In this paper, panel models for crime rates and arrest rates are utilized to separate the effect of law enforcement from several other processes that have been advanced as possible determinants of an enforcement-crime relationship. When models of this type are estimated for official index crime rates in a sample of U.S. cities for the years 1964-1970, no meaningful relationship between arrest rates and crime rates is found. This finding permits us to exclude the existence of any appreciable deterrence effect. (October 1979)
American Sociological Review
1979
Greenberg DF, Kessler RC and Logan CH
http://www.jstor.org/discover/10.2307/2094531?uid=3739256&sid=21101999149803