Abstract
Objectives. We explored differences in criminal convictions between holders and nonholders of a concealed handgun license (CHL) in Texas.
Methods. The Texas Department of Public Safety (DPS) provides annual data on criminal convictions of holders and nonholders of CHLs. We used 2001 to 2009 DPS data to investigate the differences in the distribution of convictions for these 2 groups across 9 types of criminal offenses. We calculated z scores for the differences in the types of crimes for which CHL holders and nonholders were convicted.
Results. CHL holders were much less likely than nonlicensees to be convicted of crimes. Most nonholder convictions involved higher-prevalence crimes (burglary, robbery, or simple assault). CHL holders’ convictions were more likely to involve lower-prevalence crimes, such as sexual offenses, gun offenses, or offenses involving a death.
Conclusions. Our results imply that expanding the settings in which concealed carry is permitted may increase the risk of specific types of crimes, some quite serious in those settings. These increased risks may be relatively small. Nonetheless, policymakers should consider these risks when contemplating reducing the scope of gun-free zones.
In 2007, American civilians owned 270 million firearms (88.8 firearms/100 persons), making the US population the most heavily armed civilian population in the world.1 These firearms are in the hands of approximately 40% of the population. More than 60% of gun owners own multiple firearms, with 13% of Americans owning 4 or more firearms.2,3 Estimates indicate that 39% of firearms owned by Americans are handguns, and 61% are rifles or shotguns.3
In recent decades, the number of individuals legally permitted to carry concealed handguns has increased dramatically. In 1987, 10 states allowed citizens to carry concealed handguns; by 2010 more than 40 states had concealed carry laws.4 Not only are more states allowing concealed carry, but those states are issuing many more licenses each year. In 2000, Pennsylvania officials issued 95 334 concealed carry licenses. In 2009, that number had increased by 74%, to 165 857 licenses to carry a concealed handgun.5 In Texas, 2.6% (461 724 individuals) of the population aged 21 years or older held a concealed handgun license (CHL) in 2011.6
The public heath impact of firearms on American society is a contentious issue. With the growth in CHL legislation, the legal carrying of concealed handguns has become an element in the ongoing academic and policy debates over the relationship between public health and firearms. Thus far, empirical results indicate that CHL legislation lowered crime rates,7,8 increased crime rates,9 or had no significant effect on crime rates.10 As the National Research Council concluded in 2004, “with the current evidence it is not possible to determine that there is a causal link between the passage of right-to-carry laws and crime rates.”11(p150)
However, discussion of different aspects of CHL legislation and the behavior of those who seek or have a CHL has been limited. We will not have a complete picture of the relationship between CHL legislation and criminality until we consider the behavior of individuals holding a CHL. The limited research available on this population indicates that CHL holders with a felony record are more likely to be involved in firearm-related crimes than are applicants with a felony record who are denied a license.12 Other research indicates that crime rates for CHL licensees with a previous criminal record vary significantly from criminality among CHL licensees with no criminal record.13
CHL advocates do not focus on the CHL population and differences within that population. They focus on the more general argument that those allowed to carry concealed handguns are much less likely than the average American to engage in criminality.13,14 Handguns are, therefore, being legally placed in “safe hands.” The most common evidence provided to support the safe hands argument involves a broad comparison of rates of criminality between CHL licensees and those not licensed to carry a concealed handgun. Figure 1 provides an example of this style of analysis.
FIGURE 1— Texas Department of Public Safety rates of convictions data used to prove that holders of concealed handgun licenses commit few crimes.
Note. CHL = concealed handgun license.
Source. DemocraticUnderground.com.15
Data such as these lead to statements like that of Charles Cotton on the Texas CHL Blog:
The latest comparison of crime statistics between Texas Concealed Handgun Licensees and the general population (over age 20) is shocking—shockingly good that is! … In spite of a 56% increase in the number of Texas CHL’s (258,168 in 2006; 402914 in 2009). CHL’s were 15 times less likely to commit a crime than the general public in Texas, as compared to 7 times less likely in 2006. That’s right, and [sic] already excellent track record was over twice as good.16
These data also led to commentary from Harold Nemerov, a columnist for the Texas State Rifle Association’s Sportsman magazine, who noted that CHL holders “voluntarily undergo background checks normally reserved for sensitive jobs or criminal arrests, in order to certify that they rank among the most law-abiding citizens prior to receiving their license. These data prove that trust has not been in vain.”17
Our research should add to our understanding of the relationship between CHL holders and public health because it focused on the issue of criminality among CHL holders. Previous studies supporting the safe hands argument compared overall conviction rates between CHL holders and other Texans (nonholders). We analyzed the same data source often used to promote the safe hands argument to investigate a different question: in Texas, how did the crimes for which CHL holders were convicted differ from the crimes for which other Texans (nonlicensees) were convicted?
The Texas Department of Public Safety (DPS) annually publishes data on CHL activity and criminal convictions for the population on its Web site.18 These annual data contain aggregate information on the age, gender, and race of people who hold a CHL, as well as the number of convictions for holders and nonholders of CHLs for 126 serious crimes. We analyzed data for 2001 to 2009 because this range provided a view of current criminality and yielded a large enough number of convictions of CHL holders to make the results relatively stable. We also used data on the characteristics of the prison population in Texas and national data from the Uniform Crime Reports for comparisons with DPS data.18,19
During these 9 years, 532 498 persons without CHLs were convicted of the crimes on the DPS list. Only 1022 convictions for crimes on this list involved CHL holders. After inspecting the conviction data, we deleted from the database crimes for which only CHL holders could be convicted (e.g., unlawful possession of a handgun by a CHL holder in a premise serving alcohol). We also eliminated from our comparisons crimes that CHL holders should not be able to commit (e.g., felon in possession of a firearm). We deleted several crimes that had very low prevalence and were difficult to categorize meaningfully (e.g., child abandonment, assault on a nursing home resident by a staff member). These exclusions eliminated 36 low-prevalence criminal offenses.
These deletions did not dramatically reduce the size of the database. The data, after exclusions, had information on 516 958 convictions of nonlicensees and 934 convictions of CHL holders. The convictions retained for our analyses represented 97% of the convictions for nonholders of CHLs and 91% of the convictions for CHL holders for 90 serious criminal offenses.
We aggregated the 90 crimes included in our analyses into 9 groups containing similar offenses (number of offense types): Sexual offenses (13), Weapons violations (15), Crimes involving an intentional killing (12), Robbery (2), Burglary (4), Aggravated assault (21), Assault (15), Deadly conduct (2), and Terrorist threats (6)
Two of these crime categories are not part of the usual panoply of criminal offenses investigated by researchers. Deadly conduct involves threatening a person with a firearm when that person has reason to fear that the firearm will be used to harm them. Terrorist threats involve threats to engage in criminal activities that may result in an emergency response from civil authorities (e.g., bomb threats, threats involving toxic substances).
To investigate differences in licensee and nonlicensee populations, we compared the 2001 to 2009 Texas CHL conviction data with national arrest data for the same years that provided information on both the offense alleged and the arrestees’ demographic characteristics (age, race, and gender).19 Arrests obviously differ from convictions. However, our interest was not in direct comparisons of these 2 populations but in comparing the distribution of arrests across different types of crimes and the distribution of convictions across different types of crimes.
Measures calculated from the raw conviction data were (1) for CHL holders, the percentage of total CHL convictions for each of the 9 crime categories, (2) for nonholders of CHLs, the percentage of total non-CHL convictions for each of the 9 crime categories, and (3) the ratio of the 2 percentages (CHL/non-CHL) for each category of convictions.
The DPS data were population data, but they were part of an administrative database that may contain random measurement or processing errors. For that reason, we estimated the z score for the differences in proportions in these 2 samples. However, our analysis of these population data focused more heavily on the ratio of our 2 percentages than on statistical significance.
The results of our analyses appear in Table 1. Convictions of nonholders of CHLs were concentrated in the 3 most common types of crimes: simple assaults, burglaries, and robberies. These 3 crimes accounted for 70% of all convictions for nonlicensees. Robbery and burglary accounted for 22% of nonlicensee convictions and only 3% of convictions for CHL holders. Simple assault accounted for 48% of nonlicensee convictions and 31% of convictions for CHL holders.
TABLE 1— Convictions for Serious Crimes Among Holders and Nonholders of Concealed Handgun Licenses: Texas, 2001–2009
| Convictions | |||
| Crimes | z Score | CHL Holder, No. (%) | CHL Nonholder, No. (%) |
| Burglary | −11.56 | 19 (2.0) | 81 914 (15.8) |
| Robbery | −6.69 | 10 (1.1) | 33 318 (6.4) |
| Assault | −9.92 | 293 (31.4) | 246 022 (47.6) |
| Aggravated assault | −0.40 | 53 (5.7) | 30 936 (6.0) |
| Terroristic threats | 2.20 | 49 (5.2) | 19 947 (4.0) |
| Intentional killing | 3.80 | 19 (2.0) | 4520 (0.87) |
| Sexual offenses | 10.93 | 159 (17.0) | 39 044 (7.5) |
| Weapons offenses | 14.09 | 208 (22.3) | 46 666 (9.0) |
| Deadly conduct | 19.21 | 124 (13.3) | 14 591 (2.8) |
| Total convictions | 934 (100) | 516 958 (100) | |
Note. CHL = concealed handgun license.
The ratios of the proportion of convictions in each group of offenses for CHL holders and nonholders appear in Figure 2. Ratios below 1.0 indicate that the proportion of CHL holder total convictions for the type of crime indicated was lower than the proportion of convictions for that type of crime for nonlicensees. As the results in the figure indicate, the ratios of the percentage of criminal convictions for CHL holders compared with nonholders were 0.13 for burglary, 0.17 for robbery, and 0.66 for simple assault. For 2 other groups of offenses (aggravated assault and terroristic threats), the conviction ratios were relatively close to 1.0, ranging from 0.95 to slightly higher than 1 (1.36).
FIGURE 2— Ratio of percentage of convictions for CHL holders to percentage of convictions for non–CHL holders (% CHL/% non-CHL): Texas, 2001–2009.
Note. AGG. = aggravated; CHL = concealed handgun license.
We found that a higher proportion of CHL holders’ convictions, when compared with those of nonholders, were for sexual offenses, weapons offenses, deadly conduct, and offenses involving the intentional killing of another person. For nonlicensees, 7.6% of convictions were for sexual offenses; 17% of CHL holders’ convictions were for sexual offenses. The ratio of the percentages was 2.2, meaning that a much larger proportion of the CHL holders’ than nonholders' convictions were for sexual offenses. The same held true for weapons offenses. A larger proportion of CHL holders’ than nonholders' convictions were for weapons offenses (ratio = 2.5). A larger proportion of CHL holders’ than nonholders' convictions were for deadly conduct (threatening someone with a firearm; ratio = 4.7). In our comparison of convictions among CHL holders and nonholders for offenses involving a death (which incorporated various classes of manslaughter, murder, and capital murder), we found that a larger proportion of CHL holders’ than nonholders' convictions were for these offenses (ratio = 2.3).
To get some indication of how CHL holders in Texas generally differed from individuals involved in the criminal justice process, we made 2 comparisons. In 2009, 138 768 individuals in Texas received a CHL.20 We compared new CHL licensees in that year to the Texas prison population in the same year and found that CHL recipients were more likely to be female (22% vs 10%),20,21 White (87% vs 30%),20,21 and older than 40 years.21,22 We found 7 comparable types of offenses in the national arrest data and the Texas conviction data (murder and nonnegligent manslaughter, robbery, aggravated assault, burglary, assault, weapons offenses, and sexual offenses). For these offenses, for men older than 20 years, 0.6% of the total arrests were for capital murder, murder, or voluntary manslaughter. For Texas CHL holders, these offenses constituted 2.5% of convictions for these crimes. Weapons offenses for men older than 20 years constituted 6.9% of arrests and for Texas CHL holders, 27.3% of convictions.
The percentage of convictions for robbery, burglary, and aggravated assault was much lower for CHL holders than for men older than 20 years. This pattern also held when the comparison was with all arrestees older than 39 years. We observed similar results when we separately compared our CHL holder conviction data for these offenses with the arrest data for women, Blacks, and Whites older than 20 years. This pattern is consistent with the results displayed in Table 1. These data indicate that Texas CHL holders differ from the population involved in the criminal justice process.
Texans without a CHL committed many more crimes than did Texans with a CHL. We observed dramatic differences, however, in the distribution of the types of criminal convictions for these 2 groups. That these 2 populations are usually convicted of different types of crimes should not be terribly surprising. Indeed, the regulatory requirements for acquiring a CHL in Texas were expressly devised to make CHL holders a subpopulation that differs dramatically from the average criminal offender.
These requirements have strong effects on the demographics of the CHL population and clearly distinguish it from those involved in crimes in the general population. Such differences also exist between the characteristics of licensees and those of the perpetrators of violent crimes. Some evidence of this comes from the Federal Bureau of Investigation’s data for 2006, which indicate that a solid majority (60%) of those arrested for violent crimes were younger than 30 years.23 In 2006, 91% of Texans issued a CHL were aged 30 years or older.24
Holders of CHLs in Texas differ in important ways from the general population and are responsible for very little of the state’s criminality. Acquiring a CHL in Texas requires a background check, fingerprinting, hours of instruction by a licensed instructor, a written exam, and a marksmanship test on a firing range. These requirements reduce the likelihood that a CHL holder will be someone who engages in what we might call street, or blue-collar, crime.25 The demographic characteristics of the average CHL holder in Texas who meets these requirements are clearly related to the rates of criminal activity in this group; CHL holders are mostly middle-aged Whites, many of whom are women.18 This group has not traditionally proven to be a high-crime segment of our population.
CHL holders rarely “break bad.” When they do cross the line into illegality, the types of criminality for which they are convicted differ significantly from the convictions of nonlicensees. These differences might arise from the differences in demography or from differences in the easy availability of handguns. Assault, robbery, and burglary were the crimes in almost three quarters of the convictions we reviewed. CHL holders’ convictions were much less likely to involve these high-prevalence crimes than were the convictions of nonlicensees. On their face, these differences would seem to derive from demography. Burglary, robbery, and assault are usually the province of younger, male offenders.26
By contrast, CHL holders’ convictions were much more likely than nonholders' convictions to be for a sexual offense (ratio = 2.25). Demography likely lies at the root of this difference. Many of these convictions among CHL holders involved sexual offenses against children, crimes that are almost solely the province of adult men, often those middle-aged or older.27
The most reasonable explanation for the remaining differences lies in the availability of a handgun. Weapons offenses often involve the unlawful carrying of a handgun, and CHL holders have ready access to a handgun they can easily carry in a manner that is considered criminal (e.g., into a school or a sporting event). Deadly conduct demands possession of a firearm, and CHL holders become licensed expressly so that they can carry a concealed firearm. Intentional injuries with a firearm can easily lead to death, and CHL holders apply for a license so that they can legally carry a concealed deadly weapon. These realities made it sensible to conclude that the concentration of convictions for weapons offenses, threatening someone with a firearm, and intentionally killing a person stem from the ready availability of a handgun for CHL holders.
Our study did not address the issue of whether CHL laws significantly lower or raise crime levels. Nonetheless, our results have some policy implications. Our findings indicate that the safe hands into which states put concealed handguns may not be as safe as one would hope. Ideally, further research will identify characteristics of those CHL holders who have some likelihood of engaging in firearm-related crimes and allow policymakers to develop licensure rules that would deny those individuals a CHL. Unfortunately, the political dynamics of handgun-related licensure, legislation, and research make such an outcome unlikely. Instead, future legislation concerning handgun regulation may follow the path blazed by House Bill 1220 in New Hampshire, which attempts to reduce screening criteria for the purchase of a firearm, in this case repealing the requirements for criminal history and protective order reviews prior to a sale.28
Our research may shed some light on another major issue in the CHL debate. The drive of the National Rifle Association and other organizations is to reduce the number of settings where carrying a concealed handgun is prohibited.29 Legislative bodies in many states have already approved, or are considering, legislation to reduce the scope of these prohibitions. These proposals include allowing concealed carry by students and faculty in college classrooms,30 by faculty in public schools,31 in national parks,32 in state parks,33,34 and in churches.34 The foundation for these proposals is straightforward; advocates of these changes argue that the presence of legally armed civilians in these new settings will decrease the likelihood or consequences of crime—especially the occurrence of those rare events involving multiple victims of gun violence.35
Our results imply that opening these settings to CHL holders carrying handguns may increase gun-related offenses in those previously gun-free zones. As the numbers in Table 1 indicate, these increases will not be dramatic; our results imply that the increase in the number of gun offenses or amount of gun violence committed by CHL holders in these new settings will be low. Nonetheless, policymakers should balance this likelihood against the likelihood that CHL holders will encounter and can positively affect the calamitous situations that often lead to the demand that carry restrictions be reduced.
Our study had several limitations that should also be considered in interpreting our results. The data came from 1 period in 1 state. The data were aggregated, so individual covariates could not be included in the analyses. Finally, we could not precisely decompose the differences in criminality between CHL holders and nonholders into those portions driven by demographic characteristics or CHL status.
Our findings introduce a somewhat different question into the dialogue about CHL licensing and criminality in American society. Holders of a CHL in Texas in 2001 to 2009 were almost universally a law-abiding population, like most individuals who shared their demographic characteristics. However, in those rare instances when they committed crimes—by contrast to criminality among nonholders of CHLs—they were more likely to be convicted for serious weapons-related offenses (illegally carrying a firearm, threatening persons with a firearm, or intentionally killing another person). The policy implications of this finding deserve consideration in discussions of changes in CHL policy, especially discussions of who may receive a CHL and where they can carry their handguns.
Human Participation Protection
This study was approved by the Texas A&M University internal review board.
