can yield estimates of the cost of different types of crime, the data requirements for such a project appear to be too onerous, and the property value studies to date have only estimated the cost of a generic "index crime" (rape, robbery, assault, motor vehicle theft, burglary, and larceny). Phillips and Votey (1981) took a different approach by combining estimates of the value of a statistical life with crime seriousness rankings from the public surveys of Sellin and Wolfgang (1964). Their approach was the first that attempted to account for the intangible costs of individual crimes, but it still was not able to account for many of those intangible costs.

Cohen (1988) was the first study to explicitly attempt to place monetary values on pain, suffering, and lost quality of life to victims of different crimes. Although that study had numerous data limitations (including using short-term losses in the absence of long-term data and relying primarily on NCVS incidence estimates), it used a similar methodology to that used in this study by combining jury awards and the statistical value of life to estimate intangible losses. Cohen (1990) compared these estimates to the earlier property value studies and found that the two different approaches yielded very similar estimates of the cost of an "index crime." Several followup studies have updated these estimates by adding more comprehensive information, including Cohen, Miller, and Rossman (1994); and Miller, Cohen, and Rossman (1993).

Table 9 compares both the individual crime cost estimates and the aggregate United States costs estimated here to several earlier studies. For individual crimes, this study has estimated significantly higher costs for rape than previous studies. For example, the average cost of a rape (including attempts) is estimated to be $87,000, whereas previous estimates ranged from $55,000 to $70,000. This difference is primarily due to the fact that jury awards in rape cases were analyzed to determine estimates of pain, suffering, and lost quality of life, whereas earlier studies employed even more indirect methods of estimation (such as jury awards in more generic cases of emotional distress). Further, earlier studies did not make clear distinctions between victims and victimizations. They often used victim counts even though they estimated costs on the basis of victimizations. This study's estimate of the cost per rape victim is $110,000.

This study's estimates are substantially higher than Government estimates published from the NCVS. For example, the medical, mental health care, public services, and productivity costs are more than 10 times the short-term loss components captured in the NCVS for rape and assault and 5 times the costs captured for robbery. This is primarily because this study's methodology included long-term medical and wage costs, mental health care, valuation of lost household and school hours, and the administrative (insurance and Government) costs of reimbursement for many of these losses. Nor did the researchers follow the extremely conservative approach of ignoring costs acknowledged to be real though not directly supplied by NCVS survey respondents. For example, some NCVS survey respondents reported that they did incur costs (medical, property loss, etc.) but were unsure of the dollar amount. Klaus (1994) treats these responses as a zero-cost, whereas this study assigned these cases the mean cost among similar cases with known costs.

The quality of life and mental health care costs are lower than in earlier studies. Both the quality of life costs and mental health care costs in the earlier studies were based on jury awards for auto crashes and tort cases; but data on jury awards for pain and suffering in criminal victimization cases are now available. For mental health care costs, survey data on victim use from mental health care providers are now also available.

Aggregate cost estimates have also increased significantly for some crimes. The largest increase is for rape, since the cost per incident is higher and this study's estimate of the number of rapes has increased tenfold. For example, Cohen, Miller, and Rossman (1994) estimated the aggregate cost to be $10 billion (in 1993 dollars), based on 147,000 NCVS rape victims (excluding series victims) and $68,800 per victim. This figure was subsequently updated to include series victimizations (with series=1), reported in Miller, Cohen, and Rossman (1993). The latter estimate resulted in aggregate costs of $16 billion (in 1993 dollars), based on 229,000 rape victims and $68,800 per victim. This study estimates the aggregate cost of a rape to be $127 billion, based on 1,467,000 victimizations and a cost per victimization of $87,000.