First published by ISS Today

In 1982, George L Kelling and James Q Wilson penned an article that created the idea of broken windows policing. Kelling, who died on 15 May aged 83, will be remembered as the father of broken windows policing, and for his early contributions to what has become evidence-based policing. Both are important for South Africa – a country replete with broken windows and not enough evidence-informed police practice.

The broken windows theory argues that policing should promote public order by addressing minor infractions, such as loitering and public drunkenness. It was made famous through its association with the rapid decline in crime in New York City in the early 1990s. But what is it and is it really effective?

Kelling and Wilson’s argument was based on the findings of a policing experiment to test whether increased foot patrols reduced crime. They found that while this tactic didn’t bring crime down, it did improve the public’s feelings of safety and perceptions that crime had declined, and increased trust in police. Officers walking foot beats for the experiment also reported improved morale and job satisfaction.

This is an early example of evidence-based policing – a hypothesis was formed, tested and evaluated, and its lessons used to inform police practice and urban governance. In this instance, the experiment birthed the broken windows theory of policing.

Kelling and Wilson read a number of conclusions into the study. First, they noted that feelings of fear need not be based on actual crime. Rather, perceptions of disorder alone can cause public anxiety. Second, they proposed that when signs of disorder, such as a broken window, are left unchecked, they signal lack of government or community care. This promotes rule-breaking and other anti-social behaviour resulting in more broken windows and a deteriorating environment.

In turn, people withdraw from public space, and unlawful activities increase, potentially resulting in serious criminality. Their solution was to address or remove signs of disorder before they generated others. This logic is both clear and controversial. It has been interpreted as promoting the criminalisation of the poor and vulnerable in public space, although this has been denied.

What about impact – does it actually work? A 2015 systematic review of 30 randomised experiments and quasi-experimental tests of broken windows-type policing found that their crime reduction effect was modest. Importantly, the greatest impact on crime was produced by problem-oriented interventions that targeted particular disorder in specific places. In contrast, aggressive behaviour-focused interventions were ineffective.

These are important findings. They suggest that simply arresting large numbers of perceived transgressors for petty offences doesn’t result in long-term safety. And when pursued with persistent pressure from police managers without attention to the methods used, such policing can cause harm, and result in more crime.

The South African Police Service (SAPS) is particularly vulnerable to this. It has a history of chasing arrest and crime-reduction targets while policing millions of impoverished and disenfranchised people. Instead, it should holistically tackle specific problems in specific areas with well-considered interventions that can prevent crime. This requires interventions beyond only policing, but police remain key to their success.

Police are ideally positioned to identify areas of disorder (and crime), and pull the levers of government and civil society required to address them. When resources are intentionally focused in this way and deployed according to a vision shared by all stakeholders, they are most likely to be effective.

Broken windows policing should not imply a uniform approach across a city or country, and claims of such by politicians should be resisted. Rather it should be about focused problem solving with an emphasis on minor infractions in otherwise “orderly” spaces, ideally using methods that promote trust in police. For example, police alone can’t prevent all drug dealing or assault because some factors behind these offences are beyond the scope of policing.

Police should partner with other government and civil society structures and offer more incentives for behaviour change than mere arrest and prosecution. In this way, order can be maintained without ushering people into a criminal justice system that may do them more harm than good.

Broken windows theory was born out of an experiment in policing that was carefully planned and evaluated. Although it sought to reduce crime, factored into its analysis were many more indicators than reported crime alone.

In South Africa, reported crime is often emphasised in police performance evaluations. This has made public safety primarily a police concern and limited the multi-agency interventions required. The sharp rise in serious violent crime in recent years shows that this approach has failed.

Among George Kelling’s legacies is confirmation that researchers and police can collaborate for impact. Speaking of Kelling shortly after his death, former New York police chief Bill Bratton said of the academic: “I put into practice his theories and they worked.”

To improve policing in South Africa – whether through broken windows or other approaches – we need to grow a culture and capacity of planning and evaluating police practices. We need to generate and test our hypotheses and produce evidence to know how best to fix our many broken windows while simultaneously building trust in our police. DM

Andrew Faull is Senior Researcher, Justice and Violence Prevention, ISS