ABSTRACT
We examine whether the clustering of certain types of higher-likelihood-of-recidivating parolees in neighborhoods differentially influences violent and property crime. We also test whether the relationship between the concentration of certain types of parolees and crime is moderated by disadvantage. We examine parolees released between 2000 and 2009 in Cleveland, Ohio, and neighborhood crime data. Results suggest that increases in certain types of parolees contribute to a corresponding increase in crime. This suggests that risk factors associated with reoffending might explain larger crime trends in neighborhoods. Furthermore, the broader neighborhood context compounds these risk factors, resulting in higher rates of crime.
Notes
1 Others have argued that parolees contribute to the higher crime rate indirectly through their negative impact on residential stability (Rose and Clear Citation1998).
2 Although this method cannot account for crimes committed specifically by parolees, given high parolee recidivism rates it is likely that parolees may be responsible for at least a sizeable portion of these crimes.
3 Address data were available annually between 2000 and 2003, while addresses were available every six months beginning in 2004. While it would be preferable to have address data for each move a parolee makes, residing at a residence for a relatively short period of time reduces the likelihood that a parolee will have much effect on neighborhood characteristics. However, there is some evidence that parolees are fairly stable in terms of their living situation; a recent study also in Cleveland found that approximately 72% of a sample of parolees lived at the same location 1–2 years later (La Vigne and Parthasarathy 2005). To deal with the varying frequency of address information, the data were weighted to account for the actual amount of time at a particular address in a given year. Thus, for the semiannual addresses in 2004–2009, each address was weighted to account for the potential that a parolee could have resided in more than one neighborhood in a particular year. When parolees changed residence for any reason, the address was weighted to account for this proportional influence.
4 While it would have been more desirable to capture whether an offender had ever committed a violent or serious crime, a complete criminal history was not available for all of the parolees in the study. Consequently, information on the most recent conviction offense was used.
5 Parolees classified under intense supervision are categorized with high.
6 For the purposes of this study, we include only a measure of the general parolee population in nearby neighborhoods and we do not disaggregate by parolee type for the spatial lag. Future research should assess how different types of parolees in nearby neighborhoods may differentially affect crime in the focal neighborhood.
7 In this case income inequality is measured as the distribution of income in a block group given its population relative to the income distribution for the city as a whole. The Gini coefficient of inequality (G) is calculated as: where x is an observed value, n is the total number of values and
is the mean. For more information on the Gini coefficient see Hipp (Citation2007) or Nielson and Alderson (1997).
8 Given that the racial distribution in Cleveland is predominantly Black (51%) and non-Latino white (43%), and highly segregated, other measures of race/ethnicity were not included.
9 We also correlate the disturbance terms for each outcome variable over time in adjacent years to account for any additional autocorrelation that may occur. This implies that disturbance ε1t was allowed to covary with ε1t-1, ε2t with ε2t-1, and ε3t with ε3t-1. This reduces any bias that maybe be generated by the lagged dependent variables (Nickell, 1981).
10 This was derived by multiplying the coefficient for parolees by 100. Since crime was log transformed, it is necessary to multiply the linear coefficients by 100 for an accurate interpretation of the findings.
11 Given recidivism research which indicates that rates of failure may be higher in minority reentrants, we include this measure in the model as a control. Ancillary models were tested in which the percent of Black parolees was not included in the models, and the results are fairly consistent with the models presented here. We choose to include this measure given its robustness across all models.
12 In Model 7, assessing the impact of parolees supervised average on the property crime rate, the impact of the percent parolees in general is positive, though non-significant. In all other models, the percent parolees is negatively related to crime when the percent Black parolees is included (though it is not significant in every model).
13 Full results are available by request from the authors.
14 We also tested an interaction between concentrated disadvantage and the presence of parolees generally, finding a positive and significant effect for violent crime only. For brevity, results are not shown.
Additional information
Funding
Notes on contributors
Alyssa W. Chamberlain
Alyssa W. Chamberlain is an Assistant Professor in the School of Criminology and Criminal Justice at Arizona State University. Her research focuses on neighborhood dynamics related to changes in social and demographic structure, housing, and inequality and the spatial distribution of crime. She also examines issues related to prisoner reentry and recidivism, and how incarceration and reentry both affect and are affected by community structure.
Lyndsay N. Boggess
Lyndsay N. Boggess is an Associate Professor in the Department of Criminology at the University of South Florida. Her research focuses on communities and crime, primarily how crime affects and is affected by neighborhood change, racial/ethnic composition, and economic investment in the housing market. Her work has appeared in Criminology, the Journal of Quantitative Criminology, and Crime & Delinquency.