A Longitudinal Analysis of the Welfare-Homicide Relationship

Abstract

Drawingon the social altruism and social threat hypotheses, the present investigation challenges prior macro-level studies that assume that there is a nonrecursive relationship between welfare transfers and the aggregate level of homicide. The bivariate autoregressive integrated moving average analyses of the number of individuals receiving Aid to Families With Dependent Children (AFDC) benefits and total and disaggregated counts of homicide reveal no evidence of simultaneity bias. However, as predicted by social altruism theory, the results support the conclusion that prior research is probably tainted by offense aggregation bias. That is, the number of AFDC recipients is, as predicted, negatively associated with the level of familial homicides but is not associated with either the level of nonfamily, felony-related, nonfelony-related, or total homicides. The implications of these findings for the welfare-crime relationship are discussed.