"I couldn’t be any clearer about [the police's] mission: it isn’t a thirty-point plan; it is to cut crime. No more, and no less." Theresa May, in her 2010 speech as Home Secretary to the National Policing Conference.
This post is about police accountability. Before I begin, I need to clarify what kind of accountability I am focusing on. When the police commit acts of violence and discrimination, there is often a public outcry for police accountability. The desired outcome is usually for the police officers involved to receive their just deserts and for measures to be implemented to ensure that such actions do not occur again. This is an important aspect of accountability, and it will be addressed later in this series. The primary focus of this post, however, is on how government attempt to exercise administrative accountability, thereby ensuring, to borrow a cringe-inducing phrase from UK government, 'value for money'.
I expect my conversation here to be relevant to most democratic countries, but I should note that I have developed my knowledge of policing governance in an England context, though I have also supplemented this background with research into police governance in Canada. In the UK the police are, in theory, directly accountable to elected politicians, titled Police and Crime Commissioners. In Canada police governance is more varied because authority over police governance is devolved to the provinces.
"If you can measure it, you can manage it," goes an old cliche, and the ethos is readily apparent in the above quote from Theresa May. "Cutting crime" and "being tough on crime" are common parlance in politics, and I think that a significant contributing factor is the fact that crimes are counted. Its easy to ‘manage’ the police via crime statistics because you simply point to the numbers and say "I want these numbers up, I want these numbers down." An excellent example of accountability by numbers is the 20:20:20 targets implemented by former Mayor of London Boris Johnson. The goal was to cut key crimes by 20%, boost public confidence by 20%, and cut costs by 20%. These numbers were obviously arrived at through painstaking deliberation.
A "cutting crime" narrative also compliments our cultural understanding of police work. Western media feeds us a steady diet of cop-shows and sensationalized crime stories, to such an extent that "officers may model themselves on film and television characters" (Tim Newburn, 2012). The primary police activity with which we are familiar is the apprehension of criminals. Crime statistics therefore encompass both political and public understanding of what good policing looks like – if the police are doing their jobs then there should be less crime, the logic goes.
The media further fuels the emphasis placed on police management via crime statistics. While working at the Greater London Authority as a policing policy officer in the UK, I observed how every policing statistic became a political battleground. Both politicians and the media were continuously digging up alarming crime statistics and waiving them before the public like matadors. Knife crime is up, arrest rates are down, complaints against the police are up, police numbers are down, etc, etc. Both politicians and the media are strongly incentivized to use crime figures in this way. For the media, alarming crime statistics make particularly good headlines. For politicians, crime statistics are an easy way of showing their success or their opponents' failures in a way that attracts press. Correspondingly, for politicians in power it is very important that crime figures improve to avoid negative press.
There is a significant problem, however, with police management via crime statistics. There are numerous problems, actually, but there is one in particular which I consider the most pertinent. Crime statistics tell you almost nothing about crime rates. I propose the following four reasons that this is the case:
1. The police can only record crimes reported crime or those crimes they observe directly.
2. The polices' strategic decisions can strongly impact crime figures.
3. The police are in charge of reporting crime figures while simultaneously are held to account by those same figures.
4. Crime is socially constructed.
The first, and probably the least controversial, of my arguments is that the police can only track those crimes which they observe or that are reported to them. It is therefore the case that crime statistics can be improved by decreasing the number of police officers, thereby decreasing the number of crimes observed. Further, the majority of crimes are not reported to the police. Observe the following modified diagram from the UK Office for National Statistics:
Crime in England and Wales: year ending June 2017. Office for national statistics (2017).
The bars along the bottom represent police-recorded crime in the UK. The line represents crime estimates based on the Crime Survey for England and Wales (CSEW). The CSEW is a regularly conducted survey of approximately 50,000 households in the UK which asks for participants to report how many crimes they have experienced over the past year. Though the two data-sets do not fully overlap (the sample size of the survey is insufficient to accurately estimate low volume crime (e.g. violent assault), and, for obvious reasons, a survey cannot capture individuals experience of murder), it is evident from the diagram that police crime stats have historically captured significantly fewer crimes than the crime survey. Now, the obvious next step, at least in the case of the UK, would be to point to the crime survey itself as a reliable source of police data. However, it too has significant limitations. First, due to the sample size it only reliably captures high volume crime, and only for the entirety of England and Wales – it does not, for example, tell you whether crime in London is up or down. Second, it does not capture low-volume crime. The third issue pertains to the social construction of crime, which I will get to later.
So, not all crime is counted. However, if crime figures consistently tracked crime proportionately (i.e. if it were true that crime figures rose when crime rose), than crime figures would still be valuable. There are at least two reasons to think that this is not the case, those being points 2 and 3 on the list above.
The police make strategic decisions around the deployment of resources. The police do not have the capacity to enforce all laws everywhere at all times. As noted above, one of the easiest ways to reduce crime is to get rid of police officers. In the same vein, specific types of crime will go up and down as police priorities change. For example, in the UK there has been a significant shift in emphasis from high-volume crime (e.g. theft, possession) to high harm crimes (e.g. human trafficking, protecting vulnerable people). Redistribution of resources inevitably impacts crime detection. The number of recorded incidents of crimes such as domestic abuse and human trafficking in the UK are increasing, not because they are suddenly worse, but because the police are putting more resources into them. This observation must be combined with the reality of finite police resources: increased diligence in one area must come at the expense of other areas.
The third factor that undermines crime statistics is that they are used as a management tool. If the police are held to account via crime statistics, and are simultaneously the people responsible for reporting those very same crime statistics, then there is a strong incentive for the police, as an institution, to manipulate the figures that they are reporting. This may sound like I am accusing the police of widespread corruption, so let me be clear that this is not my intent. The police manipulate crime figures because they are street-level bureaucrats who are expected to deliver on impossible expectations. Crime stat manipulation does not require a police conspiracy, though deliberate and coordinated manipulation of crime figures has also been documented in the past (link). All it requires is for managers to place pressure on police constables to meet targets. Constables, with little control over crime, then adjust their recording practices. Consider the Metropolitan Police’s remarkable delivery of Boris Johnson’s strenuously researched 20% crime reduction target on a 20% decreased budget a year ahead of schedule (link). There is an extensive literature on police manipulation of crime figures, as well as numerous news articles. I shall provide links to examples at the end of the article.
Finally, the fourth factor that undermines crime statistics is the fact that crime is socially constructed. All that I mean by this is that ‘criminal’ is not an inherent property of an event. An event only becomes ‘criminal’ when so labelled by the state and by society. Political and societal perspectives on events change over time. For example, same sex marriage was legalized in the UK in 2014. Another example, the Crime Survey for England and Wales recently introduced fraud into the types of crime that it counts – it turns out there are nearly as many incidents of fraud in a year as all other types of crime combined in the UK (approximately 5 million incidents: link). So much for the apparent convergence of the crime survey and police figures. What counts as criminal, what is counted as criminal, depends as much on convention as on law. Consider the town in the UK that introduced a Public Space Protection Order that made it illegal for more than two youth under 17 to congregate (link). Consider the self-fulfilling theory that poverty causes crime, therefore the police focus their efforts on poor neighborhoods, therefore poor people are caught committing crimes, therefore the numbers show that poor people commit crimes. The same logic applies to the theory that black people commit more crime. There are strong pressures on law enforcement to focus their attention on marginalized groups, and theories of criminality have a tendency to be self-fulfilling. It beyond the scope of this article to comment on who commits more crime, my point is that we cannot rely on crime statistics to tell us who commits more crime.
I shall summarize. The purpose of this post was to examine the use of crime statistics as a means of holding the police accountable. I began by briefly examining why police statistics are used to hold the police to account. First, its easy. Second, it appeals to the public's intuitive understanding of what good policing attains. Third, crime figures make good headlines. I then propose that the most significant flaw in using crime figures to hold the police to account is that crime figures tell you very little about crime. First, there is a lot of crime that the police do not detect. Second, increasing detection of one type crime must come at the expense of others. Third, it is problematic for the police to be simultaneously responsible for recording crime figures and being held to account through those crime figures. This strongly incentivizes the police to game the numbers, particularly in the face of unrealistic targets. Fourth, crime is socially constructed. That which is considered criminal changes over time. Social power dynamics play a role in crime theory, and crime theories can often be self-fulfilling.
I will reduce each argument in a single sentence. The police cannot observe all crime. The more attention that they give to a particular type of crime, the less they can give to others. The police are incentivized to fiddle their numbers to meet unreasonable public and political demand. There is no 'real' number of crimes to count.
If crime figures do not work, how should the police be held to account? To answer this question, we must first further explore the purpose of policing, which I shall do in my next post, "A (very brief) history of violence."
Articles on how police manipulate crime figures:
References
Newburn, Tim, ed. Handbook of Policing. Routledge, 2012.
Comentarios