Memphis and Gun Violence

From Atlantic Cities, by Richard Florida:

Gun violence made its way into the area of professional football this weekend, when Jovan Belcher of the Kansas City Chiefs murdered 22-year-old Kasandra Michelle Perkins, the mother of their 3-month old daughter, in their home. He then turned the gun on himself in front of his coaches and teammates outside Kansas City’s Arrowhead Stadium.

The Daily News placed attention where it needs to be: “After the predictable, shocked rhetoric of death about Jovan Belcher of West Babylon High and the Kansas City Chiefs, you still arrived back at the only place that mattered in the last hour of his life: the gun.”

Gun violence in America has reached epidemic proportions — over 30,000 people died by gun in 2011, according to preliminary data [PDF]. This past summer, dubbed the summer of the gun, these numbers felt all too close and all too random, with shootings near the Empire State Building and in an office in D.C. closely following mass killings in Aurora, Colorado and Oak Creek, Wisconsin.

These events prompted me to examine the geography of gun violence here on Cities this past July, based upon a prior analysis for The Atlantic. Both posts relied upon state-level data for 2008 from the U.S. Centers for Disease Control and Prevention [PDF].

A useful follow-up analysis was recently posted by Harry Moroz over at Next American City. Moroz relied on CDC data from 2006 to 2007 [PDF], which covers the 50 largest metros and their constituent center cities. This data set also breaks out two key types of gun violence — gun-related homicides and suicides — as well as total gun-related deaths. As such, these data provide a more precise take on the geography of gun-related death at the metro and city levels. Moroz writes:

Based on the CDC data, almost 60 percent of U.S. firearm homicides occur in the 62 cities of the country’s 50 largest metros. However, only 27 percent of suicides do. In 2006, firearm suicides were a primarily suburban (and non-central city) phenomenon, which is likely weighing down the relationship between firearm deaths and the city unemployment rate.

Moroz finds a close relationship between city unemployment and murder by gun (and has some nice scatter-graphs to show it). The correlation between city unemployment at the overall rate of gun deaths is considerable (.55), and the correlation between it and gun-related murders is even higher (.72).

Spurred by Moroz’s more detailed analysis and with the help of my Martin Prosperity Institute colleagues Charlotta Mellander and Zara Matheson, I decided to take a closer look at gun deaths at the metro and city levels based on these CDC data.

Map by MPI’s Zara Matheson based on data from Centers for Disease Control [PDF]

The first map above shows the overall rate of gun-related deaths per 100,000 people by metro. The rates vary substantially from a high of 32.8 in New Orleans to a low of 3.6 in Boston. Birmingham has the second highest rate with 20.5, followed by Memphis with 19.8. Las Vegas (17.6) and Jacksonville (17.5) round out the top five metro rates. After Boston, the metros with the lowest rates include San Jose (3.8), followed by Providence (4.1), New York (4.8), and Hartford (4.8).

Table data from Centers for Disease Control [PDF]

Cities have even higher rates of gun death. (While the CDC collects data on gun death by city as well as metro, the city data has missing values for particular kinds of gun-related deaths. So we report only for those cities where complete data are available.)

The rate of gun death for central cities ranges from a high of 69.1 in New Orleans to a low of four in San Jose. Detroit has the second highest rate (41.4) followed by Las Vegas (36.9), Miami (33.5), and Baltimore (33.1). St. Louis, Richmond, Memphis, Cleveland, and Philadelphia — all with rates above 20 gun deaths per 100,000 people — round out the top 10. All have a rate of gun death that is roughly two or more times higher than their metro rate. On the flip side, New York City follows San Jose with second lowest rate (4.9), followed by central San Diego (7.1), and Seattle (8.3).

Map by MPI’s Zara Matheson based on data from Centers for Disease Control [PDF]

The second map shows the pattern for rates of gun-related suicides by metro. The range is again considerable, varying from a high of 11.4 to a low of 1.5 per 100,000 people. (It is important to point out that there are missing values for a number of cities with high overall gun-related deaths, including Newark and Oakland.)

The metro pattern here is different from than that for the rate of gun-related deaths overall. The Las Vegas metro has the highest rate of gun-related suicides (11.4), followed by Nashville (9.7), Birmingham (9.4), Louisville (9.2), Phoenix (9) (which includes the city of Mesa), and New Orleans (8.7). On the other side of the ledger, New York as the lowest rate (1.5), followed by Boston (1.7), Hartford (2.1), San Jose (2.6), Providence (2.6), and Chicago (3.1).

Table data from Centers for Disease Control [PDF]

The range in central cities is also considerable — from a high of 23.4 in Las Vegas to a low of 0.9 in New York City. Las Vegas has the highest rate by far, more than double that of the next city, Tampa, with a rate of 11.1. Aurora, Colorado (10.4), Miami (9.8), Louisville (9.5), and Mesa, Arizona (9.1) round out the top five. Las Vegas and Miami stand out with city suicide rates that are roughly double those of their metros. On the opposite side of the ledger, Washington, D.C., (1.9) follows New York City with the second lowest rate, and then San Jose (2.1), Chicago (2.3), and San Francisco (2.3). Moroz notes that center cities account for just 27 percent of gun-related suicides compared to 60 percent of gun-related homicides, concluding that “firearm suicides were a primarily suburban (and non-central city) phenomenon.”

Map by MPI’s Zara Matheson based on data from Centers for Disease Control [PDF]

This third map charts the rate of gun-related homicides by metro. Again, the variation is considerable, ranging from a high of 24.1 to a low of 1.2 gun deaths per 100,000 people.

New Orleans has the highest rate (24.1), followed by Memphis (11.6), Birmingham (11.1), Baltimore (10.3), Jacksonville (9.4), and Detroit (9.3). On the opposite end of the spectrum, San Jose has the lowest metro rate with 1.2 followed by Portland, Oregon (1.4), Austin (1.5), Providence (1.5), and Minneapolis (1.8).

Table data from Centers for Disease Control [PDF]

Cities have substantially higher rates of murder by gun, as Moroz points out. New Orleans has the highest rate, 62.1 per 100,000, more than twice its metro rate. Detroit has the second highest rate with 35.9, nearly four times its metro rate, followed by Baltimore (29.7), Oakland (26.6), and Newark (25.4). St. Louis, Miami, Richmond, Philadelphia, and Washington, D.C., round out the top 10. All of these rates are considerably higher than their metro rates — in most cases more than double or triple, or in the case of Newark, nearly eight times its metro rate. On the other side of the ledger, Austin has the lowest rate of gun-related homicides (1.5), followed by San Jose (1.9), Portland (2.2), Virginia Beach (2.7), and San Diego (2.8).

We also examined the social or economic characteristics that are associated with gun deaths at the metro level.

Our previous state-level analysis (here and here) found several key factors to be associated with gun deaths overall. With images of mentally deranged killers so prominent in the media, it is frequently assumed that gun violence is a product of mental illness or drug abuse. But we found no association between mental illness, stress, or illegal drug use and gun deaths at the state level. While one would think gun violence would be higher in states with higher levels of economic anxiety related to unemployment or inequality, we found no association to either at the state level. My colleagues and I did, however, find gun deaths to be higher in states with higher levels of poverty and  lower incomes, as well as in red states and those with more blue-collar working class economies. Conversely, we found gun deaths to be less likely in states with more college graduates and stronger knowledge-based economies.

With the help of my colleague, Charlotta Mellander, we did a similar analysis of the metro level data. Mellander ran a basic correlation analysis across the three measures of gun-related deaths and a variety of economic, demographic, and social characteristics of metros. We were unable to do a similar analysis at the city level due to missing data on gun-related deaths and and the lack of readily-accessible matching data for key economic and social variables. As usual, I point out that correlation does not equal causation.

Let me start by pointing out that the overall rate of death by gun is closely related both gun-related suicide (.80) and gun-related homicides (.83). But gun-related suicides and homicides are far less related to one another (with a correlation of .35).

It is commonly thought that gun violence is higher in bigger cities and metros. But that is not what we find. Population size is negatively associated with suicide-related gun deaths (-.46) and not significantly associated with either total gun deaths or murder by gun at the metro level. Population density is not significantly associated murder by gun and negatively and significantly associated with gun-related suicides (-.67) and the overall rate of death by gun (-.46).

Poverty is a substantial factor in gun deaths by metro, as it was in our previous state-level analysis. The percentage of a metro’s population below the poverty line is significantly associated with all three types of gun death — homicide (.45), suicide (.35), and the overall rate (.49).

Inequality has been increasing in America and is a major source of economic anxiety and tension. It is associated (.33) with murder by gun (this differs from the state level where we did not find a statistically significant correlation), but not with gun-related suicides or the overall rate of gun deaths.

Unemployment is a substantial source of economic anxiety, but we find no association between it and any category of gun death at the metro level. This is in line with our earlier finding for states. But it is at odds with Moroz’s analysis that found a close association between unemployment and gun deaths in center cities. Unemployment rates, of course, are typically much higher in center cities than states or metro areas. This is one area where we see clearly different patterns at across these three basic levels of geography.

Economic advantage plays a substantial role in moderating death by gun, at the metro level as it did for states. More affluent metros have lower rates of all forms of gun death. That said, economic advantage — measured as per capita income — plays a bigger role in moderating the overall rate of gun death (-.55) and that for gun-related suicide (-.64) than for gun-related murders (-.32).

Education plays a similar role in moderating gun death, with more highly-educated metros having lower levels of all types of gun death. The share of adults that are college grads is negatively correlated with of gun death overall (-.57), suicides (-.52), and murders (-.46).

Gun death also varies by socio-economic class. Higher levels of knowledge-based, creative class work at the metro level is associated with lower levels of all three types of gun death — overall gun deaths (-. 55), gun-related suicides (-.53), and gun-related homicides (-.39). The same pattern holds for high-tech industry. A metro’s share of high-tech industry is negatively associated with overall gun deaths (-.49), gun-related suicides (-.53), and homicides (-.32). Conversely, metros with higher shares of blue-collar working class jobs experience higher rates of all three, with positive correlations to overall gun deaths (.52), suicides (.49), and murders (.37).

Interestingly, gun deaths at the metro level are related to the share of people who drive alone to work. There are significant positive correlations between the share of commuters who drive to work alone and both the overall rate of gun death (.40) and gun-related suicides (.55). However, there is no statistically significant relationship between gun murders and commuting to work alone. This is line with Moroz’s contention that gun-related suicide is less of a center city and more of a suburban phenomenon. Driving to work alone may reflect the social isolation caused by sprawl.

Criminologists have noted that crime — especially violent crime — is more likely among young males. Interestingly, this does not appear to be the case according to our analysis of metros. In fact, we find a modest negative correlation between the share of young males and the overall rate of gun death (-.10).

Race, unfortunately and tragically, factors into gun death at the metro level. The share of the population that is black is positively related to both the overall rate of gun death (.56) and even more so with gun-related homicides (.72). The pattern is similar for the share of the population that is comprised of young black males which is also positively related to the overall rate of gun death (.55) and murder by gun (.70). That said, we find no significant association between any type of gun death and the share of the population that is Hispanic. The importance of gun control cannot be minimized. The state level is the appropriate level to examine this. And our previous state level analysis found gun deaths to be significantly lower in states with stricter gun control laws. We found substantial negative correlations between the rate of gun deaths and states that ban assault weapons, require trigger locks, and mandate safe storage requirements for guns.

Moroz is right to point out that a nuanced analysis of gun violence in America requires a careful accounting for the socio-economic factors that come into play at different geographic scales. Still, the consistency of our findings across metro and state levels strongly suggest that gun violence is not just the product of troubled or deranged individuals, as is commonly portrayed, but is both associated with and embedded within the economic and social context of places. Whether looking at the state or metro level, we find strikingly consistent associations between gun violence and key markers of socio-economic disadvantage — poverty, income, education, class, and race.  Of course, center cities bear the heaviest concentrations of such socio-economic disadvantage and we are likely to find even stronger associations and more magnified patterns there, as Moroz does for unemployment.

Death by gun clearly reflects the class divides which vex America, being substantially more likely in poorer, less advantaged places. And this concentrated nature of gun violence makes it easier for those in more affluent and sheltered places to ignore its consequences. Yes, our nation is in desperate needs of strategies to bridge its burgeoning class divide, but if we truly care to limit the carnage caused by guns in our society, controlling them is the best place to start.