Broken Windows Revisited | 3: Proactive Policing

The 3rd and final part of our Broken Windows reassessment looks at the latest American research that questions the claim proactive / Zero Tolerance policing prevents minor forms of social disorder developing into major forms.


In two previous posts re-examining Broken Windows we’ve considered both its general theoretical and empirical background and its theoretical origins in ecological theories of crime. In this third and final part we assess one of Broken Windows’ key theoretical components: the claim that minor forms of social disorder, if allowed to go unchecked, result in major forms of disorder. Or, as Wilson and Kelling (1982) originally put it:

If a window in a building is broken and is left unrepaired, all the rest of the windows will soon be broken…one unrepaired broken window is a signal that no one cares, and so breaking more windows costs nothing”.

If, as its proponents claim, relatively minor forms of social disorder lead to larger, more-serious, forms of criminal disorder – as Bratton and Kelling (2015) have more-recently expressed it “A neighbourhood where minor offenses go unchallenged soon becomes a breeding ground for more serious criminal activity and, ultimately, for violence” – the way to control the latter is to prevent the former and one way of doing this, introduced by the New York Police Department (NYPD) in the late 1990s, was through a system of proactive policing.

This broadly involved police officers “on the street” taking a more-active role in either trying to prevent criminal activity before it occurred (through things like dispersing loitering groups of young men, stopping and searching potential offenders and so forth) or by immediately punishing every instance of criminal or misdemeanour activity, through things like spot-fines and arrests, the moment it occurred. This particular form of Order Maintenance Policing (OMP) has come to be known as Zero-Tolerance Policing (ZTP) because the police exhibit no tolerance towards any offender, no-matter how trivial or seemingly inconsequential the offence.

As an aside, it’s important to note that while ZTP is often seen as synonymous with Broken Windows, we shouldn’t conflate the two: Broken Windows doesn’t necessarily involve Zero Tolerance Policing, even though the two are frequently seen to be one and the same thing. There are a number of different ways of preventing the escalation of minor forms of disorder into major forms, of which ZTP is but one – although ZTP has, particularly in America, been increasingly used by forces such as the NYPD as the primary or de facto way of putting Broken Windows into practice.

One of the reasons Broken Windows has come to be influential with politicians, police and public alike is that it has a certain face validity; that is, it seems like a plausible way to both explain how and why crime develops in a particular locality and, by extension, how to prevent criminal behaviour spiralling out of control.

Part of the reason for this is that the idea of major forms of disorder stemming from unchecked minor forms is something that has a certain resonance with our everyday personal experience. If you think, for example, about a work desk that you gradually allow to become cluttered with books and papers, it eventually becomes difficult and time-consuming to find the things you need: “major” disorder, in other words, stemming from untreated minor disorders…

While you could, every once in a while, instigate a “big clean-up” it might be easier to keep things tidy while you work. This takes a bit more effort and willpower but should, all things being equal, save you time and effort in the long run…

The problem here, however, is that societies are not like individuals and “maintaining social order” can be much more difficult than maintaining a tidy desk – particularly if the area that requires maintaining is home to a wide range of poverty-stricken individuals and families who don’t necessarily maintain strong social and moral ties.

This, in terms of Broken Windows, is where proactive policing enters the picture: as a way of imposing some sort of order on a situation that tends towards the disorderly. This, on the face of things, seems to make sense in terms of our general understanding of social order and disorder but the problem we have is how to test this idea. How, for example, can we evaluate the validity of the Broken Windows argument “in the real world” of offenders and control agents?

One obvious way would be to compare an area that had been subject to proactive forms of ZTP with the same area at a later point when ZTP was no-longer in operation – and this is exactly what Sullivan and O’Keeffe (2017) were able to do thanks to an opportunity to conduct a natural experiment in New York in late 2014, early 2015.

A Natural Experiment


For reasons that needn’t detain us here, the NYPD began a “work-to-rule slowdown” on 22nd December 2014 that would eventually last for 7 weeks. Over this period police officers, according to Sullivan and O’Keeffe, “performed only the most necessary duties and refrained from proactive policing by refusing to get out of their vehicles to issue summonses or arrest people for petty crimes and misdemeanours”.

In other words, for nearly two months – encompassing the normally-busy Xmas and New Year holiday period – police patrol officers declined to issue criminal summonses aimed at “quality-of-life violations” (such as consuming alcohol in public or disorderly conduct), or to engage in speculative “stop, question and frisk” (SQF) activities that involved stopping and searching individuals for contraband. As they further note “Emblematic of the slowdown’s effects (and the change from proactive to responsive policing), zero summonses were issued for quality-of-life violations on New Year’s Eve 2014”.

This, to say the least, was highly unusual.

Equally unusual was the fact that having analysed and compared several years of NYPD data Sullivan and O’Keeffe found that, contrary to the predictions of Broken Windows, civilian complaints of major crimes such as burglary, assault and theft “decreased during and shortly after the sharp reductions in proactive policing”.

In other words, their data not only suggested the claimed relationship between allowing minor social disorders and increased types of major disorder was demonstrably false (“civilian complaints of major crimes declined by approximately 3 – 6% during the slowdown”), it also implied the reverse was true: that “aggressively enforcing minor legal statutes” served to incite more severe criminal acts – an idea compounded by the fact that, as Sullivan and O’Keeffe note, following the return to normal police working in the spring of 2015 – and the subsequent reapplication of proactive policing in New York City – the major crime rate reverted to its pre-slowdown levels…


The fact Zero Tolerance Policing and Broken Windows are not simply synonymous means that Sullivan and O’Keeffe’s research findings do not automatically falsify the basic claim underpinning Broken Windows that minor forms of social disorder, if left unchecked, escalate into more serious forms. It is possible, for example, that a version of Broken Windows could be operationalised that does not involve the use of proactive policing – but until and if such alternatives occur what we are left with is a general form of “Broken Windows policing” that tends towards being highly proactive and which, according to Sullivan and O’Keeffe’s research, does not produce the results claimed for such an approach.

In addition, they suggest a number of reasons why proactive policing is not only ineffective in meeting it’s stated aim of reducing major disorder but also contributes to the creation of the very problem it exists to solve.

  • proactive policing diverts resources away from major crimes units and, by so doing, limits their ability to act effectively against serial offenders and criminal networks.
  • proactive policing that targets large number of trivial offences and offenders is disruptive of the kinds of informal communal controls that are known to be effective against group-level offending.
  • the criminalisation of large numbers of people for relatively trivial offences impacts on both individuals – who may, for example, lose their jobs and find it difficult to find new employment – and communities by driving people off the streets. This breaks down the kinds of informal social controls that can be effective in limiting the development of large-scale forms of social disorder.
  • if officers apply aggressive forms of low-level policing inconsistently across a community this can undermine public perceptions of police legitimacy which, in turn, erodes confidence in and co-operation with the police – particularly in relation to solving more serious forms of community offending.
  • a cumulative effect of the above is one that “increases legal cynicism” in the ability of the police to apply the law consistently and effectively. Where people, for example, fear being victimised by the police if they report criminal behaviour it makes it more-likely they will rely on various forms of extra-legal sanctions (“taking the law into their own hands”) and “informal institutions of violence” (such as gangs) as a replacement.
  • Overall Sullivan and O’Keeffe conclude that “sharply reducing proactive policing in areas where it had been deployed pervasively may actually improve compliance with legal authority, thereby reducing major crimes”.


    Sullivan and O’Keeffe (2017) “Evidence that curtailing proactive policing can reduce major crime

    Wilson and Kelling (1982) “Broken Windows: The police and neighborhood safety

    Bratton and Kelling (2015) “Why we need broken windows policing

    Stay Updated

    Enter your email to be notified when we post something new:

    Archived Posts