Crime in American Cities: What the data says

Most often when we think of crime, especially crime in major cities, we think of what makes the evening news, i.e. violent and sensational murders. In reality, violent crime and murders make up nowhere near the majority of crime. For the past 13 years, in the top 75 largest American cities the property crime rate has been roughly five times that of the violent crime rate.

The murder rate is so small in comparison to other crime. The violent crime rate is roughly 70 to 90 times the murder rate. More concerning, while there’s a great disparity among murder rates and violent crime rates, the same is not true of property crime. For example, the city with highest murder rate is St. Louis at 60.94 per 100,000 residents and the lowest is Virginia Beach at 1.55 per 100,000. The highest violent crime rate is in Detroit at 2,007.82 and the lowest is in Irvine at 55.55. In contrast, the city with the highest property crime rate is Memphis at 6,405.6, only a little over five times that of the lowest rate in Irvine at 1,270.26.

Property crime does have a predictor: income. When graphing the murder and violent crime rates in comparison to income there is very little correlation. For example the murder rate spikes in 2016, a year that saw no change in the pattern of a slight increase of income per capita. The property crime rate, however, has been steadily decreasing with the increase in income. This makes logical sense; a lack of resources is an obvious reason to commit theft.

Given that property crime is the most common kind of crime and more predictable than violent crime, it would make sense for the size of police departments to correlate with property crime. Unfortunately, it does not.

There is a correlation between police officers per 10,000 residents and crime rates per 100,000 that suggests violent crime has a stronger effect on the size of the police department. Running a correlation test on the policing and crime numbers for 2018, violent crime has a correlation coefficient of .55 indicating a moderately positive relationship between the number of police officers in a major US city and the rate of violent crime. Whereas property crime has a correlation coefficient of half of that, indicating a weak relationship with the number of police.

This pattern can be difficult to see when looking at individual cities and changes in policing over time. For example, New York-- the most populous city in the United States which has one of the lowest rates of property crime-- saw a spike in policing from 2014 to 2018 when both property and violent crime declined. However, 2012 saw a spike in violent crime with violent crime increasing by 15 percent since 2009. In comparison property crime increased by 1.8 percent. The spike in policing was likely in response to the spike in violent crime not property crime.

The inverse can be seen in San Francisco, a city with a high property crime rate and a below average violent crime rate. Policing has steadily been decreasing since 2007 and the property crime rate has spiked repeatedly in recent years. At its peak in 2017, the property crime rate was 56 percent higher than it was just seven years earlier, yet policing had decreased by three percent in the same period.

There are multiple possible reasons why this correlation could be the case. Perhaps violent crimes require more manpower. But it could also be that increases in violent crime lead to increases in demands for policing and local officials running on platforms of “being tough on crime.”

Elsie Eigerman is a Data Research Intern at Truth in Accounting, a nonprofit government finance watchdog organization in Chicago, Ill.