The effects of the abolishment of the basic
student grant on income, study results and
expenditures of students in the Netherlands
Master thesis
Teun Broeders Student ID: S1760378Supervisor: Prof. dr. M.G. Knoef Second reader: Dr. J. Been 7-8-2020
Master Thesis Public Administration Economics & Governance
Faculty of Governance and Global Affairs Campus The Hague, Leiden University
2
Abstract
This thesis researches the effects of the abolishment of the basic study grant. Effects on the income of students, the study results of students and their expenditures are researched. The goal of this paper is to extend the current monitoring of trends of the effects of the abolishment of the basic study grant by trying to identify causal relationships. This information could provide an insight in the current discussion in Dutch politics whether or not to adjust the policy on the financial aid system available to students in the Netherlands.
Main findings of this thesis are that the reform has had an influence on the (financial) situation of students. On the income side, parental contributions to students education have increased. No difference is measured in the labor income students earn. However, students take out study loans more often. No negative effect on the performance of students is perceived, as the exam grades do not change significantly. In the expenditures, a decrease can be spotted in the total expenditures and a lower percentage of the students s living away from home, possibly to keep their expenses down. On the basis of these findings, the reform seems to achieve the goals intended by the policy change and in this respect there seems to be no reason to reverse the reform, as is proposed by a majority of the political parties in the Dutch parliament.
3
Table of contents
1. Introduction ... 5
1.1 Motivation for the research ... 5
1.2 Research question ... 6
1.3 Reading guide ... 7
2. Institutional design ... 8
2.1 History of financial aid to students in The Netherlands ... 8
2.2 The pre-reform situation: Wet Studiefinanciering 2000 ... 9
2.3 The 2015 reform ... 10
3. Literature review ... 12
3.1 The rationale behind government intervention ... 12
3.2 Expected effects of the reform... 13
3.2.1. Income ... 15 3.2.2. Study results ... 15 3.2.3. Expenditures ... 16 3.3. Hypotheses ... 17 3.3.1. Income ... 17 3.3.2. Study results ... 18 3.3.3. Expenditures ... 19 4. Research method ... 20 4.1 Operationalization ... 20 4.2 Model ... 21 5. Data ... 23 5.1 Dataset ... 23 5.2 Descriptive statistics ... 24 5.3 Graphical evidence ... 25 5.3.1. Income ... 25 5.3.2. Study results ... 28 5.3.3. Expenditures ... 28
4
6. Estimation Results ... 30
6.1. Income ... 31
6.1.1. Parental contribution ... 31
6.1.2. Labor income ... 31
6.1.3. Labor force participation ... 31
6.1.4. Study loan ... 32
6.2. Study results ... 32
6.3. Expenditures ... 32
6.3.1. Total expenditures ... 32
6.3.2. Living away from home ... 33
7. Conclusion and discussion ... 34
7.1 Conclusion ... 34
7.2 Discussion ... 36
7.3 Suggestions for further research ... 37
5
1. Introduction
1.1 Motivation for the research
After a long period of bargaining between different parties, Jet Bussemaker, Minister of Education, Culture and Science, announced that the Dutch system of financial aid to students would be reformed at the start of the academic year 2015/2016. With support of D66 and GroenLinks, coalition parties VVD and PVDA finally had a majority in parliament in favor for the strongly contested law on study financing of Bussemaker (NOS, 2014). This meant the basic study grant was abolished and replaced by a study loan. The additional grant remained to exist for students of lower income families. In return, savings caused by the reform would flow back to higher education (Bakker, 2015). It was expected that the reform would save the government approximately one billion euros. These funds were to be invested in improving the quality of higher education. Opponents of the plans of Bussemaker countered that especially middle class families, who earned too much to be entitled to the additional grant, would be hit hard and that there would be unequal access to higher education for students of different socioeconomic groups. Bussemaker and other supporters of the reform, justified the policy with the argument that students, after completing their higher education, would earn a wage 1,5 to 2 times as high as their age peers who did not get higher education. This was a reason for the proponents of the policy to ask for private investments of students and their families and not to only rely on public investments in students in higher education (NOS, 2014; Regeerakkoord, 2012).
In the years after the implementation, support for the 2015 reform declined. In 2019, PVDA and GroenLinks, former proponents of the policy, switched sides. In 2020, D66 also turned against the new system of student loans, making the VVD the only party in favour of the policy. The parties which altered their opinion on the policy, along with other opponents of the policy, expressed their concerns that students experienced a lot of stress due to the burden of the debt they had built up during their studies. Students were under more pressure to finish their studies without any delays and this eliminated the time for side activities. On top of that, students were thought to postpone the decision to live away from home to save money. Finally, the parties said the investments in the quality of higher education were not as large as promised (NOS, 2020).
6 Statistics Netherlands (CBS) recently published an article on the policy partly supporting the claims of the opponents of the policy. The main conclusion of the research was that students took out more loans and lend bigger amounts of money. Also, the CBS notices an increase of parental contributions and side jobs to finance the increased costs of their study. Students studied faster and do not engage in side activities as often to reduce the study length and as a result decrease their study costs (CBS, 2019). On the other hand, Broek, Cuppen, Korte & Warps (2020) find the policy on study financing to be efficient and effective. In their independently conducted policy review, at the request of the Ministry of Education, Culture and Science, they also find that students borrow more often and bigger amounts after the policy has been revised. However, the researchers point out that studying is an investment of the student in him- or herself and it is thereby expected that the student and parents make a contribution to the education. Finally, both researchers of ResearchNed (Broek et. al., 2020) and researchers of the Netherlands Bureau for Economic Policy Analysis (Bolhaar, Kuijpers & Zumbuehl, 2020) find that the accessibility of different socioeconomic groups is not in danger and that the goals of the policy are largely achieved. Also, students in secondary education move on to higher education just as often as before the reform. Bolhaar et. al. (2020) state that they see no reason for (partially) reversing the policy changes.
1.2 Research question
As a student who started his studies in the academic year 2014/2015, I am confronted with the difference in the financial aid scheme provided by the government on a daily basis. From the academic year 2015/2016, the financial aid scheme was changed. The basic study grant was abolished and replaced by a study loan. It therefore sparked my interest to research the effects of the new system on students and to investigate if there indeed are distinct differences in the financial situation of students and their study performance after the reform as claimed by the opponents of the reform. Subsequently, the findings can be compared to the effects found by earlier conducted research on this topic to try and create a better understanding of the change in position of the political parties. This paper investigates the effects on the financial situation and performance of students by means of the following research question: What are the effects of the abolishment of study grants on the income, study results and expenditures of students in the Netherlands?
7 The effects will be investigated with data provided by the yearly Studentenmonitor survey. This survey questions Dutch students about study related behaviour. Trends in the choices, motivation and behaviour of students are looked at. However, the Studentenmonitor only looks at the trends of the different indicators. This research will also test the causal effects of the abolishment of the basic study grant on income, study results and expenditures of students.
1.3 Reading guide
This thesis will be structured as follows. First, the second chapter will discuss the institutional design to create a better understanding of financial aid schemes available to students in the Netherlands. The third chapter will give an overview of existing literature is provided. Theories about the rationale of financial aid schemes provided to students are discussed and expected effects study grants and study loans studied. These expectations are converted into hypotheses. In the research design (chapter 4), the method is explained which is used to test the hypotheses and to produce the results. A brief description of the dataset is given and graphical evidence is presented in chapter 5. Next, in chapter 6 the results of the regression model are examined and explained. Finally, the conclusion (chapter 7) gives an overview of the main findings and the limitations of this research and provides suggestions for further research.
8
2. Institutional design
To create a better understanding of financial aid schemes available to students in the Netherlands, this chapter provides a historical overview of the development of the system. Subsequently, an explanation of the basics of the current Dutch system are discussed. Also, the policy changes that were made to the system, of which the effect will be investigated in this research, are indicated clearly.
2.1 History of financial aid to students in The Netherlands
Already from 1815 onwards study grants are provided to students and continued to exist in different types and forms through the years (Slaman, 2015:17). In 1986, the foundation for the current system was laid with the Wet studiefinanciering, the law on study financing (Vossensteyn, De Boer & Jongbloed, 2015). The goal of the government was to make every individual student financially independent. By providing study grants, the idea was to make parental contributions unnecessary and thus making it possible for everyone, from all socioeconomic backgrounds, to follow a study in higher education. However, shifts in the political landscape and public opinion and the outbreak of economic crises caused the system of study grants to gradually become more sober (Slaman, 2015: 237-238; Oosterbeek, 1998). For example, from 1993 onwards the tempobeurs was introduced. This was a policy to encourage students to reduce the amount of years students spend to complete their studies. The government demanded students to acquire a minimum of 25% of available study credits in an academic year. When this demand was not met, the study grant would turn into a loan. In 1995 the study grant was reduced and tuition fees were increased as a measure of retrenchment. In 1996 the study grant was turned into a prestatiebeurs. This meant that every student received their student allowance first as a loan. This loan would then turn into a gift when the student succeeded to graduate and receive a degree in higher education within the prescribed nominal study length plus two additional years (Vossensteyn, De Boer & Jongbloed, 2015).
9
2.2 The pre-reform situation: Wet Studiefinanciering 2000
In the year 2000, the Wet studiefinanciering was revised and changed into the Wet studiefinanciering 2000 (Wet Studiefinanciering 2000, 2020). The allowance consisted of five components (DUO, z.d.-a; Oosterbeek, 1998):
• Basic grant • Additional grant • Student travel product • Study loan
• Tuition fee loan (from 2008 onwards)
The basic grant consisted of a monthly allowance available to all students. The amount of the allowance was equal for all students, only differing for students living with their parents and students living away from home. The basic grant for students living away from home was significantly higher, to compensate for the additional expenses on rent, food and other expenses. There was an additional grant available to students of families who had a joint income under a set level. These students received an extra allowance every month. The amount depended on the income of the parents. This policy was in place to assist the less financially able parents who were not able to contribute as much private investments to the education of their children, thereby guaranteeing equal opportunities of education to all (Oosterbeek, 1998). Every student eligible for financial aid from the government also got access to a student travel product, to be used in public transport to travel to and from their university or from their student house to their parental home during the weekend (DUO, z.d.-a). These first three allowances were allowances with requirements according to the prestatiebeurs rules. This meant that the basic grant, additional grant and student travel product are in principle a loan. If, however, the student completes his or her education and receives a diploma in higher education within ten years, the loan would be converted into a grant. The ten year requirement was more compliant than the rules in place before 2000, when students had to finish their studies within the nominal prescribed study time plus two additional years. Students who did not manage to complete education within ten years, were required to pay back the grant (Van den Berg & Van Gaalen, 2018).
10 Aside from the basic grant, additional grant and student travel product, students had the opportunity to take out a student loan under favorable conditions from the government. On top of the student loan, students also could take out a tuition fee loan from 2008 onwards. The amount of the tuition fee loan depended on the type of study and the corresponding tuition fee (DUO, z.d.-b; Vossensteyn et. al., 2015).
In this research, the situation of Wet studiefinanciering 2000, as described above, will be the system which the students in the control group have to deal with. Therefore this system will be referred to as the ‘old system’ in the remainder of this paper. This system was revised in 2015 (Vossesteyn et. al., 2015). The details of the ‘new system’, the post-reform system of which the effects will be investigated, will now be discussed.
2.3 The 2015 reform
The biggest difference between the old and the new system on financial aid to students is the abolishment of the first component of the allowance, the basic study grant. Students in the new system no longer have access to a monthly allowance provided by the government, which would turn into a gift if a diploma in higher education was achieved within ten years.
The other four components are still in place. The additional study grant is still provided to students whose families have a low joint income, to ensure equal accessibility to education for all children in the Netherlands. To compensate for the loss of the basic study grant for students with less financially able parents, the additional study grant is increased. However, students living away from home and eligible for the additional study grand are still worse off below the line in the new system. Students living with their parents and receiving the additional grant are slightly better off in the new system. The student travel product remains available to all students, just like the additional study grant to ensure the accessibility of education (Regeerakkoord, 2012; Van den Berg & Van Gaalen, 2018). Both the additional grant and the travel product are available to students during their nominal study length and are still converted into a gift if the students complete their study within ten years, as was the case in the old system (DUO, z.d.-b).
The student loan and the tuition fee loan are still available to students and have a more important role in the financing of the studies of a lot of students adhering to the new system. According to the government, not only the government itself should invest in the education of the youth, but also students themselves and their parents (Van den Berg & Van Gaalen, 2018).
11 The repayment term of the loans is extended from 15 to 35 years, to relieve the burden of the debt of students who now have to lend more money to finance their studies. The yields of the reform are invested in a ‘quality impulse’ of education and research in higher education, according to the coalition agreement (Regeerakkoord, 2012; Vossesteyn et. al., 2015).
12
3. Literature review
In this chapter a literature review will be carried out on existing theories on different ways students finance their education. First the rationale behind the decision of governments to provide study loans or study grants to students will be looked into. After that, previous empirical research will be examined on the effects of a study loan on the income, study results and expenditures of students. Finally, the expectations raised by previous research will be formulated into hypotheses, which will be tested later in this paper.
3.1 The rationale behind government intervention
Due to capital market failure and equity considerations governments can opt to introduce government intervention in the form of financial aid schemes for students (Oosterbeek, 1998). From 1975 to 2000, Cameron & Tauber (2000) notice an expansion of student aid programmes, with the underlying assumption that borrowing constraints present an obstacle for students to be able to access higher education. The authors state that this especially holds for lower income students, as they find a large positive correlation between family income and schooling attainment. Access to credits and returns to schooling prove to be important in schooling decisions and Cameron & Taber (2000) therefore think that it is important to relieve these borrowing constraints. However, the budget of a government is limited and governments can opt to increase the contributions made by students in the form of tuition fees or to decrease the generosity of financial aid schemes provided to students (Oosterbeek, 1998). In Western Europe a trend of increasing demand for student loans is observed by Lochner & Longe-Naranjo (2011), caused by budget cuts after the financial crisis and rising costs and returns of studying. As touched upon in the chapter Institutional design, this trend is also visible in the Netherlands: In the coalition agreement of 2012, the governing parties decided to change the Dutch system of financial aid schemes for students. Instead of a grant, the system changed into a loan. This decision was made to save money. This money would in return be invested in the quality of the educational system and research possibilities (VVD & PVDA, 2012).
13
3.2 Expected effects of the reform
The Netherlands Bureau for Economic Policy Analysis (CPB, 2013) expected the reform of the system to have significant effects on the amount of study loans used by students. The CPB predicted that study debts would increase with 6000 euro per student on average, but this can rise up to 9000 euro if the loss of the study grant were to be compensated completely by borrowing. However, the CPB expects no major income effects. Only low income students who would not have taken out study loans in the old system where study grants were provided, will experience income effects. This is expected to be a small group. Oosterbeek & Van den Broek (2008) found that students with the prospect of a good income borrow more easily. Students who have easier access to financial resources borrow less often. Another factor which affects the borrowing behaviour of students is debt-aversion. Students who do not like to take risks, have an aversion to borrowing. This is supported by the findings in the Monitor Beleidsmaatregelen 2016-2017 (Van den Broek, Wartenbergh, Bendig-Jacobs, Tholen, Duysak en Nooij, 2017). The Monitor Beleidsmaatregelen looks into the effects of implemented policies, like the abolishment of the study grant. The researchers find that the most important reason not to take out a student loan in the case of students with lower educated parents, are debt-aversion and principal objections to debts. In the case of students with higher educated parents, the main reason of not borrowing money is sufficient parental contributions. Van den Broek et al. (2017) note, like the study of Oosterbeek & Van den Broek (2008), that students with higher certainty about their future and their future income take out loans more easily and do so because of the favorable loan conditions. Insufficient parental contributions and the choice to work less, or not work at all, play a role in the choice to take out a student loan or not. The choice whether or not to take out a student loan is dependent on different factors. The most important factor possibly is the income a student has. The income of students consists of different sources. The income from jobs, parental contributions and social allowances are the biggest sources. Up until recently, students in the Netherlands could rely on a study grant provided by the government as a part of their income to partly finance their education. With that part of their income being nullified after the 2015 reform, the expectation is that this is compensated in some way with other sources of income to retain the same standards of living. Also, the reform could cause a reduction in the number of students. In the article of Oosterbeek & Webbink (1995), the consumption model of Kodde is introduced to show the effects changes in grants, parental income or other forms of income can have. This consumption model is shown in Figure 1.
14
Figure 1. Kodde’s consumption model framework
In the consumption model the amount of schooling (S) is measured on the horizontal axis. The vertical axis measures the consumption of goods by a student excluding schooling (Y). Line ABSmax represents the budget constraint: the total sum of income out of grants, loans and other sources of income available to the student for consumption of schooling. The indifference curve (I) indicates the optimal amount of schooling. When a student no longer receives a study grant, the sum of income will drop from line AB to line AB’. The question is whether students will try to bring back their income to the level of before the reform of line AB by increasing their income, for example by increasing the earnings out of work or receiving higher parental contributions. The other option is that they will not reach the optimal amount of schooling indicated by indifference curve I. This could mean that students do not continue their studies after receiving a bachelor’s degree or do not start a study in higher education at all. Oosterbeek & Webbink (1995) expect that the effects of changing the student grant into a student loan will only have a modest effects in the Netherlands, although the expected effects for students with low-income parents will be significant and measures could be needed for this small group to retain equal access to higher education.
15
3.2.1. Income
As stated above, effects for low-income students are most noticeable. Keane & Wolpin (2001) have investigated effects of parental transfers and borrowing constraints on educational attainment. They found that, apart from inherited abilities, higher educated parents invest more in the education of their children than parents with lower types of education. The authors say that this is due to the preferences for education of the parents and due to their higher wealth. Linsenmeier, Rosen & Rouse (2006) warn for the danger that study loans could crowd out these parental contributions and change the financial behaviour of parents. Lower parental contributions and higher costs of studying lead to an increase of the number of hours worked by students (Kalenkoski & Pabilonia, 2010). They also find that relieved borrowing constraints will inversely decrease the number of hours worked. On top of that higher private transfers lead to a shift of time spent studying and not in working (Bachman & Boes, 2014). Been, Knoef & Dalhuijsen (2020) studied the consequences of student loans on students’ spending and employment. The authors do not find the labor earnings of university students to be significantly affected by a change in student loan. They do find a substantial effect of the student loan on the parental contributions. If the student loan decreases with one euro, the parental financial contribution increases with 55 cents.
3.2.2. Study results
Warren, Lepore & Mare (2002) describe the relationship between working and studying as a zero-sum game. With this zero-sum game Warren et. al. (2002) mean that there is a direct trade-off between the hours worked and the hours spent on studying. This would imply that the more students will start working to compensate for the loss of income of the study grant, the less time they will have less time for studying. However, Oosterbeek & Van den Broek (2008) are less certain about this relationship. They find that the number of hours worked by students does not have a direct correlation to the hours spend on studying. A reduction of working hours only has a limited positive effect on the time spend studying. On the other hand, the statement of Bachmann & Boes (2014) that in the case of higher private transfers time is shifted to studying rather than working, raises the expectation that study results might improve when students receive higher parental contributions and reduce their working hours. This would imply that parental contributions could increase after the reform to compensate for the loss of income of the study grant and that this could free up more time for studying.
16
3.2.3. Expenditures
The reform of the financial aid scheme for students in the Netherlands means a significant cut in income. As discussed above, a student can choose to compensate this decrease of income by borrowing more, working more hours or be fortunate enough that his or her parents increase their contribution. Instead, it can also lead to a decrease in the spending of a student. Carneiro & Heckman (2002) support this by stating that borrowing constraints can lead to a change in the labor supply of a student or result in a change in the spending behaviour. According to a yearly study on expenditures of students carried out by Nibud, a Dutch institute educating people on financial matters, students spend on average 867 euro per month. Logically, the average spending of students living away from home is much higher: 1082 euro per month. The biggest expenditures are rent, groceries and tuition fees (Werf, Schonewille & Stoof, 2017). Been et. al. (2020) find that a one euro decline in student loan reduces the students’ expenditure by 36 cents. The students particularly decline housing and insurance expenditures. In a research performed for the CBS, Van den Berg & Van Gaalen (2018) also notice that after the abolishment of the study grant, less students are living away from home. The authors do not see differences between socio-economic groups compared to the situation before the reform. Students with relatively wealthy parents are still more likely to move away from home. The Monitor Beleidsmaatregelen (Van den Broek et. al., 2017), which analyzes the first effects of the 2015 reform, identifies an influence of the cost of commuting to school on the choice of school for students in lower education. This does not seem to have an effect on students in higher education. However, Frenette (2006) raises the expectation that the budget of a student and the availability of student loans can influence the decision of student regarding residency and commuting choices. Although Frenette’s (2006) research was carried out in Canada, where distances are a lot bigger than in the Netherlands, the costs of living and travelling could still play an important role in the change of expenditures of students after the abolishment of the study grant.
17
3.3. Hypotheses
Figure 2 presents the theoretical model of this research. As followed from the research question What are the effects of the abolishment of study grants on the income, study results and expenditures of students in the Netherlands?, this paper will look at three different effects of the abolishment of the Study grant by the government in 2015. The three effects are the effects on income, study results and expenditures of students. For each of the three effects one or two indicators will be looked at to check for effects after the 2015 reform. For each indicator a hypothesis will be formulated.
Figure 2. Theoretical model
3.3.1. Income
The consumption model of Kodde, as explained in the article of Oosterbeek & Webbink (1995), shows that the loss of income as a result of the abolishment of the study grant, will have to be replaced by some other form of income to attain the same level of education. Keane & Wolpin (2001) show that students with parents who have attended a higher level of education, invest more in their children and are more likely to increase their parental contributions after the study grant is abolished. This is consistent with the findings of Been et. al. (2020) that the parental financial contribution increases substantially if the student loan declines. This will be tested in the first hypothesis that students will receive higher parental contributions after the abolishment of the study grant.
Hypothesis 1
H0 = no effect on parental contributions
18 Parents of lower socio-economic classes might not be able or willing to increase parental contributions to compensate for the loss of the study grant. Kalenkoski & Pabilonia (2010) state that lower parental contributions or an increase in the costs of studying will lead to more hours of work. In this way the loss of the study grant can also be compensated. Therefore the second hypothesis is that students will increase their Labor income due to the abolishment of the study grant.
Hypothesis 2
H0 = no effect on labor income
H1 = increased labor income due to the abolishment of the study grant
3.3.2. Study results
The second factor which might be affected by the reform is the study results the students obtain after their study grants have been cut. The indicator with which we will try to measure the effect is the average exam grade. In previous empirical research, studied above in the literature review, different authors identify contradicting effects which could affect the exam grades. Warren et. al. (2002) expect a negative effect, because students will have to work more to achieve the same amount of income. These ideas are shared by Oosterbeek & Van den Broek (2008), but they see a much more limited, moderate negative effect between hours worked and time spend on studying. According to the article of Bachman & Boes (2014), the expected effect would be positive, because of the fact that private (parental) transfers might increase and therefore compensate the income of the students. These private transfers would then cause students to decrease their working hours and spent more time on studying. The third hypothesis therefore is that a change in the exam grades is to be expected due to the abolishment of the study grant. However, the direction of this effect is hard to predict.
Hypothesis 3
H0 = no effect on the exam grades
19
3.3.3. Expenditures
In contrast to increasing the sources of income to compensate for the loss of income of the study grant, the reaction of students can also be to cut their spending, as is seen in the research of Been et. al. (2020). Van den Berg & Van Gaalen (2018) perceive a decrease in the number of students living away from home after the 2015 reform. This is consistent with the expectations, as the monthly costs of living away from home are twice as much as the costs of living at home, according to the Studentenonderzoek 2017 of Nibud (2017). Frenette (2006) also expects an influence of study loans on the residency and commuting choices of students. Summarized, the decrease in income due to the abolishment of the study grant will most likely have an effect on the total expenditures of students. This is formulated in the fourth hypothesis.
Hypothesis 4
H0 = no effect on the total expenditures
20
4. Research method
In this chapter the abstract concepts following from the theory will be operationalized in measurable units. After that. the methodology of the research will be discussed. In the final part of this chapter, the dataset used to carry out this research will be presented.
4.1 Operationalization
The Dutch government already decided on the reform of the system of study grants in 2014, but the reform was in effect from 2015 onwards (Wet Studiefinanciering 2000, 2020). To study the effects of the reform in the most representative way, the research sample is defined as follows. Only students who have university education are included in the sample. Students attending applied sciences education (Dutch HBO-level education) are excluded. The percentage of HBO-students living with their parents is higher, lowering their expenditures and thus reducing the dependence on study loans. The respondents have an age between 17 and 25 years old and are at most in their fourth year of their studies. Older respondents are not included in the sample, because some of these respondents follow their study part time and have a main job on the side. Due to the higher income, their borrowing behaviour is not representative for this study.
Table 1
Definition of the treatment (black) and control group (white)
In Table 1 students are categorised in different groups depending on their year of study for the academic years 2011/2012 to 2016/2017. This is done for students in the first four years of their study. The first group impacted by the abolishment of the basic study grant were the cohort starting their bachelor program in the academic year 2015/2016. Students starting a new master
21 program also lost their right to a basic study grant from the academic year 2015/2016 onwards. Both groups no longer receiving a grant are colored black. It is important to notice that for students in their fourth year of their bachelor study there is no change. The reform only applied to students starting a new study program. When continuing the same study, students could still claim the full duration of the basic study grant when continuing to follow the same study program.
In the following year, the academic year 2016/2017, again the first year bachelor students and students starting a new master program in their fourth year do not receive a grant. Also, the students who started in 2015/2016 still do not receive a grant in their second year. The cells representing these groups are colored black. Third year students, who started in the academic year 2014/2015, still receive a basic study grant in 2016/2017. Again, fourth year students still in their bachelor studies also receive the basic study grant. Therefore the cells are not colored.
Because the data of 2016 are not available, we can only compare one year of post-reform period, 2017, with the years pre-reform. This means that the treatment period is the year 2017 and the control period 2012-2015.
The groups affected by the reform are the colored cells in Table 1: the first and second year bachelor students and fourth year master students in academic year 2016/2017. These students would receive a basic study grant in previous years, but in this year the grant was abolished. This group is assigned to the Treatment group. Third year and part of the fourth year students are not affected by the reform. They received the basic study grant in the control period, and continue to receive benefits in the treatment period. This group will be referred to as the Control group in the rest of this research.
As explained in the hypotheses, the effects of the abolishment of the basic study grant will be tested on different dependent variables, being Parental contribution, Labor income, Exam grades and Total expenditures. The variables Gender and Age will be added to control for the effect of the abolishment of the study grant.
4.2 Model
To investigate the four hypotheses following from the theoretical model, presented in Figure 2, four different regressions are executed. In this way the changes in the outcome variables
22 parental contributions, labor income, exam results and total expenditures as a result of the abolishment of the basic study grant are examined.
The effect will be tested using a difference-in-differences model. To carry out a difference-in-differences method, the assumption is made that two groups follow a parallel trend up to the moment of reform. Both groups adhere to the same rules and have the same access to the basic study grant in the control period before the reform. This is called the common trend assumption. After the reform, the treatment period, one of the groups receives a treatment and the other group does not. In this case the basic study grant is abolished for the students in the treatment group. For the control group, the same rules apply as did before the reform. When these two groups are compared before and after the moment of reform, the treatment effect can be measured.
The effects are estimated in an OLS regression, using the following model:
𝐷𝑒𝑝𝑉𝑎𝑟𝑖𝑡 = 𝛽0+ 𝛽1𝐺𝑖𝑡+ 𝛽2𝑇𝑡𝐺𝑖𝑡+ 𝛽3𝐺𝑒𝑛𝑑𝑒𝑟 + 𝛽4 𝐴𝑔𝑒 + 𝛽5𝑌𝑒𝑎𝑟2013 + (… )
+ 𝛽8𝑌𝑒𝑎𝑟2017 + ∈𝑖𝑡
𝐷𝑒𝑝𝑉𝑎𝑟𝑖𝑡 The dependent outcome variable of student i in period t
𝐺 Treatment group dummy
𝑇 Treatment period dummy
𝛽1 Treatment group effect
𝛽2 Treatment effect
𝑌𝑒𝑎𝑟2013 + (… )𝑌𝑒𝑎𝑟2017 Year dummies
∈𝑖𝑡 Error term
In the OLS regression model the effects are measured on the four dependent outcome variables. Four different regressions are run for the dependent variables Parental contributions, Income out of labor, Exam results and Total expenditures.
For each regression the most important effect is described by 𝛽2. This is the treatment
effect as a result of the imposed reform. This effect is controlled for by adding the control variables 𝐺𝑒𝑛𝑑𝑒𝑟 and 𝐴𝑔𝑒.
23
5. Data
5.1 Dataset
The dataset used in this research is the Studentenmonitor Higher Education, which is part of a broad European research project called Eurostudent. In this collaboration, different European countries study comparable data on socio-economic traits of students, accessibility of higher education and internationalization of education (ResearchNed, z.d.).
The Dutch Studentenmonitor, as used in this research, is a yearly survey set out by the Dutch Ministry of Education, Culture and Science. In the survey, students are questioned about all aspects related to studying. Students are asked to answer questions about their financial situation, their working life and other time consumption. The goal of the survey is to keep track of traits and study related behaviour of students and to try and discern trends in the choices, motivation and behaviour of students. Lastly the Studentenmonitor survey tries to establish effects of policy related to education, such as financial support, study results and behaviour of the student population in The Netherlands (Studentenmonitor, z.d.).
The Studentenmonitor Eurostudent research is set out yearly since 2000. To get a trustworthy insight in the developments of the accessibility of higher education and to generate usable information for the creation of policy on education, it is important that the research method is changed as little as possible over the years. Changes in the educational systems, like the introduction of the bachelor/master system in 2002, however, forced the Studentenmonitor to change. To maintain an as high as possible continuity, this research will use the Studentenmonitor data from 2012 onwards. This is done to keep the indicators and questions used as consistent as possible (Studentenmonitor, z.d.). At the moment of writing this paper, the data of the Studentenmonitor of 2016 was unavailable and thus excluded from this research. To achieve the most representative results, only students in the first four years of their studies and with an age between 17 and 25 are included in the sample. Also, this study only analyzes effects for students attending academic education (WO), not for students attending universities of applied sciences (HBO).
24
5.2 Descriptive statistics
In this part, the descriptive statistics of the sample are observed. In Table 2 the sample statistics are displayed, giving an overview of the number of respondents (N), the percentage of male respondents and the average age of the respondents. All these statistics are given for both the control group (Treatment=0) and the treatment group (Treatment=1).
Table 2
Descriptive statistics of the sample
Treatment 2012 2013 2014 2015 2017 Total N 0 1640 1287 1137 1562 1346 6972 1 2557 2433 1589 2802 2727 12108 Male (%) 0 38.2 34.5 31.2 36.3 35.3 35.4 1 34.9 31.2 29.0 32.9 34.4 32.8 Age (avg) 0 21 22 22 22 22 22 1 19 20 20 20 20 20
As shown in Table 2, the number of respondents varies, but in all years the number of respondents is high enough to perform statistical tests. The percentage of male respondents is lower than the number of female respondents, especially in 2014. However, the percentage of male respondents is fairly stable, accounting for approximately a third of responses over the years covered in this research for both the control and treatment group. The average age of the respondents is also very comparable over the years covered by this research. The age of the respondents in the control group is a bit higher compared to the treatment group. This can be explained by the fact that the respondents in the control group are 3rd and 4th year bachelor students. In the treatment group, 1st, 2nd and 4th year students who started their master studies are included.
25 In Table 3 the distribution of the respondents study years is presented for each year of the dataset.
Table 3
Distribution of respondents study years
2012 2013 2014 2015 2017 Total 1st year 976 1050 663 1065 1178 4932 2nd year 1204 1027 595 1227 1133 5186 3rd year 1135 887 812 1002 913 4749 4th year (4th year bachelor) 377 356 331 510 416 1990 4th year (1st year master) 505 400 325 560 433 2223 Total 4197 3720 2726 4364 4073 19080
5.3 Graphical evidence
5.3.1. Income
In the survey, respondents were asked about different sources income they received. Over the years the researchers changed these different sources of income that were included in the sample, making it difficult to compare the total income of student groups of different years. For this reason three specific types of income are analyzed in this research. Parental contribution, Labor income and Study loan are discussed, being important sources of income for students.
First, Parental contribution is inspected more closely. In the survey of Studentenmonitor, the students are asked about the monthly contribution in euros they receive from their parents. The trend of the amount of the parental contributions received by students are shown in the graph in Figure 3 for the control group (CG) and the treatment group (TG). It is interesting to observe that the trends of both groups run parallel up until the moment of the reform in 2015. After the reform, a very big difference can be noted. The parental contributions received by the treatment group, who do not longer receive a study grant, increase at a much higher rate. In the case of the control group, who still receive a study grant, there is only a minor change.
26
Figure 3. Development of Parental contribution
In Figure 4 the trendlines are shown for the monthly Labor income. It can be observed that both lines of the treatment (liTG) and the control group (liCG) roughly follow the same trend during all the years included in this research. However, the amount of income from labor for the treatment group is considerably higher than the amount of labor income the control group earns. This could have been caused by the fact that the treatment group consists of students who are in their third and fourth year of their study and a therefore on average older. This could possibly allow them to have higher earning jobs.
Figure 4. Comparison of development of Labor income (left axis) and Labor force participation (right axis)
27 Subsequently, in Figure 4 the Labor income of students (li) is compared with the Labor force participation (lfp), the percentage of students who have a job. This can be both incidental or permanent jobs. It is interesting to note that the percentage of students also follow the same trends for the labor force participation. A higher percentage of students in the treatment group (lfpTG) has a job, explaining the difference in income between the two groups.
The last source of income that is investigated in this research is the Study loan. Figure 5 shows that until the reform in 2015 the treatment group (TG) and the control group (CG) follow the same trend lines. The percentage of students with a study loan during that period is a little bit higher for the control group than for the treatment group, approximately 30% for the treatment group versus approximately 20% for the control group. After the reform an increase can be seen for both groups. However, the increase for the treatment group is a lot larger, surpassing the amount of students in the control group with a study loan. In 2017, approximately 60% of the students in the treatment group has a study loan, compared to approximately 50% for the control group. It seems that to compensate for the loss of the basic study grant, there is a sharp increase of students in the treatment group that take out a study loan.
28
5.3.2. Study results
To investigate the changes in the study results of students affected by the policy change and students who were not, the average Exam grades are compared in Figure 6.
Figure 6. Development of Exam results
The lines shown in Figure 6 are convergent from 2012 to 2013. From 2013 until the reform in the academic year 2015/2016, the lines of the control group (CG) and treatment group (TG) seem to follow more or less the same trend. The average exam grades increase steadily from 2013 to 2015. From 2015 to 2017, however, there is a sharp increase in the exam grades of the treatment group, which is not the case for the exam grades of the control group. This could indicate that the students in the treatment group who were no longer entitled to the study grant, studied harder from 2015 onwards to reduce their study length and thus study debt.
5.3.3. Expenditures
Figure 7 shows the Total expenditures for both the treatment group (expTG) and the control group (expCG) in euros per month. In the graph, there are yearly changes in the height of the total expenditures of students, but the trends of both groups seem to be roughly the same for the years 2012 to 2015. In 2017 a difference can be observed. The control group shows a clear increase of their total expenditures, whereas the total expenditures of the treatment group only shows a very small increase.
29
Figure 7. Comparison of development of Total expenditures (left axis) and Living away from home (right axis)
In Figure 7, the percentage of students Living away from home (afh) is described as well. This is done to check if there is an effect of the reform on the amount of students living with their parents longer and delaying their move away from home. Over the years, no clear change can be observed in the percentage of students living away from home for both the treatment (afhTG) and control group (afhCG). The expectation raised in hypothesis 4 that students affected by the reform decrease their spending compared to the control group seems to be confirmed. However, this effect seems not to be caused by the residency decision.
30
6. Estimation Results
In this part the results will be discussed for each outcome variable independently. For each indicator, it is assessed if the null hypothesis can be rejected or not. Table 4 presents the main regression results for all dependent outcome variables.
Table 4 Regression results Parental contri-butions Labor income Labor force partici-pation Study loan Exam results Total expendi-tures Living away from home Abolishment study grant 61.96*** (10.09) 3.50 (8.70) 0.007 (0.158) 0.171*** (0.017) 0.037 (0.026) -23.38** (11.46) -0.028* (0.017) Gender 2.43 (4.23) -0.94 (3.65) 0.063*** (0.007) -0.240*** (0.007) 0.015 (0.011) 1.15 (4.70) 0.075*** (0.007) Age 1.99 (1.66) 21.44*** (1.43) 0.000 (0.003) 0.015*** (0.003) 0.005 (0.004) 45.22*** (1.83) 0.046*** (0.003) Year 2013 27.61*** (6.50) -39.52*** (5.61) -0.217** (0.010) -0.013 (0.010) -0.007 (0.016) -15.35** (7.18) 0.000 (0.011) Year 2013 8.89 (7.04) -31.23*** (6.07) 0.165 (0.011) -0.002 (0.011) 0.042** (0.018) 24.17*** (7.79) 0.003 (0.011) Year 2013 9.94 (6.25) -34.35*** (5.39) -0.024** (0.010) 0.031*** (0.010) 0.032** (0.016) -10.35 (6.91) -0.034*** (0.010) Year 2013 44.47*** (9.11) -1.70 (7.86) -0.027* (0.014) 0.198*** (0.015) 0.053** (0.023) 49.99*** (10.29) -0.041*** (0.015) Treatment group -7.10 (5.21) -8.54* (4.49) -0.038*** (0.008) 0.062*** (0.008) 0.057*** (0.013) -35.28*** (5.77) -0.040*** (0.008) Constant 110.03*** (35.62) -264,79*** (30.72) 0.694*** (0.055) 0.032*** (0.058) 7.016*** (0.089) -372.21*** (39.45) -0.335*** (0.058) N 18,101 18,102 19,080 19,080 19,080 19,080 19,080
31
6.1. Income
6.1.1. Parental contribution
In the first column of the regression results shown in Table 4 the effect of the abolishment of the study grant on Parental contribution is displayed. With a coefficient of 61.96 this is a very big effect. The coefficient indicates that the students in the treatment group, who do not longer have the right to receive the study grant, received 62 euros more parental contributions per month than students in the control group who were still receiving the study grant. The coefficient is very significant, with a p-value smaller than 0.01. This means this effect is not based on coincidence. The control variables Gender and Age were included, but are both not significant.
This means that for hypothesis 1, the null hypothesis can be rejected and the alternative hypothesis can be accepted. There is a clear increase observed in the Parental contribution caused by the abolishment of the study grant.
6.1.2. Labor income
The effect of the reform on the Labor income of students is very small, with 3.50 euro per month. In addition the coefficient is not significant and therefore the measured effect could possibly be based on coincidence. Control variable Age is significant and has a positive effect. With each year the age of the respondent increases, Labor income increases with 21 euros. The second control variable, Gender, is not significant. Based on these results, the effects of the abolishment of the basic study grant on the Labor income of students is not significantly different from zero. This means that for hypothesis 2, the null hypothesis cannot be rejected.
6.1.3. Labor force participation
The variable Labor force participation is a dummy variable, varying between 0 (no labor force participation) and 1 (labor force participation). Following from the model, there is almost no change in the Labor force participation after the 2015 reform, with an increase of only 0,6%-points. This coefficient is not significant, so the small effect could be based on coincidence. Control variable Gender has a positive effect and is significant on the p-level of 0.01. This means that male students have a 6%-points higher Labor force participation than female students. The second control variable, Age, is not significant.
32
6.1.4. Study loan
There is a very strong and statistically significant (p<0.01) increase in the dummy variable Study loan. Among students who do not longer have the right to the basic study grant study loans increase with 17%-points. Both control variables are also significant on the 1% p-level. Student loans are taken out by female students more often. For the control variable Age no effect can be measured.
6.2. Study results
In the regression results of the average Exam grades there can only be observed a very small positive effect of 0.037 points on a 10 point scale. On top of that the coefficient is not statistically significant, so there is a chance the measured effect is based on coincidence. The control variables Gender and Age are not significant as well.
Because the coefficient is very small and not significant, the null hypothesis of hypothesis 3 can not be rejected. There is no effect of the abolishment of the basic study grant on the Exam grades of students.
6.3. Expenditures
6.3.1. Total expenditures
The Total expenditures of students in the treatment group after the reform decrease with 23.38 euro. This is a significant effect with a p-level smaller than 0.05. Age has a strongly significant effect (p < 0.01) on the Total expenditures of students. For each year the students are older, the Total expenditures increase with 45.22 euro. For the second control variable, Gender, a negligible effect is measured and this effect is not significant.
The last null hypothesis, hypothesis 4, can therefore be rejected as there is a clearly significant effect of the reform on the Total expenditures of the treatment group. Students who lose the right to receive the basic study grant cut their spending by almost 24 euro per month on average.
33
6.3.2. Living away from home
The variable Living away from home is a dummy variable, students living with their parents are described with a 0 and students living away from home with a 1. The coefficient is slightly negative, indicating that after the reform a little less students live away from home. The effect measured is significant, but only on the 10% level (p<0.1). The control variables Age and Gender are both significant on the 1% level (p<0.01). The percentage of male students Living away from home is 7%-points higher than for female students. For each year the students are older, the percentage of students Living away from home increases with almost 5%-points.
34
7. Conclusion and discussion
This research tries to extend the existing monitoring of trends of the effects of the abolishment of the basic study grant by trying to identify causal relationships. Effects on the income of students, the study results of students and their expenditures were researched. For each indicator a hypothesis was formulated and in this section the main findings are presented. Afterwards, limitations of this research will be discussed and pointers for further research will be provided.
7.1 Conclusion
One of the most clear conclusions that can be drawn from this study is the effect of the abolishment of the basic study grant on the parental contribution that is received by students. A very strong effect is measured and the alternative hypothesis, that there is an increase in the parental contributions after the reform, is accepted with a strong, statistically significant result. The students in the Treatment group, who do not have the right to the basic study grant, receive 62 euros more parental contributions per month than the students in the control group, who do receive the basic study grant. The aim of the government to attract more private investments of students and their parents for their education instead of public investments seems to work.
For the second source of income examined, the labor income of students, no effect can be identified. The null hypothesis of Hypothesis 2 can not be rejected. It seems that students affected by the reform do not start working more and increasing their labor income now they no longer receive an important source of their income through the basic study grant. Compared to students who still receive the basic study grant, no significant changes can be examined. The same observation can be made when inspecting the labor force participation more closely. There is no significant difference spotted in the percentage of students who have a job before and after the reform. Furthermore, the distribution between students with an incidental job and students with a permanent job does not change. As the income of students is threatened to decrease by the abolishment of the basic student grant and is not compensated by students working more and having a higher labor income, it is expected that other sources of income will be increased. Another source of income which can be used to ease the decrease in total income is the student loan offered by the government. The results show that a sharp rise in these student loans can be noted. A significant increase is seen for students in the treatment group, affected by the abolishment of the basic grant. They take out a student loan 17% more often compared to students who were not affected by the reform. This shows that students rely on a student loan
35 to compensate for the loss of income due to the abolishment of the basic study grant and do not start working more. This could be explained due to a lack of time available to students to spend on working extra hours. It could be possible that students try to shorten the study length and in that way reduce the costs of studying as much as possible. A last remark that can be made on the income of students, is that a higher percentage of the male students have a job to finance their studies. On the contrary, female students take out study loans more often.
Hypothesis 3 examined the effect of the reform on the study results. This was done by looking at the average exam grades of students and comparing them for the Treatment group and the Control group. Although there seems to be a small positive effect on the exam grades of the Treatment group, this effect can not be presented with certainty due to the fact the regression results are not statistically significant. The effect is potentially caused by coincidence. However, the suspicion that students shorten the lengths of their studies as a result of the abolishment of the basic study grant could still be true.
The last effect investigated in this research was the effect of the abolishment of the basic study grant on the expenditures of students, as formulated in Hypothesis 4. This effect is negative and significant, so the null hypothesis can be rejected and the alternative hypothesis is accepted. This means that respondents in the treatment group, dealing with the abolishment of the basic study grant, reduce their spending with more than 23 euros per month compared to the students in the control group, who have access to the basic study grant. The age of the respondents is an important indicator for the results. Older students have higher expenditures. For each year increase in the age, the average total expenditures of a student increases with 45 euros per month. The cut in spending that can be observed can be partly explained by the decrease of the percentage of students living away from home. This number declines by 3%-points after the abolishment of the basic study grant, as was expected based on the existing literature and the first trends measured after the reform. As students have lower total expenditures, it would be expected that students remain living with their parents for a longer period of time as a way of limiting their expenditures. It then is expected that students also reduce their expenditures on other sources, but in this research no source can be identified where students save money. Possibly, students live more soberly and weigh their expenses more carefully due to the abolishment of the basic study grant.
To summarize, the abolishment of the basic study grant has had an effect on the (financial) situation of students. More private investment are required to get access to higher education. These investments are compensated by increased parental contributions, but it appears that students do not decide to work more. Instead, students take out study loans more
36 often and they spend less by cutting their expenditures. No negative effects on the performance of the students can be identified in this research. On the basis of these findings, the reform seems to achieve the goals intended by the policy change and in this respect there seems to be no reason to reverse the reform, as is proposed by a majority of the political parties in the Dutch parliament.
7.2 Discussion
This research is carried out with data collected under Dutch university students. This means it is difficult to generalize the findings for students in HBO education. Also these findings only apply to Dutch students, as students of other countries are confronted with different systems of financial aid schemes provided to students.
Unfortunately, there is a hiatus in the data used in this research. The Studentenmonitor survey was conducted in 2016, but the data was not available at the time this research was carried out. This is an important year in the data, as it was data on the first year the reform was carried out, and thus the first year of the Treatment period that could have been investigated. This could have caused the trend, as seen between the control (pre-reform) period and the treatment period, to be more pronounced then it might have been. Further research including the data on the academic year 2015/2016 is encouraged.
In addition, the results could have been influenced by the distribution of the control group and the treatment group. The first students who were affected by the reform started their first year of their studies in the academic year 2015/2016, but the data of this year were unavailable to use in this research. The data of the post-reform year 2017 were used and in this year, first and second year students and students starting their master degree in their fourth year were affected by the reform and therefore included in the Treatment group. The Control group consisted of the students unaffected by the reform in that year. This meant that the Control group only included third year students and some fourth year bachelor degree students, resulting in a higher average age of this sample group. This could have influenced the result as older students possibly have higher earning jobs. These assumptions could be checked for by including 2016 in the research and more future yearly Studentenmonitor surveys as the age distribution of the Treatment and Control group will be more similar.
Lastly, because the type of questions used in the Studentenmonitor survey over the years the information about the income of students changed slightly also. Therefore it was not possible to compare the total income of students over the years, as not all sources of income are
37 measured in each year of the survey. It now was possible to only compare the income from different sources like parental contributions and the percentage of students taking out student loans, but not the trend in the total income of students.
7.3 Suggestions for further research
This research can be extended and followed up by checking if the effects measured are consistent over future years. The Studentenmonitor survey is carried out on a yearly basis for the Ministry of Education, Culture and Science. This allows for a comparison between more pre-reform and post-reform years. It is interesting to investigate if the trends, as found in this research, continue to be present or if the effects measured are just temporary changes in the trends caused by the introduction a new financial aid system. Also, with the data of Studentenmonitor of the academic year 2017/2018 a new comparison can be drawn between a Treatment group, the third year students now affected by the reform, and the Control group. A check can be performed if the same effects can be identified. The data of the Studentenmonitor 2018 is meanwhile published, but not yet included in this research.
Furthermore, this research does not look into the effects of the accessibility of higher education. Opponents of the reform of the abolishment of the basic study grant feared that higher education would only be accessible for the rich. The first studies on this topic, such as the study carried out by the Bolhaar et. al. (2020) for the CPB, find no decline in the number of students proceeding in higher education after finishing secondary education and the accessibility of higher education for lower socioeconomic groups seems not to be in danger. However, as equal access to higher education for all is very important, this should be watched closely and frequently investigated further in the future.
Finally, the reform was carried out in the first place to save money on student grants and invest these funds into the quality of higher education and into research. This research did not look at the success of these investments as the results of these investments are not yet known. It would be very interesting to research in the future and to check if the saved spending on student grants were carried over wholly to the promised quality impulse of higher education and to see if effects are visible.
38
8. References
Bachmann, A. & Boes, S. (2014). Private transfers and college students’ decision to work. Economics of Education Review, 42, 34-42.
Bakker, M. (2015, january 20). Sociaal leenstelsel wordt ingevoerd: dit gaat er veranderen. De Volkskrant. Retrieved from https://www.volkskrant.nl/nieuws-achtergrond/sociaal-leenstelsel-wordt-ingevoerd-dit-gaat-er-veranderen~b8b915ea/
Been, J., Knoef, M. & Dalhuijsen, D. (2020). Early-life financial behavior, intergenerational transfers, and employment: Insights from a nudge in student loan policy. Retrieved from
https://www.netspar.nl/assets/uploads/P20200615_DP017_Been.pdf
Berg, L. van den, Gaalen, R. van (2018). Studeren en uit huis gaan nog haalbaar? Retrieved from
https://pure.uva.nl/ws/files/29496370/2018ST01_Studeren_en_uit_huis_gaan_nog_haalbaar.p df
Bolhaar, J., Kuijpers, S. & Zumbuehl, M. (2020). Effect Wet studievoorschot op toegankelijkheid en leengedrag. Retrieved from
https://www.cpb.nl/sites/default/files/omnidownload/CPB-Policy-Brief-Effect-Wet-studievoorschot-op-toegankelijkheid-en-leengedrag.pdf
Broek, A. van den, Wartenbergh, F., Bendig-Jacobs, J., Tholen, R., Duysak, S. & Nooij, J. (2017). Monitor Beleidsmaatregelen 2016-2017. Retrieved from
http://www.studentenmonitor.nl/rapporten/monitor-beleidsmaatregelen-2016.pdf?
Broek, A. van den, Cuppen, J., Korte, K. de & Warps, J. (2020). Beleidsdoorlichting artikel 11 Studiefinanciering. Retrieved from
https://www.researchned.nl/wp-content/uploads/2020/07/111906040-Beleidsdoorlichting-studiefinanciering.pdf
Cameron, S. & Taber, C. (2000). Borrowing constraints and the returns to schooling. NBER Working Paper Series, 7761.
39 Carneiro, P. & Heckman, J. (2002). The evidence on credit constraints in post-secondary schooling. The Economic Journal, 112, 989-1018.
CBS (2019). Studenten lenen vaker en meer. Retrieved from https://www.cbs.nl/nl-nl/nieuws/2019/41/studenten-lenen-vaker-en-meer
CPB (2013). Aflossing studieschuld bij sociaal leenstelsel. Retrieved from https://www.cpb.nl/publicatie/aflossing-studieschuld-bij-sociaal-leenstelsel
DUO (z.d.-a). Oud stelsel hbo en universiteit. Retrieved from https://duo.nl/particulier/oud-stelsel-hbo-en-universiteit.jsp
DUO (z.d.-b). Studiefinanciering. Retrieved from https://duo.nl/particulier/studiefinanciering/index.jsp
Frenette, M. (2006) Too far to go on? Distance to school and university participation, Education Economics, 14(1), 31-58.
Kalenkoski, C. & Pabilonia, S. (2010). Parental transfers, student achievement, and the labor supply of college students. Journal of Population Economics, 23(2), 469-496.
Keane, M. & Wolpin, K. (2001). The Effect of Parental Transfers and Borrowing Constraints on Educational Attainment. International Economic Review, 42(4), 1051-1103.
Linsenmeier, D., Rosen, H. & Rouse, C. (2006). Financial aid packages and college
enrollment decisions: An econometric case study. The Review of Economics and Statistics, 88(1), 126-145.
Lochner, L. & Monge-Naranjo, A. (2011). The nature of credit constraints and human capital. The American Economic Review, 101(6), 2487-2529.
NOS (2020, July 3). Evaluatie leenstelsel: geld lenen is geen drempel om te studeren. Retrieved from https://nos.nl/artikel/2339441-evaluatie-leenstelsel-geld-lenen-is-geen-drempel-om-te-studeren.html
