The City that Became Safe:

New York and the Future of Crime Control

Franklin E. Zimring

February 3, 2011

Word count: 3,992 (including tables, captions and figures)

 Over the 20 years after 1990, New Yorkers were the

beneficiaries of the largest and longest sustained drop in

street crime ever experienced by a big city in the

developed world. In less than a generation, rates of

several common crimes that inspire public fear – homicide,

stranger assault, robbery and burglary – have dropped by

80% or more. By 2008, the homicide rate in New York was

lower than it had been in 1961 despite changes in the

racial and ethnic profile of the city that normally

increase risks. Figure 1 shows the recent crime rates in

New York City in an unconventional fashion for crime

statistics – by expressing the crime rate in 2009 as a

percentage of its 1990 level.

2
3
 The risk of being killed in New York City was less

than a fifth of its 1990 rate in 2009, the risk of being

robbed was less than 1/6th the former rate and the risk of

car theft declined to 1/16th of its 1990 level.

The arithmetic of the decline is amazing but the

recent public reaction to New York’s epidemic of public

safety is rather complacent. In one sense, this is

understandable – crime statistics only command attention

when citizens are worried about crime just as people only

think about dentists when their teeth hurt.

But New York’s crime adventures demand scientific and

public attention for three reasons. First, this is a

Guinness Book of World Record’s crime drop. The size and

the length of the city’s crime decline have never been

documented in a major metropolis before. Indeed, most

professional observers would have doubted that a major city

could reduce safety crime by 80%. So New York City has

witnessed what many of us would have thought impossible for

any city.

The second noteworthy lesson about New York’s

experience is that the systematic changes which have

4
contributed a major part of the city’s crime decline are

not extremely expensive and can be adapted to conditions in

other cities. New York shows what is possible in many

other cities.

The third surprise from New York is that the city made

giant strides toward solving its crime problem without

either winning its war on illicit drug use or massive

increases in incarceration. So the great success in this

city is a challenge to the two dominant assumptions of

crime control policy in modern America.

The Crime Decline

The first nine years of decline in New York City was

part of a much broader decline in crime during the 1990s in

the United States, a close to 40% drop in crime that

started in the early 1990s and ended in 2000. This

nationwide crime decline was the longest and largest drop

in modern history. What sets New York City apart from this

general pattern is that its decline was twice as large as

the national trend and lasted twice as long.

5
 The extraordinary difference between a 40% and 80%

drop in safety crime can be seen in comparing homicide

rates in the five largest cities in the United States in

1990 and in 2009.

6
7
 The great crime decline of the 1990s reduced homicide

in all five cities and in four of those five by a

substantial amount. But New York City went from being dead

center in its homicide rates in 1990 to being by the lowest

of the major population centers – more than 30% below the

next best city and only 40% of the mean rate for the other

four places.

Even after its 80% drop, the homicide rate in New York

City is much higher than in most major European cities and

six times that of Tokyo. But rates of many property

felonies in New York City are significantly lower than in

London and Sydney, Australia, as shown in Table 1.

8
Table 1. Crime Rates in New York City, London, Sydney, and Tokyo, 2007 (Rate per 100,000).

Homicide Rape Robbery Burglary Auto Theft

New York 6 10.6 265 254 161
London 2.2 30.7 610 1,290 501
Sydney (2006) 1.5 51.4 159 1,008 461
Tokyo 1 1.8 4.7 137 12
Toronto 2 N.A. 133 362 279

9
 Homicide in the contemporary New York City is much

higher than in the comparison sites, but rates of forcible

rape are lower in New York than in Sydney and London; rates

of New York robbery are lower than in London. Burglary and

auto theft are lower in New York City by far than in the

other major Western cities in the table.

Too Good To Be True?

Because official crime statistics are generated and

verified by the same police departments that are praised

when crime rates fall and blamed when they increase, many

observers greet very good news from police statistics as

untrustworthy pronouncements by an interested party (The

New York Times, September 9, 2010). Since police have a

special interest in lowering rates, how can we be confident

that the numbers behind Figure 1’s spectacular good news

reflect the reality of street crime?

The best method to verify trends is with independent

data. Two of New York’s seven index crimes are also

counted by agencies that are independent of the police:

county health departments keep meticulous records of all

deaths and provide specific reports of what the police

10
classify as murder and non-negligent homicide. Over the 19

years when the police reported the 82% decline in Figure 1,

the agreement between the health and police reports each

year was almost perfect (a correlation of annual rate

trends of .999). For auto theft – the crime rate with the

spectacular 94% drop – insurance claims by victims are

recorded by insurance companies. I obtained theft loss

reports by year from two separate industry data bureaus.

The most complete city-wide claim loss statistics from

insurance projected a decline in theft rates (accounting

for population change) of just over 90%. So the biggest

decrease (auto theft) and the most important decrease

(homicide) are both confirmed by independent sources. And

there is confirmation as well for the big drop in robbery,

because the number of killings from robberies dropped more

than the 84% decline in non-fatal robberies.

By American standards, New York City has become a safe

and low crime urban environment. How did this happen? Can

other states and cities match the New York achievement?

Explaining the New York Difference

11
 About half of the New York City crime drop was part of

a larger national downturn in the 1990s that did not have

any distinctive local causes. The same mix of increased

incarceration, 1990s prosperity, an aging population and

mysterious cyclical influences that created the national

pattern operated as well in New York City. The decline was

not easy to tie to specific causes at the national level

(Zimring 2007) and cannot be connected any better in New

York City.

But almost half the New York City decline (and all of

the city’s decrease after 2000) is a distinctive local

phenomenon. What local conditions generated this extra 40%

crime decline that transformed the happy windfall of the

national decline into an unprecedented public safety

achievement? For much of this portion of the city’s

decline, plausible answers are easier to obtain.

There were very few drastic changes in population,

economy, schools or housing in the city during the 20 years

after 1990. One of the four major boroughs, Manhattan, had

a drop in ethnic minority population but the other three

boroughs had increases – and crime went way down in all

four. The percentage of the population in the arrest prone

12
15-29 age bracket declined at the same rate as nationally,

and economic growth did not reduce either poverty or

unemployment in New York City below the national average

during the first 18 years of the crime decline.

13
14
15
 More remarkable, the rate at which the city filled its

jails and the state’s prisons was well below the national

pattern all through the two decades after 1990. Figure 5

shows combined trends in persons imprisoned from the city

with those held in city jails.

16
17
 While the U.S. prison and jail population expanded by

half during the declining crime 1990s, New York’s

incarceration rose only 15% in the first seven years of the

decade and was under its 1990 rate by 2000. By 2007, the

rate of incarceration was down 28% in New York City. So

what had been regarded as high-risk populations in three of

the large boroughs of New York City did not change or

increased and the number of offenders locked up did not

increase while robbery, burglary and murder dropped by over

80%. In a nation where we have long believed that criminal

offenders will keep offending unless they are locked up,

the first mystery from New York City is this: Where have

all the criminals gone?

There is a second respect in which the momentous drop

in New York street crime is a policy surprise. Figure 6

shows trends in three measures of illegal drug use over

time, the New York hospital discharges from drug treatment,

drug overdose deaths and the percent of Manhattan men

arrested who test positive for an illegal drug.

18
19
 These indirect measures of the amount of illegal drugs

purchased and used in the city have not dropped with crime

rates. While the cocaine using population became older in

New York (and elsewhere) and the location and visibility of

drug sales may have changed, the overall illicit drug

volume appears relatively stable. So New York City appears

to be winning the War on Crime without winning the War on

Drugs.

Estimating Police Effects

The one aspect of crime policy where New York made big

changes was policing, and these are the only obvious

candidates to take credit for the city’s larger crime

decline. Beginning in 1990, the city added 9,000 new

uniformed police, changed its management of police efforts

and made its street policing tactics both much more

aggressive and more focused on high crime settings.

Without doubt, this combination of changes made a major

contribution to reducing street crime in the city. But how

much of the drop did the package of police changes cause

and which changes should get most of the credit?

20
 The best estimate of the total policing effect is the

difference between the New York City decline and the median

decline in other top-10 U.S. cities. To avoid giving the

police changes credit for extra crime reductions in

Manhattan from population changes, the median non-Manhattan

crime reduction from the other major boroughs is the

measure of the New York crime drop. Figure 7 shows the

“policing only” difference in crime rates for all index

crimes.

21
22
 There is some New York City advantage for all reported

index crimes, but very significant differences by type of

offense. The three classic street crimes — robbery,

burglary (access from street) and auto theft — show very

large policing effects ranging from 21% to 32% of total

1990 offense levels. Larceny and assault show very small

New York advantages of 2% and 4%--not a strong indication

of any police effect. Homicide and rape are in the middle,

with much more policing impact than larceny and assault but

much less than burglary, robbery and auto theft. This

differential pattern in Figure 7 is quite consistent with

the police having more impact on city streets and public

areas.

The sheer weight of new police on the street was

originally thought to have caused most of the distinctive

New York decline. The rate at which reported crime dropped

was higher during the 1990s when the police force was

growing quickly than after 2000 when the force was reduced

by more than 4,000 uniformed officers.

The problem with relying on the total size of the

1990s decline, however, is that was the period when crime

was declining everywhere. Figure 8 provides a second

23
method of exploring the impact of manpower changes on city

crime, by comparing the relative performance of New York

against the median city crime trend of the other nine of

the ten largest cities for average decline in the seven

“index” offenses.

24
25
 Using this strategy, what economists call a

“difference-in-difference” approach, New York City shows

the same advantage in crime during the period of 50%

expansion (21%) as during the period when police manpower

dropped by 19% (21%). Does that mean that police numbers

are irrelevant? Probably not. Even the most recent period

has substantially more police on the street than other

cities. And the cumulative effects of increased manpower

may have lasted into the decade when force levels went

down. This kind of long-term cumulative effect would be

novel in the study of policing. But the high magnitude of

effects one derives from any research strategy that

attributes the city’s distinctive drop to policing is also

unprecedented.

Which Tactics Worked?

It is difficult to determine which of the many changes

in policing should get the major credit for the policing

impacts and impossible to provide quantitative estimates

for any particular strategy.

Two of New York’s new tactics almost certainly

contributed to the crime reduction in the city. The

26
department emphasis on “hot spots” for special enforcement

was proven to work in other cities and well implemented in

New York. And the campaign against public drug markets

produced drug specific indicators of effectiveness. Three

other changes – including two of the major shifts – were

probably successful: the management and data mapping

COMPSTAT system, the expansion of police manpower and the

emphasis on guns. The case for COMPSTAT and manpower

effects is the large size of the first decade’s drop as

well as the sustained impact in the second decade. But

specific effects in New York City can’t be proven. Gun

programs also worked in Chicago, but the New York version

was not a replication of the Chicago method.

The biggest and most costly police change with an

unknown impact is the aggressive program of street stops

and misdemeanor arrests that the police use in almost every

patrol operation. There were more than half a million

stops in 2009 and two hundred thousand misdemeanor arrests.

The police believe these tactics increase the effectiveness

of preventive interventions but there is no rigorous

evidence on the value added in New York City. The contrast

between aggressive stops and “hot spots” is the clear

evidence that hot spots work in other cities with

27
scientific evaluation. What separates the street stops

verdict of “not proven” from COMPSTAT which I label a

probable success is that COMPSTAT’s lack of effectiveness

data from elsewhere is a result of its novelty. Aggressive

patrol has a history almost as long as that of street

policing, so the absence of a strong track record from

other cities is more problematic. Aggressiveness may be a

greater success in New York than in other places but this

possibility must be the subject of careful local testing.

There are also a number of labels used to describe

styles of policing that probably don’t describe what

happened in New York City. “Zero tolerance” has always

been a slogan without definition in discourse about

policing. Police in New York did increase arrests for

minor offenses but only where this was believed to prevent

more serious crime. New York never increased arrests for

prostitution and was not consistent over time in its

enforcement of gambling or vice crimes. And “broken

windows” policing as originally designed would have avoided

pouring resources in the highest crime rate areas that

became a central priority in New York’s “hot spots”

strategy. So the New York strategy did not fit the “broken

windows” prescription either.

28
 Figure 9 summarizes current knowledge on eight

different types of police policy in New York.

29
Figure 9. What Worked in New York City Policing?

Effective Hot Spots

Destruction of Public Drug

Markets
Probably Effective Increased Manpower

COMPSTAT Management and

Mapping

Gun Programs
Not Known Aggressive Arrests and Stops
Not Implemented in New York “Zero Tolerance”

“Broken Windows”

30
 The next important steps in policy analysis are to

rigorously field test the impact of additional manpower and

some of the specific field techniques that have been part

of the kitchen sink full of changes the New York police put

in place. Then there should be trial and error adaptations

to other urban settings.

Five Lessons

But even this early in the game, there are already

five lessons from New York City that should have a

significant influence on crime policy.

1. Police Matter.

For at least a generation, the conventional wisdom in

American criminal justice doubted the capacity of urban

police to make a significant or sustained impact on the

crime rates in cities. Powerful circumstantial evidence

from New York City (and preliminary confirmation from Los

Angeles) has already altered informed opinion. The details

on cost-effectiveness and best tactics have yet to be

established, but investments in policing apparently carry

at least as much promise as other branches of crime control

in the United States.

31
2. The Irrelevance of Increased Incapacitation.

The supply-side policy tactics that have dominated

crime control in the United States assume that high risk

youth will become criminal offenders no matter what we do

and that criminal offenders will continue to commit high

rates of crime unless they are locked away. These “supply-

side” theories are the central justification of the

sevenfold expansion of imprisonment in the United States,

and supply-side proponents were warning in the mid-1990s

that cities like New York with high numbers of minority

youth growing up in single-parent families would require

massive new investments in prisons and juvenile facilities.

From this perspective, the New York Story is a game-

changing example of effective crime control without

additional incarceration. Police still catch criminals in

New York City and prosecutors and judges send them to jail.

But the city has reduced its most serious crime rate by 80%

without any net increase in secure confinement. The

central assumption of supply-side crime control has been

disproved because crime has been reduced without increasing

incarceration.

32
 If three or four fewer prisoners are the equivalent of

one more uniformed police officer then 27,000 to 36,000

fewer prisoners could buy the peak period expansion of the

New York City police department, and 18,000 fewer prisoners

could finance the expansion of the police from the 1990

level to the current rate per 100,000. So the difference

between New York City’s incarceration trends and those of

the rest of the nation have more than paid for the city’s

expanded police force.

3. Crime Control without Drug Control.

New York City has been the illicit drug use capital of

North America for at least seven decades. By all accounts

it continues to be. The widespread introduction of crack

cocaine to the city in the 1980s was associated with sharp

increases in homicide. The belief by the public and

government officials that illegal drug sales and use was

closely linked to urban violence was one of the animating

theories for a drug war declared in the United States in

the decade after 1985. From the late 1980s perspective, a

major reduction in violence without massive reductions in

the sale and use of illegal drugs would be an impossible

dream. But that’s exactly what seems to have happened

since 1990 in New York City.

33
 Drug use appears to have stayed stable in New York

City while crime rates dropped, whether the drug use

indicator is overdose deaths, or hospital discharges for

drug treatment or urine tests of criminal suspects.

Indicators for all drugs other than cocaine stay stable for

the entire period. For cocaine, drug use for the young

does drop in the late 1990s, but the consistent pattern for

older users shows that the change in age profile (a

national trend) was not associated with cocaine becoming

scarce or more expensive. Meanwhile, drug-related killings

drop by 90% from peak rates over time.

4. The Poor Pay More and Get More from Crime Control

Both the costs and benefits of New York’s crime

control are spread unevenly over the city’s neighborhoods

and populations. The street stops, bullying and pretextual

arrests that are the common tactics of aggressive street

policing fall disproportionately on young men of color, in

their own neighborhoods and in other parts of the city when

they dare. So police aggressiveness is a very regressive

tax. But the benefits of reduced crime also

disproportionately favor the poor, ironically the same

young males with dark skins who suffer most from police

34
aggression. About 65% of the reduction in homicides in New

York were Black and Hispanic men between ages 15 and 44, a

group which comprised only 13% of the population. The good

news is that lower crime was the one public benefit of the

last 20 years which had its greatest impact on the people

who needed help the most. The bad news is that the

physical and dignitary harms of police aggression also fall

the hardest on those who can least afford more trouble from

the government.

5. The Inessentiality of Urban Crime

The contrast between the modest changes in the

conditions of urban life in New York and the game-changing

changes in its crime rates establishes one further point

about crime in modern cities that applies to most cities

and has nothing to do with determining the particular

causes of New York’s good news. What we know in 2010 is

that most of the homicide and muggings in modern big cities

isn’t hardwired into the populations, cultures and

institutions of these cities. We know this because serious

crimes dropped by 80% without any of these conditions being

transformed. So we don’t have to change either the nature

of cities or of urban populations to create homicide rates

below six per 100,000 compared to New York’s 30 per 100,000

35
in 1990. Of course even lower rates are desirable, but

finding that 80% of the serious crime rates prevalent 20

years ago are not a necessary part of city life is beyond

good news.

Since most big cities in the United States have

already benefited from crime drops that were part of the

1990s New York decline, the additional declines in rates of

safety crime that they can expect from catching up to New

York will be closer to 40% than to 80%, and the shifts in

policing and environmental controls that will best serve

Baltimore or Cleveland may be different from the New York

City portfolio. But modest changes can produce big effects

in all cities. And if New York’s rates drop further after

2009, this will also indicate that other cities should aim

for greater reductions.

Two decades of New York experience after 1990 doesn’t

merely suggest that there is no deep structural linkage

between urban cultural and social conditions and very high

rates of violent crime, it proves the point. The steady,

significant and cumulatively overwhelming crime decline in

New York is proof positive that cities as we know them need

not be incubators of robbery, rape and mayhem. High rates

36
of serious crime simply are not an essential part of urban

living. If traditionally high rates of killings and

robbery were essential to the fabric of American urban

life, the 80% drop in New York that did happen couldn’t

happen. That is a fundamental surprise to many students of

the American city, and the most hopeful insight of

criminological science in a century.

37
Further Reading

Blumstein, Alfred and Joel Wallman, eds. The Crime Drop in

America. New York: Cambridge University Press (2000).

Karman, Andrew. New York Murder Mystery: The True Story

behind the Crime Crash of the 1990s. New York: New York

University Press (2000).

Skogan, Wesley and Kathleen Frydl, eds. Fairness and

Effectiveness in Policing: The Evidence. Report of the

National Research Council Committee to Review Research on

Police Policy and Practices. Washington, DC: The National

Academies Press (2004).

Zimring, Franklin E. The Great American Crime Decline. New

York: Oxford University Press (2007).

Zimring Franklin E. The City That Became Safe: New York and

the Future of Crime Control. (Forthcoming, 2011).

38