Crime and Main Street
I testified before the House Small Business Committee last week on trends in crime and victimization and their effects on small businesses. As they say on the Hill, I would like to revise and extend my remarks. Well, not revise, but I’d like to make one small point that comes up a lot in the kinds of events, but that didn’t happen to come up in this one. Which is, both data and storytelling matter in crafting policy. Here’s the metaphor.
In physics, the study of the largest forces in the universe falls under the umbrella of cosmology. The study of foundational units of matter, the smallest bits of the universe, is studied in particle physics. It turns out though, that finding a unified theory that explains how we get from the smallest forces to the largest, and the largest forces to the smallest, is not that straightforward.
If you were to take the foundational principles of participle physics and apply inductive reasoning to generalize from your specific knowledge about quarks and bosuns and such, it is very unlikely that you would find yourself deriving dark matter and the Big Bang, at least not without some bootstrapping.
Similarly, if you were to study cosmic microwave background radiation and dark energy and try and deduce foundational elements of matter that are irreducible, it’s not clear you would arrive at bosuns and quarks and such.
Now, I am not a physicist, so I am sure there is much to quibble about in those statements. After all, those who can, do physics. Those who can’t, do something else. And I do something else.
The point remains, at least metaphorically. If you begin with the general and try to deduce truths about the specific, it’s not a straight line (and is often a logical fallacy). Conversely, making general conclusions from specific examples generates some spread in your confidence, and probably some error. And this is true, no matter how sophisticated your analysis.
Here’s the point. My work is at the policy level and so I lean on statistical analysis to understand average effects and the shape of the distribution of those effects. The criticism I tend to get, and the criticism others who use similar tools tend to get, is that it dehumanizes the problem and critical points are missed. To challenge my approach, I am often confronted with compelling anecdotes. The implicit criticism is: by focusing on the forest, you have missed the trees, and in the end what we are trying to fix are trees, not forests. And that’s fair, to a point.
But I don’t see these approaches as being in conflict. Each data point in a distribution around an average effect has a story attached to it. Telling those stories gives the whole structure meaning. Stories provide insight into important information that is outside the data. Stories generate new hypotheses to test. And more. But you can also learn a lot from studying the forest as a whole.
It’s a dance, not a competition.
That’s why I like the physics metaphor. Physics with cosmology alone or particle physics alone provides an incomplete understanding of the universe. Empirical data without qualitative context is similarly incomplete. And qualitative data without empirical evidence is also complete.
So, in the context of this hearing, and similar events, I welcome stories about what is actually happening behind the numbers.
But the thing is, there are multiple ways to solve difficult problems. One tack is to take a statistical approach and find policies that nurture the forest and improve the lot of the average tree. That doesn’t solve every problem or fix every tree, but on average, it makes things better. Another approach is to work on problems individually within their own context. It seems straightforward to me to do both, or at least maintain an awareness of the gap that results from only doing one or the other.
House Committee on Small Business, January 11, 2024
Crime on the Rise: How Lawlessness is Impacting Main Street America
Chairman Roger Williams, Ranking Member Nydia M. Velázquez, and Members of the Committee: I am honored by the opportunity to discuss the important issue of the effects of crime on small businesses in America. I am a senior fellow and director of the Center on Public Safety and Justice at NORC at the University of Chicago (NORC). NORC is a nonprofit economic and social research organization committed to delivering objective, nonpartisan insights and analysis that decision-makers trust. The Center on Public Safety and Justice develops actionable, evidence-based solutions to crime and victimization to empower decision-makers to make criminal justice systems fairer and more effective. Our solutions can also be used to inform community-level responses beyond policing and incarceration. In addition to my role at NORC, I also serve as the chair of the Crime Trends Working Group at the Council of Criminal Justice, which collects real-time crime data, identifies gaps in data collection and availability, and develops strategies to improve crime reporting nationwide. I chair the Federal Firearms Data Infrastructure working group at Safe States Alliance, and I co-direct the National Prevention Science Coalition to Improve Lives. I have been working on these issues for more than 25 years.
The last four years have been a time of profound uncertainty for all Americans, including small business owners and their patrons. In my field of criminal justice and economics, between 2019 and 2020, we observed the largest one-year increase in violent crime and homicide. By 2023, we saw the largest one-year decrease in violence. This period of disruption included several unusual changes in criminal activity, such as reports of increased carjackings, frightening videos of organized retail theft, and vicious images of hate crimes. Separately and together, these events have created deep concern throughout American communities. I purposefully use the word "reports" rather than "statistics" because the empirical data around each of these crimes are extremely limited; thus, our ability to craft evidence-based, data-driven policy is very limited.
My testimony expands upon three distinct areas: 1) what is known about crime and violence overall in the last four years; 2) what is known and unknown about organized retail theft; and 3) the gap in reporting of crime statistics. This last point is critical because it answers both questions at the forefront of this committee: "What is the real risk to small business owners and patrons of Main Street?" and "What is the real risk to the average American as they go about their daily routines?"
CRIME AND VIOLENCE IN THE UNITED STATES: 2020-2023
The last four years have been a time of unprecedented volatility in crime. Crime in the United States varies widely across cities and states; however, year-to-year changes at the national level tend to be small. Taking a historical perspective back to 1960, crime rates in the United States have exhibited long-term trends rather than abrupt increases or decreases. The 1960s and 1970s were a period of steadily increasing crime.[1] Crime and violence remained at all-time high rates throughout the 1980s.[2] The 1990s saw rapid declines in criminal activity throughout the United States.[3] Crime slowly declined in the 2000s and remained steady in the 2010s.[4] And while crime and violence vary substantially across cities and over time, large year-to-year changes are rare. Between 1960 and 2020, for example, the homicide rate never changed by as much as 10 percent.[5]
Beginning in March 2020, the effects of the COVID-19 pandemic began to overwhelm the daily routines of many, if not most, Americans. This disruption had a significant impact on reported crimes. For example, some types of crime and violence rapidly increased, particularly gun homicides. The changes in crime in 2020 are evident when compared with crime in 2019. During that year, aggravated assaults increased 12.1 percent and homicides increased 29.4 percent. The change from 2019 to 2020 was the largest one-year increase in homicide since the FBI began modern record-keeping in 1960.[6] But not all types of crime increased—robberies declined by 9.3 percent, reported rape by 12 percent, and property crime by 7.4 percent. As with violent crime, the decrease in property crimes was not uniform—motor vehicle thefts increased by 11.8 percent while larceny-theft declined by 10.6 percent.
Changes in reported crimes of this magnitude are extremely unusual. At the same time, Americans’ perception of their risk of victimization increased rapidly, and they proposed many hypotheses to explain the rapid rise in violence. In December 2023, the Council on Criminal Justice released a report, Trends in Homicide: What You Need to Know, that detailed six leading theories of why violence spiked in 2020.[7] These include:
Routine Activity Theory. This theory posits that, for a crime to occur, a motivated offender, a suitable target, and an absence of guardianship must converge. Social distancing during the height of the pandemic reduced the number of convergences.
Police Legitimacy. After George Floyd’s death, this theory proposes that citizen perceptions of policing authority as legitimate declined, making citizens less likely to contribute to police investigations, which reduced the clearance rate and thus reduced the perceived cost of committing crimes.
De-Policing. De-policing theory is hypothesized to have occurred in two ways. In the wake of protests following George Floyd’s death, police engaged in less proactive policing, so fewer crimes were prevented. Also, many departments directed their officers to reduce police-citizen contacts to protect officer health and safety.
Gun Sales. Between 2019 and 2020, the number of firearms purchased increased from 13 million to more than 20 million. The theory hypothesizes that legally purchased firearms become crime guns in fixed proportions, so the increase in sales led to more guns on the street.
Bail Reform and Progressive Prosecution. The theory hypothesizes that changes in prosecutorial practices, such as ending cash bail and effectively decriminalizing low-level offenses, led to increases in crime.
Drug Market Disruption. Research suggests that the increase in violence in 2020 was concentrated in places with drug markets and concentrations of high levels of violence.
Recently reported data helps to clarify the timing of the crime and violence spike. In May 2022, the Centers for Disease Control (CDC) reported that the homicide increase in 2020 was largely due to firearms: while the overall homicide rate increased 28 percent, the firearm homicide rate increased 34.6 percent.[8] In the first three months of 2020, there were 3,447 homicides—about a 7 percent increase from the same month in 2019 and a 5 percent increase from 2018. In April, firearm homicides increased by 16 percent in a single month and 18 percent in May. By the end of May, firearms homicides had increased by 33 percent. In the summer months, the firearms homicide rate stabilized, increasing about 4 percent in June, declining 2 percent in July, and declining 7 percent in August.
With respect to the timing of the homicide increase and the timing of the theoretical explanations, social distancing in response to the COVID-19 pandemic best fits the data. Social distancing measures, including pandemic-related shutdowns, were in place and widespread by the beginning of April 2020. Other macro forces that correlate to the spike in crime, such as large numbers of local government employees being separated from their jobs, began at the same time.
Crime Increases across Sectors 2021-2022
One major obstacle to understanding the crime spike and subsequent decline is the lack of national crime data for 2021. The data from 2021 are more limited than other years in the FBI national crime statistics data series. In 2020, FBI data include data from both the Summary Reporting System (SRS)—which has been the main source for decades—and the National Incident-Based Reporting System (NIBRS). In 2020, 16,546 of 18,631 law enforcement agencies submitted data. In 2021, the FBI discontinued the SRS and mandated that only NIBRS data to be reported, and the number of reporting agencies declined to 13,181.[9] In 2022, the FBI again allowed reporting agencies to submit either SRS or NIBRS data, and the number of reporting law enforcement agencies increased to 15,726. Thus, in 2021, only 65 percent of the United States population was served by a law enforcement agency reporting crime data to the FBI.[10] In prior years, about 95 percent of the U.S. population was covered by these official national crime statistics.[11]
Overall, between 2020 and 2021, crime in the United States was relatively stable, at least compared with the volatility of the prior year.[12] All violent crime declined by 1.7 percent, although homicide increased by 4.3 percent. Aggravated assault remained steady (a decline of 0.1 percent) and robbery declined substantially (-8.2 percent). Property crime declined by 3.8 percent, larceny-theft declined 4.9 percent, and burglary declined 11.4 percent, while motor vehicle theft increased 12.3 percent.
Compared to 2021, violent crime declined by 1.6 percent in 2022. Homicide declined by 7.9 percent, and aggravated assault declined by 1.4 percent. Robbery increased by 0.9 percent. Property crime rose 6.7 percent, including a 7.4 percent increase in larceny-theft, a steady burglary rate, and a third consecutive year of large increases in motor vehicle theft (10.5 percent).
The 2023 Crime Decline
The December 2023 Council on Criminal Justice Trends in Homicide: What You Need to Know report includes data from 30 cities in the United States and explains that homicide declined 9.4 percent in the first half of 2023 compared with the same period in 2022. The June 2023 homicide rate (1.39 per 100,000) is essentially the same as the June 2018 homicide rate (1.36). Crime analyst Jeff Asher reports that, as of early December 2023, homicide has declined 12.7 percent in 2023, which would be the largest one-year decline in homicide since modern records have been recorded.[13] Asher also notes that through the third quarter of 2023, property crime has declined 6.3 percent, including declines in burglary (-11.7 percent) and larceny-theft (-8.5 percent). Only motor vehicle theft continues to increase, up again in 2023 by 10.1 percent.
To summarize the period 2020 through 2023:
Homicide remains above pre-pandemic levels but has declined two years in a row.
Violent crime rates appear to have fallen below pre-pandemic levels.
Property crime rates were elevated for one year (2022) but have remained below pre-pandemic levels throughout the period.
Burglary rates have declined throughout the period and remain below pre-pandemic levels.
Larceny-theft rates increased in 2022 and are below pre-pandemic levels.
Explaining the Crime Spike and Crime Decline
Understanding why—and how much—crime first rose then fell is critical to creating evidence-based policies. Policy changes should account for both the sudden increase and the sudden decline. The explanation that best fits the data is that much of the volatility in crime and violence was a result of COVID-19-related causes.
The pandemic can be simply described as an "everything, everywhere, all-at-once" experience for most Americans. Most explanations for the initial crime spike and the subsequent crime decline are not similarly all-encompassing, and most explain the crime spike but not the crime decline. The most straightforward explanation for both the spike and the decline is the pandemic and the direct consequences of the pandemic. These consequences occurred in different ways, at different times, and together they provide a coherent narrative.
The spike in firearms homicide coincides with the early days of social distancing caused by COVID-19. This would have kept people who live in historically high-violence places within close proximity of each other during their routine activities. At the same time, guardianship was limited, due to reduced proactive policing either to protect the health and safety of officers and/or in response to public disapproval. As the social distancing abated, firearms homicides first stabilized and then declined.
The later increases of some types of violence and property crimes in 2022 appear to be related to second-order effects of the pandemic. By second-order, we mean changes that indirectly affect crime and violence rather than directly (for example, changes in routine activities can be seen as a direct effect). One large-scale indirect effect of the pandemic on crime was the substantial reduction in local government workforce in the first months of the pandemic. Hundreds of thousands of local government employees who work directly with young people (e.g., teachers, counselors, law enforcement)—who represent the highest-risk population—lost their positions.
The Effect of Local Government on the Crime Spike and Crime Decline
What almost all of the explanations for the crime decline in the Council on Criminal Justice paper and in other research have in common is that they are activities supported by local government, which includes any government below the level of state government, including county, municipal, and township. Local government experienced a dramatic reduction in workforce during the pandemic that persisted through 2022 and then rapidly reversed in 2023. After a lengthy period of underfunding and under-staffing caused by the pandemic, local governments have today, by most measures, returned to pre-pandemic levels.[14]
Local governments perform numerous activities that are directly related to fighting crime. Most of the 700,000 sworn law enforcement officers are employed by local governments, as are about 8 million teachers, including secondary schools and community colleges.[15] Local government employees include behavioral health specialists, nurses, and other health workers, as well as grant monitors who support community-based organizations. Local governments also employ bailiffs, correctional officers, and probation and parole agents (among other justice system employees). Between March and May 2020, more than 1.24 million local government employees lost their jobs. Another 300,000 state government employees lost their positions. Bureau of Labor Statistics data report that local government staffing levels did not return to pre-pandemic levels until December 2023.
ORGANIZED RETAIL THEFT: 2020-2023
It has been widely suggested that commercial larceny-theft, often referred to as shoplifting, changed substantially during the pandemic. A LexisNexis search for “organized retail theft” and “smash and grab” news stories by the Center for Just Journalism found a spike in news stories on both topics beginning in 2021.[16] A November 2023 study by the Council on Criminal Justice used open-source data from 24 cities over five years to estimate trends in retail larceny-theft.[17] The study replicated the national data described above and found that larceny-theft declined in 2020 and 2021 before increasing in 2022, likely due to changes in routine activities. There was tremendous variation across cities, but, overall, the data are very sensitive to the inclusion of New York City. With New York City in the sample, larceny-theft increased in 2023 (compared with 2020) by 16 percent, yet without New York City in the data, larceny-theft declined by 7 percent. The rates across all months were lower than all but one month in 2018 and 2019.
The challenge with understanding commercial retail theft is that national crime statistics are not designed to measure that specific event. There are many sources of retail loss—theft in the supply chain (cargo theft), paperwork errors (fraud), employee theft, as well in-store theft, which may be amateur thieves, opportunists, "professionals," or theft by organized criminal enterprises. Crimes reported to the police do not distinguish among these categories. A call for service is likely to be recorded in a computer-aided dispatch (CAD) system using NIBRS codes from the FBI.[18] In the category of property crimes, the types of larceny/theft offenses are extensive and include: pocket-picking; purse-snatching; shoplifting; theft from building; theft from coin-operated machine or device; theft from motor vehicle; theft of motor vehicle parts or accessories; and all other larceny.
Instead, many “smash and grab” crimes against retailers may be coded as burglaries, which are defined as “the unlawful entry into a building or some other structure to commit a felony or a theft.”[19] They also may be coded as a robbery, which is the taking of property by force or threat of force. Of note, according to the FBI instructions to local law enforcement agencies submitting crime data, “every Robbery includes some type of assault, but because the assault is an element integral to the crime of Robbery, the LEA should report only Robbery.”[20]
As is discussed below, national crime statistics in the United States are very limited in general, and, as is shown above, they are of limited utility in understanding the magnitude of organized retail theft. Overall, however, there is limited evidence that organized retail thefts and “smash and grab” crimes have increased nationally. Larceny-theft statistics do not show meaningful increases, except in 2022, nor do robbery and burglary statistics.
IMPROVING THE STATE OF KNOWLEDGE
One important takeaway from this discussion is the inadequacy of current national crime statistics to address these important questions. Organized retail theft is not the only type of crime subject to these limitations. Other examples of insufficient data in national statistics abound. Currently, NIBRS does not record shootings incidents or shooting victimizations—instead, NIBRS includes a category, aggravated assault with firearm (with and without injury), which includes incidents where the weapon was discharged, where the weapon was only brandished, and where the weapon was used to inflict blunt force trauma. Other categories of criminal victimization that have received intense media and public interest during the pandemic include carjacking, which is not an offense codable in NIBRS. Also, data on firearms are not just limited but specifically constrained by the Tiahrt Amendment, which restricts the ATF’s ability to share locally collected data with local jurisdictions, and the Dickey Amendment, which limits the CDC’s ability to provide grants to researchers to study firearms injury.
CONCLUSION
In recent years, the United States experienced a disruption in daily life that has few precedents in recent decades. One consequence was that crime and violence exhibited radical changes in short timeframes, resulting in concerns about victimization that deepened in all corners of the nation—from neighborhoods with persistent high levels of violence, to Main Streets with little history of victimization. When faced with such uncertainty, one reasonable response is to explore deep policy changes. But in the case of crime and violence, the key causes appear to be direct effects of the pandemic. As effects of the pandemic on routine activities return to normal, those negative effects have begun to dissipate. While there are many lessons from the pandemic, one that gets less attention is the need to strengthen our national criminal justice data infrastructure. Doing so will strengthen the evidence for future policymaking if another crisis emerges or if our policy goals shift from a focus on reducing crime and victimization to a long-term equilibrium with high rates of victimization to an equilibrium at safer levels.
[1] Pfaff, John F. "The micro and macro causes of prison growth." Georgia State University Law Review 28, no. 4 (2012).
[2] Blumstein, Alfred. "The crime drop in America: An exploration of some recent crime trends." Journal of Scandinavian Studies in Criminology and Crime Prevention 7, no. S1 (2006): 17-35.
[3] Levitt, Steven D. "Understanding why crime fell in the 1990s: Four factors that explain the decline and six that do not." Journal of Economic Perspectives 18, no. 1 (2004): 163-190.
[4] Baumer, Eric P., and Kevin T. Wolff. "Evaluating contemporary crime drop (s) in America, New York City, and many other places." In Understanding New York’s Crime Drop, pp. 5-38. Routledge, 2020.
[5] Author’s calculation.
[6] FBI National Press Office, “FBI Releases 2020 Crime Statistics.” September 27, 2021. Accessed January 7, 2024: https://www.fbi.gov/news/press-releases/press-releases/fbi-releases-2020-crime-statistics.
[7] Council on Criminal Justice. “Trends in Homicide: What you need to know”. December, 2023. Washington, DC: Council on Criminal Justice.
[8] Scott R. Kegler; Thomas R. Simon; Marissa L. Zwald; May S. Chen; James A. Mercy; Christopher M. Jones; Melissa C. Mercado-Crespo; Janet M. Blair; Deborah M. Stone; Phyllis G. Ottley; and Jennifer Dills. “Vital Signs: Changes in Firearm Homicide and Suicide Rates — United States, 2019–2020.” Morbidity and Mortality Weekly Report (MMWR) 71(19);656–663. May 13, 2022.
[9] Federal Bureau of Investigation, “Crime/Law Enforcement Stats (Uniform Crime Reporting Program)” accessed January 7, 2024: https://www.fbi.gov/how-we-can-help-you/more-fbi-services-and-information/ucr.
[10] Berzofsky, Marcus, Dan Liao, G. Lance Couzens, Erica L. Smith, and Cynthia Barnett-Ryan. "Estimation procedures for crimes in the United States based on NIBRS data." Bureau of Justice Statistics (2022).
[11] Ibid.
[12] All data in the paragraph refer to: Federal Bureau of Investigation. (undated). “The Transition to the National Incident-Based Reporting System (NIBRS): A Comparison of 2020 and 2021 NIBRS Estimates.” Accessed January 7, 2023, link available at: https://www.hsdl.org/c/abstract/?docid=871879.
[13] Asher, Jeff. “Crime in 2023: Murder Plummeted, Violent and Property Crime Likely Fell Nationally.” Accessed January 7, 2024.
[14] U.S. Bureau of Labor Statistics, All Employees, Local Government [CES9093000001], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/CES9093000001, January 7, 2024.
[15] Ibid.
[16] Center for Just Journalism. (undated). “Retail Theft: What to Know and Where to Go.” Accessed January 7, 2024: https://justjournalism.org/media/download/CJJ_TipSheet_RetailTheft_2022.pdf
[17] Lopez, Ernesto, Robert Boxerman, and Kelsey Cundiff. “Shoplifting Trends: What You Need to Know”. Washington, DC: Council on Criminal Justice. November, 2023.
[18] Criminal Justice Information Services Division Uniform Reporting Program. “2023.0 National-Incident Based Reporting System User Manual.” June 30, 2023. Accessed January 7, 2024: https://le.fbi.gov/file-repository/nibrs-user-manual-063023.pdf/view
[19] Ibid, page 30.
[20] Ibid, page 18.
Logical and persuasive as always.