My briefing to the Special Rapporteur on Contemporary Forms of Racism, Racial Discrimination, Xenophobia and Related Intolerance
Please contact me for a PDF-copy with footnotes (not shown in this text).
Written submission (deadline: 15 February 2020) in response to:
End of Mission Statement of the Special Rapporteur on Contemporary Forms of Racism, Racial Discrimination, Xenophobia and Related Intolerance at the Conclusion of Her Mission to the Kingdom of the Netherlands, The Hague (dated 7 October 2019)
Written & submitted by Tom van Messel LL.M.
The theme of this submission is limited to one of the themes highlighted in your Statement: the discourse about the (non-western) migrant background of citizens of the Netherlands.
In particular, this submission is limited to government-produced communication about crime and people with a (non-western) migrant background.
You wrote (in §15):
“Of course, there is nothing inherently offensive or discriminatory about the designation of being of a non-western migrant background. Rather, the problem is that racial and ethnic minority Dutch citizens in the Netherlands, through this continuing emphasis in public and political discourse on the non-western origin of their descendants and ancestors, become branded as perpetual foreigners”.
I would like to share a number of facts and views with you, in order to help documenting information that I think is relevant in this respect. By means of this submission, I would like to make a case for the following argument:
The government does not deal responsibly with its own communication about crime and government-made country-of-origin categorisations. It has not conducted research into the (likely) negative effects of its communication choices about this topic. Without such research, the weighing exercise between the value of this communication (and underlying statistics) and the likely negative effects cannot take place.
First of all, I would like to share this quote with you, from Patrick Simon’s “Ethnic” statistics and data protection in the Council of Europe countries: A Study Report, published by European Commission against Racism and Intolerance (ECRI) in 2007:
“Dangers of abuse and misuse
Cases in which persecution has relied on statistical files compiled for other purposes, or for that very purpose, are sufficiently well documented to remove all doubt as to their having occurred. Illicit uses include 1) identification of individuals with reference to characteristics which may expose them to discrimination, exclusion or even persecution 2) the use of statistics to stigmatise a vulnerable group. In the first case, the main dangers inherent in the collection of data and creation of files relate to the possibility of their being used to identify individuals. The data protection laws which regulate the collection, production and dissemination of personal data attempt to control and minimise those dangers. In the second, the problem is less the possibility of identifying individuals, than the isolation of variables which characterise individuals or groups. Here misuse takes the form of stigmatisation, e.g, when stereotypes are confirmed by using findings reductively, or publishing tables without explaining them or analysing the factors which account for discrepancies. This is particularly true of statistics on crime, the prevalence of behaviour regarded as deviant or social problems treated more as burden on the community than as disadvantage for those directly concerned”.
“….or publishing tables without explaining them or analysing the factors which account for discrepancies”.
This quote is quite suitable to describe how the majority of the Dutch government-publications have presented crime numbers per country of origin.
In the past ten years, I found that (at least) 49 publications were published by government organisations in which crime numbers were presented broken down in country-of-origin categories. Of these 49 publications, only 7 showed to what extent a range of social-economic factors can explain the discrepancies in crime rates, between different country-of-origin groups.
DESCRIPTION OF THE RELEVANT GOVERNMENT PUBLICATIONS
At the end of the 1980s, the Research and Documentation Centre of the Ministry of Security and Justice has started publishing reports in which registered crime (suspect) figures have been broken down into a selection of groups on the basis of country of origin. Since 2005, CBS started including second-generation categories in their publications about crime rates. The second-generation regards inhabitants born in the Netherlands, who have at least one parent who has been born in one of the countries on which the statistic categories are based: Morocco, Turkey, Netherlands Antilles, Suriname.
Furthermore, at least one CBS-publication was issued in which two of the four country-of-origin categorisations were used for crime suspect numbers among third generation inhabitants (inhabitants with at least one grandparent born in either Suriname or the Antilles).
As an illustration: in the course of 2016, government bodies published five reports in which crime suspect rates were categorised into a selection of countries of origin.
One of these publications is the annual report on integration, published in November 2016. An example of a presentation of the figures in this report:
What can be seen in this bar chart — above — is that four groups of inhabitants are separately shown, on the basis of their country of origin.
The country-of-origin categorisations are government-made, on the basis of objective criteria: the country of birth of oneself and/or one’s parents and/or grandparents.
Within these categories, most inhabitants possess Dutch nationality.
Another thing that can be seen in the above-shown bar chart is that it is broken down into two different generations and that it focuses on a specific age bracket. Another chart in this particular publication is broken down into male and female suspects (per country of origin).
As mentioned above, only 7 of the 49 publications between 2009 and 2020 have shown how crime differences between country-of-origin groups correlate to differences in social- economic status. With the help of so-called multivariate analyses, some of the publications include an attempt to answer the following question: to what extent does the factor ‘country-of-origin’ uniquely contribute to ‘the chance of being a crime suspect’ after the effects of other variables (that overlap with the predictive value of the factor ‘country-of-origin’) have been removed? The result of applying these variables is that the unique contribution of the factor ‘country-of-origin’ as an explaining factor becomes smaller. In other words, the crime suspect rates are being ‘corrected for relevant personal attributes’.
Variables that have been taken into account in publications showing the effect of social-economic variables include: (parents’) dependency on state benefits, ‘ethnic’ composition of the neighbourhood, income, family situation/composition and degree of urbanisation and sometimes - various education-related aspects.
Important to note here: the government only has a limited amount of social-economic data available.
The following bar chart taken from a 2016 publication from the Netherlands Institute for Social Research (SCP) serves as an illustration of the outcomes of such an exercise:
Among those publications in which multivariate analyses are shown, a number of publications provide theories about the question why there is still overrepresentation in crime (of certain country-of-origin groups) after correction for a number of social-economic variables.
These theories are based upon qualitative research as opposed to the quantitative approach in the multivariate analyses. The publications repeatedly refer to research on topics like the ‘matrifocal family structure of the Antillean’, ‘hustle culture among Surinamese’ and ‘cultural dissonance’ among Moroccan and Turkish youth.
In publications like these, disclaimers can often be found such as “coincidental circumstances and socio-economic background characteristics that we have not included in the analysis are probably also of influence“ .
Some of the government researchers have published such publications in English: ‘The involvement of different ethnic groups in various types of crime in the Netherlands’ and ‘On the deviant age-crime curve of Afro-Caribbean populations: The case of Antilleans living in the Netherlands’.
These two publications can serve as examples of the seven (Dutch-language) publications aimed at ‘explaining overrepresentation’ (contrasting with the 42 publications that merely provide numbers without further explanation).
On top of the publications described above, annual crime numbers are available 24/7 online, on ‘Statline’. These crime number tables show the crime numbers for a range of ‘national origin’ population categories:
• Dutch background
• Migrant background
• Western migrant background
• Non western migrant background
• Unknown
• Morocco
• Turkey
• (former) Netherlands Antilles
• Suriname
• Other non western migrant background
Visitors of this Statline website can furthermore break crime numbers down in various types of crime, and a range of suspect attributes: age, sex, level of urbanisation, possession of a Dutch home address and nationality.
One cannot break down these suspect numbers into categorisations based upon (other) social-economic variables (like education or poverty standards).
The screenshot below is meant as an illustration of the Statline crime number tables (red markings and text added by me):
MOTIVES BEHIND THE STATISTICS & COMMUNICATION, AS DECLARED BY THE DUTCH GOVERNMENT
The government has expressed a number of different reasons for creating statistics on country-of-origin and crime and reporting about it.
First of all, the reporting of these statistics has been motivated by the government’s wish to separate fact from fiction and to counter stigmatisation caused by the spreading of false information about the “nature and magnitude of crime rates among minorities”.
Another reason for the numerous publications to exist is related to the alleged need to acquire data for different purposes. Under the umbrella of this motive the following objectives can be distinguished (not entirely mutually exclusive).
Developing targeted policies aimed at enhancing integration (as ‘crime’ is seen as an indicator and/or consequence of a lack of integration); the association between crime and ethnic minorities is seen as harmful and hindering integration; reducing overrepresentation of certain country-of-origin groups in crime. This includes measures in the repressive sense and increasing the chance of apprehension. For these objectives, the government deems the following activities necessary:
Conducting research with the aim of gaining knowledge about the mechanisms that lead to change in behaviour and in order to “investigate the nature of the relationship between criminality and integration”. Also: gaining knowledge about the predictive value of particular (combinations of) factors.
Another motive that can be identified as a separate one is naming/articulating problems, although this could also be understood as a necessary condition for the before-mentioned research objectives.
Furthermore, various holders of public office have made remarks about the selection of countries-of-origin categorisations that are to be included in the publications:
“Because of pragmatic considerations, the analysis is predominantly aimed at the four biggest minority groups, namely groups and persons of Turkish, Moroccan, Antillean and Surinamese origin”.
Another indication of why some country-of-origin categorisations are consistently highlighted and others not: “We no longer register the problems with the Moluccan community, but we do notice that problems again occur among the third and fourth generations. For me this gives cause to once again consider the possibility of registering incidents in that group”.
By the way, the selection of the four country-of-origin categorisations is not based upon the size of the groups. For example, inhabitants of German or Indonesian origins reach numbers comparable with the four groups that are separately shown.
CRITICISM: A NUMBER OF CONSIDERATIONS
In random order, non-exhaustively (and not entirely mutually exclusive):
a) The vast majority of the publications do not show or mention how crime (in general) correlates to social, economic and psychological factors. The question is how do those 42 publications over the past 10 years (as referred to on page 2) contribute to the self-declared motives of the government?
b) Overrepresentation of certain country-of-origin groups in crime is the point of departure in a considerable share of the publications, instead of reporting on (the explaining factors for) all crime.
c) By including crime in a range of yardsticks aimed at measuring ‘integration’, the government provides citizens with an instruction how to understand the government’s reporting on crime and country-of-origin categorisations.
d) The majority of the publications do not (exhaustively) list the disclaimers associated with the crime numbers per country of origin (e.g. no fiscal offences, no data on police profiling, limitations to the amount of data for the purpose of multivariate analyses).
e) A number of publications draw conclusions about the alleged cultural characteristics associated with a higher risk of becoming a crime suspect. The cultural factors mentioned include: lack of loyalty to the majoritarian culture, a lack of affinity for social skills needed for integration and cultural dissonance (defined as the phenomenon that migrants live their lives between two strongly differing cultures). At the same time, the publications drawing such conclusions do not reflect on the question as to how the qualitative studies that form the basis for such ‘culture’ theories have been selected.
f) In the past 10 to 15 years, no (visible) attempts have been made to expand the number of variables to be included in the multivariate analyses (note that the current amount of data already reduces the unique contribution of the variable ‘country of origin’ to the dependent variable ‘having been a crime suspect’ to close to zero for some of the country-of-origin categorisations).
g) The abundant attention the government gives to crime in relation to people’s country of origin contrasts with the lack of attention given to actually relevant (underlaying) factors that correlate strongly with crime. This is reflected in the limited number of government publications about all crime and correlated factors (as opposed to publications focusing on ‘overrepresentation’ in crime of certain country-of-origin groups).
h) Lack of evaluation, nr. 1. No evaluation has taken place to assess the degree to which the country-of-origin categorisations in crime statistics have been indispensable for policy making.
i) Lack of evaluation, nr. 2. There are no attempts at tracking possible harmful effects of 20+ years of reporting on overrepresentation of ethnic groups in crime. No government evaluation has taken place examining what citizens remember or understand from the different types of government reporting on crime and country-of-origin categorisations. An example of how the country-of-origin categorisations in crime rates are understood (and/or justified) by some people (in this case police officers): “a clear majority of respondents finds the use of statistical data about ethnic origins accepted for considerations concerning decisions who to check or stop. 79 percent of respondents finds the use of statistical knowledge justified. A police stop/check on the basis of ethnicity is viewed as justified by 44 percent of respondents if the ethnic background is overrepresented in statistics about crime offenders”:
j) Lack of evaluation, nr. 3. There have been no government attempts to assess the degree to which citizens consent with being categorised for the purpose of government reporting on crime overrepresentation.
k) Lack of evaluation, nr. 4. The Dutch government has never made attempts to measure the support for the application of country-of-origin categorisations in crime statistics and communication among different groups in society. There are indications, however, of the lack of support among people with a (non-western) migration background. “More than half of all immigrants think the government should stop registering ethnic origins, though only a third of native Dutch people share this view”, was the conclusion of a small non- representative internet-survey.
l) Lack of evaluation, nr. 5. The Dutch government has not exhausted its options to assess the different types of government communication against international standards — in particular article 8 of the European Convention on Human Rights. The following section of this submissions provides a number of quotations from international institutions that are relevant to the question to what extent the government’s communication choices interfere with certain (human) rights.
RELEVANT QUOTATIONS FROM COUNCIL OF EUROPE INSTITUTIONS AND EP
Among the different Council of Europe institutions, it is the European Commission against Racism and Intolerance (ECRI) that has been most specific in expressing concerns about the Dutch government’s country-of-origin categorisations. In 2007 and 2013, it dedicated the following remarks to the issue:
“a considerable amount of information is available in the Netherlands broken down by “allochthony” — “allochthonous” persons [. . .] However, the extent to which this information is used to inform policy aimed at improving the situation of those persons who are found to be at particular disadvantage is not clear to ECRI, especially as social policies in recent years have been targeted less and less at specific ethnic minority groups. Instead, it has been stressed that such information is rather used to target security measures at particular minority groups. [. . .] ECRI stresses the need for such data to be used to monitor patterns of discrimination or situations of disadvantage facing minority groups. It should not be used for purposes that contribute to further stigmatising the members of such groups”.
“ECRI recalls that the collection of data in different areas of life broken down according to categories such as national or ethnic origin is legitimate if it is done to monitor patterns of discrimination or situations of disadvantage facing vulnerable groups, provided that it is carried out with due respect to the principles of confidentiality, informed consent and the voluntary self-identification of persons as belonging to a particular group. Any other type of ethnic registration would be dangerous, contrary to human rights and prone to misuse”.
“The authorities have informed ECRI that the CBS regularly publishes statistical tables on the living conditions of allochtones in various fields of life. They have further stressed that this system of data collection is not based on the principle of self-identification of persons because only a small part of “ethnic minorities” identify themselves as belonging to a certain ethnicity. Furthermore, they wish to track information for extended periods of time, whereas tracking a person’s cultural orientation over long periods of time would prove to be very difficult. ECRI notes that this system of collection of data, does not respect the principle of informed consent and the voluntary self-identification of persons as belonging to a particular group”.
Besides these remarks about Dutch government reporting, there are a couple of relevant quotations that apply generally:
“‘stigmatisation’ shall mean the labelling of a group of persons in a negative way”.
“the governments of the member states, public authorities and public institutions at the national, regional and local levels, as well as officials, have a special responsibility to refrain from statements, in particular to the media, which may reasonably be understood as hate speech, or as speech likely to produce the effect of legitimising, spreading or promoting racial hatred, xenophobia, anti-Semitism or other forms of discrimination or hatred based on intolerance”.
“Union law on the protection of privacy and personal data, the right to equality and non-discrimination, as well as the right of individuals to receive information about the logic involved in automated decision-making and profiling and the right to seek judicial redress are applicable to data processing when processing is preceded by pseudonymisation techniques or, in any case, when the use of non-personal data might impact on individuals’ private lives or other rights and freedoms, leading to the stigmatisation of whole groups of the population”.
PRESS RELEASE FROM THREE ORGANISATIONS CALLING FOR CHANGES
<<PRESS RELEASE: Three organisations from the former National Consultation Platform on Minorities call for changes to government communication on crime and countries of origin
Published: 9 April 2019
In a letter sent yesterday, 8 April 2019, the organisations Ocan (formerly known as the Dutch Caribbean Consultative Body) Cooperation Platform Moroccan-Dutch (SMN) and the Consultative Council of Turks in the Netherlands (IOT) have called on various cabinet members to reconsider how the government communicates about crime and the country of origin. Since the 1990s, numerous government publications have been published in which suspect figures are presented, classified according to a selection of countries of origin. If one of your parents is born in a certain country, you are classified as a separate group in government communication about — for example — crime rates. In at least one publication, groups were also broken down based on the country of birth of grandparents.
John Leerdam, chairman of Ocan:
“For more than twenty years, government organisations have been presenting crime rates broken down by country of origin, in all kinds of publications. But what is the added value of this? That still remains vague! Government research on crime is about uncovering underlying factors such as economic position, level of education or upbringing. Categorisations by origin are by no means necessary. It is almost impossible that this type of communication has not in some way had adverse effects for Dutch-Antilleans, for example for the position on the labor market. It is therefore time to talk about this in a society that advocates inclusion. Especially in view of equal opportunities for our youth, we believe that people’s country of origin should not be mentioned every time we talk about crime.”
The three organisations propose a number of concrete steps. Among other things, the organisations ask for a clear and complete written reconsideration, in case the responsible cabinet members and government organisations choose to continue with communications of this type.
Bouchaib Saadane, chairman of SMN:
“A few years ago, the terms ‘allochtonous’ and ‘autochtonous’ were banned from government communication, supported by extensive government assessment. One of the arguments then was: negative effects on public perception and prejudices. Without much explanation, suspect figures are still being broken down by country of origin. Why would the government’s considerations look differently for country-of-origin categorisations than for the terms ‘allochtonous’ and ‘autochtonous’? And shouldn’t you learn a little more about how these figures are perceived by citizens?”
In the past twenty years, governments have not reconsidered the continuation of the origin breakdowns in crime figures as a separate topic, nor has there been any research into the reinforcing role of government communication on negative perceptions. Such an assessment would also have to pay attention to the criticisms of ECRI, the European Commission against Racism and Intolerance, according to the three organizations. In both 2007 and 2013, ECRI emphasised that the Netherlands violates a norm with the Dutch methods of categorising and presenting data on the basis of country of origin.
If in the future government bodies continue with the country-of-origin breakdowns in suspect figures, then groups with the same socio-economic characteristics should be compared in all of the publications, according to Ocan, IOT and SMN. The vast majority of current publications that refer to suspect figures and countries of origin do not meet this condition.
Substantial progress should then be made in the publications that remain, according to the three organisations. They call on the government to collect more data with which different origin groups can be made more comparable. In addition, they call for more work to be done on research into, for example, ethnic profiling, the influence of (friends) networks and the influence of daily negative attitudes about people’s ethnic origins.
Zeki Baran, IOT chairman:
“Research into factors that correlate with crime does not have to start with the question why there are differences between different origin groups. If you omit that question, you can still acquire exactly the same knowledge about criminogenic factors. The starting point must be that criminogenic factors are characteristics that occur throughout the world, regardless of the origin group. A clear exception to this is selectivity in police arrests. We therefore welcome it if this type of selectivity deserves more attention in crime analyses and data gathering”.>>