The Demand For Protection and the Persistently High Rates of Gun Violence Among Young Black Males

why is the level of violence in the young black male population so high…The murder rates for black men are roughly 10 times higher than the rate for white men in most years…. and how has this level of violence persisted for so long?..The violence in this population is driven by guns. As we outline below, by the end of our sample, 97 percent of murders of young black males are via guns, a rate substantially larger than the rate for the general population. Surveys spanning nearly forty years suggests that young black males carry guns at high rates primarily as protection against others who are also armed...With so many people carrying weapons, disagreements are settled with deadly consequences…To derive specific predictions from our model, we specialize to the case where the two dimensions of a person’s type follow a bivariate normal distribution. If the average instrumental value of carrying a gun is low, gun carry rates are also low. As that value increases, the equilibrium carry rate shifts dramatically. The problem is that returning to a low-gun equilibrium among non-criminals is difficult. The situation can be understood as a coordination failure. Most non-criminals would be better off if everyone stopped carrying guns, but getting there is difficult without a coordinating device. Hence, gun violence persists.The comparative statics are troubling. Once a high gun carry equilibrium is reached, large drops in the benefits of carrying a gun are necessary to revert to a low gun equilibrium. Furthermore, an important asymmetry prevails. The equilibrium gun carry rate, and (by implication in our model) the level of danger and violence, is much more sensitive to changes in the average instrumental value of a gun than its average defensive value, even though the vast majority who arm do so for perceived protection. Our reading of the existing gun policy research suggests these conclusions conflict with real-world policy options in that it has been easier for policy makers to attempt to change the costs of owning guns for non-criminals than criminals. These results suggest the problem may be particularly difficult to solve through legislative means… The homicide rate has ballooned as the COVID-19 pandemic has progressed. In Figure 1, we showed that murder rates increased by 27 percent from 2019 to 2020. Much of this was due to an increase in gun murders, which rose 34 percent compared to 9.4 percent for non-gun murders. Many urban counties have experienced substantially larger increases including Jefferson County, KY, home to Louisville (81.3 percent),Milwaukee County, WI (72.3 percent), Alameda County, CA, home to Oakland (65.2 percent), and Orleans Parish, LA, home to New Orleans (64.2 percent), just to name a few. The groups with the highest murder rates already also had the greatest change in murder rates between 2019 and 2020. Breaking the population into three age groups (15-24, 25-34, 35 and up), four races (white, black, Asian/Pacific Islander, American Indian/Alaskan Native) and 2 genders, there are 24 groups. The three groups that saw the biggest increase in murder rates were white males 15-24 (36.0 percent), black males 25-34 (38.2 percent), and black males 15-24 (39.7 percent). The preliminary data for 2021 looks even worse. Murder counts for the first six months of 2021 are up 13.9 percent over the same time period in 2020.

William Evans and Maciej Kotowski, "The Demand For Protection and the Persistently High Rates of Gun Violence Among Young Black Males,"National Bureau Of Economic Research, April 2022, https://www.nber.org/papers/w29969

“…In Figure 5, the black dashed line reports the fraction of murders of young black males from 1976 to 2019 where an offender is identified in the SHR. This has declined steadily over time from 80 to about 52 percent. In the same graph, the solid line reports the fraction of murders of young black males with known offender where at least one of the offenders is from this same demographic group. This number starts at 44 percent in 1976 rises to 70 percent in 1994 and has been near 60 percent since 2002. When we increase the age range by another decade, the fraction known offenders that are black males ages 15-34 has been hovering around between 77 and 81 percent since 1992…”

“…Third, during the height of SQF use in New York City, murders of young black males fell dramatically. In Figure 15a, we report in solid lines the murder rate for black males 15-24 and 25-34 from New York City while the dotted lines of the same color represent the same numbers for the rest of the country. The police reforms that started in New York City in 1994 occur at a time when the crack epidemic was receding and murder rates for young black males were falling. This makes it difficult to assign how much of the decline is due to SQF or other factors, like mean reversion. That said, the rates for young black males in the rest of the country retreat to their pre-crack levels and do not decline, whereas the decline is dramatic in New York City. In Figure 15b, we report the time series of murder rates in New York City and the rest of the country for everyone who is not a black male 15-34. Here, there is conflicting data on the impact of the SQF program. A result consistent with those favorable to SQF is that New York City had persistently higher murder rates than the rest of the country and starting in 1994, these murder rates drop quickly and are now below the rest of the country. The percent drop in murder rates in New York City for those that are not black males 15-34 is comparable to the drop for younger black males, the primary targets of SQF. This could mean some other policy was impacting murder rates... The model developed in section 3 suggests that once a high-gun carry equilibrium is obtained, it would take a massive reduction in the protective value of guns to return the world to a low gun equilibrium. There may have been no more targeted policing strategy for gun violence among young black males that SQF in New York City. The numbers above suggest it was especially targeted to young black males, but the low yield of arrests, guns and contraband suggest the program was not targeted to likely criminals but to a much larger group. The increased risk of sanction for carrying aweapon along with improved in community safety reduce the attractiveness of guns as a protective means to non-criminals. If we were to make the generous assumption that all of the decline in murders in New York City were attributable to SQF, then this program illustrates the degree to which jurisdictions must police in order to obtain large reduction in the protective value of guns and hence a large reduction in murders. This conclusion is consistent with the relative rigidity of equilibria with respect to changes in… Large and costly interventions along this dimension are necessary to shift outcomes…

“…For completeness, we report in Figure 3 the same graph for white male. The basic time series patterns are quite similar to those of black males. The startling distinction between Figures 2 and 3 is the scale.The murder rates for black men are roughly 10 times higher than the rate for white men in most years. The time series movement in the murder rate for young black males is driven almost exclusively by gun violence. In Figure 4, we report the murder rate, the gun murder rate and the non-gun murder rate for this group since 1968. The non-gun murder rate has fallen steadily since 1968. In contrast, the overall murder rate essentially follows the gun murder rate. Moreover, the fraction of murders attributable to guns is now much higher than 50 years ago, climbing from about three-quarters of murders in 1968 to 98 percent in 2020..”

“…In Figure 2, we report the national time series from 1968 through 2020 for murder rates among black males in three age groups: 15-24, 25-34, and 35-plus. The murder rate for older black men has fallen from 98.3 in 1972 to a low of 21.2 in 2010, a drop of 78 percent. Since 2010, the murder rate in this group has edged up, but is still 63 percent below its peak and 55 percent lower than the 1968 value. The murder rate for black men in the 25-34 age category has been much more volatile, seeing peaks almost every decade. The long term trend is decidedly negative. Between 1968 and 2020, murder rates fell 26 percent. The peak (176.6 in 1972) to trough (67.7 in 2014) changes represents a 62 percent decline. The numbers for young black males follow a completely different series. Between 1978 and 1994 there was a 137 percent increase in murders, then over the next 21 years, murder rates fell by 60 percent. However, the murder rate in 2019 (77) was at about the same as it was in 1968 (81) and the rate and the rate in 2020 (108) is 33 percent higher than in 1968. This is a staggering difference compared to the rest of the country. The rate in 2020 is 17 times what it is for people not in that demographic. Although young black males are about 1 percent of the population, the fraction of murders in this group represented 17 percent of murders in 1994 and 17 percent in 2020…”

Trends in Murder Rates

“…In Figure 1, we graph the national murder rate as reported by the National Vital StatisticsData (NVSD) from 1968 through 2020.The time series was volatile for the first 23 years with peaks in 1974, 1980, and 1992. The national murder rate declined consistently from 1992 through 2015, with a substantial increase after that, especially in 2020 when murder rates for the nation rose by 27 percent, more than twice the second largest annual percent increase in the murder rate in the series. The rate in 2020 is about the same as it was in 1968 and the number for 2019 was 20 percent lower than the value in 1968. In the same graph, we also report the gun and non-gun murder rates. The variation in murders over time is driven mainly by changes in gun murders. Of the increase between 1968 and 1992, 81 percent was attributable to gun violence and of the decline from 1992 to 2015, 65 percent was due to a drop in gun murders. The murder rate for black males differs fundamentally in level and time series than the rest of the country…”