Firearms Regulation

What the evidence says

The clearest, highest-confidence result in this set is Australia's 1996 National Firearms Agreement: zero mass shootings in the 10.5 years after the law, against 13 in the 18 years before it — a result reported independently by two sources drawing on the same underlying count. The same reform is tied to an acceleration in firearm-suicide decline from 3% to 7.4% per year and in total-firearm-death decline from 3% to 6% per year, with total suicide trends flipping from a 1% annual increase to a 4.4% annual decrease (p<0.001) and total homicide trends deepening from a 0.1% to a 3.3% annual decrease. Other national and subnational reforms point the same direction on suicide: Austria's law is tied to a 9.9% reduction in the firearm-suicide trend, Norway's 1990 permit requirement is identified as the firearm law most likely contributing to lower suicide rates, and in Canada the annual rate of decline in male firearm suicide accelerated from −2.7% to −11.1% (ages 15–34) and from −1.2% to −5.6% (ages 35–64) after 1996, with no significant change detected in men 65 and older. None of these reforms show a compensatory shift to other suicide or homicide methods — the data explicitly do not support a method-substitution hypothesis in either Australia or Canada. One reform-linked finding cuts the other way: in Australia, unintentional firearm deaths rose 8.5% per year even as total firearm deaths and firearm suicides fell.

Individual firearm-purchase and carry laws diverge sharply by type. Permit-to-purchase (licensing) requirements are the most consistent protective lever measured here: Maryland-style urban counties with permit-to-purchase laws show a 14% reduction in firearm homicide, and Missouri's 2007 repeal of its permit-to-purchase law is tied to 55–63 additional homicides per year (a 25% increase in firearm-homicide rates in a separate analysis of the same repeal). Right-to-carry, shall-issue, and stand-your-ground laws move the opposite way in most of the newer, larger studies: shall-issue states show 8.6% higher firearm homicide and 6.5% higher total homicide than may-issue states, right-to-carry laws are tied to a 29% increase in firearm workplace homicide, and stand-your-ground laws are tied to a 6.8% increase in homicide (14.7% among white males). But this is not unanimous — an earlier reanalysis using 1977–1992 data found shall-issue laws associated with lower homicide rates, and a subsequent recoding of that same data found no association at all, so the direction of this particular lever is contested rather than settled.

Background-check, waiting-period, and access-restriction laws targeting specific risk situations show a more consistently protective pattern, though usually without a quantified magnitude. Domestic-violence firearm-purchase restrictions are tied to a 9% reduction in intimate-partner homicide. Child access prevention laws are tied to fewer unintentional firearm deaths among children under 15 (no measured effect for older adolescents), a 19% reduction in firearm suicides among children 15 and younger, and fewer nonfatal firearm injuries among people under 18. The Brady Act's waiting-period provision is tied to fewer firearm suicides among adults 55 and older but shows no measured effect on firearm homicide among adults of any age — a pattern replicated across two independent citations of that law. Closing the gun-show background-check loophole is tied to a lower rate of crime guns exported out of state: 7.5 per 100,000 population in states that closed the loophole versus 19.8 per 100,000 in states that did not.

At the level of aggregate state policy strength, counties in states with weak home-state and weak neighboring-state ("interstate") firearm policy scores show 38% higher firearm homicide and 32% higher overall homicide than counties with stronger scores, with no measured difference in nonfirearm homicide — indicating the effect is specific to firearm-enabled violence rather than violence generally. Stronger state firearm laws are similarly tied to lower firearm and overall suicide rates. Outside the United States, a cross-national comparison finds the number of firearms per country associated with higher firearm-suicide rates (R²=0.21) and more strongly with higher firearm-homicide rates (R²=0.35). A separate line of evidence on law enforcement finds that stricter state firearm laws, driven specifically by gun-owner-accountability provisions (a 3.7–3.8% decrease in fatal police shootings per one-point increase in that provision score), are tied to fewer fatal police shootings; the overall Giffords law-strength score shows a smaller 0.7–1% decrease per point, while broader firearm prevalence and most other law categories (background checks, other sale regulations, weapon classes, consumer safety, public-place restrictions) show no measured effect on that outcome.

Several policy levers show flatly no measured effect, and that absence is itself informative. Firearm buyback programs in St. Louis and Buffalo show no association with firearm homicide; the 1997 Australian buyback shows no association with firearm homicide either, though it is tied to a reduction in suicide proportional to the number of firearms surrendered. Laws limiting handgun purchases to one per person per month and law-enforcement undercover stings of gun dealers both show no measured effect on intrastate gun trafficking. Comprehensive background-check laws not tied to permit-to-purchase requirements are labeled inconclusive rather than protective or null. And causal attribution remains explicitly unresolved in at least one source: the study documenting Canada's post-1996 firearm-suicide acceleration states plainly that "whether this represents a causal relationship requires further study," citing a comparable decline in Scotland absent new gun-control law and concurrent economic recovery as unruled-out alternative explanations.

Where the evidence splits

Shall-issue and right-to-carry laws split on their association with homicide. An early time-series analysis of 1977–1992 Uniform Crime Reports data found shall-issue laws associated with lower homicide rates; a later reanalysis of the same data, after "recoding the data and varying model specifications," found no association. More recent panel/difference-in-differences studies (comparing shall-issue to may-issue states) find the opposite of the original result — 8.6% higher firearm homicide and 6.5% higher total homicide in shall-issue states — while an earlier cross-sectional time-series study found a similar-direction but not statistically significant estimate (RR 1.11, CI crossing 1). The claims identify one explicit cause (data recoding and model respecification) and leave the rest of the divergence — across different data periods and study designs — unexplained by the sources here.

Castle-doctrine and stand-your-ground laws split on both direction and magnitude. A time-series study using 1977–2005 data found castle doctrine laws associated with a 9% decrease in homicide. A difference-in-differences study using 2000–2010 data found these laws associated with a 6–11% increase, and a third study found a 6.8% increase overall (14.7% among white males). The claims state the differing identification strategies and study periods explicitly, which plausibly explains part of the divergence, though the sources don't quantify how much of the gap that difference accounts for.

The federal assault weapons ban splits on direction. One study found it associated with higher rates of firearm homicide; another, using 1980–1995 Uniform Crime Reports data across 15 states and adjusting for the presence of other firearm and crime laws (including in New York and California), found no association with homicide rates. The claims describe the second study's control strategy but not the first study's method, so the source of this disagreement is not established here.

The 1976 District of Columbia firearm restrictions split on effect. One analysis, using DC and adjacent Maryland/Virginia metro areas as controls, found an abrupt reduction in homicide and suicide with no similar change in the control areas. A second analysis substituted Baltimore as the control area — citing differences in prelaw homicide rates between the original control areas and DC — and found no association with abrupt or gradual changes in homicide rates. This split is explained directly by the claims: the choice of comparison area.

Minimum sentencing enhancements for firearm offenses split by scope of study. A multi-state panel found minimum sentencing enhancements associated with higher firearm homicide rates. A single-state evaluation of Massachusetts' Bartley-Fox mandatory-sentence law found no association with homicide rates — but that evaluation examined only the 6 months following implementation, a narrower follow-up window than the panel study. The sources describe this design difference but do not state that it is what caused the divergent result.

Thresholds and dose-response

For counties with low home-state firearm-policy stringency, firearm-homicide risk is lower where neighboring-state ("interstate") policy stringency is high than where it is low: IRR 1.38 (95% CI 1.02–1.88) relative to the reference category under low interstate stringency, versus IRR 1.18 (89% CI 0.89–1.54) under high interstate stringency. For counties with medium or high home-state policy stringency, this interstate gradient disappears — homicide rates (firearm, nonfirearm, and overall) show no measured association with interstate policy score, regardless of whether it is low, medium, or high.

Open empirical questions

• What magnitude of effect comprehensive background-check laws not tied to permit-to-purchase requirements have on firearm homicide — the source calls this "inconclusive" without further detail.
• Whether the accelerated post-1996 decline in Canadian firearm suicide is causally attributable to Bill C-17 specifically — the source states this "requires further study" and cannot rule out concurrent factors (a comparable Scottish decline without new gun-control law, economic recovery, other suicide-prevention efforts).
• Whether the divergence between older shall-issue/right-to-carry studies (decrease or null) and newer panel studies (increase) reflects a true change in effect over time, a shift in study period and state samples, or the model-specification differences the sources partially describe.
• What magnitude several protective associations carry — multiple claims here report direction only ("decrease," magnitude unspecified) for storage laws' effect on homicide, restrictive background-checks-plus-restraining-order laws, and several Santaella-Tenorio-cited studies, leaving effect size unmeasured in the sources on this page.