In Real Clear Policy: “Murder Isn’t a Nationwide Problem”

May 3, 2017 | Featured

Dr. John Lott has an op-ed in Real Clear Politics today on our work on the concentration of murder and gun ownership in the US.

The vast majority of murders in the United States occur in just a tiny percentage of counties. In fact, the country can be divided up into three types of places: those where there are no murders; those where there are a few murders; and those where murders are very common.

In 2014, the most recent year that a county level breakdown is available, 54 percent of counties (with 11 percent of the population) had no murders. 69 percent of counties had no more than one murder, and about 20 percent of the population and only 4 percent of all murders in the country.

The worst 1 percent of counties have 19 percent of the population and 37 percent of the murders in 2014. The worst 2 percent of counties contain 47 percent of the population and accounted for 51 percent of the murders. 68 percent of the murders occurred in only 5 percent of counties.

Perhaps surprisingly, murders used to be even more concentrated. From 1977 to 2000, on average 73 percent of counties in any given year had zero murders. This change may be a result of the opioid epidemic’s spread to more rural areas. But no one has yet shown clearly what has caused this change. . . .

According to a 2013 PEW Research Center survey, the household gun ownership rate in rural areas was 111 percent greater than in urban areas. Suburban households are 28.6 percent more likely to own guns than urban households. Despite lower gun ownership, urban areas experience much higher murder rates. One should not put much weight on this purely “cross-sectional” evidence at one point in time. But it is hard to overlook the fact that so much of the country has both very high gun ownership rates and few, if any, murders. . . .

The entire piece is available here.



  1. Brett Bellmore

    The way you’re binning this confuses matters, because the worst 2% of counties includes the worst 1%, which blurs just how fast murder rates drop off once you get out of the hot spots.

    You should do this, worst 1%, second worst 1%, third worst 1%… If you do this, it becomes clear that the worst 1% have a murder rate *four times higher* than the second worst 1%.

    The worse 1% really is insanely bad, and even doing it at the county level conceals how concentrated the problem is.

    • johnrlott

      There is a graph that does that. Isn’t that figure 2?

  2. Michael W. Perry

    How about politics? Is there a correlation between the dominant party for the local government and local murder rates? That doesn’t necessarily mean causation, i.e. having a Democratic mayor causes more murders, but it could be an indication of an ineffective response, i.e. Democratic mayors don’t know how to lower their city’s murder rates.
    One illustration of the latter would be a mayor’s obsession with an ineffective response, say strict gun control, and hostility to an effective response, say aggressive policing of even minor crimes.

    • geokstr

      Nearly all the major metropolitan areas (which is where these murders occur) have had complete and unbroken Democrat control from 50-100 years. That’s not just the mayor but the city council and all the bureau heads, too.

      And it’s not just the violence either – there’s a strong overlap with poverty, welfare, rotten school systems, illegal immigration, drug abuse and political corruption of all stripes.

      But how can this be? The Marxis…oops, sorry, the Democrats, are the party of love and peace and compassion and caring,and all the rainbows and unicorns are belong to them, right?


    • nuthinmuffin

      Is there a correlation between the dominant party for the local government and local murder rates

      • nuthinmuffin

        isn’t that a given?

    • Ken

      The correlation weakly correlates with politics. It strongly correlates with race.

  3. bobby b

    “The worst 2 percent of counties contain 47 percent of the population and accounted for 51 percent of the murders. ”

    This is a dead-even correlation in terms of murders per person.

    The only thing this stat tells us is that our counties differ greatly in number of residents.

  4. HaqJaqShellaq

    This also looks like it’s correlated with population density.

    But what the heck does opioid abuse have to do with murder? I could see it correlating with suicide, but murder?

    • johnrlott

      Thanks, HaqJaqShellaq, though as we note that the level of murders in these areas is significantly greater than their share of the population.

  5. www.Linux.Ca

    I will immediately grasp your rss feed as I can’t to find your e-mail subscription hyperlink or newsletter service.
    Do yyou have any? Kindly let me recognize in order thaqt I may subscribe.

  6. wooden pet gates

    The townspeople have done horrible things.


  1. News of the Week (May 14th, 2017) | The Political Hat - […] Murder isn’t a nationwide problem The vast majority of murders in the United States occur in just a tiny…