Sehiembre de i»w
Desarrollo y s
both property crime -robbery, burglaries and the like- and homicides2*.
What is more, the elasticity of the crime rate with respect to inequality
appears to be substantial. The original estimates by Ehrlich suggested
that, in 1960, a 1 per cent increase in relative poverty, measured by the
number of persons below half the median income, in one state increased
the erime rate by approximately 2 per cent for most property crimes.
Using more recent data, Lee (1993. cited by Freeman, 1996. ρ. 33) found
that when observations for the various states at different times were pooled
together the increase in inequality that took place during the 1980s might
have caused an estimated 10 per cent increase in crime rates. Interestingly
enough, this order of magnitude turns out to be not very different from
Ehrlich’s estimates.
Time series analyses do not seem to lead to such clear conclusions. FYeeman
(1996), still reporting Lee’s results, mentions that changes in crime rates
in US metropolitan areas during the 1980s were not Significantlycorrelated
with changes in inequality, whereas Allen (1996) finds no significant effect
of inequality -and a negative effect of absolute poverty- on the aggregate
crime rate during the last 30 years. The latter also reports insignificant
effects of poverty and inequality in other time-series analyses. Λ possible
explanation of positive results obtained with cross-section data would thus
simply be that there are some States where crime and inequality are both
higher or lower than average because of a third unobserved factor more
or less constant over time. Cross-sectional analysis would thus simply pick
up the effect of these factors and conclude to a positive relationship
between crime and inequality, even though there might not be any causal
relationship between both variables.
An important correction to time scries analysis of crime is proposed by
Freeman (1996) in view of the substantial increase in the number of
incarcerated people observed during the period where most of the increase
in inequality took place. This number doubled between 1980 and 1990
from 0.5 to 1.1 million2''. Ilis point is that if the frequency of crimes had
remained the same among all criminals, this increase in the incarceration
rate should Iogicallyhave produced a drop in the aggregate crime rate. But
no such drop occurred, so that one must conclude either that the frequency
2fi Sec the discussion of this evidence in Krecman {]996, ρ. A3).
24 SeediIulio (1996. p.13).
83