Journal of Urban Health

, Volume 91, Issue 2, pp 293–302

Effects of the Repeal of Missouri’s Handgun Purchaser Licensing Law on Homicides

Authors

    • Johns Hopkins Center for Gun Policy and ResearchJohns Hopkins Bloomberg School of Public Health
  • Cassandra Kercher Crifasi
    • Johns Hopkins Center for Gun Policy and ResearchJohns Hopkins Bloomberg School of Public Health
  • Jon S. Vernick
    • Johns Hopkins Center for Gun Policy and ResearchJohns Hopkins Bloomberg School of Public Health
Article

DOI: 10.1007/s11524-014-9865-8

Abstract

In the USA, homicide is a leading cause of death for young males and a major cause of racial disparities in life expectancy for men. There are intense debate and little rigorous research on the effects of firearm sales regulation on homicides. This study estimates the impact of Missouri’s 2007 repeal of its permit-to-purchase (PTP) handgun law on states’ homicide rates and controls for changes in poverty, unemployment, crime, incarceration, policing levels, and other policies that could potentially affect homicides. Using death certificate data available through 2010, the repeal of Missouri’s PTP law was associated with an increase in annual firearm homicides rates of 1.09 per 100,000 (+23 %) but was unrelated to changes in non-firearm homicide rates. Using Uniform Crime Reporting data from police through 2012, the law’s repeal was associated with increased annual murders rates of 0.93 per 100,000 (+16 %). These estimated effects translate to increases of between 55 and 63 homicides per year in Missouri.

Keywords

firearm policy firearm violence gun policy gun violence

Introduction

Homicide is the second leading cause of death for people aged 15–34 years in the USA and the leading cause of death for black males in this age group.1 Homicide also accounts for 5 % of the Years of Potential Life Lost (YPLL) in the USA2 and is the second leading cause of the racial disparity in life expectancy between black and white males.3 Two-thirds of all homicides in the USA are committed with firearms,1 and the firearm homicide rate in the USA is 19.5 times higher than the average firearm homicide rate in other high-income countries.4

It has been argued that weaknesses in federal and state firearms laws contribute to the unusually high homicide rate in the USA, especially the lack of background checks or record-keeping requirements for private, unlicensed sellers of firearms.5 Many perpetrators of homicide have backgrounds that would prohibit them from possessing firearms as a result of prior convictions for felony crimes6 or for misdemeanors involving domestic violence, being under a restraining order for domestic violence, young age, or other disqualifications.7 Federal law requires background checks and record keeping for sales by federally licensed firearms dealers but exempts these regulations when the firearm seller is unlicensed. Fifteen states require individuals purchasing handguns from unlicensed sellers to pass background checks, and eleven of these states require all handgun purchasers to acquire a permit-to-purchase (PTP) license.

PTP systems require prospective handgun purchasers to obtain a license verifying that they have passed a background check. All handgun sellers, both licensed dealers and private sellers, may only sell to those with a current PTP license. Most states with PTP handgun licensing require applicants to apply for the license directly at a law enforcement agency. In all other states, individuals wishing to purchase a handgun from a licensed dealer must complete a purchase application form. The dealer or dealer’s employee submits the form to the Federal Bureau of Investigation’s (FBI) National Instant Check System (NICS) or, in some cases, to state police, to determine whether the applicant is prohibited from possessing firearms.

Prior research has shown that cities and states that require background checks and record keeping for handgun sales by unlicensed sellers and stricter PTP handgun licensing laws have lower levels of guns being diverted to criminals within a year of retail sale8 and fewer guns exported to criminals across state borders.9 A recent study found a cross-sectional association between states having PTP handgun licensing or other forms of universal background check requirements for gun sales and lower homicide rates.10

Missouri repealed its PTP handgun licensing law effective August 28, 2007. Missouri’s law had been in place since 1921 and required all handgun purchasers to have a valid PTP license (good for 30 days) in order to lawfully purchase a handgun from any seller, licensed or unlicensed. Applicants applied in person at their local sheriff’s office which facilitated the background check. Webster and colleagues8 reported that immediately following the repeal of Missouri’s PTP handgun law, there was a twofold increase in the percentage of guns that had unusually short intervals between the retail sale and the recovery by police, an indicator of firearm diversion or trafficking.11 , 12 The repeal also coincided with a sharp increase in the percentage of crime guns recovered by police in Missouri that had been originally sold by in-state retailers, from 56.4 % in 2006 to 71.8 % in 2012.13

This study examines the effects of the repeal of Missouri’s PTP handgun licensing law on homicide rates. Because this change eliminated mandatory background checks for handguns sold by unlicensed sellers, it is of particular relevance for debates in the US Congress and in several states about proposals to extend background check requirements to all firearm sales.

Methods

Design

The association between the repeal of Missouri’s PTP handgun licensing law on homicide rates was estimated using a quasi-experimental research design with annual, state-level homicide rates. Homicide rates were age adjusted and stratified by those committed with a firearm versus all other methods to discern the specificity of the effects of the policy change on firearm versus non-firearm homicides.

Data and Measures

We hypothesized that the policy change would affect homicide rates but only those committed with firearms. Thus, the primary outcome measure was state-level annual firearm homicide rates, derived from death certificate and census data, age adjusted (reference year 2000) in Centers for Disease Control and Prevention's (CDC) Web-based Injury Statistics Query and Reporting System (WISQARS) Fatal Injury Reports.14 Seven states (HI, ME, NH, ND, SD, VT, and WY) were dropped from the analyses because WISQARS suppressed the data for states and years for which there were very few firearm homicides to protect the anonymity of the data. Missouri’s mean baseline rate of firearm homicides during the pre-repeal study years was approximately four to five times higher than was experienced in the seven dropped states, and none of the dropped states were geographically close to Missouri. Within Missouri, we also used county-level cause-of-death mortality data from CDC’s Wide-ranging Online Data for Epidemiologic Research (WONDER) system15 to assess the degree to which state-wide changes in age-adjusted homicide rates differed across counties.

These data from CDC’s WISQARS and WONDER systems have the advantage of complete, mandatory reporting of death certificate data and the ability to easily isolate homicides committed with firearms versus other methods. The disadvantage of these data is that they were only available through the end of 2010 at the time of this study. We also collected and analyzed state-level data on annual rates of murder and non-negligent manslaughter (which will capture virtually all homicides) from the Federal Bureau of Investigation’s (FBI) Uniform Crime Reporting (UCR) system. UCR data provided two additional years of post-PTP-law-repeal data; however, the FBI has to interpolate some data for states and years, when there is incomplete reporting from local law enforcement agencies, and rates are not age adjusted.

Although data from prior years are available, we chose 1999 as the beginning of our study period because the period 1999–2012 has been the most stable period for homicide trends in many decades. Periods of dramatic change, especially if the underlying causes for those changes cannot be easily modeled, are vulnerable to omitted variable bias in estimates of policy impact.16

Regression analyses are used to estimate policy change effects and controlled for changes in rates of unemployment, poverty, incarceration, burglary, law enforcement officers per capita, and the presence of four other types of state laws potentially most directly relevant to lethal violence for which there was significant change during the study period. These laws included so-called Stand Your Ground (SYG) laws, which give individuals an expanded right to use deadly force in potentially dangerous encounters with no duty to retreat, right-to-carry (RTC) laws which require law enforcement agencies to issue permits to carry concealed firearms to all legally qualified applicants, bans of unsafe handguns including so-called Saturday Night Specials, and firearm prohibitions for young adults resulting from convictions for serious crimes adjudicated in juvenile courts. SYG laws have been enacted in many states in recent years, including in Missouri in 2007. Prior research indicated that these laws may increase homicides.17 Early research suggested that RTC laws may reduce homicides,18 but the most rigorous studies show no evidence that RTC laws affect homicide rates.19 , 20 Maryland’s adoption of a SNS ban was associated with a reduction in firearm homicide rates,21 but this policy has not been rigorously studied in other states nor has firearm prohibitions stemming from serious juvenile offenses.

Average annual unemployment rates (per 100 population 16 years of age and older) were obtained from the Bureau of Labor Statistics.22 Poverty rates (per 100 population) were obtained from the Census Bureau’s Current Population Survey.23 Burglary rates (per 100,000 population)—an indicator of crime rates that should not be directly affected by gun laws—and the rates of law enforcement officers (per 100,000 population) were drawn from the FBI’s UCR program.24 Incarceration rates (per 100,000 population) are from the Sourcebook of Criminal Justice Statistics.25 The repeal of Missouri’s PTP handgun licensing law was measured as the proportion of days in a year when the state had no PTP handgun law, i.e., 0 for the years the PTP law was in place (1999–2006), 0.263 in 2007, and 1 for 2008–2012.

Analytic Methods

Pre-repeal versus post-repeal differences in mean age-adjusted homicide rates were tested for statistical significance using t-tests. To estimate the independent association between the repeal of Missouri’s PTP handgun licensing law and age-adjusted homicide rates, we used generalized least squares regression models. The models included state- and year-fixed effects to control for baseline differences in states’ homicide rates and yearly fluctuations that occurred nationally as well as changes in the covariates described above. Standard errors for model coefficients were adjusted to account for clustering by state and for heteroskedasticty using the Eikert–Huber–White adjustment.26 Analyses were conducted using Stata IC v 11.0.27

Results

From 1999 to 2007, Missouri’s firearm homicide rate was relatively stable, fluctuating around a mean of 4.66 per 100,000 population per year (Fig. 1). In 2008, at the first full year after the permit-to-purchase licensing law was repealed, the firearm homicide rate in Missouri increased sharply to 6.23 per 100,000, a 34 % increase from the baseline mean. For the post-repeal period of 2008–2010, the mean annual firearm homicide rate was 5.82, 24.9 % higher than the pre-repeal mean (t = 4.38, df = 10, p = .001). Within Missouri, firearm homicide rates per 100,000 increased sharply between the pre- and post-repeal periods in each of the three large central metro counties/jurisdictions—by 30 % in Jackson County (11.2 to 14.7), 47 % in St. Louis County (5.0 to 7.4), 27 % in St. Louis City (21.7 to 27.5), and 34 % overall in the nine Missouri counties designated as large metropolitan fringe counties (3.1 to 4.2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11524-014-9865-8/MediaObjects/11524_2014_9865_Fig1_HTML.gif
FIG. 1

Age-adjusted firearm homicide rates in Missouri, states bordering Missouri (population-weighted averages), and the USA, 1999–2010.

This sharp increase in firearm homicide rates in Missouri beginning in 2008 was out of sync with changes during that period nationally and in states bordering Missouri (Table 1). The mean age-adjusted firearm homicide rate in the USA declined 5.5 % from 4.03 per 100,000 during 1999–2007 to 3.81 for 2008–2010. The population-weighted mean firearm homicide rates across the eight states bordering Missouri changed little between these two time periods (4.15 to 4.06, −2.2 %; p = .480, Fig. 1), and there were no statistically significant changes in any specific state that bordered Missouri.
TABLE 1

Mean firearm homicide rates before (1999–2007) and after Missouri repealed its permit-to-purchase handgun licensing requirement for handgun sales by licensed and unlicensed sellers (2008–2010)

 

Mean before Missouri’s PTP handgun law repealed 1999–2007

Mean after Missouri’s PTP handgun law repealed 2008–2010

% Change

Probability 2 means are equal

Missouri

4.67

5.82

+24.9

.001

Population-weighted mean for states bordering Missouri

4.15

4.06

−2.2

.480

Arkansas

5.12

5.23

+2.1

.691

Illinois

5.10

4.77

-6.6

.335

Iowa

0.93

1.00

+7.8

.627

Kansas

3.95

3.85

−3.4

.757

Kentucky

3.26

3.29

+1.0

.898

Nebraska

1.75

2.28

+30.0

.096

Oklahoma

3.80

3.93

+3.5

.618

Tennessee

5.42

5.23

−3.5

.553

Controlling only for baseline differences across states and year effects nationally (model 1, Table 2), the repeal of Missouri’s PTP handgun licensing law was associated with an increase in firearm homicide rates of 1.32 per 100,000 (p < .001), a 29.4 % increase above rates projected without the repeal. After controlling for changes in rates of unemployment, poverty, burglary, incarceration, and law enforcement officers along with other state laws, the estimated increase in annual firearm homicide rates associated with the repeal of Missouri’s PTP handgun law was 1.09 per 100,000 population per year (p < .001; 95 % confidence interval (CI) 0.81 to 1.38), a 23 % increase.
TABLE 2

Estimates of effect of the repeal of Missouri’s permit-to-purchase handgun law from generalized least squares regression models on states’ age-adjusted firearm, non-firearm, and all-cause homicide rates, 1999–2010, and murder and non-negligent manslaughter rates, 1999–2012

Outcome variable

β

Robust S.E.

P value

95 % CI for β

Firearm homicide rates, 1999–2010 R2 within = .208, R2 overall = .948

1.09

0.14

<.001

0.81 to 1.38

Non-firearm homicide rates, 1999–2010 R2 within = .162, R2 overall = .583

−0.08

0.10

.446

−0.28 to 0.12

Total homicide rates, 1999–2010 R2 within = .177, R2 overall = .943

1.00

0.18

<.001

0.66 to 1.35

Murder and non-negligent manslaughter rates, 1999–2012 R2 within = .183, R2 overall = .908

0.93

0.23

<.001

0.48 to 1.38

All models controlled for rates of unemployment, poverty, burglary, incarceration, law enforcement officers, “Stand Your Ground” laws, right-to-carry laws, bans of Saturday night special (junk) handguns, and firearm prohibitions of young adults with prior serious criminal offenses adjudicated in juvenile courts. Estimates for each of these covariates can be found in the Supplemental Tables.

The increase in homicide rates following the repeal of Missouri’s PTP handgun licensing law occurred only for homicides committed with firearms. Following similar trends nationally, Missouri’s age-adjusted rate of non-firearm homicides declined from a pre-repeal (1999 to 2007) mean of 2.19 to a post-repeal (2008 to 2010) mean of 1.88 (−14 %). Regression analyses indicated that Missouri’s repeal of its PTP handgun law was associated with no change in the age-adjusted non-firearm homicide rate (β = −0.077, p = .446) and an increase in annual homicide rates for all methods of 1.00 per 100,000 (Table 2, p < .001, 95 % CI 0.66 to 1.35).

Using UCR data from police reports for 1999-2012, the difference in the annual murder rate in Missouri minus that of the U.S. as a whole grew from 0.60 per 100,000 population during the pre-PTP-repeal period to 1.82 during the 5 years after the repeal of the PTP law (data not shown, t=4.12, df=12, p<.001). A model which only controlled for state- and year-fixed effects estimated a 1.34 increase in annual murder rates associated with the repeal of the PTP handgun law (β = 1.34, p = .001, 95 % CI 0.58 to 2.11); however, the estimated effect of the policy change was reduced to an increase of 0.93 murders per 100,000 population per year after all covariates were included in the model (Table 2, β = 0.93, p < .001, 95 % CI 0.48 to 1.38), a 16 % increase relative to the counterfactual.

Firearm homicide, total homicide, and murder rates were positively associated with burglary rates and negatively associated with poverty rates. New unsafe handgun bans adopted in California and Massachusetts were associated with an increase in total homicide rates (β = 0.46, p = .008, 95 % CI 0.12 to 0.80). No other covariate reached statistical significance at the .05 level (Supplemental Tables).

Discussion

This study provides compelling evidence that the repeal of Missouri’s PTP handgun licensing law, which required all handgun purchasers to pass a background check even for purchases from private sellers, contributed to a sharp increase in Missouri’s homicide rate. Our estimates suggest that the law was associated with an additional 55 to 63 murders per year in Missouri between 2008 and 2012 than would have been forecasted had the PTP handgun law not been repealed.

Our analyses ruled out several alternative hypotheses to explain the relatively large and highly statistically significant increase in firearm homicides in Missouri following the repeal of its PTP handgun licensing law. We controlled for changes in unemployment, poverty, policing levels, incarceration rates, trends in crime reflected in burglary rates, national trends in homicide rates, and several kinds of other laws that could affect homicides. That Missouri’s sharp increase in firearm homicides was unique within the region, specific to firearms, and was observed in metropolitan jurisdictions across Missouri suggests that unmeasured unique local circumstances (e.g., gang activity and changes in social norms) are unlikely to have biased our estimates of the impact of the policy change. Estimates of the effects of the repeal of Missouri’s PTP handgun law were similar for firearm homicides and total homicides using death certificate data for 43 states through 2010, and for murders and non-negligent manslaughters using police reports for all 50 states through 2012. This suggests that the data source and time period studied are unlikely to have biased the findings.

Causal inferences from quasi-experimental studies can be strengthened with empirical evidence supporting the proposed causal chain between the intervention, mediators, and the outcomes under study. Handgun purchaser licensing and universal background checks are hypothesized to affect homicide rates by reducing gun diversions to criminals and other prohibited groups. The evidence that Missouri’s increase in firearm homicides was fueled by the state’s repeal of its PTP law is bolstered by data indicating that the repeal was immediately followed by a twofold increase in the percentage of crime guns that were recovered by police soon after the guns’ retail sales and an unusually large increase in the percentage of Missouri’s crime guns that had been purchased from Missouri gun dealers.9 These finding are consistent with prior research showing that states that regulated handgun sales by unlicensed sellers had fewer guns diverted to criminals shortly after in-state retail sales,8 and that states with the most comprehensive handgun sales laws including PTP licensing requiring direct interface with law enforcement have proportionately fewer guns used in crime that were originally sold by in-state retailers.28 , 29 Having a large percentage of crime guns that originate from out-of-state sales, as was the case in Missouri prior to the repeal of its PTP law, is indicative of a restricted supply of guns available to criminals from in-state sources. Restrictions from local suppliers increase prices in the underground gun market and attract suppliers from states with fewer legal impediments to gun diversion.30 , 31

The weakening of Missouri’s gun laws may have also contributed to gun trafficking to border states that regulate handgun sales by all sellers via PTP licensing. The number of guns sold in Missouri and later recovered by police in Illinois and Iowa, two border states with handgun purchaser licensing laws, increased 37 % (from 133 to 182) from 2006 (just before Missouri’s PTP law was repealed) to 2012 when the overall number of crime guns recovered by police in those states actually declined by 6 %.12

A potential threat to the validity of our estimate of the impact of the repeal of Missouri’s PTP law is confounding by the simultaneous adoption of a Stand Your Ground law in Missouri. Controlling for the effects of SYG laws across all states, our estimate of the effect of the repeal of Missouri’s PTP law on homicide rates declined slightly but was still substantial and statistically significant at p < .001. A separate analysis of justifiable homicide data from the FBI’s Uniform Crime Reports revealed that there were approximately three additional justifiable homicides per year in Missouri following the adoption of the state’s Stand Your Ground law above pre-SYG-law levels—less than 1 % of the total number of gun homicides during 2008–2010.

Critics could question the use of a relatively short pre-repeal baseline period used for this study. Using more longitudinal observations can potentially produce more accurate forecasts of the counterfactual in interrupted time-series impact studies. However, the period from 1985 to 1998 included dramatic increases and decreases in US homicide rates. Experts believe that these changes were driven by factors that could not be directly measured (e.g., dynamics of the crack cocaine market, and changes in social norms)32 and thus controlled statistically and that these unmeasured forces appear to have been uneven across states.19 Such conditions pose considerable challenges for deriving unbiased estimates of policy impacts. By limiting the analyses to the relatively stable period of 1999–2012, we minimized the potential for omitted variable bias that would have likely been introduced by including data from this earlier time period.

The generalizability of our findings to other states with PTP handgun laws is unknown. Data from a recent cross-sectional study indicated that PTP licensing laws and universal background check requirements were associated with lower homicide rates after controlling for other population risk factors;9 however, the lack of longitudinal data weakens causal inference from that study. We caution, however, that passage of a PTP handgun licensing law with mandatory background checks and record keeping for all handgun sales may not result in as immediate and large a reduction in firearm homicides as occurred in reverse when Missouri’s law was repealed. Although our findings indicate that Missouri benefited from the protective effects of its PTP law before the law’s repeal, the beneficial effects of new laws of this type may be more gradual as enforcement practices are put in place, awareness of the law increases, and the stock of guns available in the underground market is depleted. Additional methodologically rigorous research of the impact of other laws of this type is warranted.

Acknowledgments

This research was funded by a grant from The Joyce Foundation to the Johns Hopkins Center for Gun Policy and Research. Impact of Permit-to-Purchase Handgun Licensing Laws on Violent Crime, 12-34373.

Supplementary material

11524_2014_9865_MOESM1_ESM.docx (15 kb)
ESM 1 (DOCX 15 kb)
11524_2014_9865_MOESM2_ESM.docx (17 kb)
ESM 2 (DOCX 17 kb)

Copyright information

© The New York Academy of Medicine 2014