Raiding Crack Houses: The Kansas City Experiment

The study design focused on indicators of public order and safety in testing the effectiveness of raids. These indicators were limited to offense reports and calls for police service. Blocks were selected for study that had at least 5 calls for police service in the 30 days preceding an undercover crack buy. Results showed 98 raids prevented a total of only 35 reported crimes (1 crime for every 3 raids) and 85 calls for police service (less than 1 call per raid). Even these small effects were extremely short-lived, disappeared after 12 days, and were further reduced by displacement. Only 23 of the 98 raids produced any arrests, but effects of the raids were substantially the same regardless of whether arrests were made. Raids had more effects on calls for police service in the winter than in warmer weather. Effects on offense reports, however, did not change with the seasons. With a 40-officer squad producing at most 800 raids a year, the full labor cost per raid, including training, undercover buys, and preparation, was about 2.6 officer-weeks. With 2 officers shot in the first 1,895 raids, the estimated risk-benefit ratio was 284 crimes prevented for every officer shot. Citizen calls to the police drug hotline increased after the raids. Implications of the results for police crime prevention and drug control policies are discussed. 9 references, 2 tables, and 3 figures