The Washington PostDemocracy Dies in Darkness

Why Do Some States Have More Gun Violence Than Others?


Gun violence has been on the rise in the US. After decreasing by roughly 1% annually from 2006 to 2014, firearm homicide rates increased by 6% from 2014 to 2019. But the increase was far from uniform across the 50 states.

A team of researchers at RAND, led by economist Rosanna Smart, analyzed firearm-related homicides from 2006 through 2019. They found that states with high firearm homicide rates in 2014 also had the largest relative increases in firearm homicides after 2014. The Midwest and South-Central US were regional hot spots, experiencing increases of 75% to 115% after 2014. Missouri had the highest relative increase, followed by Alaska, New Mexico, Kentucky and Alabama. The five states with the lowest relative increases were Connecticut, New York, Nebraska, New Jersey and Massachusetts.

I interviewed Smart about the research by email last week. An edited transcript follows.

Francis Wilkinson: What do you consider the most significant result from your study?

Rosanna Smart: We show a striking divergence of firearm and non-firearm homicides starting in 2014, such that more than three of every four homicides are now due to firearms. But these increases weren’t felt equally across states, and not just because states differ demographically (there are well-documented differences in firearm homicide rates across demographic groups). One of our most significant results shows that, controlling for demographic differences, there was a more than 100-fold difference in relative increases in firearm homicide rates between 2014 and 2019, with much of the rise concentrated in South-Central and Midwestern states.

Relatedly, the places and people for which we see the most serious worsening trends are largely the places and people that already had some of the highest firearm homicide rates. So, the groups already subject to the highest rates of fatal interpersonal gun violence are experiencing some of the worst increases.

Wilkinson: Can you explain what “controlling for demographic differences” entails? I think many people are not entirely clear on what that means and how you accomplish it.

Smart: We estimated trends in firearm homicide rates across groups using regression analysis, which allows us to consider how factors such as age, sex, race/ethnicity and geography are associated with differential trends. By “controlling for demographic differences” across different states, we are essentially comparing trends in each state to the trends in demographically similar states. For example, we can compare predicted firearm homicide rate trends in California versus those trends in North Dakota by treating those very different states as if they have similar populations.

Wilkinson: Cultural, economic, political, demographic and other differences were apparent among many of these states prior to 2014. Significant differences in rates of firearm possession were also apparent. Do you see any clues suggesting why firearm homicide rates suddenly start to diverge so sharply after 2014?

Smart: Our study doesn’t try to answer why firearm homicide rates reversed course in 2014. But by looking at where, and for whom, these increases were greatest, we can perhaps start to sort through potential explanations. For example, there have been theories about increasing murder rates post-2014 due to de-policing, deteriorating police-community relations or growth in illicit opioid markets. I’m not sure any of those theories point to increases in firearm homicides specifically (versus homicides more generally), and none seems to align with the state-specific patterns we find. None of the South-Central or Midwestern states that experienced large increases are known for being liberal hotbeds of police or prosecutorial reform. Conversely, the places where we see relatively lower rates of increase are largely in the Northeast, an area that has been hard-hit by fentanyl.

It does appear that firearm homicide trends correlate with state-level gun ownership. States with higher gun ownership rates tend to have higher firearm homicide rates. But the differences in gun ownership, and the correlation between firearm ownership and homicides, predate 2014. It may be that there is some sort of interaction between levels of gun ownership (or other pre-existing features correlated with gun ownership) and other factors. For example, changes in policing or economic conditions could have had different effects in places with many guns versus places with few guns. Our study can’t answer this question, but it is an interesting area for future research.

Wilkinson: There is a pretty glaring difference in your study among states with weak gun laws that basically encourage gun ownership and those states with stricter gun regulations. The states with minimal regulation have the greatest homicide spikes. Other studies have confirmed a correlation between weak laws and high rates of gun violence. Has that correlation reached the status of a settled fact among credible gun researchers? Or is there still a question about that?

Smart: I would say the positive correlation between overall firearm homicide rates and gun ownership is fairly well-accepted, as is the correlation between firearm homicide rates and more permissive gun laws. Our study results suggest a less well-documented correlation with recent firearm homicide spikes. However, correlation isn’t causation, and I don’t think it’s a settled question among researchers whether these correlations represent a causal impact of gun laws on gun violence. Some think differences in gun laws drive differences in gun violence, while others would posit that states with more permissive gun laws differ from other states in a variety of other ways — culturally, politically, economically — that lead them to have different firearm homicide rates.

As an aside, there’s also been some work suggesting that the correlation of firearm homicide rates and gun ownership doesn’t hold as strongly for certain subtypes of firearm homicide, such as nondomestic homicides.

Wilkinson: The bottom line, however, is that research repeatedly confirms that high rates of gun ownership and weak gun laws consistently correlate to higher gun homicides. Yet many states have spent the past couple decades promoting increased gun ownership while weakening their gun laws. Do you ever feel that reality-based firearm research is a tad frustrating?

Smart: It’s a research area where emotions and politics often seem to trump the evidence, and there are probably always going to be some people on both sides of the gun policy debate who will make claims that aren’t actually backed by science. As somebody hoping to build the evidence base in order to inform policy, yes, this can definitely be frustrating. But there are a multitude of other reasons to be frustrated working in this space! There are huge gaps in important data that researchers would like to have, like information on nonfatal gun violence, defensive gun use, gun policy implementation and enforcement, and gun purchasing or carrying behavior.

Wilkinson: A previous research paper that you co-authored, which was published in 2020, estimated that thousands of lives could be saved with the implementation of just three policies: (1) laws to inhibit children’s access to firearms; (2) restrictions on the right to carry and (3) repeal of stand-your-ground laws. Could you describe that research?

Smart: The study, led by Terry Schell, sought to estimate how three types of gun laws regulating how people use, carry, and store firearms (stand-your-ground laws, concealed carry laws, and child access prevention laws) affect firearm death rates. Our study differed from earlier work in a few key ways. For one, we estimated effects of each of the individual policies I mentioned, but we also estimated their combined effects — the better to examine likely consequences of moving from a less restrictive gun-policy regime to a more restrictive one.

Looking at these particular policies, we found that child access prevention laws were most likely to be effective at reducing firearm deaths, estimating about a 6% reduction. Our findings for restricting concealed carry and stand-your-ground laws were more uncertain, although the results suggest that more permissive carrying and self-defense laws likely increase firearm deaths rather than decrease them. And if we look at the three policies together, comparing states with the most restrictive regime (those with a child access prevention law, restrictions on right-to-carry, and no stand-your-ground law) to the least restrictive regime, we find the more restrictive regime is associated with about an 11% decrease in firearm deaths.

Wilkinson: An 11% reduction from a baseline of some 40,000 annual firearm deaths would mean an awful lot of saved lives. In your more recent study, the five states where firearm homicides have increased most since 2014 are Missouri, Alaska, New Mexico, Kentucky and Alabama. The Giffords Law Center, a gun-safety organization, grades state gun laws on a scale from A to F. Each of those 5 states earned an F, except for New Mexico, which earned a C.

The states that experienced lower rates of firearm homicides since 2014 were also mostly unsurprising. Connecticut, New York, Nebraska, New Jersey, Massachusetts all received at least an A- from Giffords — except for Nebraska, which earned a C-. Nebraska, the only red state in that group, is an interesting outlier. Does the research suggest why Nebraska is more successful at limiting gun homicides than other red states? Or how we might create more Nebraskas (and fewer Missouris)?

Smart: Nebraska is an outlier on several dimensions — they have a different structure to their state government, somewhat different gun laws than most red states (they have a handgun permitting system), but really it could be almost anything. I’m not aware of research that would convincingly explain gun violence in Nebraska specifically. In general, it’s probably less instructive to pick out an individual state. It’s better to focus on overall patterns such as the widening disparities among states. It’s kind of a rich-get-richer/poor-get-poorer story in that the places that had more gun murders to start with are getting more murder-y.

Wilkinson: Research on gun violence has really picked up in recent years. But the “guns everywhere for everybody” ethos has simultaneously been gaining ground both in conservative states and in the Supreme Court. How do you view the role of gun violence research in public policy right now, and — if you would hazard a guess — what do you think the future impact of research might be?

Smart: I actually think the recent Supreme Court decision in New York State Rifle & Pistol Association v. Bruen in some way highlights the need for gun violence research in public policy. Since the decision, we’ve already seen action by states to change concealed carry regulations. It will be important to try to understand how these changes influence outcomes — not just about gun violence but also changes in who owns and carries firearms, perceptions of safety and law enforcement behavior. Other policies may also change in light of Bruen; similar questions about these impacts are important to answer. Research can help inform how we craft regulations and implement policy within the constitutional bounds that courts will have to delineate, to best ensure public safety.

Finally, I see a real role for better data collection, more descriptive research, and evaluations of the impact of broader policies (such as economic or health-care policies) on gun violence. Recent increases in federal and private-foundation funding for gun violence research offer an opportunity not only to answer questions, but to better understand the full scope of gun violence and to offer policymakers a suite of potential solutions for reducing it.

More From Bloomberg Opinion:

• An Executive Order That Might Stop Gun Violence: Timothy O’Brien

• The Flaw in the Progressive Stance on Guns: Matthew Yglesias

• The Gun Debate Needs to Break Old Patterns: Ramesh Ponnuru

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Francis Wilkinson is a Bloomberg Opinion columnist covering U.S. politics and policy. Previously, he was an editor for the Week, a writer for Rolling Stone, a communications consultant and a political media strategist.

More stories like this are available on

©2022 Bloomberg L.P.