). Among the 64% of young people leaving NEET status, 44% did not experience recurrence within 18 months, as is the case for 20% of them (table 1). It is important to note that the recurrence frequency after 18 months estimated here corresponds to a minimum value insofar as not all young people are observed over an identical range of time. Young people who no longer have NEET status in September 2012 are observed for 18 months while those no longer in NEET status in March 2013 and September 2013 are observed for 12 months and 6 months respectively, due to a lack of necessary data. In other words, at least 20% experience recurrence within 18 months of emerging from NEET status. ii) iii) In 16% of cases, NEETs who were evaluated were in the same situation two years later, in March Persistence in NEET status affects about two out of ten young people. Finally, 20% of the NEETs left the country between March 2012 and March This figure is estimated from the number of young people who disappear from administrative files, i.e., those who were affiliated with the Luxembourg social security system between September 2011 and March 2012 and who, two years later, no longer are, in any form whatsoever (there is no data to identify with certainty whether they left Luxembourg). This estimation is subject to two biases. The first, which is marginal because it concerns very few cases, is linked to the fact that some of the young people who leave the social security system are actually on the streets, homeless and have no relationship with their families. This explains why they no longer benefit from social security cover that they would have when living with their parents. The second bias is that there is no disaffiliation obligation with regard to social security in the event of leaving Luxembourg. In other words, as with the recurrence figure, our estimated figure (20%) is the lower limit: at least 20% of NEETs left the country. If this figure proved to be underestimated, the proportion of persistent NEETs would be reduced 14. Table 1. Main results Persistence in NEET status for at least 24 months 16% Leave Luxembourg 20% Leaving NEET status for employment, or training, with no recurrence Leaving NEET status for employment, or training, with recurrence 44% 20% Sources: IGSS-ADEM-MENJE-CEDIES 14 Among the young people who leave the social security system, 80% are of foreign nationality. Amongst whom, a large majority does not seem to return to Luxembourg. Indeed, of the 186 young people who disappeared from administrative files between March 2012 and September 2012, 60% had not returned to Luxembourg by March 2014, 18 months later. 20

21 Trajectories of NEETs: analysis of persistence and recurrence 4.2 DISCUSSION These results inspire different comments when looking at them in terms of public policies and supporting young people in difficulty. i) The NEET rate calculated in European indicators overestimates the gravity of the problem. Temporary situations that will most often be resolved without public intervention cannot be differentiated from situations of greater concern, which will continue and reduce the chances of finding employment or returning to school. Computing dynamic indicators is complex. However, the figures presented here clearly illustrate the difference in perspective that static and dynamic indicators can provide and encourage reflection on how to take a dynamic approach to the NEET issue into account. ii) iii) Overall, the persistence and recurrence of NEET status affect at least 36% of the analysed cohort. So, more than a third of NEETs surveyed face one of two situations. It is important to note that, by using data that only covers a two-year period, as is the case here, it is only possible to observe cases of recurrence on the one hand and persistence on the other. However, it would also be interesting to prolong the observation period in order to assess if and, where relevant, how such persistence and recurrence are combined in the same trajectory. Several situations can be considered. One can indeed imagine events in which persistence in NEET status may occur after several non-virtuous recurrences; these don t allow the young to to accumulate skills and experience; on the contrary, he is faced to a dynamic of exclusion which, ultimately, leads him to drop out the labor market. Persistence and recurrence may also correspond to different situations in which the two phenomena are not necessarily linked. If persistence occurs immediately after becoming NEET, it can actually be a voluntary choice on the part of some young people who decide to travel, for example, or young mothers raising their child(ren) (for these young mothers, the voluntary dimension of their situation is undoubtedly less systematic). The assumption about young mothers seems to be confirmed by the fact that among persistent NEETs, 57% are women while among recurring NEETs, only 41% are women (Table 1). If we accept that we need to combat the persistence and recurrence of NEET status, we need to identify determinants so the trajectory of young people can be anticipated and public support can be provided to those who are most exposed to these risks. From this perspective, administrative files are very poor and provide little in terms of characteristics that can help such identification. However, some descriptive elements, shown in the following table, can be established. In particular, they highlight that the risk of persistence is stronger for women than for men and for foreigners than for Luxembourgers (women represent 57% of persistent NEETs, 41% of recurring NEETs and represent 46% of the cohort - foreigners account for 59% of persistent NEETs, 48% of recurring NEETs and represent 56% of the cohort). The survey conducted by LISER, which provides extensive information on the characteristics of young people, will complement previous statistics. Note that IGSS is trying to improve our analysis of NEETs using other administrative bases, which remain unexplored for scientific and statistical purposes. In particular, the exploratory work aims to understand the situation of parents; certain studies have shown it has a significant impact on the situation of children. This aim to supplement the administrative data used is part of the strategy that we hope to implement for this project: while the survey highlights certain risk determinants for the persistence and recurrence of NEET status, the objective is to determine the extent to which these determinants or proxies can be identified from administrative data; this would provide a permanent monitoring tool that can better target public intervention and evaluate effectiveness. Table 1. Characteristics of the young people 21