17 January 2016

Denver Murders In 2015 By Race And Ethnicity

Denver Homicides per the FBI for years 1985-2014 and per the Denver Police Department for 2015, via the Denver Post.  Note that this understates the extent to which homicide rates have declined in Denver in that time period because Denver's population increased by 31% from 1985-2014 and continued to grow from 2014 to 2015.  For example, it would take about 117 murders in 2015 to match the murder rate implied by the 90 murders in Denver in 1986.  The 50 murders experienced by Denver in 2015 correspond to a murder rate that would have produced 38 murders in 1985.

Murders In Denver In 2015

In 2015, there were 50 homicides not committed by police (police killed 7 people) in Denver, 28 black, 12 Latino and 10 white. As of 2014, Denver was 53.4% non-Hispanic white, 10.4% black, and 30.8% Hispanic (the remainder consists of Asians, Pacific Islanders, Alaska Natives, Native Americans and mixed race individuals who are non-Hispanic).

The murder rate per 100,000 people by race in Denver in 2015 was as follows:

* Non-Hispanic White 2.82
* Latino 5.87
* Black 41.35

There were no murders in other racial or ethnic categories in 2015, although some individuals classified as black or Latino in Denver's murder statistics may have been classified as having more than one race in census statistics.

The lack of murders in some of these other racial categories isn't too surprising, because the numbers of people in racial categories like Native American (2.0%) is small enough that one would not necessarily expect to see a murder in those racial categories every year in Denver, particularly if those racial categories have murders rates closer to non-Hispanic whites or Latinos than the blacks.

Small numbers in absolute terms, such as the number of murders in the City and County of Denver in a given year, are inherently subject to greater statistical fluctuations than large numbers, such as the number of homicides in the entire United States in a given year, due to the law of averages.

UPDATE January 18, 2016: Of the seven killed by police six were shot (several by non-Denver police officers) and one died in the jail. Four were Hispanic, two were black and one was Native American. Those deaths are not included in the murder rate, but including them would further broaden the racial disparity.

National data on deaths at the hands of police also indicate a racial disparity with almost exactly half of the victims being white based upon a Guardian newspaper survey of 776 such deaths in the U.S. in part of 2015, a figure which is much smaller than the percentage of Americans who are non-Hispanic white (96% of the victims were male). END UPDATE

The Larger Context

Some Minor Caveats

Of course, central cities like Denver almost always have crimes rates that are high relative to their resident population, in part, because the number of people in a central city is almost always higher than its residential population during the working day and during the evening and weekend hours when people come to central cities for a wide variety of entertainment, sports, dining and shopping opportunities. In contrast, many suburban areas have populations that are below their resident populations during a large share of all waking hours during an average week.

The distinction between resident population and the average number of people in the city during the waking hours is a particularly important factor, for example, in the high homicide rate of Las Vegas, Nevada, whose average population of non-resident tourists is very large relative to its resident population, even before considering any other factors at play in its high homicide rate.

A not insignificant share of the urban-suburban disparity between homicide rates would disappear if this adjustment were made, although this adjustment would still not explain all, or even most of the difference.

Historically, but less so recently, central cities have also been home to an outsized share of the deeply poor who are much more likely to be involved in gangs and commit serious crimes for reasons that are much more complex.

Race Greatly Influences Homicide Rates

In the entire United States from 2010 through 2012, according to polling and statistics site 538:
[T]he annual rate of homicide deaths among non-Hispanic white Americans was 2.5 per 100,000 persons, meaning that about one in every 40,000 white Americans is a homicide victim each year. By comparison, the rate of homicide deaths among non-Hispanic black Americans is 19.4 per 100,000 persons, or about 1 in 5,000 people per year.
Denver was above both of those averages in 2015, by about 12% in the case of non-Hispanic whites, and by about 100% in the case of blacks.  Since the base number of homicides in a single city is so small, the variability of this number is large in percentage terms, but the overall trend of homicide rates that vary dramatically by the race of the victim, is very robust.

The unfortunately reality, if you do linear regression comparisons of murder rates in various geographic areas to all manner of plausible statistics about those geographic areas is that the racial makeup of a geographic areas is the single best first order predictor of it murder rate when compared to comparable geographic areas (e.g. states to states, cities to cities, etc.). Race predicts homicide rates much more accurately, for example, than poverty rates or education level or the percentage of undocumented immigrants in an area.  Other measures like gun ownership rates or gun control laws in force, death penalty laws in force, or even police staffing levels, are similarly second or third order effects in predicting the homicide rate, at most.

Denver's murder rate in 2015 was still below average for the 30 largest cities in the United States, which isn't terribly surprising, given the longer term historical trends and given Denver's racial composition compared to many of the other cities on that list.

The murder rate experienced by blacks in Denver, for example, is roughly the same as the murder rate in Detroit, which is 83% black and 8% non-Hispanic white.  This is a quite unflattering comparison, considering that Denver, unlike Detroit, does not have a massively understaffed police department, does not have huge tracts of abandoned housing in its high crime areas (about 28% of housing units in Detroit are unoccupied), and has not experienced decades of economic decline.

The very low homicide rate in Seattle, Washington, in contrast (about 4.1 per 100,000), in significant part reflects Seattle's racial composition (7.9% black, 66.3% non-Hispanic white and 13.8% Asian).

Of course, that doesn't mean that racial composition is a terribly accurate predictor of murder rates. For example, Baltimore, Maryland has a much higher homicide rate (about 55 per 100,000) than Detroit (about 40 per 100,000), despite having that fact that it is 63% black and 28% non-Hispanic white (a significantly smaller percentage of black residents and larger percentage of non-Hispanic white residents than Detroit).  One factor that might lead to the distinction between Baltimore and Detroit could be that Baltimore is a metropolitan area in which a far larger share of middle class blacks than most metropolitan areas live in the suburbs rather than in the central city and a high cost of living, leaving the central central with an African-American population that has much higher rates of poverty that in turn lead to crime, but almost any explanation is to some extent speculation.

Incidentally, Seattle (pop. 668,342), Detroit (pop. 680,250), Baltimore (pop. 622,793) and Denver (pop. 663,832) are all the central cities of their urban areas with roughly similar populations in 2014 according to the Census Bureau.

Why Does A Victim's Race So Strongly Influence The Homicide Rate?

The Denver Post story on 2015 murders linked above didn't discuss the race of the homicide suspects, appropriately, because the perpetrator is unknown in 44% of those murders. But, nationally, over many years, the percentage of murders that are intraracial is very high, about 91% of blacks who are murdered are murdered by a black perpetrator, and about 84% of whites who are murdered are murdered by a white perpetrator.

This largely reflects segregation both socially and residentially.  Most murders are committed by acquaintances and family members, and most murders of strangers are committed by people who don't live very far from each other. For example:
"Because of the racial homogeneity of most neighborhoods, moreover, it is even true that most stranger killings are intraracial — 67 percent for white victims and 89 percent for blacks" based on data between 2000 and 2009, Fox wrote in his book The Will to Kill: Making Sense of Senseless Murder.
Put another way, the high murder rate for blacks in Denver and nationwide, is largely a function of the fact that murderers are disproportionately black, and that black Americans are much more likely to have acquaintances, family members and neighbors who are also blacks.

The Role of Gang Homicides

Another important factor, which the statistics quoted above don't demonstrate, but which appears to be true from other data is that a disproportionately share of variability in homicide rates from place to place, and from year to year, is attributable to differences in the number of gang related killings by young minority men (mostly black) by other young minority men (mostly black). It is no coincidence, for example, that 90% of murder victims in 2015 in Denver were men, and that the number of murders among people 21-30 was much higher than any other age range.

According to the Denver Post story linked above, just under half of murders in Denver were gang related, and a post at this blog from April of 2015 identified some reasons why gang related violence might be on the upswing in Denver.  In a nutshell, many historically black neighborhoods in Northeast Denver have seen the black middle class depart to the suburbs, and gentrification has pushed gangs that remain in Northeast Denver to compete for control of a dwindling territory of low income black neighborhoods.

Gang related killings are disproportionately likely to go unsolved. According to the Denver Post story linked above:
Of the 50 homicides, arrests have been made in 28 cases, a 56 percent closure rate, said Sonny Jackson, a police department spokesman. The closure rate drops when looking at the gang-related killings with suspects charged in only 40 percent of those cases.
In contrast, more than two-thirds of killings that are not gang related are closed.

In 2007, the head of Colorado's prison system estimated that 40% of its inmates were gang members.

Why Does Race Have Such An Outsized Impact On Homicide Rates?

Of course, that doesn't answer the $64,000 question of why murderers are disproportionately black, and any answer to that question is almost certainly complex and a full answer is certainly beyond the scope of this post.

An important clue, however, can be found in the fact that the rate at which non-Hispanic whites commit and are victims of murder is higher in the American South than it is elsewhere in the United States.  This fact is particularly notable considering that virtually all African-Americans outside the American South are migrants from that region within the last two or three generations, with much of that time period marked by intense de jure or de facto racial segregation.  This relatively recent history of migration, followed by prolonged segregation, has allowed the culture and regional dialect of the South to remain relatively undiluted in these communities compared to culture and dialect of migrants to other regions who are immersed in populations of natives of the region.

In the same vein, children of Korean immigrants from Buffalo, New York are generally much more assimilated into American culture than children of Korean immigrants from Los Angeles, California, where the Korean community has enough critical mass to resist assimilation into the larger American culture.  The more residentially segregated a community is, the less its members need to assimilate into the larger community where they live.

Anthropologists have described the cultural traits of the American South associated with high rates of violence both in the United States and in regions with similarly cultural values around the world as a "culture of honor."

Another clue that suggests that culture, rather than race (and by association, genetics), per se, is a key factor is that the murder rates now seen in African-American communities in the United States, are quite similar to those seen in early modern London and Amsterdam, for example, that have since seen dramatic declines in their murder rates rates to near global lows, despite experiencing only very modest tweaks to their racial composition during the time period when the murder rates declined.

Racism in the criminal justice system may account for a significant share of the disparity, but it is unlikely that a large percentage of the difference arises from disparities in homicide prosecutions per se.  There is no doubt that blacks are disproportionately victims of wrongful criminal convictions. But, the best evidence is that less than 10% convictions following a criminal trial are wrongful and that the percentage of people who plead guilty of a serious crime despite being factually innocent of that crime or another similar but perhaps less serious crime, is even smaller.  This percentage is not large enough to account for the roughly 100 times greater disparity in homicide conviction rates between African Americans and non-Hispanic whites, and likewise can't explain the huge disparity in homicide victimization rates by race.

Instead, if racism in the criminal justice system gives rise to some of the disparity, it is likely due to racial disparities in enforcement rates for lesser offenses (for example, African Americans are much more likely to be prosecuted for using marijuana than non-Hispanic whites), which taint young African-Americans (especially men) with criminal records and time spent in jails and prisons.  This, in turn, reduces their economic prospects in legitimate endeavors and also foster associations with criminal gangs, socializes them with experienced criminals and develops social expectations that this is their future as well.  It is much harder, however, to quantify the impact of this kind of racism as the amount of information needed to evaluate the impact of each individual offender's prior experiences is huge.  Basically, racism in the criminal justice system may facilitating the networking that is critical to the perpetuation of criminal gangs.  The latest catch phase to describe this phenomena has been "the school to prison pipeline."

Put another way, gangs may be more prevalent in African-American communities than they are in white communities, because few whites have enough exposure to the criminal justice system at a young age together with other youths who are their neighbors, to have the critical mass of convicted young criminals to make gang formation viable.

Falling into the habit of committing serious crimes is much more likely to happen if you are part of a criminal gang that socially supports that conduct than if you are not part of a gang.

No comments: