Reducing Crime Through Environmental Design: Evidence from a Randomized Experiment of Street Lighting in New York City

‍

This paper presents experimental evidence demonstrating that crime can be effectively reduced by altering the situational environment encountered by potential victims and offenders. We examine the impact of street lighting, a pervasive yet surprisingly under-researched feature of urban landscapes, and provide the first experimental evidence on its effect on crime. Through a unique public partnership in New York City, temporary streetlights were randomly allocated to public housing developments from March through August 2016. Our findings reveal that communities receiving additional lighting experienced significant reductions in crime. After accounting for potential spatial spillovers, we estimate that the provision of streetlights led to at least a 36 percent reduction in nighttime outdoor index crimes.

‍

Acknowledgments: We extend our gratitude to the New York City Police Department for providing the data used in this study. These data belong to the NYPD, and any further use must be approved by the NYPD. We also thank the New York City Mayor’s Office of Criminal Justice for coordinating this study, the New York City Housing Authority for logistical support and communication with residents, and the Laura and John Arnold Foundation for their generous support of the University of Chicago Crime Lab and this project. We are grateful to Valentine Gilbert, Melissa McNeill, and Anna Solow-Collins for their exceptional research assistance, as well as to Roseanna Ander, Robert Apel, Monica Bhatt, Monica Deza, Jennifer Doleac, Katy Falco, Justin Gallagher, David Haftez, Zubin Jelveh, Jacob Kaplan, Max Kapustin, Mike LaForest, Jens Ludwig, John MacDonald, Vikram Maheshri, Aurélie Ouss, Greg Ridgeway, and Nick Sanders for their valuable comments on earlier versions of the manuscript. The opinions expressed in this paper are those of the authors and do not necessarily reflect those of the Laura and John Arnold Foundation or the New York City Police Department.

‍

Introduction

‍

For decades, social scientists have debated the relative influence of individual characteristics—shaped by a combination of genetics and environmental exposures during childhood and adolescence—versus situational aspects of the environment in which judgments, decisions, and behaviors are made. To what extent is behavior driven by "the person versus the situation," as Lee Ross and Richard Nisbett famously posed the question?

‍

A substantial body of research in behavioral economics and psychology suggests that small environmental changes can have surprisingly large effects on human behavior. This principle is evident in Stanley Milgram’s classic obedience studies and, more recently, in various policy-relevant applications involving "nudges" (see Bertrand, Mullainathan, and Shafir 2006). Low-cost interventions that leverage minor environmental changes are particularly appealing as they have the potential to alter costly behaviors without the need for resource-intensive, invasive, or difficult-to-scale interventions. Nowhere are such interventions more urgently needed than in criminal justice, a domain where unwanted behaviors result in enormous social costs, potentially as high as $1 trillion per year (Chalfin 2015), and where the primary intervention—incapacitation—has been employed so extensively that the United States now has the highest incarceration rate in the world. Beyond the substantial financial burden on the government, the collateral harms associated with incarceration, including direct economic and social impacts on families and communities, disproportionately affect low-income, racially segregated neighborhoods (Western, Kling, and Weiman 2001; Aizer and Doyle 2015; Mueller-Smith 2016).

‍

Can crime be influenced by relatively minor situational changes? Public policy has largely overlooked this possibility, focusing instead on costly enforcement actions or resource-intensive social programs, which can be difficult to evaluate and scale. However, two well-documented empirical regularities provide grounds for optimism. First, the significant geographic concentration of crime, particularly violent crime, suggests that the social and physical features of the urban landscape might play a crucial role in the crime production function (Weisburd 2015). Second, evidence from psychology and economics indicates that offenders are often myopic and tend to have high social discount rates (Lee and McCrary 2017). In a world where potential offenders are shortsighted, small environmental design changes experienced in the present may have a disproportionately large impact on behavior compared to the uncertain prospect of future prison sentences.

‍

The idea that crime is difficult to curb through light-touch interventions is perhaps best captured by Robert Martinson’s seminal essay, "What Works" (1974), which argued that criminal behavior is resistant to change, as evidenced by high recidivism rates and the failure of rehabilitative efforts to significantly reduce reoffending. Martinson’s essay received widespread attention and, as crime rates continued to climb during the 1970s and 1980s, mass incarceration seemed, to many, like the only viable solution.

However, the importance of the environment is implicit in research such as the Moving to Opportunity study of the early 2000s (Katz, Kling, and Ludwig 2005), which continues to be a subject of debate (Sampson 2008; Chetty, Hendren, and Katz 2016; Chyn 2017).

‍

With public safety in mind, cities have recently renewed their interest in modifying the physical design of public spaces. A small body of research suggests that certain property crimes are surprisingly responsive to environmental factors, such as the availability of criminal opportunities (Ayres and Levitt 1998; Cook and MacDonald 2011) and physical disorder, as evidenced by a study in Science reporting that physical disorder begets further disorder (Keizer, Lindberg, and Steg 2008). However, it remains unclear whether changes to the physical environment can effectively reduce serious crimes, particularly violent crimes, which contribute the most to social costs.

‍

In this study, we provide the first experimental evidence that a minor change in environmental design—specifically, street lighting—can reduce violence. Street lighting, a common yet understudied element of urban infrastructure, has been relied upon by cities to maintain public safety for over two centuries. While many communities may lack the resources to convert vacant lots into parks or repair abandoned buildings, every city faces darkness each night. By enhancing visibility, lighting has the potential to influence crime through various mechanisms, including empowering potential victims to better protect themselves and signaling to potential offenders that a public space is observed or that police are present.

There is some evidence that other environmental strategies, such as increasing the availability of trees and green space (Branas et al. 2011; Kondo et al. 2016), and securing abandoned buildings (Branas et al. 2016), may reduce violence. However, the evidence is mixed and non-experimental (Bogar and Beyer 2016). Lighting, like many traditional behavioral interventions, provides participants in the crime market with information—typically obscured by darkness. Lights may also complement traditional deterrence strategies, such as police patrols (Sherman and Weisburd 1995), surveillance cameras (Priks 2015), and "eyes on the street" (Carr and Doleac 2017), potentially driving down crime through additional mechanisms.

‍

Despite wide interest in street lighting among scholars, policymakers, and the public, no randomized experiment has yet studied its effectiveness in controlling crime. This study, facilitated by a unique partnership between the New York City Mayor’s Office for Criminal Justice (MOCJ), the New York City Police Department (NYPD), and the New York City Housing Authority (NYCHA), offers the first experimental evidence on the effectiveness of street lighting in reducing street crime, particularly violent crime.

Our field experiment, conducted in 2016 in NYC—a city where crime has declined significantly over the last three decades, partly due to innovations in policing—focuses on public housing communities, where violent crime remains disproportionately high. NYC’s public housing communities were thus selected as the preferred setting for the intervention.

With any place-based experiment, two core research challenges must be addressed: statistical power and sensitivity to modeling assumptions. Intervening in large places, such as housing projects, is costly. Although more than 80,000 people live in the areas studied, only a small number of locations could be treated. Furthermore, index crimes, while elevated, are still relatively rare and variable. To maximize statistical power, we employed a block randomized design that randomized the dosage of lighting received by each community. This design allows for greater variation and, therefore, maximizes statistical power.

‍

We also addressed the sensitivity of treatment effects to different modeling assumptions. Given the large number of potential covariates relative to observations, we used LASSO regression to select predictors with the greatest out-of-sample predictive power, minimizing researcher discretion and guarding against finite sample bias due to imperfect randomization.

‍

We estimate that the introduction of additional lighting reduced outdoor nighttime index crimes by approximately 60 percent, and by at least 36 percent once potential spatial spillovers were accounted for. These findings provide the first evidence that the physical environment of cities and communities is a key determinant of serious crime.

‍