Crime in the era of COVID-19

The onset of the COVID-19 pandemic had an enormous impact on nearly every aspect of our day-to-day lives, ranging from economic distress, to disrupted schooling, and public health impacts. Relatedly, the pandemic has impacted crime in different ways, but there is still a lot of confusion and disagreement regarding this relationship. On its face, the onset of the pandemic was initially correlated with large drops in many types of crime. However, this finding comes with a caveat: while overall crime rates are lower than they have been in previous years, homicides and shootings are higher than normal, and this trend appears to be continuing into 2021.

Researchers have attempted to unpack the complex relationship between COVID-19 and crime, but doing this is easier said than done — some crimes are up, some are down, some are going unreported, and some have likely displaced to other locations. To better understand the more nuanced patterns, we need to better pinpoint crime pockets and attempt to examine the micro-level factors associated with them. In the research, micro-level factors are things like time of day, geography, environment, street activity, and spatial-temporal characteristics. Considering the wide variation in crime patterns that we have seen during the COVID-19 pandemic, micro-level factors warrant consideration in this conversation.

In the first six months of the pandemic, David Abrams from the University of Pennsylvania examined crime rates for 25 of the largest U.S. cities in comparison to numbers from the previous five years. He documented major drops in overall crime, including drops of at least 35% in major cities such as New York City, Chicago, Washington D.C., Philadelphia, and San Francisco. However, he argued that this crime drop might be masking  a potentially more important finding: a large increase in homicides and shootings, spiking after the death of George Floyd in May 2020. Moreover, decreases in crime at the 6-month mark were driven largely by a decrease in drug crimes (-65%). Given that drug crimes are typically reported directly by the police (rather than citizens), Abrams suggests that the reduction could be related to changes in policing strategy that reduce enforcement of certain crimes. Abrams also explains out how COVID-19 policy restrictions substantially affected the flow and volume of people on the streets, which may increase or decrease crime depending on the situation. For example, things like stay-at-home orders, social distancing, business closures, and decreases in mass transit created dramatic shifts in what is referred to as “population mobility,” i.e., the lifestyles and routine activities of people that live and work in an area.

The National Commission on COVID-19 and Criminal Justice has continued to monitor the situation through the present by examining monthly crime rates for ten violent, property, and drug offenses in 34 U.S. cities, with a special emphasis on homicide and violent crimes. The year-end 2020 report found that robberies, as well as property and drug crimes (with the exception of motor vehicle theft) continued to fall throughout 2020. In comparison to 2019, residential burglaries had decreased by 24%, drug offenses decreased by 30%, larceny decreased by 16%, and robbery decreased by 9%. And while Abrams’ earlier estimates showed an initial increase in commercial burglaries (+24%), this effect faded through the latter half of the year, ultimately falling about 7% from 2019 to 2020. Most non-violent crime rates continued to decrease through the  first quarter of 2021 — residential burglaries, commercial burglaries, larceny, and drug offenses all decreased (by 16%, 7%, 16%, and 24%, respectively) relative to the first quarter of 2020.

In contrast, 2020 saw increases in motor vehicle theft (+13%) among the 34 cities. This trend seemed to continue through the first quarter of 2021, at which time motor vehicle thefts were up by 28% in comparison to the same time period in 2020. After the initial spike in May 2020, homicide rates remained elevated throughout the summer, with increases in homicide rates amounting to 30% on average from 2019-2020. This trend declined at the end of 2020 through March 2021, but early 2021 rates are still higher than the same period from years prior. Specifically, the first quarter of 2021 saw homicide rates that were 24% higher than 2020 and 49% higher than 2019. The same goes for violent crimes such as assaults and gun assaults, which were up 6% and 8% (respectively) during the first quarter of 2021 relative to the first quarter of 2020.

There is a fair degree of contention regarding exactly how COVID-19 impacted the prevalence, reporting, and displacement of  crime. Concurrently, there is fair room for debate as to how increased political polarization, protests, and distrust of the government/police has affected crime throughout the pandemic. Among all of these effects, there is one very important area that impacts all of these: the impact of spatial-temporal and geographical factors on crime, as we briefly mentioned above. Dramatic shifts in lifestyles, routine activities, and access to particular goods/services affects the nature and frequency of interactions between potential victims and motivated offenders, therefore impacting crime.

But shifts in population mobility are unlikely to affect all crime types in the same way. For instance, stay-at-home measures might decrease opportunities for residential burglaries as people are more likely to be home during the day, but opportunities for domestic violence incidents or child abuse might increase for this same reason. Similarly, reductions in street activity might increase opportunities for commercial burglaries if stores are left quiet and unguarded (aka, fewer ‘eyes on the street’). One thing we can do to better understand these patterns is to assess them at the micro-level (using data), and subsequently inform predictive policing and/or surveillance efforts. Beyond looking at the direction and magnitude of larger trends, it helps to study the patterns of crime in terms of time, place, and time/place. This might help us pinpoint  specifically where increases are occurring and/or if they are being displaced to new locations (i.e., crime displacement).

As argued in a recent report by Nicole Gelinas of the Manhattan Institute, it seems that violent crime has been spotted specifically in the New York subway system relative to other parts of the city. The report explains how the finding is correlated with lower ridership, which may be explained by rational choice theory: when ridership is lower, there is increased opportunity to commit crime without fear of witnesses. Another study done in New York City by Christopher Hermann, Andrew Maroko, and Travis Taniguchi (albeit pre-COVID) supports this idea. In that study, researchers examined the impact of subway ridership on robbery rates, finding that robberies tended to occur within closer proximity of subway stations and were more likely to decrease during the times/days that subway stations were closed.

Some approaches that police might consider are hot-spots policing, intelligence-led policing, or third-party policing. Hot-spots policing is a policing strategy that incorporates place-based and time-variant factors to predict the approximate time of day and area where crime occurs (i.e., a ‘hot spot’). The idea is that if we can predict where crime is likely to occur, police can be more proactive in preventing crime. Intelligence-led policing is similar to hot-spots in that it is data-driven, but it focuses less on place/time and more on social networks to predict what types of offenders are more likely to commit certain types of crime. Third-party policing is another type of intelligence-led policing that draws on the social control mechanisms held by other government and community actors — such as subway workers, for example. These community actors can then act as “eyes and ears” for the police and also impose their own legal levers to help control crime. All of these approaches have been evaluated to some extent with promising effects. For more information on the effectiveness of various policing strategies, see, where results from various studies are collated, reviewed, and given ratings regarding their efficacy.

While data-driven policing strategies are not completely without unintended effects (e.g. the potential for crime displacement or enforcement disparities), research suggests they are promising in preventing crime. Unfortunately though, we have seen less police proactivity in the recent months which has been exacerbated by COVID-19, and policing remains a contentious topic in the political arena. However, it is likely that the police could do their job better with more resources, particularly if they were dedicated to data-driven policing.