Assessing “Broken Windows” Theory

Randall G. Shelden

Introduction: An Overview of “Broken Windows” Theory

The early 1990s in New York City greeted William Bratton as the new commissioner of the New York Police Department. Bratton centered his attention on the New York subways, and reinvigorated the subway police.  He was able to acquire new equipment and weapons for officers working below the streets.  In an all out war against fare evasion and the homeless people who live in the New York subway tunnels, he authorized sweeps of the subway trains and eventually claimed that the subway had been recaptured for citizens.

 His justification for such action stemmed from the broken windows theory which posits that more serious crimes evolve from minor infractions.  They stated the theory as follows:


If the first broken window in a building is not repaired, the people who like breaking windows will assume that no one cares about the building and more windows will be broken. Soon the building will have no windows. (Wilson and Kelling, 1982). 


Crime flourished because of lax law enforcement, the classic deterrence argument.  Among other assertions, they contended that if rude remarks by loitering youth were left unchallenged, such youth will be under the impression that no one cares and their behavior will likely escalate to more serious crimes. In other words, take minor problems and “nip them in the bud” before things get worse.  Who knows, maybe the rude teenager of today will be the “superpredator” of tomorrow.  This view ignores the fact that millions of teenagers engage in all sorts of "rude" and other obnoxious behavior in front of adults yet never become career criminals.  One result of such a view is the infamous policy known as “zero tolerance,” although supporters of this view have denied that broken windows is the same as zero tolerance (Kelling and Sousa, 2001).  However, some do assert that they are the same, including some inside New York’s police department, such as the NYPD Captain who stated that, “I think we’ve gone overboard with zero tolerance.  Cops are being held to unrealistic expectations to arrest people, and the result is that the community now sees us as occupiers, not problem solvers” (Wynn, 2001:107, emphasis added).  Note the emphasis on the word occupiers, for this describes one of the major historical roles of the police in class societies (Shelden, 2001).

Mayor Rudolph Giuliani came to power in New York City also during the 1990s.  The Giuliani-Bratton team moved the war on crime from the subways to the streets, instituting a policy of “broken windows.”  War had been declared on window washers (those unemployed individuals who washed windshields at city intersections – the so-called “squeegee men”), prostitutes, truant children (many who were truant because they had problems coping with inner city school violence), and homeless people living under bridges.  The people of New York rejoiced.  After all, it is unsettling to be employed and be required to contend with, or see, the dangerous population of homeless and unemployed people on a daily basis in the streets - it give capitalism a bad name (Parenti, 1999; Duneier, 1999).  One police Captain, who worked in the Times Square area, recalled Bratton telling police officers, “You’re not just hanging out.  Go find someone doin’ something.”  Such as “feel free to arrest someone for drinking a beer on the street?” this officer asked.  “It was more like, you better arrest somebody for drinking in public” (Wynn, 2001: 111).  This may be a slight exaggeration by this police Captain for, as will be noted later, police officers still had plenty of discretion and only arrested or gave a summons to those who drank in public where there were a lot of people (e.g., an entertainment area), but for those who were drinking outside their apartment, for example, the police merely told them to go indoors (Kelling and Sousa, 2001). 

The trouble with broken windows theory is that it is really no theory per se, for we have to ask the simple question: what is the theoretical basis for it? In other words, what is the rational for claiming that if we start arresting minor offenders like window washers, then serious crime will go down?  Another important question: "Why are windows in poor communities broken in the first place?" (Another reasonable question would be: “How many broken windows have been repaired since the policy was instituted?”)  The broken windows "theory" seems to be saying that the primary causes of crime are broken windows, or window washers, or homeless people sleeping in subways, etc. More specifically, the contention is that these minor “disorders” may lead to more serious kinds of crimes. Also, there is the assumption that those who engage in graffiti tagging, jumping turnstiles in the subways, or kids just hanging out are either criminals or future criminals.  In short, it “reduces the city’s quality of life and creates a culture that encourages more serious crime” (Erzen, 2001: 19).

How is “disorder” defined?  It is “behavior that violates widely accepted standards and norms of behavior, and about which a broad consensus exists, in spite of racial, ethnic and class differences (Kelling and Coles, 1996: 4, emphasis added).  Criminologist Wesley Skogan gets more specific, distinguishing between “social disorder” and “physical disorder.”  The former includes specific behaviors like “catcalling,” “drinking,” prostitution, sexual harassment, while the latter include things like “ill-kept buildings,” trash, alleys with rats in them.  Disorder is also distinguishable from more serious crimes like rape and murder by such minor offenses like vandalism, public urination, unsolicited washing of windows, and “aggressive panhandling” (Erzen, 2001: 21). 

Some of these claims seem absurd and perhaps should not even be taken seriously.  But we must take it seriously, as it has been embraced by "law and order" politicians and police chiefs all over the country.[1]  Now Bratton has taken his approach to Los Angeles, where he hopes to stem the tide of a recent increase in violence.

One may seriously question the claims about a “broad consensus” on “disorder,” both social and physical.  Including “catcalling” and “drinking” seem curious, as many of all ages engage in such behaviors (construction workers whistling at passing women, college students getting drunk and disorderly on fraternity or sorority row or during spring break).  As for prostitution, apparently supporters of the broken windows theory are complaining about street hookers, rather than more expensive call girls (in Las Vegas the distinction between these two are quite dramatic, for we have available for “high rollers” very expensive women, including those who advertise in both the Yellow Pages and in alternative newspapers as “escorts” ).  Why is one type of prostitution indicative of “social disorder” and not the other one?  Moreover, there appears to be a class bias operating, as behaviors in the inner cities seem to be preludes to more serious crime, but not similar behavior elsewhere.  (Given the huge toll white collar and corporate crime has on the public, why not focus on “social disorder” on Wall Street or inside the boardrooms on the top floors of corporate headquarters?)

Some of the proclamations by Bratton seem humorous, as when he began a policy of abandoning the use of “desk appearance tickets” (DATs), where people accused of committing minor crimes were issued tickets to appear in court on some future date.  Bratton stated his policy as follows:  “No more DATs.  If you peed in the street, you were going to jail.  We were going to fix the broken windows and prevent anyone from breaking them again.” (We wonder how many have been fixed.) Indeed, the NYPD was going to “reclaim the public spaces of New York,” quoting Bratton again (quoted in Eck and Maguire, 2000:225).

Another important aspect of broken windows theory is the almost complete dismissal of the various social causes of crime, so well documented by social science research during the past 100 years.  Many supporters of broken windows theory often display contempt for sociological theories of crime, with derisive statements, such as referring to such theories as “root causes” theories.  Kelling and Sousa (2001: 2, emphasis in the original) make the following point as they respond research that questions the value of certain police policies: 


These empirical findings became grist for the mill of new theorists who posited that crime was the result of collective “root causes” like social injustice, racism, and poverty.  The practical implication of such root-cause theory was that crime could only be prevented if society itself were radically changed.  These views became memorialized in President Lyndon Johnson’s Commission on Law Enforcement and Administration of Justice and became the virtual dogma of criminal justice thinking.


The tone of the above comments should be self-evident.  The contention that “root causes” became the “virtual dogma” of the thinking among academics seems exaggerated.  They further claimed that academic criminologists were “confounded” over the drop in crime during the 1990s, because “their root-cause theories had predicted the exact opposite.” What was this “exact opposite”?  They claim that criminological theories had predicted that the “echo baby boom” would cause an increase in crime (Kelling and Sousa, 2001: 3).  The only criminologist they cite was James Fox, but they could have cited John DiIulio as well.  Other than these two, few criminologists had gone on record with such wild claims.  Besides, this claim has been thoroughly discredited (Males, 1999).

At the conclusion of their analysis, Kelling and Sousa once again rather derisively mock the “root causes” approach, stating that “despite the root-causes theories that have dominated criminological, criminal justice, and much popular thinking about crime control, police can have a significant impact on crime levels in neighborhoods and communities” (Kelling and Sousa, 2001: 18).  From reading the works of Kelling and other supporters (e.g., James Q. Wilson), the causes of crime have little or nothing to do with various social and even psychological factors.  Crime, apparently, just suddenly springs forth out of nowhere.  Consistent with “rational choice” theory, they apparently believe that people simply “choose” to commit a crime because they think they can get away with it or perhaps because the police are not around. Those who support such a view apparently never bother to ask the most obvious question of why people would “choose” crime and not refrain from crime.

The Impact of Bratton’s Policies

The official crime rate did in fact drop significantly in New York City, with murder decreasing by 72 percent between 1990 and 1998, while total violent crime went down by 51 percent. Not surprisingly, there was rejoicing by conservatives all over the country as they claimed that the repressive measures they have consistently supported work and that we don’t have to be concerned with getting at the “root causes” of crime, like poverty and inequality (DiIulio, 1995).  More specifically, in New York City two main factors were alleged to have caused the drop in crime: (1) changes in police tactics and (2) greater use of incarceration.  In the case of the former, it was NYPD’s “strict crackdown on nuisance offenses, and its use of computer-drawn maps to reveal local hot spots and patterns of criminal activity…”  There was also a new “corporate-style devolution of responsibility by headquarters to precinct commanders” who were given the main responsibility to “show results” (Karmen, 2001).  They also began using a new form of crime mapping called “Compstat,” a computer system introduced in 1994 that provides data for each precinct on arrests, complaints and other information about crime.  It was claimed to have been part of the reason for the drop in crime in New York City.  This assertion will be discussed shortly.

In short, Giuliani and Bratton claimed to have cleaned up New York City, but where did they put the "rubbish" (the homeless and the unemployed)?  Did they simply disappear or vanish from the face of the earth, or should this be of concern to us?  During the Giuliani-Bratton war against the poor, little mention was given to policing insider trading violations, fraud, or other white collar crimes occurring in New York at the time (costing consumers billions of dollars each year).  Police patrol operations demonstrate what is commonly referred to as selective law enforcement, whereby administrative policies prioritize the types of laws to be enforced.  Selective law enforcement is also linked to police discretionary power and practices - who will be arrested and who will simply receive a warning.

It is now time to review some of the research that has been done on the crime drop in New York City and elsewhere, and specifically to assess the evidence concerning the broken windows theory. 

The Evidence

Before reviewing the research on broken windows, some more general comments are in order about the effectiveness of the police when it comes to crime.  A complete review of these studies is beyond the scope of this report.  A short version of the conclusions of all this research can be provided from quotes from two of the top researchers on the police, David Bayley and Lawrence Sherman.  Bayley stated simply that “The police do not prevent crime.”  He furthermore asserted that this “is one of the best kept secrets of modern life” (Bayley, 1994, quoted in Eck and Maguire, 2000: 207).  Sherman arrived at the same conclusion, although with some slight differences, noting that the hiring of more cops (always the solution offered by politicians and law enforcement officials, even though they know better), reacting arrests, unfocused random patrol “does not prevent serious crime.”  Sherman did note that certain kinds of “direct patrols, proactive arrests, and problem solving at high-crime ‘hot spots’ has shown substantial evidence of crime prevention” (Sherman, 1997, quoted in Ibid.).

After more than 30 years studying the crime problem, I have arrived at a similar conclusion, but with a slightly different take on the subject.  It seems to me that all the talk about various police strategies (and of the entire criminal justice system for that matter) is like talking about ways to prevent natural disasters, like tornadoes, hurricanes, earthquakes and the like.  It is almost as if crime is without specific causes and is a “natural event” like hurricanes.  It is almost as if there is nothing we can do to prevent it, so the next best thing is to develop procedures to respond as quickly and as efficiently as possible, like improving the delivery of emergency services to the victims of the afore-mentioned natural disasters.  This may be an exaggeration on my part, but if we simply “follow the money” (to use a famous line), it will tell us a lot, for it is an established fact that for every dollar on addressing some of the major causes of crime (e.g., inequality, poverty, joblessness, family problems), we spend at least ten times as much on the criminal justice system (more than $150 billion according to the latest figures; Shelden and Brown, 2003).  Asking about the success of the police and other components of the criminal justice system nevertheless is important, if for no other reason than to see “where our money went.”

Concerning the drop in crime in New York City, Andrew Karmen, who has studied this problem carefully, notes that the increase in the number of officers on patrol did not begin to rise until after 1995 – and well after the crime rate had already started its downward spiral (in the late 1980s).  Then the number of police officers began to decrease during the next several years, and yet homicides kept on decreasing.  Moreover, murders committed indoors, where the police can have little or no influence, also decreased throughout the nineties (Karmen, 2001).

Another problem has to do with “clearance rates” and the assertion that increases in this rate result in a lowering of crime. The clearance rate for murder did not improve notably in New York City during the 1990s (there was a “paper rise” in clearance rates, to 90%, but only because the police were allowed to count the arrests made in old cases dating back to the 1980s). Moreover, in New York City the motor vehicle theft rate went down by 70 percent during the 1990s, even though the clearance rate for this crime remained at around 7 percent, versus a national average of 14 percent (Karmen, 2001). ). Several studies point out that in other major cities, such as Washington, DC and Chicago, homicide rates also went down, with no change in their clearance rates (Main, 2000; Ferkenhoff and Possley, 2001; Karmen, 2001).  Even in New York City, one study noted that non-gun homicides declined between 1985 and 1986, while homicides with guns began declining in 1991, with an especially sharp decline in 1993, prior to Bratton’s changes (Fagan, Zimring, and Kin, 1997).  Incidentally, it should be noted that when comparing the national clearance rate in the early 1970s with the current rate, there has been virtually no change (Shelden and Brown, 2003: 41).

The idea that arresting people on minor offenses will lead to a decrease in major offenses is patent nonsense, contradicted by the official statistics often cited by supporters of this view.  A cursory review of any annual FBI report will show that the overwhelming majority of persons arrested have always been charged with minor offenses (the Part II offenses in the FBI report).  This is what the police have always done and no one ever claimed that making such arrests will lead to decreases in serious crime.  Further, in every major city over the past 100 years or so there have been periodic crackdowns on mostly minor offenses - vagrancy, prostitution, minor drug dealing, skid row drunks, etc. (Shelden and Brown, 2003).  In fact, even in New York City, the police made more arrests during 1986 and 1987 than in the 1990s, without causing a decline in overall crime (Karmen, 2001).

The real test of this view comes from looking at other cities that experienced declines in crime during the same time period, while employing other methods besides the broken windows and zero tolerance approach.  In fact, taking a longer view of crime in America, it should be noted that the crime rate in the year 2000 (4,124) was almost exactly the same as it was in 1971 (4,168).  There were ups and downs along the way, but essentially by the turn of the century it was sort of "back to the future" as far as crime is concerned.  It should be noted that during this same 30-year period, criminal justice expenditures went up by about 1,500 percent, while the imprisonment rate increased by about 500 percent.  The rate of violent crime during this time went up from 396 to 506 (Shelden and Brown, 2003: 35).  During this period of time, there were numerous law enforcement policies instituted in virtually every major city, with plenty of variations of "broken windows" and "zero tolerance."  Yet no real change in the overall crime rate.  How can "success" be claimed with such overwhelming evidence to the contrary?  (It should be noted that during the last couple of years the rate has either remained the same or risen slightly in most cities.)

More specific to so-called “quality-of-life” enforcement – a big part of the “broken windows” approach – is that empirical research has leveled a serious challenge to this view.  General, this research finds little evidence in support of the broken windows strategy.  One critic charged that any success can be attributed to increased surveillance that is part of this strategy, but is part of any strategy where there is an increase in the concentration of officers in certain areas (Harcourt, 1998).  Indeed, if there is a cop standing on a street corner, it is highly unlikely someone will come along and urinate in front of him!  (But if he has to go bad enough, he will look for a place where there are no cops.) 

Another critic (a former police chief) said that Bratton was merely lucky to have come along after the crime rate in New York City was already decreasing (Bouza, 1997, quoted in Eck and Maguire, 2000: 226). Still another critic called this approach “harassment policing” (Panzarella, 1998, quoted in Eck and Maguire, 2000: 226).  Numerous reports have noted the most obvious result of Bratton’s style in that the targets have been overwhelmingly racial minorities, with the resulting increase in civil law suits (we should mention, at least in passing, the horrific death of Amadou Diallo, who was shot 41 times because he acted “suspiciously” when in fact he was completely innocent of any wrong-doing).  Other critics warn that even with a short-term drop in crime, there is the potential of a long-term decline in respect by community members and the legitimacy of the police in general, especially in minority communities (Goldstein, 1990; Panzarella, 1998).  Perhaps even more important, with so many arrests on minor charges, more and more people (especially young minority males) will have a record, thus hindering their job prospects in the future and perhaps even propelling them into more crime, especially drug crimes (Sherman, 1997).

The literature review by Eck and Maguire cause them to arrive at the following conclusion:  “In general, there is little evidence that generic changes in policing are responsible for reducing violent crime.  There is greater evidence for focused policing strategies contributing to the drop in violent crime.” They further note that two well-publicized strategies, zero tolerance and the use of Compstat, “are the least plausible candidates for contributing to the reduction in violent crime” (2000: 245).

A recent analysis by Kelling and Sousa needs to be discussed, for they bring forth data that seem, at least on the surface, to demonstrate that broken windows policies were the major reason that serious crime declined in New York City during the 1990s.  The dependent variable they used was reported violent crimes, while the independent variables were misdemeanor arrests and the number of young males in the population, hospital discharges for cocaine-related episodes and the number of unemployed.[2]  Rather than comparing different cities with New York, they took all 75 precincts within the city and compared them. Through statistical analysis they found that the strongest correlation was between misdemeanor arrests and violent crime, although unemployment was also statistically related. 

Using misdemeanor arrests as a proxy for broken windows enforcement, they concluded as follows:


The average NYPD precinct during the ten-year period studied could expect to suffer one less violent crime for approximately every 28 additional misdemeanor arrests made (Kelling and Sousa, 2001:9).


They follow this statement with yet another negative “root cause” statement:  “Contrary to ‘root cause’ advocates, overall declines in violent crime cannot be attributed to our measure of the economy, changes in drug use patterns, or demographics” (Ibid.).

            Closer examination of their data and methods reveals some important problems.  First, the measurement of crime, always an inexact science, relies solely upon official record keeping.  Second, crime is measured according to misdemeanor arrests and violent (and other index crimes) reported to the police.  As every beginning criminology student knows, in the annual FBI report, crime is divided into “Index” and “Part II” offenses.  The FBI report shows “crimes known to the police” - meaning the index crimes that citizens report to the police – and “persons arrested” – meaning the number of arrests for both Index and Part II offenses.  Therefore, in New York City, during the 1990s, the total amount of crime actually increased, since there are far more Part II offenses than Index crimes.

            A third point is that Kelling and Sousa note that, according to their field observations, out of almost 250 police-citizen encounters they witnessed, only about half involved a minor offense, and less than 70 resulted in an arrest or summons (2001: 18).  In other words, about one-fourth of all police-citizen encounters resulted in an arrest.  Could it be possible that instead of misdemeanor arrests being the cause of the drop in violent crime, an increase in police-citizen encounters was even more highly correlated?[3]

            Fourth, we still don’t know what the connection is between specific types of disorders and violent crime.  Why should a concentration on public urination or public drinking prevent violent crimes?  Where is the logical connection?

            Fifth, if broken windows policies lead to a drop in serious (especially violent) crime, why did serious crime drop in every major city in the country, where policies other than broken windows were used?  If broken windows is more strongly related to crime than the “root causes” then why does it not apply in other cities?  Supporters like Kelling could argue that broken windows policies applies to New York City but nowhere else, but this would hardly be a “theory” of crime, nor a serious rebuttal to sociological theories of crime.

            Finally, for the country as a whole, there were several “social indicators” (which indicate the overall health and well-being of the country) that showed improvement.  These include infant mortality, the number of low birthrate infants, high school dropouts, unemployment, child poverty, youth suicide and teenage births (Miringoff and Miringoff, 1999).  Some or all of these could be causative factors for the decrease in crime, in New York City and elsewhere.



Bayley, D. H.  1994.  Police for the Future.  New York: Oxford University Press.


Blumstein, A. and J. Wallman (eds.). 2000.  The Crime Drop in America.  New York: Cambridge

University Press.   


Bouza, A.  1997.  “NYPD Blues: Good, Lucky, or Both?” Law Enforcement News, January 31, 8-10.


Butterfield, F. 2000.  “Cities reduce crime and conflict without New York-style hardball.”  New York

Times, March 4, pp. A1, B4.


Cohen, T., N. Kauder, and B. Ostrom, 2000. “A renewed interest in low-level crime.”  Caseload

Highlights 6: 1-5.


DiIulio, J. J.  1995.  “Arresting Ideas: Tougher Law Enforcement is Driving Down UrbanCrime.”

Policy Review 74: 12-16.


Duneier, M.  Sidewalk. 1999. New York: Farrar, Straus and Giroux.


Eck, J. and E. Maguire. 2000.  “Have Changes in Policing Reduced Violent Crime.”  In Blumstein

and Wallman, The Crime Drop in America.


Fagan, J., F. E. Zimring, and J. Kim. 1998 “Declining Homicide in New York City: A

Tale of Two Trends.”  National Institute of Justice Journal 237: 12-13.


Ferkenhoff, F. and M. Possley. 2001.  “As murders drop, fewer cases solved.”  Chicago Tribune,

April 12, pp. A1, A5.


Goldstein, H.  1990.  Problem-Oriented Policing.  New York: McGraw-Hill.


Greene, J.  1999.  “Zero Tolerance: A Case Study of Police Policies and Practice sin New York

City.”  Crime and Delinquency 52: 171-197.


Harcourt, B. E. 1998.  “Reflecting on the Subject: A Critique of the Social Influence Conception of

Deterrence, the Broken Windows Theory, and Order Maintenance Policing New York Style.”  Michigan Law Review 97: 291-389.


Joanes, A.  2000.  “Does the New York City police department deserve credit for the decline in New

York City’s homicide rates?: A Cross-city comparison of policing strategies and homicide rates.”  Columbia Journal of Law and Social Justice 33: 265-303.


Karmen, A. 2001. “Smarter Policing and Stepped-Up Imprisonment as the Primary Causes of

Falling Crime Rates in New York City: The Emergence of an Urban Legend?”  Justice Policy Journal 1, 1 (on-line:


Main, F. 2000.  “Cops catching fewer killers. Chicago Sun Times, December 21, p. A8.


Males, M.  1999.  Framing Youth.  Monroe, ME: Common Courage Press.


Miringoff, M. and M. L. Miringoff.  1999.  The Social Health of the Nation. New York: Oxford

University Press.


Panzarella, R.  1998.  “Bratton Reinvents ‘Harassment Model’ of Policing.”  Law Enforcement

News, June 15-30, 13-15.


Parenti, C. 1999. Lockdown America.  New York: Verso.


Shelden, R. G. and W. B. Brown. 2003.  Criminal Justice in America: A Critical View.  Boston:

Allyn and Bacon.


Sherman, L.  1997.  “Policing for Crime Prevention.”  In Preventing Crime: What Works, What

Doesn’t, What’s Promising – A Report to the Attorney General of the United States.  Washington, DC: U.S. Department of Justice.


Wilson, J. Q. and G. E. Kelling. 1982.  “Broken Windows: The Police and Neighborhood Safety.” 

Atlantic Monthly, March.




[1]   It is curious that whenever there is a drop in crime in a given location, the police tend to take credit for it because of some new policy, but when the crime goes up they place the blame elsewhere.


[2]  It should be noted that there are other ways of measuring the extent of drug abuse and drug use, namely through the numerous annual surveys conducted in the United States.  According to one survey, Monitoring the Future (University of Michigan), drug use among 12th graders began to decline in the late 1970s, but began to rise by the early 1990s, while drug use among 8th and 10th grade students rose during the 1990s (Miringoff and Miringoff, 1999: 118-120).


[3]   In this particular study Kelling and Sousa give some illustrations of the application of broken windows by officers in the field, based upon ride-along observations.  They gave four examples, two of which had to do with public urination.  The other two were of drinking in public.  In one case of urination, they issued a summons to two white males who they observed urinating in front of a dumpster.  They issued the summons because of the fact that they committed this act in front of the dumpster, rather than behind it out of public view.  One officer stated afterwards that “The dumpster was right there, if he had gone behind it that would have at least been something---instead he decides to do it in full view of everybody” (Kelling and Souse, 2001: 17, emphasis added).  The other case of public urination we done by two black males who drove down a secluded alley and did their urination out of public view.  They were stopped, asked for ID and told to move on after a verbal warning.  One might reasonably wonder if this represents the state of modern-day police work, where uniformed officers, under the guise of preventing serious crime, stop people who are peeing in public! 


© All rights reserved. No part of this may be reproduced without permission from the author.