Skip to main content

High-Social Capital Cities Not Immune to Homicide Spike

The Wall Street Journal recently reported on the homicide spike afflicting major U.S. cities. Its analysis (based on crime data available as of July 27) found that, in addition to a rise in overall shootings, homicides have risen by 24 percent across America’s 50 most-populated cities. In the Journal’s telling, the increase in murders is, in part, a story of core institutions receding amid the coronavirus pandemic:

Schools let out young adults in March because of the pandemic and after-school activities largely stopped. Churches and other social institutions were restrained for the sake of social distancing. Police first were hit by coronavirus and then blowback in the neighborhoods they patrol after the killing in Minneapolis of George Floyd, a Black man, while in police custody.

In The Geography of Social Capital in America, the Social Capital Project cited a long history of research that found a strong negative correlation between violent crime and social capital. Drawing from the work of Robert Sampson and others, the Project views violent crime as a measure of “collective efficacy…[t]he idea is that communities high in social capital are better positioned to informally police their own community and enforce pro-social norms, and their residents are less likely to do harm to one another.”

Is this borne out in the current murder spike seen in cities across America? Using data obtained from the Journal combined with our social capital index for the county in which the bulk of a given city’s population resides, we find a negative overall relationship between major cities’ annualized 2020 homicide rates and their social capital levels. In other words, cities with higher murder rates this year tend to have lower overall social capital scores.

Because a measure of violent crime is included in our overall county-level index, we need to separate out our sub-indices in order to ensure the relationship is not spurious. Indeed, the association between higher homicides rates and lower social capital appears to be driven by cities with more relative family instability. This shouldn’t be surprising. In fact, in our original report, the strongest correlation (0.47) between the Project’s county-level sub-indices is between the family unity and collective efficacy sub-indices. (Our family unity sub-index is a normalized score made up of the share of all births to unmarried women, the percentage of women currently married, and the percentage of children living with single parent. Collective efficacy is measured such that counties with lower violent crime rates have higher—i.e. more positive—collective efficacy scores, and vice versa.)

Murders in 2020 and 'family unity' sub-index in top 50 cities by population
Per 100,000 residents at annualized rate

Note: NYC excluded. Source: SCP Index: Wall Street Journal


The data, however, do not necessarily show that growth in murders is related to lower social capital; there is, if anything, a slight positive relationship between percentage growth in the homicide rate from 2019 to 2020 and both of our family unity and community health sub-indices.

Percentage growth in murders 2019-2020 and 'family unity' sub-index in top 50 cities by population
Growth in murders per 100,000 residents, year to date

Note: NYC excluded. Source: SCP Index: Wall Street Journal


This could be due at least partially to the fact that higher-social capital places generally had lower base murder rates in 2019 and thus stood to experience larger percentage increases in homicides. To illustrate, low-social capital Baltimore (the blue dot furthest left) had a homicide rate of 30.0 per 100,000 in 2019 and its 2020 year-to-date rate is just slightly higher at 30.7 per 100,000, an increase of just 2 percent. Yet Seattle (the second rightmost gray dot), located in social capital-rich King County, WA, had a murder rate of only 1.6 per 100,000 population in 2019, which has risen to 2.3 per 100,000 in 2020 thus far—but this represents a 42 percent increase.

Some caveats to this analysis are in order:

  • Our social capital index is based on county-level, not city-level, measures, so a violent city in an otherwise high social capital county could skew the analysis by showing a positive relationship between high crime and social capital
  • The Journal’s data represent a snapshot in time based on data cities had available as of July 27; most of the data are current to mid- to late-July, but a few cities’ data are current only to May 31
  • New York City is excluded because its crime data is aggregated at the city level, but its social capital data is split across five counties (boroughs)
  • DC is excluded from the community health analysis because it is an outlier due to the numerous non-profits headquartered there
  • The Journal’s data show that other violent crimes such as robberies have fallen even as homicides have risen (the Journal’s report notes this is probably a result of more people staying home due to coronavirus lockdowns)

Caveats aside, the relationship between homicide rates and social capital is one worthy of consideration for social capital-minded policymakers and researchers. Places relatively high in social capital, such as Colorado Springs, Omaha, and Minneapolis, have not been immune to the recent homicide spike, and this further underscores the importance of studying social capital in the context of the pandemic. Perhaps the places with the strongest institutions of civil society and social capital-forming behaviors have the most to lose when these entities and activities are sidelined.

Patrick Brown
Senior Policy Advisor

Vijay Menon
Policy Advisor

Latest News