The Seasonality of Crime

This blog is a part of our new series, "Perspectives in Crime". We'll be exploring leading academic studies that touch on crime data.

The seasonality of crime is an often referred to and taken for granted aspect of modern criminology. The relation of crime rates to the seasons, and analysis into the various cofactors that might exacerbate or attenuate crime at a specific time of year, has been a topic of statistical analysis and academic interest since the 1800s.

The history of seasonal crime studies

In 1831, Adolphe Quetelet, a Belgian statistician, first brought attention to seasonal patterns in crime. In 1842 he proposed a “thermic law of delinquency,” stating that violent crimes were more common in hotter climates and seasons, and that crimes against property were more frequent in colder climates and seasons. Quetelet is also known for initially developing the concept of the “average man” to determine a sort of statistical baseline for human attributes and behavior and thereby document exceptions from the norm. He is also one of the first quantitative social scientists to observe the propensity of varying age groups to commit crime. Quetelet and many other 19th century statisticians and social scientists mapped crime across the U.S. and western Europe, and began what was referred to as “moral science” which helped establish the infrastructure of modern criminology.

The academic consensus on the seasonality of crime is that summer months produce more violent crimes and that winter months produce more property crime. Among a series of explorations throughout the late 19th and early 20th century, the general argument held that warmer temperatures and a greater degree of public life increase the incidence of crimes against other persons, whereas the hardship of winter is a greater drive towards property crimes. Longer periods of darkness are also held to provide greater windows of opportunity for burglary.

Modern data shows crime seasonality

Modern data and scholarship have moved in a different direction, positing that property crime also increases in the summer months to a greater degree than in the winters. Criminologists, social scientists, and statisticians argue that analysis and understanding of crime seasonality can be an invaluable resource to firms and communities providing guidance on the allocation of guardianship resources in anticipating criminal threats. (Falk, 1952; Block, 1984)

At the national level, both violent crime and property crime are strongly seasonal, with Pinkerton data scientists finding a 12% difference between peak and trough in property crime risk. Figure 1 shows the seasonal behavior of both violent and property crime risk as compiled from Pinkerton Crime Index data. The Pinkerton Crime Index is built from continuously sourced data from federal, state, and local law enforcement agencies, as well as crime victimization surveys, among other sources.

While the seasonal behavior of crime is evident nationally, some communities display stronger seasonal crime swings than others. Cities with a larger temperature change over the course of the year, like Boston and Chicago, consistently show a strong seasonal effect where some warmer cities, like Los Angeles, show less pronounced seasonal effects. While Los Angeles does not experience the same annual swing in temperature as northern cities, it is still subject to changes in activities and thereby criminal opportunities. Key among the activities that appear to influence seasonal variation in crime, are no school during the summer and substantial summer tourism, providing a boost in population of potential offenders and potential victims alike.

Figures 2 and 3 show the seasonal oscillation of crime in Cook County (Chicago), Illinois and Los Angeles County, California. In both cases the pattern of seasonality is visually apparent, but it is more pronounced in Cook County.

Hipp et al., (2004), studied fluctuations in crime from 1990-92 across a majority of American communities (96% of the population), testing the seasonality hypotheses of Temperature/Aggression and Routine Activities against the data collected. Temperature/Aggression theory and Routine Activities theory are the two most common causal mechanisms put forth by researchers to explain the seasonal change of crime rates.

Additional theories about crime seasonality

Taking from Quetelet, the Temperature/Aggression theory holds that hot temperatures create greater aggression and reduce inhibitions, causing a greater degree of violent and destructive behaviors, both interpersonally and with regards to property. According to this understanding, the characteristic heat of summer is primarily to blame for the rise in criminal behavior. This logic implies that places where the temperature grows hotter should become more violent, all else held equal.

Routine Activities is an opportunity-based theory of crime, holding that crime occurs when a motivated offender finds a suitable target in the absence of capable guardianship. When applied to the problem of seasonality, Routine Activities theory places importance on the movements and congregation of populations. This mechanism suggests that the summer months witness a surge of crime largely because of an increase in public activities. Factors such as tourism, summer vacation from school, and populous downtown areas all provide a greater intermingling of would-be criminals and potential victims and proponents of the theory understand this as the primary driver in seasonal crime fluctuations. 

Testing statistical coincidence of temperature and crime

Focusing on both within community short-term seasonal fluctuations and between community differences in crime rates, Hipp et al (2004) built multiple statistical tests of the expected behavior of crime given the causal mechanisms of Temperature/Aggression and Routine Activities. Leveraging the FBI’s Uniform Crime Reporting data and temperature data from the National Climatic Data Center researchers were able to build a dataset to observe the statistical coincidence of temperature and crime. 

Among their results, they found frequent and strong support for Routine Activities theory in driving seasonal crime changes regarding both Violent and Property crime types, where a confluence of new crime opportunities were created in warmer months. This manifests both in the greater copresence of potential offenders and victims in public gathering, and also in the decline of guardianship at home as people spend more time outdoors, creating new opportunities for commission of property crimes. While Temperature/Aggression models successfully predicted the rise of violent crime, they did not perform well in predicting the oscillations of property crime.

Temperature/Aggression theory also posits that communities of greater population density are more susceptible to seasonal changes in violent crime. The mechanisms of greater aggression and lowered inhibitions should increase the rate of violence in more densely populated areas that join offenders and victims in space. Hipp et al (2014) found no statistical support for this proposition in data analysis. Conversely, Routine Activities theories suggest that communities with a higher distribution of entertainment venues would experience greater crime, as these venues congregate populations and produce more criminal opportunities. Summarizing the statistical evidence, Hipp et al (2004) note: “adding 32 of these establishments per 100,000 population increases violent crime almost 5% and property crime nearly 23%... increas[ing] the seasonal oscillation of violent crime 6.7% and property crime 11.1%.”

While crime seasonality is among the oldest notions in scientific criminology, modern scholarship has advanced understanding of the mechanisms that underwrite the seasonal behavior of crime rates, pointing us in the direction of controllable factors like the strengthening of guardianship in summer months with changes in routine behavior that increase target suitability for motivated offenders.

Curious about crime risk in your area? Check out the Pinkerton Crime Index to learn more.