Benchmarking Correctional Facilities in the United States: An Undesirable Output Data Envelopment Analysis Model

Abstract

This study evaluated the operational efficiency of 741 U.S. correctional facilities using Data Envelopment Analysis (DEA) with five inputs and two outputs, including undesirable outputs like disciplinary reports. An undesirable-output DEA model identified 128 efficient and 616 inefficient facilities. Efficient benchmarks averaged 825 capacity, 36.3 years in age, 31 professional staff, 189 security staff, 898 inmates, 1,043 disciplinary reports, and $27.6 million in annual costs. Hypothesis testing revealed that GED, ESL, vocational training, and health services were associated with higher efficiency. The DEA model offers a valuable tool for policymakers to benchmark and improve correctional facility operations nationwide.