Broken windows theory holds that visible signs of disorder, such as broken windows, graffiti, and public drinking, signal that a neighbourhood is unmonitored and thereby invite serious crime. Proposed by James Q. Wilson and George Kelling in 1982, it became one of the most influential ideas in criminal justice and one of the most heavily criticised.
The argument is about signals rather than about the offences themselves. An unrepaired broken window communicates that nobody is paying attention; that perception emboldens minor infractions; minor infractions further erode informal social control as residents withdraw from public space; and serious crime follows into the vacuum. The policy conclusion is that addressing small disorders prevents large ones.
New York City adopted it most prominently in the 1990s under Commissioner William Bratton, who had worked with Kelling, pursuing fare evasion, graffiti, and public disorder alongside the CompStat system of data-driven deployment. New York's violent crime fell dramatically over the decade, and the theory was widely credited. It was exported to police forces internationally.
The New York result is contested. Crime fell across the United States in the same period, including in cities that adopted no such strategy, which suggests national causes. Candidate explanations include the decline of the crack epidemic, demographic change, economic conditions, increased incarceration, and changes in lead exposure. Franklin Zimring, who studied the New York decline closely, concluded that policing contributed but that the size of the contribution attributable specifically to broken windows is unclear.
The direct tests are unfavourable. Bernard Harcourt's reanalysis of the original supporting data found the relationship between disorder and serious crime largely disappeared once neighbourhood poverty and other conditions were controlled for. Robert Sampson and Stephen Raudenbush found that both disorder and crime were driven by underlying structural conditions, particularly concentrated disadvantage and weak collective efficacy, rather than disorder causing crime.
What actually works may be narrower. Reviews suggest that community problem-solving approaches to disorder show benefits, while aggressive misdemeanour arrest strategies show little effect. If so, the successful component is not the one most forces implemented.
The costs fell unevenly. Order-maintenance policing in practice meant very large numbers of stops and misdemeanour arrests concentrated on young men in poor and minority neighbourhoods. In New York, stop-and-frisk was found unconstitutional in its application in 2013. Critics argue the theory supplied a respectable rationale for practices that damaged the relationship between police and the communities most affected, which may itself reduce the informal social control the theory says matters.
Kelling maintained that the theory was widely misapplied, that it was never intended to license mass arrests, and that its point was maintaining order in partnership with residents rather than aggressive enforcement against them. Critics reply that a theory's real-world implementation is part of its assessment. The underlying intuition that environments influence behaviour retains support; the specific causal claim, and the enforcement strategy built on it, do not command consensus.
