The prosecution of bias crime in the Netherlands and the problem of net-widening: fundamental limits to criminal liability | SpringerLink
This service is more advanced with JavaScript available, learn more at http://activatejavascript.org
The prosecution of bias crime in the Netherlands and the problem of net-widening: fundamental limits to criminal liabilityAllard RingnaldaRenée KoolOpen AccessArticle
Crimes that are committed with bias motives are categorised as ‘hate’ or bias crimes and are punished more severely than nonbias crimes. However, bias crime laws are often applied to offences where there is no clear evidence of a bias motive. Based on the results of 318 case studies into bias crime prosecutions in the Netherlands, this paper demonstrates that the causes of net-widening should be sought in the action-oriented nature of criminal law reasoning. Decision makers rely on objective behavioural indicators to infer motives. However, these are rarely reliable. We argue that this process results in a transformation of bias crime laws. They are no longer used to punish harmful motives. Rather, they are used to combat behaviour that is considered socially harmful on account of its perceived intolerant, racist or xenophobic message. This forces us to reconsider the justification behind trying to punish motive.
fulltext PDFOf the many conceptual and practical difficulties of bias crime laws [, ], net-widening is particularly urgent from a legal, normative point of view. What makes bias crimes different from ‘ordinary’ crimes and what justifies penalty enhancement is the presence of a bias or prejudice motive as a cause of the offence – the fact that the offender acted because of his dislike of a particular racial, ethnic, religious or otherwise ‘other’ group of people. The European Union (EU) 2008 Framework Decision on combating certain forms and expressions of racism and xenophobia and the Council of Europe (CoE) 2002 Recommendation on national legislation to combat racism and racial discrimination both recommend more severe punishment for offences that are committed with a ‘racist or xenophobic motivation’ [, ]. However, it has been observed that such penalty enhancement is frequently applied to offences that are not motivated by bias [, , ]. Then, the penalty enhancement loses its justification.The literature identifies the causes of net-widening as police officers, prosecutors and judges being inadequately trained or otherwise disinclined to recognise and pursue bias crimes [, , , , ]. Others have pointed to the difficulties that decision makers experience in proving bias motives, which is a novel and alien element in criminal law and specific to bias crimes [, , , ]. We want to push this last point a bit further. In this article, we suggest that punishing motive conflicts with the action-oriented nature of criminal law reasoning. Reporting the results of 318 case studies of offences that were prosecuted as bias crimes in the Netherlands [], we show that decision makers – police officers, prosecutors and judges – try to work around the problematic concept of motive by relying on more objective and observable criteria to identify bias motives. Not only does this lead to net-widening, it actually transforms the concept of bias crimes to something that is no longer related to the punishment of motive.It should be noted at the outset that we use normative, legal discourse and the peculiarities of legal reasoning to explain a phenomenon that might at first seem to be essentially criminological and empirical. However, the normative and the empirical are intrinsically related. It is both the law and legal reasoning that have the definite say on what is and what is not a bias crime, whatever the perceptions of victims and communities may be. It is also the law that prescribes how a potential bias crime should be prosecuted, proved and sentenced, and it is a specifically legal kind of reasoning that is employed in doing so. Therefore, our understanding of the problems in prosecuting and sentencing bias crimes is likely to benefit from an internal, legal perspective on the decision making process.The article opens in section two with an overview of the European and Dutch legal frameworks on bias crime. Section three outlines our case study method in which we used a categorisation of cases to say something about the presence and role of bias motives in individual cases. In section four, we show how Dutch bias crime cases that were prosecuted between 2000 and 2004 can be categorised. In section five, we argue that substantial net-widening occurs in the prosecution of Dutch bias crimes. In section six, we analyse how the effect of net-widening can be explained. The article concludes in section seven with a critical discussion of what this means for our current and future understanding of dealing with social problems of bias by means of criminal law.Bias or ‘hate’ crime is a concept that originated in the United States to describe crimes that are committed with a prejudice or bias motive. Bias crime has been introduced into the European legal sphere through the EU Framework Decision and CoE Recommendation, though many European states had previously included bias crime provisions in their laws (as did the Netherlands). It has been addressed at the European level to express the importance of tolerance [, ]. The wording of the relevant preparatory works and policy documents resonate a historical fear by terming racism and xenophobia as a ‘scourge’ that troubles Europe [:item 15], causing the need to create ‘a culture of tolerance, embracing both State and society’ [:item 6]. Besides recommending the introduction of a wide range of tolerance-related offences – such as holocaust denial, inciting violence or hate against groups, and bias speech – the policy documents also suggest that ordinary crimes should be punished more severely if committed with a bias motive. Though the European policy refers to racism and xenophobia, it should be taken to include ‘otherness’ more generally. The Recommendation defines six categories of target groups of bias, based on race, colour, language, religion, nationality, and national or ethnic origin []. The Framework Decision states descent, religion and hatred as the three key elements. States are allowed to designate additional target groups, for instance those based on social status or political convictions [:article 7a]. In the Netherlands, the bifurcated approach of the European instruments is followed. There is a distinction between statutory bias crimes and bias-motivated crimes. The first category consists of specifically defined behaviour that is considered to involve bias without requiring any proof of the offender’s bias motives. These are included in the penal code as offences in their own right. They include incitement to hate (S. 137d Dutch Penal Code (DPC)), distributing hate speech (S. 137e DPC), financing or participating in organisations that discriminate against particular target groups (S. 137f DPC), and the discrimination against target groups in the course of professional activities (S. 137g DPC). The most prevalent statutory bias crime is bias speech (S. 137c DPC). Bias speech is defined as intentionally insulting a target group through speech, writings or symbols, expressed in public. The relevant target groups are those defined by race, religion and faith, sexual preference, and physical or mental handicap. The main difference with common, non-bias penal insults is that negative qualities are attributed to a target group as a whole rather than an individual that belongs to the group. Simply calling somebody a ‘stupid Turk’ would amount to common insult, which carries a maximum penalty of 6 months imprisonment, or a fine (S. 267 DPC). However, if it is clear from the context that the author suggested that the whole group of Turkish people is stupid, the offence can be classified as bias speech with a maximum penalty of 1 year.The second category of bias crimes are bias-motivated crimes. These consist of any sort of common, already defined crime (e.g. assault, murder or vandalism) that is committed with a bias motive. Although the Recommendation and Framework Decision recommend that a bias motive be included in the penal code as an aggravating circumstance, the Dutch legislature has opted for taking measures to ensure that courts take account of the bias motive when determining the sentence []. There is no basis for bias-motivated crimes in the statute, but prosecutors are required through prosecutorial guidelines to increase their sentencing demand by at least 50% in case of a racial or xenophobic motive []. The guidelines do not specify what is meant by bias crime other than stating that these involve ‘aspects of bias’, usually in the form of a ‘bias motive’. The relevant target groups are again defined on the basis of race, religion and faith, sexual preference and handicap. Although the judge is not required to follow the prosecutor’s sentence demand, he often does and can only deviate if he can justify his decision to do so.The distinction between the two types is important for the way in which cases are prosecuted and judged. A statutory bias crime such as bias speech is proven if all elements of the statutory definition are present in the offender’s behaviour. The motive of the offender need not be proven. Bias-motivated crimes, by contrast, do require that the prosecutor prove that the offender acted out of a bias motive. From a legal point of view, the objectively observable behaviour of a bias-motivated crime does not differ from that of its nonbias parallels. The crime of assault, for instance, consists of elements such as physical contact, the lack of consent, and intent, whether or not the assault is motivated by bias. However, to justify a penalty enhancement, the prosecutor will need to prove not only that all elements constituting the offence were present, but also that the behaviour was motivated by bias or prejudice against one of the target groups. In contrast to statutory bias crimes, proving a bias-motivated crime requires the prosecutor to do more than apply the law to the observable facts. The prosecutorial decision-making process may thus be expected to differ significantly between both types.What both types share, however, is the idea that bias crime forms a distinct problem that warrants stronger punishment than similar crimes that do not involve bias or prejudice. From a legal point of view, this penalty enhancement requires a without this, there would be a risk of unequal treatment of offenders by applying different levels of punishment to similar cases. Furthermore, the principles of culpability and legal certainty require that the distinctive features of bias crimes are clear, well-defined, foreseeable, and avoidable: offenders must be able to be aware that their behaviour could, under certain conditions, be considered a bias crime.It is widely acknowledged in the literature that the bias or prejudice motives of the offender justify more severe punishment and constitute the defining difference between bias and non-bias crime. Two strands of normative theory can be identified []. The first argues that the presence of a bias motive causes greater harm. This harm can exist in a greater degree of victimisation [:88], harm to the target group, or harm to society as a whole by undermining fundamental values of equality and pluralism [:342–348]. The second argues that a bias motive is a greater moral wrong, which justifies an enhanced penalty []. Either way, the legal rationale for punishing bias more severely depends on the presence of a bias in the mind of the offender. There must also be a strong link between that bias and the offence. An offender can only be ‘guilty’ of bias if he actually decided to act on his sentiments. The bias must therefore be a motive in the literal sense, a primary causal reason for the offence. The offender must have come to his actions because of his disdain or dislike of a particular target group of which the victim was a member.Net-widening occurs if offenders are prosecuted for and convicted of a bias crime while there is no motivational relation between bias or prejudice (if any) against a target group and the offence committed. Whether net-widening occurs is ultimately a normative question, depending as it does on the type of bias or prejudice that we consider harmful and the strength of the causal link that are required to justify enhanced punishment. In order to allow a discussion on what we consider to be net-widening, we developed a categorisation that takes account of the wider context of the offence and the characteristics of the offender and thus can aid our discussion on whether, in our opinion, the cases evidenced sufficient motive and causality to justify enhanced punishment from a legal-normative point of view. We developed our categories on the basis of typologies and classification schemes that have been suggested in literature. Lawrence has pointed out that one should distinguish between three broad categories of offenders: those whose acts have been motivated those who have some racist or xenophobic sentiments that they do not usually express, but may nonetheless emerge in situations of anger and those who have committed what appears to be a bias crime without being aware that they have done so [:65–72]. Jacobs and Potter, too, have argued that one should distinguish between lesser and higher degrees of prejudice and stronger and weaker causal links [:145, ]. These models do not, however, provide any criteria on how concrete cases can be classified. A more evidence-based typology of bias crime offenders is offered by McDevitt et al., who distinguish between motives of thrill, defence, retaliation and mission [, ]. However, this typology is based on cases that have already been marked by the police as hate crimes without acknowledging the possibility of marking errors, and it is therefore not directly suitable to discuss net-widening.We conducted 318 case studies of offences that were prosecuted as bias crimes between 2000 and 2004 in three urban and two rural jurisdictions in the Netherlands. We studied the court files, which usually contained police reports, notes on the prosecutor’s assessment of the case, reports of the evidence, arguments and pleas for the prosecution and the defence, information on the offender (sex and age, previous convictions, and sometimes a psychological or social worker’s report), and the judge’s decision. In the inquisitorial system of Dutch criminal procedure, this court file – the dossier – serves as the basis for the judge’s decision. It should contain all information relevant to the case, including material for the defence. If arguments or evidence are submitted at trial, these too should be included in the dossier. In effect, the dossier ought to contain all relevant information that was available to police, prosecutor and trial judge. Therefore, we were able to review independently on what grounds decision makers concluded that the degree of bias and strength of the causal relation were sufficient to warrant penalty enhancement. Based on the information in the dossiers we studied and while continuously comparing our impressions with the literature, we developed the following case categories. Bias crimes occurring in conflicts between private citizens. Such conflicts can be of an incidental or more structural nature. In both cases, a conflict unrelated to any racist or xenophobic incident, was most likely the primary cause for the chain of events that led to the crime. The bias motive, if present, would only be contingently related to the crime, emerging only as a result of the conflict. In our opinion, there was no evidence of a malignant belief or sentiment that could have encouraged the offender to commit the offence. Signs of bias occurred only on the side, and were not, or not clearly, ancillary to the offence.
Bias crimes occurring in conflicts with public authorities. The offender was prosecuted for a bias crime that was committed against a person who was charged with some public authority, usually a police officer or an officer of public transport. The bias motive is again likely to be subsidiary in the causal chain of events, emerging as a consequence of the conflict rather than being a cause of it.
Bias crimes occurring as the result of malignant bias and/or ideological beliefs. There is clear evidence that the offender had malignant bias sentiments or convictions, opinions and beliefs. In some cases, it is even clear that the offender held a coherent set of negative beliefs about a target group that amounted to an ideology. These malignant sentiments or ideological beliefs were a primary reason for him to commit the crime: had it not been for his strong bias, the crime would most likely not have happened. Bias must clearly have been a primary motive.
Bias crime occurring without any apparent motive. Though the file contains sufficient information about the context of the offence and the offender, there is no particular evidence of a bias motive, but there is also clearly no suggestion that other motives (as is the case in conflict situations) are likely to have played a role. In these cases, one feels that the crime was committed without any motive, occurring out of the blue.
Other contexts, including work, entertainment or sports-related offences and thrill seeking. In these contexts, the motive is less clear. There may be a malignant bias or ideological motive involved, but work-related issues (stress, anger, conflicts) or thrill-seeking (for instance among party-goers or sports fans) are also likely to have played a significant role.
If the files contained no information whatsoever relating to the context of the offence or offender, we marked it as ‘unknown’. The difference with crimes committed without apparent motive is that the unknown category was used if the files contained too little documentation to be able to say anything about the context of the offence or offender.
It is important to note that the aim was not to develop exclusive categories. As the categories overlap, some cases could be included in more than one category. If so, a decision was made according to the most likely motive. We always adopted a high threshold, requiring that a motive be proven beyond reasonable doubt, as the prosecutor would be required to do. To warrant interrater reliability, we discussed all 318 case studies in order to decide on a def the initial disagreement rate was less than 10%. In order to increase the external validity of our findings, we referenced bias crime literature that reports comparable categorisations in other countries.Through the case studies and the prosecution case information database, we collected a wide variety of variables besides case categories. We also tracked various offender characteristics, circumstances of the offence, known variables about the victim, use of drugs and alcohol, previous convictions, et cetera. Variables were coupled to discover significant relations using chi-square analysis.We also used the case-study method to obtain information about the possible causes of any net-widening we discovered. Studying the case files allowed us to identify criteria that decision makers had used and the process that was applied to mark an offence as a bias crime. As the case studies yielded only indirect information on decision-making criteria (the case files did not usually contain an explicit explanation on why a case had been considered to be a bias crime), we also conducted individual and group interviews with 14 police officers, 15 prosecutors (specialised in bias crime) and 7 criminal law trial judges, which allowed us to obtain a view of the decision making process. In the group interviews, participants were asked to discuss whether various hypothetical cases should be marked as bias crime, allowing us to observe the analysis, criteria and arguments used by decision makers. We also sent questionnaires to all public prosecution offices in the Netherlands, asking about procedures and criteria to identify and prosecute potential bias crime. We received 34 responses out of the 100 questionnaires we sent. Our set of 318 cases breaks down as follows. Using the prosecution service’s database, 229 cases involving statutory bias crimes were identified. These were all the statutory bias crime cases that were prosecuted in the five jurisdictions covered by our study. As there was no similar registration system for bias-motivated crime for the period covered by our study, it was not possible to identify all of them. However, we were able to identify 89 cases in which the prosecutor had used the computerized case decision system and marked the case as bias-motivated, applying the 25% increase in sentencing demand. There are some caveats to this method of tracing bias-motivated crime. First, while the use of the computerised decision making system is strongly recommended to prosecutors, it is not mandatory and the degree of use differs between jurisdictions. Therefore, our selection does not cover all cases prosecuted as bias-motivated crime, but there are no reasons to assume that the computerised decision making system is applied in certain types of cases only. Second, as the computerized decision making system applies to relatively minor crimes only (public violence, threats, assault), our selection does not contain any serious offences. However, a review of secondary sources, such as newspaper reports and data collected by the prosecution authorities, suggest that serious crimes that are identified as having been committed with a bias motive are extremely rare. It may thus be assumed that these relatively minor offences constitute the largest and most significant part of bias-motivated crime that is actually prosecuted. Therefore, we submit that our selection of cases is sufficiently representative to study the decision making process for prosecuting bias-motivated crimes. We acknowledge, of course, that there is a potentially large dark number due to underreporting for both bias speech and bias-motivated crime [:115]. Our findings relate only to the cases that enter the criminal justice system and what happen we do not draw any conclusions about the social problem of bias crimes as such.In our collection of 89 cases that were prosecuted as bias-motivated crime, we found only one instance in which we considered that there was clear and convincing evidence of a malignant or ideological bias motive as the primary cause for the offence. We classified most cases as conflicts, because even if some signs of bias might have been present, they were unlikely to be the primary re the conflict, unrelated to any bias, was the primary cause of the offence. There were also a disconcertingly high number of cases where the crime appeared to have been committed without any particular motive. Table
reports the results.Table 1 Frequency of categories in sample of bias-motivated crime
Conflict with Authorities
No apparent motive
Malignant/ideology
Most statutory bias crimes take the form of bias speech: insults directed at a target group. Other types of statutory bias crime (such as incitement to hate or funding discriminatory activities) were so rare and diverse that we do not include them in our discussion. Table
provides an overview of the categories we found in our sample. Again, we see a predominance of conflicts, making up 55% of the sample. The occurrence of offences that were clearly motivated by bias is considerably higher than in the sample of bias-motivated crime, but at 14% it remains relatively low.Table 2 Frequency of categories in sample of bias speech
Conflict with Authorities
No apparent motive
Malignant/ideology
While we have now seen how Dutch bias crime cases are distributed over our categories, it remains to be seen what this says about net-widening. Whether and to what extent our figures indicate that net-widening occurs depends on a normative question: when do we conclude that the bias is sufficiently strong and sufficiently causally related to the offence to warrant enhanced punishment? As said, we answer this question from a normative rather than a sociological or criminological perspective. Treating bias crimes differently from ‘ordinary’, non-bias crimes (by punishing them more severely) requires a justification. The European bias crime policy refers to the presence of a prejudice or bias motive and the harm that such motives cause to society if they are acted upon. This justification only holds if the motive was actually a cause of the crime. Mere sentiments held by the offender cannot distinguish between bias and non-bias offences. On this view, penalty enhancements are only justified if it is clear that an offender acted because of his negative attitudes towards a particular target group. Both motive and causality need to be well-established. This may be termed the ‘but for’ test: were it not for the motive, the crime would not have occurred.The issue of net-widening is more complicated in regard to statutory bias crimes. As we have seen, these crimes do not require that the presence and causality of a bias motive be proven. The harm caused by the behaviour itself, as defined in the law, is thought to justify special treatment. However, we submit that motive also plays a role in this justification. A harmful bias motive is simply assumed to be implied by the behaviour. Bias speech, for instance, can be considered to deserve more severe punishment because the offender can be assumed to have made his remarks because of a bias motive. What makes such speech ‘biased’ is not simply that an entire group is insulted (‘stupid Turks’), but that such an expression implies that the offender attributes biased negative qualities to a particular group of people because of their ‘otherness’. The offender’s attitudes towards the group can here be considered to be the relevant bias motive.Offences that were clearly motivated by bias (i.e. malignant or even ideological motivations) were rare for both bias-motivated crime and bias speech. Given the centrality of bias motives in the concept of bias-motivated crime, it is remarkable that we found only one such case (Table ). They were somewhat more frequent in the set of bias speech crimes, but still made up no more than 14% of cases (Table ). Are these cases truly rare, are they unlikely to enter the criminal justice system as bias crime, or may there be other reasons for the low score? [cf. :897]. As mentioned, we applied a strict ‘but for’ test to classify a crime as ideologically motivated. However, this by no means suggests that we assumed a clear, causal motive only if the offender was an extremist motivated by clear-cut hate or otherwise held true ideological beliefs about particular target groups – in which case the causal motive is beyond dispute. Though we did find a number of such clearly ideological offenders, we also found causal, malignant motives in more ordinary members of the public. The difficulty, however, is to prove such motives if the offender is not an ideologist.The single case of bias-motivated crime in which we thought that the test was met concerned a man who, over a long period of time, harassed and threatened his homosexual neighbour. When asked at the trial why he did so, the man plainly answered that he hated his neighbour because he was gay. Though we might have marked this case as a (structural) conflict between neighbours, the statements made by the offender clearly showed that he initiated the conflict and harassed the victim because of the latter’s homosexuality – even if he was not an active extremist. As one prosecutor said: this is the type of evidence of a bias motivation that you would like to see but hardly ever get [:177, cp. , ].Bias speech crimes are more frequently committed out of bias motives. One reason is that evidence of causal bias motive is usually more readily available. Bias-motivated offenders of bias speech crimes use text, pictures, graffiti, or verbal means to express their often racist or xenophobic sentiments. The statements, in particular when these are put in writing, are usually quite explicit. Some offenders are also known members of extremist organisations or have a collection of extreme racist or xenophobic pamphlets or paraphernalia. Some run their own websites through which they express their ideas. Many offenders do not generally try to hide their convictions, and in some cases they try to justify their acts by relying on the freedom of speech. These circumstances usually gave us a clear indication that a malignant bias or even an ideological motive was the primary cause for the speech offence. Another possible reason for the higher prevalence is that bias motivated offenders might be more likely to commit bias-motivated offences in combination with bias speech. They may, for instance, use racist slurs while committing a bias-motivated assault. These cases ended up in our set of bias speech crimes. There were also cases where a bias motive was clear even if the offender was not a clear extremist, for instance if the offender explained his motives or if, from the expressions involved, it was quite clear that no other motive was likely to have been involved.While we found convincing evidence of causal bias motives in all these cases, two points of observation cast doubt on whether these motives actually played a sufficiently important role to justify enhanced punishment. Studying the 31 cases of ideologically motivated bias speech, we found, firstly, that the offenders seem more frequently to operate in groups: 42% of ideologically motivated offences were group offences, compared to 23% of all other contextual types (N = 227) [:293, table AL]. Similar findings are reported in literature on other jurisdictions [:87, , ]. Extremist groups or organisations that organise collective activities such as the public manifestation of their ideas may explain the prevalence of group crimes. In our study, however, the majority of the group offences were of a spontaneous rather than organised nature. That might suggest that rather than being truly caused by bias motives, most of these offences are the result of peer dynamics, the desire for group status, or thrill seeking [:30, ]. Nevertheless, other studies relate the presence of a group to de-individuation, which leads to a diffusion of responsibility and decreasing social restraints, making it easier for group members to act upon their ideological convictions and go on a ‘mission’ [:88, :201, :312].Secondly, we found that ideologically motivated offenders of bias speech are significantly younger than those in other categories: 29% of offenders are between 12 and 17, and 42% between 18 and 25. The rates for all other categories are 10% and 32% respectively (N = 229; p & 0.05) [:294, table AM]. This again points to thrill seeking, or at least away from serious, malignant bias motives. Similar findings have emerged in foreign studies [:96–97, ]. A German study has shown that many offenders of seemingly bias-motivated crimes were young, relatively uneducated and hardly aware of the historical or political significance of their ideas. The crimes they committed seemed to be the result of group pressure coupled with the use alcohol [:202–203, ]. Other studies suggest a prevalence of mental illness among seemingly ideologically inspired offenders []. However, in our survey it emerged that ideological offenders of bias speech are significantly less often under the influence of alcohol: only 17% of malignant bias offenders was under influence while committing the bias speech, compared to 45% of the offenders in the other categories (N = 198; p & 0.05).But even if not all cases in this category were in fact clearly motivated by bias, the set of crimes that we classified as having been committed without apparent motive may contain a number of cases that actually did involve such motives. At times, a police investigation yields no information about the ideological background of the suspect, and if the defendant makes no relevant statements, one is left unaware of any possible motive. Notwithstanding that, it is clear that in these cases there was no extraneous cause for the crime, such as a conflict between victim and offender. A striking example is the case (prosecuted as bias-motivated crime) in which a white man, who was queuing behind a black man who was using the ATM, said that the latter was too slow and that he ‘should go back to his home country’. He thereupon hit the black man in his face. At trial the suspect said that he thought that he knew the victim and that he was making a joke, using his ‘morbid sense of humour’. One may feel that this case may very well involve some sort of ideological motive, but there is not enough objective evidence to prove this beyond reasonable doubt, as is required under criminal law.Net-widening is a more serious issue in cases that we marked as conflicts, and these made up the absolute majority of both bias-motivated crimes and bias speech crimes. In our view, these case files did not yield sufficient and persuasive evidence that a bias motive was a primary cause in the chain of events, even if bias sentiments may have had a more indirect bearing on the crime. In these cases, offenders often explain to the police or in court that while they were aware that they have committed an offence, they were not aware of the fact that they did so out of bias. In our view, such claims are often credible: there is no suggestion in the case information that the offender held any bias beliefs that led him to evoke a conflict. This is not to say that the possibility of a causal bias motive is entirely absent. There may be more than meets the eye, but in all cases marked as conflicts, we found that the evidence did not meet the ‘but for’ test.Some typical examples may serve as illustrations. Conflicts with authorities, which are the second largest category in our samples, usually occur when a police officer is making an arrest or issuing a fine. The addressee feels frustrated and resists the authority exercised by the officer, for instance by being violent or insulting. If the police officer is of a different race or sex, the addressee – now offender – may refer to that which gives the offence a racist, xenophobic or sexist appearance. Conflicts between civilians may be similarly incidental. Take, for instance, the case of a Moroccan family who blocked the entrance of a parking space with their car, much to the annoyance of a white couple who wanted to enter it. A verbal argument back and forth developed and at some point the white male tells the Moroccan family to return to their home country. Later, some physical violence evolves, and the Moroccan woman gets injured. The white male is prosecuted for assault with a bias motive. In these instances, we felt that bias sentiments may have come to the fore as a result of conflict and frustration, but that they were not the primary motive behind the conflict and the offence.Some conflicts are of a more enduring nature. Take, for instance, a case involving two neighbouring Dutch and Moroccan families who had fallen out with one another over a longer period of time due to mutual nuisance, which eventually led to bias insults an complaints were filed with the police by both sides. In other cases of structural conflict, the victims of the bias crime had no apparent role of their own, for instance in the case of a foreign family that, after moving into a predominantly white neighbourhood, was continuously harassed by a group of neighbours, and eventually felt forced to move to another city.Conflicts do not exclude the possibility of a bias motive. There is evidence to suggest that structural conflicts between groups might cause ideological bias to develop and eventually result in bias-motivated violence [, , ]. A notable example is the perception of ‘stranger danger’ when immigrant families move into a neighborhood and old residents start to ‘defend their turf’ [, ]. Whether or not that should in theory be considered as bias motivated remains open for discussion [:370–372]. In any event, the structural conflicts we saw lacked clear evidence of bias being a primary motive (failing to meet the ‘but for’ test); they rather seemed to be the cumulative result of long-standing (mutual) nuisance, irritation, and conflict that appeared unrelated to bias.Net-widening occurs if bias crime penalty enhancements are applied to cases that lack the causal bias motive that justifies penalty enhancement. We scored only one malignant or ideological bias offender of bias-motivated crime (Table ) and 31 of bias speech (Table ). The absolute majority of cases are conflicts. The problem here is one of causality. Conflict crimes are motivated primarily by frustration or anger that is caused by the conflict, but often not by bias sentiments. It is not impossible that a bias motive played a role, but in our view the evidence did not meet the ‘but for’ test that we applied. Had the victim not been ‘different’, the violence or insult would most probably nonetheless have occurred. Biased remarks are made, but these are insults that refer to race, sex, or sexual preference, rather than expressions of a bias motive. Such insults are easily and readily made on account of the victim’s appearance. The offender will probably know that a bias insult may hurt the victim more, and that is precisely his reason for using it. In cases of conflicts prosecuted as bias speech (Table ), references to the victim’s appearance (race or skin colour) were significantly more frequent than in the other categories, as is demonstrated in Table . The motive in these cases is to hurt the victim by referring to his membership of a target group, not the dislike of or disdain towards that target group as such. These cases are serious and harmful, but do not involve bias and intolerance towards target groups as a motive for action.Table 3 Distribution of bias references over case categories for bias speech
Bias reference in speech*
Conflicts with authorities
Malignant/ideology
Anti-Semitism
Other religion
Turkish or Arab background
Skin colour
Homosexuality
These findings suggest that many bias crime offenders normally control some bias sentiments, but express them in situations of anger or stress. Many others are not even aware of the fact that their behaviour could constitute a bias crime. There is frequent use of insults that refer to bias or prejudice, but that does not always imply that the offender was motivated by his dislike of a particular target group. These cases fail to meet the justification of enhanced penalty that refers to the harmfulness of bias motives as a cause of action. Offences that were clearly motivated by bias do occur, but they are the minority of the cases that are prosecuted as bias crime. We therefore conclude that most cases prosecuted under Dutch law as bias crimes are a symptom of net-widening. How can this be?When prosecuting bias crime, two fundamental legal problems seem to arise. Firstly, it is difficult to prove that an offender held a particular bias motive [, ]. Secondly, even if the presence of such sentiments can be demonstrated, it is even more difficult to prove that it in fact caused the offender to commit the offence. It is particularly difficult to assess the role of a bias motive if it appears merely to be among the various reasons for the offender’s behaviour [, , ]. The significance of these problems is best explained by an example. In one of the cases in our study, a white male hailed a taxi while the Turkish cab driver was just on his way to another call and could not oblige. This angered the white man, and he called the driver a ‘stupid Turk’, and hit him in the face through the open cab window, knocking one of his teeth out. During trial, the offender said that he did not know why he behaved as he did, suggesting that he ‘needed to get rid of some anger’ [:154]. This case is clearly one of assault and insult. But are the offences motivated by bias? The insult used does certainly evidence some bias, but it does not necessarily indicate that bias was the motive for the offender’s action. He might have been angered by the fact that he could not get a ride, and had the cab driver been a white male, he might have reacted in a similar fashion except for the bias slurs. Notwithstanding the impact on the victim, who might feel that the offender acted out of bias, the case is not a clear-cut bias-motivated crime.Requiring proof of a bias motive and its causality sits uneasily with the action-oriented nature of criminal law. Under criminal law doctrine, the ground for punishment is the harm that is caused by particular behaviour, not by one’s thoughts or reasons. Criminal law punishes actions rather than thoughts or beliefs. Under Dutch law, motives are not an element of any definition of crimes, nor are they thought to be relevant for criminal liability: it is only the objectively observable behaviour that counts [:358–359, :247, , :70]. Punishing motive is therefore a novelty, and a problematic one at that. Motive and causality refer to the subjective state of mind of the offender. Unlike intent, motive is not a normative, legal concept. To establish intent (i.e. to establish whether the offender had wilfully wanted the consequences of his behaviour that constitute the crime), one need not look into the mind of the offender. Legal doctrine has developed objective criteria under which an offender can be held to have acted intentionally. We might say that certain behaviour, such as drunk driving, is so dangerous and socially unacceptable that a driver should be held to have acted intentionally if any accident occurs, so that he might be punished for intentional manslaughter or injury. But this normative approach cannot be applied to motive. Motives are subjective reasons for action. Punishing motive assumes that these can make the crime a greater wrong, even if the behaviour shows no signs of the motive.The reluctance in criminal law doctrine to take the offender’s motive into account must be understood against the backdrop of liberalism and the limits of criminal liability [, , , , ]. Opponents of bias crime legislation argue that motive in general cannot (and usually does not) constitute a ground for criminal liability or enhanced punishment. Penalisation is legitimised only if a particular behaviour – an action – causes harm. But a motive is not an action. It is rather an emotional state of mind, which is not deliberate or intentionally chosen by the offender [:1066, :226]. To sanction the motive would therefore be to punish bad character instead of harmful action, and that would be at odds with the principle of guilt: one can be guilty of intentional behaviour, but can one be guilty of merely having certain thoughts or beliefs? Additionally, one can argue that it is unclear why a bias motive would or could be more harmful than others, as crimes are usually committed out of morally abhorrent motives anyway [, ]. These principled limits to criminal liability are deeply ingrained in legal reasoning and discourse, and may therefore be expected to have significant influence on how decision makers deal with the problem of establishing motive and causality.In the discussions and legislation on bias crime, two approaches to these problems can be distinguished [:297–299, :326]. On the one hand, the Discriminatory Selection Model defines bias crime as an offence for which the victim was not selected randomly or on the basis of individual characteristics, but because of his membership of a group. Direct evidence of the mo the selection of victims on account of their membership of a target group is considered to constitute an additional harm and indirect evidence of the ‘motive’. Here, the criminally relevant harm is thought to be in the behaviour itself: selecting victims because of their membership of a target group is thought to be more harmful, either to the victims themselves (greater emotional damage) or to society (public display of intolerance). Though this approach does away with the need to establish motive, the problem of causality remains. For it will still need to be proven that the victim was selected because of his membership of a target group rather than for other reasons. That is where the criminally relevant harm lies.On the other hand, the Racial Animus Model does require that it be established that the offence was motivated by the offender’s dislike of a target group. Evidence of that motivation can be drawn directly from statements and declarations made by the offender, or from circumstantial evidence such as his membership of an extremist organisation [:94] or past interest in extremist material (such as websites, books or pamphlets) [:540–541]. When it comes to circumstantial evidence, the challenge for the prosecutor is not merely to argue convincingly that the presence of a motive is proven, but also that it was the primary cause of the behaviour at issue. This has to be proven, because the criminally relevant harm according to this model lies in the nature of the offender’s motive, in his ‘racial animus’.The importance of the distinction between the Racial Animus and Discriminatory Selection models can be demonstrated by an example. The Netherlands has seen a number of instances of robberies aimed at gay men, in particular at public meeting places. Such offences are most likely not inspired by hatred of homosexual men, but rather by the belief that they carry cash and would be too embarrassed to come forward to the police and report the crime [, ; for Germany: ; for the UK: ]. Under the Discriminatory Selection Model, crimes as these would be considered prima facie bias crimes, as the victim has been selected because of his membership of a target group. The Racial Animus Model does not exclude the possibility of classifying these as bias crimes, but requires that the presence of a motive be established i.e., that the offender acted out of hate or bias towards homosexual men, rather than out of opportunistic reasons.It is important to note that the two models are not mere methods of establishing motive and causality. They also imply distinct lines of reasoning on why sentence enhancement in cases of bias crime is justified – and, in fact, on what constitutes a bias crime. The Racial Animus Model stays close to the European policy by assuming that the criminally relevant harm lies in the presence of a bias motive. The behaviour that constitutes the offence is objectively no different from that of a nonbias parallel, but the fact that the offender acted on bias motives justifies enhanced punishment. However, the Discriminatory Selection Model, by circumventing the need to prove a bias motive and its causality, allocates the criminally relevant harm in the selection of the victim as member of a target group. The harm here lies with damage done to victim and society because of the offender's public display of intolerance, evidenced by his selection of a victim because of the latter’s membership of a disliked target group.Under Dutch law, an offence can be marked as a bias-motivated crime if the offence was motivated by bias – thus employing the Racial Animus Model. If so, the prosecutor should increase the sentencing demand by 50%. How, then, do decision makers conclude that a bias motive played a causal role in the offence? When do they assume sufficient bias and causality to justify penalty enhancement? In our interviews and questionnaire responses, police officers, judges and prosecutors indicated that the most important indicators of bias are the expressions made by the offender during or in relation to the offence. These are usually verbal, but can also be symbols (swastika’s, Nazi salute) or writings. If such expressions are absent, decision makers rely on even more indirect indicators such as clothing worn by the offender, known extremist affiliations, or the fact that offender and victim belonged to different ethnic groups that are generally in conflict with one another.It is more difficult to establish when these indicators are considered to be conclusive evidence of both motive and causality. The threshold differs between individual decision makers. They do generally acknowledge a need to distinguish between ‘stupidity’ and ‘truly ideological motives’, but the borders are far from evident. While most decision makers agreed that indicators such as bias slurs could be used to identify a motive, they had more difficulties to agree on standards on how to establish its causality. Regarding the case of the Turkish taxi driver described above, the following discussion between three prosecutors ensued [:155].(1) ‘If an offender gives that explanation [i.e. uncontrolled anger], then he may attack a Turkish man one time. Another time there’ll be an old lady in his way and he’ll pull her away, too, insulting her as an ‘old crow’. That is the way this offender expresses himself.
(2) ‘But you cannot be sure. He can excuse himself from bias by saying: I would have done it anyway, whether the victim was Turkish or Dutch, I was just feeling angry and stressed. But you have to consider the facts and the circumstances. You have to objectify. Apparently, the looks of the victim were his reason to use bias slurs.’
(3) ‘But wasn’t his anger and stress the reason to use those expressions?’
(1) ‘Yes, but the ethnic looks of the victim were his reason to use those particular [bias] expressions’.
Another case shows similar disagreement between police officers (Po), prosecutors (Pr) and judges (J) on how and when to assume a causal bias motive that justifies penalty enhancement [:155–156]. While leaving a nightclub, a group of Dutch young men came into conflict with a group of Moroccan youths, accusing them of flirting with their girlsfriends. A massive fight breaks out (it is unclear who started it), and when the police arrive one of the Dutch men shouts a bias slur relating to the Moroccan origin of the others. He is prosecuted for bias-motivated assault. Some decision makers thought that this was enough evidence of bias motivation: (Po) ‘Just shouting something like that is enough. That always hurts. He could have just said ‘bastards’ or something rather, but he specifically decided to go with the bias slur.’
(Pr) ‘Having regard to the context of the event there may not be a bias motive, but there is certainly a bias aspect to it. One might say that the original motive was the girl, and that a new motive was added when the offender realised the others were Moroccans. To me it seems like a new, bias motive.
Most decision makers, however, did not think that the concept of bias-motivated crimes ought to apply here. (Po) ‘Those groups go to a nightclub, drink too much and then get into a fight about a girl. Is that really bias-motivated crime? On a next occasion they’ll probably be together in the nightclub again. Without any problems.’
(Pr) ‘What if these were city folk against provincial farmers? What if the city youth had shouted “stupid farmers, keep your hands off our girls!”? On this analogy, would you then mark it as bias-motivated?’
(J) ‘Fifty years ago you had people from one village fighting those from the next village, or from one school against the other. It seems to be part of human nature to act on such differences.’
These discussions show that decision makers look for objective indicators of a bias motive. As they are trained under criminal law doctrine and its action-oriented nature, one should expect that they fail to deal with motive as a legal concept. This is in fact the underlying cause of many of the obstacles that exist in the prosecution and judgment of bias-motivated crime. Because a bias motive cannot be grasped in objective, behavioural terms, decision makers cannot deal with it. Offering proof is even more difficult. On the one hand, the concept of bias-motivated crime requires that the presence of a motive be established, and that it is ascertained that the motive was a primary cause of the crime. On the other hand, criminal law doctrine is action-oriented, and therefore excludes the possibility of criminal liability for thoughts or character. This contradiction can hardly be resolved in practice, even if action makers were better trained.In practice, the contradiction is resolved by referring to objective, provable circumstances as indirect indicators of the presence of a bias motive. That is to say, based on objective behaviour, such as words that were used during the offence, the offender’s membership of extremist organisations [:94, ], or his known affiliations or ideological convictions, the existence of both motive and causality are inferred. For similar reasons, the FBI has introduced a set of bias indicators, which are currently used for the purposes of registering bias crimes []. However, our case studies show that this approach leads to a move away from the Racial Animus model. Objective indicators can never help to overcome the difficulty of motive: how to prove that the offender had a certain subjective state of mind that was the primary cause of his committing the offence? Even if the decision makers that we interviewed generally adopted a high standard for inference and focused on ‘ideological motives’, our case studies – and in particular the high number of conflicts – show that most police officers and prosecutors are comfortable to infer a bias motive on the basis of bias slurs alone, without further discussion of causality. Accordingly, one could be convicted of a bias-motivated crime even if a bias motive is not, with certainty, proven to be a primary cause of action. This leads to a widening concept of bias-motivated crime. It includes not only crime motivated by bias, but also, and predominantly, conflicts in which bias slurs were used without these clearly implying a causal bias motive on the part of the offender.This explanation would be harmonious with another phenomenon. In 30% of the cases in our sample where the imposed penalty could be compared with the prosecutor’s sentence recommendation (N = 81), the judge imposed a penalty that was lower than the one requested by the prosecutor [:173]. That suggests that, in the eyes of the judge, prosecutors often failed to prove the bias motive that justifies an increased penalty in one third of the cases that have been prosecuted as bias crime [cp. 8]. Judges may be less willing to rely on objective indicators to infer that a crime was motivated by bias, and require straightforward evidence instead. In our interviews, judges explain that in the determination of a sentence they pay little attention to the (alleged) bias motive and apply the ‘ordinary’ criteria, such as the seriousness of the violence and probability of recidivism [, , :183]. Judges may consider the possibility of a bias motive in coming to a sentencing decision, but they are reluctant to explicitly mention this or to mark a case as bias motivated. They feel that objective inferences from expressions or symbols are not usually sufficient to warrant conclusions about the presence or role of a motive.Under Dutch law, the offender’s motive is not relevant for the proof of bias speech. It suffices to be proved that the offender intentionally insulted a target group as a whole. The drafting of these laws is therefore close to the Discriminatory Selection Model. From a dogmatic point of view, the harm of statutory bias crimes, such as bias speech, lies in the behaviour as such. The offender’s motive is not technically relevant to prove a statutory bias crime. But that is not to say that the offender’s motives are also irrelevant from a normative point of view. As we saw, normative reasoning does presuppose that bias speech always has some sort of bias taint. For statutory bias crimes, the objective inference of the offender’s motive is laid down in the law. The Discriminatory Selection Model is employed as a technique to define statutory bias crimes by assuming that bias speech, wherein a target group as a whole is insulted, will always be motivated by bias or prejudice towards the target group.However, our study shows that this statutory objective inference fails to differentiate between bias-motivated behaviour and conflicts. In a substantial number of conflict cases, racist or xenophobic slurs are used that refer to a target group as a whole, instead of merely to the individuals implicated in the dispute. But this does not necessarily reflect a motive of bias towards that group. In the majority of cases that were prosecuted as bias speech, other reasons than bias led to the crime. Our collection of bias speech crimes largely consists of slurs and insults that were used in cases of conflict, frustration and anger. Because of objective inference, the distinction between an ordinary insult and bias speech becomes blurred. The use of a plural can make the difference. Net-widening is a direct consequence of the objective inference that is built into the law.There is a paradox in that legal reasoning on the one hand requires that a causal bias motive be established to justify more severe punishment, while on the other hand it focuses on harmful action and thus precludes punishment of thoughts. This paradox forces decision makers in the practice of the criminal justice system to rely on objective inferences to assume the presence of a bias motive. They feel they have to find objective, behavioural signs of motive in order to treat bias crimes differently. However, these indicators are often unreliable. Most cases that are prosecuted as bias crimes are not motivated by bias, but are rather the result of conflicts. The presence of certain words or other signs, such as racist or xenophobic slurs, is relied upon as evidence of a bias motive. But such words or signs may be used out of rudeness, frustration or unawareness, and do not necessarily reflect bias motive. As a final example, take one typical case from our sample. A Moroccan teenager tried to flirt with a white girl on the street. When she ignored his courtship, he insulted her by calling her a ‘vuil Jodenkind’ [‘dirty child of a Jew’]. Such language may of course be insulting for Jewish people as a group or to the victim (who, in this case, was not Jewish). But it is not necessarily reflective of a bias motive. In this case, the insult was not even targeted at any Jewish person in particular. It rather seems to have been used as a sort of colloquial, rude insult. Most of the incidents we studied seemed to turn on undesirable yet increasingly ‘natural’ language, expressing incivility, rudeness, and carelessness [:20]. One can probably only discern with certainty those cases in which bias crimes were committed by ideologically motivated offenders or by those who are otherwise inclined to provide certainty about their motives. In many other cases, mostly conflicts, bias motives may have played a role, but the evidence we found was inconclusive as to their presence and causality. If the application of bias crime laws were limited to such clearly ideologically motivated offenders, disappointment among victims and society is likely, for it would mean that many perceived and real bias-motivated crimes will go unpunished. The significance of bias crime legislation would be reduced to symbolism, sending messages only about the most extreme types of bias crimes.As the law stands, net-widening is unavoidable. Objective inference is needed because of criminal law is action-oriented and requires objective proof (beyond reasonable doubt) of motive and causality. One ought therefore to question whether criminal law can be used as an instrument to combat crimes that are committed with a particular motive, as the European bias crime policy presupposes. The answer is negative. Bias crime legislation fails to differentiate between crime that is committed because of bias, and crime that is committed for other reasons.Whatever one should think of net-widening and its acceptability, we conclude that in the practice of criminal justice the concept of bias crime is transformed. Because of objective inference, the meaning of bias crime changes. It is no longer related exclusively to motive. Instead, it has become a concept that is applied to all sorts of behaviour that is interpreted by victims and the community in general as being unacceptable expressions of bias, regardless of the offender’s motives and reasons. Then, however, a new justification for stronger punishment is required. It turns to punishing words or slang – the most common objective indicators of bias – that are held to be undesirable because they are associated with bias. The criminally relevant harm is in the damage done to society by the use of such words, not in the offender’s motive, which is practically rendered irrelevant. We use bias crime laws primarily to send messages, both to society in general [:154–155] and to offenders, telling them how their acts are perceived []. We use criminal law to ‘clean up’ rude or bad language that we find unacceptable because of its association with bias. This may be perfectly acceptable. We regularly punish behaviour because of the social meaning we attribute to it. For bias speech, this is already more or less accepted: we punish it more severely because of the perceived social harm of expressions that display intolerance, rather than because of motive.To continue to speak of bias crime as a matter of crime motivated by bias would therefore be a matter of mere rhetoric. Legally, it obscures the justification that we rely on to punish this type of crime more strongly than others. If we acknowledge that bias crime is more about the harm caused by what are taken to be bias expressions than about the harm caused by bias motives, new normative questions arise. Does the behaviour that is now considered to constitute bias crimes really cause social harm, even absent bias motives? Does such social harm, if any, justify a stronger punishment? And is the bias perception given to such behaviour sufficiently foreseeable to respect the principle of guilt and legal certainty? Socially, the rhetoric of motive groups together all sorts of situations that may, on closer inspection, involve a wide variety of social problems and dynamics. Thinking beyond motive would allow us to consider these more clearly. We recognise that the use of bias slang may very well be considered to be a social problem involving social tension, fears of strangers, and much more implicit and less malignant racist or xenophobic sentiments [:263, :202–203]. But they are not always bias-motivated crimes. Therefore, we should think openly about means to address such problems, without reverting to the idea that we can punish motive. To do so crosses fundamental limits to criminal liability, with net-widening as a consequence.This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.1.Discriminatie, A. (2007). Issued by the Dutch collegium of Prosecutors-general, 29 October 2007. Staatscourant, 2007, 233.2.Bell, J. (2002). Policing hatred: Law enforcement, civil rights and hate crime. New York: New York University Press.3.Blee, K. M. (2007). The microdynamics of hate violence. American Behavorial Scientist, 51(2), 258–270.4.Bowling, B. (1998). Violent racism: Victimization, policing and social context. Oxford: Oxford University Press.5.Bowling, B., & Phillips, C. (2002). Racism, crime and justice. Harlow: Longman.6.Brants, Ch, Kool, R., & Ringnalda, A. (2007). Strafbare discriminatie. The Hague: Boom Juridische uitgevers.7.Buijs, L., Hekma, G., & Duyvendak, J. W. (2008). Als ze maar van me afblijven. Een onderzoek naar antihomoseksueel geweld in Amsterdam. Amsterdam: University of Amsterdam, Amsterdam School for Social Science Research.8.Burney, E. & Rose, G. (2002). Racist offences: how the law is working. The implementation of legislation on racially aggravated offences in the Crime and Disorder Act 1998. London..9.Chakraborti, N., & Garland, J. (2009). Hate crime: Impact, causes and responses. Los Angeles: Sage.10.Chakraborti, N. (2010). Crimes against the ‘other’: conceptual, operational and empirical challenges for hate studies. Journal of Hate Studies, 8(1), 9–28.11.Craig, K. M. (2002). Examining hate-motivated aggression: a review of the social psychological literature on hate crime as a distinct form of aggression. Agression and Violent Behavior, 7, 85–101.12.Cronin, S., McDevitt, J., Farrell, A., & Nolan, J. (2007). Bias crime: organizational responses to ambuiguity, uncertainty, and infrequency in eight police departments. American Behavorial Scientist, 51, 213–231.13.Dillof, A. M. (1997). Punishing bias: an examination of the }