The direct estimate approach to the costs of crime to victims is unsatisfactory, since it is bound to underestimate such costs: many of the goods and property stolen or destroyed as a consequence of crime have a true value far above their market value. In personal injury crimes, for example, the loss of earnings is neither the only, nor the major cost involved: there may be additional expenses caused by heightened fear of crime (e.g., riding a taxi instead of public transportation). In addition, the fear of crime is a true cost. The examination of the impact of crime on rents and property values can, in principle, capture all these subjective costs: the housing market is basically an implicit market, on which people reveal the costs of crime as they themselves perceive them. A regression estimate analysis, based on the 1970 Census of Housing and on Chicago Police Department Crime Summaries for various precincts, made it possible to compute the effects of the crime coefficient, modified by such factors as nonincome and income, upon property values and rents. Mean crime rates over determinate periods in the precincts chosen, and half that rate, were the values set by the experimenter to calculate the incremental cost of one rise in crime from one-half the current mean rate to the mean. The results of these calculations are tabulated in the text. Since none of the observations in the samples had a zero crime rate, the results only had to be extrapolated to crime rates within the range of the samples themselves. The results given are, therefore, partial crime costs and not total crime costs, because it would have been practically impossible to answer the larger question concerning the cost of the total amount of existing crime on the basis of the Police Department's, incomplete records. Ten explanatory footnotes are appended.