- Mark Derham
The study of crime in all its forms by scholars is a mainstay of academia. Research on this topic spans across a wide breadth of related issues, but the main concern of this essay is crime and the effect concealed handgun laws have on it. These laws provide citizens a legal means by which to carry a concealed handgun by having them undergo a background check through their local law enforcement agency to ensure they have no criminal history and are not suffering from a mental health illness (Lott and Mustard 1997).This particular issue was brought to the forefront of the academic debate by Lott and Mustard (1997) in an attempt to apply empirical, quantitative analysis to the issue to assist in the formulation of public policies regarding handguns and the right-to-carry. Following the publishing of Lott and Mustard (1997), numerous other scholars joined in the debate on concealed handgun laws and the effect such laws have on the crime rate (see Bartley and Cohen 1998; Black and Nagin 1998; Ludwig 2000; Plassman and Tideman 2001; Ayres and Donohue III 2003; Plassman and Whitley 2003; Moody and Marvell 2008; etc.). What has resulted is a collection of work showing a divide amongst scholars and a debate on whether or not concealed handgun laws have a negative, neutral or positive effect on crime rate and the proper method for analyzing the available data.
School shootings, gunmen attacking innocent civilians in public areas, and other forms of gun violence continue to put people on high alert when it comes to the issue of guns, their availability, and the ease in which one can lawfully carry said guns. These forms of crime in relation to gun laws also continue to play a big role in local, state and national elections, with party lines having already been drawn at all levels. As a result of the national attention crime receives and the impact it has on political debates, it is critical to understanding the issue of crime, and how local, state and national laws impact crime levels. With this information in mind, this essay is specifically concerned with the following question: why does the crime rate vary? With such a question, it is expected that there are numerous reasons that variance is to be found, and several variables will be set up and reviewed to determine their effect on the crime rate.
The literature surrounding concealed carry laws is extensive due in large part to Lott and Mustard (1997) providing a spark for other scholars to join the debate. The results in the literature also reveal that scholars are deeply divided on the effects concealed carry laws have on the crime rate, and the methods by which researchers believe this issue should be addressed also widely vary. While there are not necessarily key findings due to the divide amongst scholars, there are several generalizations that can be gathered from the literature. First, crime data can be difficult to interpret due to inconsistency in reporting and classification of crimes over time (Lott and Mustard 1997; Ludwig 2000; Ayres and Donohue III 2003). Second, the independent variables used to control for other factors that affect the crime rate can have a significant impact on the end results, which comes as no surprise (Ludwig 2000; Ayres and Donohue III). Third, the unit of analysis used to study the issue has been shown to have a significant impact on the results due mainly to crime reports and different independent variables affecting different regions (Ayres and Donohue III; Black and Nagin 1998; Lott and Mustard 1997; Ludwig 2000). Fourth, the majority of the literature maintains even when lower violent crime rates cannot be attributed to concealed carry laws, there is also little evidence of higher violent crime rates as a result of these laws (Bartley and Cohen 1998; Black and Nagin 1998; Lott and Mustard 1997; Ludwig 2000; Moody and Marvell 2008; Plassman and Tideman 2001; Plassman and Whitley 2003).
With such variance over all aspects of concealed handgun laws effects on crime rate, there is also a large variance in the results that scholars have achieved. Lott and Mustard (1997) originally determined that concealed handgun laws reduce murders, rapes, and aggravated assaults; property crimes increase when the laws go into effect; and the laws deter all types of crime with murder, rape, and aggravated assault being the most affected. Bartley and Cohen (1998) found a reduction in violent crime rates and no substitution effect for property crimes. Plassman and Tideman (2001) saw a deterrent for murder, rape and robbery in ten of the states they analyzed, but also recognize the potential for an increase in crime for the other half of states they reviewed. Plassman and Whitley (2003) conclude the laws do reduce murder rates, and Moody and Marvell (2009) also conclude there is generally a reduction in overall crime. Thus, even amongst the scholars that see a positive effect on concealed carry laws, the results are varied on what the precise effect is.
The other side of the debate generally recognizes that the laws maintain a neutral effect on overall crime, although, Ayres and Donohue III (2003,2009) argue that their model shows an increase in crime once concealed handgun laws are put in place.They also reject the majority of the literature supporting a reduction in crime based on issues with the original data sets utilized by Lott (2000). Once the data sets are extended to include the years up to 2000 and corrections are made to original coding issues, the results of previous studies (Lott 2000; Plassman and Whitely 2003) are reversed (Ayres and Donohue III 2009). Once the data sets extend up to 2006, the issue becomes more complicated and only shows an increase in aggravated assaults with no significant results for murder and robbery (Ayres and Donohuee III 2009b).Black and Nagin (2008) argue in their analysis that there is too much sensitivity for the model and sample size and qualify this by showing that Lott and Mustard’s (1997) results are dependent on Florida being included in the analysis. Once Florida is removed from the sample, there is no deterrence effect for murder and rape and assaults are unaffected. They conclude that Lott and Mustard’s results should not affect public policies due to the sensitivity of the analysis (Black and Nagin 2008). Finally, when analyzing only homicide rates and attempting to control for unobserved variables by using juveniles as a control group, there is a positive effect for concealed carry laws, but it is not significant (Ludwig 2000). Once again, the varied results complicate the overall analysis and provide questions on the overall reliability and accuracy of the results.
The data used to study this issue were derived by pulling statistics and information from the Federal Bureau of Investigation’s (FBI) Uniform Crime Report (UCR) and then supplementing that data with state reported statistics obtained through various state agencies for conviction rates, sentencing length, and concealed handgun permits issues (Lott and Mustard 1997). This data pool is criticized to an extent due to the known issues with the UCR data being a victim of non-reporting, incomplete reporting, and under reporting (Ludwig 2000). However, there appears to be an agreement that this is the best available data since the majority of the available literature utilized Lott and Mustard’s data to conduct their own analysis (Ayres and Donohue III; Bartley and Cohen 1998; Black and Nagin 1998; Lott and Mustard 1997; Ludwig 2000; Moody and Marvell 2008; Plassman and Tideman 2001; Plassman and Whitley 2003). There has also been work completed to supplement Lott and Mustard’s data by including additional years of statistics up to 2006 (Ayres and Donohue III 2009b; Moody and Marvell 2008).
There is also the issue of obtaining data regarding a state’s concealed handgun permits. Since there only six states that had any reportable data available regarding the number of concealed handgun permits issued, there was also agreement that a dichotomous value was required to assess the status of a states concealed handgun laws (Ayres and Donohue III 2003; Bartley and Cohen 1998; Black and Nagin 1998; Lott and Mustard 1997; Ludwig 2000; Moody and Marvell 2008; Plassman and Tideman 2001; Plassman and Whitley 2000). This continues to be a contentious point amongst scholars since there is no measure for the number of issued concealed weapon permits by state.
The unit at which the studies were analyzed varied, and there is some level of debate as to which unit of analysis is best able to interpret the available data. Lott and Mustard (1997) argue that the county is the best unit of analysis based on the heterogeneity of states and the effects this carries over to state level analysis. Additionally, they argue that the city level isn’t appropriate due to the lack of time-series data once the laws are put in place (Lott and Mustard 1997). However, the county as the unit of analysis is rejected by several other scholars. Black and Nagin (1998) use the county as their unit of analysis but reject Lott and Mustard’s results based on their acceptance of all counties, which results in the dropping of a large amount of data due to some crimes not being present in all counties each year. Thus, Black and Nagin (1998) set a population limit on the counties being accepted into the analysis. On the other hand, Plassman and Tideman (2001) find that limiting counties by size results in a loss of significant value to the models and instead use the state to discover a significant variation in results across multiple states. Finally, Ayres and Donohue III (2003, 2009b) run models utilizing both state and county levels to test for reliability and accuracy of the available data. Additionally, the state level data was able to be extended using the UCR through 2006. Utilizing county level data is dangerous due to the sheer inaccuracy of it, amongst other issues (Ayres and Donohue III 2003).What is clear from each of these studies is that the unit of analysis can certainly vary, and it is necessary to weigh the advantages and disadvantages of each before moving forward.
The methods used by scholars are the true dividing force in the debate over concealed handgun laws. There seems to be little agreement on which methods produce the most accurate results and each subsequent scholar that joins the debate finds a new method to interpret the available data. The only consensus that is to be found in the literature supports a cross-sectional, time series study. Following that, the models used to analyze the data vary widely. Lott and Mustard (1997) utilized a least squares regression model along with Black and Nagin (1998) and Bartley and Cohen (1998). Plassman and Tideman (2001) found this causes unreliable results since the least squares method is not suitable for the low levels of occurrence for some of the crimes. They argue that a Poisson-lognormal model is much more accurate based on the available data (Plassman and Tideman 2001). Ludwig (2000) utilizes only the homicide data, due to the unreliability of other nonviolent crime data, to run a difference-in-difference-in-difference (DDD) model at the state level. Ayres and Donohue III (2003) also reject the least squares regression model based on the re-running of the analysis with their additional data sets and instead favor a hybrid model that incorporates state trends, state post-law dummies, and state post-law trends. This latter model has also been rejected due to the restriction of Ayres and Donohue III’s post passage analysis to only five years. When the model was extended by one more year, the results shifted completely (Moody and Marvell 2009).This is also dangerous since there was only one to three years worth of data available for the 25 states originally reviewed, so extending the analysis an additional year could have caused errors in the results (Ayres and Donohue III, 2009).
The available research on right to carry laws and their effects on the crime rate show that there is much disagreement surrounding the available data, at what level the data should be analyzed, measures utilized, and models used for the regression analysis. As a result, there is still room to provide input into this issue and no true key findings amongst scholars exist, though, general findings can be surmised. With regard to the data, there will always be inconsistencies with how crime is reported and provided to the public in the UCR. Regardless, this is the best available source for crime data and the real debate should be on what level that data is analyzed and how to control for the inaccuracy of the available data. All of the research has used either state or county level data due to the lack of city level data prior to 1985. The measures utilized also show a discrepancy amongst research as to the most appropriate variables necessary to account for other factors that affect crime. This is certainly a problem all of its own due to the demonstrated effects on results of removing certain variables and accounting for others. Thus, more work is necessary to discover the most appropriate variables to control for during the analysis. Finally, a host of models have been used to run regression analysis with each finding differing results. This calls into question the reliability of the results and the need for additional research.
With this information in mind, this study will approach the issue of the crime rate in a quantitative method. The dependent variable will be the crime rate since the issue at hand is what causes variance in the crime rate. The units of analysis will be states and time represented over a calendar year. Treatment and control states will be utilized to account for other factors that can be expected to impact crime rates. The primary variable this article will analyze is the presence of concealed handgun laws. The second variable this study expects to affect crime rate variance is the metropolitan population percentage of a state.
With the dependant and independent variables having been established, there are two hypotheses this study puts forward. First, if concealed handgun laws are in place, then the crime rate should be reduced over time. Second, if a state’s metropolitan population percentage decreases over time, then the crime rate should decrease over time. Thus, for the independent variable of concealed handgun laws, this study expects to see a positive impact to the crime rate. For the metropolitan population percentage variable, it is expected that a lower percentage over time will have a positive impact to the crime rate, and a higher percentage over time will have a negative impact.
Aikens, Steven and Slider, Gary. 2014. “US State Pages.” Accessed on March 3, 2014.http://www.handgunlaw.us/.
Ayres, Ian and Donohue III, John J. 2003.“Shooting down the “More Guns, Less Crime” Hypothesis.”Stanford Law Review 55 (April): 1193-1312.
––––. 2009. “Yet Another Refutation of the More Guns, Less Crime Hypothesis – With Some Help From Moody and Marvell.” Econ Journal Watch 6 (January): 35-59.
––––. 2009b. “More Guns, Less Crime FailsAgain: The Latest Evidence from1977 – 2006.” Econ Journal Watch 6 (May): 218-238.
Bartley, William A. and Cohen, Mark A. 1998. “the Effect of Concealed weapons Laws: An Extreme Bound Analysis.” Economic Inquiry 36 (April): 258-265.
Black, Dan A. and Nagin, Daniel S. 1998. “Do Rightâ€toâ€Carry Laws Deter Violent Crime?” The Journal of Legal Studies 27 (January): 209-219.
Lott, John R. and Mustard, David B. 1997. “Crime, Deterrence, and Rightâ€toâ€Carry Concealed Handguns.” The Journal of Legal Studies 26 (January): 1-68.
Lott, John R. 2000. More Guns, Less Crime. Chicago: University of Chicago Press.
Ludwig, Jens. 2000. “Gun Self-Defense and Deterrence.” Crime and Justice 27: 363-417.
Moody, Carlisle E. and Marvell, Thomas B. 2008).“The Debate on Shall-Issue Laws.”Econ Journal Watch 5 (September): 269-293.
Plassman, Florenz and Tideman, Nicolaus. 2001. “Does the Right to Carry Concealed Handguns Deter Countable Crimes? Only a Count Analysis Can Say.” Journal of Law and Economics 44 (October): 771-798.
Plassman, Florenz and Whitley, John. 2003. “Confirming “More Guns, Less Crime”.” Stanford Law Review 55 (April): 1313-1369.
Cite This Work
To export a reference to this article please select a referencing stye below: