Understanding 2020's Homicide Spike
This past year, there was an unprecedented homicide spike; over one-third across urban America. These increases followed an upward trend since 2014, the year in which Michael Brown was killed by a police officer in Ferguson, Missouri. However, underpinning the national trend has been the growing proportion of homicides of black men. In 2019, black male homicides were 60 percent more than white male homicides, up from 21 percent more in 2005. By contrast, among women, the ratio remained remarkably stable with black female homicides generally close to 56 percent of white female homicides. Thus, the 2020 spike is not simply an anomaly brought about by the pandemic but a continuation of a persistent trend driven by gun violence in black communities.
There are a number of explanations offered for the 2020 increases that I have earlier explored: the deepening of poverty and deprivation resulting from the pandemic; changes in policing and criminal justice; and an unintended consequence of the increased anger among black youth as a result of the sustained demonstrations against police actions. This essay will attempt to identify the factors which can explain the variation in homicide rate across urban America.
Many social and economic measures are available for the largest 50 cities. Among them, there are eight cities in which the black population share is at least 36 percent. Each had a 2020 homicide rate of at least 27 (per 100,000) while the average across the other 42 cities was only 14. Here are the eight cities with their 2020 homicide rates: Milwaukee (33), Detroit (46), Memphis (47), Philadelphia (31), Atlanta (30), Washington DC (27), Baltimore (57), and Cleveland (50). Only one other city (Kansas City) had a homicide rate above 30 while another two (Louisville and Indianapolis) had rates above 25.
On a range of measures, these eight are quite distinct from the other 42 cities: 62% versus 39% of children live in single-parent households; 46% versus 18% of children live in high-poverty neighborhoods; 16% versus 11% of youth 18-24 years old are disconnected from work or school; and 16% versus 8% live in extreme poverty. These circumstances generate extreme deprivation that may contribute to violent behaviors. Among next largest 25 cities for which I was able to gather homicide data, there were three more with black population shares above 36 percent and high homicide rates: Cincinnati (29), New Orleans (50), and St. Louis (89).
We might gain some insight by looking at all cities that have high poverty rates. Across the 75 cities, the average poverty rate was 17.3 percent, but except for Washington DC, the other 10 cities had rates of at least 21 percent. There are another five cities that have such high poverty rates: Fresno, Houston, Tucson, Tampa, and Toledo. And yet, all five had homicide rates below 14. All of these cities except Toledo have a large Latino population.
The fact that poor Latino neighborhoods have much lower levels of violent behaviors is well documented. Darrell Steffensmeier and colleagues analyzed 232 New York and California census tracts, each was virtually exclusively black, white, or Hispanic. Both black and white census tracts had higher violent crime rates as the levels of census tract poverty increased. However, at each poverty concentration level, the violent crime rate was substantially higher in black than white census tracts. Just as important, there were many Latino census tracts with high poverty concentration levels but relatively low rates of violent crime.
The analysis so far has identified three factors that might help explain the variation in homicide rates across the 50 largest cities: poverty rates and Latino and black population shares. I tested a group of educational, family structure, and employment measures that were available for each of the 50 largest cities. Once the poverty rate was included, there were only two measures that maintained statistically significance: the black employment rate and the share of children living in male single-parent households but not the share living in female single-parent households. Fortunately, unlike almost all the other variables, these two were available for all the 75 cities surveyed.
I used the 2013-17 five-year average of the share of children living in male single-parent households and it was strongly statistically significant. So was the poverty rate, and the Latino population shares, but not the black employment measure. Even after taking these measures into account, a city’s black population share remains strongly statistically significant, suggesting that there are distinct black behaviors that impact on homicide rates.
The most surprising finding was the damaging role of male single-parent households. In a recent survey of the literature Roberta Coles found:
[With] respect to externalizing behavior (e.g., antisocial and violent behavior) and substance use (e.g., cigarette smoking, alcohol, drugs), parental gender effects become more salient, with children of single fathers consistently showing higher levels of both over children of single mothers. … [Older studies] found that having a cohabiting partner in the household, which is more common among single fathers than single mothers, was associated with higher levels of virtually every problematic outcome they measured: poorer conflict resolution skills, substance use, school deviance, antisocial behavior, and lower grades and effort at school.
The evidence that the poverty rate is an important determinant of homicide rates is the only evidence that is important to social justice advocates. Their response to the persistent gun culture has been to reduce illegal gun possession to a misdemeanor and to weaken the ability of police to act. We must get beyond the political debate that exclusively focuses on poverty on the left and policing on the right. While both are important, there is much more to fully understand the rise of gun violence.
The finding that male, not female, single-parent households are associated with violent behaviors should lead to a more nuanced understand of single-parent households. All too often researchers associate adverse outcomes with female single-parent households. This is one situation in which that assumption would have been wrong.
Moreover, it should also cause a reassessment of the view that anti-social behavior stems from a lack of father-figure in the lives of black boys. My study indicates that living in father single-parent households is more toxic; and so does the data on children raised in female single-parent households when there is a male present who is not her child’s father. As a result, the structures in which a father-figure is present, other than in two-parent households, have worse outcomes, on average, than for those children living in female single-parent households with no male present. This is why I recently advocated for changes that would reduce the marriage penalty faced by low-waged single parents and believe policies that support parental cohabitation should be expanded.
Robert Cherry is a recently retired Brooklyn College economics professor and a member of 1776 Unites.