Informing youngsters and employers about labour market prospects

Tim Peeters, Didier Fouarge, Jessie Bakens

Every two years the Research Centre for Education and the Labour Market (ROA) at Maastricht University produces medium-term labour market forecasts for 21 industry sectors and more than 100 occupational groups and education types. These forecasts, together with indicators for the current labour market situation, are gathered in an online database: the Labour Market Information System (AIS). In this article we describe our approach to labour market forecasting and present the content of the AIS. Although the project was initially started to help prospective students with their choice of study, the set of indicators has been broadened over time to include data and forecasts that are useful to employers, policy makers and labour market mediators. In this article, we focus on the relevance for prospective students and, to a lesser extent, employers.

Relevance of labour market information

As part of the Education and Labour Market Project (POA), ROA develops a number of research activities aimed at gaining a better understanding of the medium-term developments in supply and demand on the Dutch labour market.1 The main focus of these activities is the development of labour market indicators for the current friction between supply and demand, and labour market forecasts of supply and demand by industry, occupation, education, and region for the coming six years.

The idea behind making labour market forecasts and presenting them in an online database is that labour market information creates a more transparent labour market, which in turn could lead to better-informed labour market choices. At the micro level, this could result in fewer graduates being unemployed or regretting their decision for a particular field of study. At the macro level, it could improve the allocation of labour supply and demand.

Young people who are in the process of choosing an education or further education programme are an important target group of POA. Another target group consists of job seekers who want to switch to a different occupation, or would like to re-educate or retrain themselves. Although educational choices are not always made consciously and they are among the most common regrets in life (Roese/Summerville 2005), human capital theory (Becker 1962) does suggest that the expected returns from investment in schooling do play a role in the field-of-study choice.

Recent field experiments show that youngsters are often misinformed about the wage returns associated with different fields of study, and that providing them with information changes their expectations in the desired direction (Conlon 2017). Because prospective student’s beliefs about the returns to education have been shown to relate to the decisions they actually make (Wiswall/Zafar 2016), informing prospective students is expected to lead to more optimal investment decisions. Evidence also suggests that youngsters who take labour market considerations into account when choosing their field of study have more realistic wage expectations, and end up in a field with higher returns than those who primarily base their decision on the match with their personal interests (Hastings et al. 2016).

Another way in which labour market information can improve the labour market outcomes (e.g. unemployment, salary, type of contract, etc.) of youngsters and job seekers, is that it can implicitly inform them about general changes in the demand for labour, e.g., as a result of technological change. Because of changes in the skill demands of employers due to globalisation and technological developments, the demand for certain qualifications is declining (e.g., routine and administrative skills), while the demand for other qualifications is increasing (e.g., STEM skills) (Autor 2015; Autor/Dorn 2013). If youngsters or job seekers are unaware of such changes in labour demand or if they are not sensitive to these market signals, they risk unemployment or low-paid work, while employers risk running into labour shortages. Healthcare, technical and teaching occupations already experience such shortages in the Netherlands (ROA 2017).

Labour market forecasts are also relevant for employers, labour market mediators, and policy makers. Employers can adapt their recruitment strategy or terms of employment because of a predicted imbalance between supply and demand, while labour market mediators and related organisations can use the information to adjust their integration and training programmes. Policy makers can use labour market information to design policies that aim to reduce expected personnel shortages in certain sectors.

Structure of the forecasting model2

ROA’s labour market forecasts are derived from estimates of the expected medium-term flows to and from the labour market. The model itself is based on (explanatory) econometric models developed by ROA, and computes the expected labour supply and demand for the coming six years. It then uses those results to derive friction indicators for 113 occupations and 90 types of education. These educational and occupational classifications were created by ROA, cover the entire Dutch labour market, and are documented in ROA (2016), and ROA/CBS (2015), respectively. Each "type of education" reflects the combination of a qualification level and a field-of-study. The following Information Box lists the most important data inputs of the forecasting model. The details of the forecasting model are discussed in Bakens et al. (2018).

Data input for the forecasting model

The data used as input for the econometric models for labour market forecasting include:

  1. Time series data from the Labour Force Survey (LFS) conducted by Statistics Netherlands. The data used consist of time series of employment by industry sector, occupation, type of education, and their combination (period 1996-now).
  2. Employment, investment and added value forecasts by industry sector (by SEOR).
  3. Baseline forecasts from the Ministry of Education, Culture and Science (OCW) for the expected inflow of students in the labour market for broad categories of education.
  4. Current education-specific data from ROA’s School Leaver Information System (SIS).
  5. Forecasts for the short- and mid-term economic growth, and (gross) participation rates by education level, gender and age cohort by the Bureau for Economic Policy Analysis (CPB).
  6. Forecasts for the total size of the labour force by CPB.


The demand side of the model consists of three components that are estimated independently: expansion demand, replacement demand and substitution demand. The expansion demand is the expected employment creation (or destruction in case of economic decline) in an occupational group or type of education. To calculate this component, the model takes macroeconomic projections of employment growth in 21 industry sectors, and translates them into expected changes in employment by occupation. The calculation takes into account sectoral shifts in the occupational structure, as well as the possibility that, within industries, certain occupational groups might grow faster than others. Based on these occupational projections, the expected expansion demand by education type is determined, taking into account shifts in the educational composition within occupational groups (Dupuy 2006). If the current supply of individuals with a certain required degree falls short of employers’ demand, employers will seek to fill their vacancies by hiring people with a related degree or qualification. These substitution processes (substitution demand) are explicitly modelled.

Finally, the demand side of the labour market also consists of replacement demand that results from, for example, (early) retirement, outflow due to disability, temporary withdrawal from the labour market and occupational mobility. These processes are estimated using a cohort-component model.



The labour supply consists of the expected future inflow of graduates in the labour market during the forecasting period, as well as the short-term unemployed at the start of the forecasting period. Based on historical evidence, the assumption is made that the long-term unemployed (defined as those who have been looking for a job for longer than one year) are no serious competitors for recent graduates.



For each of the 90 types of education programmes we compare the expected labour supply for the next six years to the expected labour demand in order to quantify the future labour market prospects. Our main friction indicator, the Indicator Labour Market Prospects (ITA), is then calculated as the ratio of labour supply to labour demand. The ITA gives an indication of how much effort school-leavers will have to put into finding a job. If the labour supply is smaller than or equal to the labour demand, the ITA is smaller than or equal to 1.00 and the labour market outlook is defined as ‘good’. If the value of the ITA is smaller than or equal to 0.85, the labour market outlook is classified as ‘very good’. When the ITA has a value between 1.00 and 1.05 – and the excess supply is thus not much larger than what can arguably be regarded as friction – the labour market prospects are ‘fair’ (slight oversupply that results in frictional unemployment). Larger values of the ITA correspond to ‘poor’ labour market prospects, and a value greater than 1.15 is regarded as ‘bad’.

In addition to the prospects by type of education, a friction indicator is also constructed for the 113 occupational groups, the so-called Indicator Future Staffing Bottlenecks by Occupation (ITKB). This indicator represents the chance that the desired educational composition of the personnel structure within a given occupational group can actually be realised given the predicted supply and demand dynamics of the underlying education types in that occupation. The computation of the ITKB thus combines the forecasts by education type with information on the educational structure of the occupations. The values of the ITKB are categorised as either ‘no‘, ‘almost no’, ‘some‘, ‘large‘ or ‘very large‘ bottlenecks. Note that the ITKB is interpreted from the viewpoint of employers, while the ITA takes the viewpoint of job seekers.

Content of the AIS database

The labour market forecasts for 113 occupational groups, 90 types of education and 21 industry sectors are published in ROA’s online statistical database: the AIS (http://roastatistics.maastrichtuniversity.nl/) . Note that these forecasts are indicative of future job prospects and expected personnel shortages. Although the odds of finding a job is an important aspect young people might care about when choosing an education programme, it is typically not the only aspect they consider, and it is not the only aspect that matters for a good ‘match’ on the labour market. Wages, the typical number of working hours in jobs, and other employment characteristics such as the type of contract also determine how good (or bad) labour market prospects are. That is why the AIS also contains indicators that characterise the current labour market conditions for each education type and occupational group. The data for the current state of the labour market are updated on a yearly basis, while the labour market forecasts for the next six years are estimated every two years.

Content of the AIS database

The AIS includes the following information:

  1. Forecasts for the expected labour demand (six years ahead) and its components: expansion demand and replacement demand, by type of education and occupation.
  2. Forecasts for the expected labour supply (six years ahead): expected inflow of graduates by type of education.
  3. Friction indicators based on the projected supply and demand by education and occupation: the ITA for educations and the ITKB for occupations.
  4. Indicators for the cyclicality of employment by type of education and occupation, as well as indicators for the way in which the employment by type of education is distributed across occupations and industry sectors, and the educational composition of occupations.
  5. Indicator for the extent to which types of education compete for jobs in the same occupation.
  6. Industry sector, type of education and occupation-specific indicators that characterise the current situation on the labour market (e.g., total employment, composition of the workforce by age and gender, wages, hours worked).
  7. Indicators for the employment characteristics of recent graduates by type of education (e.g. job finding rate at graduation, quality of the match, and wage 1.5 years after graduation).

(cf. the full list of indicators in Bakens et al. 2018)

Although the AIS also contains occupational information, we demonstrate the content of the database and its usage with a few examples for education types.

Note: The selection is indicated in grey. ‘Mbo 4’ refers to higher-level secondary vocational education.

The figure shows a selection of variables for the labour market prospects of (vocational education level 4) ‘mbo 4 technical installation’ (technische installatie) and "mbo 4 healthcare" (gezondheidszorg). The ITA of 0.98 for mbo 4 healthcare implies that for this type of education the demand for labour is expected to be larger than the supply of labour, and the labour market prospects for graduates and job seekers with this degree are therefore ‘good’.

In order to gain more insight into the reasons behind these favourable labour market prospects, it is also possible to select the expected expansion and replacement demand, as well as the inflow of school-leavers. Because job prospects are not the only indicator job seekers or future students consider, it is also possible to inspect information such as the expected wage growth during the career, the sensitivity to the business cycle, and the width of the spectrum of occupations that can be entered given that type of education (uitwijkmogelijkheden).

On top of that, a wide range of labour market indicators for the current labour market situation are included in the AIS. These include, for each type of education and occupation: the share of female workers, the share of non-western migrants, the percentage of workers with a permanent contract, the average hours worked per week, the percentage of workers who work full-time, the unemployment share and the average gross hourly wage. The AIS shows that despite the ‘good’ labour market prospects for ‘mbo 4 healthcare’, the average gross hourly wage is not particularly high and only 19 per cent of workers with this degree actually work full-time.

Some final thoughts

The AIS is freely accessible to everyone after registration, and its data is used and disseminated by a large number of stakeholders. First, a number of major information websites (www.studiekeuze123.nl) and guides (www.keuzegids.org) aimed at youngsters choosing their field of study, make part of the information accessible to the broader public. Secondly, various independent organisations in the field of education, such as the Cooperation Organisation for Vocational Education, Training and the Labour Market (SBB) (www.kansopwerk.nl), also use this data in their dissemination of information. Thirdly, labour market mediators and related organisations use the AIS information to adjust their integration and training programmes. For example, organisations such as the employment agency Randstad Netherlands, and the Employee Insurance Agency (UWV) use the ROA-forecasts in their mediation activities.

Furthermore, organisations responsible for the accreditation of new study programmes in higher education (www.cdho.nl) use the ROA-forecasts in their efficacy evaluations to avoid the start-up of new education programmes in fields characterised by an excess-supply of graduates.

The AIS is a very rich database in terms of its content. This makes it difficult to comprehend for new users. We are currently actively looking at potential ways to visualise the database, and thus make it easier to comprehend.

The project is financed by The Netherlands Initiative for Education Research (NRO), which is part of the National Science Foundation (NWO) with additional funding from UWV, SBB and Randstad Netherlands. In order to safeguard the quality and relevance of the project, the project is supervised by and external monitoring committee that meets three times a year to discuss and reflect on ROA’s work. Moreover, we periodically evaluate the forecasting model and engage with stakeholders to assess the quality of the forecasts (see the project’s web page for details).

Although our data are broadly used, it is hard to establish whether such data has behavioural effects on the field-of-study-choice of youngsters or the recruitment strategy of employers, However, a growing body of (quasi-) experimental evidence suggest that youngsters, especially those with a low socioeconomic status (Jensen 2010), are sensitive to the labour market information that is provided to them. We review this literature in Fouarge (2017). Research also shows that the availability of labour market information improves the transition from school to work of youngsters (Saniter et al. 2019). Currently, ROA is conducting a randomised control trial experiment to further investigate the effect of labour market information provision on the educational and occupational choices of youngsters.


AUTOR, D. H.: Why are there still so many jobs? The history and future of workplace automation. In: Journal of Economic Perspectives 29 (2015) 3, pp. 3-30

AUTOR, D. H.; DORN, D.: The growth of low skill service jobs and the polarization of the US labor market. In: American Economic Review 103 (2013) 5, pp. 1553-1597

BAKENS, J. et al.: Methodiek arbeidsmarktprognoses en -indicatoren 2017-2022 [Methods for the labour market forecasts and indicators 2017-2022]. Maastricht 2018: ROA-TR-2018/4

BAKENS J.; FOURAGE, D.; PEETERS, T.: Labour market forecasts by education and occupation up to 2022. Maastricht 2018: ROA-TR-2018/3

BECKER, G. S.: Investment in human capital: A theoretical analysis. In: Journal of Political Economy 70 (1962) 5, pp. 9-49

CONLON, J. J.: Major malfunction: A field experiment correcting undergraduates' beliefs about salaries. Federal Reserve Bank of New York 2017

FOURAGE, D.: Veranderingen in werk en vaardigheden [Changes in work and skills]. Inaugural speech, Maastricht University 2017

HASTINGS, J. S. et al.: (Un)informed college and major choice: Evidence from linked survey and administrative data. In: Economics of Education Review 51 (2016) April, pp. 136-151

HUNTINGTON-KLEIN, N.: College choice as a collective decision. In: Economic Inquiry 56 (2018) 2, pp. 1202-1219

JENSEN, R.: The (perceived) returns to education and the demand for schooling. In: The Quarterly Journal of Economics, 125 (2010) 2, pp. 515-548

ROA: ROA opleidingsindeling 2015 [ROA educational classification 2015]. Maastricht 2016: ROA-TR-2016/3

ROA: De arbeidsmarkt naar opleiding en beroep tot 2022 [The labour market by education and occupation until 2022.] Maastricht 2017: ROA-R-2017/10

ROA/CBS. Beroepenindeling ROA-CBS 2014 (BRC 2014) [Occupational classification ROA-CBS 2014 (BRC 2014)]. Maastricht 2015: ROA-TR-2015/5

ROESE, N. J.; Summerville, A.: What we regret most... and why. In: Personality and Social Psychology Bulletin 31 (2005) 9, pp. 1273-1285

SANITER, N.; SCHNITZLEIN, D. D.; SIEDLER, T.: Occupational knowledge and educational mobility: Evidence from the introduction of job information centers. Economics of Education Review, 69 (2019), pp. 108-124

WISWALL, M.; ZAFAR, B.: Human capital investments and expectations about career and family (No. w22543). National Bureau of Economic Research 2016

MSc, Researcher at the Research Centre for Education and the Labour Market (ROA), Maastricht University, Netherlands

Prof. Dr., Research Centre for Education and the Labour Market (ROA), Maastricht University

Dr., Research Centre for Education and the Labour Market (ROA), Maastricht University


Translation from the German original (published in BWP 4/2019): Beverly Rudd, Exact! Sprachenservice, Mannheim