High Gun Death Perspectives

Whenever you see breathless reporting of the same core story across multiple media outlets, odds are there is more to the tale. In this case, the recently recirculating headlines noting that “Gun deaths at highest level in 40 years, CDC says” and “More than 100 people killed every day in the US as gun deaths spiral to highest level in 40 years.”

Naturally, we at the Gun Facts project had to dig below the headlines to see what the heck was going on.

The big take-aways

  • Firearm suicides are up, but in the key group, the rate of suffocation suicides is growing fastest
  • Homicides are up, but astoundingly so for young blacks
  • The upward trends started at different times and for different reasons

First, let’s take the long view

As is well known – because this has been rehashed more often than the corned beef at the local diner – the two primary causes of someone dying from a gun are suicides and homicides, in that order. The other categories – accidental deaths, legal interventions, etc. – have such small rates of occurrence that for today, we will simply ignore them … especialy given the steep multi-decade decline in accidental gun deaths.

Firearm Homicides and Suicides - United States - 1981 thru 2016To get past media hyperventilation, we first turn to the source data cited in the report, namely the Center for Disease Control (CDC) mortality databases. We do indeed see upticks in firearm suicides and homicides, but they are not covariant. Suicide rates started climbing with the onset of the Great Recession, triggered by the American housing crash crisis. This makes sense because suicide rates rise whenever there are severe psychological stresses. Suddenly being out of work and homeless certainly qualifies. It will be interesting to see if recent economic revitalization forces the suicide trendline back down in the coming years.

Firearm homicides? Well, that topic is both understandable and speculative. American firearm homicide rates started rising in 2014 – six years into the housing crisis. According to some reports 1 2014 was likely the inflection year for Latin American drug cartels shifting their production from marijuana (rapidly legalizing in the U.S. and thus becoming unprofitable to smugglers) and cocaine to opioids. In the following few years, old orders within the street drug distribution industry were disrupted and turf battles for new, rapidly growing and evolving distribution chains erupted.

Summarizing this a bit: people depressed over economic woes started killing themselves in 2008 and street-level drug dealers started killing each other in 2014.

Homicides: Age, race and not-necessarily-metro

The more interesting of the two trends is the sudden and steep rise of firearm homicides starting around 2014. This is roughly when both Chinese and Mexican traffickers started making an opioid called fentanyl. It was sold in raw form as well as being added to heroin to cut costs while boosting highs.

Based on past trends with drugs and street violence, we assumed most of the new firearm homicides would be inner-city related as they had in the past with marijuana, cocaine, crack, etc. What we found were predictable manifestations concerning the race of firearm homicide victims (statistically poor, young and urban, and thus more highly associated with street gangs and retail drug trade).

Well, we got two out of three right.

Slope –
pre 2014
Slope – 2014-2016
15-24 years -0.16 1.72
25-34 years 0.09 1.58
35-44 years 0.09 1.52
45-54 years -0.01 0.75
55-64 years 0.01 0.57
65-74 years -0.01 0.07
SOURCE: CDC WONDER database, rates averaged, firearm homicides

Gangs and drugs are a young man’s game. When we look at years before and after, we see that before 2014, most age groups were holding relatively steady in terms of firearm homicides. But in the years following, the lower age groups had rather meteoric rises. It is the before/after contrast that is important. The 15-year trend before 2014 showed small increases in firearm homicides and even some decreases, depending on the age bracket. But starting in 2014 all non-child age brackets went up, and the bands where most street gang members come from – ages 15 to 34 – rose twice as rapidly.

Black White Hispanic Average
Large Central Metro 0.99 0.11 0.37 0.49
Large Fringe Metro 1.08 0.11 0.28 0.49
Medium metro -0.52 0.16 0.17 -0.06
Micropolitan 1.21 0.12 -0.15 0.39
Small metro 1.56 0.17 0.1 0.61
Nonmetro 1.61 0.15 -0.14 0.54
Average 0.99 0.14 0.11
SOURCE: CDC WONDER database, rates averaged, firearm homicides, 2014-2017

Likewise, the firearm homicide spike followed recurring race lines as well. What was not following long-term trends was locality. With the exception of micropolitan regions (an urban area with a population of at least 10,000 but less than 50,000), firearm homicides rose almost uniformly, with less urban areas taking the lead. Along race lines, there was a noticeable increase among urban Hispanics but a solid jump among blacks across all degrees of urbanization.

The suicide slopes

Before the Great Recession, suicide rates were either stable or declining among all races except for Native Americans, and for all age groups. Starting with the housing collapse, suicide rates for blacks and Hispanics did not jump skywards, but for middle aged whites it did. Significantly.

This is sadly predictable. People who have the most to lose tend to take it the hardest. During the Great Depression of the 1920s, it wasn’t the janitor who jumped out of the skyscraper – it was the well-heeled stock trader who had lost everything. Likewise, the housing bubble burst led more middle-aged whites to exit from this world.

Average Firearm Suicide Slope Average Annual Suffocation Slope
White 0.20 0.26
Black 0.09 0.06
Hispanic 0.06 0.14
SOURCE: CDC WONDER database, rates averaged, ages 25-54

None of that is interesting. What is interesting is how media hyperventilation concerning “firearm deaths” masked this as the relative reality of the many modes of suicide. Among those ages where income and status mattered, we see a post-crash suicide rates for suffocation rising faster than the suicide rates for firearms (the rate of other modes of suicide are small enough to ignore). What this means is that the rise in firearm suicides from 2008 to 2017 lagged the overall suicide rate for the part of the population that had the highest increase. The breathless headlines didn’t bother to look for, much less show, the underlying cause.

Media and math

I don’t expect the average overworked reporter at an American news outlet to know this stuff. Though not arcane, it is neither in their area of expertise or within their deadlines. It is important though that they learn, then turn to people who can cut through the muck for them (like you). This will help slow public stampeding over horrifying and misleading headlines.


  1. How Did We Get Here? Heroin and Fentanyl Trafficking Trends A Law Enforcement Perspective, Dismukes, June 2018


High Gun Death Perspectives — 4 Comments

  1. What counts as suffocation? Hanging? Monoxide/exhaust?

    Nobody jumps anymore? That was the big Canadian substitution.

  2. Hi, We had corresponded several months ago on ratio of total deaths to suicide and other causes. Suicide has a clear BATFE stat, and I had “posited” that gang/drug/related might be another 30-35% of the total (covering 90-95% of the total US gun deaths when including suicide). You responded well that BATFE does not categorize homicide as gang/drug associated – shared the actual BATFE categorization info – appreciated.

    Am still on that question of the share of US homicide as drug/gang related – has your recent work highlighted in this post shed light? or any other thoughts or conclusions along that line?



    • “has your recent work highlighted in this post shed light?”

      Not to my satisfaction. It still goes to the root problem that the cause has to be definitive. A detective showing up well after the fact, and often with conflicting, unreliable or non-existant witnesses cannot make a definitive determination of cause. With out this, it gets classified as “Unknown” which (IIRC) constitute about 40% of the UCR stats for homicide circumstances.

Leave a Reply

Your email address will not be published. Required fields are marked *