Revisiting “More Guns, Less Crime”: What the Research Actually Shows

“More guns, less crime” is a research claim, not a slogan invented for a bumper sticker — it comes from a specific body of academic work, and understanding what it does and doesn’t establish requires separating the finding from the framing.

Where the phrase actually comes from

Economist John Lott’s research, most prominently a study published in the late 1990s and later expanded into a book of the same name, examined county-level crime data following the spread of “shall-issue” concealed carry laws across U.S. states from the late 1970s through the 1990s, and reported an association between shall-issue laws and reductions in certain violent crime rates.

What the original research claimed

The core claim: counties that adopted shall-issue concealed carry laws saw a measurable decline in violent crime relative to counties that hadn’t, controlling for other factors. It became one of the most cited — and most disputed — empirical claims in the gun policy debate.

What critics of the research argue

The research has faced sustained methodological criticism, most notably from economists John Donohue and Ian Ayres, who reran the analysis with extended data and different model specifications and reported the effect disappearing or reversing depending on the model used. A National Research Council review of the broader body of “right-to-carry” research in the mid-2000s concluded the evidence was too inconsistent across specifications to support a confident causal claim in either direction.

Where the empirical debate actually stands today

Two decades later, this remains a genuinely contested empirical question in criminology, not a settled one on either side. Isolating the effect of one policy change from everything else moving crime rates over the same decades — policing shifts, incarceration rates, the rise and fall of the crack epidemic, economic conditions — is a hard identification problem, not a simple counting exercise. Anyone citing this research, for or against, as an open-and-shut case is overstating what the data actually supports.