The claims are so serious that education minister Robbert Dijkgraaf says he wants a “far-reaching” investigation into the way student finance body Duo carries out the checks, which are partly based on algorithms, NOS reported.
NOS and Investico spoke to 32 lawyers who have represented students accused of lying about where they lived over the past 10 years. Students who live away from home are entitled to a higher grant, to help offset living costs.
Of the 376 cases the lawyers were involved in, 97% involved students with ethnic minority roots. The figures are “shocking”, law professor Gijs van Dijck told Investico.
Duo has used an algorithm to look for potential fraud since 2012 based on potential risk indicators such as age and education and trained using the “experience” of Duo investigators. Once a suspect has been identified, Duo fraud investigators decide if they should be checked out, and they take other factors into account.
For example, youngsters who live with family members, such as a brother or aunt, form a potential risk, even though they are entitled by law to a grant as a live-out student – some €200 more a month than for home students.
Once a suspect has been identified, Duo carries out home checks, talks to neighbours and decides if fraud has been committed. It then contacts the student demanding the cash be repaid and possibly a fine on top.
“The word ‘student’ is not written on my forehead,” hbo college student Mohammed, who helped pay for his studies by driving a taxi, told Investico. “People think a student is a blonde lad with blue eyes. I would come home late at night, in my taxi driver suit.”
Over the past 10 years, 10,000 students have been accused of fraud, NOS said. Of them, 6,000 have appealed and 20% of them won. There were also almost 1,500 court cases, of which the students won 25%.
Lawyer Gülsüm Çekiç, who has supported several students accused of fraud, said she questioned some of the indicators used by Duo. “For example, students with ethnic minority roots are more likely to live with a brother or aunt,” she said.
The process is open to prejudice, Lucas Haitsma, who specialises in algorithms and discrimination, told NOS.
“If you look at a specific group, you are only going to find fraud,” he said. “And if you base conclusions on that data, then it will only intensify your tunnel vision. The idea that these youngsters are more likely to commit fraud is then confirmed.”
Duo told NOS that the algorithm does not use ethnicity or nationality as a basis. “We don’t know what a student’s background is and we do not select on that basis,” a spokesman said.
The spokesman said Duo had not carried out any official evaluation of its fraud strategy since 2012 and that the recent childcare benefit scandal, in which ethnic minority parents were accused of fraud based on an algorithm, was no reason to re-evaluate their own methods.
The benefit scandal resulted in tens of thousands of Dutch parents being incorrectly accused of fraud and unjustly ordered to pay back thousands of euros in childcare benefit by the Dutch tax office. The compensation process is still ongoing.
Since 2015, student grants have only been available for students from low-income households and students at vocational training colleges (mbo) but they are set to be reintroduced for all students from next year.
The new basic grant is €274.90 per month for all students who live out and €110.30 for those who live at home. College and university students from households where the annual income is less than €70,000 will also be able to apply for a top-up grant of up to €416 per month.
In addition, all live-out students will be given an extra monthly payment of €164.30 to offset the cost of higher energy and food bills during the next year.