Strong governance of skills data is essential for helping policy makers and stakeholders navigate the complexity and uncertainty associated with the design and implementation of skills policies. This chapter explains the importance of skills data governance in Luxembourg and provides an overview of current practices and performance. It then explores two opportunities for strengthening skills data governance in Luxembourg: improving the quality of Luxembourg’s skills data collection; and strengthening co‑ordination of, and synergies between, skills data within and beyond Luxembourg.
OECD Skills Strategy Luxembourg
Assessment and Recommendations
OECD Skills Studies
5. Strengthening the governance of skills data in Luxembourg
Abstract
The importance of strengthening the governance of skills data
Strong skills data governance is essential for helping policy makers and stakeholders navigate the complexity and uncertainty associated with the design and implementation of skills policies. Skills policies are complex as they fall at the intersection of multiple policy fields, including education, labour market, innovation, industrial and migration policy. At the same time, skills policies are developed in the context of substantial uncertainty as they are significantly impacted by megatrends, such as globalisation, automation, digitalisation, demographic change and climate change (OECD, 2019[1]), many of the implications of which are being accelerated by the COVID‑19 pandemic. In the context of uncertainty and rapid change, strong skills data governance facilitates the provision of timely and relevant information, which is necessary for ensuring that governments and stakeholders can effectively design and implement skills policies and make informed choices leading to better skills outcomes. Such evidence-based policy making can also help to strategically target investments and generate higher returns on skills investments.
In this chapter, skills data is understood as all data relevant for skills policy making, most importantly, education and training data and labour market data. Strong governance of skills data refers to: 1) collecting skills data effectively and efficiently (i.e. collecting high-quality data and coordinating within government and with non-governmental stakeholders in the data collection process, respectively); and 2) facilitating the analysis, exchange and co‑ordination of skills data (e.g. via data interoperability, data exchange platforms, etc.). Using skills data (and information on education and training opportunities) effectively and efficiently for career guidance is explored in Chapter 3.
Strong skills data governance provides the foundation for the successful design and implementation of skills policies and programmes in all of Luxembourg’s skills policy priority areas described in this Skills Strategy. Skills data are important for better aligning the adult education and training offer to fast-changing labour market needs (see Chapter 2) and informing the design and implementation of guidance and financial incentives that help steer skills choices (see Chapter 3). In addition, skills data are necessary to generate information on current and future skills needs, which is key for recruiting the right foreign talent (see Chapter 4).
In the long run, strong skills data governance supports building integrated skills information systems (Figure 5.1),1as strong data governance helps mobilise data and enhance data management and evaluation processes. In Luxembourg, the importance of strong skills data governance is further underscored by the fact that relying solely on national skills data sources, in most cases, does not fully capture the complexity of Luxembourg’s skills system due to the high reliance on labour sourced from the Greater Region (see Chapter 4).
This chapter is structured as follows: the following section provides an overview of the current skills data governance practices in Luxembourg. The next section describes Luxembourg’s skills data governance performance. The last section conducts a detailed assessment and provides targeted policy recommendations in two opportunities for strengthening the governance of skills data in Luxembourg: improving the quality of Luxembourg’s skills data collection; and strengthening co‑ordination of, and synergies between, skills data within and beyond Luxembourg.
Figure 5.1. The four building blocks of strong skills system governance
Source: Adapted from OECD (2019[1]), OECD Skills Strategy 2019: Skills to Shape a Better Future, https://dx.doi.org/10.1787/9789264313835-en.
Overview and performance
Overview of Luxembourg’s current skills data governance practices
Collection of skills data in Luxembourg
Table 5.1. Key sources of administrative skills data in Luxembourg
Note: ABBL: Luxembourg Banker’s Association; HORESCA: National Federation of Hoteliers, Restaurateurs and Café Owners of the Grand Duchy of Luxembourg; ISCO: International Standard Classification of Occupations; NACE: Statistical classification of economic activities in the European Community; ROME: Operational Directory of Trades and Jobs; ESCO: European Skills, Competences, Qualifications and Occupations Classification; ISCED: International Standard Classification for Education; IVET: Initial vocational education and training.
The Joint Social Security Centre (CCSS) is Luxembourg’s key source of data for analysing the structure of its national labour market and its evolution. Employers are required to make a declaration to the Common Centre for Social Security when they wish to recruit a new employee. Data collected by CCSS include information on the employee’s occupation at the moment of hiring, sector, the type of contract and duration and the employee’s place of residence.
Luxembourg’s public employment service (ADEM) is its key data source on vacancies and job seekers. ADEM collects data on job vacancies by economic activity (NACE2) and occupation (ROME3), as well as on job seekers by gender, age, duration, resident status and education level. ADEM vacancy and job seeker data make it possible to analyse the jobs available and the occupations sought by job seekers, as well as the mismatch between the two. In addition, vacancy data are also collected by certain sectoral institutions in Luxembourg, such as the Luxembourg Banker’s Association (ABBL) (financial sector) and the HORESCA Federation, which run their own job boards. The job board of the HORESCA Federation also contains data on job seekers in the HORESCA (hospitality) sector. In addition, data on vacancies are contained on Luxembourg’s private job portals, which are web-scraped and centralised by the European Centre for the Development of Vocational Training (CEDEFOP)’s Skills-OVATE (online vacancy analysis tool for Europe) tool. Co‑ordination between Luxembourg and CEDEFOP is facilitated by the Luxembourg Institute of Socio-Economic Research (LISER), serving as a national contact point for CEDEFOP.
Administrative data on participation in general education (ISCED 2011 Level 1‑8) are collected centrally by the Ministry of Education, Children and Youth (MENJE) and the Ministry of Higher Education and Research (MESR). These data also cover enrolment, new entrants and graduates and are reported to international institutions for the purpose of tracking data and indicators on education systems (OECD, 2021[4]). Data on adult education and training participation are collected in a decentralised manner by key public providers (e.g. the National Centre for Continuing Vocational Training [CNFPC], University of Luxembourg Competence Centre [ULCC], etc.), sectoral institutions providing training (e.g. ABBL, Chamber of Commerce [CC], Chamber of Skilled Trades and Crafts [CdM], Chamber of Employees [CSL], etc.), as well as private training providers. In addition, ADEM collects data on job seekers’ training participation.
Administrative data on education and training outcomes in Luxembourg are produced and collected, respectively, by the National Institute for the Development of Continuing Vocational Training (INFPC) and the University of Luxembourg (UoL). In addition, the Training Observatory of the INFPC regularly collects data on outcomes of initial vocational education and training (IVET) as part of the School – Active Life Transition (TEVA) annual study by linking data from MENJE, MESR and CCSS and on companies’ training practices.
At the higher education (HE) level, the UoL carried out a one-off graduate tracking exercise in 2021, which collected information on graduates’ first and current employment via LinkedIn and through graduates’ academic supervisors.
Luxembourg’s collection of administrative skills data (Table 5.1) is further complemented by skills data collected through surveys (Table 5.2).
Table 5.2. Key skills-related surveys in Luxembourg
The National Institute for Statistics and Economic Studies (STATEC) is in charge of administering Luxembourg’s household surveys relevant for the analysis of skills policies. The European Union Labour Force Survey (LFS), implemented in Luxembourg, collects, among others, information on respondents' (residents only) employment status (employed/unemployed/inactive), occupation (ISCO code), as well as participation in formal and non-formal education (except for guided on-the-job training) in the last four weeks. With the European Union (EU) Adult Education Survey (AES) implemented in Luxembourg, STATEC collects, among others, information on public participation in formal and non-formal education and training and informal learning in the last 12 months. Information on obstacles to training participation and time spent on education and training is also available via AES. STATEC is equally in charge of carrying out the Continuing Vocational Training Survey (CVTS), which measures enterprises' activity in the provision of continuing vocational education and training (CVET), and the Structure of Earnings Survey (SES), which collects data on earnings of all firms with at least ten employees in all sectors by economic activity (NACE) and occupation (ISCO). Similarly to LFS and AES, both CVTS and SES are EU surveys, which STATEC implements in Luxembourg.
The Training Observatory of the INFPC collects data on the learning offer of adult learning providers through a regular survey and launched another survey in 2021 on the provision of CVET in small companies. In addition, the UoL conducts an annual survey of its graduates, collecting self-reported data on graduates’ education and training outcomes (e.g. employment status, earnings, perceived mismatch, use of skills on the job, etc.).
In addition to surveys administered by governmental actors, non-governmental and sectoral actors in Luxembourg actively collect data on workforce needs, recruitment challenges and firm-based training provision for some of the sectors of Luxembourg’s economy through their own surveys.
The ABBL conducted a one-off survey with member companies on the difficulties and duration of recruitment in the financial sector in 2018. The Competence Centre of the Building Services Engineering and Completion Engineering (GTB/PAR), which provides training to craft companies, conducts a survey twice a year to enquire about the skills/training needs among employers in the construction sector. The survey collects information on skills and training needs by level of education and by job type. In addition, the Business Federation Luxembourg (FEDIL) conducts a survey every two years on the manufacturing and information and communication technology (ICT) sectors’ future qualification needs.
The CC, in collaboration with the Luxembourg Institute for Social Research (Institut Luxembourgeois de Recherches Sociales, ILRES), a private survey company, carries out its annual Barometer of the Economy study that collects information on the key concerns among Luxembourgish employers. Each Barometer of the Economy has a different thematic edition. In 2019 and 2021, the thematic editions enquired about employers’ recruitment difficulties, and firm-based training provision, respectively. In addition, the CdM conducted a survey in 2019 on skilled labour needs and shortages in the crafts sector, supplemented by a qualitative analysis based on interviews with company managers from the various groups of craft activities. The survey was repeated in 2022.
Since 2013, the CSL, in collaboration with the UoL, has conducted the Quality of Work Index survey on an annual basis with questions on employees’ working conditions and the quality of work in Luxembourg. Topics cover work demands and workloads, working hours, co‑operation between colleagues, possibilities for further training and advancement, participation in decision making in businesses, and more.
Analysis, exchange and co‑ordination of skills data in Luxembourg
Several key institutions in Luxembourg engage in the analysis, exchange or co‑ordination of the collected skills data (as mentioned above). These institutions include ADEM, the General Inspectorate of Social Security (IGSS), INFPC, STATEC and the Labour Market and Employment Research Network (Réseau d’études sur le marché du travail et de l’emploi, RETEL). In addition, the “Trends” working group, established between ADEM and the Union of Luxembourg Companies (UEL), helps increase the transparency of existing skills data and to further consolidate and improve them.
ADEM publishes labour market analyses regularly, including analyses of vacancies reported to, and job seekers registered with, ADEM (ADEM, 2022[15]). In 2022, ADEM also elaborated sectoral studies, which analyse labour market trends across seven sectors in Luxembourg. Based on ADEM vacancy data, the sectoral studies were developed in collaboration with employers to provide information about Luxembourg’s growing, declining and emerging occupations, as well as labour shortages and surpluses, among others (ADEM, 2022[16]).
IGSS engages in the analysis and publication of labour market data. On a regular basis, IGSS publishes data on labour market trends in Luxembourg (including net employment creation trends by sector, age, gender, nationality, and place of residence, among others) (IGSS, 2022[17]). Every month, IGSS also publishes a snapshot of Luxembourg’s labour market situation (including the size of the labour force and labour force growth comparison with the preceding year, among others) (IGSS, 2022[18]).
INFPC’s Training Observatory conducts research in the field of education and training to support policy makers, employers and education and training providers in improving education and training policies and practices. The Training Observatory analyses employees' access to employer-sponsored training, firms' training activities and the state's financial contribution to firms' training plans, among others. The Training Observatory is a member of ReferNet, the European network of reference and expertise on vocational education and training (INFPC, 2021[19]).
STATEC carries out and publishes relevant analyses often (but not only) on the basis of the surveys it administers (see Table 5.2). For example, relevant publications have presented the results from the AES (STATEC, 2018[20]) and the CVTS (STATEC, 2018[21]) for Luxembourg.
Data exchange and collaboration between IGSS, ADEM and STATEC are supported via RETEL. RETEL was established to commission new, and centralise existing, labour market research in Luxembourg. RETEL aims to support the development of synergies between the work of various governmental entities and stakeholders engaged in Luxembourg’s primary labour market data collection and the analysis of labour market data. Through its work, RETEL seeks to foster an efficient and transparent approach to data-driven production of labour market information and evaluation of labour market policies in Luxembourg (Government of Luxembourg, 2018[22]).
MESR launched an initiative to set up a National Data Exchange Platform (NDEP) in Luxembourg to facilitate smoother data exchanges, covering data in various fields beyond skills. The NDEP project is developed in co‑operation with various other governmental institutions in Luxembourg (see more in Opportunity 2).
Finally, to increase the transparency of the existing skills data (as summarised in Table 5.1 and Table 5.2) and further consolidate and improve them, Luxembourg established the Trends working group within the framework of the long-term collaborative effort between ADEM and UEL (ADEM/UEL, 2021[23]). Apart from ADEM, the Trends working group includes various members of the UEL, including representatives of the ABBL, CC, CdM, Luxembourg Trade Confederation (CLC), Federation of Artisans (FDA), FEDIL, Competence Centre for Building Services Engineering (GTB) and the HORESCA Federation (ADEM/UEL, 2021[23]).
Luxembourg’s skills data governance performance
As foreshadowed at the beginning of this chapter, strong skills data governance requires that skills data collection processes are effective and efficient (i.e. skills data collection is well-coordinated, resulting in the generation of high-quality data) and that skills data exchanges are easily facilitated.
Skills data governance in Luxembourg could be strengthened on several fronts. First, there is room to improve the quality of Luxembourg’s skills data collection in terms of accuracy (i.e. the degree to which the data correctly estimate or describe the quantities or characteristics they are designed to measure) (OECD, 2011[24]), coverage (i.e. completeness of the data) and granularity (i.e. level of detail) (OECD, 2014[25]), as well as to expand the range of skills data collected.
Moreover, there is space to better facilitate skills data co‑ordination and exchanges both within the government and with relevant stakeholders in Luxembourg. With the labour market extending beyond national borders, Luxembourg could equally benefit from building stronger synergies with international data sources, especially to better understand the changing skills supply and demand in the Greater Region (see Chapters 1 and 4 for a discussion of the strategic importance of the Greater Region for Luxembourg).
The cross-border nature of Luxembourg’s labour market further complicates skills data governance processes. As almost half of Luxembourg’s labour force is constituted by cross-border workers (see Chapter 4), Luxembourg does not participate in the OECD Survey of Adult Skills (Programme for the International Assessment of Adult Competencies, PIAAC). As a result, Luxembourg cannot benefit from the survey’s assessment of the literacy, numeracy and problem-solving skills possessed by adults or how adults use their skills at home, at work and in the wider community. Similarly, other international surveys that Luxembourg takes part in (e.g. the EU LFS, AES, etc.), which, among others, provide information on the occupational distribution of Luxembourg’s workforce or participation in adult education and training, cover residents only. It is possible to include cross-border workers in the analysis of Luxembourg’s occupational distribution after combining variables in the LFS, but only with certain caveats (e.g. limited disaggregation is possible). A similar exercise is not possible in the context of the AES.
Existing skills data governance challenges affect Luxembourg’s skills assessment and anticipation (SAA) exercises.4 For example, given that ADEM’s sectoral studies (see above) were based solely on ADEM vacancy data, which themselves face important coverage challenges (due to the incomplete declaration of job vacancies by employers; see more in Opportunity 1), the results of the sectoral studies have to be interpreted with caveats, bearing in mind the limitations of the data sources used.
In addition, Luxembourg’s SAA exercises tend to be restricted to identifying current skills needs (e.g. by analysing skills demanded in job vacancies) and changing labour market conditions (e.g. growth or decline of certain occupations). However, little is done to anticipate Luxembourg’s future skills needs. Stakeholders have indicated that, to a large extent, the absence of skills anticipation exercises is linked to the quality and co‑ordination challenges of Luxembourg's existing skills data sources.
It is welcome that the recently established Trends working group led by ADEM and UEL (see above) has recognised many of the skills data challenges highlighted above and intends to work collaboratively on solving them. In addition, LISER has similarly made improving the “data infrastructure and, in particular, access to high-quality data in and for Luxembourg” one of its core objectives (LISER, 2020[26]).
Opportunities for strengthening the governance of skills data in Luxembourg
The performance of Luxembourg’s skills data governance arrangements reflects many factors. These include individual and institutional factors and broader socio-economic factors (i.e. Luxembourg’s heavy reliance on cross-border labour). However, two key opportunities for improvement have been identified based on a literature review, desk analysis and data, and input from officials and stakeholders consulted in conducting this OECD Skills Strategy.
The OECD considers that Luxembourg’s main opportunities for improvement in the area of skills data governance are:
1. improving the quality of Luxembourg’s skills data collection
2. strengthening co‑ordination of, and synergies between, skills data within and beyond Luxembourg.
Opportunity 1: Improving the quality of Luxembourg’s skills data collection
As foreshadowed earlier in this chapter, Luxembourg collects a wide variety of quantitative and qualitative data, which can be used to inform the design of skills policies, including labour market data and education and training data. These data sources, and their combination, can provide different types of insights essential for skills policy making.
Labour market data (e.g. social security data, vacancy data, etc.) are commonly used as proxies for the analysis of evolving skills needs because they allow for the observation of growth or decline in employment in specific occupations (OECD, 2016[27]). Vacancy data collected by public employment services (PES) can also provide valuable information on skills needs and shortages by facilitating the analysis of the changes in the number of declared and (un)filled vacancies and the duration of vacancy filling, among others (McGrath and Behan, 2017[28]). In addition, insights from employer surveys can be used to complement administrative data to assess skills needs and shortages by providing information not available in administrative datasets.
Education and training data (e.g. on attainment, participation, outcomes, expenditure, curricula, etc.) provide important information on the access, quality and relevance of the supply of education and training opportunities. This information helps policy makers identify challenges and opportunities in the skills system and target programmes and funding where the expected impacts will be highest.
To improve its quality, Luxembourg’s skills data collection could benefit from:
improving the accuracy, coverage and granularity of Luxembourg’s labour market data
expanding the range and strengthening the granularity of Luxembourg’s education and training data.
Improving the accuracy, coverage and granularity of Luxembourg’s labour market data
As shown in Table 5.1 and Table 5.2, Luxembourg collects a wealth of labour market data, which provides valuable information on the evolving skills needs in Luxembourg’s economy. However, to allow data users (both in and outside of the government) to unlock the full potential of Luxembourg’s labour market data, Luxembourg should work on addressing the accuracy, coverage and granularity challenges of certain labour market data sources.
Luxembourg’s social security data, collected by the CCSS, are Luxembourg’s only source of occupational data covering both Luxembourgish residents and cross-border workers. Given that cross-border workers account for 46% of Luxembourg’s labour market (see Chapter 4), the importance of CCSS data cannot be overstated. CCSS data can inform the assessment of the occupational distribution in Luxembourg’s labour market and serve as a proxy for assessing current skills needs by making it possible to observe growth or decline in specific occupations. Should Luxembourg consider developing exercises for anticipating future skills needs (as supported by stakeholders consulted in the context of this project), such as Canada’s Occupational Projection System (COPS) (Government of Canada, 2017[29]) or the United Kingdom’s Working Futures projections (Warwick Institute for Employment Research, n.d.[30]), CCSS data would be valuable for producing occupational projections (i.e. projections showing the expected future number of workers to be employed in different occupations, thereby producing insights about the future occupational growth or decline).
However, CCSS occupational data face important accuracy challenges due to incorrectly declared occupational codes by employers or their service providers (e.g. payroll managers or outsourced accounting firms). As mentioned above, each newly hired employee has to be reported to the CCSS (Government of Luxembourg, 2021[31]). When making the declaration to CCSS, the occupation of the new employee needs to be specified via a code defined in the International Standard Classification of Occupations (ISCO). However, there is evidence suggesting that in many cases, companies’ payroll managers or accounting firms to whom payroll management can be outsourced input incorrect ISCO codes in the employment declarations (ADEM/UEL, 2021[23]), making CCSS data unusable for the analysis of Luxembourg’s labour market at the occupational level.
During a data check carried out in 2016, the IGSS found that the ISCO codes were entered correctly in about only 30% of cases. More specifically, IGSS found that a large share of occupations belonging to ISCO groups 1‑3 (managers, professionals, and technicians and associate professionals, respectively) were reported under ISCO group 4 (clerical support workers). In part, the reason for inputting incorrect ISCO codes stems from the French translation of the ISCO classification, where the name of ISCO group 4 is translated as “administrative employees” (employés administratifs), which does not fully reflect the meaning of the original English name “clerical support workers”.
In order to support employers and their service providers in declaring the correct ISCO codes, IGSS published an online guide in French, where the distinction between the ISCO groups was clearly delineated. IGSS also sent out a letter to employers to highlight the existence of the problem, underscoring the importance of correctly declaring the ISCO codes. Moreover, CCSS developed an occupational code search engine on its website, accessible in French, aiming to help identify the correct ISCO codes more easily. However, it was discussed in stakeholder consultations that in a recent data check, IGSS found that these communication efforts did not yet have the desired effect, as the share of ISCO codes declared correctly had not improved.
Occupational data (declared via ISCO codes) is used for purely analytical purposes in Luxembourg, which means that employers do not have a strong incentive to ensure that the declared ISCO codes are correct. Going forward, Luxembourg could incentivise employers to declare correct occupational codes by designing practical tools, which would have clear value-added for employers without increasing employers’ administrative burden. For example, based on the CCSS data, CCSS and IGSS could design an online occupations dashboard, where employers would log in and track the occupational structure and evolution within their company, together with additional features (e.g. employee salary distribution, age profiles, etc.). Such a dashboard could bring significant benefits, especially to small and medium-sized enterprises (SMEs), which tend to have constrained administrative capacity for human resources (HR) management. Furthermore, to increase the value-added of the dashboard, employers in Luxembourg could be, in the long term, asked to regularly update the declared ISCO codes, given that the codes are currently not being updated at all.
To further facilitate ISCO code declarations, CCSS could improve the search functions of the existing occupational code search engine. For example, the occupational coding tool developed by the University of Warwick in the United Kingdom features probability scores reflecting the perceived match between the keywords input into the tool and the suggested occupational codes, as well as descriptions of tasks of the occupations suggested to match one’s search, entry routes, associated qualifications and related (similar) job titles (Box 5.1). To support the user-friendliness of the occupational coding tool, CCSS could develop a chatbot allowing real-time responses to queries of employers or their service providers who might be unsure of which occupational code to declare. CCSS could also consider making the search engine accessible in English in addition to French, given the challenges related to correctly translating the ISCO codes from English to French highlighted above. In the long run, making the search engine accessible in German and Luxembourgish could also be considered, given the variety of languages used in Luxembourg and its labour market (see Chapter 4).
Box 5.1. Relevant international example: Occupational coding tool
United Kingdom
In the United Kingdom (UK), the Warwick Institute for Employment Research at the University of Warwick developed the Computer Assisted Structured Coding Tool (“Cascot”) capable of occupational coding and industrial coding to the UK standards developed by the UK Office for National Statistics. The standards are the Standard Occupational Classification (SOC) and the Standard Industrial Classification (SIC).
Cascot assigns a code to a piece of text and calculates the degree of certainty (denoted by a score from 1 to 100) of the assigned score being correct. When Cascot encounters text that does not allow it to clearly match it with any specific occupation or industry, it will attempt to suggest a code but will limit the assigned score to below 40 to signal the uncertainty associated with the suggestion made.
Cascot also features descriptions of tasks of the occupations suggested to match one’s search, entry routes, associated qualifications and related (similar) job titles. A multilingual version of Cascot (Cascot International) is available in 15 languages, including French and German. Cascot International facilitates occupational coding to the ISCO 08 classification and a relevant national occupational classification, where available.
Cascot has been evaluated by comparing the tool’s suggestions with a selection of manually coded data of high quality. The results showed that 80% of records received a score greater than 40, and of these, 80% were matched to manually coded data. The performance of Cascot is dependent on the quality of input data.
Source: Warwick Institute for Employment Research (2018[32]), Cascot: Computer Assisted Structured Coding Tool, https://warwick.ac.uk/fac/soc/ier/software/cascot/details/; Warwick Institute for Employment Research (2018[33]), Cascot International, https://warwick.ac.uk/fac/soc/ier/software/cascot/internat/.
While the existing awareness-raising activities mentioned above (i.e. developing an online guide, sending a letter to employers) undertaken by CCSS/IGSS are steps in the right direction, they could be further strengthened. Information on common mistakes in inputting ISCO codes, as well as an overview of the existing resources and tools for helping employers select the correct ISCO code (e.g. the proposed occupations dashboard, CCSS’ occupational code search engine, IGSS’ online guide, etc.) could be regularly distributed to employers and the existing guidance tools more prominently featured on IGSS/CCSS websites. CCSS/IGSS could expand their awareness-raising activities to include associations of human resources (HR) professionals (e.g. POG – HR Community of Luxembourg) and associations of payroll experts (e.g. Luxembourgish Association of Accounting and Tax Consultancies, ALCOMFI) to target actors directly in charge of employment entry declarations. Employer representatives (e.g. UEL) should equally play a key role, and promote the importance of, and existing support tools for, correctly declaring new hires to the CCSS among employers. In addition to undertaking efforts to strengthen the accuracy of CCSS data, IGSS and other stakeholders in Luxembourg should still make use of the data to the extent possible (e.g. for the analysis of general growth trends across occupations).
Beyond addressing accuracy issues with social security occupational data, there is space to improve the coverage and granularity of Luxembourg’s vacancy data. In Luxembourg, a legal obligation exists for employers to declare all job vacancies to ADEM. Employers can report vacancies online via guichet.lu (Luxembourg’s e-government platform) or by filling out a PDF form which needs to be sent to ADEM via email or post. A vacancy declaration can be completed in English, French or German (ADEM, 2021[34]). Despite these efforts, a comparison of the number of vacancies declared to ADEM to the number of actual recruitments (based on employment entry declarations to the CCSS) shows that the vacancies reported to ADEM cover less than 30% of all actual job creations in Luxembourg’s labour market (ADEM, 2021[35]). The number of declared vacancies also varies by sector. For example, while the estimated average share of job vacancies reported to ADEM was approximately 35% in the ICT sector between January 2020 and January 2021, the figure stood at 18% in the construction sector (ADEM, 2021[35]). While a certain degree of caution should be used when interpreting the results of the comparison of vacancy and recruitment data, since not every recruitment is preceded by a job posting, it can still provide a rough estimate of the coverage of the vacancy data. Given the relatively low number of vacancies reported to ADEM, it can be challenging to use ADEM vacancy data to reliably assess Luxembourg’s labour market needs and shortages.
Stakeholders have mentioned several potential reasons for the relatively low number of vacancies reported to ADEM. For example, employers, and especially SMEs, without a dedicated HR department might struggle to find the time and resources to report job vacancies. Stakeholders have also suggested that certain employers do not see great value in reporting vacancies to ADEM since the pool of candidates (registered job seekers) is often unlikely to match the exact requirements of the declared vacancy. Moreover, some recruitments are based on informal referrals rather than officially created and published job postings.
It should be noted that ADEM has taken a number of concrete steps to incentivise employers to improve job vacancy reporting rates. Penalties for enforcing the mandatory vacancy reporting obligation are not applied, and ADEM prefers to use positive incentives instead (Goffin, 2015[36]). For example, ADEM has simplified the vacancy reporting forms (Goffin, 2015[36]) and has been running awareness-raising campaigns for employers within the framework of the ADEM/UEL partnership, underscoring the fact that vacancy reporting is as important for analytical as for recruitment purposes. ADEM also sends its employer service staff to visit companies and highlight the added value of regular vacancy reporting (Goffin, 2015[36]). Moreover, ADEM has started offering the possibility to automatically import a vacancy from a company's job database to ADEM, which several large employers have already implemented. In addition, ADEM has begun creating partnerships with certain private local job portals to automatically import their online job advertisements (OJAs) into ADEM's job board. Finally, ADEM has been developing a range of new services, so it is more attractive for employers to approach ADEM. For example, in April 2021, ADEM’s job board was for the first time opened to job seekers who are not registered with ADEM (ADEM, 2021[37]). The results of such efforts are yet to be seen.
Data from OJAs, as already partly explored by ADEM, could serve as a useful additional source of vacancy data for Luxembourg, as OJA data capture even those vacancies not declared to ADEM, but that appear on private job portals. CEDEFOP’s Skills-OVATE tool uses web-scraping techniques to collect OJAs from private job portals, public employment service portals, recruitment agencies, online newspapers and corporate websites (CEDEFOP, 2021[38]). In addition, Skills-OVATE provides a granular view of skills most sought after by employers in their vacancies, classified according to the European Skills, Competences, Qualifications and Occupations (ESCO) classification (see more below). While Skills-OVATE data for Luxembourg cover online private job boards, it does not include ADEM’s vacancy data. This is because Skills-OVATE only sources information from publicly accessible job boards, and ADEM’s job board was not public when the Skills-OVATE tool was designed. Since then, however, ADEM has agreed to publish its vacancies (with employers’ permission) on the publicly accessible European Employment Services (EURES) job board, with the view of supporting EU job mobility. Therefore, a large share of ADEM’s vacancies is now accessible on the EURES public job board and can be scraped by Skills-OVATE.
LISER, the national contact point for CEDEFOP, should thus consider reopening conversations with ADEM about including their vacancy data in the Skills-OVATE tool. This would allow for more comprehensive information on both occupational and skills needs trends in Luxembourg by having access to data from vacancies declared both to ADEM and at Luxembourg’s private job portals. As Skills-OVATE is updated four times a year, skills needs information could be obtained regularly. Partnering with CEDEFOP would equally help ensure that the data does not contain duplicate vacancies (i.e. vacancies declared both to ADEM and on private job boards) using the tools CEDEFOP had developed. Making greater use of OJA data could help improve both the coverage and granularity of Luxembourg’s vacancy data. Nonetheless, ADEM should still strive to increase the share of job postings created and declared by employers. As noted above, since not every recruitment is preceded by a job posting, even greater use of OJA data might not provide a comprehensive picture of Luxembourg’s labour market needs.
Going forward, ADEM could consider further expanding the range of support tools and services for employers to strengthen the incentives for employers to approach ADEM to declare job postings. For example, ADEM could develop an online vacancy dashboard for employers, similar to the occupations dashboard (see above), based on vacancy data reported to ADEM. Upon logging in, the vacancy dashboard could give the employer an overview of the vacancies (open and filled) over time within the company and how quickly different vacancies are being filled. In the future, consideration could be given to further developing the technical functions of the vacancy dashboard, for example, to calculate the “attractiveness scores” (i.e. probabilities of finding a suitable candidate for each vacancy based on ADEM’s job seekers database) of each vacancy. Stakeholders consulted in the context of this project agreed that employers, especially SMEs, would benefit from such an online vacancy dashboard. In addition, CCSS and ADEM could work on creating links between CCSS employment entry declarations and ADEM vacancy declarations (see more on ADEM vacancy declarations below) to make the reporting process more straightforward for employers. For example, a vacancy reported by an employer to ADEM could be assigned a unique identifier, which, once a suitable candidate has been found for the advertised job posting, could be used to pre-fill the CCSS employment entry declaration.
Evidence shows that the capacities of small employers to engage in formal recruitment, and thereby vacancy reporting, tend to be restricted the most owing to resource constraints or organisational culture (European Commission, 2018[39]). Therefore, to incentivise SMEs to approach ADEM and encourage higher vacancy reporting rates, ADEM could consider developing dedicated support services for SMEs. For example, ADEM could provide support to SMEs in drafting job postings according to their needs, on top of providing additional services (e.g. sign-posting the available reskilling/upskilling opportunities for SME employees). Box 5.2 details the recruitment and support services provided to SMEs in Germany by way of example.
Box 5.2. Relevant international example: Recruitment and support services for SMEs
Germany
In Germany, the public employment service (Bundesagentur für Arbeit) helps SMEs set up vacancies and recruit new personnel. Additionally, SMEs can receive financial support for new recruits who do not yet meet the competence standards for their work. The Bundesagentur also publishes “Faktor A”, a magazine and podcast targeted at SMEs with articles focused on recruitment and training, leadership development, as well as the future of work.
The parent organisation of the Bundesagentur, the Federal Ministry of Labour and Social Affairs (Bundesministerium für Arbeit und Soziales), provides the “Leveraging the human factor programme” (unternehemensWert: Mensch, uWM), a consulting service funded via the European Social Fund that supports SMEs in developing long-term human resource strategies. uWM’s consulting services focus on personnel management, skills development, equal opportunities and diversity. Government grants cover up to 80% of the consulting costs for companies with fewer than 10 employees, while companies with 10‑249 employees can receive grants for up to 50% of their costs.
In Berlin, the Organisation for Education and Participation (Gesellschaft für Bildung und Teilhabe mbH or GesBiT) and the Senate Department for Integration, Labour and Social Affairs (Senatsverwaltung für Integration, Arbeit und Soziales) maintain an Office for Qualification Consulting for SMEs. The consulting services offer SMEs advice on assessing companies’ skills needs, available reskilling/upskilling opportunities, as well as recruitment and funding opportunities of which they may be unaware.
Source: Investitionsbank Berlin (2021[40]), 2021/2022 Business Support Guide, www.businesslocationcenter.de/fileadmin/user_upload/Wirtschaftsstandort/files/foerderfibel-en.pdf; OECD (2020[41]), Preparing the Basque Country, Spain for the Future of Work, https://doi.org/10.1787/86616269-en.
Employer representatives in Luxembourg should equally play a role in encouraging higher vacancy reporting rates via active awareness-raising. Members of the UEL (including associations, federations and confederations and chambers of commerce) could raise awareness about the importance of vacancy reporting in their engagement activities with employers (e.g. information sessions, newsletters, prominent featuring on their respective websites, etc.).
In order to further improve the assessment of skills needs, Luxembourg’s vacancy data could be further enriched by insights on skills and workforce needs collected by directly surveying employers. As outlined above, several surveys of employers’ current and future workforce needs have been carried out by several non-governmental stakeholders (e.g. ABBL, CC, CdM, FEDIL, etc.) in Luxembourg in recent years. CdM and FEDIL implement employer surveys regularly. However, due to the varying structure of questions, sector coverage, frequency of implementation and definition of skills/workforce needs, each survey provides only a piece-meal picture of employers’ skills needs. In addition, stakeholders for this review mentioned that too many surveys risk creating “survey fatigue” among employers in Luxembourg, given the size of the country and the administrative and time burden associated with completing each survey.
Going forward, Luxembourg should review the existing employer surveys, in tandem with administrative data on skills needs (e.g. ADEM and Skills-OVATE vacancy data; see Recommendation 4.2) to identify information that: 1) would add value to Luxembourg’s labour market data collection but is currently not available in administrative datasets; and 2) is collected by existing employer surveys irregularly, inconsistently or not at all. For example, while Luxembourg’s administrative datasets currently do not hold information on future skills needs, CdM and FEDIL surveys collect such insights but only cover three sectors (crafts, industry and ICT). The review of administrative data and existing employer surveys would help Luxembourg assess the need to introduce a national, cross-sectoral employer survey of skills needs and/or gaps. Such a national survey should reduce the need for multiple employer surveys in the long run and mitigate against the risk of “survey fatigue” in Luxembourg. Box 5.3 describes how the United Kingdom implements a nationwide Employer Skills Survey. In developing such a national employer survey, Luxembourg could follow CEDEFOP’s (2013[42]) “User guide to developing an employer survey on skill needs”, where relevant.
Box 5.3. Relevant international example: Nationwide employer survey
United Kingdom
Since 2011, the United Kingdom has carried out large-scale, cross-sectoral surveys of employers’ skills needs. The United Kingdom’s employer survey is one of the largest employer surveys in the world. The survey is essential for the work of the Department for Education in England (UK) and its partners within national and local government.
The latest completed iteration of the survey was carried out in 2019 via telephone interviews and asked employers questions about: 1) recruitment difficulties and skills lacking from applicants; 2) skills lacking from existing employees; 3) underutilisation of employees’ skills; 4) anticipated needs for skill development in the next 12 months; 5) the nature and scale of training, including employers’ monetary investment; and 6) the relationship between working practices, business strategy, skill development and skill demand.
Across different sectors, the survey enquired about employers’ skills needs at the level of occupations as well as specific skills (e.g. management, leadership, digital, people or specialist job-related skills).
The success of the survey depends on the willingness of employers to participate. More than 81 000 randomly selected employers across England, Northern Ireland and Wales took part in the survey in 2019. The underlying datasets are available through the UK Data Service by special licence access.
The next iteration of the survey, to be published in 2023, was commissioned to three market research agencies by the Department for Education (DFE) in England, the Welsh Government and the Northern Ireland Executive, and covers England, Wales and Northern Ireland. Again, the survey will be conducted via telephone interviews. If selected, employers can choose the time of their interview. Each employer participating in the survey will be asked if they would like to receive a summary report with the survey findings once the survey has been completed.
The survey results will be published on the gov.uk website in 2023.
Source: Government of the United Kingdom (2020[43]), Employer Skills Survey 2019, www.gov.uk/government/collections/employer-skills-survey-2019; Department for Education (2022[44]), Employer Skills Survey 2022, www.skillssurvey.co.uk/index.htm.
Finally, to further improve the granularity of Luxembourg’s labour market data collection, Luxembourg could benefit from tools that make it possible to link the information on occupations in vacancies declared by employers to specific skills. In the future, ADEM plans to move towards skills-based matching, where job seekers would be matched with occupations best suited to their skills, which could facilitate a closer match. However, ADEM’s current job-matching system does not yet facilitate skills-based matching. At present, ADEM matches job seekers to vacancies by matching the occupation (ROME) code associated with a specific vacancy, with the occupation (ROME) code indicated to be of interest by a job seeker, as well as according to language or education requirements, among other criteria.
Several OECD countries make it possible to link occupations and skills by using skills-based occupational classifications. For example, O*NET in the United States is a database containing detailed information about the knowledge requirements of more than 800 occupations. In Canada, the National Occupational Classification describes the world of work and occupations, including the skills required by each of the 500 occupational unit groups (OECD, 2016[27]). In 2021, Australia launched its own comprehensive skills classification, identifying around 600 skills profiles for occupations in the Australian labour market based on the O*NET database (National Skills Commission, 2021[45]). In 2017, the European Commission introduced the ESCO classification, which links occupations and skills relevant for the EU labour market and education and training (Box 5.4).
Box 5.4. Relevant international example: Skills-based occupational classification
The European Skills, Competences, Qualifications and Occupations classification (ESCO)
The European Skills, Competences, Qualifications and Occupations classification (ESCO) is the European multilingual classification of skills, competences and occupations.
ESCO offers a “common language” on occupations and skills that can be used by a wide variety of stakeholders to help connect people with jobs, employers and education and training providers and promote job mobility across Europe.
ESCO works as a dictionary, describing, identifying, and classifying professional occupations and skills relevant for the EU labour market and education and training. As of 2022, ESCO provided descriptions of 3 008 occupations and 13 890 skills associated with these occupations, translated into all official EU languages, as well as Icelandic, Norwegian, Arabic and Ukrainian.
For each occupation, ESCO defines an occupational profile, listing (among others) the skills considered relevant for this occupation. With respect to skills, ESCO distinguishes between skills, competences and knowledge. Skills refer to the “ability to apply knowledge and use know-how to complete tasks and solve problems”; competences are understood as the “proven ability to use knowledge, skills and personal, social and/or methodological abilities, in work or study situations and in professional and personal development”; and knowledge is defined as “body of facts, principles, theories and practices that is related to a field of work or study”. For example, working as a civil airline pilot requires the competence to combine knowledge on emergency procedures and equipment malfunctions with skills on reading position coordinates and following the air route.
ESCO is free and available as linked open data.
Source: European Commission (2022[46]), What is ESCO, https://esco.ec.europa.eu/en/about-esco/what-esco; European Commission (2022[47]), ESCOpedia, https://esco.ec.europa.eu/en/about-esco/escopedia/escopedia.
In Luxembourg, no common, comprehensive skills classification is currently used. Links between occupations and specific skills in Luxembourg's labour market have been defined only for occupations in certain sectors and a non-coordinated manner. In some sectors, the content of occupations is described in collective agreements. For example, the collective agreement covering the banking sector describes the tasks, knowledge and technical skills, among others, required for occupations (ABBL, ALEBA, OGBL and LCBG-SEFS, 2018[48]). In addition, some sectors, such as crafts (CdM, n.d.[49]), have developed more detailed skills profiles for their occupations based on research carried out in working groups or through labour market monitoring (ADEM/UEL, 2021[23]). The limitation of the skills classifications developed at the sectoral level is that they are not linked to a recognised national or international occupational classification (e.g. ISCO), which confines their potential use to the internal needs of the sector. Moreover, the definition of “skills” used by the different classifications varies between sectors, while most sectors in Luxembourg have not developed skills classifications at all (ADEM/UEL, 2021[23]).
Stakeholders in Luxembourg have agreed that rather than developing its own common skills classification, ADEM should consider leveraging and adapting existing and internationally interoperable skills classification for classifying its own vacancy data. The ESCO classification (Box 5.4), designed to reflect the specificities of the EU labour market, might be particularly well-suited to Luxembourg's needs. The ESCO skills classification permits establishing direct links with occupational information classified according to the ISCO classification or indirect links with classifications that map onto ISCO (such as ROME), which are used in many EU member states (European Commission, 2021[50]), including Luxembourg.
ADEM would not be the first public employment service to adopt ESCO. In 2018, Iceland became the first country to adopt ESCO on a national level when the public employment service started using ESCO to revamp its job board (Box 5.5). To facilitate skills-based matching, ESCO is now being used by PES in seven countries (Albania, Finland, Greece, Iceland, Ireland, Israel and Malaysia) (European Commission, n.d.[51]). Most recently, ESCO has been adopted by Greece (Box 5.5). Beyond adopting ESCO at ADEM to support skills-based matching, ESCO could be equally used to help systematically define learning outcomes of adult learning courses, which would have important benefits for Luxembourg’s education and training data collection (see the section below on education and training data; also see Recommendation 1.3 in Chapter 2).
Box 5.5. Relevant international example: Adopting the ESCO classification by a public employment service
Iceland
Iceland became one of the first countries to introduce the ESCO classification at a national level.
In 2018, ESCO started being used by the public employment service to revamp job platforms in the country, hitherto relying on the ISCO‑88 classification for job matching.
Using the ESCO classification has helped Iceland to more effectively link employers with the right job applicants. As a result, employers are able to better communicate their needs and job seekers better able to see what exactly employers are looking for in applicants.
The use of ESCO also facilitates that job vacancies are described more accurately, thanks to the provision of multilingual information on the knowledge, skills and competences required for the job. The multilingual feature of ESCO has been particularly useful for Iceland, given that more than 10% of Iceland’s labour force is foreign-born.
Greece
The Greek Public Employment Service, OAED (Manpower Employment Organisation), adopted ESCO in 2021. It is the first Greek organisation to adopt ESCO.
To adopt ESCO, OAED has implemented the “MATCHING” (Mapping Arrangements and Transcoding Change In Greece) project, co-funded by the European Programme for Employment and Social Innovation (EaSI). The project included training for OAED’s employment counsellors (among others), peer learning from other relevant national and European organisations, as well as undertaking awareness-raising activities about ESCO.
ESCO is intended for use by: 1) OAED’s staff members to analyse the profiles of job seekers and details of job vacancies through OAED’s Integrated Information System; and 2) OAED’s clients (job seekers and employers) through OAED’s e-Services. Job seekers will be able to annotate their curriculum vitaes (CVs) with ESCO skills, and employers will be able to more accurately describe their job offers.
Source: European Commission (2018[52]), How ESCO supports the Public Employment Service of Iceland, https://audiovisual.ec.europa.eu/en/video/I-162745; OAED (2021[53]), OAED adopts the new ESCO classification of occupations and skills, www.oaed.gr/storage/oaed-logo/enimerotiko-entipo-esco-matching-en.pdf; European Commission (2022[54]), The adoption of ESCO by the Greek Public Employment Service (OAED), https://esco.ec.europa.eu/en/news/adoption-esco-greek-public-employment-service-oaed.
However, certain caveats will need to be kept in mind. For example, stakeholders have expressed concerns about the extent to which ESCO accurately reflects the skills requirements of certain occupations (especially occupations heavily reliant on digital skills). Nonetheless, for comparison, the inclusion of digital skills in the O*NET classification is even more limited. In fact, digital skills covered by the O*NET classification are limited to: 1) knowledge of computers and electronics; and 2) programming skills (Lassébie et al., 2021[55]). Stakeholders in Luxembourg have broadly agreed that despite certain shortcomings, ESCO was the most suitable skills classification, which could be adapted for use in Luxembourg’s context.
Recommendations
4.1 Improve the accuracy of occupational social security data by creating targeted incentives for employers, strengthening existing guidance tools for identifying the correct occupational codes, and conducting targeted awareness raising. IGSS and CCSS could design an online occupations dashboard using CCSS data, where employers could track the occupational, salary and/or demographic structure and evolution within their company. CCSS and ADEM could also work on creating links between CCSS employment entry declarations and ADEM vacancy declarations, whereby parts of CCSS declarations could be automatically pre-filled based on ADEM declarations to make the reporting process more straightforward for employers. Further, CCSS should improve the search functions of the existing occupational code search engine (e.g. by including descriptions of tasks related to occupations, developing a chatbot for real-time responses to queries, etc.) and consider making the search engine accessible in English, in addition to French. IGSS and CCSS should feature the existing guidance tools (i.e. the proposed occupations dashboard, occupational code search engine, etc.) more prominently on their respective websites and, in collaboration with employer representatives (e.g. UEL), design a regular reminder outlining the “common mistakes” of, and existing guidance tools for, identifying the correct occupational codes. In addition, IGSS and CCSS could expand their awareness-raising activities to include associations of HR professionals or payroll experts to target actors directly in charge of employment entry declarations. In turn, employer representatives (e.g. members of the UEL) should equally underline the importance of correctly declaring new hires to the CCSS in their own engagement activities with employers. In addition to undertaking efforts to strengthen the accuracy of the CCSS data, IGSS and other stakeholders in Luxembourg should still make use of the data to the extent possible (e.g. for the analysis of general growth trends across occupations).
4.2. Explore the possibility of including ADEM’s vacancy data in CEDEFOP’s Skills-OVATE tool. With ADEM’s vacancy data now published on a publicly accessible job board, CEDEFOP (via LISER) should explore the possibility of including ADEM data in CEDEFOP’s Skills-OVATE tool. As a result, it would be possible to regularly obtain information on occupational and skills trends in Luxembourg based on vacancy data from both ADEM’s own job board and online private job boards. The risk of counting certain vacancies twice (i.e. vacancies declared both to ADEM and on private job boards) would be minimised through CEDEFOP tools developed specifically for such a purpose.
4.3. Increase the share of job postings created and reported by employers to ADEM, by designing targeted incentives and services for employers, especially SMEs, and through employer-led awareness raising. ADEM could develop an online vacancy dashboard based on vacancy data declared to ADEM, which could provide employers with an overview of their vacancies (open and filled) over time and how quickly they are being filled. ADEM could also consider developing dedicated support services for SMEs (e.g. support with drafting job postings according to SMEs’ needs, sign-posting the available reskilling/upskilling opportunities for SME employees, etc.), whose capacities to engage in vacancy reporting tend to be restricted the most. Employer representatives (e.g. members of the UEL) should also further raise awareness about reporting vacancies to ADEM in their engagement activities with employers, in addition to highlighting the importance of correctly declaring new hires to CCSS (see Recommendation 4.1).
4.4. Review existing employer surveys, together with administrative data on skills needs to help assess the need for a national employer survey. Luxembourg should review the existing employer surveys carried out by non-governmental actors, in tandem with reviewing administrative data on skills needs (e.g. ADEM and Skills-OVATE vacancy data; see Recommendation 4.2) to identify information which: 1) would add value to Luxembourg’s labour market data collection but is currently not available in administrative datasets; and 2) is collected by existing employer surveys irregularly, inconsistently or not at all. The review of administrative data and existing employer surveys would help Luxembourg assess the need for introducing a national, cross-sectoral employer survey of skills needs and/or gaps. Such a national survey should reduce the need for multiple employer surveys in the long run and mitigate against the risk of “survey fatigue” in Luxembourg.
4.5. Adopt a skills-based occupational classification to link occupations to skills. ADEM should adopt a comprehensive and internationally recognised skills classification to facilitate matching job seekers to vacancies. For this purpose, ADEM could consider the use of ESCO, which links occupations to specific skills. ESCO could be equally used to help systematically define learning outcomes of adult learning courses, which would have significant benefits for Luxembourg’s education and training data collection (see Recommendation 1.4 and Recommendation 4.8).
Expanding the range and strengthening the granularity of Luxembourg’s education and training data
Luxembourg’s skills data collection includes a variety of education and training data that provides valuable information on Luxembourg’s skills supply (see Table 5.1 and Table 5.2). Going forward, stakeholders consulted during this review agreed that Luxembourg’s education and training data collection could be further expanded, and the granularity of the existing data further strengthened.
Data on education and training outcomes help shed light on the alignment of the education and training offer and the demands of the labour market (OECD, 2016[27]), helping policy makers to better tailor the education and training offer to labour market needs while supporting individuals in making informed skills choices (see Chapter 3 for an extended discussion on guiding and incentivising skills choices in Luxembourg). The key opportunities for strengthening Luxembourg’s data on education and training outcomes exist at the levels of higher education (HE) and adult learning.
At the HE level, graduate tracking is overseen and implemented by the University of Luxembourg. UoL is Luxembourg’s only public HE institution, with the UoL student population representing 99% of the total student population in higher education in Luxembourg (Eurostat, n.d.[56]).5 Since 2018, UoL has been collecting data on graduates’ (former master's and bachelor’s students) outcomes. UoL uses a graduate survey, implemented six months after graduation, asking graduates about their employment status, earnings, perceived skills mismatch, and use of skills on the job, among others. The survey is extended to all graduates (former master's and bachelor’s students). Around 30% of contacted graduates typically respond. As of 2022, UoL planned to include PhD graduates in the survey as well.
In 2020, UoL expanded its graduate tracking exercise by conducting a one-off “employment study” of graduates’ employment pathways using LinkedIn data. Based on LinkedIn data and anecdotal evidence from graduates’ supervisors, UoL was able to collect information for approximately 55% of its graduates (former master's and PhD students) having graduated between 2015 and 2019. UoL assembled a team of consultants who manually checked graduates’ LinkedIn profiles, while academic supervisors provided insights on graduates' employment pathways based on their knowledge. The evidence from academic supervisors was checked via desk research and only used if validated. Information on bachelor’s students was not collected due to time and resource constraints.
While UoL’s graduate tracking efforts are steps in the right direction, more could be done. For example, current graduate tracking efforts do not make it possible to gather data on how long it takes HE students to find a job post-completion of their studies, which could be useful (not only) to better facilitate the transitions of former international students into the labour market (see Chapter 4). It is also important to note that UoL’s survey results rely on self-reported information by graduates, while the survey’s response rate could be further increased. Moreover, the use of LinkedIn data for the purposes of the employment study can impact the representativeness of the graduate tracking results. Evidence suggests that low‑income individuals and those outside of the workforce tend to be under-represented on LinkedIn, while LinkedIn tends to be used frequently by individuals in knowledge-intensive sectors such as management, marketing, HE and consulting (Blank and Lutz, 2017[57]; van Dijck, 2013[58]). In addition, via the employment study, UoL gathered information on graduates’ first and current job titles and employers through LinkedIn, missing out on collecting other valuable data, such as information on earnings. As of early 2022, UoL does not plan to repeat the employment study.
Going forward, UoL could consider using administrative data for the purposes of graduate tracking and combining it with existing survey data (OECD, 2019[59]). Administrative data can provide a longitudinal view of graduates’ outcomes and could provide evidence on the time graduates take to transition into the labour market. Box 5.6 details how England (United Kingdom) combines administrative and survey data in its graduate tracking of HE graduates. In Luxembourg, administrative data (e.g. CCSS data) could provide comprehensive information (e.g. on the employment status, sector and earnings of graduates),6 at least for graduates who have remained in Luxembourg following the completion of their studies.7
Presently, the use of administrative data for graduate tracking in Luxembourg is precluded by the lack of necessary legal basis, which is required under the EU General Data Protection Regulation (GDPR). However, in the context of the development of the NDEP (see Opportunity 2), Luxembourg is aiming to grant the necessary legal basis to the NDEP, thereby making the use of administrative data possible for graduate tracking purposes.
Should NDEP be capable of establishing connections to administrative databases of other countries, especially in the Greater Region (see Recommendation 4.14), the collection of information about outcomes of HE students receiving financial aid from the Government of Luxembourg but studying outside Luxembourg8 or graduates who had studied in Luxembourg but left the country after the completion of their studies, could equally be facilitated.
To complement administrative data with information that is not available in administrative databases (e.g. how useful the university experience was in preparing graduates for the labour market, skills used on the job, etc.) (OECD, 2019[60]), UoL could link administrative data to its existing graduate survey by using unique identifiers. It could also consider expanding the survey to gather information from former international students about reasons for leaving Luxembourg post-completion of their studies in order to support efforts to better foster the transition of former international students into Luxembourg's labour market (see Chapter 4). LinkedIn data could be considered as an additional source of information to cross‑check the UoL graduate survey and administrative data and fill potential gaps in the latter two.
Box 5.6. Relevant international examples: Higher education graduate tracking
England (United Kingdom)
England tracks the outcomes of HE graduates by combining administrative and survey data.
The Longitudinal Employment Outcomes (LEO) dataset links school census, Further Education, Higher Education Statistics Agency and Her Majesty’s (HM) Revenues and Customs data to obtain information on graduates’ employment and earnings. From 2019/20, a Graduate Outcomes Survey (GOS) of HE graduates was commissioned to complement LEO. LEO and GOS are linked via unique identifiers, creating a merged dataset.
Conducted 15 months after graduation, GOS is a census survey that complements LEO by collecting self-reported, qualitative information from graduates about their job roles (e.g. their duties and responsibilities) and the extent to which they use the skills developed during their studies. GOS is managed by the Higher Education Statistics Agency and is carried out by a combination of online and telephone interviews. Higher education institutions (HEIs) support GOS implementation by providing up-to-date contacts of their graduates.
GOS contains core and optional questions. The core questions cover: 1) job title and duties, including supervision responsibilities; 2) how the current job was found; 3) what further study have they conducted since graduation; 4) skills from their course that they are using in their work; and 5) current activities and future plans. HEIs can include optional questions in GOS for a fee.
HEIs can access the raw survey data for business planning purposes and to evaluate HEI teaching standards in the Teaching Excellence Framework in England.
Source: European Commission (2020[61]), Graduate tracking - A "how to do it well" guide, https://op.europa.eu/en/publication-detail/-/publication/5c71362f-a671-11ea-bb7a-01aa75ed71a1/language-en/format-PDF/source-search.
While there is space for improving Luxembourg’s data on education and training outcomes of HE graduates (see above), there are no measures in place for tracking the outcomes of graduates from adult learning, either publicly or privately provided.
Tracking the outcomes of adult learners is a challenge in many OECD countries. Nonetheless, several relevant international examples could inspire Luxembourg. In Denmark (Box 5.7), data on all publicly provided adult learning courses are collected in a centralised course register and can be combined with administrative data thanks to a unique identifier number (similar to the identifier number of Luxembourg’s National Registry of Physical Persons), making it possible to gain detailed information on the course participants’ outcomes. In Portugal, accredited providers (both public and private) of adult learning courses delivered in the framework of Portugal’s National Qualifications System (SNQ) are required to register the training activities of individual learners in the Integrated Information and Management System for Educational and Training Supply (SIGO). SIGO issues a certificate upon training completion to the learner and automatically registers training information in the learner’s own educational and training record (the Qualifica Passport) (Box 5.7). Luxembourg could create a centralised training register similar to Denmark’s and Portugal’s. To obtain information on learners’ outcomes following the conclusion of training, information in Luxembourg’s centralised course register could be linked to administrative data (e.g. CCSS data on employment status, sector, earnings, etc.) through the NDEP (see Opportunity 2). When initiating the tracking of outcomes of adult learners, Luxembourg could restrict the exercise to collecting data on participants in continuous professional development (formation professionnelle continue) and professional retraining (reconversion professionnelle) courses, rather than personal development (développement personnel) courses, as the latter are not of direct labour market relevance.
After exploring the use of administrative data for the purposes of adult learning graduate tracking, should Luxembourg wish to collect additional information not available in administrative datasets (e.g. job progression, use of skills on the job, etc.), a survey of adult learning outcomes could be considered. For example, France implements a survey of adult learners’ outcomes every year (Box 5.7). In Luxembourg, INFPC and STATEC could collaborate on designing and implementing such a survey.
Box 5.7. Relevant international example: Adult learning graduate tracking
Denmark
In Denmark, data on publicly provided adult and continuing education are collected centrally in the “Cross-sectional course register”, managed by Statistics Denmark (Denmark’s national statistical office). Data in the register are collected via the civil registration number (CPR number) to monitor the population’s participation in adult and continuing education. In addition, combining the register with other administrative data makes it possible to obtain detailed information on the outcomes of course participants.
Statistics Denmark makes the register data available via its online databank and decides which data should be published. The possibility also exists to buy more detailed datasets or to access the microdata via paid access, or via researcher or ministerial agreements. These agreements outline the data security rules that must be followed. For example, data disaggregation may be limited to avoid identifying specific individuals.
In order to improve co‑ordination and collaboration between Statistics Denmark and data users, Denmark’s Contact Committee for education statistics organises annual or biannual meetings where representatives from selected ministries and non-governmental stakeholders can provide feedback to Statistics Denmark on data quality and any related issues.
Portugal
In Portugal, the Integrated System for Information and Management of Educational and Training Offer (SIGO) is co‑ordinated by the Directorate-General for Education and Science Statistics (DGEEC) to help manage the education and training offer for adults and young people. SIGO collects information from all accredited adult learning providers (public or private) providing courses under the National Qualifications System, whether their qualifications are part of the National Qualifications Catalogue or not. Information is collected on the training activities of individual learners and the number of learners who access and complete learning programmes. When a training activity is completed, a certificate is issued by SIGO that is automatically registered in the learner’s Qualifica Passport. The Qualifica Passport is a user-oriented online tool and platform that provides information on individuals’ own educational and training records and directs learners to potentially relevant learning opportunities based on the qualifications they have already acquired.
While SIGO has not been designed for the sole purpose of tracking adults’ outcomes, in the context of continuous improvement of SIGO, Portugal has started working on establishing links between SIGO data and social security data to be able to track adults’ outcomes in the future.
France
Every year, the Ministry of Labour, Employment and Economic Inclusion of France carries out the “Survey on the future of candidates with professional qualifications” (Enquête sur le “Devenir des candidats aux titres professionnels”). The objective of the survey is to generate quantitative and qualitative information on the education and training outcomes of job seekers who had completed a course leading to professional qualification. The survey is carried out six months after the qualification completion. The survey inquires about the former course participants’ employment status (current and in the past six months), nature of the employment contract (i.e. full-time vs part-time work) where relevant, relevance of the training completed for the individuals’ job, as well as the use of skills acquired during training. In 2019, the survey collected responses from more than 35 000 respondents, with the survey results being considered statistically significant.
Source: OECD (2021[62]), Strengthening Quality Assurance in Adult Education and Training in Portugal: Implementation Guidance, www.oecd.org/portugal/Strengthening-Quality-Assurance-in-Adult-Education-and-Training-in-Portugal-Implementation-Guidance.pdf; OECD (2018[63]), Skills Strategy Implementation Guidance for Portugal: Strengthening the Adult-Learning System, https://dx.doi.org/10.1787/9789264298705-en; OECD (2018[64]), OECD Reviews of School Resources: Portugal 2018, https://dx.doi.org/10.1787/9789264308411-en; European Commission (2018[65]), Mapping of VET graduate tracking measures in EU Member States, https://op.europa.eu/en/publication-detail/-/publication/00d61a86-48fc-11e8-be1d-01aa75ed71a1; Ministry of Labour (2019[66]), Annual report 2019 - Survey on the future of candidates with professional qualifications [Rapport annuel 2019 - Enquête sur le “Devenir des candidats aux titres professionnels” ”], https://travail-emploi.gouv.fr/IMG/pdf/rapport_devenir_candidats_2019.pdf.
Finally, Luxembourg’s education and training data collection would benefit from more granular and structured information on skills developed in adult learning courses. Clearly outlining the skills that individuals can expect to gain from a specific course and defining the skills using a “common language of skills” (i.e. following a common skills classification) would facilitate the interoperability of modular training across adult learning providers (see Chapter 2); support adult learning providers in requesting the alignment of non-formal qualifications with one of the levels of the Luxembourg Qualifications Framework (Cadre luxembourgeois des qualifications, CLQ) (MESR/MENJE, 2020[67]); and help develop automatic tools for guiding individuals towards the training courses most suitable to their needs. For example, the Luxembourg Institute of Science and Technology (LIST) has developed a skills e-assessment tool, Cross-Skill, allowing individuals to identify the skills gaps they would have to fill to move between different jobs (LIST, 2022[68]). Should LIST have access to a catalogue of training courses with learning outcomes defined in terms of specific skills, the results of such an e-assessment could be used to suggest to individuals the training courses best matched to fill their skills gaps.
Luxembourg’s online portal for lifelong learning (lifelong-learning.lu at www.lifelong-learning.lu/Accueil/en), managed by the INFPC, offers adult learning providers the opportunity to register on the portal and present their courses to the public. All providers registered on the lifelong-learning.lu portal should define the “objectives” (i.e. learning outcomes) of their courses. However, the level of granularity with which adult learning providers describe the skills that individuals are expected to develop in their courses varies widely, as there is no common definition of skills that providers follow in their descriptions.
Going forward, Luxembourg’s adult learning providers should be incentivised to upload their training offer and describe its learning outcomes on the lifelong-learning.lu portal. Accreditation of adult learning courses could be made conditional on publicising the training offers on lifelong-learning.lu, and on describing the course learning outcomes in a non-structured manner (i.e. not having to follow a skills-based occupational classification) yet in sufficient detail (e.g. by setting a minimum word count requirement) (see Chapter 2). Luxembourg could then consider developing an information technology (IT) tool, integrated on lifelong‑learning.lu, which could use text mining and natural language processing (NLP) techniques to help providers articulate the skills developed in their programmes in a structured manner. The IT tool would analyse the non-structured descriptions of course learning outcomes specified by providers and suggest skills best suited to reflect the course content, which the providers could check and validate. The generated skills suggestions would appear on the lifelong-learning.lu portal, together with the non-structured descriptions. The generated skills suggestions should follow a common, internationally recognised skills‑based occupational classification. For example, should ADEM adopt the ESCO classification for classifying its own vacancy data (see Recommendation 4.5), ESCO would be well‑suited to help adult learning providers annotate the learning outcomes of their courses with specific skills, too.
Recommendations
4.6. Consider combining graduate surveys in higher education with administrative data. The University of Luxembourg should consider the feasibility of using administrative data (e.g. CCSS data) for the purposes of graduate tracking. Luxembourg’s foreseen NDEP (see Opportunity 2) would provide the necessary legal basis for linking administrative datasets. UoL’s existing graduate survey could then be used to gather information not available in administrative databases. The survey and administrative data could be combined by using unique identifiers. UoL’s existing survey could also be expanded to ask graduates about reasons for leaving Luxembourg post-completion of their studies to support efforts seeking to better foster the transition of former international students into Luxembourg's labour market (see Chapter 4). In the long run, the possibility of creating linkages between administrative datasets within and beyond Luxembourg via NDEP (see Opportunity 2) would facilitate collecting information about outcomes of HE students receiving financial aid from the Government of Luxembourg but studying outside of Luxembourg and graduates who had studied in Luxembourg but left the country after the completion of their studies.
4.7. Initiate tracking of adult learners’ outcomes by creating a centralised training register interoperable with administrative data and by considering introducing a nationwide survey of adult learners’ outcomes. Luxembourg should start by creating a training register centralising data on participants in adult learning courses (continuous professional development and professional retraining courses) of accredited public and private providers. The data from the centralised training register could be linked through NDEP (see Opportunity 2) to administrative labour market data (e.g. CCSS data) in order to obtain information on learners’ outcomes following course completion. To complement the information gathered via administrative data (e.g. employment status, sector, earnings, etc.), Luxembourg could consider introducing a nationwide survey of adult learners’ outcomes to gather additional qualitative information (e.g. on job progression, use of skills developed on the job, etc.). INFPC and STATEC could collaborate in designing and implementing such a survey.
4.8. Incentivise adult learning providers to describe the learning outcomes of their courses in sufficient detail and develop an IT tool helping to extract structured information on skills from the learning outcomes’ descriptions. Adult learning providers in Luxembourg should be incentivised to upload their training offer on the lifelong-learning.lu portal, where they should describe and update descriptions of the learning outcomes of their courses in a non-structured manner (i.e. not having to follow a skills-based occupational classification) yet in sufficient detail (e.g. by setting a minimum word count requirement). Certification of adult learning providers with a quality label could be made conditional on publicising the training offer on lifelong-learning.lu, and on providing non-structured descriptions of the course learning outcomes (see Chapter 2). INFPC could then develop an IT tool that could be integrated on the lifelong-learning.lu portal and use text mining and NLP to help providers articulate the skills developed in their programmes in a structured manner. The IT tool would make skills suggestions to adult learning providers by analysing the non-structured learning outcomes descriptions of their courses. Adult learning providers could check and validate the suggestions made by the IT tool. The skills suggestions made by the IT tool would appear on the lifelong-learning.lu portal, together with the non-structured descriptions. The skills suggestions made by the IT tool should follow a common, internationally recognised skills-based occupational classification. The use of the ESCO classification could be considered for this purpose should it be adopted by Luxembourg’s labour market institutions (e.g. ADEM) (see Recommendation 4.5).
Opportunity 2: Strengthening co‑ordination of, and synergies between, skills data within and beyond Luxembourg
A strategic and co‑ordinated approach to collecting and managing public and even private data are important for taking full advantage of the social, scientific, economic and commercial potential of data resources. Skills data (i.e. labour market and education and training data, among others) is no exception. All too often, skills data, whether public or private, are collected in an uncoordinated manner by the various actors gathering data for their own internal needs. However, data are of limited value if they cannot be cross-referenced with other data from different sources. The usefulness of collecting data could be further enhanced by integrating them with different types of data sources. If skills data are collected and used in silos because their existence is not known or their accessibility is limited, the return on investment made in generating these data is suboptimal.
Data producers, including governmental institutions, can foster synergies between existing skills data by making relevant datasets available for use within the government or for free use by non-governmental stakeholders through "open data" or existing data platforms. However, several barriers need to be addressed to facilitate effective data co‑ordination and exchange. Such barriers include: the lack of data collected based on international data standards (semantic interoperability); the lack of standards and technological solutions for data interoperability (technical interoperability); the lack of data documentation; the lack of metadata; the lack of reusability; the lack of tools to assess and optimise the quality of available data; difficulties in assessing the value of datasets; the risk of losing competitive advantage; and intellectual property risk or risk of privacy violation (OECD, 2021[69]; European Environment Agency, 2019[70]).
Stakeholders in Luxembourg have highlighted that Luxembourg lacks a co‑ordinated approach to skills data collection and management. There is low awareness of all the existing skills data sources, as well as their scope and accessibility criteria, which limits the extent to which they can be fully leveraged to inform the design and implementation of Luxembourg’s skills policies. Efforts for improving Luxembourg’s skills data collection and use similarly often occur in an uncoordinated fashion. In addition, skills data exchanges between government institutions and stakeholders remain limited, not least because laws and technical infrastructure allowing for such exchanges have yet to be introduced.
Moreover, Luxembourg is not taking full advantage of international skills data sources, especially those of neighbouring countries. However, exploring synergies with international data sources seems crucial for Luxembourg, given the cross-border nature of its labour market, and could help Luxembourg assess the potential skills supply and demand in the Greater Region.
Luxembourg could better foster co‑ordination and synergies between skills data within and beyond its national borders by:
strengthening the co‑ordination of labour market and education and training data flows within government and with stakeholders
building synergies between Luxembourg’s and neighbouring countries’ data sources to improve the skills data availability for the Greater Region.
Strengthening the co‑ordination of labour market and education and training data flows within government and with stakeholders
Luxembourg does not currently have a single government body with a specific mandate to collect data on skills and co‑ordinate all the actors that collect data on skills demand and supply. The institutions working on skills data currently operate with different working definitions, methods and practices that reflect their different mandates.
In Luxembourg, co‑ordination can be improved between four existing “levels of institutions” involved in the skills data collection process: 1) the national statistical office (STATEC) that has the authority to access and co‑ordinate all data in the country; 2) the institutions (ADEM, CCSS/IGSS and INFPC) that collect key data on the labour market and training (with limited interoperability); 3) RETEL, which commissions new, and centralises existing, labour market research in Luxembourg (with limited capacity to carry out its role and collaborate with other institutions); and 4) stakeholders (e.g. CC, CdM, federations, private training institutes, etc.) that are collecting information on training participation in their own training courses, or employers’ skills needs.
In Luxembourg, STATEC co‑ordinates national reporting efforts with respect to labour market indicators (e.g. Eurostat, LFS) but does not have a specific mandate to collect data on existing skills in Luxembourg or the Greater Region area or companies’ skills needs. STATEC has the authority to access any government data sources within Luxembourg as long as it operates within GDPR boundaries. However, STATEC does not compile skills data collected by stakeholders (e.g. on training participation courses organised by stakeholder associations, on employers’ skills shortages, etc.). Such stakeholder data are not currently centralised nor accessible for use by government authorities or other stakeholders. Furthermore, there is no central strategy, mandate or tool to assemble different data sources, streamline different data flows or improve skills data interoperability. In view of this, Luxembourg may be inspired by Norway, whose Future Skills Needs Committee supports effective skills data co‑ordination and analyses (see Box 5.8).
Box 5.8. Relevant international example: Skills data co-ordination and analysis
Norway
The purpose of Norway’s Future Skills Needs Committee (Kompetansebehovsutvalget) is to provide the best possible evidence-based assessment of Norway’s future skills needs to support labour market authorities' strategic planning and decision making.
More specifically, it is tasked to “generate and organise the evidence base” on Norway’s future skills needs based on already available data and “stimulate the development of new evidence”. For this purpose, the Future Skills Needs Committee is also able to fund its own research via its budget financed by the Ministry of Education. Concerning the types of data used, the Future Skills Needs Committee is expected to use a variety of both qualitative and quantitative data sources.
The Future Skills Needs Committee is comprised of representatives from social partners (employers and employees), the involved ministries, as well as experts, and is headed by a renowned economist (19 members, plus 3 in the Secretariat).
Since its creation in 2017, the work of the Future Skills Needs Committee has largely been based on national and international research and has resulted in three official reports. With the new mandate covering the period 2021‑27, the Future Skills Needs Committee will not produce official reports but will target its work on specific topics with analyses and assessments instead.
Source: OECD (2020[71]), Strengthening the Governance of Skills Systems: Lessons from Six OECD Countries, https://dx.doi.org/10.1787/3a4bb6ea-en.
The ADEM/UEL Trends working group mentioned above is a promising attempt at establishing an informal platform to think collectively about ways to improve information on skills and labour market data through existing stakeholder data sources. This working group, which relies on stakeholder participation, aims to co‑ordinate efforts to develop and use skills data (both government and stakeholder) in Luxembourg. It includes representatives from ADEM and the UEL covering several sectors (ABBL for finance, FEDIL for construction and ICT, etc.) and seeks to gain an overview of the various data sources that provide information on skills needs and identify ways to improve these data sources.
The partnership between ADEM and UEL involves stakeholders in developing data on skills demand and supply in Luxembourg. These institutions are mapping existing data sources but have not yet assigned specific responsibilities for data collection and do not yet have a common objective in this regard. In addition, relevant stakeholders like CCSS, MENJE, MESR and STATEC are not currently involved in this project, suggesting that a broader government approach would be desirable.
Similarly, Luxembourg would benefit from an increase of data interoperability and could clarify institutions’ roles in data collection. Currently, skills data exist and are collected in a decentralised way; thus, their potential is underutilised. A government mandate, objectives and other issues related to the co‑ordination could be clarified and formalised with a skills data charter (or national skills data strategy), as in the cases of the United States Department of Education or the Worldwide Initiatives for Grantmaker Support (WINGS) philanthropic network (see Box 5.9). While many government services and stakeholders actively promote and develop data collection on skills demand and supply, co‑ordination can be improved. This is due to the lack of a clear government mandate on skills data collection and some shared objectives for the relevant institutions.
Box 5.9. Relevant international example: Data strategies and charters
United States - Department of Education Data Strategy
This United States (US) Department of Education (DoE) recognised the strategic and critical need for consistent governance and management of its data assets, in line with the 2018 Foundations for Evidence-Based Policymaking Act, the 2019 Federal Data Strategy, and the Federal Data Strategy 2020 Action Plan.
As a key milestone under Action 2 ("Institutionalize Agency Data Governance") of the 2020 Action Plan, the DoE established the inaugural Department of Education Data Strategy in December 2020, through its Data Governance Board. The purpose, expectations, roles and responsibilities, and procedures governing the work of the Data Governance Board are outlined in the Data Governance Board Charter. The Department of Education Data Strategy describes the department's vision for accelerating progress toward becoming a data-driven organisation and fully leveraging the power of data to advance the department's mission of ensuring equal access and fostering educational excellence for the nation's learners.
United States - Global Philanthropy Data Charter
Worldwide Initiatives for Grantmaker Support (WINGS) is a network of foundations and philanthropy networks. In 2012, the US Foundation Centre and WINGS launched the Global Philanthropy Data Charter, which was intended as a framework to guide organisations in the sector as they set out to improve philanthropy data, acknowledging the diversity of context, culture and legal framework within which these organisations operate.
WINGS and the US Foundation Centre noticed that many foundations were collecting data on their activities but in different and unstructured ways. The US Foundation Centre worked on improving the data governance in the philanthropic sector in the United States, and similarly, the WINGS network worked globally by introducing an international data charter to standardise the existing information collected by private foundations on individual private grants and improve data interoperability.
Source: US Department of Education (2019[72]), Data Governance Board Charter, www.ed.gov/sites/default/files/cdo/dgb-charter.pdf; US Department of Education (n.d.[73]), Foundations for Evidence-Based Policymaking, www.ed.gov/data; OECD (2018[74]), Private Philanthropy for Development, https://dx.doi.org/10.1787/9789264085190-en.
A national skills data charter, intended as a formal strategic document able to support the national co‑ordination of skills data collection, could clearly set out objectives to improve Luxembourg’s skills data governance going forward. The charter could allocate roles and responsibilities to the different key actors in Luxembourg’s skills data system for improving Luxembourg’s skills data governance while leveraging potential synergies. The charter could also be complemented by an action plan that covers specific time periods, sets out the necessary actions to be taken in the short term and envisages those to take in the medium term.
Data exchanges seem to take place only to comply with national and international reporting obligations and on a case-by-case basis to tackle specific coverage or granularity issues. For example, IGSS exchanged data with ADEM and STATEC to learn about their respective issues related to data coverage (see Opportunity 1). However, many data exchange exercises have to be supported by specific agreements (or memoranda of agreement, MOUs) to comply with GDPR restrictions, which limits the possibility of exchanging data even between government agencies. These rules, therefore, limit data exchange between public authorities. The matter could be simplified if data requests by the same agency could be grouped together or if the rights to access a certain type of data could be valid for a certain period of time or could be automatically granted to a requesting institution if the data request matches certain pre-specified requirements. This latter option to simplify bureaucracy would be equivalent to establishing “institutions’ data passports”.
Recently, the Ministry of Higher Education, together with five other ministries, launched a project initiative to set up a common National Data Exchange Platform (NDEP), covering all data sectors and going beyond skills. This project materialised in the form of an Economic Interest Group (EIG) in July 2022. The recruitment of human resources has started, and it is expected that the structure will become operational in the coming months, hence moving the co‑ordination of national data production in the right direction.
The NDEP project involves six main government entities (MESR, Ministry of State’s Department of Media, Connectivity and Digital Policy, Ministry for Digitalisation [MinDigital], Ministry of Finance, Ministry of the Economy and Ministry of Social Security’s IGSS) as well as the Luxembourg Institute for Health (LIH) and LISER. It is also planned that UoL and LIST will join in the course of 2022. The NDEP project aims to set up a broad data exchange environment that provides the IT infrastructure and legal basis to facilitate data sharing between government agencies while complying with GDPR requirements. This intragovernmental partnership is currently working on refining the list of participating authorities, setting out the list of priority fields that would benefit from this platform (the top priority is health, followed by energy data) and developing the legal framework. The Ministry of State is involved, as it is responsible for the publication of the government’s open data, while the other government partners have transversal responsibilities related to data governance. Other stakeholders, such as STATEC, although not part of the governance structure, will be involved in the operational activities.
The NDEP is a promising project that could help facilitate Luxembourg’s skills data exchanges. The inclusion of skills as a priority for data exchange (in addition to health and energy), covering both government and stakeholder skills data on the NDEP, should thus be considered. A similar project in Estonia led to the harmonious integration of data sources of many government agencies. It could be examined as a possible model for Luxembourg to follow (see Box 5.10).
Box 5.10. Relevant international example: Data exchange solution
Estonia
The X-Road software-based solution is the backbone of e-Estonia. Estonia’s e-solution environment includes a full range of services for the general public, and since each service has its own information system, they all rely on X-Road to ensure secure transfers, that all outgoing data are digitally signed and encrypted, and that all incoming data are authenticated and logged.
X-Road connects different information systems that may include a variety of services. It has developed into a tool that can also communicate with multiple information systems, transmit large datasets and perform searches across several information systems simultaneously. X-Road was designed with growth in mind, so it can be scaled up as new e-services and new platforms come online.
X-Road has the advantage that access to individual databases can be controlled and regulated, with only authorised users able to enter the databases and receive pre-defined content. Over 150 public sector institutions are connected to X-Road, and it is used indirectly by hundreds of enterprises and institutions and over 50 000 organisations.
In the field of education and skills, X-Road serves to support data exchanges between, for example, the Estonian Education Information System (EHIS) and other databases and platforms, such as the population register or the Health and Insurance Fund. Such connections have several benefits. For example, the connection between EHIS and the population register enables EHIS to access data on a student’s place of residence without the need to collect such information from students themselves. In turn, the Health Insurance Fund uses EHIS data to determine who is eligible for student health insurance.
Today, the X-Road software is implemented in Finland, Iceland, Japan and other countries (see Box 5.12). Estonia and other countries that adopted the approach to exchanging public administrations’ data benefitted from the introduction of common technology to facilitate the technical aspects of the data exchange.
Source: OECD (2020[71]), Strengthening the Governance of Skills Systems: Lessons from Six OECD Countries, https://dx.doi.org/10.1787/3a4bb6ea-en.
The authorities participating in the development of the NDEP should analyse the potential of working with guichet.lu, the official IT tool to carry out tasks with Luxembourgish administrations. The guichet.lu (https://guichet.public.lu/fr.html) platform is managed by the recently created MinDigital and the State Information Technology Centre (CTIE). It has the double function of providing information on public administration services to Luxembourgish citizens, residents and workers and of allowing them to carry out a number of administrative tasks. It is not used to collect data, except for the National Registry of Physical Persons, which is the only dataset kept by CTIE. The guichet.lu portal covers all adults resident in Luxembourg and the active population of the Luxembourgish labour market, including job seekers (who, for example, plan to move to Luxembourg). For this reason, for example, the government is using this IT tool to carry out the latest population census (2021). While the platform does not currently collect data, it has the potential and legal basis to access and connect all government data sources with unique identifiers. Therefore, the potential value-added of the guichet.lu platform in facilitating data exchange and interoperability should be considered in developing Luxembourg’s NDEP.
The value of skills data sources increases when they are made available to the broader public. Direct benefits of open skills data can include, for example, reduced costs for firms related to real-time information on the skills supply and faster, improved information on career guidance for students, apprentices, or job seekers, among others. Open data increase the performance and transparency of public sector organisations and can also be a driver for the demand for data and the development of existing and new data sources. Open data can contribute to the development of innovative services and new research models. Moreover, it can help the institutions responsible for data collection make more informed decisions based on existing resources (Carolan et al., 2016[75]; European Data Portal, 2020[76]).
Luxembourg’s Ministry of State is responsible for the publication of open government data. At least some labour market and skills data are already public. The data.public.lu portal (https://data.public.lu/) currently displays a number of datasets and data flows covering several domains, such as COVID‑19, public transport, etc. ADEM also shares key labour market data updated each month through this portal. However, the user-friendliness of the data.public.lu could be improved to allow for better access and the analysis of complex statistics, as, at the moment, the portal is not meant for the general public but for specific users who are comfortable with this type of open data in conducting their own analysis.
While currently co‑ordination between the relevant authorities seems to be the main priority, making part of the skills data exchanged via the NDEP available for open data publication for the broader public (e.g. skills shortages and mismatches) could increase the value of data collections. If the government were to make further efforts in this direction to increase the value of existing data on skills and the labour market – and collect new data – Luxembourg could envisage greater adherence to open data principles and support further open skills data publication. New Zealand is an example of a government that has embraced a cross-government open data approach and adopted an open data charter (see Box 5.11).
Box 5.11. Relevant international example: Open data principles
New Zealand
Open data are data that are available for everyone to access, use and share. The concept of open data is not new, but it has become increasingly important to focus on making data accessible to everyone. Open data's three most important characteristics are: 1) availability and access; 2) re-use and distribution; and 3) universal participation.
Open data have many benefits when shared freely, including benefits that can be specific and pertain to a particular category (cultural, scientific, environmental, governmental). For example, benefits for researchers include: greater access to data; the ability to build upon and create new research from publicly accessible data; enhancing the visibility of one’s research; the ability to verify and reproduce experiments; increased researcher authenticity; and reduced academic fraud.
By opening data, governments increase the value of data by making them more accessible to the greater public. Other arguments in favour of adopting an open data approach include:
The use of public money to fund the data generation should also be reflected in universally available results.
The rate of discovery in scientific research is accelerated by better access to data.
Opening government data can lead to improving education, governments and other real-world problems by reinforcing the link between researchers and policy makers.
Arguments against adopting an open data approach include:
Privacy concerns may require limited access to data for specific users or data subsets.
To avoid data misuse, the publishing authorities need to invest resources in quality management, dissemination and metadata development.
New Zealand is an example of a country that has made a commitment to open data. New Zealand’s Open Data Partnership provides government agencies with a set of principles and actions to guide and govern the release of open government data, which are not personal, classified or confidential (anyone can use and share them). All data should be open by default, timely and comprehensive, accessible and usable, comparable and interoperable, for improved governance and citizen engagement, and for inclusive development and innovation. New Zealand’s open government data cover, for example, data related to land, health, education and population, among others.
Source: Carolan et al. (2016[75]), Open data, transparency and accountability: Topic guide, www.nationalarchives.gov.uk/doc/open-government-licence; Orvium (2022[77]), What Is Open Data?; https://blog.orvium.io/what-is-open-data/; European Data Portal (2020[76]), The Economic Impact of Open Data Opportunities for Value Creation in Europe; https://data.europa.eu/en/datastories/economic-impact-open-data-opportunities-value-creation-europe; Government of New Zealand (2022[78]), Open data policies, www.data.govt.nz/toolkit/open-data/open-data-policy/.
Recommendations
4.9. Designate a single entity to lead the co‑ordination of Luxembourg’s skills data collection and analysis efforts. Luxembourg could identify a single national governmental body to lead the co‑ordination of all national efforts to collect, analyse and publish data on skills demand and supply. Such an institution could set up and lead a national consortium that includes the main institutions that operate in this domain. This body should have a clear mandate, a feasible timeline and a proactive approach to ensure the national co‑ordination on skills data collection, definitions and working practices. Institutional users of labour market data, such as INFPC, RETEL, LISER, IGSS, ADEM or STATEC, are good candidates to receive a national mandate to initiate this process of co‑ordination and set up a national consortium to set out a skills data charter, an action plan and contribute to the NDEP (see below).
4.10. Develop a national skills data charter and an action plan with clear roles, responsibilities and procedures for government and stakeholders to co-ordinate improving the relevance and quality of skills data in Luxembourg in the short and medium term. The national skills data charter (a formal strategic document) should support the work of the national skills data co‑ordinating body by clearly setting out the objectives and specific actions to improve Luxembourg’s skills data governance going forward (e.g. the types of SAA exercises to be developed, specific data sources to be improved). The charter should include the description of the data collected by each institution, together with accessibility conditions, as well as information on each actor’s current or foreseen efforts (e.g. the Trends working group, NDEP, etc.) for improving Luxembourg’s skills data governance. The charter should clearly allocate the roles and responsibilities of the different key actors in Luxembourg’s skills data system (including STATEC, ADEM, IGSS, CCSS, MENJE, MESR, MinDigital, sectoral stakeholders, etc.) for improving Luxembourg’s skills data governance going forward, while leveraging potential synergies. Main institutional data users, such as RETEL, INFPC, and LISER, should also have observer/advocacy roles. The skills data charter should also define the main governance rules and processes and should be regularly reviewed and updated based on inputs from the relevant actors involved. The charter could also be complemented by an action plan that covers specific time periods, sets out the necessary actions to be taken in the short term and envisages those to take in the medium term.
4.11. Support the development of the National Data Exchange Platform, and advance discussions on the inclusion of skills data among the main priorities of the project. The NDEP project could be developed in collaboration with all the relevant actors covered by the national skills data charter (e.g. STATEC, ADEM, IGSS, CCSS, MTEESS, MENJE, MESR, MinDigital, and sectoral stakeholders). Consideration should be given to assigning unique identifiers to skills and labour market data integrated on the platform to support data interoperability. Limitations imposed by the GDPR should be mapped out in order to set up a legal framework to support collaboration between the different actors involved. The potential of the guichet.lu platform should be explored in order to connect different data sources in relation to the NDEP.
4.12. Establish a simplified protocol for sharing existing labour market data by introducing agency-specific “data passports” to reduce the bureaucratic burden for recurrent institutional data users. Until the NDEP is fully operational, data passports could be assigned to institutions working on skills and labour market data to pre-approve recurrent data requests within the GDPR framework.
4.13. Increase the value of the existing and new skills data collections by better facilitating further publication of open data on skills. Luxembourg could consider greater adherence to open data principles and support further open skills data publication by improving the user-friendliness of the data.public.lu portal. In the future, making part of the skills data exchanged via the NDEP available as open data could be considered.
Building synergies between Luxembourg’s and neighbouring countries’ data sources to improve the skills data availability for the Greater Region
The economies of Luxembourg’s Greater Region are characterised by a high level of economic interdependence, including aspects relating to labour, employment, skills demand and supply. This applies to workers commuting between the four countries and businesses operating both nationally and across borders.
In the Greater Region, it is therefore important to make information on residents, workers and businesses more readily available to relevant government partners to perform routine checks, provide work permits and maintain labour market and skills data quality across borders. Furthermore, governments (and governments’ data flows) need to keep pace with the increasing economic and market interdependence of the Greater Region, aiming to better estimate skills outflows and inflows in addition to skills shortages, demand, supply and mismatches.
Similarly, given that Luxembourg draws its labour supply from the Greater Region, it is necessary to consider the skills supply and demand within the Greater Region and not only that of Luxembourg when designing and implementing Luxembourg’s skills policies. Although Luxembourg is a net labour importer, significant numbers of workers cross borders from Luxembourg and to Luxembourg every day (see Chapter 4). A large share of Luxembourg’s workforce and job seekers (potential workforce) resides in Lorraine, Wallonia, Rhineland-Palatinate and Saarland. Skills that are not available in the Luxembourgish labour market may be available beyond Luxembourg’s borders and within the Greater Region. Similarly, the Greater Region’s skills demand should be considered in addition to Luxembourg’s own. Skills that may be in demand in Luxembourg may also be in demand in its neighbouring regions, impacting the types of skills that Luxembourg can expect to draw upon in the future.
Luxembourg should thus further explore how it can draw upon data on skills and labour supply from Luxembourg’s Greater Region (Saarland, Lorraine, Luxembourg, Rhineland-Palatinate and Wallonia). For example, France’s Occupations’ Observatories (Observatoires des Métiers) and the Lorraine ALFOREAS‑IRTS Observatory undertake work on skills data, skills shortages and skills needs anticipation in their respective regions. France’s Ministries of National Education, Early Childhood and Sports and of Higher Education and Research collect data on graduates’ profiles, including on their fields of study. In addition, the French Pôle Emploi Grand-Est collects data on job seekers, their skills and training. Similarly, Forem (Wallonia’s Employment Service) and the Brussels Employment Observatory regularly publish analyses of occupations and skills shortages and work on skills needs anticipation. Germany’s federal and state employment agencies collect and publish labour market indicators by qualifications. It would be important that Luxembourg, as well as its neighbours in the Greater Region, have ways of securely accessing and sharing such data with each other.
The Interregional Labour Market Observatory (IBA-OIE) regularly gathers already-existing information from the main data-collecting institutions in Belgium, France, Germany and Luxembourg and publishes indicators on the situation of the employment market in the Greater Region. IBA-OIE was created in 2001 as a network of institutes specialised in the field of employment: the INFO-Institut (advocacy and research institute), ADEM, LISER, France’s Cross-Border Mission of the Grand-Est Regional Council, Grand-Est EURES/Cross-border Resource and Documentation Center, Belgium’s Ostbelgien Statistics and Walloon Institute for Evaluation, Forecasting and Statistics. IBA-OIE is an institutional user of labour market data in the Greater Region. However, it does not have a specific mandate to work on skills data (e.g. sectoral shortages/excess studies) or to develop data collection. However, it represents a good example of an institution that exploits synergies between neighbouring countries’ institutions and whose focus is Luxembourg’s Greater Region rather than limiting its scope to the national borders.
In this context, a promising project has been developed in Luxembourg by LISER and its Data Centre, which are working on establishing a “safe room” to access Belgium, France, Germany and Luxembourg’s administrative data on workers and firms. The setup of a data room would consist of a secure space to streamline and simplify organisations’ workflows and to enable them to access each other’s data while meeting compliance standards and ensuring high levels of security. Another project will link administrative data from Germany and Luxembourg, which would enable the analysis of employment profiles of past and current cross-border workers in these two countries. Such efforts could be scaled up and further developed. For example, Estonia and Finland have established systems allowing to exchange data across borders (Box 5.12).
Box 5.12. Relevant international example: Cross-border data exchange
Estonia and Finland
The Governments of Estonia and Finland are exchanging data across borders in an X-Road Trust Federation. Estonia and Finland, respectively, had already developed and implemented their national data exchange systems to enable interoperability across government agencies of the same country.
X-Road is an open-source software solution that provides unified and secure data exchange between organisations. It is also a digital public good, verified by Digital Public Goods Alliance. The X-Road Trust Federation is an ecosystem for cross-border data exchange allowing for interoperability between two X-Road-based data exchange layers from different jurisdictions – be these regions, federal states or countries.
By linking the two X-Road-based systems initially developed separately in Estonia and Finland, government agencies in the two countries are now exchanging relevant census and labour market data on population and businesses via the Internet across borders. Data are exchanged across borders between Estonia and Finland and between national commercial registers, tax boards and population registers. Benefits are immediately identifiable in reduced bureaucracy (workload and request processing times); public officials can access data more seamlessly, shorten waiting times, increase information availability, and grant data accuracy. Citizens, workers and businesses enjoy a more efficiently operative public sector, have their data automatically available for the government entities they must interact with, and do not need to submit updates of personal or business information as these change in their home country.
Estonia, Finland, and other countries benefitted from the introduction of a national platform to facilitate data exchange between public agencies. The fact that the platform (and the technology) used ensured the compatibility of Estonia and Finland’s systems facilitated the setup of the data exchange between the two countries.
Source: X-Road (2020[79]), Estonia and Finland launch automated data exchange between population registers — X-Road® Data Exchange Layer, https://x-road.global/estonia-and-finland-launch-automated-data-exchange-between-population-registers.
A joint data exchange platform for Luxembourg’s Greater Region could work similarly to how the NDEP plans to function at the national level and build on the existing initiatives like LISER’s “safe room” and IBA‑OIE’s work. A cross-border data exchange platform could foster data exchange between each country’s public authorities, integrate the Greater Region’s stakeholders and simplify data requests in relation to GDPR compliance. As this work would involve public institutions from different countries, it would also be more difficult to create the political stimulus to carry out this project: it would therefore need to be led by a single body, corresponding to a public authority or a consortium of public authorities in order for it to function effectively.
Recommendations
4.14. Develop a comprehensive mapping of the neighbouring countries' data sources on skills. The sources mapped should be complementary to Luxembourg's sources on skills and labour supply (e.g. neighbouring regions’ data on trainings, vocational education and training, HEIs and unemployed/job seekers). Luxembourg authorities and the Greater Region need to identify relevant governmental and non-governmental actors in Belgium, France and Germany that collect or might collect data on skills available within Luxembourg’s Greater Region or skills demand within the region that might be attracting skills available within Luxembourgish borders. Given the large number of skills data sources available within the Greater Region, a co‑ordinated effort by research institutes in the four countries could be envisaged to lead the mapping exercise. For example, similar co‑ordinated research efforts have taken place at the European level through European Research Infrastructure Consortiums (ERICs). Among Luxembourgish institutions, the Ministry of Family Affairs, Integration and the Greater Region, the IBA-OIE Observatory, LISER and its data room are well placed to take the initiative on this matter.
4.15. Promote the establishment of a data exchange platform for skills data within the Greater Region. Luxembourg and the Greater Region would strongly benefit from increased collaboration between the four countries' public authorities and from the institutionalisation of data exchange covering labour market data, skills and other related domains. Exchanging data and, subsequently, the increased availability of data could lead to better knowledge and smoother information flows on the labour market. As political incentives do not lie within a single country’s borders, government authorities should identify (or create) a responsible body to co‑ordinate efforts to improve the availability and exchange of data within the Greater Region. The potential of using Luxembourg’s own foreseen NDEP for this purpose (i.e. making the platform interoperable internationally) should be explored.
Overview of recommendations
Strong governance of skills data is essential to help policy makers and stakeholders navigate the complexity and uncertainty associated with the design and implementation of skills policies. Two opportunities have been selected indicating where the governance of skills data in Luxembourg can be strengthened:
1. improving the quality of Luxembourg’s skills data collection
2. strengthening co‑ordination of, and synergies between, skills data within and beyond Luxembourg.
This chapter presented 15 recommendations to seize these opportunities in the area of skills data governance. A high-level overview of the recommendations can be found in Table 5.3. This selection is based on input from a literature review, desk research, discussions with the Luxembourg National Project Team and broad engagement with a large variety of stakeholders, including two workshops in Luxembourg and various related meetings and group discussions.
Two recommendations have been selected that could be considered to have the highest priority based on potential impact, relevance in the current context in Luxembourg, as well as overall support for implementation. To strengthen the governance of skills data, the OECD recommends that Luxembourg, bearing in mind its administrative capacity, should:
Improve the accuracy of occupational social security data by creating targeted incentives for employers, strengthening existing guidance tools for identifying the correct occupational codes, and conducting targeted awareness raising (Recommendation 4.1).
Develop a national skills data charter and an action plan with clear roles, responsibilities and procedures for government and stakeholders to co-ordinate improving the relevance and quality of skills data in Luxembourg in the short and medium term. (Recommendation 4.10).
Table 5.3. High-level overview of recommendations to strengthen the governance of skills data
Note: ADEM: Agence pour le développement de l’emploi; CEDEFOP: European Centre for the Development of Vocational Training; CCSS: Centre commun de sécurité sociale; IGSS: Inspection générale de la sécurité sociale; RETEL: Réseau d’études sur le travail et l’emploi; LISER: Luxembourg Institute of Socio-Economic Research; UEL: Union des Entreprises Luxembourgeoises; STATEC: Institut national de la statistique et des études économiques; MESR: ministère de l'Enseignement supérieur et de la Recherche; INFPC: Institut national pour le développement de la formation professionnelle continue; MFAMIGR: ministère des Affaires familiales, de l’Intégration et de la Grande Région; IBA-OIE: Observatoire interrégional du marché de l’emploi.
References
[48] ABBL, ALEBA, OGBL and LCBG-SEFS (2018), Collective Labor Agreement for Banking Employess 2018-2020 [Convention collective de travail des salariés de banque 2018-2020], http://www.ogbl.lu/syndicat-banques-et-assurances/files/2018/07/20180705-CCT-version-finale-final.pdf (accessed on 29 January 2022).
[15] ADEM (2022), Chiffres-Clés de l’ADEM, https://adem.public.lu/fr/marche-emploi-luxembourg/faits-et-chiffres/chiffres-cles-adem.html (accessed on 26 August 2022).
[16] ADEM (2022), Sectoral studies, https://adem.public.lu/en/employeurs/futureskills/etudes-sectorielles.html (accessed on 21 January 2022).
[34] ADEM (2021), How to submit a declaration of vacant position, https://adem.public.lu/en/employeurs/declarer-poste-vacant/comment.html (accessed on 2 June 2022).
[37] ADEM (2021), Job vacancies declared to ADEM become accessible for all jobseekers, https://adem.public.lu/en/actualites/adem/2021/04/jobboard2021.html (accessed on 5 May 2022).
[35] ADEM (2021), Vacancy data.
[23] ADEM/UEL (2021), ADEM/UEL Employment Partnership - Trends Working Group - Working document.
[57] Blank, G. and C. Lutz (2017), “Representativeness of Social Media in Great Britain: Investigating Facebook, LinkedIn, Twitter, Pinterest, Google+, and Instagram:”, https://doi.org/10.1177/0002764217717559, Vol. 61/7, pp. 741-756, https://doi.org/10.1177/0002764217717559.
[75] Carolan, L. et al. (2016), Open data, transparency and accountability: Topic guide, http://www.nationalarchives.gov.uk/doc/open-government-licence (accessed on 13 May 2022).
[49] CdM (n.d.), Occupations and qualifications [Métiers et formations], https://www.handsup.lu/fr/metiers-formations (accessed on 29 January 2022).
[38] CEDEFOP (2021), Skills-OVATE, https://www.cedefop.europa.eu/en/tools/skills-online-vacancies (accessed on 4 December 2021).
[42] CEDEFOP (2013), “User guide to developing an employer survey on skill needs”, No. 35, Publications Office of the European Union, Luxembourg, https://doi.org/10.2801/25552.
[13] Chambre de Commerce Luxembourg (2021), Focus thématique: Skills & Training, Chambre de Commerce Luxembourg, https://www.cc.lu/fileadmin/user_upload/tx_ccpublications/BE_S1-2021_Publication.pdf (accessed on 2 December 2022).
[12] Chambre de Commerce Luxembourg (2019), Focus thématique: Recrutement, Chambre de Commerce Luxembourg, https://www.cc.lu/fileadmin/user_upload/tx_ccpublications/Barometre_economie_S1_2019.pdf.
[14] Chambre des salariés Luxembourg (2021), Pourquoi Quality of Work Index ?, Quality of Work Index Luxembourg, https://www.csl.lu/fr/sante-et-bien-etre-au-travail/quality-of-work-index-luxembourg/ (accessed on 2 December 2022).
[44] Department for Education (2022), Employer Skills Survey 2022, https://www.skillssurvey.co.uk/index.htm (accessed on 15 February 2022).
[47] European Commission (2022), ESCOpedia, https://esco.ec.europa.eu/en/about-esco/escopedia/escopedia (accessed on 13 June 2022).
[54] European Commission (2022), The adoption of ESCO by the Greek Public Employment Service (OAED), https://esco.ec.europa.eu/en/news/adoption-esco-greek-public-employment-service-oaed (accessed on 27 August 2022).
[46] European Commission (2022), What is ESCO?, https://esco.ec.europa.eu/en/about-esco/what-esco (accessed on 2 June 2022).
[50] European Commission (2021), National occupational classifications, https://ec.europa.eu/esco/portal/escopedia/National_occupational_classifications (accessed on 4 December 2021).
[61] European Commission (2020), Graduate tracking - A “how to do it well” guide, https://op.europa.eu/en/publication-detail/-/publication/5c71362f-a671-11ea-bb7a-01aa75ed71a1/language-en/format-PDF/source-search (accessed on 6 December 2021).
[39] European Commission (2018), Engaging with and improving services to employers, https://ec.europa.eu/social/main.jsp?langId=en&catId=89&newsId=9158&furtherNews=yes (accessed on 7 December 2021).
[52] European Commission (2018), How ESCO supports the Public Employment Service of Iceland, https://audiovisual.ec.europa.eu/en/video/I-162745 (accessed on 7 December 2021).
[65] European Commission (2018), Mapping of VET graduate tracking measures in EU Member States, https://op.europa.eu/en/publication-detail/-/publication/00d61a86-48fc-11e8-be1d-01aa75ed71a1 (accessed on 4 December 2021).
[51] European Commission (n.d.), Job-matching ESCO use case, https://esco.ec.europa.eu/en/job-matching-esco-use-case (accessed on 2 June 2022).
[76] European Data Portal (2020), The Economic Impact of Open Data Opportunities for Value Creation in Europe, https://doi.org/10.2830/63132.
[70] European Environment Agency (2019), “Open data and e-government good practices for fostering environmental information sharing and dissemination”.
[56] Eurostat (n.d.), Education and Training - National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRS), https://ec.europa.eu/eurostat/cache/metadata/en/educ_uoe_enr_esqrs_lu.htm (accessed on 31 January 2022).
[10] FEDIL (2021), Les qualifications de demain dans l’industrie 2021, FEDIL, https://fedil.lu/en/publications/les-qualifications-de-demain-dans-lindustrie-2021/ (accessed on 2 December 2022).
[11] FEDIL (2018), The Qualifications of Tomorrow in the ICT Sector 2018, https://fedil.lu/en/publications/les-qualifications-de-demain-dans-les-tic-20172018/ (accessed on 2 December 2022).
[36] Goffin, S. (2015), Less than half of the vacancies reported [Moins de la moitié des postes vacants déclarés], http://www.lessentiel.lu/fr/luxembourg/story/moins-de-la-moitie-des-postes-vacants-declares-11103562 (accessed on 7 December 2021).
[29] Government of Canada (2017), Canadian Occupational Projection System (COPS), http://occupations.esdc.gc.ca/sppc-cops/w.2lc.4m.2@-eng.jsp (accessed on 30 January 2022).
[80] Government of Luxembourg (2021), Applying for state financial aid for higher education (AideFi), https://guichet.public.lu/en/citoyens/enseignement-formation/etudes-superieures/aides-logement/aide-financiere.html (accessed on 14 February 2022).
[31] Government of Luxembourg (2021), Making a declaration of start of employment to the social security — Business — Guichet.lu - Administrative Guide - Luxembourg, https://guichet.public.lu/en/entreprises/sante-securite/declaration-secu/affiliation/declaration-entree.html (accessed on 3 November 2021).
[22] Government of Luxembourg (2018), Labour Market and Employment Research Network (RETEL), https://adem.public.lu/fr/marche-emploi-luxembourg/acteurs/retel.html (accessed on 4 December 2021).
[43] Government of the United Kingdom (2020), Employer Skills Survey 2019, https://www.gov.uk/government/collections/employer-skills-survey-2019 (accessed on 4 December 2021).
[78] Government, N. (2022), Open data policies, https://www.data.govt.nz/toolkit/open-data/open-data-policy/ (accessed on 23 February 2022).
[18] IGSS (2022), Publications, http://igss.gouvernement.lu/fr/publications.html (accessed on 26 August 2022).
[17] IGSS (2022), Tableau de bord janvier 2021- janvier 2022, http://www.adem.public.lu/fr/marche-emploi-luxembourg/faits-et-chiffres/statistiques/igss/Tableaux-interactifs-flux-emploi/index.html (accessed on 26 August 2022).
[19] INFPC (2021), About the Training Observatory [À propos de l’Observatoire de la formation], https://www.lifelong-learning.lu/Detail/Article/Accueil/observatoire-de-la-formation/fr (accessed on 3 February 2022).
[40] Investitionsbank Berlin (2021), 2021/2022 Business Support Guide, https://www.businesslocationcenter.de/fileadmin/user_upload/Wirtschaftsstandort/files/foerderfibel-en.pdf (accessed on 26 October 2022).
[55] Lassébie, J. et al. (2021), “Speaking the Same Language: A Machine Learning Approach to Classify Skills in Burning Glass Technologies Data”, https://www.oecd.org/fr/emploi/speaking-the-same-language-adb03746-en.htm (accessed on 2 December 2021).
[26] LISER (2020), Annual Report 2020, https://liser.elsevierpure.com/ws/portalfiles/portal/32702683/2021_LISER_Annual_Report_2020.pdf (accessed on 4 December 2021).
[68] LIST (2022), Cross Skill, https://crossskill.list.lu/ (accessed on 22 January 2022).
[28] McGrath, J. and J. Behan (2017), “A comparison of shortage and surplus occupations based on analyses of data from the European Public Employment Services and Labour Force Surveys”, Labour shortages and surpluses.
[67] MESR/MENJE (2020), Report on referencing the Luxembourg qualifications framework to the European Qualifications Framework for Lifelong Learning and to the qualifications framework in the European Higher Education Area, https://europa.eu/europass/system/files/2020-06/Luxembourg%20Referencing%20Report.pdf (accessed on 17 January 2022).
[66] Ministry of Labour, E. (2019), Annual report 2019 - Survey on the future of candidates with professional qualifications [Rapport annuel 2019 - Enquête Devenir des candidats au Titre professionnel].
[45] National Skills Commission (2021), Australian Skills Classification BETA release discussion paper, https://www.nationalskillscommission.gov.au/sites/default/files/2021-03/AUSTRALIAN%20SKILLS%20CLASSIFICATION%20BETA%20DISCUSSION%20PAPER.pdf (accessed on 4 December 2021).
[53] OAED (2021), OAED adopts the new ESCO classification of occupations and skills, https://www.oaed.gr/storage/oaed-logo/enimerotiko-entipo-esco-matching-en.pdf (accessed on 2 June 2022).
[9] Observatoire de la formation and INFPC (2021), Offre de formation: L’offre de formation continue en chiffres-enquête 2020, https://www.lifelong-learning.lu/bookshelf/documents/l_offre-de-formation-continue-en-chiffres---enquete-2020.pdf.
[4] OECD (2021), Education at a Glance 2021: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/b35a14e5-en.
[69] OECD (2021), “Mapping data portability initiatives, opportunities and challenges”, OECD Digital Economy Papers, No. 321, OECD Publishing, Paris, https://doi.org/10.1787/a6edfab2-en.
[62] OECD (2021), Strengthening Quality Assurance in Adult Education and Training in Portugal: Implementation Guidance, OECD Publishing, Paris, https://www.oecd.org/portugal/Strengthening-Quality-Assurance-in-Adult-Education-and-Training-in-Portugal-Implementation-Guidance.pdf.
[41] OECD (2020), Preparing the Basque Country, Spain for the Future of Work, OECD Reviews on Local Job Creation, OECD Publishing, Paris, https://doi.org/10.1787/86616269-en.
[71] OECD (2020), Strengthening the Governance of Skills Systems: Lessons from Six OECD Countries, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/3a4bb6ea-en.
[59] OECD (2019), Monitoring Learning Outcomes of Adult Learning Programmes: A Review of European Best Practices on Monitoring Instruments, https://epale.ec.europa.eu/sites/default/files/oeso-rapport-2019-monitoring-learning-outcomes-of-adult-learning-programmes.pdf.
[1] OECD (2019), OECD Skills Strategy 2019: Skills to Shape a Better Future, OECD Publishing, Paris, https://doi.org/10.1787/9789264313835-en.
[60] OECD (2019), OECD Skills Strategy Poland: Assessment and Recommendations, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/b377fbcc-en.
[64] OECD (2018), OECD Reviews of School Resources: Portugal 2018, OECD Reviews of School Resources, OECD Publishing, Paris, https://doi.org/10.1787/9789264308411-en.
[74] OECD (2018), Private Philanthropy for Development, The Development Dimension, OECD Publishing, Paris, https://doi.org/10.1787/9789264085190-en.
[63] OECD (2018), Skills Strategy Implementation Guidance for Portugal: Strengthening the Adult-Learning System, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264298705-en.
[27] OECD (2016), Getting Skills Right: Assessing and Anticipating Changing Skill Needs, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/9789264252073-en.
[25] OECD (2014), Measuring and Managing Results in Development Co-operation: A Review of Challenges and Practices among DAC Members and Observers, https://www.oecd.org/dac/peer-reviews/Measuring-and-managing-results.pdf (accessed on 25 January 2022).
[24] OECD (2011), Quality Dimensions, Core Values for OECD Statistics and Procedures for Planning and Evaluating Statistical Activities, OECD Publishing, Paris, https://www.oecd.org/sdd/21687665.pdf (accessed on 13 December 2021).
[3] OF, INFPC (2022), Pratiques de formation: Analyse quantitative de l’activité de formation des enterprises bénéficiaires de l’aide financière de l’État en matière de formation professionnelle continue en 2019, lifelong-learning.lu, https://www.lifelong-learning.lu/bookshelf/documents/pratiques-de-formation-avril-2022.pdf (accessed on 2 December 2022).
[2] OF, INFPC (2019), TEVA: Premiers pas dans la vie active des jeunes issus de la formation professionnelle (TEVA: First steps in the working life of young people from vocational training), https://www.lifelong-learning.lu/bookshelf/documents/infpc_teva_indicateurs_2014-2017.pdf.
[77] Orvium (2022), Whats Is Open Data?, https://blog.orvium.io/what-is-open-data/ (accessed on 23 February 2022).
[6] STATEC (2022), Adult education survey (AES), https://statistiques.public.lu/en/enquetes/enquetes-particuliers/adultes-education.html (accessed on 2 December 2022).
[7] STATEC (2022), Continuing Vocational Training Survey (CVTS), https://statistiques.public.lu/en/enquetes/enquetes-entreprises/professionnelle-education.html.
[5] STATEC (2022), Labour force survey (LFS), https://statistiques.public.lu/en/enquetes/enquetes-particuliers/eft-travail.html (accessed on 2 December 2022).
[8] STATEC (2022), Structure of Earnings Survey, https://statistiques.public.lu/en/enquetes/enquetes-entreprises/salaires-travail.html.
[20] STATEC (2018), Bulletin n° 2/2018 L’éducation des adultes, https://statistiques.public.lu/fr/publications/series/bulletin-statec/2018/bulletin-02-18.html (accessed on 26 August 2022).
[21] STATEC (2018), Regards N° 04/2018 - Formation dans les entreprises: Le Luxembourg progresse, https://statistiques.public.lu/fr/publications/series/regards/2018/04-18.html (accessed on 26 August 2022).
[72] US Department of Education (2019), Office of the Chief Data Officer Data Governance Board Charter, https://www.ed.gov/sites/default/files/cdo/dgb-charter.pdf (accessed on 23 February 2022).
[73] US Department of Education (n.d.), Foundations for Evidence-Based Policymaking, https://www.ed.gov/data (accessed on 13 May 2022).
[58] van Dijck, J. (2013), “‘You have one identity’: Performing the self on Facebook and LinkedIn:”, Media, Culture & Society, Vol. 35/2, pp. 199-215, https://doi.org/10.1177/0163443712468605.
[33] Warwick Institute for Employment Research (2018), Cascot International, https://warwick.ac.uk/fac/soc/ier/software/cascot/internat/ (accessed on 1 May 2022).
[32] Warwick Institute for Employment Research (2018), Cascot: Computer Assisted Structured Coding Tool, https://warwick.ac.uk/fac/soc/ier/software/cascot/details/ (accessed on 1 May 2022).
[30] Warwick Institute for Employment Research (n.d.), Working Futures, 2017-2027, https://warwick.ac.uk/fac/soc/ier/wf7downloads (accessed on 9 May 2022).
[79] X-Road (2020), Estonia and Finland launch automated data exchange between population registers — X-Road® Data Exchange Layer, https://x-road.global/estonia-and-finland-launch-automated-data-exchange-between-population-registers (accessed on 23 February 2022).
Notes
← 1. Integrated information systems are systems that collect and manage the data and information that governments and stakeholders produce, analyse and disseminate to ensure that policy makers, firms, individuals and others have access to accurate, timely, detailed and tailored information. Relevant data and information include, among others, the results of skills assessment and anticipation exercises, information on where to access learning opportunities, as well as information from evaluations of public policies (OECD, 2019[1]).
← 2. The Statistical Classification of Economic Activities in the European Community, abbreviated as NACE, is the classification of economic activities in the European Union (EU); NACE is a four-digit classification providing the framework for collecting and presenting a large range of statistical data according to economic activity in the fields of economic statistics.
← 3. The ROME is the "Operational Directory of Trades and Jobs", which was created in 1989 by the ANPE (French National Agency for Employment). It is mainly used for classification and identifying trades based on associated skills. The ROME code is often used by administrations, employment services to classify occupations, announcements and requests from employers.
← 4. Skills assessment and anticipation (SAA) exercises are tools to generate information about the current and future skills needs of the labour market (skills demand) and the available skills supply. SAA exercises include general labour market information systems, sectoral/occupational/regional studies and forecast-based projections, among others (OECD, 2016[27]).
← 5. In addition to the graduate tracking undertaken by the University of Luxembourg, certain private higher education institutions in Luxembourg undertake their own graduate tracking studies. In 2022, the Luxembourg National Research Fund also completed a one-off tracking study of PhD students in public-private partnership programmes.
← 6. CCSS occupational data could also be used for the purposes of graduate tracking, conditional upon improving the quality of the CCSS occupational data (see Recommendation 4.1).
← 7. The results of the “employment study” carried out by the University of Luxembourg (see more above) show that 53% of master’s graduates from Luxembourg, 52% of master’s graduates from the European Union (excluding Luxembourg) and 49% of master’s graduates from third (i.e. non-EU) countries have found employment in Luxembourg following graduation. The figure stands at 38% (Luxembourg), 31% (European Union excluding Luxembourg) and 36% (third countries) for PhD graduates during the same time period. Lower-bound estimates are used in both cases due to missing information in the UoL employment study (see more above).
← 8. In Luxembourg, both resident and non-resident students (full-time or part-time) in higher education (HE) are eligible for financial aid for higher education (AideFi), subject to certain conditions. For example, a non-resident student is eligible for AideFi as long as the student’s parent(s) has been working in Luxembourg for at least five years over a ten-year reference period preceding the AideFi application, or for at least ten years at the time of the application (Government of Luxembourg, 2021[80]). Stakeholders have indicated that tracking the outcomes of HE students receiving AideFi but studying outside of Luxembourg is a challenge.