In contrast to a posterior analysis given a particular sampling model, posterior model probabilities in the context of model uncertainty are typically rather sensitive to the specification of the prior. In particular, “diffuse” priors on model-specific parameters can lead to quite unexpected consequences. Here we focus on the practically relevant situation where we need to entertain a (large) number of sampling models and we have (or wish to use) little or no subjective prior information. We aim at providing an “automatic” or “benchmark” prior structure that can be used in such cases. We focus on the Normal linear regression model with uncertainty in the choice of regressors. We propose a partly noninformative prior structure related to a Natural Conjugate g-prior specification, where the amount of subjective information requested from the user is limited to the choice of a single scalar hyperparameter g0j. The consequences of different choices for g0j are examined. We investigate theoretical properties, such as consistency of the implied Bayesian procedure. Links with classical information criteria are provided. More importantly, we examine the finite sample implications of several choices of g0j in a simulation study. The use of the MC3 algorithm of Madigan and York (1995), combined with efficient coding in Fortran, makes it feasible to conduct large simulations. In addition to posterior criteria, we shall also compare the predictive performance of different priors. A classic example concerning the economics of crime will also be provided and contrasted with results in the literature. The main findings of the paper will lead us to propose a “benchmark” prior specification in a linear regression context with model uncertainty.

Abstract not found

Abstract—The labor market prospects of young, unskilled men fell dramatically in the 1980s and improved in the 1990s. Crime rates show a reverse pattern: increasing during the 1980s and falling in the 1990s. Because young, unskilled men commit most crime, this paper seeks to establish a causal relationshi p between the two trends. Previous work on the relationship between labor markets and crime focused mainly on the relationshi p between the unemployment rate and crime, and found inconclusive results. In contrast, this paper examines the impact of both wages and unemployment on crime, and uses instrumental variables to establish causality. We conclude that both wages and unemployment are signi �-cantly related to crime, but that wages played a larger role in the crime trends over the last few decades. These results are robust to the inclusion of deterrence variables, controls for simultaneity, and controlling for individual and family characteristics.

Current research suggests that the crime-combating instrument "sentence probability " is more e#ective than "sentence severity". However, the focus of the simultaneous (or very short-run) impact of the law enforcement policy on crime impedes a comparison of these two instruments with respect to their long-term e#ectiveness. With Swiss data, we investigate the dynamic interrelationships between enforcement policy and crime and find that overall, sentence severity is just as e#ective as sentence probability (and even more e#ective for certain crimes). Furthermore, we show how the VAR-modeling technique can be exploited to conveniently distinguish between deterrence and incapacitation e#ects. Therefrom, concrete recommendations for a cost-e#ective crime-reduction policy can be derived. 1 Patricia Funk & Peter Kugler, WWZ (Department of Economics), Petersgraben 51, CH-4003 Basel. Tel.: ++41 61 267 33 52, Fax: ++41 61 267 33 40, E-Mail: Patricia.Funk@unibas.ch, Peter. Kugler@unib...

$5.00 Up-to-date information about CPR’s research projects and other activities is available from our World Wide Web site at www-cpr.maxwell.syr.edu. All

This paper uses data on Finnish juvenile punishment experiment to identify the effect of sanctions on recidivism, education and employment of convicted adolescents. The experiment conducted in certain municipalities sent adolescents, who- on the basis of their criminal record, were considered unsuitable to be sentenced in prison, yet no longer eligible for parole- to a rehabilitative program that aimed to improve their social skills and attachment to labor markets. We use unique data on sentences and punishments in the years 1990-2004. The criminal data is merged with the longitudinal population census file which entails detailed information on criminal activities and socio-economic background of individuals with a criminal record. We use a differences-in-differences-in-differences approach where we control for the possibility that the overall effect of punishment is different in the experiment municipalities from that of the control municipalities. We find that juvenile punishment did not have a significant effect on recidivism, but it increased the likelihood of continuing schooling or being employed.

Although it is a widely held belief that racial bias exists in law enforcement, the courts have found it difficult to prove that an officer singled out a suspect because of race. This paper models the interaction between motorists and police and develops an alternative empirical approach to measure racial bias. In previous work, perfect observability by police is implicitly assumed, and it further assumed that motorists have control over discretionary characteristics and there are uniform regional costs of police assessment. We construct an empirical test that compares officer find rates while accounting for additional signals used by officers. We use fixed effects generated at the agency level that account for deviations in the regional cost of assessment as well as mitigate the impact of omitted variable bias. The empirical application of the model is based on a unique dataset containing the results of all breathalyzer tests administered from 2003 through 2006 by the Washington State Patrol. We test for the existence of racial bias exhibited by officers conducting assessments for driving under the influence of alcohol (DUI). Both the conditional means tests and our new parametric test indicate that officers are less accurate with Black and Asian motorists. However, the conditional means test suggests police officers are more accurate with Hispanic motorists than white motorists, which is consistent with bias in favor of Hispanic motorists. In contrast, using our parametric test that incorporates fixed effects, the effect of

Abstract: In the 1990s, while there was a large decline in property-related crime along the U.S.-Mexico border, violent crime rates began to converge to the national average. At the same time, legal and illegal immigration from Mexico surged and border enforcement rose to unprecedented levels. In this paper, we investigate the relationship between border county crime rates, immigration and enforcement since the early 1990s. We find that while the volume of illegal immigration is not related to changes in property-related crime, there is a significant positive correlation with the incidence of violent crime. This is most likely due to extensive smuggling activity along the border. Border enforcement meanwhile is significantly negatively related to crime rates. The bad news is that the deterrent effect of the border patrol diminishes over this time period, and the net impact of more enforcement on border crime since the late 1990s is zero. JEL classification: J61, K42

Abstract: In the 1990s, while there was a large decline in property-related crime along the U.S.-Mexico border, violent crime rates began to converge to the national average. At the same time, legal and illegal immigration from Mexico surged and border enforcement rose to unprecedented levels. In this paper, we investigate the relationship between border county crime rates, immigration and enforcement since the early 1990s. We find that while the volume of illegal immigration is not related to changes in property-related crime, there is a significant positive correlation with the incidence of violent crime. This is most likely due to extensive smuggling activity along the border. Border enforcement meanwhile is significantly negatively related to crime rates. The bad news is that the deterrent effect of the border patrol diminishes over this time period, and the net impact of more enforcement on border crime since the late 1990s is zero. JEL classification: J61, K42

$5.00 Up-to-date information about CPR’s research projects and other activities is available from our World Wide Web site at www-cpr.maxwell.syr.edu. All recent working papers and Policy Briefs can be read and/or printed from there as