The study hypothesizes that individuals' decisions regarding crime depend on perceived sanction levels that, in turn, depend on sanction levels realized in past periods. The study's econometric results support this view. Although crime rates appear to be little affected by sanction levels, they certainly are not random phenomena. They are related to a number of exogeneous economic and demographic variables accounting for a major share of observed variation in property crime rates. In particular, rising unemployment accounts for about 10 percent of the 1970-77 average increase in burglary rates and about 15 percent of the average increase in larceny rates. Thus, burglary and larceny could probably be controlled as much by reducing unemployment as by increasing the risk and severity of criminal sanctions. Footnotes and about 30 references are included. (Author abstract modified)