3 minute read

Unemployment and Crime

Dependence On Labor Force Attributes

One variable that greatly affects crime and employment relationships is the age of persons who are unemployed. Government statistics regularly indicate that juveniles and young adults greatly exceed older persons in rates of arrest for burglary, robbery, and other crimes of taking property belonging to others. A 1959 pioneer study found that this inverse relationship of such crimes to unemployment was most pronounced for young persons, and less evident for older persons out of work (Glaser and Rice). In 1968 a British study yielded similar findings, showing that the relations between crime and unemployment are most intense for youths who were out of school as well as work (Farrington et al.). The old adage that "idle hands are the devil's workshop" seemed to be confirmed.

These findings also appear to support the 1999 assertion by Bruce Western and Katherine Beckett that the U.S. penal system is "a labor market regulating institution." They justified this statement by pointing out that confinement institutions remove able-bodied but idle young men from the workforce, and that once these men have a record of imprisonment, their subsequent job prospects are greatly diminished (imprisoned women are not sufficiently numerous to affect the total female labor force significantly).

Unfortunatly, any statistical generalizations on the linkage of crime to unemployment, as well as to age or other personal attributes of offenders and nonoffenders, can only be tested with imperfect data. The completeness of our knowledge on lawbreakers necessarily varies with the extent to which they are caught, and with the use of imprisonment rather than alternative penalties for those convicted. Data on employment, age, and various other attributes of persons committing crimes is usually reported for those offenders who are arrested, but their total number, and information on them is somewhat diminished (although presumably made more accurate) if one studies only those arrestees who are subsequently convicted of the crimes for which they were arrested. Furthermore, data on the personal attributes of those convicted are often not compiled in as much detail for those fined or released on probation as for those who are imprisoned.

In 1939 two European scholars, Georg Rusche and Otto Kirchheimer, refugees from Hitler's Germany, published what proved to be a classic volume of historical scholarship, Punishment and Social Structure. In it they treated the variations in reaction to crime from ancient to contemporary periods, and generalized that the types of penalties used—such as executions, transportation to distant colonies, torture, mutilation, confinement in idleness, of forced labor—depended greatly on economic conditions, particularly on the current value of labor. The cruelest penalties became most frequent, they asserted, when unemployment was extensive, making labor cheap.

These rather vague assertions were subsequently formalized and tested statistically by others, with diverse data, as can be illustrated by summarizing a few of the most methodologically sophisticated studies. Mathematician David Greenberg concluded from multivariate analysis in the 1970s, that "oscillations in the rate of admissions to prison in Canada in recent years have been governed almost entirely by changes in the unemployment rate. The same relationships appear to hold in the United States as well" (p. 651). Sociologists Andrew Hochstetler and Neal Shover, however, found in the 1990s that the intensity of this relationship in the United States varied greatly in different historical periods, and in various regions.

Hochstetler and Shover note that the incarceration rate in the United States, defined as the number of imprisoned adults per 100,000 population, was 462 for the southern states in 1994, but only 291 for the northeastern states. From a regression analysis of 1990 data for a sample of 269 U.S. counties that they showed were highly representative of all counties, they conclude that R-square (the variance explained) was only 0.14. But regressing 1990 imprisonment data for these counties with the 1980 unemployment rates of the same counties yielded an R-square of 0.74. They interpret this finding of a lag in the impact of unemployment on crime rates by pointing to the fact that the highest rates of known offenses, particularly unspecialized street crimes, occurs among teenagers, and in those whose childhood was spent in the most poverty-stricken urban areas, the slums. Their main conclusion is:

Change in violent street crime, in the proportionate size of the young male population, and in labor surplus, contribute to change in the use of imprisonment, while changing levels of property crime do not. These relationships persist even when street-crime rates and other presumed correlates of imprisonment are controlled. . . . The criminal justice system grows in creasingly punitive as labor surplus increases. The fact that our findings were achieved using both a unit of analysis more appropriate theoretically than measures employed by most investigators, and a longitudinal design, only strengthens confidence in them.

Additional topics

Law Library - American Law and Legal InformationCrime and Criminal LawUnemployment and Crime - Dependence On Labor Force Attributes, Other Factors In Crime-unemployment Relationships, Bibliography