this information when it was missing. Following the procedure in Miller (1993), they estimated days of household work lost from days of work lost. Based on employment status and age reported in NCVS, they estimated the value of housework days lost by victims from the regression equations in Douglass et al. (1990). Since the NCVS only covers short-term costs, the resulting data do not reflect lost productivity that occurs more than a few months after the victimization. Although most crime victims have few long-term productivity losses, some suffer substantial wage losses. Thus, the researchers used the expected lifetime earnings (absent an injury) from a standard age-earnings model similar to the model described in Rice et al. (1989) and permanent disability probabilities from workers' compensation claims (by body part and nature of injury) to estimate long-term productivity losses by body part and nature of injury. Those estimates were then applied to the distribution of physical injuries in criminal victimizations.

Because of age-related differences in recuperative power, the probabilities may underestimate the disability impacts of crime against the elderly and infants or overestimate the impacts of crime against youths. They also may underestimate disability probabilities due to the greater force and targeting of vulnerable areas involving an intentional injury as compared to a typical workplace injury. For example, a knife that slips while slicing a bagel is likely to inflict less damage than a knife that disarms a robber. This has a dramatic effect on estimated wage losses, with long-term losses averaging 5 to 10 times the short-term losses. For example, short-term productivity loss for the average assault victim was estimated to be $356, while long-term productivity losses were estimated to be $2,035.

Several smaller loss components were also estimated. NCVS includes data on the number of hours of work and wages lost by family members aiding victims and being involved with victims in the legal system. The first component adds about $5 to $10 per victim and the second adds about $10 to $20 per victim. For victims under age 12, this study assumed that an equal number of days of parent productivity are lost when a child victim is medically treated as when a family member loses work due to victimization of an unmarried, unemployed relative. The estimated losses average $65 to $85 per child victim. Employers also suffer indirect losses due to their employee's victimization, such as management time hiring replacements and otherwise dealing with the emotional trauma associated with an employee victimized by crime. Based on earlier studies by Miller (1992) and Miller and Rossman (1990), this study very roughly estimated these losses to be approximately $120 to $140 per rape, robbery, or assault victim.

Finally, school-age children who are victimized may be absent from school while they are recuperating. Based on the number of work days lost for comparable crimes and injuries reported in the NCVS, the number of school days lost for child victims were estimated. The researchers adopted a straightforward and conservative approach to valuing a lost school day by using the average amount spent on education per child per day (see Cohen, Miller, and Rossman, 1994, for details). Placing dollar values on the returns to education is a more appropriate measure of the loss but is a complicated theoretical and empirical exercise that is beyond the scope of this project. However, since the average number of school days lost is relatively small, this cost category will be a small portion of overall costs regardless of the valuation approach.

Intangible losses

Unlike tangible losses such as medical expenses or lost wages, pain, suffering, and reduced quality of life do not have a market price and cannot be bought and sold. Nevertheless, these losses are real. Victims would pay dearly to avoid them.

Economists use varied methods to place monetary values on the intangible losses (see Cohen, Miller, and Rossman, 1994). Monetary estimates of lost quality of life due to fatalities are generally based on the amount people routinely spend (in dollars or time) to reduce their risk of death. (Miller, 1990; and Viscusi, 1993, review the extensive literature on this subject.) Typical studies have assessed the market for safety by examining the increased demand for smoke detectors as prices dropped, the demand for safer cars, and the differential in worker pay due to different levels of risk exposure. Estimates of the monetary value of a fatality (including lost wages and other tangible losses) range from $500,000 to $7 million. This study adopted the mean estimate from the literature review and synthesis by Miller (1990-adjusted to reflect victims' life expectancies), a total loss of about $2.7 million (inflated to 1993 dollars), with the lost quality of life component being about $1.9 million.