Lincoln (Neb.) Police Chief Tom Casady — whose agency patrols a city of about 75 square miles and protects a population of a quarter of a million people — sees a the potential for predictive policing to lead to the end of one paradigm in law enforcement that’s been omnipresent for many decades. He highlighted his belief by posing a rhetorical question to the attendees at IACP 2010 in Orlando.
“What is the number one activity of police officers in U.S. cities? What is the number one thing that they do?” Casady paused before answering his own question. “It’s driving around aimlessly, burning fossil fuels, waiting for the next call from the dispatchers. For those of you in the room who are Chiefs, how many times have you heard your officers say, ‘We’re going from call to call to call’ and you know that that’s not true. There is an awful lot of driving around aimlessly waiting for something to happen. I don’t think this can last.”
Innovative, Effective, and Inexpensive
It’s certainly not news to state that police agencies of every size across the United States are facing diminishing budgets and increasing demands for police services. It’s equally well known that tax revenues for most cities and states have gone down or remained flat, while the cost of doing business for police agencies has increased. Casady, along side co-presenter Jim Mallard of the Arlington (Texas) Police Department, showed that predictive policing can be used by just about any size agency to help alleviate the budgetary burden and actually increase the effective outcomes of your agency’s efforts.
“We’ve got to use our resources more effectively,” Casady explained, “and that means targeting our efforts more intensely on efforts that do not involve simply driving around waiting for something to happen. We’re going to be forced to do more with less, and predictive policing has the potential to help us be more productive and more efficient.”
Following Casady’s opening remarks, Mallard explained that predictive policing is an affordable, doable thing for any agency, no matter what the budget constraints may be. “I’m not saying that you can’t go out to the Expo and spend money on some sort of software with sophisticated mathematical algorithms. You can certainly do that — there are some applications that do some pretty incredible things. But to get into the realm of predictive policing does not necessarily require a huge capital investment in extremely sophisticated software for which there’s no analyst to provide interpretation and context. You can use tools that are less expensive — you can do these types of things and you can be proactive in other ways.”
The Punxsutawney Phil of Policing
In a very real sense, predictive policing is really bout risk assessment — not unlike the type of analysis a lender might do when considering a loan applicant, or the weather service would do when evaluating atmospheric conditions to predict the proability of a storm.
“I think we can draw a lot from the business community, and credit scores are a good example — credit reporting companies are really good at being able to predict who’s a risk and who’s not. They are able to determine who is most likely to pay their bills and who’s not. We’re talking about identifying behaviors and conditions that make it ripe for future crime to occur, and that’s where weather comes in. We know that when a cold front and a warm front meet over a flat plain on a hot summer day that we might get thunderstorms. That doesn’t mean we will get thunderstorms — it only means that the conditions are more ripe in that location than in other locations,” Mallard said.
Mallard opined that predictive policing is all about risk assessment — trying to predict what your risks are in your city and in specific parts of your city. The truly strategic element of predicative policing is not truly about policing at all, but about city planning. It’s about working with city planners when building developments are proposed, indentifying the risks inherent in a proposed project and mapping police resources toward addressing those risks.
“There are a lot of different types of risks,” Mallard said, “There are risks of victimization, risk of future crime, and risk to your own internal resources. What can we anticipate the impact that the construction of a new development will have on our departmental resources? And that’s why we need a forward-looking strategy rather than being reactive”
Not Just Cops on Dots
Mallard used the example of auto theft to illustrate the forward-looking mindset that goes with the predictive policing model.
“We know, for example, that certain makes and models of cars are more likely to be stolen than others. So why not go to the DMV and get the addresses of where those vehicles are likely to be? That’s a forward-looking, data-driven, process to make a deployment decision that helps prevent crime. You can arm yourself with information with a non-traditional source to determine what your deployment strategy should be. You’re not waiting until tons of vehicle thefts have occurred and look back saying, ‘Well, we had a bunch of Hondas taken in this area, so let’s go put a bunch of cops there so maybe we won’t have so many stolen Hondas anymore.”
Casady added, “We know an awful lot about crime and place. We’ve really ratcheted up our ability to predict where crime is going to occur. Give me a bar, tell me what their business model is, tell me what their clientele is going to be like, who they’re going to market themselves to, and I think I can predict pretty accurately whether we will or will not have a hot spot for assaults and disturbances. And we know more about crime and people. When it comes to criminal conduct, past performance is the best predictor of future performance, and we know fairly accurately who the frequent fliers are in our communities. The men and women who are most likely to continue to involve themselves in criminal activity are those men and women who have involved themselves in criminal activity in the past.”
Casady explained further that a recent development in predictive policing is a vastly increased knowledge base about the victims of crime. For example, a person who has been in a relationship in which they were victims of domestic violence, there’s a much greater probability that they’re get into another relationship in which they were victims of domestic violence.
“This is not new learning so much as the coalescing of that learning with new technology, which has made it much faster for us to predict where crime is going to occur. Those of you who have been in policing a while know well what it means when a quarter-million-square-foot big box retail facility goes in on a street corner in your city, for example. You know what that’s going to mean in terms of demand on police services and crimes like theft and fraud and identity theft. You know that because you know exactly how it happened at the other big box retail store across town. So we know a lot — where crime is likely to occur, who is likely to commit those crimes, and who are the people most prone to be victims. With predictive policing, we’re using that knowledge about potential hot spots, crime victims, and the criminals themselves to make deployment decisions that actually prevent crime, not just respond to it.”
Measuring What Matters
For decades, elected officials — and to some extent the public — have been heavily invested in measuring law enforcement outputs not law enforcement outcomes. For example, writing a ticket is an output not an outcome. The purpose of writing that ticket is to effect the outcome of making that street safer. But we don’t measure typically the value of the property destroyed in traffic collisions year over year, or the numbers of deaths and injuries due to those collisions, as much as we continue measuring the numbers of tickets written or arrests made. That, they said, is an unsustainable model for the future.
Predictive policing gives agencies a set of tools to do more with less. As a practical matter, if you know there is a particular neighborhood in your city that is particularly prone to crime — whether it be because of the nature of the housing there, or the predominance of paroles, or the presence of some other risk factor — and all you do with that knowledge is to deploy more troops to make more misdemeanor arrests, you likely will have missed an opportunity to do something strategically impactful with that knowledge.