Gary L. Shoesmith

Crime, Teenage Abortion, and Unwantedness

This article disaggregates Donohue and Levitt’s (DL’s) national panel-data models to the state level and shows that high concentrations of teenage abortions in a handful of states drive all of DL’s results in their 2001, 2004, and 2008 articles on crime and abortion. These findings agree with previous research showing teenage motherhood is a major maternal crime factor, whereas unwanted pregnancy is an insignificant factor. Teenage abortions accounted for more than 30% of U.S. abortions in the 1970s, but only 16% to 18% since 2001, which suggests DL’s panel-data models of crime/arrests and abortion were outdated when published. The results point to a broad range of future research involving teenage behavior. A specific means is proposed to reconcile DL with previous articles finding no relationship between crime and abortion.

Insights into the contestability and sustainability of the U.S. petroleum refining industry

Contestable market theory has been advanced in recent years by several authors, the most thorough treatment being the book by Baumol, Panzar, and Willig (1982) (BPW). Although several applications of the theory have appeared in the literature, few (if any) follow the step-by-step procedure set forth in Chapter 16 of BPW. This paper represents one of the first attempts to follow the general procedure in investigating the contestability and sustainability of an industry. The long-run cost structure of petroleum refining is first estimated using a translog multiproduct cost function and company refining data from the Financial Reporting System maintained by the U.S. Energy Information Administration (EIA). The estimated cost function is then used to approximate the optimal (cost-minimizing) industry structure for petroleum refining. Given that the existing structure is significantly different than the optimal structure, the principles of contestable market theory are used in evaluating the contestability and sustainability of the industry in light of adjustments in industry structure and the performance of various classes of refining companies between 1981 and 1987.

State Prisoner Elasticities and Integrated State Policies Using Non-Prison Alternatives

Research Summary: This study estimates national and state prisoner elasticities for violent crime, property crime, murder, rape, robbery, assault, burglary, larceny, and auto theft, benchmarking the results against previous studies. The state elasticities vary widely across states, implying that the use of national estimates at the state level will often lead to misallocations of state funds. In modeling crime, experiments are conducted to assess the use of first-differences and lagged dependent variables. The initial results here suggest that it may be possible to closely approximate national and state prisoner elasticities using lagged dependent variables as substitutes for control variables. Policy Implications: The main purpose of the paper is to demonstrate an integrative approach to state policymaking aimed at reducing crime and costs. For each state, the prisoner elasticities are combined with current data on crime and imprisonment, prison maxouts, time served, and high school education in an effort to formulate coordinated policies based on recent research on the criminogenic effects of prison, deterrence, social and other non-prison alternatives, and compulsory education. Broad-brush policies are proposed for 19 states, with special attention to Louisiana and Florida.