Ch 14. Broken windows

Any cave, any house, any car that looks watched-over or well tended is likely to seem less vulnerable than places and possessions that appear to be neglected or abandoned. The concept was vividly demonstrated in 1969 by the psychologist Philip Zimbardo, whom we met in Chapter 7.

It was Philip Zimbardo who set up a prison experiment in the psychology department at Stanford University where volunteer guards and prisoners quickly let their play-acting get out of hand.




Intrigued, he and two students took a sledgehammer to the Palo Alto car and smashed a window to see if that would make a difference. It did, immediately. People came across to join the fun. Within a few hours that car too had been wrecked and finished up resting on its roof.

P. G. Zimbardo, The human choice: Individuation, reason, and order versus deindividuation, impulse, and chaos. Nebraska Symposium on Motivation, 17, 1969, pp. 237–307.




Some years later, his experiment caught the attention of two Harvard academics, James Q. Wilson and George Kelling, who wrote a popular article with, most importantly, a catchy title: ‘Broken Windows’.

James Q. Wilson and George L. Kelling, ‘Broken Windows’, The Atlantic Monthly, March 1982, Vol. 249, No. 3, pp. 29–38[].




In 1990 America’s iconic city had been described on the cover of Time magazine as ‘The Rotten Apple’. That year, New York State recorded 2,600 murders

In other words, in 1990 the murder rate in New York State was almost seven times that in Greater London.




History has conflated two quite different ideas: broken windows and zero tolerance. It has also neglected a fundamental feature first discovered by Sir Francis Galton: regression to the mean. What goes up most, usually comes down most.

Technically, regression to the mean is a statistical phenomenon whereby variables which are outliers on a first measurement are likely to be closer to normal on subsequent measurements. A similar feature, sometimes called regression to the norm, is frequently observed in personal, social and political settings where extreme variations often to go into reverse. Thus, I may get an exceptional amount done one day but over time my output will average out. So in policing there is a problem of establishing whether a big crime challenge declined due to a specific intervention or because of natural fluctuations.




Therein also lay the question of who should take the credit, for none of the four main characters, Giuliani, Bratton, Timoney and Maple lacked their share of ego or ideas, and all felt at times that others took more than their fair share of glory.

Jack Maple used to say that the first thing John Timoney did was to try to have him fired.




Bratton needed to reclaim the streets and reassert police authority. ‘Responding to 911 calls was random behaviour. We had lost all control over our police resources,’ he recalls. Instead his cops were to be proactive. Following the broken windows model, police in large numbers were put back on the beat and took no nonsense. ‘We had to take the beachhead. New York was an incredibly violent place so it took a very robust form of community policing at first. Later you can have a very different style of policing.’

Conversation with author.




But while most people celebrated, and conservatives exulted, some on the left of politics were dismayed.

Zero tolerance appealed to law and order sentiments. Furthermore one of the Broken Windows authors, James Q. Wilson had been chairman of a White House Task Force of Crime in 1967 and was well known as a conservative commentator who despaired of sociological theories or of reforming criminals – he preferred incarceration to keep repeat offenders off the streets. But he is thoughtful and no dogmatist. His co-author, George Kelling, has a background in social work and in raising money for alternatives to prison. Even so, as a colleague in Kelling’s later university, Rutgers, conceded to me, ‘Liberals don’t like either of them.’




… the trend persisted, down from 2,262 murders in 1990 to 629 in 1998. As Bill Bratton later wrote, ‘It’s hard to hide that many bodies.’

William Bratton and George Kelling, ‘There are no cracks in the broken windows’, National Review, 28 February 2006.




… police spending swelled while school budgets were being cut (police numbers eventually increased by 50 per cent).1  More disconcerting to liberals was what they saw as the price in civil liberties. The NYPD was accused of being overbearing.2

1  Sworn officer numbers increased from 26,500 in 1990 to 40,864 in 2000.

2 Complaints of police heavy-handedness were loud for a time, but this may be what I call the ‘rising standards fallacy’ – in other words, it might be that officers’ behaviour deteriorated occasioning more protests, or it might be that things improved so much that grievances were worth pursuing. In time official complaints against police declined, and by 1998 police shootings had reached their lowest level since data were first recorded.





They saw the city’s deliverance from its crime pandemic not as victory but defeat. Bernard E. Harcourt, in Illusion of Order: The False Promise of Broken Windows Policing, wrote:

What we are left with today is a system of severe punishments for major offenders and severe treatments for minor offenders and ordinary citizens, especially minorities, a double-barrelled approach with significant effects on large numbers of our citizenry. The problem, in a nutshell, is that order-maintenance crackdowns permeate our streets and our police station houses while severe sentencing laws pack our prisons. We are left with the worst of both worlds.

Bernard E. Harcourt, Illusion of Order: The False Promise of Broken Windows Policing, Harvard University Press, 2002, p. 6. Harcourt was then a young lawyer and political-cum-social theorist at the University of Chicago whose liberal leanings recoiled at the idea of oppressive policing, and the title of his fight-back book and its cover picture of prison bars left little room for doubt on where he stood. He acknowledged that more police can reduce crime; it was the aggressive focus on minor crime and perceived disorder that he objected to. ‘Despite repeated claims there is no reliable evidence that the broken windows theory works’ (p. 8). In a later paper he dismissed broken windows as ‘a cute slogan good for marketing’ and subscribed to the regression to the mean explanation of what happened in New York: ‘The data support what we call Newton’s Law of Crime: what goes up, must come down, and what goes up the most, tends to come down the most.’ Bernard E. Harcourt and Jens Ludwig, ‘Broken windows: new evidence from New York City and a five city social experiment’, Chicago Public Law and Legal Theory Working Paper No. 93, June 2005[]. For a splendid debate involving Harcourt, one of the most knowledgeable critics of Bratton’s claims, with one of the most articulate defenders, David Thacher, a professor of Public Policy and Urban Planning at the University of Michigan, see: []




… despite all the talk of zero tolerance, incarceration rates remained pretty much the same.

Franklin E. Zimring, The City That Became Safe, Oxford University Press, 2012, pp. 164–70.




In fact zero tolerance was a red herring. The phrase was never mentioned in ‘Broken Windows’ and if it conjures images of robocops swarming on the streets enforcing diktats from city hall it could scarcely be further from what Kelling and Wilson originally envisaged.1 Nor was the expression ever used by Bratton, except sometimes to disparage it. It belittles the complexity of policing, smacks of over-zealousness and is not a credible policy because it promises too much to law-abiding citizens and cries wolf to offenders. ‘Consequently, zero tolerance is neither a phrase that I use nor one that captured the meaning of what happened in New York City.’2


1 In their original 1982 essay Kelling and Wilson proposed community-oriented policing as a means to improve public confidence and were at pains not to promise the strategy would cause substantial cuts in major crime. But in the light of New York’s experience Kelling, in particular, changed his mind and did claim that restoring order directly reduces crime. He even went so far as to put a figure on it: ‘Over 60,000 violent crimes were prevented from 1989 to 1998 because of “broken windows” policing.’ George L. Kelling, William H. Sousa, Jr., ‘Do Police Matter? An Analysis of the Impact of New York City’s Police Reforms’, Civic Report, No. 22, December 2001. James Q. Wilson has always remained more equivocal, and was quoted in one interview saying, ‘God knows what the truth is.’


2 Curiously, the term ‘zero tolerance’ caught on more in the UK than in the US, perhaps because it was a simple slogan that appealed to journalists and politicians. Many people remain undaunted in its use even though in his contribution to a book of essays entitled Zero Tolerance, Bill Bratton so vehemently disavows the term. Zero Tolerance, Policing in a Free Society, Norman Dennis (ed.), Institute for Economic Affairs, London, 1998, ISBN 0-255 36432-6, pp. 42–3.





… New York was anyway emerging from its darkest hour. What had happened in the late 1980s had been an aberration. The city had been through an economic and employment crash, appalling fiscal and political mismanagement and staggering levels of police corruption.

While New York’s financial malaise was an aberration, corruption of its police had been endemic from the start. The force was supposed to be modelled on London’s Peelers but from its creation in 1845 the New York version was plagued with antipathy, migrant rivalry and political intrigue. The first recruits were drawn almost entirely from Irish immigrants who wore plain clothes at first, insisting they would not be ‘liveried lackeys’. The word cop or copper probably derives from their copper badge of office. Blue uniforms were finally issues in 1853 but Irish dominance of policing continued, led by a clique of Irish Democrats who ran the city. In 1857 Republicans who governed New York State sought to wrest authority from the cabal in Manhatten and formed an alternative force of Anglo-Dutch extraction. There was chaos and rioting as the old Municipal officers, supported by Irish gangs, clashed with the new Metropolitan Police. It took the National Guard to restore order and the Republicans were eventually outmanoeuvred by an Irish-American political faction, the Tammany Society, which controlled the Democratic Party nominations and whose headquarters, Tammany Hall, became a byword for corruption. Police brutality and bribery were just part of a grand degeneracy which manipulated the courts, the legislature, the treasury, the ballot box and, above all, contracts for public works. Despite repeated clean-ups, a series of scandals involving violence, pay-offs and protection rackets dogged New York policing throughout the century that followed. In the 1970s the Knapp Commission uncovered systemic pay-offs to police by organised crime. Twenty years after that an inquiry initiated by Mayor Dinkins reported that cops themselves were now driving the bribery deals, extorting money from drug dealers, stealing and selling drugs, beating up suspects and falsifying reports. The NYPD leadership had turned a blind eye and ‘at times concealed corruption rather than rooting it out’.




Homicides and recorded robberies had dropped a third in the two years before Bratton’s appointment, a rate of decline slightly faster than anything that followed.

Franklin E. Zimring, The City That Became Safe, Oxford University Press, 2012, pp. 21 and 23.




Compstat can be a rough business. I never experienced one led by Jack Maple but I have seen others including the nearest thing: one under John Timoney when he moved to be chief of police in Philadelphia.

John Timoney personally drove me round town in the early hours, stopping every now and then to chat to his officers (or rather, to keep them on their toes), and when next day he took me to Philadelphia’s magnificent train station to bid farewell, he was plainly a celebrity; citizens came up to applaud him and a congressman wanted to shake him by the hand. Timoney went on to take over as police chief in Miami and then as a police consultant to Bahrain. In 2000 Esquire magazine called him ‘America’s best cop’. Bill Bratton, meanwhile, eventually became chief of the Los Angeles Police Department, is in much demand on the lecture and seminar circuit and was touted as a possible Commissioner of London’s Metropolitan Police. Jack Maple died tragically of cancer aged forty-nine; his role in New York’s turnaround rarely celebrated outside police circles. Giuliani, of course, went on to become front-runner for a while as Republican candidate for the presidency of the United States.




…it is rare that police chiefs inherit such a massive increase in resources.

NY policing numbers had grown from 32 per 10,000 citizens in 1980 to 50 per 10,000 in Bill Bratton’s time, so they had more officers at their disposal than any other city in the country.




The public, the politicians and the media had become so used to rising rates of murder, rape and robbery that no one could believe it when lines on the charts began to turn. Falling crime was not part of the script.

Much the same happened in the UK when crime rates fell from 1995. Journalists and political commentators denounced improving figures as crude manipulation by the government, and persisted in their disbelief even when the falls became consistent across different measures and over several years.




As one thorough study concluded, ‘New York’s homicide trend during the 1990s did not differ significantly from those of other large cities.’1 It wasn’t uniform, it wasn’t predictable. Some places did better with murder, some with robbery, some with drug offences, some with burglary. Yet it happened across the board.2

1 Richard Rosenfeld, Robert Fornango, Eric Baumer, ‘Did Ceasefire, Compstat, and Exile reduce homicide?’, Criminology & Public Policy, Vol. 4, No. 3, September 2005.[]

2 Arguably other cities also contributed to New York’s salvation, not least Detroit with its new theft-resistant automobiles – see Chapters 2 and 3.




New York still has five times the homicide rate of London. But taking the long view, it has seen a reduction twice as fast and twice as long as the average for major cities in America. Whether broken windows theory was in part responsible we shall never know; it is impossible to disentangle so many things that happened simultaneously and interacted unpredictably.

The best textbooks on the subject run through lots of possible contributory factors – incarceration rates, police tactics, social theories, drug fashions demographics (New York’s population of young males actually increased during the drop in crime) – and end up shrugging their shoulders in bewilderment. What is clear is that New York demonstrates ‘the shortcomings of traditional criminological theory’.Franklin E. Zimring, The City That Became Safe, Oxford University Press, 2012, p. 216. See also John Conklin, Why Crime Rates Fell, Allyn and Bacon, New York, 2003.




There are examples of how we follow social cues, going back at least to pioneering work by the psychologist Solomon Asch in the 1950s, who used stooges to induce students to give wrong answers to simple questions. Another psychologist, Robert Cialdini, once got several people to look up at the sky and, as more and more passers-by joined in, the crowd began to spill into the street and stop the traffic.

Robert Cialdini, Influence: The Psychology of Persuasion, Harper Collins, 2007. Thus it is hardly surprising that if we see litter or broken windows – cues that other people don’t care – we are less inclined to care ourselves, so that mouldering environmental problems can lead to a spiral of antisocial behaviour.

The extent to which we are socialised is vividly demonstrated in the Milgram and Zimbardo experiments which show how easily we conform to models of bad behaviour. But equally there are studies which show how we act better if cued to comply with positive social norms. A splendid little experiment in Newcastle University found people put twice as much cash in an honesty box when under the steady gaze of a pair of eyes, even though the eyes were just a picture. Each week for ten weeks researchers changed posters above tea- and coffee-making kit in their department. Sometimes there would be flowers, but on alternate weeks there were different faces showing close-ups of a stare. And the eyes have it. When the honesty box was cashed up flower-power was a flop but, presumably unconsciously, dons were prodded into paying when being spied on by watchful poster.


Money paid per litre of milk consumed as a function of week and image type

Melissa Bateson, Daniel Nettle and Gilbert Roberts, Cues of being watched enhance cooperation in a real-world setting, Biology Letters, Royal Society, London, doi:10.1098/rsbl.2006.0509




A stamped addressed envelope clearly containing cash was left hanging from a post box. Most passers-by ignored it, and others pushed it through into the mail, but 13 per cent took it out and pocketed it. After the letter box was sprayed with graffiti or surrounded with discarded rubbish, twice as many people stole the money.

Kees Keizer, Siegwart Lindenberg and Linda Steg, ‘The Spreading of Disorder’, Science, December 2008, Vol. 322, No. 5908, pp. 1681–85. [doi: 10.1126/science.1161405.]