Estimating the Causal Effect of Gun Prevalence on Homicide Rates: A Local Average Treatment Effect Approach
Tomislav Kovandzic , Mark E. Schaffer & Gary Kleck
Objectives: This paper uses a “local average treatment effect” (LATE) framework in an attempt to disentangle the separate effects of criminal and noncriminal gun prevalence on violence rates. We first show that a number of previous studies have failed to properly address the problems of endogeneity, proxy validity, and heterogeneity in criminality. We demonstrate that the time series proxy problem is severe; previous panel data studies have used proxies that are essentially uncorrelated in time series with direct measures of gun relevance. Methods: We adopt instead a cross-section approach: we use US county-level data for 1990, and we proxy gun prevalence levels by the percent of suicides committed with guns, which recent research indicates is the best measure of gun levels for crosssectional research. We instrument gun levels with three plausibly exogenous instruments: subscriptions to outdoor sports magazines, voting preferences in the 1988 Presidential election, and numbers of military veterans. In our LATE framework, the estimated impact of gun prevalence is a weighted average of a possibly negative impact of noncriminal gun prevalence on homicide and a presumed positive impact of criminal gun prevalence. Results: We find evidence of a significant negative impact, and interpret it as primarily “local to noncriminals”, i.e., primarily determined by a negative deterrent effect of noncriminal gun prevalence. We also demonstrate that an ATE for gun prevalence that is positive, negative, or approximately zero are all entirely plausible and consistent with our estimates of a significant negative impact of noncriminal gun prevalence. Conclusions: The policy implications of our findings are perhaps best understood in the context of two hypothetical gun ban scenarios, the first more optimistic, the second more pessimistic and realistic. First, gun prohibition might reduce gun ownership equiproportionately among criminals and noncriminals, and the traditional ATE interpretation therefore applies. Our results above suggest that plausible estimates of the causal impact of an average reduction in gun prevalence include positive, nil, and negative effects on gun homicide rates, and hence no strong evidence in favor of or against such a measure. But it is highly unlikely that criminals would comply with gun prohibition to the same extent as noncriminals; indeed, it is virtually a tautology that criminals would violate a gun ban at a higher rate than noncriminals. Thus, under the more likely scenario that gun bans reduced gun levels more among noncriminals than criminals, the LATE interpretation of our results moves the range of possible impacts towards an increase in gun homicide rates because the decline in gun levels would primarily occur among those whose gun possession has predominantly negative effects on homicide.
Heterogeneity in the Frequency Distribution of Crime Victimization
Tim Hope & Paul A. Norris
Objectives:
Tests the idea that the frequency distribution typically observed in crosssectional crime victimization data sampled from surveys of general populations is a heterogeneously distributed result of the mixing of two latent processes associated, respectively, with each of the tails of the distribution. Methods: Datasets are assembled from a number of samples taken from the British Crime Survey and the Scottish Crime Victimization Survey. Latent class analysis is used to explore the probable, latent distributions of individual property crime and personal crime victimization matrices that express the frequency and type of victimization that are self-reported by respondents over the survey recall period. Results: The analysis obtains broadly similar solutions for both types of victimization across the respective datasets. It is demonstrated that a hypothesized mixing process will produce a heterogeneous set of local sub-distributions: a large sub-population that is predominantly not victimized, a very small ‘chronic’ sub-population that is frequently and consistently victimized across crime-type, and an ‘intermediate’ sub-population (whose granularity varies with sample size) to whom the bulk of victimization occurs. Additionally, attention is paid to the position of very high frequency victimization within these sub-populations. Conclusions: The analysis supports the idea that crime victimization may be a function of two propensities: for immunity, and exposure. It demonstrates that zero-inflation is also a defining feature of the distribution that needs to be set alongside the significance that has been attached to the thickness of its right tail. The results suggest a new baseline model for investigating population distributions of crime victimization.
The Incapacitation Effect of First-Time Imprisonment: A Matched Samples Comparison
Hilde Wermink , Robert Apel , Paul Nieuwbeerta & Arjan A. J. Blokland
Objectives: The logic of incapacitation is the prevention of crime via the forced removal of known offenders from the community. The challenge is to provide a plausible estimate of how many crimes an incarcerated individual would have committed, were s/he free in the community rather than confined in prison. The objective of this study is to provide estimates of the incapacitation effect of first-time imprisonment from a sample of convicted offenders. Methods: The data are official criminal records of all individuals convicted in The Netherlands in 1997. Two different analytical strategies are used to estimate an incapacitation effect. First, the offending rate of the imprisoned individuals prior to their confinement in 1997 provides a “within-person counterfactual”. Second, imprisoned offenders are paired with comparable non-imprisoned offenders using the method of propensity score matching in order to estimate a “between-person counterfactual”. Incapacitation estimates are provided separately for juvenile imprisonment (ages 12–17) as well as adult imprisonment (ages 18–50), and for male and female offenders. Results: The best estimate is that 1 year of incarceration prevents between 0.17 and 0.21 convictions per year. The use of additional data sources indicates that this corresponds to between roughly 2.0 and 2.5 criminal offenses recorded by the police. Conclusions: The current results suggest that, insofar as imprisonment is used with the primary goal of reducing crime through incapacitation, a general increase in the use of incarceration as the sanction of choice is not likely to yield major crime control benefits.
The Effect of Incarceration on Re-Offending: Evidence from a Natural Experiment in Pennsylvania
Daniel S. Nagin & G. Matthew Snodgrass
Objectives: This paper uses a sample of convicted offenders from Pennsylvania to estimate the effect of incarceration on post-release criminality. Methods: To do so, we capitalize on a feature of the criminal justice system in Pennsylvania—the county-level randomization of cases to judges. We begin by identifying five counties in which there is substantial variation across judges in the uses of incarceration, but no evidence indicating that the randomization process had failed. The estimated effect of incarceration on rearrest is based on comparison of the rearrest rates of the caseloads of judges with different proclivities for the use of incarceration. Results: Using judge as an instrumental variable, we estimate a series of confidence intervals for the effect of incarceration on one year, two year, five year, and ten year rearrest rates. Conclusions: On the whole, there is little evidence in our data that incarceration impacts rearrest.
Prisons and Crime, Backwards in High Heels
William Spelman
Objectives: Prisons reduce crime rates, but crime increases prison populations. OLS estimates of the effects of prisons on crime combine the two effects and are biased toward zero. The standard solution—to identify the crime equation by finding instruments for prison—is suspect, because most variables that predict prison populations can be expected to affect crime, as well. An alternative is to identify the prison equation by finding instruments for crime, allowing an unbiased estimate of the effect of crime on prisons. Because the two coefficients in a simultaneous system are related through simple algebra, we can then work backward to obtain an unbiased estimate of the effect of prisons on crime. Methods: Potential instruments for crime are tested and used to identify the prison equation for the 50 U.S. states for the period 1978–2009. The effect of prisons on crime consistent with this relationship is obtained through algebra; standard errors are obtained through Monte Carlo simulation. Results: Resulting estimates of the effect of prisons on crime are around −0.25 ± 0.15. This is larger than biased OLS estimates, but similar in size to previous estimates based on standard instruments. Conclusions: When estimating the effect of a public policy response on a public problem, it may be more productive to find instruments for the problem and work backward than to find instruments for the response and work forward.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.