Ethnic profiling—the use of racial, ethnic, or religious characteristics as a way of singling people out for police or security checks. In the United States, debate over its fairness and effectiveness has raged for years. In the United Kingdom, it's in the headlines.
In most of the rest Europe, it's a different story. Despite its widespread use by police forces and the festering resentment in affected communities, ethnic profiling is rarely examined; aside from the UK, no European government collects information on the ethnicity of those stopped by police.
But the Open Society Justice Initiative has documented the prevalence and impact of profiling in Europe, assembling evidence that shows it is not only unfair, but an ineffectual way of fighting crime.
We are also taking the fight to the courts. On the eve of filing a new constitutional challenge in France, we asked Rachel Neild, senior advisor with the Justice Initiative, about the program’s work.
How did the Open Society Justice Initiative get involved in working on ethnic profiling?
We’ve been working on ethnic profiling in Europe since 2005. We met a woman in Spain, Rosalind Williams, who had been stopped by a police officer. And when she asked him why, he said, “Well, it’s because you’re black.” Being a Spanish citizen but of African-American origin she challenged this and took it to court in Spain, where the Spanish Supreme Court actually ruled that it was just fine to use ethnic appearance to judge who might not be a national of that country. We went a step further and took Rosalind Williams’s case to the UN Human Rights Committee, which ruled against Spain and said that Spain was violating its international treaty obligations not to discriminate against people in criminal justice.
And that’s really the bottom line in this—that all people deserve to be equally treated by the law. That’s one of the most basic principles in play here. And to discriminate against someone because of things that are personal traits—your appearance, your race, your ethnicity, your religion—or because of stereotypes or associations, is morally unacceptable and also unlawful. It really goes against this principle. It’s also just wrong.
So the Rosalind Williams case provided a landmark judgment to work from?
It was the first judgment by an international court specifically on ethnic profiling. And it’s very important, because the ruling was very clear that the use of ethnic or racial appearance in stops—including immigration stops—is not lawful police practice. It affirmed the basic principle that people should only be targeted by police and other law enforcement officers on the basis of their personal behavior—what they’ve done, not who they are.
What was your next step?
At that point we realized just how little information there was about the practice in Europe. A lot of continental Europe has less tradition of studying policing and a lot less data about policing is made public than we are accustomed to in Britain and America. So, our first approach was to see what sort of data we could actually produce.
We’ve done this through a number of reports We did an initial report where we conducted semi-structured interviews with police officers in three countries—Bulgaria, Spain, and Hungary—which went through how they use stop-and-search and then asked them who they think commit what kinds of offenses. And they were all remarkably open, about their stereotypes and in some cases just listed: “Oh, Roma do this, and Bulgarians do that; Eastern Europeans are this, and Latin Americans are that.” And it was just really obvious.
What was also obvious in that report is that none of the police forces were seriously evaluating the effectiveness of those powers like stop-and-search. You know, who are they stopping? Where? Why? Those questions just really weren’t built into the system at all. So here is a power that is profoundly intrusive on people’s lives and yet the police have no idea at a management level as to what they’re really up to.
So, we got the cooperation of police in the three countries to collaborate with us and generate ethnic data on their stop practices, which was quite revolutionary in Europe. So we introduced stop forms in those countries—to create a record of who was being stopped and where and so on—and we actually produced ethnic data on stop-and-search, which showed very clear disproportionality in all the data we could gather.
That certain groups are, in fact, targeted more often than others.
Exactly. Not surprising, in itself, but now we were working from real data.
Now, we’ve also worked on documenting ethnic profiling in the context of counter-terrorism practices. We had a very strong sense in the mid-2000s that not only was ethnic profiling a longstanding problem for people because of their appearance, but that in the wake of terror attacks that Muslims in particular were being increasingly targeted. And we documented that in a 2009 report.
You also conducted a major observational study on profiling in Paris. Was that done in collaboration with the police?
No, we wanted to avoid an observer effect. We wanted police to conduct stops as they normally would. So, with an observational study like that, the first step is to look at the population in the areas where you’re documenting—we were largely in the Paris metro and close by outside it—and we took data for 32,000 individuals, looking at their age, their gender, their racial characteristics, broadly, their clothing, and whether they had large bags or not. The bags related to concerns about terrorism. Then the observers followed police officers in the same locations and documented 525 stops, using the same set of criteria. And the disproportionate number of stops of blacks and Arabs was very, very clear.
No one in France, I have to say, was very surprised by the results—that blacks and Arabs are systematically stopped at far greater rates than other French people. But the generation of the ethnic statistics to demonstrate that caused a tremendous amount of discussion. The study has been invaluable because it produced a data set that everybody has recognized as about as objective as you can be.
Isn’t law enforcement inclined to push back on your findings and say, okay, we see the pattern, but isn’t this reasonable? Doesn’t it make sense, in policing terms, to target populations where incidences of crime are higher?
Does it make sense? No. When you test stereotypes, particularly on stop-and-search but in fact with all the police powers we’ve looked at, it just doesn’t add up. Just about everywhere we have been able to get good data, we have found that ethnic minorities, when stopped and searched, are less likely to be found committing an offense than members of the indigenous population. And that holds true in the British statistics and the American statistics. In Britain, for example, it’s very clear that when you give more discretion to a police officer to stop and search—that is, powers that don’t require “reasonable suspicion”—that officers stop more minorities and have lower hit rates, meaning that their stops are productive less often. And that holds true for all the settings for which we’ve been able to generate solid data.
There’s cumulative evidence here, too. In the United States, for example, data on marijuana shows that use is pretty even across racial and ethnic groups—if anything it’s higher among white kids. But if you look at marijuana arrests in New York City, they’re overwhelmingly of black kids. It’s a form of social control policing. And it’s more acceptable to police certain groups in certain ways—which are fundamentally uneven and unjust—precisely because of these stereotypes. Of course, the stereotype is used to justify the policing. But really it’s a self-reinforcing cycle.
As part of your work, do you take data to police departments and say, look, what you’re doing actually isn’t effective?
We do. We show officers the data that we have.
What kind of reaction do you get?
They say, “This is fascinating, but this isn’t us.” They want to see really specific data, the sort of stuff we produced in the Paris research to see what they’re actually doing. Which is a little frustrating, because, frankly, if they’re interested in this, they ought to produce the data themselves. And ultimately they need to be looking at their own powers and how they use them and be concerned that they use them well.
Can you gauge the impact when powers like stop-and-search aren’t used with care?
The evidence is extremely powerful that policing that is unfair—and is experienced as unfair—is not only inefficient for the police but also generates widespread hostility and a loss of credibility for law enforcement. It causes people to be more reluctant to cooperate with police officers. And the reality is that policing is not like the TV shows we see with forensics all over the place. Almost all cases are solved because of witness testimony and people reporting crimes. So, if you alienate people to the point that you reduce cooperation, you’re really undermining your own efficiency in very profound ways. But these are things that many police departments fail to adequately recognize and measure.
I think there’s a sense that this is just an inconvenience for people. But if you’re black or brown, and every time you go to the metro you’re being stopped and often treated quite rudely, and this happens again and again—it can be demeaning. It can be embarrassing. It can be humiliating, even traumatizing. It’s being done to little kids, as well as adults. And I think we’re seeing the results in the kind of hostility that’s demonstrably building up in certain places.
It seems that’s what’s really at stake here—that breakdown of the relationship between communities and police, that cycle of mistrust.
It’s a big issue and certainly something we argue a great deal with police officers. But I think as human rights advocates there are often qualms about arguing you shouldn’t do this purely for reasons of your own effectiveness. Because, frankly, you shouldn’t do this because it’s wrong.
It is wrong to treat someone in a certain way because they belong to a group you have a stereotype about, when you know nothing about them individually. Just because they live in a neighborhood with relatively high rates of crime does not mean that they’re offending. And to treat people that way is wrong and unlawful. That said, this is also going to generate problems for you institutionally as a police force, so we do make both those arguments.
Where do you take the work from here?
We will continue to try to persuade police departments to adopt good practice. There are demonstrated good practices out there that can help police do their work in a way that is both fairer and more efficient. We’ve put together a handbook of good practice that pulls together examples of these and offers a conceptual framework for good practice. And we will continue to look for ways to collaborate with police departments on piloting these best practices.
There are certain countries where we have just hit a wall in terms of trying to engage with governmental authorities and with police. Unfortunately, European attitudes right now, driven by economic recession and fear of terrorism, are increasingly xenophobic. That’s driving more discriminatory policing, more ethnic profiling.
So, in some of those settings, we’re actually looking at bringing cases against the government and trying to litigate to get clear legal standards at the national level that would require some change in practice. We’re undertaking this in France, where we’re going to file a constitutional challenge to the legality of current police powers to stop and search people under the French criminal code.