Criminal Justice

iResearchNet




Criminal Justice > Criminology > History of Criminology > The Crime Drop of the 1990s

The Crime Drop of the 1990s




The crime drop of the 1990s—one of the most dramatic and unexpected declines in violent crime in recorded American history—reduced homicide and violent crime by roughly one-third over a decade, generating competing explanations that have reshaped criminological theory and crime prevention policy ever since.

This article examines the magnitude and character of the 1990s crime decline, the principal competing explanations advanced by criminologists and economists, the specific evidence bearing on each explanation, the international dimensions of the crime drop and what cross-national variation reveals about causal mechanisms, and the post-2000 trajectory of American crime rates within the broader History of Criminology. The argument advanced here is that the 1990s crime drop was almost certainly over-determined—produced by the convergence of multiple causal factors rather than any single mechanism—and that its primary theoretical significance lies not in identifying which single cause was responsible—the evidence increasingly suggests no single cause was dominant—but in what the evidence for multiple simultaneously operating contributing causes reveals about the underlying complexity of crime rate determination, the systemic interdependence of seemingly disparate social conditions, and the limits of single-factor criminological theories as guides to explanation or prevention.

Introduction

American violent crime peaked in the early 1990s following two decades of elevated rates and then declined with a consistency and magnitude that surprised virtually every criminologist who had studied the period of rising crime. From 1991 to 2001, the homicide rate fell by approximately 43 percent, robbery by 44 percent, and aggravated assault by 28 percent. These were not modest statistical fluctuations but dramatic sustained declines across virtually all major crime categories, all demographic groups, all regions, and most American cities—declines that continued largely unabated through the first decade of the twenty-first century before leveling off and eventually reversing in some categories.




The crime drop was not simply a return to some pre-existing baseline. It represented a real improvement in the security and wellbeing of American communities, particularly the disadvantaged urban communities that had borne the greatest burden of the crime increase that preceded it. Fewer people were killed, robbed, raped, and assaulted than would have been the case had the early 1990s rates been sustained, and the communities most affected by violent crime—predominantly Black and Latino neighborhoods in major cities—experienced some of the most dramatic improvements. The practical significance of the crime drop was enormous: it affected many millions of lives, transformed the lived experience of urban life across the country, and influenced policing strategy, urban development, and political discourse for an entire generation.

The theoretical significance of the crime drop was equally substantial. Its timing—beginning in the early 1990s and continuing through the 2000s—was inconsistent with several popular explanatory frameworks. The demographic explanation predicting that the aging of the baby boom generation would reduce crime had anticipated crime declines beginning in the 1980s, not the 1990s. The incarceration explanation predicting that the dramatic expansion of imprisonment would reduce crime could not account for the fact that incarceration had been rising for two decades before crime began to fall. And the policing explanation could not easily account for crime declines in cities with very different policing strategies or for simultaneous declines in countries with very different criminal justice systems.

The crime drop also challenged criminological theory’s capacity to explain not only why crime is distributed as it is at any given moment but why it changes over time. Most criminological theories—social disorganization, strain, social learning, social control—are essentially static theories that explain the distribution of crime across social groups and contexts rather than dynamic theories that explain changes in aggregate crime rates over time. The discipline had relatively few analytical tools for understanding why a country’s or city’s crime rate might increase or decrease substantially within a decade, and the crime drop exposed this gap in criminological theory in ways that have subsequently motivated significant theoretical development and methodological innovation in longitudinal analysis of crime trends.

Magnitude and Character of the Crime Decline

The Statistics and Their Interpretation

The magnitude of the 1990s crime drop is well-established across multiple data sources—Uniform Crime Reports, National Crime Victimization Survey, hospital emergency room data, and mortality statistics—that use different measurement approaches and are subject to different biases. The convergence of these independent data sources on similar estimated magnitudes of decline provides confidence that the crime drop was a genuine behavioral phenomenon rather than a statistical artifact of changes in reporting rates or recording practices.

The UCR data, subject to potential manipulation through changes in police recording practices, showed homicide declines of 43 percent from 1991 to 2001. The NCVS, which measures victimization independently of police reporting, showed violent crime declines of approximately 50 percent over the same period. Hospital emergency room data and mortality statistics, which are independent of both police recording and victim survey methodology, confirmed the homicide decline with comparable magnitude. The convergence of these independent measures establishes that the crime decline was not primarily an artifact of changed measurement but reflected genuine reductions in violent victimization that Americans experienced in their daily lives.

The spatial distribution of the decline was not uniform but concentrated in the cities and neighborhoods that had experienced the most dramatic crime increases during the crack cocaine era of the late 1980s and early 1990s. New York City’s crime decline was among the most dramatic—homicide fell by 72 percent from 1990 to 2000—but comparable declines occurred in Boston, Los Angeles, Dallas, and dozens of other cities where the crack market had driven violence to historically unprecedented levels. This concentration of the decline in the cities most affected by the crack epidemic suggested that the resolution of the crack market crisis was a major contributor to the crime drop, even if it was not the only one.

The Timing and Its Theoretical Significance

The timing of the crime drop’s beginning—in the early 1990s, well into the period of the crack cocaine crisis rather than before or immediately after—provides important evidence about its causal mechanisms. The crack cocaine epidemic, which had driven sharp increases in homicide particularly among young Black men in major cities from the mid-1980s through the early 1990s, appears to have peaked in the early 1990s and then begun to decline as the crack market matured and the violence associated with market establishment and competition diminished. The timing of the crime drop’s beginning aligns closely with the apparent exhaustion of the crack epidemic’s most violent phase, suggesting that the epidemic’s resolution was at minimum a contributing factor.

The timing also argues against several popular explanations. The dramatic expansion of American incarceration began in the late 1970s and had been proceeding for more than a decade before crime began to fall—making it difficult to attribute the timing of the decline specifically to incarceration expansion that was already well underway. The community policing and problem-oriented policing strategies that received most credit from police departments and their political allies in explaining the crime drop were largely implemented in the early 1990s, coincident with the decline’s beginning rather than preceding it—making it difficult to establish that the policing innovations caused the decline rather than coinciding with it.

Table 1. Principal Explanations for the 1990s Crime Drop: Mechanism, Evidence Strength, and Key Research

Explanation Causal Mechanism Evidence Strength Key Researchers
Incarceration expansion Incapacitation of active offenders; deterrence Moderate — accounts for 25-30% of decline per some estimates Spelman (2000); Levitt (2004)
Crack cocaine epidemic decline Resolution of crack market violence; generational deterrence Strong for homicide timing and magnitude Blumstein & Rosenfeld (1998); Zimring (2012)
Police strategies Hot spots policing; problem-oriented policing; CompStat Moderate for specific cities; difficult to generalize Eck & Maguire (2000); MacDonald (2002)
Lead hypothesis Reduced childhood lead exposure reducing adult impulsivity Strong correlational; causal mechanism plausible Nevin (2007); Reyes (2007)
Legalized abortion Reduced cohort of high-risk births from 1970s onward Contested; methodological debate ongoing Donohue & Levitt (2001, 2020)
Economic factors Improved labor markets reducing economic strain Weak — crime fell in 1990s recession and continued during 2000s expansion Raphael & Winter-Ebmer (2001)
Demographic change Aging of baby boom reducing proportion of crime-prone young men Weak for timing — demographic shift preceded decline by decade Levitt (2004)

Competing Explanations and the Criminological Debate

The Incarceration Hypothesis

The incarceration hypothesis—that the dramatic expansion of American imprisonment from approximately 200,000 prisoners in 1972 to over 1.5 million by 2000 substantially reduced crime through incapacitation of active offenders and deterrence of potential offenders—received serious scholarly attention and some empirical support. William Spelman’s estimate that increased incarceration accounted for approximately 25 percent of the crime decline, and Steven Levitt’s higher estimate of 32 percent, provided quantitative assessments that established incarceration as at minimum a contributing factor even if not the dominant one.

The theoretical mechanism was not disputed—incapacitating active offenders does prevent the crimes those offenders would otherwise commit, and the scale of incarceration expansion was large enough to have meaningful aggregate effects. The empirical debate centered on the size of the effect and on whether the incapacitation effect was large enough to offset the criminogenic effects of incarceration identified by labeling and social learning theories—the destruction of legitimate employment opportunities, family ties, and community connections that reentry research documented as drivers of recidivism. Bruce Western’s research demonstrating that the expansion of incarceration had produced substantial collateral damage to the communities most affected—increased family instability, reduced social capital, and disrupted labor markets in Black neighborhoods—suggested that the incapacitation benefit may have been partially offset by these downstream costs.

The Crack Cocaine Hypothesis

Alfred Blumstein and Richard Rosenfeld’s analysis of the specific dynamics of the crack cocaine epidemic provided the most analytically precise account of the crime drop’s timing and geographic distribution. They proposed that the crack epidemic had driven the crime surge of the late 1980s through a specific mechanism: the recruitment of young men as street-level drug market participants, the arming of those participants with firearms that were necessary for drug market competition, and the diffusion of those firearms from drug market participants into the broader youth population through social networks. The resolution of the crack epidemic—as markets matured and stabilized, as drug war enforcement removed key participants, and as the epidemic’s effects became visible to subsequent generations—produced a corresponding decline in the youth violence that the epidemic had driven.

The crack hypothesis is supported by the geographic concentration of the crime decline in cities where the crack epidemic had been most severe, by the timing of the decline’s beginning in the early 1990s as the crack market was apparently peaking, and by the demographic concentration of the decline among young Black men in urban areas—precisely the population most directly affected by crack market recruitment and associated violence. The hypothesis does not easily account for crime declines in cities and countries that were not significantly affected by crack cocaine markets, suggesting that other factors were operating alongside the epidemic’s resolution to produce the broader decline.

Franklin Zimring’s detailed analysis of New York City’s crime decline—documented in The City That Became Safe (2012)—provided the most rigorous single-city case study of the crime drop and generated conclusions that both supported and challenged the crack hypothesis. Zimring found that New York’s extraordinary decline (72 percent homicide reduction from 1990 to 2000) could not be explained primarily by incarceration expansion, since New York’s prison population did not expand as rapidly as many other states’ during the period. He argued that policing innovations—particularly the NYPD’s adoption of CompStat, hot spots deployment, and focused enforcement strategies—were more important contributors to New York’s specific experience than national-level analyses suggested. But Zimring also acknowledged that New York’s pattern of crime types that declined did not match well with what policing innovation alone would have predicted, suggesting that the resolution of the crack epidemic contributed substantially even in New York. The conclusion—that city-specific factors combined with broader national and international trends to produce the specific pattern observed in each jurisdiction—supports the over-determination hypothesis rather than any single causal account.

The Abortion Legalization Hypothesis

John Donohue and Steven Levitt’s controversial paper proposing that the legalization of abortion following Roe v. Wade in 1973 contributed substantially to the crime decline of the 1990s—by reducing the number of children born into high-risk circumstances who would have reached peak crime-committing age in the early 1990s—generated the most heated debate in the crime drop literature and among the most technically complex methodological exchanges in empirical criminology. The central argument was that abortion legalization had disproportionately reduced births to women who were young, unmarried, poor, and poorly educated—populations whose children were at elevated risk of criminal involvement—and that the resulting reduction in the size and composition of the high-risk birth cohort reaching adolescence in the early 1990s contributed to the crime decline of that decade.

The causal mechanism proposed by Donohue and Levitt was plausible in terms of what was known about risk factors for criminal behavior, and the timing prediction—that crime effects would emerge approximately eighteen to twenty years after abortion legalization, concentrated in states that legalized abortion before Roe—matched the observed pattern of the crime decline in ways that alternative explanations did not. Subsequent exchanges, including a vigorous replication debate in which John Joyce challenged the original methodology and Donohue and Levitt responded with revised estimates, did not definitively resolve the empirical question, but the most recent econometric analyses have generally been more supportive of the original hypothesis than the methodological critics suggested. The abortion hypothesis remains contested but not dismissed—a genuine empirical uncertainty about the contribution of a specific policy decision to aggregate crime rate trends.

The Lead Hypothesis

Rick Nevin’s analysis of the relationship between childhood lead exposure and subsequent crime rates—finding strong correlations between national blood lead levels lagged by approximately twenty-two years (the developmental period during which lead exposure affects brain development and subsequent behavior) and crime rates across multiple countries—provided a novel and controversial explanation for the crime drop that connected it to the phase-out of leaded gasoline during the 1970s. On Nevin’s account, the dramatic reduction in childhood lead exposure following the Clean Air Act’s mandate to remove lead from gasoline produced a cohort of young adults in the 1990s with lower average levels of the neurological impairment that lead exposure causes—impairment that increases impulsivity, reduces cognitive control, and thereby increases criminal propensity.

Jessica Reyes’s more rigorous econometric analysis, which exploited variation in the timing of state-level gasoline lead phase-outs to identify causal effects, found that the phase-out of leaded gasoline accounted for approximately 56 percent of the crime decline—a substantial effect if the identification strategy is valid. The lead hypothesis has the significant advantage of explaining the cross-national character of the crime drop—countries that eliminated lead from gasoline at similar times showed crime declines at similar lags—which country-specific explanations like policing strategies, incarceration expansion, and abortion legalization cannot easily account for.

The biological mechanism connecting childhood lead exposure to criminal behavior is well-established in the neuroscientific literature: lead exposure damages the prefrontal cortex and other brain regions involved in impulse control, executive function, and long-term planning, producing deficits that persist through adulthood and that increase propensity for impulsive, risk-taking behavior—including criminal behavior. The plausibility of the causal mechanism substantially strengthens the evidential value of the statistical associations, since a credible causal story connecting exposure to outcome reduces the probability that the observed correlations are spurious. The lead hypothesis also connects the crime drop to a specific public policy decision—the regulation of lead in gasoline—that was made on environmental and health grounds without any expectation that it would affect crime rates, providing an instance of inadvertent crime prevention through environmental policy that illuminates the complex, systemic nature of the conditions that produce and prevent crime.

Policing Innovations and the Place-Based Revolution

CompStat, Hot Spots, and Evidence-Based Deployment

The policing innovations of the early 1990s—particularly the introduction of CompStat in New York City under Police Commissioner William Bratton and the diffusion of hot spots policing strategies informed by Lawrence Sherman and colleagues’ experimental research—received enormous credit from police departments and their political allies for driving the crime decline. The temporal coincidence of policing innovation and crime decline in New York City was striking: homicide began falling sharply in 1990, coincident with major changes in NYPD strategy, and the city’s extraordinary 72 percent reduction through 2000 seemed to demand an explanation in terms of what New York specifically had done differently.

CompStat’s contribution to crime reduction was plausible in theoretical terms: by identifying the specific locations and patterns of crime through systematic data analysis, deploying patrol resources to those locations, holding commanders accountable for crime trends in their precincts, and tracking the effectiveness of specific tactical responses, CompStat instantiated the hot spots policing framework in an operational system that connected research evidence to resource allocation decisions. The experimental evidence for hot spots policing’s effectiveness—demonstrating that concentrated patrol at high-crime locations reduced crime significantly—provided theoretical support for the proposition that resource allocation informed by data could produce better crime reduction outcomes than traditional patrol patterns.

Community Policing and Problem-Oriented Strategies

Community policing and problem-oriented policing—the strategic frameworks that emphasized police-community partnership and the systematic analysis and resolution of specific crime problems rather than reactive response to incidents—also received substantial credit for the crime decline, particularly in cities like Boston where David Kennedy’s focused deterrence strategy produced dramatic reductions in gang homicide. The focused deterrence model—communicating directly to gang members that future violence would produce overwhelming enforcement consequences—represented a theoretically grounded alternative to both mass incarceration and traditional patrol that achieved large crime reductions in specific high-violence contexts.

The causal contribution of policing innovations to the broader national crime decline remains contested. John Eck and Edward Maguire’s systematic review found that the evidence for policing’s contribution to the national decline was weaker than the evidence for specific locally implemented strategies, because crime declined across cities with very different policing approaches and because the timing of policing innovations did not consistently precede the crime decline. The most reasonable conclusion is that policing innovations contributed to the decline in specific cities where they were most effectively implemented—particularly New York—while other factors accounted for the decline in cities where policing changed little and in countries where American policing innovations were not replicated.

The policing debate illustrates a broader challenge in crime drop research: establishing causal contribution from temporal coincidence is difficult when multiple factors changed simultaneously, when the counterfactual (what would have happened without the specific intervention) is unobservable, and when the mechanisms proposed are plausible but not distinctive—that is, when multiple different mechanisms all predict the same observed outcome. The methodological sophistication required to disentangle these competing explanations has driven substantial innovation in criminological methods, producing the difference-in-differences designs, regression discontinuity approaches, and synthetic control methods that contemporary criminology uses to address causal identification challenges that observational data alone cannot resolve. The crime drop debate has in this respect been a methodological forcing function for the discipline—creating demands for causal identification that have imported methods from econometrics and biostatistics into criminological practice, raising the evidentiary standards for causal claims about crime determinants, and producing a generation of criminologists who are more methodologically sophisticated in causal inference than their predecessors. This methodological improvement has consequences extending well beyond the specific crime drop debate to all areas of criminological research where causal claims are made on the basis of observational evidence.

The International Crime Drop

Cross-National Evidence and What It Reveals

The 1990s crime drop was not exclusively an American phenomenon but occurred simultaneously across virtually all Western developed nations—a cross-national character that substantially constrains the range of plausible explanations. Countries that experienced crime declines comparable in magnitude to the United States included Canada, England and Wales, Australia, the Netherlands, Germany, and most Scandinavian countries. These countries had very different criminal justice systems (with dramatically lower incarceration rates than the United States), very different policing strategies, no crack cocaine epidemic of comparable scale, and different demographic profiles—making the explanations specific to American institutional arrangements inadequate to account for the international scope of the decline.

The cross-national evidence most strongly supports the lead hypothesis—which predicts crime declines wherever leaded gasoline was phased out at comparable times—and suggests that the crack cocaine epidemic’s resolution, while real, was supplementary to more universal factors rather than the primary driver of the decline even in the United States. It also supports the possibility that cultural and normative changes associated with the maturation of generations that came of age during the second half of the twentieth century—changes in attitudes toward violence, in family formation patterns, and in the social norms governing conflict resolution—contributed to crime declines across diverse national contexts in ways that no single institutional explanation can fully capture.

Graham Farrell, Nick Tilley, and Andromachi Tseloni’s security hypothesis provides a further cross-national explanation: the rapid improvement in security technologies from the early 1990s onward—vehicle immobilizers, home security systems, mobile phone networks that facilitated rapid reporting, and improved cash management practices that reduced the rewards available from robbery—substantially reduced the opportunities for the property crimes (vehicle theft, burglary, robbery) that declined most dramatically in the international crime drop. The security hypothesis is supported by the specific pattern of property crime decline—concentrated in crimes where improved security most directly reduced opportunity—and by the timing of the decline relative to the diffusion of specific security technologies. It offers a distinctive causal mechanism from both the lead hypothesis and the crack epidemic resolution, and is not mutually exclusive with either: all three mechanisms may have operated simultaneously to produce the convergent international decline.

Post-1990s Trends and the Persistence of the Decline

The crime decline that began in the early 1990s proved far more durable than most criminologists had anticipated—extending through the 2000s and into the 2010s, surviving the severe economic recession of 2008–2009 that conventional wisdom suggested would reverse the trend, and continuing at lower levels well into the current decade. This persistence challenged the economic explanations for the crime drop—if improved labor markets during the 1990s boom had driven crime down, the sharp deterioration of those markets during the Great Recession should have reversed the trend. The fact that crime rates continued to decline (or at worst stabilized) through the recession suggested that the structural changes underlying the crime drop were more fundamental than cyclical economic conditions.

The post-2015 partial reversal—the uptick in homicide rates that occurred in many American cities beginning in 2015 and accelerating through the COVID-19 pandemic years—has reopened the debate about what had caused the original decline, as the factors proposed to explain the reversal (decreased policing, reduced incarceration, increased firearm availability, social disruption from the pandemic) implicitly make claims about which factors had been responsible for the preceding decline. The partial reversal confirmed that the crime decline, while durable, was not permanent—that the structural and contextual factors maintaining lower crime rates required active maintenance and could be disrupted by specific adverse developments. Understanding the mechanisms of both the decline and the subsequent reversal remains one of the most important empirical challenges facing contemporary criminology.

Conclusion

The 1990s crime drop defies any single-factor explanation, and the attempt to find one has been among the most productive exercises in criminological research of the past three decades—productive not because it has produced consensus but because the debate has generated rigorous empirical analysis of mechanisms whose complexity the pre-crime-drop theoretical landscape had underestimated. The most defensible conclusion is that the crime drop was almost certainly over-determined: the resolution of the crack cocaine epidemic, the long-delayed effects of childhood lead exposure reduction, the incapacitation effects of expanded incarceration, improvements in policing strategy, and possibly other factors all contributed to the decline, with different factors predominating in different cities and for different crime types.

The crime drop’s theoretical legacy for criminology extends beyond the specific debate about its causes to a broader lesson about the complexity of crime rate determination. The dramatic increase in crime from the late 1950s through the early 1990s, and its equally dramatic subsequent decline, occurred against a background of relatively stable social conditions—poverty rates, urbanization, inequality—that structural theories identify as the primary determinants of crime rates. The crime rate’s dramatic variation in the face of structural stability suggests that crime rates are more sensitive to specific contextual factors—drug market dynamics, exposure to environmental toxins, the precise character of policing strategies—than structural theories alone predict. This lesson in the complexity and contingency of crime rate determination is among the most important that the crime drop debate has produced.

The crime drop also carries a practical lesson of equal importance: that crime rates can fall dramatically and sustainably over a decade without requiring the kind of mass incarceration that the United States pursued in the belief that only more imprisonment could make communities safer, and that the specific mechanisms of crime reduction—environmental policy, drug market evolution, security technology improvement, focused deterrence—may be more powerful determinants of crime rates than the blanket criminal justice responses that receive the most political attention and absorb the most public resources. This lesson has not fully penetrated American crime policy, which continues to reflexively reach for incarceration and enforcement responses when crime rises. But the evidence from the crime drop—that a 43 percent reduction in homicide could be achieved without reference to any single criminal justice policy initiative—suggests that the most consequential crime prevention policies may operate outside criminal justice altogether: in public health, environmental regulation, community development, and the management of specific criminal markets that drive violence.

Patrick Sharkey’s Uneasy Peace (2018) documented the social dividends of the crime decline with particular vividness—the revitalization of urban neighborhoods, the improvement in children’s educational outcomes, the expansion of economic opportunity in communities previously devastated by violence—in ways that underscored the human significance of crime rate changes that the technical debates about causal mechanisms can obscure. Understanding why the crime drop happened matters not only as an intellectual puzzle but as a practical question whose answer could guide efforts to sustain the gains of the 1990s and to reverse the partial increases that have occurred since 2015. The debate about the crime drop’s causes is therefore both one of the most intellectually demanding and practically consequential exercises in criminological research of the past three decades—a sustained scholarly controversy whose partial resolution has already shaped criminological theory and crime prevention policy, and whose remaining uncertainties continue to motivate some of the discipline’s most methodologically rigorous work.

References

  1. Blumstein, A., & Rosenfeld, R. (1998). Explaining recent trends in U.S. homicide rates. Journal of Criminal Law and Criminology, 88(4), 1175–1216. https://doi.org/10.2307/1144139
  2. Blumstein, A., & Wallman, J. (Eds.). (2006). The crime drop in America (rev. ed.). Cambridge University Press.
  3. Cook, P. J., & Laub, J. H. (2002). After the epidemic: Recent trends in youth violence in the United States. Crime and Justice, 29, 1–37. https://doi.org/10.1086/652220
  4. Donohue, J. J., & Levitt, S. D. (2001). The impact of legalized abortion on crime. Quarterly Journal of Economics, 116(2), 379–420. https://doi.org/10.1162/00335530151144050
  5. Eck, J. E., & Maguire, E. R. (2000). Have changes in policing reduced violent crime? An assessment of the evidence. In A. Blumstein & J. Wallman (Eds.), The crime drop in America (pp. 207–265). Cambridge University Press.
  6. Farrell, G., Tilley, N., & Tseloni, A. (2014). Why the crime drop? Crime and Justice, 43(1), 421–490. https://doi.org/10.1086/678081
  7. Gould, E. D., Weinberg, B. A., & Mustard, D. B. (2002). Crime rates and local labor market opportunities in the United States: 1979–1997. Review of Economics and Statistics, 84(1), 45–61. https://doi.org/10.1162/003465302317331919
  8. Levitt, S. D. (2004). Understanding why crime fell in the 1990s: Four factors that explain the decline and six that do not. Journal of Economic Perspectives, 18(1), 163–190. https://doi.org/10.1257/089533004773563485
  9. MacDonald, J. M. (2002). The effectiveness of community policing in reducing urban violence. Crime & Delinquency, 48(4), 592–618. https://doi.org/10.1177/001112870204800405
  10. Marvell, T. B., & Moody, C. E. (1994). Prison population growth and crime reduction. Journal of Quantitative Criminology, 10(2), 109–140. https://doi.org/10.1007/BF02221155
  11. Nevin, R. (2007). Understanding international crime trends: The legacy of preschool lead exposure. Environmental Research, 104(3), 315–336. https://doi.org/10.1016/j.envres.2007.02.008
  12. Ousey, G. C., & Lee, M. R. (2002). Examining the conditional nature of the illicit drug market-homicide relationship: A partial test of the theory of contingent causation. Criminology, 40(1), 73–102. https://doi.org/10.1111/j.1745-9125.2002.tb00950.x
  13. Raphael, S., & Winter-Ebmer, R. (2001). Identifying the effect of unemployment on crime. Journal of Law and Economics, 44(1), 259–283. https://doi.org/10.1086/320275
  14. Reyes, J. W. (2007). Environmental policy as social policy? The impact of childhood lead exposure on crime. B.E. Journal of Economic Analysis & Policy, 7(1). https://doi.org/10.2202/1935-1682.1796
  15. Rosenfeld, R. (2004). The case of the unsolved crime decline. Scientific American, 290(2), 82–89. https://doi.org/10.1038/scientificamerican0204-82
  16. Sharkey, P. (2018). Uneasy peace: The great crime decline, the renewal of city life, and the next war on violence. Norton.
  17. Sharkey, P., Torrats-Espinosa, G., & Takyar, D. (2017). Community and the crime decline: The causal effect of local nonprofits on violent crime. American Sociological Review, 82(6), 1214–1240. https://doi.org/10.1177/0003122417736289
  18. Spelman, W. (2000). The limited importance of prison expansion. In A. Blumstein & J. Wallman (Eds.), The crime drop in America (pp. 97–129). Cambridge University Press.
  19. Western, B. (2006). Punishment and inequality in America. Russell Sage Foundation.
  20. Zimring, F. E. (2012). The city that became safe: New York’s lessons for urban crime and its control. Oxford University Press.
  21. Zimring, F. E., & Hawkins, G. (1997). Crime is not the problem: Lethal violence in America. Oxford University Press.
  22. Weisburd, D., & Majmundar, M. K. (Eds.). (2018). Proactive policing: Effects on crime and communities. National Academies Press.
  23. Welsh, B. C., & Farrington, D. P. (Eds.). (2012). The Oxford handbook of crime prevention. Oxford University Press.