Tag Archives: predictive policing

Ghettoside

I’m reading Ghettoside: A True Story of Murder in America, a book about homicide in Los Angeles between the late 1980s and early 2000s. The word “ghetto”, by the way, might seem loaded but it is how residents and police referred to the neighborhood that is the focus of the book. It’s a very interesting and also disturbing book. It tells a little bit different story than what I have been reading in books and the media elsewhere. In the view of this book, a central factor in high homicide rates, at least in Los Angeles at the time covered in the book, is that police departments don’t solve murders of black men and boys at the same high rates that they solve murders of other groups. This leads to a situation of lawlessness where a sort of “law of the street” develops. In this view, people would actually like more help from the authorities if they felt it was fair and professional, but they don’t believe they can get it so they take matters into their own hands.

The book talks about disputes and arguments among men and boys getting out of hand and leading to cycles of revenge and retaliation. Homicide detectives do their best, but even the best homicide detectives have limited capacity, and training new ones is difficult. When there is a spike in homicides, the supply of good homicide detectives does not increase in kind. Cases get rushed and a smaller fraction of the total get solved. People correctly learn that they are likely to get away with murder, and that contributes to the feedback loop. In Los Angeles at the time, the situation escalated to the point that total strangers were murdering each other simply for being in the wrong neighborhood or wearing the wrong color clothing.

The book argues that Los Angeles at the time was diverting resources from investigating and solving homicides to “violence prevention” and “predictive policing” programs, which were politically popular but less effective than simply solving more cases would have been. It also argues that people can feel harassed and overpoliced at the same time they might support more investigation and solving of violent crime cases if they felt it was fair and effective. I hear echoes of this in the media during the current homicide wave we are experiencing in many U.S. cities. Maybe the violence prevention approaches have improved and have more evidence behind them, but we do hear both that homicide is way up and that the clearance rate is down. And we perpetually hear about the idea of a lack of trust and respect between police and residents of primarily black neighborhoods.

It’s interesting that the crimes discussed in the book are almost all gun crimes, but this is not a book that focuses on guns. Nor does it focus on the drug trade. It focuses on the people involved and their motivations on all sides, from victims to perpetrators to police. It mentions a few police shootings of suspects in passing, but this is also not a focus of the book.

predictive policing

Here’s an interesting article on predictive policing from Motherboard. People are concerned that if a particular area has been overpoliced in the past, that is where the algorithms are going to predict crime in the future and they will continue to be overpoliced. Others just don’t like the idea of proprietary algorithms. I think any of these concerns could be badly depending on how it is implemented, but I don’t see why the tool itself could not be implemented in a fair way. In fact, I don’t see why measures to prevent discrimination couldn’t be built into the algorithms themselves. If the algorithms say people in a particular area or in a particular demographic group are being arrested at higher rates, it could help the search for route causes and preventive measures to help a particular group revert back to the mean. Transparency seems good in principle, maybe publishing some generalized statistics and maps, but of course if it is too predictable exactly where the police are going to be and when, people could take advantage of that. You could try to get around this by balancing random and targeted patterns within the algorithm.