Reference estimate

Ministry of Education, Culture and Science

Forecast of national numbers of pupils and students in the Dutch funded education system

Last change on 8th of October 2024, at 7:16 (CET) | Publication Standard 1.0

• Publication category: Other algorithms
• Impact assessment: Field not filled in.
• Status: In use

General information

Theme

Education and Science

Begin date

04-2011

Contact information

[e-mail]

Link to publication website

https://www.rijksoverheid.nl/documenten/rapporten/2024/04/24/referentieraming-ocw-2024

Responsible use

Goal and impact

The Reference Forecast shows the expected development of the numbers of pupils and students at institutions in funded education in the Netherlands. The estimate primarily serves to substantiate the education budget of the Ministry of Education, Culture and Science (OCW).

Besides underpinning the budget, the estimate is used to describe recent and expected developments in education. The annual report explaining the Reference Estimate and www.ocwincijfers.nl are the main products for this purpose.

Finally, the Reference Estimate is also used as input for other estimates. For some of these estimates, data from this reference estimate up to 2070 is provided. The main users for follow-up estimates are:

- OCW's study funding estimate;

- institution forecasts from DUO;

- labour market estimates from Project Education Labour Market by ROA;

- education staff labour market estimates by Centerdata;

- estimates for student housing by ABF Research;

- the Mobility Assessment by KiM, the Dutch Knowledge Institute for Mobility Policy.

Considerations

The Reference Estimate has proven accurate enough to underpin the budget of the Ministry of Education, Culture and Science. Also, the Reference Estimate is not too complex: this allows a new estimate to be calculated in reasonable time. This is important because there is limited time to make the Reference Estimate. There are several months between receiving the latest data and delivering the estimate for budgeting purposes. However, an incorrect estimate can lead to a deficit or surplus in the budget for all of OCW.

No personal data are used to prepare the Reference Estimate and no individual educational careers are included. Only total numbers of individuals in education streams are used. For example: total number of children who went from group 7 to group 8.

Because individual educational careers are not included, valuable information might be missed in the current estimate. An alternative to the current model would be, for example, a microsimulation model, which does include information on individual pupils and students. However, this method is more complex than the current method. This requires more computer power, making the process slower. It also requires detailed data, where data quality is more difficult to guarantee. More complex behaviour can be part of a more detailed estimation model, which can improve the quality of the estimate. It just also takes more time in performing the estimation. Because of the time constraints for preparing a new estimate, these vulnerabilities are not desirable.

Human intervention

One of the most valuable aspects of the algorithm is its steerability. This means that the estimate can be adjusted by an analyst based on additional information and take into account expected developments in education, such as effects of policies and the impact of unemployment on educational participation.

For steering the reference estimate, advice is sought from education experts. Based on their advice, a decision is made on how and how much to steer.

Making the reference estimate is an iterative process, with the results being analysed after each step. First, an estimate without controls is made. Then, one by one, controls are added. At each step, it is checked whether the controls generate unexpected or unrealistic outcomes.

Once the iterative process is completed, the estimate is submitted to the so-called CLR (Pupil Estimates Coordination Group). The CLR includes policy directorates of the Ministry of Education, Culture and Science. They are the policy directorates of the various sectors (primary education, secondary education, etc.). They look critically at the forecast and advise the Knowledge Director of the Ministry of Education, Culture and Science to adopt the forecast.

Only after the forecasts are set are they actually used for the ministry's budget.

Risk management

• When making the estimate, a 4-eye check is performed. This involves 2 analysts separately using the same agreed settings to make the estimate. They then check whether the results are the same. This 4-eye check is important to avoid errors during the estimation process.
• Before any changes are made to the methodology, advice is sought from the Pupil and Student Estimating Supervision Committee (BLS). This committee consists of experts from, among others, CBS, SCP, CPB and educational organisations. The tasks and composition of the BLS are laid down in a formal establishment decision, which can be found here: https://wetten.overheid.nl/BWBR0047356/2022-10-25.
• The projections are coordinated with the Coordinating Group for Pupil Estimates (CLR). The CLR includes the policy directorates of the various sectors (such as primary education, secondary education, etc.). These are policy directorates of the Ministry of Education, Culture and Science. From their domain expertise, they can advise on the results.

Legal basis

Preparing the budget of the Ministry of OCW is a public interest task. Without a budget, the Ministry of OCW would not be able to serve the public interest in all its ways. Article 105(1) of the Constitution describes that a State budget must be prepared. Section 2.5 of the Comptabiliteitswet 2016 describes that an estimate is a required part of the budget. Section 15 of the WRO describes that basic data from the education participants register may be used for budget preparation by the minister. The data in the education matrix are all taken from (the basic data from) the register of education participants and are used to prepare the estimate.

The education matrix is created by DUO based on the education participant register (ROD). The original collection purpose of the data in the ROD is to provide data for a variety of purposes, which includes budget preparations (see WRO Article 5). The purpose purpose of processing the education matrix and reference estimate is budget preparation.

Links to legal bases

• Artikel 105 van de Grondwet: http://wetten.overheid.nl/jci1.3:c:BWBR0001840&hoofdstuk=5&paragraaf=2&artikel=105
• Artikel 2.5 van de Comptabiliteitswet 2016: http://wetten.overheid.nl/jci1.3:c:BWBR0039429&hoofdstuk=2&paragraaf=2&artikel=2.5
• Artikel 5 van de Wet register onderwijsdeelnemers: http://wetten.overheid.nl/jci1.3:c:BWBR0042012&hoofdstuk=2&paragraaf=2.1&artikel=5

Operations

Data

• CBS's Population Forecast: this forecast contains the expected numbers of births, deaths, immigration and emigration.
• The education matrices from DUO's Information Products Department: the education matrices show the numbers of people per flow in, through and out of education. The education matrices are made on the basis of the enrolment data of educational institutions and provide insight into how students and pupils moved through education in a given year.
• CPB's unemployment forecast. This is used to estimate what proportion of MBO students take a convex course and what proportion take a blended course.

Links to data sources

• Bevolkingsprognose CBS: https://www.cbs.nl/nl-nl/longread/statistische-trends/2023/bevolkingsprognose-2023-2070-minder-geboorten-meer-migratie
• Macro Economische Verkenning CPB: https://www.cpb.nl/macro-economische-verkenning-mev-2024

Technical design

The reference estimate is carried out using an age-dependent flow model. This estimates flows of people into, through, out of and out of education to arrive at expected, future numbers of pupils and students.

The starting point for the reference estimate is the education matrix. This lists the numbers of persons flowing into, through and out of education. From this, transition probabilities, the probability of someone moving from one level or grade of education to another, are derived. The numbers of individuals and probabilities are used to estimate future flows.

For example: if 70 out of 100 vwo6 students moved on to wo last year, the transition probability is 70% to move from vwo6 to wo. Suppose the current number of vwo6 students is 130, then the model assumes that 70% of them (91 students) will go on to WO the following year.

Basically, the most recent transition probability is used to estimate future numbers. Of some transition probabilities, the trend of recent years is considered. These are the probabilities of 'strategic transitions'. These are transitions between education types that are most affected by policy changes, such as from primary to secondary education and from secondary education to further education. The transition probabilities for these flows are determined with extrapolation based on the last 12 observations (years). However, these extrapolated transition probabilities are first evaluated. In case of good reason, the extrapolation of a particular transition probability is waived.

The estimate grows with the population in the Netherlands. Using the CBS population forecast, the number of births is taken into account and corrected for mortality, immigration and emigration.

For more information, see section 5.1 of the public report at https://www.rijksoverheid.nl/documenten/rapporten/2024/04/24/referentieraming-ocw-2024