The paper features an analysis of a study conducted for the Association of Chief Police Officers that had three objectives. One objective was to determine as accurately as possible, using existing published information sources, the nature, extent, and cost of fraud to the public and private sectors. A second objective was to assess the availability and quality of existing evidence on fraud; and the third objective was to recommend appropriate strategies for facilitating the comprehensive and consistent recording of fraud data. An overview of the study's approach focuses on problems with existing data sources. This is followed by a discussion of the nature of fraud, which addresses the definition of various types of fraud, identification of fraud victims, and the application of a typology to fraud. A section on the assessment of the extent and costs of fraud notes the costs of fraud considered in the research. They included victims' direct losses, the costs of preventing fraud, and the costs of responding to fraud after it occurs. Fraud losses affecting the private and public sectors are discussed separately. A section on weaknesses in the current fraud data focuses on the definition of "fraud," accounting for differing purposes of each data-collection exercise, readiness to participate in fraud surveys, ensuring full survey participation by victims, determining the unit of analysis, determining where the fraud was committed, accounting for the time lapse between the commission and reporting of fraud, aggregating data and deriving nationwide estimates, and irregularity in data-collection procedures. 2 tables and 45 references